Introduction
The distal region of chromosome 8p is prone to various genomic rearrangements, mainly due to the presence of low copy repeats (LCRs). The LCRs flank an approximately 5 Mb region on 8p23.1 and contain two olfactory receptors gene clusters or defensin repeats, repeat-distal (REPD) in the distal 8p23.1 region and repeat-proximal (REPP) in the proximal 8p23.1 region (1,2). These LCRs are believed to mediate a U-type exchange, commonly resulting in a recurrent interstitial 8p23 deletion, a terminal 8p23 deletion and a variably sized interstitial inverted duplication in 8p with an 8p terminal deletion (1,3–6). The distal deletions of 8p23, including either the terminal or the sub-terminal 8p region, appear as an isolated finding or as a part of a complex chromosomal rearrangement. Both interstitial or terminal 8p23 deletions are relatively frequent among the structural variations associated with 8p, and patients with a deletion in this region are generally characterized by developmental delay (DD)/intellectual disability (ID), mildly dysmorphic features, microcephaly, congenital heart defects (CHD), congenital diaphragmatic hernia (CDH), seizure, and neuropsychiatric disorders (4,5,7,8). The deleted region often spans a large genomic segment containing numerous genes. Therefore, the genotype-phenotype correlations associated with the deletion are largely unknown. However, an increasing number of studies have contributed to narrowing the critical region (CR) to search for the candidate genes associated with the clinical features. A deletion at 8p23.1 was proposed as a CR for CHD, CDH and other common clinical features (4,5,8–10). Although genes in the deleted region may interact or participate in the same pathways, specific candidate genes are expected to play important roles in certain phenotypes. Notably, haploinsufficiency of GATA4, NEIL2 and/or SOX7 located at 8p23.1 were shown to be associated with CHD and CDH in several studies (8,11–13). Another gene in the region, MCPH1, has also been implicated as a candidate gene for primary microcephaly (14–16).
In addition to the common 8p23.1 deletion, a terminal deletion more distal to the REPD deletion involving 8p23.2 to 8pter has increasingly been reported as a CR associated with DD/ID, seizure and neurobehavioral disorders, such as autism spectrum disorder (ASD) and attention-deficit/hyperactivity disorders (ADHD) (7,17–19). Isolated 8p23.2-pter deletions are a rare chromosomal aberration. The recurrence of an isolated distal deletion involving the region from 8p23.2 to 8pter in patients with similar phenotypes is consistent with a distinctive microdeletion syndrome. To the best of our knowledge, cytogenetic and molecular characteristics have only been reported for four patients (7,17,19). The locations and sizes of the deletions detected in this region vary, similar to most of the sub-telomeric microdeletion syndromes. The majority of the examined patients with an 8p23 deletion were characterized using traditional techniques, including G-banded karyotyping, fluorescence in situ hybridization (FISH), microsatellite marker genotyping and multiplex ligation-dependent probe amplification (MLPA), with limited resolution and low efficiency. Therefore, an accurate definition of the locations, sizes and genes contained in the deletions has been difficult to obtain (5,9,20–23). Nevertheless, few microarray-based molecular studies have been performed to investigate these deletions. Current high-resolution chromosomal microarray analyses (CMA), including array comparative genome hybridization and single-nucleotide polymorphism (SNP) arrays, can now accurately determine the locations, sizes and genes contained in 8p23 deletions. This technique allows researchers to obtain more precise genotype-phenotype correlations and aids in identifying CRs and candidate genes. Here, we present a new case study of a patient with an isolated 8p23.2-pter deletion that was determined using a SNP array and compare the clinical features with patients with isolated deletions involving the region from 8p23.2 to 8pter that were previously identified using CMA or recorded in the Database of Chromosomal Imbalance and Phenotype in Humans using Ensembl Resources (DECIPHER; https://decipher.sanger.ac.uk/). The accumulation of patients with isolated 8p23.2-pter deletions at various locations and of different sizes will help in further defining the CRs and candidate genes for certain phenotypes.
Results
Routine G-banded karyotyping showed normal karyotypes in the patient (46, XY, Fig. 2C) and his parents (46, XX and 46, XY, respectively). The SNP array revealed an 8p23.2-pter terminal deletion in the patient, represented as arr[hg19] 8p23.3p23.2(158, 048 - 6, 004, 205)x1 (Fig. 2A and B). The deleted region was ~6.0 Mb in size and contained 23 RefSeq genes, including 11 protein-coding genes and 12 non-coding genes; 6 of the protein-coding genes were annotated in OMIM (FBXO25, DLGAP2, CLN8, ARHGEF10, MYOM2, and CSMD1) (Fig. 3A). Other copy number variants related to pathogenicity or of unknown significance were not identified. Furthermore, the two-colour FISH analysis of the patient's metaphase chromosomes using specific probes identified a terminal deletion of chromosome 8p (Fig. 2D), confirming the 8p terminal deletion identified by SNP array. However, FISH studies of the parents were normal, excluding cryptic rearrangements of the 8p terminal region in the parents and indicating the de novo deletion of 8p23.2-pter in the patient.
To date, only four patients with isolated deletions involving the region from 8p23.2 to 8pter that were identified using CMA have been reported in previous studies; all of them shared common clinical features. In addition, three patients were included in DECIPHER harbored deletions from 8p23.2 to 8pter that were recognized as definitely or likely pathogenic (Table I). The locations, sizes and genes contained in the deleted region in these patients were compared based on the genetic map (Fig. 3). In addition, we also performed a phenotypic comparison between the 8p23.2-pter deletion and 8p23.1 deletion (Table II).
Discussion
Deletions of varying sizes at the distal region of the 8p chromosome are frequently reported, particularly the 8p23 deletion, but most have not been investigated using molecular cytogenetics. With the increasing use of high-resolution microarray technologies, a better understanding of the genotype-phenotype correlations of the distal 8p deletion have been obtained with the identification of CRs or candidate genes. Among distal 8p deletions, the 8p23.1 deletion is frequently observed, mainly due to the presence of LCRs. The recurrence of CHD and CDH in patients with various 8p23.1 deletions has led to the identification of a CR at 8p23.1 for CHD and CDH (4,5,8–10). Furthermore, the GATA4, NEIL2 and SOX7 genes in the deleted region have been recognized as candidate genes for CHD and CDH (8,11–13). In addition, a more distal deletion to 8p23.1 is the 8p23.2-pter deletion, a relatively rare aberration. Patients with an 8p23.2-pter deletion usually do not present CHD, CDH and urogenital abnormalities, but instead appear to have milder phenotypes than patients with an 8p23.1 deletion (Table II). The diversity in phenotypes suggests that the 8p23.2-pter deletion does not encompass the CR or candidate genes (such as the GATA4, NEIL2 and SOX7 genes) on 8p23.1 associated with these clinical features. Moreover, based on these data, the 8p23.1 region plays an important role in the manifestation of these clinical features.
However, common clinical features, including DD/ID, mildly dysmorphic features, microcephaly and neurobehavioral disorders, are usually shared by patients with either an 8p23.2-pter deletion or an 8p23.1 deletion (not only individuals with a large terminal 8p23.1 deletion but also in individuals with small interstitial 8p23.1 deletions) (Table II). Either an isolated 8p23.1 deletion or an isolated 8p23.2-pter deletion seems to contribute to these common clinical features, suggesting that the isolated 8p23.2-pter region may harbour other CRs and candidate genes associated with these features. The present study compared the phenotypic features between the present patient and four previously reported patients with isolated terminal deletions involving the region from 8p23.2 to 8pter that were characterized by CMA (7,17,19). Of these five patients, the most common phenotypes included growth retardation (5/5), microcephaly (4/5) and mildly dysmorphic features (4/5), followed by ID (3/5), language delay (3/5) and behavioural problem (3/5, including ASD and ADHD). In the DECIPHER database, three patients with isolated deletions involving the 8p23.2 to 8pter region were classified as definitely or likely pathogenic and all of them presented with ID. Taken together, although the phenotypes listed in DECIPHER might be incomplete and limited, ID was present in 6 out of the 8 patients with a deletion involving the region from 8p23.2 to 8pter reported in the literature and DECIPHER. A terminal 2.1 Mb deletion of 8p23.3 reported by Wu et al (19) was the smallest segment observed in the five patients from the present study and the literature. This deletion might be proposed as a CR for DD/ID, dysmorphic features and microcephaly (Fig. 2). The CR contains 5 OMIM genes, including FBXO25, DLGAP2, CLN8, ARHGEF10 and MYOM2. Most of these genes are known to be expressed in the brain.
The product encoded by FBXO25 is a member of the F-box protein family and is involved in phosphorylation-dependent ubiquitination. Fbxo25 may act as a ubiquitin E3 ligase to degrade cardiac transcription factors and control the transcriptional activity of the cardiac transcription factors Tbx5 and Nkx2-5, which play key roles in human cardiac development (24). The protein encoded by DLGAP2 is a membrane-associated protein that is biallelically expressed in the brain and plays roles in synapse organization and signalling in neuronal cells. It has been implicated as a candidate gene in neurobehavioral disorders, such as ASD and schizophrenia (25). Li et al (26) identified 5 rare, isolated variants and a DLGAP2 haplotype (CCACCAACT) in a cohort of patients with schizophrenia and showed that increased expression of the DLGAP2 gene is associated with schizophrenia. As shown in a study by Chien et al (17), an individual with DD/ID, ASD and seizures carried an 8p23.2-pter deletion. The DLGAP2 gene located in this region was suggested to be a candidate gene for ASD. Subsequently, the authors performed deep exon resequencing of DLGAP2 and detected two SNPs and some rare missense DLGAP2 mutations that might contribute to ASD (27). In a large autism case-control study, a de novo 817 kb duplication that partially overlapped with DLGAP2 was detected in a child with non-syndromic autism, and the authors also proposed DLGAP2 as a candidate gene involved in the pathogenesis of ASD (28). Notably, two SNPs in DLGAP2 (rs6558484 and rs7014992) were also revealed to be associated with alterations in brain volume (white matter volume of the prefrontal cortex) (29). CLN8 is recognized to have roles in lipid synthesis, transport, and sensing, thus contributing to cell proliferation during neuronal differentiation and defence against cell death stimuli (30). Mutations in this gene alter the levels of sphingolipids and phospholipids in the brain (30,31). A homozygous CLN8 mutation is one of the causes of autosomal recessive neuronal ceroid lipofuscinoses, which are mainly associated with progressive seizures and DD/ID (30). Allen et al (32) reported a patient with seizures, microcephaly, DD and behavioural problems, which were thought to result from the co-existence of a heterozygous CLN8 mutation and a de novo 8p23.3 terminal deletion. Notably, the 8p23.3 deletion was only identified by a FISH analysis using a probe (D8S504) located on the telomeric region of the chromosome (8:1017556-1017754) that did not cover the CLN8 gene. Therefore, the authors did not clearly determine whether this patient truly had a CLN8 gene deletion. The actual size and genes contained in the 8p23.3 deletion should be further investigated to obtain a better understanding of the genotype-phenotype correlation in this patient. ARHGEF10 encodes a Rho guanine nucleotide exchange factor and may play a role in neural morphogenesis (33). Mutations in the ARHGEF10 gene may cause demyelination and re-myelination of the peripheral nervous system and are associated with a slower nerve conduction velocity through an autosomal dominant pattern (33). Some SNPs in ARHGEF10 might confer a risk of developing schizophrenia (34). Notably, in DECIPHER, the 8p23.3-p23.2 and 8p23.3 deletions presented by two patients (nos. 295087 and 338097, respectively) with ID covered DLGAP2, CLN8 and ARHGEF10. Among all the patients carrying these genes (literature and DECIPHER), the deleted segment in patient no. 338097 was the smallest. Further studies of patients with similar phenotypes are required to determine whether the deleted region in patient no. 338097 is a potential CR for ID (only this phenotype was presented in DECIPHER). Although limited clinical features are listed in DECIPHER, the two patients also provided evidence for the potential role for either the deleted region or these genes in the pathogenesis of ID.
Furthermore, among the five patients with an 8p23.2-pter deletion from the present study and the literature, only the deletions reported by both Chien et al (17) and Wu et al (19) did not contain the CSMD1 gene. CSMD1 encodes a large multi-domain protein that mainly consists of CUB and Sushi domains. This protein may function as an important regulator of complement activation and inflammation in the developing central nervous system and is involved in brain mechanisms associated with memory and learning (35,36). Based on the evidence from previous studies, a SNP (rs10503253) located within the CSMD1 gene may be associated with the risk of multiple neurodevelopmental phenotypes, such as schizophrenia, seizure, speech delay, cognitive impairment and learning difficulties (37–39). Regarding cognitive ability, the ‘A’ allele of rs10503253 is considered to have deleterious impacts on verbal or performance intelligence, verbal episodic memory, and spatial working memory (35). In DECIPHER, patient no. 280996 had a maternally inherited 134 kb deletion that partially overlapped with CSMD1, and the deletion was recognized as a likely pathogenic variant. Both the patient and his mother showed moderate ID. Thus, CSMD1 might be a potential candidate gene for ID.
Microcephaly is caused by a congenital insufficiency during foetal brain (mainly cerebral cortex) development and is frequently observed in patients with either an 8p23.2-pter deletion or an 8p23.1 deletion (7,16,19). A candidate gene for microcephaly, the MCPH1 gene (16,40), is located on 8p23.1, but no candidate genes have been proposed for the region from 8p23.2 to 8pter. However, the DLGAP2, CLN8, ARHGEF10 and CSMD1 genes mentioned above are all expressed in the brain and implicated as being associated with multiple neurodevelopmental phenotypes. These genes should not be excluded as possibly having a role in microcephaly, particularly DLGAP2, which has been shown to be associated with alterations in brain volume alterations (29).
However, haploinsufficiency of each gene alone has not proven sufficient to explain the clinical phenotypes. These genes might function together, forming an interaction network in some pathway or may be modified by other genetic or environmental factors (27). Thus, the deletion of all these genes might conceivably have adverse effects on neurodevelopment. Compound inheritance of a rare null mutation and a common risk haplotype should also be considered in studies aiming to understand genotype-phenotype correlations. According to the study by Wu et al (41), compound inheritance of a heterozygous deletion and a common risk haplotype of the TBX6 allele are responsible for congenital scoliosis. Therefore, some genes in the deleted region, including DLGAP2, CLN8, ARHGEF10 and CSMD1, might contribute to DD/ID, microcephaly and neurobehavioral disorders, although the pathogenic mechanisms remain unclear.
In summary, the current study reported a new case study of a patient with a rarely isolated 8p23.2-pter deletion, and performed a review of patients with deletions in the region from 8p23.2 to 8pter that have been reported in the literature and the DECIPHER database. Similar clinical characteristics were observed among patients with the deletions, which were consistent with a distinctive microdeletion syndrome. Although a refined CR for DD/ID, microcephaly and neurobehavioral disorders was not conclusively determined due to the currently limited number of patients, the deleted region reported by Wu et al (19) as being responsible for these features is regarded as the smallest deletion among the five deletions reviewed in the present study. Furthermore, after a detailed review of the potential correlations between the genes located in 8p23.2 to 8pter region and their clinical significance, we hypothesized that DLGAP2, CLN8, ARHGEF10 and CSMD1 might be plausible candidate genes for DD/ID, microcephaly and neurobehavioral disorders. However, further high-resolution studies of patients with small isolated, overlapping and interstitial deletions involving the region from 8p23.2 to 8pter may allow us to obtain genotype-phenotype correlations for specific genes crucial to 8p23.2-pter.