- Split View
-
Views
-
Cite
Cite
Daryl A. Scott, Merel Klaassens, Ashley M. Holder, Kevin P. Lally, Caraciolo J. Fernandes, Robert-Jan Galjaard, Dick Tibboel, Annelies de Klein, Brendan Lee, Genome-wide oligonucleotide-based array comparative genome hybridization analysis of non-isolated congenital diaphragmatic hernia, Human Molecular Genetics, Volume 16, Issue 4, 15 February 2007, Pages 424–430, https://doi.org/10.1093/hmg/ddl475
- Share Icon Share
Abstract
Non-isolated congenital diaphragmatic hernia (CDH+) is a severe birth defect that is often caused by de novo chromosomal anomalies. In this report, we use genome-wide oligonucleotide-based array comparative genome hybridization (aCGH) followed by rapid real-time quantitative PCR analysis to identify, confirm and map chromosomal anomalies in a cohort of 26 CDH+ patients. One hundred and five putative copy number changes were identified by aCGH in our cohort of CDH+ patients. Sixty-one of these changes (58%) had been previously described in normal controls. Twenty of the remaining 44 changes (45%) were confirmed by quantitative real-time PCR or standard cytogenetic techniques. These changes included de novo chromosomal abnormalities in five of the 26 patients (19%), two of whom had previously normal G-banded chromosome analyses. Data from these patients provide evidence for the existence of CDH-related genes on chromosomes 2q37, 6p22–25 and 14q, and refine the CDH minimal deleted region on 15q26 to an interval that contains COUP-TFII and only eight other known genes. Although COUP-TFII is likely to play a role in the development of CDH in patients with 15q26 deletions, we did not find COUP-TFII mutations in 73 CDH samples. We conclude that the combination of oligonucleotide-based aCGH and quantitative real-time PCR is an effective method of identifying, confirming and mapping clinically relevant copy number changes in patients with CDH+. This method is more sensitive than G-banded chromosome analysis and may find wide application in screening patients with congenital anomalies.
INTRODUCTION
Congenital diaphragmatic hernia (CDH) [Online Mendelian Inheritance in Man (OMIM) no. 142340] is a common birth defect—with an estimated incidence of one in 2500 births—and accounts for 8% of all major congenital anomalies (1,2). Mortality remains high among severely affected children and long-term complications in survivors are common (3–5). CDH can present as an isolated birth defect or in combination with other non-hernia-related anomalies (CDH+). Chromosomal anomalies are common in CDH+ and are typically screened for using G-banded chromosome analysis (6–8). Recent reports suggest that some causative chromosomal anomalies may go undetected in infants with CDH+ screened using standard cytogenetic techniques (9,10). These reports emphasize the need for screening methods with greater sensitivity since the detection of a chromosome anomaly can impact prognosis, the selection of additional screening tests, treatment decisions and recurrence risk counseling (7,11). Identifying cryptic chromosomal abnormalities in patients with CDH+ may also help to focus research efforts on regions of the genome that are likely to harbor one or more genes that play a role in the development of CDH.
Array comparative genomic hybridization (aCGH) is a powerful technique for identifying deletions and duplications in genomic DNA (12). Recently, regional BAC-based aCGH analysis has become available on a clinical basis (13,14). BAC-based arrays can provide improved resolution over G-banding but their maximal resolution is limited by the relatively large size of BAC clones (∼100–200 kb). Although oligonucleotide-based arrays—with probe sizes typically ranging from 45 to 80 bp—have the potential for much higher resolution, they have not been used extensively in the clinical setting (15).
One obstacle to the wider use of oligonucleotide-based arrays as a diagnostic tool is the need for an independent method for evaluating small copy number changes below the resolution of FISH. In this report, we describe the first use of a combination of high-resolution, genome-wide oligonucleotide aCGH and quantitative real-time PCR as a screening method for identifying, confirming and mapping copy number changes in patients with CDH+. Our results suggest that this combination of techniques is more sensitive than G-banded chromosome analysis and may find wide application in the screening for copy number variations in patients with congenital anomalies.
RESULTS
Twenty-six subjects with CDH were screened for chromosomal anomalies by aCGH. Three of the 26 patients (12%) had chromosome anomalies previously identified by G-banded chromosome analysis or FISH including Patient N4 whose de novo chromosomal anomalies have been published previously (Table 1). These chromosomal anomalies served as positive controls for the array analysis. One hundred and five putative copy number changes were identified in these patients including all changes previously detected by G-banding or FISH (Fig. 1).
Patient . | Sex . | CDH type . | Associated anomalies . | De novo changes detected by aCGH . | Previously identified by G-banding/FISH? . | Confirmed by . | Breakpoint positionsa . |
---|---|---|---|---|---|---|---|
TX19 | F | Left-sided posterolateral | Bilateral ptosis, dysmorphic facial features, microcephaly failure to thrive, developmental delay | Partial duplication of 6p | Yes; 46,XX inv dup(6)(p25.2p22.2) | qPCR, FISH | Tel 2,977,284–3,010,031 Cent 25,000,597–25,041,232 |
TX20 | F | Left-sided Posterolateral | Microtia, unilateral microphthalmia, corneal clouding, congenital cataract, cleft palate, ASD, thickened aortic and pulmonary valves, patent urachus, bilateral kidney malrotation, ulnar deviation of the hands, long thin fingernails, fifth finger clinodactyly | Duplication of 14q | No; Counting of additional cells after aCGH analysis revealed an isochromosome 14q in 2% of cells | FISH performed on EBV-transformed lymphocytes used as a DNA source for aCGH identified three 14q signals in 60/60 (100%) of cells | N/A |
TX21 | F | Right-sided Posterolateral | Ear anomalies, wide-spaced nipples, ASD | Deletion of X chromosome | Yes; 45,X | N/A | N/A |
N4 | M | Left-sided Posterolateral | Dysmorphic facial features, ventricular septal defect, micropenis, hypoplastic toe nails, loose skin | Partial duplication of 11q, partial deletion of 12q | Yes; 46,XY,der(12)t(11,12) (q23.3;q24.3) | FISH | Previously described (31) |
N9 | F | Left-sided Posterolateral | Double outlet right ventricle, hypoplastic nails | Partial duplication of 2q, partial deletion of 15q | No | qPCR, FISH | 2q 235,322,898–235,351,819 15q 92,219,239–92,231,673 |
Patient . | Sex . | CDH type . | Associated anomalies . | De novo changes detected by aCGH . | Previously identified by G-banding/FISH? . | Confirmed by . | Breakpoint positionsa . |
---|---|---|---|---|---|---|---|
TX19 | F | Left-sided posterolateral | Bilateral ptosis, dysmorphic facial features, microcephaly failure to thrive, developmental delay | Partial duplication of 6p | Yes; 46,XX inv dup(6)(p25.2p22.2) | qPCR, FISH | Tel 2,977,284–3,010,031 Cent 25,000,597–25,041,232 |
TX20 | F | Left-sided Posterolateral | Microtia, unilateral microphthalmia, corneal clouding, congenital cataract, cleft palate, ASD, thickened aortic and pulmonary valves, patent urachus, bilateral kidney malrotation, ulnar deviation of the hands, long thin fingernails, fifth finger clinodactyly | Duplication of 14q | No; Counting of additional cells after aCGH analysis revealed an isochromosome 14q in 2% of cells | FISH performed on EBV-transformed lymphocytes used as a DNA source for aCGH identified three 14q signals in 60/60 (100%) of cells | N/A |
TX21 | F | Right-sided Posterolateral | Ear anomalies, wide-spaced nipples, ASD | Deletion of X chromosome | Yes; 45,X | N/A | N/A |
N4 | M | Left-sided Posterolateral | Dysmorphic facial features, ventricular septal defect, micropenis, hypoplastic toe nails, loose skin | Partial duplication of 11q, partial deletion of 12q | Yes; 46,XY,der(12)t(11,12) (q23.3;q24.3) | FISH | Previously described (31) |
N9 | F | Left-sided Posterolateral | Double outlet right ventricle, hypoplastic nails | Partial duplication of 2q, partial deletion of 15q | No | qPCR, FISH | 2q 235,322,898–235,351,819 15q 92,219,239–92,231,673 |
aThe position of each breakpoint interval is given based on NCBI build 36.
Patient . | Sex . | CDH type . | Associated anomalies . | De novo changes detected by aCGH . | Previously identified by G-banding/FISH? . | Confirmed by . | Breakpoint positionsa . |
---|---|---|---|---|---|---|---|
TX19 | F | Left-sided posterolateral | Bilateral ptosis, dysmorphic facial features, microcephaly failure to thrive, developmental delay | Partial duplication of 6p | Yes; 46,XX inv dup(6)(p25.2p22.2) | qPCR, FISH | Tel 2,977,284–3,010,031 Cent 25,000,597–25,041,232 |
TX20 | F | Left-sided Posterolateral | Microtia, unilateral microphthalmia, corneal clouding, congenital cataract, cleft palate, ASD, thickened aortic and pulmonary valves, patent urachus, bilateral kidney malrotation, ulnar deviation of the hands, long thin fingernails, fifth finger clinodactyly | Duplication of 14q | No; Counting of additional cells after aCGH analysis revealed an isochromosome 14q in 2% of cells | FISH performed on EBV-transformed lymphocytes used as a DNA source for aCGH identified three 14q signals in 60/60 (100%) of cells | N/A |
TX21 | F | Right-sided Posterolateral | Ear anomalies, wide-spaced nipples, ASD | Deletion of X chromosome | Yes; 45,X | N/A | N/A |
N4 | M | Left-sided Posterolateral | Dysmorphic facial features, ventricular septal defect, micropenis, hypoplastic toe nails, loose skin | Partial duplication of 11q, partial deletion of 12q | Yes; 46,XY,der(12)t(11,12) (q23.3;q24.3) | FISH | Previously described (31) |
N9 | F | Left-sided Posterolateral | Double outlet right ventricle, hypoplastic nails | Partial duplication of 2q, partial deletion of 15q | No | qPCR, FISH | 2q 235,322,898–235,351,819 15q 92,219,239–92,231,673 |
Patient . | Sex . | CDH type . | Associated anomalies . | De novo changes detected by aCGH . | Previously identified by G-banding/FISH? . | Confirmed by . | Breakpoint positionsa . |
---|---|---|---|---|---|---|---|
TX19 | F | Left-sided posterolateral | Bilateral ptosis, dysmorphic facial features, microcephaly failure to thrive, developmental delay | Partial duplication of 6p | Yes; 46,XX inv dup(6)(p25.2p22.2) | qPCR, FISH | Tel 2,977,284–3,010,031 Cent 25,000,597–25,041,232 |
TX20 | F | Left-sided Posterolateral | Microtia, unilateral microphthalmia, corneal clouding, congenital cataract, cleft palate, ASD, thickened aortic and pulmonary valves, patent urachus, bilateral kidney malrotation, ulnar deviation of the hands, long thin fingernails, fifth finger clinodactyly | Duplication of 14q | No; Counting of additional cells after aCGH analysis revealed an isochromosome 14q in 2% of cells | FISH performed on EBV-transformed lymphocytes used as a DNA source for aCGH identified three 14q signals in 60/60 (100%) of cells | N/A |
TX21 | F | Right-sided Posterolateral | Ear anomalies, wide-spaced nipples, ASD | Deletion of X chromosome | Yes; 45,X | N/A | N/A |
N4 | M | Left-sided Posterolateral | Dysmorphic facial features, ventricular septal defect, micropenis, hypoplastic toe nails, loose skin | Partial duplication of 11q, partial deletion of 12q | Yes; 46,XY,der(12)t(11,12) (q23.3;q24.3) | FISH | Previously described (31) |
N9 | F | Left-sided Posterolateral | Double outlet right ventricle, hypoplastic nails | Partial duplication of 2q, partial deletion of 15q | No | qPCR, FISH | 2q 235,322,898–235,351,819 15q 92,219,239–92,231,673 |
aThe position of each breakpoint interval is given based on NCBI build 36.
To determine if these putative changes had been described previously in normal controls, we searched for similar deletion or duplications in the Database of Genomic Variants hosted by the Center for Applied Genomics. Of the 105 putative changes, 61 (58%) had been previously described in normal control individuals and were, therefore, unlikely to represent de novo CDH+ causing changes.
Of the 44 remaining changes, four (9%) were chromosomal anomalies previously identified by G-banded chromosome analysis and/or FISH (Table 1). To determine if the remaining 40 putative changes represented true deletions or duplication, we used quantitative real-time PCR as an independent evaluation of copy number (Fig. 2). Sixteen of the 40 putative changes (40%) were confirmed, 23 (58%) were found to be associated with a normal copy number (false positives) and one change on chromosome 22p11 (2%) could not be interrogated using this method due to high levels of homology to other chromosomes throughout the minimally deleted region.
As expected, the percentage of false positive results decreased with increasing numbers of putatively deleted or duplicated probes within a region of interest (Fig. 3).
To further determine if failure to confirm putative changes was due to a true lack of copy number change (false positive result on array) or a lack of sensitivity on the part of the quantitative PCR assay (false negative result on quantitative PCR), we performed a second round quantitative PCR using new PCR primer pairs on 10 of the false positive loci. In each case results from the second round of quantitative PCR were in agreement with the original results (false positive result on array). This suggests that although the sensitivity of quantitative real-time PCR might, in some cases, be improved using multiple primer pairs, the interrogation of most loci can be accurately completed using a single carefully selected primer pair.
When parental DNA samples were available, we used quantitative PCR or standard cytogenetic techniques to determine the inheritance pattern of the confirmed deletion or duplication (Fig. 2). We were able to determine the inheritance pattern of 13 of the 16 confirmed anomalies (81%) of which three (23%) represented de novo changes. These changes represent the likely etiology of CDH+ in two patients, both of whom had previously normal G-banded chromosome analyses.
Patient N9 was found to carry an unbalanced translocation resulting in a duplication of chromosome 2q and a deletion of chromosome 15q (Table 1). Patient TX20 was found to have a duplication of chromosome 14q by aCGH. Scoring of additional cells from the original chromosome analysis confirmed mosaicism for an isochromosome 14q in 2% of cells. However, FISH analysis performed on cells from the EBV-transformed lymphocyte culture used as a DNA source for the aCGH studies revealed that 100% of cells carried the isochromosome. This suggests a transformation or survival bias for cells containing the isochromosome 14q.
FISH is commonly used to define the boundaries of chromosomal anomalies identified by G-banded chromosome analysis. In an alternative method we used quantitative PCR in combination with aCGH data to rapidly map the location of the key breakpoints involved in the 6p duplication identified in Patient TX19, and the 2q duplication and 15q deletion identified in Patient N9 (Table 1, Fig. 4).
To determine if de novo mutations in the COUP-TFII are a common cause of CDH we screened 73 patient samples—35 with isolated CDH and 38 with non-isolated CDH—for COUP-TFII mutations. No sequence changes were identified within the coding sequence and the surrounding intronic splice donor and splice receptor sequences. Assuming an 80% mutation detection rate, this sample set would be sufficiently large to give a 95% probability of detecting a causative mutation if mutations in COUP-TFII were responsible for 5% of all CDH cases.
DISCUSSION
This study marks the first time that the combination of oligonucleotide-based aCGH and quantitative real-time PCR has been used to identify and confirm copy number changes responsible for non-isolated CDH. Using this combination of techniques, we identified a total of 81 copy number changes in our cohort of 26 CDH+ patients. These changes included de novo changes in five patients, two of whom had previous normal G-banded chromosome analyses. We conclude that the combination of oligonucleotide-based aCGH and quantitative real-time PCR is an effective method of identifying and independently confirming clinically relevant copy number changes in patients with CDH+. This combined method is more sensitive than G-banded chromosome analysis and may find wider application in screening for copy number variations in patients with congenital anomalies.
In addition, we have demonstrated that the combination of oligonucleotide-based aCGH and quantitative real-time PCR can be used to rapidly determine the location of the breakpoints surrounding a chromosomal copy number anomaly. As more is learned about the function of individual genes—and the biological consequences of altering their copy number—accurately defining the location of breakpoints will have increasing clinical importance and may ultimately allow counseling, therapy and surveillance to be individualized based on the genes and regulatory regions affected.
Mapping the boundaries of chromosomal anomalies can also be an important step towards identifying genes responsible for various clinical phenotypes. Recently, Klaassens et al. (16) defined a minimally deleted region for CDH on chromosome 15q26 using FISH and array CGH data from patients with CDH+. This interval was based on a CDH patient with an interstitial deletion of chromosome 15q26 and an individual without CDH who had a terminal deletion starting at 15q26.2 (Fig. 5). Of the genes in this region, COUP-TFII was considered a particularly strong candidate since it is regulated by the retinoid signaling pathway which has long been implicated in the development of CDH (17–19). Following this report, Castiglia et al. (20) presented data from several patients with relatively large 15q terminal deletions without CDH and suggested that defining the CDH minimally deleted region using data from patients without CDH could be unwise since haploinsufficiency of the CDH locus may be characterized by reduced penetrance. We agree that the most conservative approach to defining a CDH minimal region would be to use data only from patients with CDH. With this in mind, data from Patient N9 can be combined with data from Patient 8 presented by Klaassens et al. to define a new minimally deleted region on 15q26 that contains COUP-TFII and only eight other known genes (Fig. 5).
This interval could be further refined by a CDH patient carrying a ring chromosome 15 described by Tümer et al. (21). However, establishing a clear genotype/phenotype relationship in a ring chromosome carrier can be difficult owing to the potential instability of ring chromosomes which may result in the gain or loss of genetic material in different tissues (20,21). In either case, COUP-TFII would be located within the minimal deleted region and the proposed candidate genes CHD2 and RGMA would be located outside the interval (22). It is impossible, however, to exclude the possibility that the minimal deleted interval contains control elements required for the normal function of these genes.
Additional in vivo evidence of the role of COUP-TFII in the development of CDH comes from the targeted ablation of Coup-TFII in mice. You et al. (23) recently showed that homozygous ablation of Coup-TFII in the foregut mesenchyme results in a left-sided, posterolateral CDH similar to that seen in patients with 15q26 deletions. Taken together, these data strongly suggest that COUP-TFII plays a role in the development of CDH in individuals with 15q26 deletions.
As previously mentioned, CDH is not seen in all individuals with 15q26 deletions involving COUP-TFII (20). It is likely, therefore, that other genetic and/or environmental influences either raise or lower the threshold for CDH development in individuals with this deletion. Given that Patient N9 has both a 15q26 deletion and a 2q37 duplication, it is possible that over-expression of genes on 2q37 also played a role in the development of CDH in this patient. The hypothesis that over-expression of one or more genes on 2q37 may influence diaphragm development is supported by three other reported cases of CDH involving duplications of 2q37 (11,24,25).
Although it is likely that COUP-TFII plays an important role in the development of CDH in individuals with 15q26 deletions, we did not identify CDH-related mutations in the COUP-TFII coding region and surrounding splice donor/slice acceptor sequences in our screen of 73 CDH patients. Although this suggests that mutations in COUP-TFII are unlikely to be associated with >5% of all CDH cases, it is possible that mutations in this gene are more common in a specific subset of CDH patients such as those with heart defects or other anomalies seen in patients with 15q26 deletions.
Literature reviews have identified several other chromosomal regions that are recurrently duplicated or deleted in patients with non-isolated CDH and are likely to harbor one or more CDH-related genes (6–8). FOG2, for example, is located on chromosome 8q23.1—a region recurrently deleted in CDH patients—and has been shown by Ackerman et al. (26) to cause diaphragmatic eventrations in mice and humans. Data from our patients provide additional evidence for the existence of CGH-related regions on chromosomes 6p22–25 and 14q. Patient TX19 is the first CDH patient to be described with an isolated duplication of 6p but CDH has also been reported previously in a patient with partial trisomy 6p and partial trisomy 22 resulting from 3:1 meiotic disjunction of maternal (6p;22q) translocation (27). Although Patient TX20 is, to our knowledge, only the second CDH+ patient described with mosaic trisomy 14, partial duplications involving 14q32 have been previously described in two other patients with CDH (7,28,29). These data suggest that over-expression of one or more genes in these regions may predispose to the development of CDH.
MATERIALS AND METHODS
Subject accrual for aCGH studies
Seventy-five subjects with CDH were ascertained from patients seen at tertiary care centers in Houston, Texas and Rotterdam, The Netherlands in accordance with IRB-approved protocols. Thirty-five of the 75 patients (47%) were identified as having non-isolated CDH based on documentation of CDH and at least one other non-hernia-related anomaly. Lung hypoplasia, abnormalities in cardiac position, intestinal malrotation and patent ductus arteriosus are examples of hernia-related defects and were not considered as grounds for the diagnosis of non-isolated CDH. Twenty-six of the 35 patients with CDH+ (74%) were selected for further study based on the availability of a sufficient quantity of high-quality DNA (∼4 µg). Three of the 26 patients (12%) had chromosome anomalies that were previously identified by G-banded chromosome analysis or FISH and acted as positive controls.
Oligonucleotide-based aCGH
aCGH was performed using Human Genome CGH 44B Oligo Microarray Kits (Agilent) according to the manufacturer's protocol version 2.0. Arrays were scanned using an Agilent DNA Microarray Scanner. Data was extracted using Feature Extraction Software 8.1 (Agilent) and analyzed using CGHAnalytics3.2.32 Software (Agilent). Control DNA consisted of commercially available pooled control DNA (Promega) or DNA from a healthy male and female control with no personal or family history of CDH. These individual control samples proved particularly useful since each contained small, rare, copy number variations which served as internal positive controls.
Putative copy number changes were defined by intervals of two or more adjacent probes with log 2 ratios suggestive of a deletion or duplication when compared with the log 2 ratios of adjacent probes. These changes were identified with the assistance of the Aberration Detection Method 1 algorithm contained within the CGHAnalytics 3.2.32 Software.
Quantitative real-time PCR
Primer pairs for quantitative real-time PCR were designed from unique sequences within the minimal deleted or duplicated region of each putative copy number change using Oligo 6.0 software (Molecular Biology Insights). For regions <1 kb, primers were chosen from a region that included 5 kb of flanking sequence. Quantitative PCR confirmed array findings in each of these cases. The nucleotide–nucleotide BLAST algorithm at NCBI (http://www.ncbi.nlm.nih.gov/BLAST/) was used to confirm that each PCR amplification product was unique and did not contain sequences that were significantly similar to those contained in Alu repeats. The presence of primer dimmers and non-specific amplification products in PCR reactions was minimized by careful primer design and annealing temperature optimization. Quantitative PCR analyses were performed using a LightCycler 1.1 instrument in combination with LightCycler FastStart DNA Master SYBR Green I kits (Roche Molecular Diagnostics). Experiments were designed in a manner similar to the standard curve method described by Boehm et al. (30) with a region of the 3′-portion of the TCOF1 gene serving as a control locus. All experimental values were determined in triplicate. At the locus of interest, copy number values between 0.8 and 1.2 were considered normal while values ≥1.3 and ≤0.7 were considered evidence of duplication and deletion, respectively.
Identification of previously reported changes
To determine if a putative change identified by an aCGH had been described previously in normal controls, we searched for similar deletion or duplications in the Database of Genomic Variants hosted by the Center for Applied Genomics (http://projects.tcag.ca/variation/). To be considered as previously reported, a putative change had to be of the same type (deletion or duplication) and involve the same approximate interval.
COUP-TFII mutation screening
Seventy-three DNA samples from patients with CDH were screened for mutation in the coding sequence of COUP-TFII and the surrounding intronic splice donor and splice receptor sequences by direct sequencing of purified PCR amplification products. The sequences for primers used in this screen are available on request. Sequence changes in patient samples were identified by comparison with control DNA sequences using Sequencher 4.6 software (Gene Codes Corporation).
ACKNOWLEDGEMENTS
The authors would like to thank the patients and family members who cooperated in this study, Scott Vacha for his technical expertise, Svetlana Yatsenko and Pawel Stankiewicz for their help with FISH analyses and the Baylor Human Genome Sequencing Center for providing BAC clones used in FISH confirmation studies. This study was funded in part by the Sophia Foundation for Scientific Research, Rotterdam, The Netherlands (SSWO project 441), the Baylor College of Medicine Child Health Research Center (NIH, HD41648), NIH grant HD-050583 and the Howard Hughes Medical Institute.
Conflict of Interest statement. Authors have no financial interests or connections, direct or indirect or other situations that might raise the question of bias in this work.
REFERENCES
Author notes
The authors should be identified as having contributed equally.
- polymerase chain reaction
- congenital abnormality
- chromosome abnormality
- mutation
- chromosomes
- cytogenetic analysis
- genes
- genome
- oligonucleotides
- comparative genomic hybridization
- hernia, congenital diaphragmatic
- chromosome analysis
- quantitative real-time polymerase chain reaction
- array-based comparative genomic hybridization