Original Paper Received: January 6, 2015 Accepted after revision: May 7, 2015 Published online: July 10, 2015
Fetal Diagn Ther DOI: 10.1159/000431320
Prenatal Diagnosis of DNA Copy Number Variations by Genomic Single-Nucleotide Polymorphism Array in Fetuses with Congenital Heart Defects Shaohua Tang a, b Jiaojiao Lv a Xiangnan Chen a Lili Bai b Huanzheng Li b Chong Chen b Ping Wang a Xueqin Xu b Jianxin Lu a
b
Key Laboratory of Medical Genetics, School of Laboratory Medicine and Life Science, Wenzhou Medical University, and Key Laboratory of Birth Defects, Department of Genetics, Wenzhou Central Hospital, Wenzhou, China
Key Words Prenatal diagnosis · Congenital heart disease · Genetic counseling
Abstract Objectives: To evaluate the usefulness of single-nucleotide polymorphism (SNP) array for prenatal genetic diagnosis of congenital heart defect (CHD), we used this approach to detect clinically significant copy number variants (CNVs) in fetuses with CHDs. Methods: A HumanCytoSNP-12 array was used to detect genomic samples obtained from 39 fetuses that exhibited cardiovascular abnormalities on ultrasound and had a normal karyotype. The relationship between CNVs and CHDs was identified by using genotype-phenotype comparisons and searching of chromosomal databases. All clinically significant CNVs were confirmed by real-time PCR. Results: CNVs were detected in 38/39 (97.4%) fetuses: variants of unknown significance were detected in 2/39 (5.1%), and clinically significant CNVs were identified in 7/39 (17.9%). In 3 of the 7 fetuses with clinically significant CNVs, 3 rare and previously undescribed CNVs were detected, and these CNVs encompassed the CHD candidate genes FLNA (Xq28 dup), BCOR (Xp11.4 dup), and RBL2 (16q12.2 del). Conclusion: Compared with conventional cytogenetic genomics, SNP array analysis provides significantly improved detection
© 2015 S. Karger AG, Basel 1015–3837/15/0000–0000$39.50/0 E-Mail [email protected] www.karger.com/fdt
of submicroscopic genomic aberrations in pregnancies with CHDs. Based on these results, we propose that genomic SNP array is an effective method which could be used in the prenatal diagnostic test to assist genetic counseling for pregnancies with CHDs. © 2015 S. Karger AG, Basel
Introduction
Congenital heart defects (CHDs) are a major cause of mortality in the perinatal period and affect an estimated 10% of spontaneous miscarriages [1]. Furthermore, CHDs are responsible for more deaths in the 1st year of life than any other etiology [2]. Recently, the Ministry of Health of China reported that CHDs account for 26.7% of birth defects that occur in China [3]. A complex set of environmental and genetic factors are potentially etiologic in the pathogenesis of CHDs. Correspondingly, more than 1,700 genes have been estimated to play important roles in cardiac development in mice (http://www.informatics.jax.org/). The homologs of these genes may be involved in cardiac development in humans [4], and 55 human genes have been identified to have roles in CHDs [5]. It is possible that variations in critical genes lead to severe birth defects which are lethal to a fetus, whereas the CHDs Jianxin Lu Key Laboratory of Medical Genetics School of Laboratory Medicine and Life Science, Wenzhou Medical University Ouhai District, Wenzhou, Zhejiang 325000 (China) E-Mail jxlu313 @ 163.com
Downloaded by: Univ. of California Santa Barbara 198.143.33.34 - 7/10/2015 11:35:20 AM
a
Methods Samples This prospective study was conducted from January 2011 to February 2014. A total of 364 pregnant women with a gestational age ranging from 18 to 32 weeks underwent invasive testing. A genomic SNP array analysis was subsequently performed at the Center of Wenzhou Prenatal Diagnostics. There were 45 fetuses with cardiovascular abnormalities that were detected by ultrasound and confirmed by a pediatric cardiologist. In some cases, other ultrasound anomalies were detected as well. An abnormal karyotype in a conventional G-band karyotype analysis led to the exclusion of 6 of these 45 fetuses. As a result, 39 fetuses were enrolled in this study, including 18 fetuses with isolated CHD and 21 fetuses with CHD plus other ultrasound anomalies. Both ultrasound findings and karyotype were supported at the Wenzhou Central Hospital (Wenzhou, Zhejiang, China). This study was approved by the Dingli Clinical School Ethics Committee of Wenzhou Medical University, and informed consent was obtained from each prenatal couple. DNA Extraction Genomic DNA was extracted from umbilical cord blood samples (n = 33) and amniotic fluid samples (n = 6) using Qiagen Blood Mini Kits (Qiagen, USA) and Biochain DNA Extraction Kits
2
Fetal Diagn Ther DOI: 10.1159/000431320
(Tissue and cells; Biochain, USA), each according to the manufacturers’ instructions. When it was necessary to exclude maternal contamination and to assist in the interpretation of the CNVs, parental peripheral blood samples were collected and genomic DNA was extracted using Qiagen Blood Mini Kits. SNP Array Platform Genomic DNA samples were subjected to an SNP array analysis using the HumanCytoSNP-12 array v1.0 (Illumina Inc., USA). 298,563 overlapping SNP probes provide a useful content for cytogenetic analysis. This includes a dense coverage of around 250 genomic regions commonly screened in cytogenetics laboratories, including subtelomeric regions, pericentromeric regions, sex chromosomes, and targeted coverage in around 400 additional disease-related genes. The average resolution of this array is 31 kb. DNA digestion, ligation, PCR amplification, fragmentation, labeling, and hybridization for each sample were performed according to the manufacturer’s protocols. The array slide was scanned using an Iscan Reader (Illumina). CNV Identification and Assessment To analyze the generated data, the GenomeStudio v2011 software (human genome build 37/Hg19 for analysis) and Illumina cnvPartition (v3.2.0) were used, respectively. All detected gain-ofcopy number and loss-of-copy number CNVs were compared with CNVs listed in our in-house database and the following publically available databases: Database of Genomic Variants (DGV, http:// dgv.tcag.ca/dgv/app/home), International Standards for Cytogenomic Arrays (ISCA) consortium (https://www.iscaconsortium. org/), UCSC Genome Bioinformatics (http://genome.ucsc.edu/), DECIPHER Database (https://decipher.sanger.ac.uk/index), CHDWiki (http://homes.esat.kuleuven.be/∼bioiuser/chdwiki/index.php/Main_Page), Online Mendelian Inheritance in Man (OMIM; http://www.omim.org/), and PubMed (http://www.ncbi. nlm.nih.gov/pubmed). If a CNV was described in a CNV database for healthy individuals (DGV), or in our local normal CNV database, it was considered to be benign. If a CNV was described in the DECIPHER pathogenic chromosomal database, or had been reported to be a pathogenic CNV in the literature, it was considered pathogenic. CNVs identified as morbid OMIM genes or other important functional genes for CHD were considered potentially pathogenic CNVs. Parental samples were further analyzed by SNP array when variants of unknown significance (VOUS) were detected in a fetus. If a CNV was inherited from a healthy parent, the finding was considered benign. The pathogenicity of the CNVs was determined according to the 2010 ISCA consortium criteria [12]. All clinically significant CNVs were further confirmed by real-time PCR.
Results
In this study, genomic DNA from 39 fetuses with cardiovascular abnormalities detected by ultrasound and a normal karyotype were examined. A total of 123 CNVs were identified (81 copy gain, 42 copy loss). These CNVs were distributed in 38/39 (97.4%) fetal genomes, and the size ranged from 1.6 kb to 3.8 Mb. In addition, 36 fetuses Tang/Lv/Chen/Bai/Li/Chen/Wang/Xu/Lu
Downloaded by: Univ. of California Santa Barbara 198.143.33.34 - 7/10/2015 11:35:20 AM
identified in surviving children are less severe [5]. Accordingly, a number of unknown human cardiac development genes could be identified by investigating the prenatal ultrasound diagnosis of the fetal heart defect. In the past few years, our understanding of the molecular pathway involved in cardiac development has markedly improved. A number of mutations in mendelian disease genes that result in heart defects have been identified, as well as chromosomal aberrations associated with CHDs [6]. In addition, it has been reported that microscopically visible chromosomal aberrations could be observed in 8–18% of CHD patients [7–9]. Furthermore, recent advances in genome-wide microarrays, such as comparative genomic hybridization arrays and singlenucleotide polymorphism (SNP) arrays, have led to the discovery of a large number of submicroscopic copy number variants (CNVs) associated with CHDs [10, 11]. For some of these CNVs, the corresponding regions have been found to contain critical genes involved in cardiac development. In order to explore the relationship between CNVs and CHDs, an SNP array analysis was performed for genomic samples obtained from 39 CHD fetuses with or without other ultrasound anomalies, and these fetuses all showed a normal karyotype. We predicted that these results may further expand our understanding of CHD etiology.
Table 1. Characteristics of 39 CHD, normal karyotype fetuses with clinically significant CNVs (cases 1 – 4) or VOUS (cases 8 – 9) detected
by genome SNP array Loci
Copy CNV coordinate
Size
Ultrasound abnormalities cardiac defect
1
22q11.21
loss
18877787 – 21462353
2584567
2
22q11.21
loss
18844632 – 21462353
2617722
3
22q11.21
loss
18895227 – 21409018
2513792
4
22q11.21– loss 11.22
21798907 – 22922798
1123892
5 6
16q12.2 Xp11.4
loss gain
53213419 – 54075698 39019742 – 40288074
862280 1268333
7
Xq28
gain
152040308 – 153886394 1846087
8 9
19p13.2 19p13.3 13q34
gain loss loss
319874 8067125 – 8386998 1019732 1657281 – 2677012 114495479 – 115106996 611518
associated anomalies
ventricular septal defect, aortic stenosis tetralogy of Fallot ventricular septal defect, tricuspid atresia, pulmonary stenosis and hypoplastic right heart anomalous pulmonary venous drainage, enlargement of the right heart atrioventricular septal defect aortic stenosis, ventricular septal defect
single ventricle, single atrium, persistent superior vena cava, aortic override visceral ectopic septal defect
Known syndrome and significant gene 22q11.2 deletion syndrome 22q11.2 deletion syndrome 22q11.2 deletion syndrome
cleft lip
Dandy-Walker malformation, bilateral renal agenesis, single umbilical artery
distal monosomy 22q11.2 syndrome RBL2 BCOR
FLNA
cleft lip bilateral multiple choroid plexus cyst, increased nuchal translucency
had more than 1 CNV, and the largest number of CNVs found in a fetus was 8 (online suppl. table 1, www.karger. com/doi/10.1159/000431320). Of these CNVs, VOUS were detected in 2/39 (5.1%) cases. Clinically significant CNVs were detected in 7/39 (17.9%) cases, with known syndromes identified in 4 cases [22q11.2 deletion syndrome (n = 3), distal monosomy 22q11.2 (n = 1)], and potential pathogenic CNVs identified in 3 cases (Xq28 dup, Xp11.4 dup, and 16q12.2 del) (table 1). Cases 1–3 harbored a deletion in chromosome 22q11.2 that encompassed TBX1 and contributed to 22q11.2 deletion syndrome (velocardiofacial/DiGeorge syndrome). Case 1 was characterized by a ventricular septal defect, aortic stenosis, and the deletion of 2.58 Mb (18877787– 21462353). In case 2, tetralogy of Fallot was accompanied by a deletion of 2.62 Mb (18844632–21462353). In case 3, a ventricular septal defect, tricuspid atresia, pulmonary stenosis, and hypoplastic right heart was accompanied by a deletion of 2.51 Mb (18895227–21409018). Common phenotypes associated with 22q11.2 deletion syndrome include conotruncal cardiac anomalies, moderate immune deficiency, developmental delay and behavioral problems, facial dysmorphia, palatal dysfunction, and feeding difficulties [13, 14]. However, the ultrasound re-
sults obtained only presented variable abnormalities of the heart. It is possible that the ultrasonic technology was restricted, and the size of the deletions contributed to the variable phenotypes that were associated with the 22q11.2 deletion and DiGeorge syndrome cases [15]. In case 4, ultrasound anomalies were accompanied by anomalous pulmonary venous drainage, enlargement of the right heart, and a cleft lip in a fetus of 30 + 4 weeks’ gestational age. The fetus had a 1.12-Mb deletion of 22q11.21–q11.22. This deletion has previously been associated with distal monosomy 22q11.2 syndrome [16]. Case 5 included a septal defect and a 0.68-Mb deletion in chromosome 16q12.2. This region of the genome harbors five genes (RPGRIP1L, FTO, CHD9, RBL2, and AKTIP) that are implicated in cell proliferation, differentiation, and apoptosis. Parental samples were also examined by SNP array, and the paternal genome included a 0.50-Mb deletion in chromosome 16q12.2 involving RPGRIP1L and FTO (fig. 1). The fetal deletion was larger than his father’s and contained three additional critical genes for cell proliferation. Thus, we considered this CNV may be pathogenic. For case 6, an ultrasound examination was performed at 23 + 1 weeks of gestational age. Aortic stenosis, a ven-
Diagnosis of DNA CNVs by Genomic SNP Array in Fetuses with CHDs
Fetal Diagn Ther DOI: 10.1159/000431320
3
Downloaded by: Univ. of California Santa Barbara 198.143.33.34 - 7/10/2015 11:35:20 AM
Patient
Color version available online
a
b
4
Fetal Diagn Ther DOI: 10.1159/000431320
Tang/Lv/Chen/Bai/Li/Chen/Wang/Xu/Lu
Downloaded by: Univ. of California Santa Barbara 198.143.33.34 - 7/10/2015 11:35:20 AM
Fig. 1. Genome analysis and ultrasound findings for case 5. a An SNP array showed a 0.68-Mb deletion in chromosome 16q12.2. The five genes harbored in this deletion include: RPGRIP1L, FTO, AKTIP, CHD9, and RBL2. b An SNP array analysis of the paternal genomic DNA for case 5 showed a 0.5-Mb deletion in 16q12.2, and this deletion included RPGRIP1L and FTO. c An atrioventricular septal defect was observed by ultrasound in case 5. (For figure 1c see next page.)
tricular septal defect, Dandy-Walker malformation, bilateral renal agenesis, and a single umbilical artery were observed. Genomic samples from this fetus indicated that a 1.26-Mb duplication was located in Xp11.4 (fig. 2). This region harbors a morbid gene, BCOR, which has been linked to oculofaciocardiodental (OFCD) syndrome and Lenz microphthalmia syndrome [17]. In case 7, a 1.84-Mb duplication was present in chromosome Xq28. In addition, a single ventricle, single atrium, persistent superior vena cava, and aortic override were observed by ultrasound. Chromosome Xq28 is a gene-rich region (fig. 3). In particular, the duplication includes the MECP2 gene, which is usually associated with Xq28 (MECP2) duplication syndrome in living individuals [18]. Nevertheless, there have been few studies that have linked this syndrome with CHDs. Among the other genes identified, FLNA is a plausible contributor to CHDs. The ultrasound examination performed in case 8 showed a fetus with visceral ectopic and cleft lip at a gestational age of 25 weeks. SNP array analysis revealed a 0.31-Mb duplication in chromosome 19p13.2. The region contains seven genes, ELAVL1, NDUFA7, RPS28, CD320, CERS4, CCL25, and FBN3. These genes are mainly involved in cell anabolism, immunoregulation, and cancer development, yet no study has reported that they are implicated in cardiac development. The CNVs were not found to be inherited from the parents of the fetus. In case 9, the ultrasound revealed a ventricular septal defect, a bilateral multiple choroid plexus cyst, and increased nuchal translucency at 26 weeks’ gestational age. The results of the SNP array analysis revealed a 1.01-Mb
SNP array represents an important prenatal genetic diagnostic tool for exploring the possible genetic causes of CHDs in fetuses. By providing whole-genome screening, chromosomal imbalances are obtained at a much higher resolution than by conventional karyotyping. Therefore, SNP arrays can provide additional information for prenatal genetic counseling and risk assessment. Only a few studies have evaluated the usefulness of genome-wide microarrays for the prenatal diagnosis of fetuses with CHDs. In 2012, Schmid et al. [21] performed a genome-wide microarray analysis of samples from 12 CHD fetuses with a normal karyotype and negative for 22q11.2 deletion syndrome. The detection rate for potentially causal CNVs was 25% (with 1 Mb resolution). In 2013, MademontSoler et al. [22] reported that 6.4% of fetuses with CHDs had 22q11.2 deletion syndrome. However, when this microdeletion syndrome and abnormal karyotype were excluded, the detection rate of pathogenic CNVs was 2.0% in fetuses with a CHD and no detectable VOUS. In 2014, Yan et al. [19] reported the detection rate of pathogenic CNVs to be 6.6% for 76 fetuses with CHDs without the 22q11.2 deletion and a normal karyotype, and the VOUS rate was 5.3%. Moreover, when Liao et al. [23] subjected 99 samples from fetuses with CHD and a normal karyotype to SNP array, the detection rate for clinically significant CNVs was 19.2%, and the proportion of VOUS was 3% (300-kb resolution). Recently, in a large sample research, Donnelly et al. [24] reported a subgroup analysis
Diagnosis of DNA CNVs by Genomic SNP Array in Fetuses with CHDs
Fetal Diagn Ther DOI: 10.1159/000431320
Discussion
5
Downloaded by: Univ. of California Santa Barbara 198.143.33.34 - 7/10/2015 11:35:20 AM
Color version available online
1 c
deletion in chromosome 19p13.3 and a 0.61-Mb deletion in chromosome 13q34. The deletion in 19p13.3 contains 28 genes, but current studies found that no gene was associated with cardiac development. The deletion in 13q34 was found to be inherited from the healthy father of the fetus. There were no differences in the frequency of chromosomal aberrations that were detected between the cases with CHDs with other anomalies and the cases with CHD alone [n = 5/18 (27.8%) vs. n = 4/21 (19.0%); p > 0.05]. These results are consistent with a previously published report [19]. Moreover, when Xiang et al. [20] studied chromosomal aberrations associated with mental retardation, most of the confirmed pathogenic CNVs were >500 kb in size. In the present study, CNVs >500 kb were more likely to be pathogenic compared with the CNVs that were
Fetal Diagn Ther DOI: 10.1159/000431320
Prenatal Diagnosis of DNA Copy Number Variations by Genomic Single-Nucleotide Polymorphism Array in Fetuses with Congenital Heart Defects Shaohua Tang a, b Jiaojiao Lv a Xiangnan Chen a Lili Bai b Huanzheng Li b Chong Chen b Ping Wang a Xueqin Xu b Jianxin Lu a
b
Key Laboratory of Medical Genetics, School of Laboratory Medicine and Life Science, Wenzhou Medical University, and Key Laboratory of Birth Defects, Department of Genetics, Wenzhou Central Hospital, Wenzhou, China
Key Words Prenatal diagnosis · Congenital heart disease · Genetic counseling
Abstract Objectives: To evaluate the usefulness of single-nucleotide polymorphism (SNP) array for prenatal genetic diagnosis of congenital heart defect (CHD), we used this approach to detect clinically significant copy number variants (CNVs) in fetuses with CHDs. Methods: A HumanCytoSNP-12 array was used to detect genomic samples obtained from 39 fetuses that exhibited cardiovascular abnormalities on ultrasound and had a normal karyotype. The relationship between CNVs and CHDs was identified by using genotype-phenotype comparisons and searching of chromosomal databases. All clinically significant CNVs were confirmed by real-time PCR. Results: CNVs were detected in 38/39 (97.4%) fetuses: variants of unknown significance were detected in 2/39 (5.1%), and clinically significant CNVs were identified in 7/39 (17.9%). In 3 of the 7 fetuses with clinically significant CNVs, 3 rare and previously undescribed CNVs were detected, and these CNVs encompassed the CHD candidate genes FLNA (Xq28 dup), BCOR (Xp11.4 dup), and RBL2 (16q12.2 del). Conclusion: Compared with conventional cytogenetic genomics, SNP array analysis provides significantly improved detection
© 2015 S. Karger AG, Basel 1015–3837/15/0000–0000$39.50/0 E-Mail [email protected] www.karger.com/fdt
of submicroscopic genomic aberrations in pregnancies with CHDs. Based on these results, we propose that genomic SNP array is an effective method which could be used in the prenatal diagnostic test to assist genetic counseling for pregnancies with CHDs. © 2015 S. Karger AG, Basel
Introduction
Congenital heart defects (CHDs) are a major cause of mortality in the perinatal period and affect an estimated 10% of spontaneous miscarriages [1]. Furthermore, CHDs are responsible for more deaths in the 1st year of life than any other etiology [2]. Recently, the Ministry of Health of China reported that CHDs account for 26.7% of birth defects that occur in China [3]. A complex set of environmental and genetic factors are potentially etiologic in the pathogenesis of CHDs. Correspondingly, more than 1,700 genes have been estimated to play important roles in cardiac development in mice (http://www.informatics.jax.org/). The homologs of these genes may be involved in cardiac development in humans [4], and 55 human genes have been identified to have roles in CHDs [5]. It is possible that variations in critical genes lead to severe birth defects which are lethal to a fetus, whereas the CHDs Jianxin Lu Key Laboratory of Medical Genetics School of Laboratory Medicine and Life Science, Wenzhou Medical University Ouhai District, Wenzhou, Zhejiang 325000 (China) E-Mail jxlu313 @ 163.com
Downloaded by: Univ. of California Santa Barbara 198.143.33.34 - 7/10/2015 11:35:20 AM
a
Methods Samples This prospective study was conducted from January 2011 to February 2014. A total of 364 pregnant women with a gestational age ranging from 18 to 32 weeks underwent invasive testing. A genomic SNP array analysis was subsequently performed at the Center of Wenzhou Prenatal Diagnostics. There were 45 fetuses with cardiovascular abnormalities that were detected by ultrasound and confirmed by a pediatric cardiologist. In some cases, other ultrasound anomalies were detected as well. An abnormal karyotype in a conventional G-band karyotype analysis led to the exclusion of 6 of these 45 fetuses. As a result, 39 fetuses were enrolled in this study, including 18 fetuses with isolated CHD and 21 fetuses with CHD plus other ultrasound anomalies. Both ultrasound findings and karyotype were supported at the Wenzhou Central Hospital (Wenzhou, Zhejiang, China). This study was approved by the Dingli Clinical School Ethics Committee of Wenzhou Medical University, and informed consent was obtained from each prenatal couple. DNA Extraction Genomic DNA was extracted from umbilical cord blood samples (n = 33) and amniotic fluid samples (n = 6) using Qiagen Blood Mini Kits (Qiagen, USA) and Biochain DNA Extraction Kits
2
Fetal Diagn Ther DOI: 10.1159/000431320
(Tissue and cells; Biochain, USA), each according to the manufacturers’ instructions. When it was necessary to exclude maternal contamination and to assist in the interpretation of the CNVs, parental peripheral blood samples were collected and genomic DNA was extracted using Qiagen Blood Mini Kits. SNP Array Platform Genomic DNA samples were subjected to an SNP array analysis using the HumanCytoSNP-12 array v1.0 (Illumina Inc., USA). 298,563 overlapping SNP probes provide a useful content for cytogenetic analysis. This includes a dense coverage of around 250 genomic regions commonly screened in cytogenetics laboratories, including subtelomeric regions, pericentromeric regions, sex chromosomes, and targeted coverage in around 400 additional disease-related genes. The average resolution of this array is 31 kb. DNA digestion, ligation, PCR amplification, fragmentation, labeling, and hybridization for each sample were performed according to the manufacturer’s protocols. The array slide was scanned using an Iscan Reader (Illumina). CNV Identification and Assessment To analyze the generated data, the GenomeStudio v2011 software (human genome build 37/Hg19 for analysis) and Illumina cnvPartition (v3.2.0) were used, respectively. All detected gain-ofcopy number and loss-of-copy number CNVs were compared with CNVs listed in our in-house database and the following publically available databases: Database of Genomic Variants (DGV, http:// dgv.tcag.ca/dgv/app/home), International Standards for Cytogenomic Arrays (ISCA) consortium (https://www.iscaconsortium. org/), UCSC Genome Bioinformatics (http://genome.ucsc.edu/), DECIPHER Database (https://decipher.sanger.ac.uk/index), CHDWiki (http://homes.esat.kuleuven.be/∼bioiuser/chdwiki/index.php/Main_Page), Online Mendelian Inheritance in Man (OMIM; http://www.omim.org/), and PubMed (http://www.ncbi. nlm.nih.gov/pubmed). If a CNV was described in a CNV database for healthy individuals (DGV), or in our local normal CNV database, it was considered to be benign. If a CNV was described in the DECIPHER pathogenic chromosomal database, or had been reported to be a pathogenic CNV in the literature, it was considered pathogenic. CNVs identified as morbid OMIM genes or other important functional genes for CHD were considered potentially pathogenic CNVs. Parental samples were further analyzed by SNP array when variants of unknown significance (VOUS) were detected in a fetus. If a CNV was inherited from a healthy parent, the finding was considered benign. The pathogenicity of the CNVs was determined according to the 2010 ISCA consortium criteria [12]. All clinically significant CNVs were further confirmed by real-time PCR.
Results
In this study, genomic DNA from 39 fetuses with cardiovascular abnormalities detected by ultrasound and a normal karyotype were examined. A total of 123 CNVs were identified (81 copy gain, 42 copy loss). These CNVs were distributed in 38/39 (97.4%) fetal genomes, and the size ranged from 1.6 kb to 3.8 Mb. In addition, 36 fetuses Tang/Lv/Chen/Bai/Li/Chen/Wang/Xu/Lu
Downloaded by: Univ. of California Santa Barbara 198.143.33.34 - 7/10/2015 11:35:20 AM
identified in surviving children are less severe [5]. Accordingly, a number of unknown human cardiac development genes could be identified by investigating the prenatal ultrasound diagnosis of the fetal heart defect. In the past few years, our understanding of the molecular pathway involved in cardiac development has markedly improved. A number of mutations in mendelian disease genes that result in heart defects have been identified, as well as chromosomal aberrations associated with CHDs [6]. In addition, it has been reported that microscopically visible chromosomal aberrations could be observed in 8–18% of CHD patients [7–9]. Furthermore, recent advances in genome-wide microarrays, such as comparative genomic hybridization arrays and singlenucleotide polymorphism (SNP) arrays, have led to the discovery of a large number of submicroscopic copy number variants (CNVs) associated with CHDs [10, 11]. For some of these CNVs, the corresponding regions have been found to contain critical genes involved in cardiac development. In order to explore the relationship between CNVs and CHDs, an SNP array analysis was performed for genomic samples obtained from 39 CHD fetuses with or without other ultrasound anomalies, and these fetuses all showed a normal karyotype. We predicted that these results may further expand our understanding of CHD etiology.
Table 1. Characteristics of 39 CHD, normal karyotype fetuses with clinically significant CNVs (cases 1 – 4) or VOUS (cases 8 – 9) detected
by genome SNP array Loci
Copy CNV coordinate
Size
Ultrasound abnormalities cardiac defect
1
22q11.21
loss
18877787 – 21462353
2584567
2
22q11.21
loss
18844632 – 21462353
2617722
3
22q11.21
loss
18895227 – 21409018
2513792
4
22q11.21– loss 11.22
21798907 – 22922798
1123892
5 6
16q12.2 Xp11.4
loss gain
53213419 – 54075698 39019742 – 40288074
862280 1268333
7
Xq28
gain
152040308 – 153886394 1846087
8 9
19p13.2 19p13.3 13q34
gain loss loss
319874 8067125 – 8386998 1019732 1657281 – 2677012 114495479 – 115106996 611518
associated anomalies
ventricular septal defect, aortic stenosis tetralogy of Fallot ventricular septal defect, tricuspid atresia, pulmonary stenosis and hypoplastic right heart anomalous pulmonary venous drainage, enlargement of the right heart atrioventricular septal defect aortic stenosis, ventricular septal defect
single ventricle, single atrium, persistent superior vena cava, aortic override visceral ectopic septal defect
Known syndrome and significant gene 22q11.2 deletion syndrome 22q11.2 deletion syndrome 22q11.2 deletion syndrome
cleft lip
Dandy-Walker malformation, bilateral renal agenesis, single umbilical artery
distal monosomy 22q11.2 syndrome RBL2 BCOR
FLNA
cleft lip bilateral multiple choroid plexus cyst, increased nuchal translucency
had more than 1 CNV, and the largest number of CNVs found in a fetus was 8 (online suppl. table 1, www.karger. com/doi/10.1159/000431320). Of these CNVs, VOUS were detected in 2/39 (5.1%) cases. Clinically significant CNVs were detected in 7/39 (17.9%) cases, with known syndromes identified in 4 cases [22q11.2 deletion syndrome (n = 3), distal monosomy 22q11.2 (n = 1)], and potential pathogenic CNVs identified in 3 cases (Xq28 dup, Xp11.4 dup, and 16q12.2 del) (table 1). Cases 1–3 harbored a deletion in chromosome 22q11.2 that encompassed TBX1 and contributed to 22q11.2 deletion syndrome (velocardiofacial/DiGeorge syndrome). Case 1 was characterized by a ventricular septal defect, aortic stenosis, and the deletion of 2.58 Mb (18877787– 21462353). In case 2, tetralogy of Fallot was accompanied by a deletion of 2.62 Mb (18844632–21462353). In case 3, a ventricular septal defect, tricuspid atresia, pulmonary stenosis, and hypoplastic right heart was accompanied by a deletion of 2.51 Mb (18895227–21409018). Common phenotypes associated with 22q11.2 deletion syndrome include conotruncal cardiac anomalies, moderate immune deficiency, developmental delay and behavioral problems, facial dysmorphia, palatal dysfunction, and feeding difficulties [13, 14]. However, the ultrasound re-
sults obtained only presented variable abnormalities of the heart. It is possible that the ultrasonic technology was restricted, and the size of the deletions contributed to the variable phenotypes that were associated with the 22q11.2 deletion and DiGeorge syndrome cases [15]. In case 4, ultrasound anomalies were accompanied by anomalous pulmonary venous drainage, enlargement of the right heart, and a cleft lip in a fetus of 30 + 4 weeks’ gestational age. The fetus had a 1.12-Mb deletion of 22q11.21–q11.22. This deletion has previously been associated with distal monosomy 22q11.2 syndrome [16]. Case 5 included a septal defect and a 0.68-Mb deletion in chromosome 16q12.2. This region of the genome harbors five genes (RPGRIP1L, FTO, CHD9, RBL2, and AKTIP) that are implicated in cell proliferation, differentiation, and apoptosis. Parental samples were also examined by SNP array, and the paternal genome included a 0.50-Mb deletion in chromosome 16q12.2 involving RPGRIP1L and FTO (fig. 1). The fetal deletion was larger than his father’s and contained three additional critical genes for cell proliferation. Thus, we considered this CNV may be pathogenic. For case 6, an ultrasound examination was performed at 23 + 1 weeks of gestational age. Aortic stenosis, a ven-
Diagnosis of DNA CNVs by Genomic SNP Array in Fetuses with CHDs
Fetal Diagn Ther DOI: 10.1159/000431320
3
Downloaded by: Univ. of California Santa Barbara 198.143.33.34 - 7/10/2015 11:35:20 AM
Patient
Color version available online
a
b
4
Fetal Diagn Ther DOI: 10.1159/000431320
Tang/Lv/Chen/Bai/Li/Chen/Wang/Xu/Lu
Downloaded by: Univ. of California Santa Barbara 198.143.33.34 - 7/10/2015 11:35:20 AM
Fig. 1. Genome analysis and ultrasound findings for case 5. a An SNP array showed a 0.68-Mb deletion in chromosome 16q12.2. The five genes harbored in this deletion include: RPGRIP1L, FTO, AKTIP, CHD9, and RBL2. b An SNP array analysis of the paternal genomic DNA for case 5 showed a 0.5-Mb deletion in 16q12.2, and this deletion included RPGRIP1L and FTO. c An atrioventricular septal defect was observed by ultrasound in case 5. (For figure 1c see next page.)
tricular septal defect, Dandy-Walker malformation, bilateral renal agenesis, and a single umbilical artery were observed. Genomic samples from this fetus indicated that a 1.26-Mb duplication was located in Xp11.4 (fig. 2). This region harbors a morbid gene, BCOR, which has been linked to oculofaciocardiodental (OFCD) syndrome and Lenz microphthalmia syndrome [17]. In case 7, a 1.84-Mb duplication was present in chromosome Xq28. In addition, a single ventricle, single atrium, persistent superior vena cava, and aortic override were observed by ultrasound. Chromosome Xq28 is a gene-rich region (fig. 3). In particular, the duplication includes the MECP2 gene, which is usually associated with Xq28 (MECP2) duplication syndrome in living individuals [18]. Nevertheless, there have been few studies that have linked this syndrome with CHDs. Among the other genes identified, FLNA is a plausible contributor to CHDs. The ultrasound examination performed in case 8 showed a fetus with visceral ectopic and cleft lip at a gestational age of 25 weeks. SNP array analysis revealed a 0.31-Mb duplication in chromosome 19p13.2. The region contains seven genes, ELAVL1, NDUFA7, RPS28, CD320, CERS4, CCL25, and FBN3. These genes are mainly involved in cell anabolism, immunoregulation, and cancer development, yet no study has reported that they are implicated in cardiac development. The CNVs were not found to be inherited from the parents of the fetus. In case 9, the ultrasound revealed a ventricular septal defect, a bilateral multiple choroid plexus cyst, and increased nuchal translucency at 26 weeks’ gestational age. The results of the SNP array analysis revealed a 1.01-Mb
SNP array represents an important prenatal genetic diagnostic tool for exploring the possible genetic causes of CHDs in fetuses. By providing whole-genome screening, chromosomal imbalances are obtained at a much higher resolution than by conventional karyotyping. Therefore, SNP arrays can provide additional information for prenatal genetic counseling and risk assessment. Only a few studies have evaluated the usefulness of genome-wide microarrays for the prenatal diagnosis of fetuses with CHDs. In 2012, Schmid et al. [21] performed a genome-wide microarray analysis of samples from 12 CHD fetuses with a normal karyotype and negative for 22q11.2 deletion syndrome. The detection rate for potentially causal CNVs was 25% (with 1 Mb resolution). In 2013, MademontSoler et al. [22] reported that 6.4% of fetuses with CHDs had 22q11.2 deletion syndrome. However, when this microdeletion syndrome and abnormal karyotype were excluded, the detection rate of pathogenic CNVs was 2.0% in fetuses with a CHD and no detectable VOUS. In 2014, Yan et al. [19] reported the detection rate of pathogenic CNVs to be 6.6% for 76 fetuses with CHDs without the 22q11.2 deletion and a normal karyotype, and the VOUS rate was 5.3%. Moreover, when Liao et al. [23] subjected 99 samples from fetuses with CHD and a normal karyotype to SNP array, the detection rate for clinically significant CNVs was 19.2%, and the proportion of VOUS was 3% (300-kb resolution). Recently, in a large sample research, Donnelly et al. [24] reported a subgroup analysis
Diagnosis of DNA CNVs by Genomic SNP Array in Fetuses with CHDs
Fetal Diagn Ther DOI: 10.1159/000431320
Discussion
5
Downloaded by: Univ. of California Santa Barbara 198.143.33.34 - 7/10/2015 11:35:20 AM
Color version available online
1 c
deletion in chromosome 19p13.3 and a 0.61-Mb deletion in chromosome 13q34. The deletion in 19p13.3 contains 28 genes, but current studies found that no gene was associated with cardiac development. The deletion in 13q34 was found to be inherited from the healthy father of the fetus. There were no differences in the frequency of chromosomal aberrations that were detected between the cases with CHDs with other anomalies and the cases with CHD alone [n = 5/18 (27.8%) vs. n = 4/21 (19.0%); p > 0.05]. These results are consistent with a previously published report [19]. Moreover, when Xiang et al. [20] studied chromosomal aberrations associated with mental retardation, most of the confirmed pathogenic CNVs were >500 kb in size. In the present study, CNVs >500 kb were more likely to be pathogenic compared with the CNVs that were
