Association between chromosomal aberration of COX8C and tethered spinal cord syndrome: array-based comparative genomic hybridization analysis.

Copy number variations have been found in patients with neural tube abnormalities. In this study, we performed genome-wide screening using high-resolution array-based comparative genomic hybridization in three children with tethered spinal cord syndrome and two healthy parents. Of eight copy number variations, four were non-polymorphic. These non-polymorphic copy number variations were associated with Angelman and Prader-Willi syndromes, and microcephaly. Gene function enrichment analysis revealed that COX8C, a gene associated with metabolic disorders of the nervous system, was located in the copy number variation region of Patient 1. Our results indicate that array-based comparative genomic hybridization can be used to diagnose tethered spinal cord syndrome. Our results may help determine the pathogenesis of tethered spinal cord syndrome and prevent occurrence of this disease.

Introduction

Tethered spinal cord syndrome (TCS) is a neurodevelopmental disorder that results in spinal cord malformation (Payne, 2007; Cearns et al., 2016). TCS is classified as a neural tube defect, and although the incidence of neural tube defects is approximately 1% worldwide (Feuchtbaum et al., 1999; Tunçbilek et al., 1999; van der Put et al., 2001; Khoshnood et al., 2015; Atta et al., 2016), infants born with neural tube defects account for 20–25% of all congenital malformations (Laharwal et al., 2016). The causes of neural tube defects are multivariate, yet to date there is no convincing mechanistic evidence for their occurrence. Some possible contributing factors include gene mutations, chromosomal abnormalities, and environmental factors (Bassuk and Kibar, 2009; Joó, 2009a, b; Molloy et al., 2009; Wen et al., 2009). Recent studies have revealed novel risk factors for neural tube defects including heterozygous missense mutations in the genes, VANGL1 and FUZZY (Bartsch et al., 2012; Seo et al., 2015), as well as maternal folic acid deficiency (Bartsch et al., 2012; Seo et al., 2015). Altered methylation of MGMT, a DNA repair gene, is also associated with neural tube defects (Tran et al., 2012). Moreover, abnormal expression of genes coding for zinc finger proteins is reported to be risk factors (Grinberg and Millen, 2005; Costa-Lima et al., 2008).

Previous studies have shown that chromosomal imbalances due to genomic instability are closely associated with neural developmental disorders (Au et al., 2010; Zhao et al., 2013). Copy number variations (CNVs) are found in patients with neural tube abnormalities in cerebral and spinal sections (Bassuk et al., 2013; Chen et al., 2013). Array-based comparative genomic hybridization (aCGH) is a modern technique for molecular karyotype analysis that combines conventional comparative genomic hybridization and microarray analysis (Saberi et al., 2014). In contrast to conventional hybridization, aCGH does not detect metaphase chromosomes. Instead, it targets genomic DNA to perform high-throughput screening of the whole genome for CNVs (Vissers et al., 2003). The aCGH approach can accurately locate CNVs on chromosomes, and clearly calculate CNV length and identify genes within variant fragments (Mosse et al., 2005). Nowadays, aCGH is commonly used for cancer and genetic disorder research (Kallioniemi, 2008; Sireteanu et al., 2012). In this study, we used aCGH to detect CNVs in three children with TCS and two healthy parents. In order to examine TCS pathogenesis at the chromosome and gene levels, we determined the relationship between these chromosomal aberrations and TCS, and consequently detected CNVs linked with occurrence and development of TCS.

Subjects and Methods

Subjects

Three children diagnosed with typical TCS based on clinical criteria (Filippidis et al., 2010) by the Department of Neurosurgery at the Chinese PLA General Hospital and the Second Artillery General Hospital, and the healthy parents of Patient 1 were enrolled in the study. Peripheral blood samples were collected from the patients and healthy controls. Before initiation of the study, written consent was obtained from the guardians of all children. The study (Project ID: S2013-117-01) was approved by the ethics committee of the Chinese PLA General Hospital, China.

Case 1 was a 2-year-old girl with a sacrococcygeal mass and right foot deformity. The sacrococcygeal mass was identified at birth. Physical examination revealed spina bifida. Strephenopodia of the right foot and a second enlarging sacrococcygeal mass were first observed at 8 months of age. The patient was diagnosed with TCS with myelomeningocele.

Case 2 was a 12-year-old boy who presented with a lumbosacral mass at the age of 8 months. The patient was diagnosed with TCS with spinal cord lipoma. Surgical treatment was performed. Urinary abnormality occurred 11 years after surgery, along with urinary incontinence, nocturnal enuresis, urinary frequency, and urinary urgency. A further surgery was performed because magnetic resonance imaging showed spinal cord lipoma and recurrence of TCS.

Case 3 was a 5-year-old girl with abnormal hair growth in the lumbosacral region at birth. Physical examination revealed a partial spinal canal defect. Because the hair growth increased, magnetic resonance imaging examination was performed. The results revealed a tethered spinal cord and split cord malformation (Type I). Surgery was performed to correct the malformation.

aCGH analysis

aCGH is a specific array-based genomic hybridization method that uses different fluorescent dyes to label DNA from patients and controls, to identify differences between the two groups (Sealfon and Chu, 2011; Brady and Vermeesch, 2012). By comparing the ratio of two different fluorescence signals at each target spot in the microarray, CNVs are detected in specific sequences or genes between two genomes (Gijsbers et al., 2011; Shoukier et al., 2013).

Total DNA was extracted from peripheral whole blood using a commercially available DNA-isolation kit (BioChain Inc., Beijing, China), according to the manufacturer's protocol. For each aCGH experiment, purified DNA and normal sex-matched DNA (1 µg each; Promega, Madison, WI, USA) were digested with AluI and RsaI (10 U each; Promega), and differentially labelled with cyanine-5 and cyanine-3 fluorescent dyes using a Genomic DNA Enzymatic Labeling Kit (Agilent, Santa Clara, CA, USA). aCGH analysis was performed using the Agilent 8 × 60K commercial array. This platform contains 60-mer oligonucleotide probes spanning the entire human genome with an overall mean probe spacing of 50 kb. After hybridization, arrays were scanned using a dual-laser scanner (Agilent), and images extracted and analyzed using the Feature Extraction (Agilent) and Workbench genomics software, respectively. Changes in test DNA copy number at specific loci were considered only if they were < −0.38 (deletion) or > 0.38 (amplification) of the log2 ratio values from at least five consecutive probes.

TCS-related CNV analysis

Removal of polymorphic CNVs using the Database of Genomic Variants

CNV fragments were scanned against the Database of Genomic Variants (Iafrate et al., 2004; Wong et al., 2007). CNVs that completely matched those in the database were removed as they represent common polymorphic variants present in the normal population. Partially overlapping (< 40%) CNVs were considered non-polymorphic and retained for further analysis. In addition, discontinuous polymorphic fragments appearing within CNV sequences (total fragment length was shorter than half-lengths of detected CNVs) were not treated as common polymorphisms and were also retained for further analysis.

Comparison of non-polymorphic CNVs with DECIPHER

The non-polymorphic CNV fragments selected above were searched against the DECIPHER database (Firth et al., 2009). Cases were identified with CNVs similar to those reported in previously tested samples (partial overlap > 60%) or containing documented CNVs. Additionally, chromosomal abnormalities, related phenotypes, and syndromes associated with these cases were identified.

Gene function enrichment analysis

Entire genes incorporated in non-polymorphic CNVs were identified using the University of California, Santa Cruz (UCSC) Genome Browser database (http://genome.ucsc.edu/). Gene function enrichment analyses were performed for the genes identified, including Gene Ontology (GO) (http://geneontology.org/) and Kyoto Encyclopedia of Genes and Genomes (KEGG) Pathway (http://www.genome.jp/kegg/) analyses.

Enrichment P-values for each GO term or KEGG pathway were calculated using the hyper-geometric distribution method. P-values were then corrected for multiple hypotheses testing using the false discovery rate method. A P-value of 0.05 was set as the threshold value for significant gene enrichment for each GO term or KEGG pathway.

Results

Gene micro-repeat fragment location in TCS patients

Results of the aCGH analysis for all three patients and two parents are shown in . Three micro-repeat fragments were detected in DNA isolated from Patient 1. A micro-deletion fragment was detected in Patient 2, while a micro-deletion and micro-repeat were detected in Patient 3. The father of Patient 1 had a normal karyotype, whereas the mother's chromosome map showed micro-deletion and micro-repeat fragments. The micro-deletion fragment in Patient 2 and micro-repeat fragment in Patient 3 were located in the same region: 15q11.1q11.2 ().

Chromosome maps of the three patients with tethered spinal cord syndrome.

(A1–A3) The three CNVs identified in Patient 1 were: 1p21.2, 9p13.3, and 14q32.12. (B) The CNV identified in Patient 2 was: 15q11.1q11.2. (C1, C2) The two CNVs identified in Patient 3 were: 2p11.2 and 15q11.1q11.2. Blue squares in the figure indicate the location of chromosome micro-variations. CNVs: Copy number variations.

Database searching of CNVs

The eight identified CNVs were searched against the Database of Genomic Variants. The results showed that four CNVs were normal chromosomal polymorphisms, specifically, the 1p21.2 micro-repeat in Patient 1, 2p11.2 micro-deletion in Patient 3, and 7q11.22q11.23 micro-deletion and 19p12 micro-repeat in the mother of Patient 1.

Investigation of the other four non-polymorphic CNVs in DECIPHER revealed eight specific CNVs in these regions (). Non-polymorphic CNVs in Patients 2 and 3 (ID 4 and 6 in ) shared the same chromosomal initiation site, indicating that multiple CNVs occur in the same location. Further analyses revealed that these CNVs are associated with two syndromes (Angelman and Prader-Willi) and one phenotype (microcephaly) ().

Gene function enrichment analysis

Within the four non-polymorphic CNVs regions, 13 genes were identified by the UCSC Genome Browser (). Function enrichment analysis of GO terms and KEGG pathways were performed for these genes. The results included a number of biological functions (e.g., gamete generation), molecular functions (e.g., ubiquitin-protein ligase activity), two cellular components (mitochondrial inner membrane and integral membrane component), as well as eight KEGG pathways, including viral myocarditis, cardiac muscle contraction, Parkinson's disease, oxidative phosphorylation, ubiquitin-mediated proteolysis, Alzheimer's disease, Huntington's disease, and olfactory transduction. From these results, we found that the COX8C gene is closely related to neural system diseases such as Parkinson's disease, Alzheimer's disease, and Huntington's disease (Tables ).

Discussion

Advantages of using aCGH for detection of rare chromosomal micro-variations

Chromosomal sub-microscopic variations are strongly associated with human disease (Feuk et al., 2006). In clinical settings, the definite diagnosis of several diseases cannot be achieved using existing techniques. Consequently, some rare syndromes are labelled idiopathic or unexplained. Most of these syndromes are due to genomic imbalances created by chromosomal micro-variations such as micro-deletions and micro-repeats (D’Angelo et al., 2014). The aCGH approach efficiently detects chromosomal micro-aberrations and aids elucidation of idiopathic or unexplained diseases.

Significance and limitations of aCGH analysis

The main objective of this study was to identify non-random CNVs and evaluate their association with TCS. The main questions regarding the CNVs we identified are: (1) whether the CNVs are inherited; (2) whether they are found in the normal population; (3) whether their lengths are sufficient to contain genes with functional annotations; (4) whether they are linked to diseases in DECIPHER; and (5) whether any are unreported, unidentified, or novel. Although the Database of Genomic Variants and DECIPHER, which are globally representative databases, were used to determine the type of CNVs identified, ethnic differences are inevitable when using international databases.

Diseases similar to TCS that are associated with COX8C

CNVs similar to the ones we detected are found in the DECIPHER database. These CNVs are associated with Angelman and Prader-Willi syndromes, and microcephaly. All of these disorders involve significant neural abnormalities (Mabb et al., 2011; Mahmood et al., 2011; Cassidy et al., 2012). Furthermore, gene function analysis indicated a close association between COX8C and certain diseases including Parkinson's, Alzheimer's, and Huntington's diseases, all of which are typical nervous system diseases (Bassil and Mollaei, 2012; Pogledić and Relja, 2012; Gazewood et al., 2013). By comparing the CNVs from Patient 1 with those identified in her parents, we excluded the possibility of TCS being hereditary. Thus, we propose that the condition may be acquired during neural development.

Conclusion

In this study, we used high-resolution aCGH to identify pathogenic CNVs in samples from patients with typical TCS. Our findings suggest an association between certain CNVs and nervous system disease. Our data may be used in the future as a reference for the integration of available data, or for further studies with larger sample sizes. Ours study demonstrates specific transformation research, and shows that a molecular method can be used to clinically diagnose TCS. Our findings may help to shed new light on the pathogenesis of TCS.

Table 1

Array-comparative genome hybridization analysis of TCS patients and controls

Table 2

DECIPHER search results for non-polymorphic copy number variations (CNVs)

Table 3

Syndromes and clinical phenotypes linked to non-polymorphic copy number variations

Table 4

Genes contained in non-polymorphic copy number variations

Table 5

Enrichment results for gene ontology (GO) analysis

Table 6

Gene enrichment analysis