|Home | About | Journals | Submit | Contact Us | Français|
Recurrent interstitial deletion of a region of 8p23.1 flanked by the low copy repeats 8p-OR-REPD and 8p-OR-REPP is associated with a spectrum of anomalies that can include congenital heart malformations and congenital diaphragmatic hernia (CDH). Haploinsufficiency of GATA4 is thought to play a critical role in the development of these birth defects. We describe two individuals and a monozygotic twin pair discordant for anterior CDH all of whom have complex congenital heart defects caused by this recurrent interstitial deletion as demonstrated by array comparative genome hybridization. To better define the genotype/phenotype relationships associated with alterations of genes on 8p23.1, we review the spectrum of congenital heart and diaphragmatic defects that have been reported in individuals with isolated GATA4 mutations and interstitial, terminal, and complex chromosomal rearrangements involving the 8p23.1 region. Our findings allow us to clearly define the CDH minimal deleted region on chromosome 8p23.1 and suggest that haploinsufficiency of other genes, in addition to GATA4, may play a role in the severe cardiac and diaphragmatic defects associated with 8p23.1 deletions. These findings also underscore the importance of conducting a careful cytogenetic/molecular analysis of the 8p23.1 region in all prenatal and postnatal cases involving congenital defects of the heart and/or diaphragm.
Published deletions involving chromosome 8p23.1 range from large terminal deletions that are easily detectable by routine chromosome analysis to small interstitial deletions which are best identified using fluorescence in situ hybridization (FISH) or molecular techniques such as array comparative genomic hybridization (aCGH) [Paez et al., 2008; Reddy 1999; Devriendt et al., 1999]. Recurrent deletions of a region of 8p23.1 flanked by low copy repeats 8p-OR-REPD (distal) and 8p-OR-REPP (proximal) are associated with a spectrum of anomalies including congenital heart malformations, congenital diaphragmatic hernia (CDH), developmental delay, and neuropsychiatric findings [Digilio et al., 1998; Ciccone et al., 2001; Shimokawa et al., 2005; Ciccone et al., 2006; Zuffardi et al., 2006].
This region contains the transcription factor gene GATA4 which is known to play a key role in heart development in humans [Garg et al., 2003; Okubo et al., 2004; Sarkozy et al., 2005]. Disruption of Gata4 in mice has also been shown to cause heart defects and anterior CDH [Molkentin et al., 1997; Kuo et al., 1997; Jay et al., 2007]. Mice homozygous for a Gata4 null allele failed to form a heart tube and died between E7.0 and E9.5 [Kuo et al., 1997; Molkentin et al., 1997]. More recently, Jay et al. reported that a significant fraction of C57BL/6 mice heterozygous for a Gata4 deletion in exon 2—which removed the translation start site and the N-terminal activation domain—died within 1 day of birth, and that developmental defects in these heterozygotes included atrial septal defects (ASD), ventricular septal defects (VSD) and atrioventricular septal defects (AVSD) [Jay et al., 2007].
In this report, we describe two individuals and an identical twin pair discordant for anterior CDH all of whom were diagnosed prenatally with complex congenital heart defects and were subsequently shown to have interstitial 8p23.1 deletions. We review the spectrum of congenital heart and diaphragmatic defects that have been reported in individuals with isolated GATA4 mutations and interstitial, terminal, and complex chromosomal rearrangements involving 8p23.1 and discuss the implications of this data on our understanding of the gene(s) within this region which may play a role in heart and diaphragm development.
Patient 1 is a male born at 39 3/7 weeks gestation to a 26-year-old G2P1→2 mother. Both parents were Hispanic and nonconsanguineous. Prenatal ultrasound revealed pulmonary and tricuspid atresia. No amniocentesis was performed. Birth weight was 2.82 kg (10th centile), length was 48.5 cm (50th centile) and frontooccipital circumference (OFC) was 31.5 cm (25th centile). In the neonatal period he was noted to have a two-vessel umbilical cord, a very large anterior fontanelle, downslanting palpebral fissures, mild micrognathia, a short neck, slender fingers, and proximally implanted thumbs. Head ultrasound revealed normal anatomy with possible thinning of the corpus callosum.
Shortly after birth he was admitted to the neonatal intensive care unit for respiratory distress and tachypnea. Postnatal cardiac evaluation revealed a complex single ventricle anatomy consisting of a double outlet right ventricle, unbalanced complete atrioventricular septal defect (AVSD) with left atrioventricular (AV) valve atresia and left ventricular hypoplasia, pulmonary atresia, and D-transposition of the great vessels. Despite stenting of his ductus ateriosus, he continued to have limited pulmonary outflow. During surgery to create a systemic-to-pulmonary artery (Blalock-Taussig) shunt and to ligate the stented patent ductus arteriosus (PDA), he was found to have bilateral superior vena cavae with absence of a bridging innominate vein.
Approximately 6 weeks after birth, he experienced a cardiopulmonary arrest requiring resuscitation and emergent use of extracorporeal membrane oxygenation (ECMO). While on ECMO he developed seizures. He was weaned, decannulated and maintained for a short period of time on high-dose inotropic support before dying at age one month.
Patient 2 is a female born via cesarean at 35 weeks to a G2P1→3 mother as part of a twin gestation. Prenatal ultrasound showed complex congenital heart disease and intra-uterine growth retardation. No anomalies were identified in the male twin. No amniocentesis was performed.
The family history was significant for postaxial polydactyly in father, paternal half brother and maternal half sister. A paternal half brother had aortic insufficiency but this was felt to be secondary to rheumatic heart disease.
Birth weight was 1320 gm (<5th centile), length was 39 cm (<5th centile) and frontooccipital circumference (OFC) was 27 cm (<5th centile). His weight, length, and OFC were at approximately the 50th centile level for a 29-30 week gestation infant, consistent with symmetric intra-uterine growth retardation. Postaxial polydactyly of the left hand was noted along with prominent heels. No facial dysmorphic features were noted.
Shortly after birth, the patient developed respiratory distress requiring intubation and mechanical ventilation. Cardiac evaluation by ECHO and cardiac angiography revealed a complex cardiovascular malformation consisting of an unbalanced AVSD, left AV valve atresia, hypoplastic left ventricle, severe tricuspid regurgitation, absent intrahepatic inferior vena cava and hemiazygos continuation to a left superior vena cava. Head and abdominal ultrasounds were normal.
Considering the severity of the cardiac condition, life-support was withdrawn with parental consent and the patient died at 88 days of life.
Patient 3 is a male infant born at 34 5/7 weeks via cesarean to a 33-year-old G1P0→2 mother as part of a monozygotic twin gestation (see Patient 4). Prenatal ultrasound at 20-weeks gestation revealed an AVSD with a large inlet ventricular septal defect (VSD) extending to the perimembranous septum, aortic hypoplasia, a dominant pulmonary artery, and mild tricuspid regurgitation. An amniocentesis was performed and aCGH and chromosome analyses revealed a deletion on chromosome 8p.
Birth weight was 1835 grams (25th centile), length was 44.5 cm (50th centile), and frontooccipital circumference (OFC) was 29 cm (10th centile). At birth, examination revealed a holosystolic murmur, mild dysmorphic facial features, minor digital anomalies, and a sacral dimple. Spinal cord tethering was ruled out by spinal ultrasound.
Postnatal cardiac evaluation revealed a balanced type A complete AVSD with mild AV regurgitation, moderate atrial septal defect (ASD), hypoplastic aortic arch, a narrow left ventricular outflow tract, a large PDA, and good biventricular function. At surgery, he was found to have a right atrium with two appendages and absence of a bridging innominate vein.
Shortly after birth, the patient was transferred to the neonatal intensive care with tachypnea and desaturations. Further evaluation revealed evidence of a large anterior congenital diaphragmatic hernia with herniation of the liver, levoposition of the heart and a hypoplastic, posteriorly rotated left lung.
The patient remained hospitalized until approximately 12 weeks of age and is presently 5-months-old.
Patient 4 is the male monozygotic twin of patient 3. Prenatally, he was found to a balanced complete AVSD, a small aortic arch, and a large PDA. An amniocentesis was performed and aCGH and chromosome analyses revealed a deletion on chromosome 8p.
He was born at 2020 grams (25th centile), with a length of 44.8 cm (30th centile), and frontooccipital circumference (OFC) of 29.5 cm (10th centile). Additional findings on postnatal cardiac evaluation included a small aortic valve, bilateral peripheral pulmonary stenosis, and mild right ventricular dilatation. He was also found to have a small right choroid plexus cyst and a sacral dimple. Spinal cord tethering was ruled out by spinal ultrasound.
He was discharged home on day of life 38, with a discharge weight of 2960 g, a length of 48.5 cm and an OFC of 33 cm.
All four patients were referred for clinical cytogenetic testing. Institutional review board approved informed consent was obtained for high-resolution studies.
Standard laboratory procedures were used for metaphase preparations. G-banded chromosome analysis and FISH analysis were performed in either Medical Genetics Laboratories at Baylor College of Medicine or by the referring center. DNA was extracted from peripheral blood [Ou et al., 2008] and amniotic fluid samples [Bi et al., 2008] as previously described. Patients 1 and 2 were examined in the neonatal period and Patients 3 and 4 in the prenatal period. In the case of Patients 1, 3, and 4, aCGH testing and chromosome analysis were performed concurrently at the Medical Genetics Laboratories at Baylor College of Medicine. For Patient 2, aCGH analysis was performed following normal chromosome analysis results. All clinical aCGH analyses were performed using Baylor College of Medicine (BCM) Chromosomal Microarray version 6 (CMA V6) manufactured by Agilent Technologies (Santa Clara, CA) according to the manufacturer's instructions with modifications [Ou et al., 2008].
DNA from Patients 1 and 4 was further evaluated by high resolution genome-wide aCGH using Human Genome CGH 244K Oligo Microarray Kits G4411B (Agilent Technologies, Santa Clara, CA) according to the manufacturer's protocol version 2.0. Arrays were scanned using an Agilent DNA Microarray Scanner (Agilent Technologies, Santa Clara, CA). Data extracted using Feature Extraction Software Version 9.1.3 (Agilent Technologies, Santa Clara, CA) was analyzed using CGH Analytics 3.4.40 Software (Agilent Technologies, Santa Clara, CA) with copy number changes identified with the assistance of the Aberration Detection Method 2 algorithm (threshold 6.0). Control DNA consisted of DNA from a healthy gender-matched reference individual with no personal or family history of heart or diaphragm defects.
PCR amplification of genomic fragments was performed using previously described primer pairs [Okubo et al., 2004]. PCR-fragments were cleaned and sequenced commercially (Agencourt Bioscience, Beverly, MA) and DNA changes were identified by comparing the published GATA4 sequence to sequence data obtained from patient samples using Sequencher 4.5 software (GeneCodes, Ann Arbor, MI).
Visible deletions of 8p were identified on the chromosome analysis for Patients 1, 3 and 4 (Fig 1). Chromosome analysis and FISH for 22q11.2 deletions performed on a peripheral blood sample from Patient 2 at an outside facility were reported as normal.
Clinical aCGH detected an identical 8p23.1 deletion in all four patients with a minimal deletion size of 2.945 Mb (8,850,913 to 11,796,333) and a maximal deletion size of 6.352 Mb (6,436,314 to 12,788,647) (Fig 1). Parental studies confirmed that all of the deletions were de novo.
High resolution aCGH analyses confirmed the deletions previously seen in the clinical studies. In each case the deletion was flanked by the low copy repeats 8p-OR-REPD and 8p-OR-REPP with differences in apparent deletion size between Patients 1 and 4 being attributable to known copy number variant regions within and/or directly adjacent to the low copy repeats. The minimal deleted interval encompassed twenty-two known genes/open reading frames including GATA4 (Fig 2).
To determine whether disruption of both GATA4 alleles could explain the severe cardiac defects and CDH seen in our patients, we screened the entire GATA4 coding sequence and intron/exon boundaries of the remaining GATA4 allele in Patients 1 and 4. No variations from the published GATA4 reference sequence were identified in the coding sequence or the intron/exon boundaries of these DNA samples.
We have described the phenotype associated with the recurrent deletion of 8p23.1 in four patients with prenatally diagnosed complex congenital heart defects. The GATA4 gene resides within this recurrently deleted interval and has been implicated as the gene responsible for heart defects associated with 8p23.1 deletions. The role of GATA4 in heart development is supported by mouse models [Kuo et al., 1997; Molkentin et al., 1997; Jay et al., 2007] and studies of patients with GATA4 gene mutations.
Some of the most compelling evidence that mutations in GATA4 are associated with heart defects in humans come from studies of seven families in which ASD was found to segregate with heterozygous GATA4 mutations with varying levels of penetrance (Table I) [Garg et al., 2003; Hirayama-Yamada et al., 2005; Sarkozy et al., 2005; Okubo et al., 2004]. The GATA4 gene has also been screened in individual patients with ASD, VSD, AVSD, and tetralogy of Fallot (TOF). Although a number of non-synonymous changes were identified in these studies, it was often difficult to conclude whether these changes were causal due to a lack of parental/population controls and/or functional studies. This is particularly true of changes that are inherited from a non-affected parent. The most convincing of these changes include a de novo E216D missense mutations in two individuals with TOF that was shown to have reduced transcriptional activity in a reporter assay, and a frameshift mutation (Pro226fs) and a premature stop codon mutation (Arg266Ter)—which severely truncate the protein—each found in a patient with atrioventricular septal defects (AVSD) (Table II) [Nemer et al., 2006; Reamon-Buettner et al., 2007].
The complex cardiac phenotypes seen in our patients led us to question whether the spectrum of heart defects associated with 8p23.1 deletions was more severe than that seen in patients with GATA4 mutations. With this in mind we reviewed previous reports of interstitial, terminal and complex deletions of 8p23.1 (Table III) to determine the spectrum of cardiac phenotypes associated with these deletions. Previously reported patients with interstitial deletions were found to have a spectrum of cardiac defects that included ASD, VSD, AVSD, pulmonary stenosis, pulmonary valve stenosis, tetralogy of Fallot, and/or combinations of these defects. Cardiac defects in patients with terminal deletions extending to at least 8p23.1 included AVSD, hypoplastic left heart, hypoplastic right ventricle, pulmonary atresia/stenosis, pulmonary valve stenosis, partial anomalous pulmonary venous return, subaortic stenosis, transposition of the great arteries, double inlet/double outlet right ventricle, double inlet left ventricle and tetralogy of Fallot.
When combined with data from the four patients described in this report, it would appear that the spectrum of heart defects is more severe in interstitial and terminal deletions involving 8p23.1 when compared to defects seen in patients with heterozygous GATA4 mutations. Although this increase in severity could be due to deleterious mutations in the remaining GATA4 allele, we did not detect such mutations in Patients 1 and 4. A similar evaluation carried out by Paez et al. also failed to identify mutations in the remaining GATA4 allele in two patients with 8p23.1 deletions involving GATA4 who also presented with heart defects [Paez et al., 2008].
An alternative explanation for the increase in severity would be the existence of another gene(s) in the recurrent 8p23.1 deletion region that impact heart development. Of the genes deleted along with GATA4, SOX7 is one of the most likely candidate genes. SOX7 is expressed in mouse and human adult heart, and in the early cardiogenic region of Xenopus embryos [Takash et al., 2001; Taniguchi et al., 1999; Zhang et al., 2005]. SOX7 mRNA injection induced cardiogenic marker expression in Xenopus animal cap explants whereas knockdown of SOX7 using morpholinos decreased the expression of cardiogenic markers MHCα and Nkx2.5; and marker expression was rescued by injection of RNA encoding a SOX7 transcript [Zhang et al., 2005]. Interestingly, silencing of Sox7 in mouse F9 embryonal carcinoma cells blunts the increase in Gata4 mRNA levels seen after treatment with all trans-retinoic acid/dibuterol cAMP [Futaki et al., 2004]. In contrast, silencing of Gata4 did not result in decreased Sox7 expression in the same system. This suggests that Sox7 lies upstream of Gata4, and that haploinsufficiency of SOX7 may exacerbate the cardiac phenotype of individuals with GATA4 deletions.
Genes outside the GATA4 region may also be involved in abnormal cardiac development based on reports of patients with inverted duplication deletion events involving 8p. The mechanism involved in creation of the most common inverted duplication deletion events of 8p has been well described [Ciccone et al., 2006]. These events typically result in loss of copy number (deletion) of 8p distal to 8p-OR-REPD, normal copy number of the GATA4 region between 8p-OR-REPD and 8p-OR-REPP, and a variable region of increased copy number (duplication) proximal to 8p-OR-REPP. The number of cardiac malformations associated with patients carrying the inverted duplication deletion 8p was 11/39 (28.2%), a much lower proportion than 17/18 (94.4%) and 45/60 (75%) associated with interstitial and terminal deletions, respectively (Table III). In general, the spectrum of cardiac malformations in these cases is also milder but still includes ASD, VSD, right aortic arch, and pulmonary stenosis (Table III). If we assume normal expression of GATA4 and other genes between 8p-OR-REPD and 8p-OR-REPP, we are left to conclude that decreased expression of one or more genes distal to 8p-OR-REPD and/or over expression of one or more genes proximal to 8p-OR-REPP can also contribute to the development of some heart defects.
Congenital diaphragmatic hernia is also common in patients with a deletion encompassing 8p23.1 with, at least, nine previously reported cases [Holder et al 2007]. CDH is associated with 4/18 (22.2%) of reported interstitial deletions and 5/60 (8.3%) of terminal deletions but has not been described in patients with inverted duplication deletions. This is, presumably, due to the fact that terminal deletions may not always include the CDH minimal deleted interval and that this interval has a normal copy number in individuals with inverted duplication deletions.
The majority of CDH patients with 8p23.1 deletions have been described as having left sided (assumedly-posterior) CDH. Indeed, Patient 3 is the only individual described to date with an anterior CDH associated with an 8p23.1 deletion. This is surprising since the heterozygous Gata4 mice described by Jay et al. have anterior CDH similar to that seen in our patient [Jay et al., 2007]. Using array data presented here and previously reported molecularly-defined deletions associated with CDH, the minimal deleted region for CDH on 8p can be defined as the region bounded by 8p-OR-REPD distally and 8p-OR-REPP proximally (Fig 3) [Faivre et al., 1998; Slavotinek et al., 2004; Shimokawa et al., 2005].
Although the use of array comparative genome hybridization can aid in the identification of 8p23.1 deletions in patients with heart and/or diaphragm defects, caution must be used in the interpretation of these findings. Chromosome 8p contains many copy number variant regions whose potential contribution to the development of birth defects has not been adequately studied. Recently Chen et al.  described a patient with a Fryns-like phenotype including congenital diaphragmatic hernia, macrocephaly, brachytelephalangy, nail hypoplasia, short webbed neck with redundant posterior nuchal skin, coarse face, flat and broad nasal bridge, hypertelorism, macrostomia, microretrognathia, and low-set ears. The patient's phenotype was attributed to a de novo 0.7 MB deletion within 8p23.1 [Chen et al., 2007]. However, this deletion lies entirely within a known copy number variant region making it difficult to determine if this deletion is causal. Without careful evaluation, physicians may erroneously quote sibling and offspring recurrence risks that are either too low or too high based on an incorrect assumption of causality.
When used properly, aCGH detects detrimental submicroscopic changes that can be easily missed on routine chromosomal analysis especially in prenatal samples where the band resolution may be compromised [Pecile et al., 1990; Wu et al., 1996; Faivre et al., 1998]. This is illustrated well in Patient 2 where chromosome analysis failed to identify an 8p deletion which was easily detected by aCGH. This, and the high frequency of cardiac and diaphragmatic defects associated with 8p23.1 interstitial deletions, leads us to recommend that aCGH be performed on all prenatal and postnatal cases with congenital cardiac and/or diaphragm defects. At the present time we do not recommend that the GATA4 gene be sequenced in patients with cardiac defects since the percentage of cases with protein-altering changes in GATA4 is likely to be small— approximately 1% based on the articles reviewed—and the clinical significance of the majority of GATA4 changes remains unclear.
The authors thank the patients and family members who participated in this study and Zhiyin Yu for her technical assistance. This study was funded in part by NIH grant HD-050583 (DAS).