Use of array comparative genomic hybridization for prenatal diagnosis of fetuses with sonographic anomalies and normal metaphase karyotype

Abstract

Objective

To prospectively study the addition of array comparative genomic hybridization (CGH) to the prenatal evaluation of fetal structural anomalies.

Methods

Pregnant women carrying fetuses with a major structural abnormality were recruited at the time of invasive procedure for chromosome analysis. Only women whose fetuses had a normal karyotype (n = 50) were subsequently evaluated by array CGH using one of two arrays (1887 clones covering 622 loci or subsequently 4685 clones covering 1500 loci).

Results

The mean gestational age of the fetuses was 24.5 weeks (range 11–38 weeks). The most prevalent anomalies were cardiac, central nervous system, skeletal, and urogenital. The median turnaround time for culturing and array CGH diagnosis was 18 days (range 2–72). Four of 50 fetuses had abnormal array results. One (2%) was clinically significant and three (6%) were inherited or benign variants.

Conclusions

Array CGH studies in fetuses with sonographic anomalies and normal metaphase karyotype detected clinically significant copy number alterations in 1 of 50 cases. This percentage (2%) is consistent with prior prenatal reports. Further studies are warranted to more precisely identify which fetal anomalies are associated with copy number alterations of clinical significance.

INTRODUCTION

Pregnant women who present with a sonographically detected fetal anomaly are frequently offered metaphase karyotype analysis, either by chorionic villus sampling (CVS) in the first trimester or by amniocentesis in the second or third trimester, if the anomaly is suspected to be associated with chromosomal aberrations. Prior studies demonstrate that metaphase karyotype provides a diagnostic explanation for the fetal anomaly in approximately 16–30% of cases, depending on the gestational age of the fetus and the presence of single or multiple anomalies (Wilson et al., 1992; Rizzo et al., 1996). If the structural anomaly is accompanied by a high clinical suspicion for a particular genetic condition, i.e. velocardiofacial syndrome in patients with cardiac anomalies, additional targeted fluorescence in situ hybridization (FISH) testing may be performed.

The application of array-based comparative genomic hybridization (CGH) analysis to fetuses with sonographic anomalies may improve the diagnosis of genetic abnormalities compared to metaphase karyotype alone. After elective termination or spontaneous miscarriage, Le Caignec et al. (2005) found that array CGH analysis of fetuses with multiple malformations identified genomic rearrangements, which had not been observed by karyotype analysis in ∼16% of cases. Approximately 10% were considered causative of the fetal malformation (Le Caignec et al., 2005). In the prenatal setting, the use of subtelomeric FISH probes was evaluated in a series of 48 fetuses with sonographically detected structural anomalies. Four per cent of fetuses had clinically significant rearrangements (Gignac et al., 2006). The use of molecular methods, such as subtelomere FISH, has been evaluated more extensively in pediatric populations in which subtelomeric chromosomal abnormalities account for approximately 5% of cases of idiopathic mental retardation (Flint et al., 1995; Knight et al., 1999; De Vries et al., 2003). More recently, Shaffer et al. (2008) found an 11.4% rate of clinically relevant cytogenetic abnormalities in 1,375 newborns and infants (age 0–84 days) who were tested by array CGH because of dysmorphic features, structural anomalies or the presence of developmental delay (Shaffer et al., 2008). The recently published results by Van den Veyver et al. (2008) revealed that 4/84 (4.8%) of fetuses with abnormal findings on prenatal ultrasound examination had abnormal array CGH findings. Of these four, two fetuses had aneuploidy (45,X and 45,X/46,XY). One fetus had a deletion of unclear significance and the remaining fetus had a 15q26.3 deletion that may have caused the sonographically detected malformation (Van den Veyver et al., 2008). In addition, Shaffer et al. (2008) found two clinically significant abnormalities in 151 prenatal cases tested by array CGH. The purpose of the present study was to evaluate the role of array CGH analysis in a more narrowly defined population, i.e. fetuses with sonographic abnormalities and a normal metaphase karyotype. Our hypothesis was that the addition of array CGH to the evaluation of fetuses with sonographically identified anomalies would yield additional diagnoses.

MATERIALS AND METHODS

Pregnant women were recruited at two institutions following a sonographically identified fetal abnormality in which further karyotype assessment was undertaken by CVS or amniocentesis. We included only patients carrying fetuses with significant structural malformations and/or intrauterine growth restriction (defined as an estimated fetal weight below the 10th percentile for gestational age) (Table 1). We excluded women whose fetuses had only ‘soft markers’ for aneuploidy, i.e. intracardiac echogenic foci or mild hydronephrosis. Fetuses with multiple anomalies were classified according to the most severe malformation present.

Table 1.

Fetal sonographic anomalies observed in the study population

Anomaly	BWH	TMC	Total
Cardiac	14	10	24
Pulmonic atresia	—	—	—
AV canal malformation	—	—	—
Hypoplastic left heart	—	—	—
Cardiac ventricle asymmetry	—	—	—
Ventricular septal defect	—	—	—
Double outlet right ventricle	—	—	—
Transposition of the great vessels	—	—	—
Aortic stenosis	—	—	—
CNS	4	2	6
Ventriculomegaly	—	—	—
Agenesis of the corpus callosum	—	—	—
Holoprosencephaly	—	—	—
Encephalocele	—	—	—
Polymicrogyria/macrocephaly	—	—	—
Lissencephaly	—	—	—
Skeletal	2	4	6
Unilateral cleft hand	—	—	—
Skeletal dysplasia	—	—	—
Scoliosis	—	—	—
Urogenital	2	2	4
Moderate bilateral hydronephrosis	—	—	—
Bladder outlet obstruction	—	—	—
Ambiguous genitalia	—	—	—
Echogenic kidneys	—	—	—
Cleft Lip/Palate	1	1	2
Cystic Hygroma/Enlarged NT/Hydrops	2	1	3
Growth disorder	0	3	3
Intrauterine growth restriction	—	—	—
Gastrointestinal	0	2	2
Cloacal anomaly	—	—	—
Congenital diaphragmatic hernia	—	—	—
Echogenic bowel	—	—	—
Multiple anomalies	7	10	17
Isolated	18	15	33
Total	25	25	50

CNS, central nervous system; NT, nuchal translucency; BWH, Brigham and Women’s Hospital; TMC, Tufts Medical Center

Fifty women whose fetuses had a normal metaphase karyotype were recruited. Of these, 47 women had amniocenteses and 3 had CVS. Women diagnosed with a fetal structural anomaly on ultrasound examination who opted for diagnostic amniocentesis for metaphase karyotype had an additional 5 cc of amniotic fluid drawn. Patients undergoing fetal cardiac therapy, i.e. in utero aortic valve dilation for evolving hypoplastic left heart syndrome or in utero atrial septostomy for established hypoplastic left heart, underwent amniocentesis at the time of the procedure and approximately 10 cc of amniotic fluid was collected. In the case of cystic hygromas, cultured chorionic villi were evaluated. Parental blood samples were obtained simultaneously when possible. Study participants were counseled that they would be informed if deletion and duplication syndromes were identified and also about clinically significant copy number variants (CNVs). Our approach to counseling patients about CNVs of unknown significance differed at the two institutions. At Tufts, the findings of CNVs without known clinical significance and those shared by one of the parents were not disclosed. At Brigham and Women’s Hospital, patients were counseled regarding all CNVs and their likely origin—de novo, inherited, or previously shown to be benign. Institutional review boards at both Tufts Medical Center and Brigham and Women’s Hospital approved these protocols.

Clinical data related to each fetus, including the fetal anomaly, gestational age at the time of invasive procedure, metaphase karyotype results, and array results were collected. Specific information related to processing requirements for each sample, including the duration of time from sample collection to array results, and whether amniotic fluid was directly evaluated or required cell culture, was collected.

Previous articles have described the method of array CGH in detail (Ward, 2006). The array platform used for the first 26 cases was the Signature Chip 4.0 (processed by Signature Genomic Laboratories, Spokane, WA). This array has 1887 clones covering 622 loci. The subsequent 24 cases were processed on the Signature Whole Genome (WG) chip, which includes 4685 clones targeted to over 1500 regions with an average gap of 1.6 Mb across the genome. The resolution on either microarray within the targeted regions is 80–100 kilobases. An analysis was done as previously described (Bejjani et al., 2005) and data were displayed using the software tool Genoglyphix® (http://www.signaturegenomics.com/genoglyphixFAQ.html).

RESULTS

Clinical and laboratory demographics

Samples were collected over a 19-month period from April 2007 until November 2008. The mean gestational age of the fetuses at the time they were evaluated was 24.5 weeks, with a range of 11–38 weeks. The most prevalent fetal anomalies were cardiac, central nervous system, skeletal and urogenital (Table 1). Of the total of 50 cases studied, 40% were from fetuses with multiple anomalies and 60% had isolated anomalies.

The turnaround time for array CGH diagnosis ranged from 2 to 72 days, with an average of 17 days. The median was 18 days. There was one outlier specimen that required a repeat amniocyte culture; this took 72 days. Direct DNA isolation was performed on 6/47 amniotic fluid specimens. The remainder of specimens (n = 41) was cultured either due to the need for additional testing at the referral center or due to maternal red blood cell contamination of the amniotic fluid (n = 31). The mean turnaround time for amniotic fluid specimens, which were evaluated directly, was nine days. In those requiring culture it was 18.7 days.

The mean gestational age at the time of procedure for specimens that had direct extraction of DNA from amniotic fluid was 27 weeks (SD = 5.1 weeks). The mean gestational age for specimens requiring culture was 25.5 weeks (SD = 6.6 weeks; p = 0.61). The average amniotic fluid volume submitted for analysis was 11.3 mL (range 5.0–48.0 mL). Specimens that could be directly evaluated had a mean volume of 16.2 mL (SD = 12.7 mL). Those requiring culture had a mean volume of 10.4 mL (SD = 11.2 mLp = 0.26). There was no statistically significant difference between samples directly evaluated and those requiring culture with respect to age or volume submitted.

Results of microarrays

In the 50 fetuses with normal metaphase karyotype, 4 had abnormal array results. Only one case (2%) had clinically significant results following array analysis. Two others had CNVs that differed from the reference sample, but were shared by one of the parents (cases 2 and 4). One additional case had a well-described benign CNV (case 3). These cases are as follows: Case 1, the one with the clinically significant abnormal array finding, was a fetus with cardiac ventricular asymmetry and pleural effusions. The fetus had a normal karyotype by CVS. The array subsequently performed on amniotic fluid at 36 weeks showed mosaic trisomy 22q. The extra chromosome was approximately 33.8 Mb in size (Figure 1) (microarray result: arr cgh 22(RP11-701M12->GS1-99K24)×3 and FISH result: ish 22q13.31(RP11-625J4×3)[11]/22q13.31(RP11-625J4×2)[19]. The neonatal peripheral blood karyotype was normal. Parental karyotypes and array studies were normal. Further information about the newborn is not available.

Figure 1.

Analysis of mosaic trisomy 22 identified by microarray and FISH. (A) Microarray plot showing a normal chromosome 22. Clones are ordered on the x axis according to physical mapping positions with proximal 22q to the left and distal 22q to the right. The blue line represents the ratios for each clone from the first experiment (control Cy5/patient Cy3), and the pink line represents the ratios for each clone obtained from the second experiment in which the dyes have been reversed (patient Cy5/control Cy3). (B) Microarray plot showing chromosome 22 in the fetus. BAC clones are ordered as in (A). Pink shading indicates a gain over all clones. (C) Metaphase FISH analysis using BAC clone RP11-625J4 to the 22q13.31 region showed two copies in 19 out of 30 cells. (D) Metaphase FISH showed three copies of BAC clone RP11-625J4 in 11 out of 30 cells. These results are consistent with mosaic trisomy 22

Case 2 was a 380 Kb deletion of 7q31.1 in a fetus with polymicrogyri and macrosomia. This same deletion was also seen in the phenotypically normal mother. The clinical significance of this finding is unclear.

Case 3 was a fetus with bilateral hydronephrosis and clubbed feet that was found to have a duplication of a single clone (NPHP1) on 2q13, a locus associated with nephronophthisis. This locus is associated only with medullary kidney dysfunction rather than a malformation of the collecting system, and is a well-described benign CNV (Baris et al., 2006).

Case 4 was a fetus with a hypoplastic left heart. The fetus had a 3-clone duplication of 4q28.2. This same 349 Kb duplication was present in a parent. In the absence of dysmorphic features in the parent, this finding was again suspected to be a benign variant.

All of the cases with variation from the reference genome were detected using the SignatureChipWG, although Cases 1 (mosaic 22q) and 3 (NPHP1 duplication) would have also been detected using the SignatureChip v4.0.

DISCUSSION

In this focused group of prospectively studied fetuses with major sonographic anomalies and a normal metaphase karyotype, 3/50 (6%) had CNVs, of which 2 were shared by one parent and 1 is a well-described benign variant. The study yielded additional genetic diagnostic information in 1/50 (2%) of subjects, although the clinical significance of this finding was unclear. Microdeletion of 22q13.3 is well-described (Phelan, 2008), but there are no prior reports of a clinical phenotype associated with duplication of this region. We speculate that this is a clinically relevant finding.

Although this study is small, it compares favorably with a recently published study (Van den Veyver et al., 2008) in which 300 prenatal cases were analyzed by array CGH for all indications, including sonographic abnormalities. This group found that in 1 out of 43 cases (2.3%) array CGH contributed clinically relevant new information. In addition, our results are similar to those previously reported by Shaffer et al. (2008), showing detection of 2 clinically significant findings in 151 prenatal cases (1.3%). Taken together, our studies and others suggest that with increasing numbers of study subjects the specific added value of array CGH in prenatal diagnosis is becoming clarified. Although we detected copy number alterations in 8% of cases, only 2% are clinically significant. The detection rate may have been higher if all cases with abnormal sonographic findings received microarray analysis rather than focusing on fetuses with major structural anomalies. These major anomalies may be more strongly associated with aneuploidy. As many genetic syndromes present postnatally with growth disorders, we hypothesize that perhaps evaluating a subset of fetuses with growth abnormalities and another anomaly may yield more microdeletions and duplications. In addition, we recognize that prenatal ultrasound examination has a limited ability to identify features that would be more readily apparent in the postnatal population, such as facial dysmorphia and abnormal mental development.

Of additional concern, although not previously explored in the medical literature to date, is whether or not the finding of an identical CNV in the parent is truly benign. Careful evaluation of parents carrying the same CNV as the fetus would be helpful in distinguishing true benign variants from disease-causing alterations.

Further collection of cases with other malformations and growth disorders will provide more robust data on the role of array CGH in the evaluation of other structural anomalies. The ongoing trial funded by the National Institute of Child Health and Human Development (NICHD), ‘Chromosomal Microarray Analysis for Prenatal Diagnosis: A Prospective Comparison with Conventional Cytogenetics’ (Wapner et al., 2008), will significantly increase numbers and statistical power. Studies like the one presented here and the NICHD-sponsored clinical trial will likely delineate the benefits and limitations of microarray analysis in pregnancies with or without fetal anomalies.