Rapid and novel prenatal molecular assay for detecting aneuploidies ...

PRENATAL DIAGNOSIS

Prenat Diagn (2011) Published online in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/pd.2674

Rapid and novel prenatal molecular assay for detecting aneuploidies and microdeletion syndromes Susan J. Gross1,2 *, Komal Bajaj2,3 , David Garry1 , Susan Klugman3,4 , Barry M. Karpel1 , Anne Marie Roe1,3 , Brian J. Wagner5 , Jenny Zhan2 , Stephen D. Apfelroth6 and Nicole Schreiber-Agus2 1

Department of Obstetrics and Gynecology, North Bronx Healthcare Network, Albert Einstein College of Medicine, Bronx, NY, USA 2 Human Genetics Laboratory at Jacobi Medical Center, North Bronx Healthcare Network, Bronx, NY, USA 3 Department of Obstetrics and Gynecology and Women’s Health, Montefiore Medical Center, Albert Einstein College of Medicine, Bronx, NY, USA 4 Department of Obstetrics and Gynecology, Bronx Lebanon Hospital Center, Bronx, NY, USA 5 North Shore—Long Island Jewish Healthcare System, New Hyde Park, NY USA; Albert Einstein College of Medicine, Bronx, NY, USA 6 Laboratory Medicine, North Bronx Healthcare Network, Albert Einstein College of Medicine, Bronx, NY, USA Objectives To develop a targeted aneuploidy and microdeletion detection platform for use in the prenatal setting, to assess the integrity of the platform with a robust validation system, and to prospectively determine the performance of the platform under routine clinical conditions. Methods To generate proxies for the various disorders assessed by the assay for analytical validation purposes, cells from ten microdeletion syndromes as well as from common aneuploidies were spiked into cleared amniotic fluid. Genomic DNA was isolated, labeled, and hybridized to microbeads that have been coupled to DNA derived from Bacterial Artificial Chromosome (BAC) from the relevant regions targeted by the array. Beads were read using a flow cytometric multiplex bead array detection system. In the prospective part of the study, 104 amniotic fluid samples were collected and analyzed. Results All microdeletion syndromes and aneuploidies were validated in a blinded fashion. In the prospective study, the total number of readable samples was 101 of 104 (97%). All sample results were confirmed independently. Conclusion The bead array approach is a rapid and reliable test for detecting aneuploidies and microdeletions. This assay has the potential to provide the benefit of expanded molecular cytogenetic testing to pregnant women undergoing invasive prenatal diagnosis. This approach may be especially useful in parts of the world where cytogenetic personnel and facilities may be limited. Copyright  2011 John Wiley & Sons, Ltd. Supporting information may be found in the online version of this article. KEY WORDS:

molecular cytogenetics; prenatal diagnosis

INTRODUCTION For several decades, the mainstay of prenatal diagnostics has remained the karyotype, obtained through an invasive procedure such as amniocentesis in the second trimester or chorionic villus sampling in the first trimester. Using standardized staining methodology and light microscopy, significant structural abnormalities such as inversions, translocations, duplications, and deletions can be identified, to a resolution of approximately 5 Mb (Gardner and Sutherland, 2004). However, while the value of identifying such abnormalities cannot be overemphasized, serious disorders such as microdeletion syndromes can result from smaller structural changes. Thus, the next breakthrough *Correspondence to: Susan J. Gross, Department of Obstetrics and Gynecology, Jacobi Medical Center, 1400 Pelham Parkway S., Building 1, BS 26, Bronx, NY 10461, USA. E-mail: [email protected]

Copyright  2011 John Wiley & Sons, Ltd.

was the advent of molecular technologies, specifically fluorescence in situ hybridization (FISH), quantitative fluorescence-polymerase chain reaction (QF-PCR), multiplex ligation-dependent probe amplification (MLPA), and array comparative genomic hybridization (array CGH). These technologies allow clinical laboratories to identify smaller duplications and deletions, with the potential advantages of rapid turnaround time and automation (Shaffer and Bui, 2007). However, despite the fact that we have had these molecular tools to diagnose significant microdeletion syndromes for several years, the vast majority of women undergoing an invasive procedure have not had access due to technical and/or cost issues (ACOG committee opinion, 2009). Additionally, these disorders may elude prenatal diagnosis because they may not have associated ultrasound manifestations or family history that would in turn prompt the molecular analysis. We conducted this study with the aim of designing an accessible, targeted molecular assay that could detect Received: 16 May 2010 Revised: 22 October 2010 Accepted: 24 October 2010

S. J. GROSS et al.

clinically significant microdeletion syndromes in addition to the common aneuploidies. In the first part of our study, we developed an approach to validate the platform and ensure accuracy of detection of the aneuploides as well as the selected microdeletions, which as a group are quite common, but are individually rare. Thereafter, we obtained fresh amniocentesis samples to determine DNA requirements and the reliability and robustness of the assay in actual clinical practice.

METHODS

Selection of disorders represented on the array Probes for chromosomes 13, 18, 21, X and Y were incorporated on the array, which is currently the baseline standard for all rapid molecular cytogenetic tests. Criteria for selection of the microdeletion syndromes (Table 1) included significant morbidity and mortality with known natural history and where the majority of cases are accounted for by well-characterized deletions. We also sought to include disorders where there was a significant chance that critical physical findings may be missed on detailed sonography. For purposes of confirmation of positive results, we made sure that clinical FISH probes

for these disorders were readily available. We generally included disorders that were relatively common such as 22q11.2 deletion syndrome. However, if a disorder met the above criteria (particularly severity) and was easily confirmed with accessible FISH probes, it was likewise included (such as Miller–Dieker syndrome). For the microdeletion syndromes, the probes on the array were usually localized to the critical region of the respective autosome (Appendix S1, Supporting information). Autosomal probes for regions that have no or few Online Mendelian Inheritance in Man (OMIM) references associated with them as well as negative control probes (the latter were 70 mers that have no significant homology in the human genome) were also included in the array for purposes of normalization and assay quality determination. All chromosomal probes (4–8 per disorder) comprised PCR-amplified, sequence-verified products (averaging 300–1000 bp in length) generated using degenerate primers off of BAC clones; these PCR products were then immobilized onto Luminex (Luminex Corporation, Austin, TX, USA) carboxy-polystyrene beads (reagents prepared and provided by PerkinElmer Wallac OY, Turku, Finland). Each bead, thus representing sequences from a single BAC, is distinguishable by its unique ‘color code’ when read by the Luminex analyzer. Eighty-five beads in total were used for this panel (Appendix S1). Each DNA sample tested was exposed to the entire set, in duplicate hybridization wells.

Table 1—Aneuploidy and microdeletion syndromes covered by the array (incidence refers to live births) Aneuploidies Trisomy 13; Patau syndrome Trisomy 18; Edwards syndrome Trisomy 21; Down syndrome Sex chromosome abnormalities Microdeletion syndromes 22q11.2 deletion syndrome DiGeorge syndrome 2 (10p13-p14 deletion syndrome) Williams syndrome Prader–Willi syndrome Angelman syndrome Smith–Magenis syndrome Wolf–Hirschhorn syndrome Cri du Chat syndrome Langer–Giedion syndrome Miller–Dieker syndrome

Chromosomal region involved/critical region

Incidence 1/22 700a 1/7500a 1/830a 1/400–1/650a

Whole Whole Whole Whole

chromosome chromosome chromosome chromosome

13 18 21 X or Y

Incidence

∼% cases caused by deletion

Chromosomal region involved/critical region

1/2000–4000a

93–95

1.5–3.0 Mb on 22q11.2

1/200 000b

Unknown

5 Mb 10p13 pter

1/7500c 1/10 000–1/30 000c 1/12 000–1/20 000c 1/25 000c

>95–99 70–75 68 >90

1.5–1.8 Mb on 7q 11.23 15q11.2-q13 (paternal) 15q11.2-q13 (maternal) 17p11.2

1/50 000c

>90

4p16.3 region

1/20 000–1/50 000d

>99

Unknowne

>99

Unknownf

>90

Varies, only 5p15.2 band to entire 5p arm Chromosome 8, commonly involving 8q23.2-8q24.1 17p13.3

a Nussbaum et al. (2007). b Dasouki et al. (1997). c http://www.genetests.org. d http://www.ncbi.nlm.nih.gov/entrez/dispomim.cgi?id=123450. e http://www.ncbi.nlm.nih.gov/entrez/dispomim.cgi?id=150230. f Kuwano et al., 1991.


Prenat Diagn (2011) DOI: 10.1002/pd

PRENATAL ANEUPLOIDY AND MICRODELETION ASSAY

Sample procurement

Assay overview and data analysis

Part A: Aneuploidy/microdeletion validation

DNA isolation

Currently, both American and European standards for validation of molecular cytogenetic testing for constitutional chromosomal abnormalities require the use of reference DNA (http://www.acmg.net/AM/Template.cfm? Section=Laboratory Standards and Guidelines& Template=/CM/ContentDisplay.cfm&ContentID=2509; http://www.eurogentest.org/web/files/public/unit1/J.%20 Vermeesch.pdf). Although these guidelines were written for microarray analysis, they are in keeping with general standards for using stored DNA samples for validations studies, with The American College of Medical Genetics advocating for the ‘use of at least one reference DNA with a known abnormality for each disorder’. For our novel, molecular-based approach, we went one step further by creating a ‘proxy’ model validation system that would more closely emulate in vivo amniotic fluid samples. Rather than starting from the point where DNA has already been extracted, cell lines representing the syndromes covered by the array were obtained from the Coriell Institute for Medical Research and cultured according to the recommended conditions, and then 200 000 cells were added to 2 cm3 of cleared amniotic fluid supernatant and stored at room temperature overnight. To ensure that the fluid was indeed ‘cleared’, it was run through the labeling process (see below), and there was no residual labeling detected, suggesting that there were minimal or no residual cells/cell-free DNA present. The specific number of cells to be spiked in was chosen based on previous reports that determined the number of cells in the amniotic fluid of varying gestational ages (Wahlström, 1974; Elejalde et al., 1990). We also chose to use more than one reference sample, using three cell lines for all syndromes when available.

Part B: Prospective performance assessment of assay in routine clinical setting This part of the study was conducted at five prenatal diagnostic centers in the greater New York City area between April and June 2009. Subjects were recruited from women who had already consented and agreed to undergo amniocentesis for indications commensurate with standard of care. Following appropriate recruitment and consent procedures for this Institutional Review Board (IRB)-approved study, the women had an additional 5 cm3 of amniotic fluid aspirated. Data collection included gestational age, indication for testing, and fetal karyotype. Standard cytogenetic studies were performed in a Clinical Laboratory Improvement Amendments (CLIA)-approved cytogenetics laboratory as per the study subject’s provider, and management was independent of this study. All amniotic fluid samples were transported to the Human Genetics Laboratory at Jacobi Medical Center in sterile tubes at room temperature. Copyright  2011 John Wiley & Sons, Ltd.

DNA was isolated with the QIAamp DNA Blood Mini Kit (QIAgen 51106).

Validation study: Cell pellets were made by spinning the 2 mL spiked samples in microcentrifuge tubes for 4 min at 15.7 k rcf. Pellets were resuspended in 200 µL of 1× PBS, and 20 µL of proteinase K and 200 µL of AL buffer were added. After a 10-min incubation at 56 ◦ C, 200 µL of 100% ethanol was added and the sample was transferred to a QIAamp spin column. Samples were washed with 500 µL of AW1 buffer followed by 500 µL of AW2 buffer, and elution was with 100 µL of AE buffer. Prospective study: DNA was isolated as per above with the following modifications: the initial spin was performed in a 15 mL conical at 1320 rcf for 10 min, and the elution was performed with 50 µL of AE buffer. DNA from these amniotic fluid samples was made between day 1 through day 5 after the amniocentesis. These differences in waiting time did not significantly affect the DNA yield (average total yield is around 1.1 µg/5 mL sample) or the sample’s performance in the assay. Amniotic fluid DNA was stored at −20 ◦ C. On the day of the assay, the optical density was determined.

Assay methodology For the assay, 125 ng of genomic DNA was biotindCTP-labeled, purified, and then hybridized in duplicate wells overnight to the BAC-containing Luminex beads (reagents obtained from PerkinElmer Wallac OY). Normal male and female DNAs (Promega, Madison, WI, USA) were also labeled and hybridized as references (total of four wells of each reference per assay). After a series of washes before and after streptavidin–phycoerythrin reporter incubation, the beads were read using the Luminex xMAP system. A more detailed protocol can be found in Appendix S2 (Supporting information).

Data analysis Luminex output files of median fluorescence intensities were imported into an Excel-based program. Four categories of beads are included in the bead set: negative controls (sequences without known homology to the human genome), autosomal controls, syndrome probes (microdeletions and common aneuploidies), and sex chromosome probes. Normalization occurs in two steps—subtraction of signals against negative controls to account for background noise, followed by autosomal controls, where the ratio is set at ‘1’. After these normalization steps, ratios are calculated across all probes Prenat Diagn (2011) DOI: 10.1002/pd

S. J. GROSS et al.

for all syndromes for each sample over the male reference and for each sample over the female reference; these references are assayed in the same runs as the test samples, in their own wells. An initial ‘flag’ or threshold is determined based on median ratios for specific syndromes deviating beyond 2 standard deviations (SDs). These probe signals are then excluded to determine a baseline ‘trimmed’ SD [SD same as coefficient of variation (CV) for normalized data]. Each clone is now assessed using a threshold whereby a ratio of n–1 or greater clones of the syndrome that deflect beyond the trimmed SD × 2 is considered a positive call for nonsex chromosome syndromes. For syndromes involving the X or Y chromosome, significant deflections entailed that the ratios of n–1 or greater clones of the syndrome were outside trimmed SD × 3 (as the ratio responses are larger, the threshold needs to be greater). The ‘n–1 or greater’ rule was based on results obtained from the known syndromic cell lines (Appendix S3, Supporting information). In addition to calling significant deflections for the syndromes, the gender of the fetus was also determined. Sample acceptance criteria required a trimmed CV of 6% or less. Run acceptance criteria required a signal/noise ratio for the reference controls of 3.0 or greater; the signal to noise ratio represents the average autosomal probe signals for these controls after subtracting the average negative values/average negative values.

RESULTS

Aneuploidy/microdeletion validation All microdeletion syndromes and aneuploidies were validated on DNA from cells (Coriell) representing the various syndromes that had been spiked into cleared amniotic fluid. These spiked samples, which represent ‘proxies’ for the syndromes, were run in a blinded fashion. Upon analysis, all were significantly positive only for the gains/losses expected based on the information provided for the respective cell lines in the Coriell repository database (Appendix S3). In addition to the perfect correlation with respect to the expected aneuploidies/microdeletions (i.e. lack of false negatives and also false positives), all of the genders of the spiked cell samples as determined by the assay were identical to the genders reported for the cell lines in the Coriell database. Representative graphs of results for a spiked sample serving as a proxy for 22q11.2 deletion syndrome (Coriell sample GM13325) are shown in Figure 1 (additional representative graphs of other spiked samples are shown in Appendix S6 (Supporting information). The top graph shows the ratios of sample/female reference (pink or lighter line) and sample/male reference (blue or darker line) for the various probes covering the syndromes. The bottom graph shows the median ratios of the sample/female reference (pink or lighter bar) or the sample/male reference (blue or darker bar) for the various syndromes. As all five probes for this DGS1/22q11.2 region deviate greater than 2 SDs, this would be called Copyright  2011 John Wiley & Sons, Ltd.

a ‘positive result for 22q11.2 deletion syndrome’. Note that the representative sample is a female, because when compared to a female it has the same amount of X signal but when compared to a male it has more X; it also has less Y than the male reference.

Prospective performance assessment of assay in routine clinical setting A total of 104 samples were obtained (Appendix S4, Supporting information) from women undergoing clinically indicated amniocentesis under an IRB-approved protocol. Five samples were deemed ‘no calls’ as they failed to meet sample acceptance criteria by giving trimmed CVs above our cut-off of 6% maximum (marked in bold in Appendix S5, Supporting information; Trimmed CV column). Two of these five samples (nos. 15 and 28) were repeated and ran successfully on the repeat (marked in bold in Appendix S5; sample no. column). Three of the five samples were not repeated due to insufficient DNA amounts (nos. 18, 47, and 54; marked as ‘fail’ in Appendix S4). In sum, this brings the total number of readable samples up to 101 of 104, for a success rate of 97%. Results from the runs were compared to clinical cytogenetic results obtained through chromosome analysis that was performed as part of the subject’s routine clinical care in an independent laboratory. For the 101 readable samples, all of the assay calls for gender of the fetus correlated with the gender as determined by karyotyping. Ninety-six samples were normal by our assay (syndrome median ratios across all syndromes of ∼1.0) and by karyotype (representative normal sample no. 96 shown in Figure 2A, and also refer to Appendices S4 and S5). The prospective sample set was not specifically intended to demonstrate ‘positive’ specimens, but two aneuploidies were captured, likely because the cohort included women referred due to high risk findings on ultrasound or serum screening. These two samples that were trisomic by our assay (no. 68, trisomy 18, and no. 97, trisomy 21; syndrome ratio for the trisomy 18 sample was 1.36, and for the trisomy 21 sample was 1.35; Appendix S5) were confirmed to be so by chromosome analysis (refer to representative graph for trisomy 21 sample no. 97 in Figure 2B). In each of the trisomic cases, all five probes for the specific aneuploidy region on the array were significantly deflected, thus fulfilling our criteria for designation of ‘positive result’ (n–1 or greater rule). Finally, three samples that were called normal by our assay were shown to harbor inversions/balanced translocations on final karyotype (nos. 74, 80, and 103; Appendix S4); such chromosomal alterations would be undetectable in this assay. DISCUSSION In this study, based on current standards, we have analytically validated a new platform for the prenatal detection of the major aneuploidies found at birth as well as of several significant microdeletion syndromes, and also tested Prenat Diagn (2011) DOI: 10.1002/pd

PRENATAL ANEUPLOIDY AND MICRODELETION ASSAY A

Cell line GM13325 2

1.8 1.6 1.4 1.2 1 0.8 0.6 22q11.2 Region

0.4 0.2 0 13

18

21

AUTO

CDC

DGS1 DGS2

Sample/female reference

B

LGS

MDS

PWS

SMS

WS

WHS

X

Y

Sample/male reference

Cell line GM13325 2

1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 13

18

21

AUTO CDC DGS1 DGS2 LGS Sample/female reference

MDS PWS SMS

WS

WHS

X

Y


Figure 1—22q11.2 deletion syndrome proxy sample. Representative graphs of results for a spiked sample serving as a proxy for 22q11.2 deletion syndrome (made from Coriell cell line GM13325) are shown (additional representative graphs for the other syndromes on the panel are shown in Appendix S6). In panel A, the ratios of sample/female and sample/male for the various probes covering the syndromes are graphed. In panel B, median ratios of the sample/female or the sample/male for the various syndromes are graphed. As all five probes for this DGS1/22q11.2 region deviate greater than 2 standard deviations (SDs), this would be called a positive result for 22q11.2 deletion syndrome. Also note that the sample is female. AUTO, autosomal; DGS1, DiGeorge syndrome 1 (22q11.2 syndrome); DGS2, DiGeorge syndrome 2; WS, Williams syndrome; PWS, Prader–Willi syndrome; AS, Angelman syndrome; SMS, Smith–Magenis syndrome; WHS, Wolf–Hirschhorn syndrome; CDC, Cri du Chat syndrome; LGS, Langer–Giedion syndrome; MDS, Miller–Dieker syndrome

its performance in actual clinical practice. Currently, the three major molecular cytogenetic technologies in use worldwide in the prenatal diagnostic setting are FISH, QF-PCR, and MLPA. FISH has played a dual role in this setting—(1) it is currently used to provide a rapid result for the common aneuploidies (involving chromosomes 13, 18, 21, X and Y) (Aviram-Goldring et al., 1999; Bink et al., 2000; Luquet et al., 2002; Locatelli et al., 2005) primarily in North America, and (2) individual FISH probes are used for diagnostic purposes when a particular microdeletion syndrome is suspected (Kuwano et al., 1991). QF-PCR and MLPA are more commonly used outside the United States and both have the advantage of not requiring time-consuming cell culture. QF-PCR is used for rapid results and can be performed at very low cost (Verma et al., 1998; Bili et al., 2002; Cirigliano et al., 2006; Ochshorn et al., 2006). Because this method has been so successful, serious consideration has been Copyright  2011 John Wiley & Sons, Ltd.

given to using it as the test of choice (Caine et al., 2005; Leung and Lao, 2005; Ogilvie et al., 2005). However, QF-PCR is limited to the common aneuploidies and as such may miss other potentially severe cytogenetic abnormalities. MLPA has also been used for rapid testing and is available for microdeletion syndromes as well. Nevertheless, to date it has been used primarily for the detection of the common aneuploidies (Slater et al., 2003; Hochstebach et al., 2005; Boormans et al., 2010). To address this common shortcoming of the above methodologies where detection is essentially limited to major aneuploidies, array CGH is being advocated as a solution. Array CGH has the ability to identify very small changes by comparing normal with test DNA. This technology is proving valuable in the pediatric setting wherein children with birth defects and/or developmental delay are found to indeed have fine chromosomal Prenat Diagn (2011) DOI: 10.1002/pd

S. J. GROSS et al. Sample #96

A 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 13

18

21

AUTO CDC DGS1 DGS2 LGS MDS PWS SMS Sample/female reference

WS

WHS

X

Y

WS

WHS

X

Y


Sample #97

B 2 1.8 1.6 1.4 1.2 1 0.8 0.6

Chr 21 Region

0.4 0.2 0

13

18

21

AUTO CDC DGS1 DGS2 LGS MDS PWS SMS Sample/female reference


Figure 2—Normal male and trisomy 21 female samples. Representative graphs of results for a prospective normal male sample (no. 96; panel A) and a female sample that tested positive for trisomy 21 (no. 97; panel B). For sample no. 97, as all five probes for the Chr21 region deviate greater than 2 standard deviations (SDs), this would be called a positive result for trisomy 21 syndrome. This result was confirmed by routine karyotyping

abnormalities that previously were beyond the resolution of standard karyotyping (de Vries et al., 2005). Studies have reported the use of array CGH in the prenatal setting, with the majority of these samples coming from fetuses with birth defects (Le Caignec et al., 2005; Sahoo et al., 2006; Shaffer et al., 2008; Coppinger et al., 2009; Kleeman et al., 2009; Van den Veyver et al., 2009). However, the use of this technology in the prenatal setting remains controversial (Pergament, 2007; ACOG committee opinion, 2009; Choy et al., 2010). Although these arrays have the ability to detect chromosomal structural changes at a very fine level and thus potentially can uncover associated syndromes and disorders, they may also identify copy number variants (CNVs) of uncertain significance, posing a very difficult situation for the clinician and for the parents. In addition, in the case of CNVs, a sample from the father is Copyright  2011 John Wiley & Sons, Ltd.

required for further work-up, thus complicating management in certain cases. Furthermore, at the present time, the associated cost is still considerable and potentially limiting in a setting of limited health resources. Also, validation and quality assurance are more complex to accomplish due to the large number of disorders and loci. As a result, while this technology may be beneficial in the setting of fetal anomalies, it has not been adopted as a routine approach suitable for all women undergoing invasive testing at this time and awaits further study (ACOG committee opinion, 2009). The targeted, bead-based molecular approach described here addresses multiple concerns associated with the currently available technologies. We have included microdeletion disorders that were selected on the basis of their having natural histories that are well known to the genetic community, thereby making counseling Prenat Diagn (2011) DOI: 10.1002/pd

PRENATAL ANEUPLOIDY AND MICRODELETION ASSAY

straightforward and unambiguous. Additionally, features of many of these disorders may be missed or not present on prenatal ultrasound. From a laboratory perspective, we specifically selected syndromes for which FISH probes are readily available for confirmatory purposes, with the benefit of not requiring any additional sample from the putative father. Such an approach would fulfill our aim of developing a laboratory test that could easily be integrated into a common practice and routine use. Although monosomy 1p36 is the most common terminal deletion syndrome, this syndrome was not selected for the panel as there is significant complexity in that region such that a positive result may not be straightforward, and confirmation would likely require expert follow-up to delineate precise breakpoints and consequent prognosis. This being said, the bead array system is entirely flexible, and disorders could be added or removed based on local public health needs as well as on availability of requisite expertise in interpretation. Along this line, while for the current study we chose a Luminex system that allowed for multiplexing of a maximum of 100 analytes (microspheres)/sample well, there is also a Luminex system available that allows for discrimination of up to 500 different microspheres. Other particular benefits include that Luminex technology has been used for many years very effectively in clinical serologic assays and could therefore be adoptable by the laboratory community. This ready availability of the bead array platform may be especially advantageous in regions of the world where trained cytogeneticists are limited. Our study also sought to address the very legitimate concerns of the lay and professional community with respect to validation. All of the disorders on our array have been validated using a ‘proxy amniocentesis’ system, which could serve as a valuable approach as we grapple with new molecular technologies. Prospective validation studies for several even relatively common microdeletion syndromes would require tens of thousands of collected amniocentesis specimens in order to achieve an appropriate sample size, rendering such studies difficult; this may also delay technology from reaching the public in a timely fashion. To address this problem, stored DNA samples that are at least one step removed from the actual collected specimen are used to validate relatively rare disorders. Although our ‘proxy’ system is not precisely the same as assaying a fresh amniocentesis sample, it has allowed us to not only meet current professional standards but also to more thoroughly assess factors related to DNA extraction and storage that may affect assay performance. In consideration of the fact that indeterminate results in the prenatal setting could lead to the termination of a normal fetus, we have set very rigorous criteria when calling a sample ‘positive’, and accordingly there may be false negatives. In that vein, the assay should be able to detect changes, such as duplications in microdeletion regions (such as in the 22q11.2 region) or deflections in n–2 or fewer probes for a given region, that may also be associated with abnormal outcomes. These types of variations will not be reported out by us until there is further validation of their performance in the assay and Copyright  2011 John Wiley & Sons, Ltd.

their significance to the fetus (this will require larger prospective studies that could also assess the sensitivity and specificity of the assay in low risk situations). Furthermore, a negative result does not fully exclude the diagnosis of any of these syndromes, as other etiological bases may exist for these disorders that are not covered in the assay design, such as point mutations. Also, the assay has the same limitations as the other quantitative molecular approaches, including inability to detect balanced translocations or inversions. Our overall approach is based on a mindset of ‘first do no harm’, and reflects the fact that the test is designed for all women undergoing invasive testing, in particular those with normal appearing fetuses. It is thus in agreement with the European guidelines for molecular karyotyping by microarray which state that ‘the development of arrays with a limited number of well known targets’ may be a solution to the prenatal scenarios where there is an absence of obvious ultrasonographic fetal abnormalities (http://www.eurogentest.org/web/files/public/unit1/J.% 20Vermeesch.pdf). Finally, while the platform does not have the broad coverage of BAC or oligonucleotide arrays, the assay does provide rapid (results can be obtained the day after DNA is prepared), accurate testing, like FISH and QF-PCR, while yielding additional important, validated information without any additional risk to the fetus. Although we do not address or calculate costeffectiveness directly, the National Institute for Health Research (http://www.hta.ac.uk/execsumm/summ710. shtml) (Grimshaw et al., 2003) determined that incorporating molecular technologies, specifically FISH or QF-PCR, in a rational way has the potential to save significant health care dollars in a public health-based system. We anticipate the cost of this assay to be perhaps slightly more than some technologies that offer fewer disorders (such as QF-PCR), but less than other technologies that have the ability to detect more alterations (e.g. array CGH) but that have an associated increase in difficulty of interpretation. Larger comparative studies will be needed to address issues of clinical utility and cost advantage and could also determine the relative performance in terms of the minimum amount of DNA needed, the threshold for detecting mosaicism, among other variables. CONCLUSION The bead array approach is a rapid and reliable test for detecting aneuploidies and microdeletions. This assay has the potential to provide the benefit of expanded molecular cytogenetic testing to pregnant women undergoing invasive prenatal diagnosis; it also may be applicable to those having chorionic villus sampling according to preliminary data (Grati et al., 2010). The methodology we describe may be especially useful in parts of the world where cytogenetic personnel and facilities may be limited. If this approach is adopted on a broad scale, including for non-anomalous fetuses, it would help uncover the actual incidence of the microdeletion disorders, as many of these children now die without a Prenat Diagn (2011) DOI: 10.1002/pd

S. J. GROSS et al.

diagnosis. An international consortium employing this approach could ultimately provide such data, generating future avenues for research into underlying mechanisms, and perhaps eventually therapies as well. ACKNOWLEDGEMENTS

The authors would like to thank Howard Nadel of the North Bronx Healthcare Network, Dr Michele Caggana, Wadsworth Center, Albany, NY, and Dr Aleksandr Fuks and Sara Delgado, Bronx Lebanon Hospital Center. Dr Gross reports receiving grant support from PerkinElmer Inc and Genzyme Corp. This study was supported by a research grant provided by PerkinElmer Inc. REFERENCES ACOG Committee Opinion No 446. 2009. Array comparative genomic hybridization in prenatal diagnosis. American College of Obstetricians and Gynecologists. Obstet Gynecol 114(5): 1161–1163. Aviram-Goldring A, Daniely M, Dorf H, et al. 1999. Use of interphase fluorescence in situ hybridization in third trimester fetuses with anomalies and growth retardation. Am J Med Genet 87(3): 203–206. Bili C, Divane A, Apessos A, et al. 2002. Prenatal diagnosis of common aneuploidies using quantitative fluorescent PCR. Prenat Diagn 22(5): 360–365. Bink K, Pauer HU, Bartels I. 2000. Interphase FISH test as a rapid test for trisomies in amniotic fluid—results of a prospective study. Z Geburtshilfe Neonatol 204(1): 8–13. Boormans EM, Birnie E, Oepkes D, et al. 2010. MLP and karyotyping evaluation (M.A.K.E.) study group. Comparison of multiplex ligationdependent probe amplification and karyotyping in prenatal diagnosis. Obstet Gynecol 115(2 Pt 1): 297–303. Caine A, Maltby AE, Parkin CA, et al. 2005. Prenatal detection of Down’s syndrome by rapid aneuploidy testing for chromosomes 13, 18, and 21 by FISH or PCR without a full karyotype: a cytogenetic risk assessment. Lancet 366(9480): 123–128. Choy K, Setlur S, Lee C, Lau T. 2010. The impact of human copy number variation on a new era of genetic testing. BJOG 117(4): 391–398. Cirigliano V, Voglino G, Marongiu A, et al. 2006. Rapid prenatal diagnosis by QF-PCR: evaluation of 30,000 consecutive clinical samples and future applications. Ann N Y Acad Sci 1075: 288–298. Coppinger J, Alliman S, Lamb AN, et al. 2009. Whole-genome microarray analysis in prenatal specimens identifies clinically significant chromosome alterations without increase in results of unclear significance compared to targeted microarray. Prenat Diagn 29(12): 1156–1166. Dasouki M, Jurecic V, Phillips JA III, et al. 1997. DiGeorge anomaly and chromosome 10p deletions: one or two loci? Am J Med Genet 73(1): 72–75. de Vries BB, Pfundt R, Leisink M, et al. 2005. Diagnostic genome profiling in mental retardation. Am J Hum Genet 77(4): 606–616. Elejalde BR, de Elejalde MM, Acuña JM, et al. 1990. Prospective study of amniocentesis performed between weeks 9 and 16 of gestation: its feasibility, risks, complications and use in early genetic prenatal diagnosis. Am J Med Genet 35(2): 188–196.


Gardner RJM, Sutherland GR. 2004. Chromosome Abnormalities and Genetic Counseling (3rd edn). Oxford University Press: New York; 11. Grati FR, De Toffel S, Marcato L, et al. 2010. BACS on BEADS for rapid detection of fetal microdeletionis and aneuploidies: validation study on prenatal samples. ISPD Conference (Amsterdam), Abstract 7–2. Grimshaw GM, Szczepura A, Hultén M, et al. 2003. Evaluation of molecular tests for prenatal diagnosis of chromosome abnormalities. Health Technol Assess 7(10): 1–146. Hochstenbach R, Meijer J, van de Brug J, et al. 2005. Rapid detection of chromosomal aneuploidies in uncultured amniocytes by multiplex ligationdependent probe amplification (MLPA). Prenat Diagn 25(11): 1032–1039. Kleeman L, Bianchi D, Shaffer LG, et al. 2009. Use of array comparative genomic hybridization for prenatal diagnosis of fetuses with sonographic anomalies and normal metaphase karyotype. Prenat Diagn 29(13): 1213–1217. Kuwano A, Ledbetter SA, Dobyns WB, et al. 1991. Detection of deletions and cryptic translocations in Miller-Dieker syndrome by in situ hybridization. Am J Hum Genet 49(4): 707–714. Le Caignec C, Boceno M, Saugier-Veber P, et al. 2005. Detection of genomic imbalances by array based comparative genomic hybridisation in fetuses with multiple malformations. J Med Genet 42(2): 121–128. Leung WC, Lao TT. 2005. Rapid aneuploidy testing, traditional karyotyping, or both? Lancet 366(9480): 97–98. Locatelli A, Mariani S, Ciriello E, et al. 2005. Role of FISH on uncultured amniocytes for the diagnosis of aneuploidies in the presence of fetal anomalies. Fetal Diagn Ther 20(1): 1–4. Luquet I, Mugneret F, Athis PD, et al. 2002. French multi-centric study of 2000 amniotic fluid interphase FISH analyses from high-risk pregnancies and review of the literature. Ann Genet 45(2): 77–88. Nussbaum RL, McInnes RR, Willard HF. 2007. Thompson and Thompson Genetics in Medicine (7th edn). Oxford Saunders: Philadelphia. Ochshorn Y, Bar-Shira A, Jonish A, Yaron Y. 2006. Rapid prenatal diagnosis of aneuploidy for chromosomes 21, 18, 13, and X by quantitative fluorescence polymerase chain reaction. Fetal Diagn Ther 21(4): 326–331. Ogilvie CM, Lashwood A, Chitty L, et al. 2005. The future of prenatal diagnosis: rapid testing or full karyotype? An audit of chromosome abnormalities and pregnancy outcomes for women referred for Down’s syndrome testing. BJOG 112(10): 1369–1375. Pergament E. 2007. Controversies and challenges of array comparative genomic hybridization in prenatal genetic diagnosis. Genet Med 9(9): 596–599. Sahoo T, Cheung SW, Ward P, et al. 2006. Prenatal diagnosis of chromosomal abnormalities using array-based comparative genomic hybridization. Genet Med 8(11): 719–727. Shaffer LG, Bui TH. 2007. Molecular cytogenetic and rapid aneuploidy detection methods in prenatal diagnosis. Am J Med Genet C Semin Med Genet 145C(1): 87–98. Shaffer LG, Coppinger J, Alliman S, et al. 2008. Comparison of microarraybased detection rates for cytogenetic abnormalities in prenatal and neonatal specimens. Prenat Diagn 28(9): 789–795. Slater HR, Bruno DL, Ren H, et al. 2003. Rapid, high throughput prenatal detection of aneuploidy using a novel quantitative method (MLPA). J Med Genet 40(12): 907–912. Van den Veyver IB, Patel A, Shaw CA, et al. 2009. Clinical use of array comparative genomic hybridization (aCGH) for prenatal diagnosis in 300 cases. Prenat Diagn 29(1): 29–39. Verma L, Macdonald F, Leedham P, et al. 1998. Rapid and simple prenatal DNA diagnosis of Down’s syndrome. Lancet 352(9121): 9–12. Wahlström J. 1974. The quantity of viable cells at various stages of gestation. Humangenetik 22(4): 335–342.

Prenat Diagn (2011) DOI: 10.1002/pd