Identification and characterization of a novel 43-bp deletion mutation of the ATP7B gene in a Chinese patient with Wilson’s disease: a case report

Wilson’s disease (WD, OMIM #277900), an autosomal recessive disorder characterized by abnormal copper accumulation and related toxicities, is caused by mutations in the ATP7B gene (OMIM *606882) [1]. The ATP7B gene is located on 13q14.3 spanning ~ 100 kb, including 21 exons and 20 introns; it encodes a transmembrane copper-transporting P-type ATPase of 1465 amino acids, which plays a crucial role in maintaining body copper homeostasis and is involved in copper transport into the plasma by ceruloplasmin as well as copper excretion from the liver [2‐4]. ATP7B gene mutations lead to ATP7B protein dysfunction, which in turn causes accumulation of copper in the liver, brain, kidneys and corneas, with a wide range of clinical symptoms, including hepatic disorders, neuronal degeneration of the brain, and Kayser-Fleischer rings at the corneal limbus [5, 6]. The prevalence of WD is estimated at 1 in 30,000, and the heterozygous carrier rate approximates 1 in 90 in many populations [7, 8]. Without timely diagnosis and treatment, WD can be fatal [9]. Early detection and intervention is critical in preventing disease progression and irreversible sequelae. Genetic testing for the detection of biallelic ATP7B mutations do not only help diagnose WD, but can be also used for prenatal screening. About 540 disease-causing variants have been reported in ATP7B according to the HUGO database, and are mostly point mutations and small insertions or deletions distributed in all exons/introns (http://​www.​wilsondisease.​med.​ualberta.​ca/​database.​asp). Screening of all exons and flanking intronic regions for ATP7B mutations is usually performed by direct sequencing, e.g. Sanger sequencing and next-generation high-throughput sequencing. However, mutations are identified in only one allele or none in a substantial number of WD patients [10]. A possible reason could be that partial gene deletions of one or more exons are not detected by current methods, for example c.4021 + 87_4125-2del2144, c.51 + 384_1708-953del8798, and c.52-2671_368del3039 of the ATP7B gene [11‐13]. For unexplained WD cases, in whom no or one allele is mutated, detecting partial or whole gene deletions/duplications by MLPA assay, QPCR, and other methods is required. Here we report a novel 43-bp deletion by direct Sanger sequencing of all exons in the ATP7B gene, which is by far the biggest exonic deletion within a single exon of the ATP7B gene detected in a WD patient. Meanwhile, we comprehensively analyzed the possible mechanisms underlying the formation of this deletion mutation, and a replication-based mechanism named fork stalling and template switching/microhomology-mediated break-induced replication (FoSTeS/MMBIR) seems to explain it clearly [14‐16].

The proband was the first child of healthy non-consanguineous Chinese parents from Jiangsu Province. He was born at 39 weeks of gestation, and the perinatal period was normal. He was referred to our hospital at the age of 10, for paroxysmal headache of an approximate one-month duration. Paroxysmal forehead pain in this patient consisted of daily episodes at different times, even up to dozens of times, automatically disappearing after 1–2 min. Before the onset of these symptoms, his developmental milestones were normal. Physical examination was normal. Neurological examination revealed clear consciousness, normal language use, normal muscle tone and tendon reflexes of the limbs, and normal gait. However, he presented with rapid alternating movements of slightly clumsy hands and slightly unstable finger-to-nose movements. Blood routine examination for red blood cell count, white blood cell count, platelet count, and hemoglobin was normal. Serum total bilirubin, direct bilirubin, total bile acids albumin and plasma ammonia levels were in control ranges, as well as serum alanine transaminase and aspartate aminotransferase. Normal serum total cholesterol, triglyceride, plasma glucose and homocysteine levels were observed. Serum creatinine and blood urea nitrogen were normal, as well as serum electrolytes, including potassium, sodium and chloride. Serum ceruloplasmin and copper levels were 20.6 mg/L (normal 200~ 600 mg/L) and 1.53 μmol/L (normal 11–22 μmol/L). Kayser–Fleischer rings in both corneas were observed under slit-lamp. Abdominal ultrasonography showed mild spleen enlargement and slightly increased reflectiveness of the liver and both kidneys, but no hepatomegaly. Cranial magnetic resonance imaging (MRI) showed a decreased signal intensity in T1WI and an increased signal intensity in T2WI in bilateral caudate nucleus. The diagnosis of WD was clinically made based on low serum ceruloplasmin levels and Kayser–Fleischer rings, with the Wilson’s disease scoring system applied as WD diagnostic criteria [17]. Improvement was observed after 4 weeks of treatment with D-penicillamine in another hospital. To further confirm the diagnosis of WD, DNA sequence analysis of the ATP7B gene and bioinformatics analysis were performed by the methods in Additional file 1. The CARE guidelines were followed in reporting this case.

Sanger direct sequencing of the 21 exons of the ATP7B gene revealed two heterozygous mutations in the proband: one missense variant, c.3517G > A (p.Glu1173Lys) and one deletion variant of 43 base pairs, c.532_574del. His mother was heterozygous for the c.3517G > A mutation, while the father was heterozygous for c.532_574del (Fig. 1). Accordingly, the proband was compound heterozygous for c.3517G > A and c.532_574del mutations in both alleles inherited from his mother and father, respectively. The missense mutation p.Glu1173Lys has been reported three times in other cases (Table 1) [18‐20]. The c.532_574del mutation was predicted to cause a premature tarnation codon (p.Leu178PhefsX10) in the N-terminal region. According to previous reports and the WD Mutation Database, the c.532_574del mutation in our WD patient was unknown so far. Owing to definite genetic cause of WD in our patient, the parents was offered prenatal diagnosis during their second pregnancy. The fetus inherited the wide-type maternal allele and the deletion-bearing paternal allele. The parents were counseled that the fetus was predicted to be an unaffected carrier of WD.

The 193 nucleotide sequence including the 43 nucleotide deletion fragment between two breakpoints plus 75 nucleotide fragments surrounding the two breakpoints was used for extensive bioinformatics analysis, assessing the involvement of local genomic architecture in the predisposition to DNA breakage, e.g. repetitive elements, sequence motifs, and non-B DNA conformations, which may initiate the formation of deletions [21]. Known repeats were not found at both breakpoints of the 193-nucleotide sequence by the CENSOR and RepeatMasker tools. Seven sequence motifs previously associated with DNA breakage, including 2 “Deletion hotspot consensus”, 2 “DNA polymerase arrest site”, 1 “Ig heavy chain class switch repeat 1”, 1 “Ig heavy chain class switch repeat 2” and 1 “Vaccinia topoisomerase I consensus”, were observed in the 193-nucleotide sequence using Fuzznuc. No motifs forming left-handed Z-DNA were predicted by nBMST. Direct, inverted, and mirror repeats capable of forming slipped hairpin, cruciform, and triplex structures, respectively, were not found in the above sequence by RepeatAround. The QGRS software revealed oligo(G)n tract forming tetraplex structures in the above nucleotide sequence.

The deletion mutation c.532_574del (p.Leu178PhefsX10) in the ATP7B gene was reported for the first time in present study, and was observed in proband’s parent, was therefore not a de novo mutation. The 43-bp deletion occurring in exon 2 of the ATP7B gene caused a frameshift mutation to yield a truncated ATP7B protein, with 1279 amino-acid residues deleted from the C-terminal segment compared with the mature protein consisting of 1465 amino-acid residues. The ATP7B protein includes five important functional areas that contribute to copper transport, including N-terminal metal-binding, transmembrane, adenosine triphosphate (ATP)-binding, phosphatase, and phosphorylation domains [22]. The deletion mutation removed almost completely these five functional domains (Fig. 1). On the other hand, it resulted in almost 87% amino-acid residues missing from the entire mature protein. The variant would be considered as a pathogenic variant with more severe phenotype than the other missense pathogenic variants. Indeed, the phenotype of the proband with p.Leu178PhefsX10/p.Glu1173Lys compound heterozygote mutations in this study was likely more severe than two reported WD patients whose genotypes were separately p.Pro992Leu/p.Glu1173Lys and p.Gln1200Ter/p.Glu1173Lys (Table 1) [18‐20].

After analysis of the c.532_574del mutation in the ATP7B gene, two microhomology (DNA direct repeat sequences, TCTCA) were observed on both deletion breakpoints. Based on the short microhomology stretches at breakpoint junctions, a replication-based mechanism named fork stalling and template switching/ microhomology-mediated break-induced replication (FoSTeS/MMBIR) or a non-replicative repair mechanism named microhomology-mediated end-joining (MMEJ) seemed to be able to explain the formation of small deletion mutations [16, 23]. However, MMEJ associated deletions usually have imperfect or interrupted microhomologies at the breakpoints; for instance, distinct 8~ 22 bp pair imperfect microhomologies are frequently observed in the MMEJ events of S. cerevisiae and Drosophila [10, 24‐26]. Meanwhile, MMEJ occasionally generates small scars or causes insertion of nucleotides at the junction site of the deletion in mammalian cells [27, 28]. Indeed, current reports suggest that microhomology is not necessarily essential for MMEJ [28, 29]. MMEJ is currently used to explain the possible formation of large deletions belonging to structural variations (SVs), commonly referred to as copy number variants (CNVs) which are generally defined as DNA regions of approximately 50 bp and larger in size, and not small deletions belonging to indels which are small insertions or deletions generally between 1 and 50 bp in size [25]. The occurrence of FoSTeS/MMBIR events is stimulated by DNA strand breaks in the DNA replication phase. The DNA motifs around the deletion breakpoints would likely cause DNA strand breaks. For example, oligo(G)n tracts have the potential to adopt stable G-quadruplex structures prone to cause genomic alterations or replication fork stalling. Accordingly, the FoSTeS/MMBIR mechanism seemed to better explain the origin of the novel deletion mutation c.532_574del of the ATP7B gene in this study.

Several mechanisms of ATP7B gene deletions have previously been reported. Alu-mediated MMEJ is a possible mechanism underlying the formation of the 3827 bp deletion in the ATP7B gene (c.3134_3556 + 689del) [30]. Meanwhile, Alu-mediated non-allelic homologous recombination (NAHR) seems to favor the formation of the 2776 bp deletion in the ATP7B gene (c.52-2461_366del2776) [13]. Therefore, we here report the first case with c.532_574del of the ATP7B gene mediated by FoSTeS/MMBIR. Since exactly 1-bp microhomology at the junctions can cause a FoSTeS/MMBIR event, and up to four FoSTeS/MMBIR events occur in the DNA replication phase [14‐16], all reported deletion mutations of the ATP7B gene were reviewed in Additional file 2: Table S1, and the formation of 65.5% (91/139) small deletions as well as 100% (4/4) gross deletions could be explained by single template switch event in MMBIR using the FoSTeS/MMBIR mechanism; the remaining small deletions could result from two FoSTeS/MMBIR events. Although the FoSTeS/MMBIR mechanism still is imperfect, it provides a perspective to explore the occurrence of deletion mutations in the ATP7B gene.

Published studies indicated that FoSTeS/MMBIR is a mitotic event rather than a meiotic one, e.g. somatic mutations associated with the PMP22 and FMR1 genes [16, 31]. Therefore, c.532_574del in the ATP7B gene could have occurred through a single FoSTeS/MMBIR event during the mitotic cell division of somatic cells in grandparents or ancestors of the proband (Fig. 2).