Customized high resolution CGHarray for ... - Wiley Online Library

4 downloads 5678 Views 2MB Size Report
Jun 24, 2013 - Customized High Resolution CGH-Array for Clinical. Diagnosis Reveals Additional Genomic Imbalances in Previous Well-Defined Pathological ...
RESEARCH ARTICLE

Customized High Resolution CGH-Array for Clinical Diagnosis Reveals Additional Genomic Imbalances in Previous Well-Defined Pathological Samples Elena Vallespı´n,1,2 Marı´a Palomares Bralo,1,2 M. ´Angeles Mori,2,3 Rube´n Martı´n,1,2 Sixto Garcı´a-Min˜au´r,2,4 Luis Ferna´ndez,1,2 M. Luisa de Torres,2,3 Fe Garcı´a-Santiago,2,3 Elena Mansilla,2,3 Fernando Santos,2,4 Victoria E. M-Montan˜o,1 M. Carmen Crespo,1 Sol Martı´n,1 Victor Martı´nez-Glez,1,2 Alicia Delicado,2,3 Pablo Lapunzina,2,4 and Julia´n Nevado1,2* 1

Section of Functional and Structural Genomics of Instituto de Gene´tica Me´dica y Molecular (INGEMM)-IdiPAZ, Hospital Universitario La Paz, Madrid, Spain 2 CIBERER, Centro de Investigacio´n Biome´dica en Red de Enfermedades Raras, Madrid, Spain 3 Section of Cytogenetics of Instituto de Gene´tica Me´dica y Molecular (INGEMM)-IdiPAZ, Hospital Universitario La Paz, Madrid, Spain 4 Section of Clinical Genetics of Instituto de Gene´tica Me´dica y Molecular (INGEMM)-IdiPAZ, Hospital Universitario La Paz, Madrid, Spain Manuscript Received: 27 June 2011; Manuscript Accepted: 3 March 2013

High-resolution array comparative genomic hybridization (aCGH) is a powerful molecular cytogenetic tool that is being adopted for diagnostic evaluation of genomic imbalances and study disease mechanisms and pathogenesis. We report on the design and use, of a custom whole-genome oligonucleotide-based array (called KaryoArray1v3.0; Agilent-based 8  60 K) for diagnostic setting, which was able to detect new and unexpected rearrangements in 11/63 (17.5%) of previous known pathological cases associated with known genetic disorders, and in the second step it identified at least one causal genomic imbalance responsible of the phenotype in 20% of patients with psychomotor development delay and/or intellectual disability. To validate the array, first; we blindly tested 120 samples; 63 genomic imbalances that had previously been detected by karyotyping, FISH and/or MLPA, and 57 sex-matched control samples from healthy individuals; secondly a prospective study of 540 patients with intellectual disabilities, autism spectrum disorder and multiple congenital anomalies were evaluated to confirm the utility of the tool. These data indicate that implementation of array technologies as the first-tier test may reveal that additional genomic imbalances could co-exist in patients with trisomies and classical del/dup syndromes, suggesting that aCGH may also be indicated in these individuals, at least when phenotype does not match completely with genotype. Ó 2013 Wiley Periodicals, Inc.

Key words: chromosomal imbalances; arrayCGH; customized array; cytogenetics; microdeletions; microduplications; KaryoArray

INTRODUCTION Genomic imbalances are a major cause of congenital and developmental abnormalities observed in patients with dysmorphic

Ó 2013 Wiley Periodicals, Inc.

How to Cite this Article: Vallespı´n E, Palomares Bralo M, Mori MA, Martı´n R, Garcı´a-Min˜au´r S, Ferna´ndez L, de Torres ML, Garcı´a-Santiago F, Mansilla E, Santos F, M-Montan˜o VE, Crespo MC, Martı´n S, Martı´nez-Glez V, Delicado A, Lapunzina P, Nevado J. 2013. Customized high-resolution CGH-array for clinical diagnosis reveals additional genomic imbalances in previous well-defined pathological samples. Am J Med Genet Part A 161A:1950–1960.

features, intellectual disability (ID) autistic sprectrum disorders (ASD) and/or multiple congenital anomalies (MCA). Array Comparative Genomic Hybridization (aCGH) is a powerful molecular tool to detect and study genomic imbalances, disease mechanisms and pathogenesis, which is rapidly becoming as a new gold standard Additional supporting information may be found in the online version of this article. Conflict of interest: none. Grant sponsor: FIBHULP (Redes/FIBHULP08.Nevado).  Correspondence to: Julia´n Nevado, Ph.D., MBA, INGEMM—Instituto de Gene´tica Me´dica y Molecular Hospital, Universitario La Paz, Paseo de la Castellana 261, 28046 Madrid, Spain. E-mail: [email protected] Article first published online in Wiley Online Library (wileyonlinelibrary.com): 24 June 2013 DOI 10.1002/ajmg.a.35960

1950

VALLESPI´N ET AL. method for clinical genetics. In fact, using conventional cytogenetics techniques, chromosomal imbalances such as aneuploidies and segmental aneusomies must be larger than 3–5 Mb in size to be detected by GTG-banding at standard 450–600 band resolution [Shaffer and Lupski, 2000]. In the past two decades, classical GTG-banding has been combined with targeted molecular technologies to improve the resolution at which one can detect genomic changes. Fluorescence in situ hybridization (FISH) [Trask, 1991], comparative genomic hybridization (CGH) [Rao et al., 1998], multiplex-FISH [Speicher et al., 1996], spectral karyotyping (SKY) [Lu et al., 2002] and MLPA [Schouten et al., 2002] have been used to detect additional complex or submicroscopic abnormalities; however, such technologies are not suitable for whole genome scans in routine clinical testing because they either lack the necessary resolution or are too time consuming, labour intensive, and/or expensive [Harris et al., 2003; Lau et al., 2004; Ravnan et al., 2006; Stankiewicz and Beaudet, 2007]. Chromosomal microarray analysis (CMA) has higher resolution and excellent throughput when compared with conventional and molecular cytogenetic techniques [Pinkel and Albertson, 2005; Speicher and Carter, 2005]. The main aim of this work is to evaluate the performance of a custom whole genome CMA tool into our clinical routine genetic laboratory. To do this, we designed an Agilent-based 8  60 K high-resolution genome-wide array, called KaryoArray1 based in our clinical/laboratory expertise.

MATERIALS AND METHODS Design of a Customized Array (KaryoArray1) This focused oligonucleotide chip combines genome-wide coverage with targeted enrichment for covering more than 380 clinically relevant regions of genomic imbalances, including all subtelomeric and pericentromeric regions, and regions responsible for welldefined microdeletion/microduplication syndromes, ID, and ASD (for a summary of the list of disorders, see Table I of Supplemental Data in Supporting information online). The features were selected from Agilent’s eArray (Agilent, https://earray.chem.agilent.com/earray) probe library in a custom high-resolution format of 8  60 K containing 62,976 distinct 60-mer oligonucleotides probes. It also includes control probes spanning coding and non-coding sequences. Thus a total of 46,609 region-specific features covering targeted regions (see above) and 16,367 probes of backbone were included in the array. The average density of the probe coverage is 43 Kb; the average density of known pathogenic regions is 7 Kb (3–27 Kb). Each subtelomeric or pericentromeric region has a minimum coverage of 5 Mb. Initially, the probe sequences and gene annotation were based on NCBI Build 36 of the human genome and UCSC version hg18 and eventually have been translated and converted to NCBI Build 37/UCSC version hg19. The current, revised version of the array is v.3.0, and for designing the targeted oligonucleotidebased array, we consulted the Database of Genomic Variants (http://projects.tcag.ca/variation/) and DECIPHER (http://www. sanger.ac.uk/PostGenomics/decipher/) to avoid CNVs with no apparent clinical relevance. Most clinical tests fall short of this ideal. In fact, one of the most important challenges of an array CGH experiment in a clinical setting is trying to avoid false positives or

1951 false negatives [Vermeesch et al., 2007]. We defined sensitivity and specificity parameters [Altman and Bland, 1994] for our array validation. The sensitivity relates to the test’s ability to identify positive results, whereas the specificity relates to the ability of the test to identify negative results. Arrays were manufactured with Agilent’s Sure-Print Inkjet technology (Agilent Tech, Santa Clara, CA). In addition, an important consideration when selecting an array CGH platform is the price, especially when studying a large series of samples, such as for clinical implementation. We choose to customize eight specimens in one slide format (8  60 K). Finally, several key variables were also used to evaluate the quality of the array experiments and to test the quality of the dataset as a whole. The following cutoffs were used to pass our quality-control testing: DLRSpread 0.30, median signal intensity 50, background noise in both channels 10, and SNR 15. None of the samples failed quality-control testing due to poor chip quality or problems with hybridization.

Clinical Samples for Validation Although there is no perfect negative control for this technique, using both control samples, healthy and abnormal samples may help to define operatively and viability of this tool, in order to maximize the detection of true positives and minimize falsenegative and false-positive calls, as experts recommended [Vermeesch et al., 2012]. In fact, they have recently recommended that 50 samples are the reasonable number of controls to be tested for validation purposes, comprising both normal and known abnormal samples. Normal samples are also beneficial in defining the optimum calling thresholds because apparently normal individuals also harbor benign CNVs. To validate KaryoArray1, we blindly tested 120 DNA samples. Two of us (PL and JN) selected from our database 120 DNA samples: 63 of them were wellcharacterized pathogenic samples and 57 were specimens from healthy individuals that were used as negative controls for validation of the tool. The control samples had previously been tested by karyotype and MLPA for telomeric and recurrent genomic rearrangements (SALSAs p036 þ p070 and p245, respectively; MRC-Holland; Amsterdam, the Netherlands). The rest of the researchers did not know the diagnosis of each sample and did not participate in the selection of them. Informed consent was obtained from all individuals involved in the study, which was approved by the IRB of our Center. Among the pathogenic specimens, we included 19 patients with different aneuploidies (including sexual chromosomes), 10 patients with complex rearrangements (duplication of part of one chromosome and deletion in the other one), 11 patients with recurrent microdeletion syndromes (e.g., Williams-Beuren, Sotos, Smith-Magenis, WolfHirschhorn, 1p36, 22q11.2, and 9q34 deletions, etc.), two chromosome rings, and eight patients with small gains or losses involved in new microdeletion or microduplication syndromes, including several samples with interstitial rearrangements. In general, samples included for this validation covered rearrangements in the majority of chromosome arms (see Supplemental data Table II in Supporting information online). In this validation study very diverse specimens (peripheral blood, fetal blood, amniotic fluid, saliva, and different tissue samples) were evaluated.

1952

Clinical Samples for Routine Implementation Studies In a second step, we included 540 samples with diagnosis of ID, ASD, and MCA referred for CMA to our Institute from August 2010–October 2012. In many cases, CMA analysis has been already used as the first-tier test and no other molecular or cytogenetic studies were done.

Microarray Experiments DNA samples were extracted preferentially from peripheral blood lymphocytes with Puregene DNA Isolation Kit following manufacturer’s recommendations (Gentra Systems, Minneapolis, MN). DNA extraction from ammniotic fluid and saliva was done using Masterpure kit (Epicentre Biotech. Madison, WI) and Oragene DNA (DNAGenotek, Otawa, Canada), respectively. Array experiments were performed as recommended by the manufacturer (Agilent Technologies). DNAs (500 ng) from the specimen and a reference of the same sex (Promega, Madison, WI) were doubledigested with RsaI and AluI for 2 hr at 37˚C. After heat inactivation of the enzymes at 65˚C for 20 min, each digested sample was labeled by random priming (Genomic DNA Enzymatic Labelling Kit Agilent catalogue # 5190-0449) for 2 hr using Cy5-dUTP for patient DNAs and Cy3-dUTP for reference DNAs. Labeled products were column-purified (Microcon Ym-30 filters, Millipore Corporation). After probe denaturation and pre-annealing with Cot-1 DNA, hybridization was performed at 65˚C with rotation for 24 hr. After two washing steps, the array was analyzed with the Agilent scanner using the Feature Extraction software (v9.1 Agilent Technologies).

Data Analysis The analysis and visualization of KaryoArray1 data was performed using Genomic Workbench Standard Edition 5.0 software. Comprehensive description of the statistical algorithms is available in the user’s manual provided by Agilent Technologies. The Aberration Detection Method 2 (ADM-2) quality weighted interval score algorithm identifies aberrant intervals in samples that have consistently high or low log ratios based on their statistical score. The score represents the deviation of the weighted average of the normalized log ratios from its expected value of zero calculated with derivative log2 ratio standard deviation algorithm. A fuzzy zero algorithm is applied to incorporate quality information about each probe measurement. To make a positive call, our threshold settings for the CGH analytics software were 6.0 for sensitivity, 0.30 for minimum absolute average log ratio per region. Three consecutive probes with the same polarity were required for the minimum number of probes per region. All arrays were scanned at 3 mm resolution using the Agilent scanner.

RESULTS Validation of Previous Known Samples by aCGH KaryoArray1 detected known regions of genomic imbalances in 63 samples with almost 100% concordance (with exception of one in v.1.0; where X-chromosome pseudoautosomal region-I (PAR-I)

AMERICAN JOURNAL OF MEDICAL GENETICS PART A was not completely covered, which is considered as false negative). The array also refined the genomic intervals involved in the imbalances. This tool also established three false positives (some areas on the array that might not report the expected results due to hybridization artifacts); two in chromosome arm 9p and other at 7p, in v.1.0 (checked by MLPA of subtelomeres and pericentromeric regions). The calculation of the specificity and sensitivity and other analytic validation parameters was done as follows: Sensitivity: true positives/true positives þ false negatives: 98.41%; Specificity: true negatives/true negatives þ false positives: 95.00%; Positive predictive value, defined as the probability of the array to detect an aberration when it is really a genomic aberration: true positives/ true positives þ false positives: 95.38%; Negative predictive value, defined as the array’s probability in not detecting an aberration when it is not real: true negatives/true negatives þ false negatives 98.27%. On the other hand, another parameter to validate the array is defined, as the number of targets will not respond to copy number changes in the expected way. This may reduce the effective resolution of an array and may lead to false negative results. The number of non-responding targets in an otherwise changed region should be documented for each hybridization. In general, it is widely accepted that a global minimal retention value of 90% is acceptable in many other designs [Vermeesch et al., 2007, 2012]. Median average of retention value for KaryoArray1 was 99.76%. All 57 control samples gave normal results for both karyotype and MLPA and served as negative controls for validation. The array did not detect significant imbalances in any of the 57 normal controls, with the exception of several reported benign CNVs (data not shown). In fact, average of CNV detection (including v1.0–v3.0) in normal controls were 2.56/sample, and remarkable in 31.58% of these samples none CNVs were detected. Interestingly, median average of CNV detection in known pathogenic samples were slightly superior to normal controls, around 3.30/sample, and remarkable in 33.33% had only one annotation per sample. The sensitivity and specificity are influenced considerably by the choice of the log R-ratio-calling threshold cutoff values for discriminating deletions and duplications from normal diploid copy number. A high calling threshold will give a higher specificity and can strongly decrease the (technical) false-positive rate. However, it will also lower the sensitivity of the test [Vermeesch et al., 2012]. Our values of DLRs are on very acceptable ranges: average 0.19. Other quality variables; the median signal intensity of both channels, background noise for both channels (mean was 6.21) and signal-to-noise ratio (SNR mean was 17.3) were also analyzed. None of the samples failed quality-control testing due to poor chip quality or problems with hybridization. Interestingly, during the KaryoArray1 validation process we detected 17.5% (11 out of 63 cases see Supplementary data file in Supporting Information online) of additional genomic aberrations (samples 3, 19, 26, 33, 42, 45, 54, 57, 58, 62 and 63; Table I and Fig. 1) in previous well defined pathological samples diagnosed by karyotyping, FISH, or MLPA. Among them, we remark sample #19 in which besides a known deletion on 17p12 (1.4 Mb), two novel aberrations were observed: a deletion at 15q13.2 of 50 Kb and a gain at 15q13.3 of 530 Kb (Table I and Fig. 1b). In sample #42 there was a known deletion at 13q33.2–q34 of 9.7 Mb and a new interstitial de novo deletion of 550 Kb was detected at 16p11.2 (Table I and

VALLESPI´N ET AL.

1953

TABLE I. Size Distribution of Additional Aberrations Found During KaryoArray1 Validation Process Case 3 19 26 33 42 45 54 57 58 62 63

Origin Unknown De novo De novo Unknown De novo Unknown Unknown De novo Unknown Inherited De novo

Size 254 Kb del 50 Kbdel/530 Kbdup 4.1 Mb dup 150 Kbdup 550 Kb del 280 Kb del 581 Kb dup 1.5 Mb dup 428.3 Kb dup 540 Kb dup 8.234 Mb del

Chromosomal region 12q24.33 15q13.2–15q13.3 17q22–23.2 22q11.21 16p11.2 22q11.22 3q29 8q24 1p13.2 16p11.2 7q22.1–q22.3

Fig. 1e). In sample #54, KaryoArray1 confirmed a known complex rearrangement involving a deletion/duplication at 10pter, and also showed a subtelomeric duplication of 581 Kb at 3q29 (Table I). In case #62 a known duplication of SHOX was accompanied with an unexpected 16p11.2 duplication of 540 Kb (Table I). In this case the 16p11.2 duplication was paternally inherited and segregated with the clinical findings, mainly behavior features (aggressivity). Finally, in case #63 routine cytogenetic analysis suggested a de novo interstitial deletion of the long arm of chromosome 7 with the breakpoints in 7q21.2 and 7q31.2. However, KaryoArray1 detected two 7q interstitial deletions (8,057 and 8.234 Mb, respectively) with an intermediate region between the two deleted segments, which had not been lost (Table I). All of these additional findings were reconfirmed by means of MLPA and/or FISH.

CNVs Detected by KaryoArray Setting

1

in the Clinical

The CNVs have been assigned as outlined elsewhere [Miller et al., 2010; Vermeesch et al., 2012; Spanish Consensus Document, 2012]. Briefly, we interpret and classify any detected CNV into one of the following three general categories: (1) pathogenic or clinically relevant; (2) uncertain clinical significance or VOUS; and (3) likely benign, polymorphic. As many of imbalances reported so far are unique, proof of pathogenicity of the imbalance will depend on several additional criteria. Thus, we proceeded to interpret CNVs with differing degrees of confidence depending on a number of factors including prior knowledge, size, gene content and inheritance status as it was previously suggested [de Ravel et al., 2007]. Thus, when a CNV cannot confidently be classified as pathogenic or benign, we reported it as a variant of uncertain clinical significance (VOUS) (Fig. 2). Parental samples are necessary in some cases for a full interpretation of the proband’s array results. Array testing may identify private familial variants that have not been previously observed in studies of apparently normal individuals or in patients with phenotypes [Itsara et al., 2009]. Additional familial samples may also be required to determine

whether a particular CNV is segregating with the phenotype within a family. We are aware that there are numerous reports of inherited pathogenic deletions/duplications that display reduced penetrance and/or variable expression [Cooper et al., 2011; Vermeesch et al., 2012]. In addition, a deletion inherited from an apparently normal parent may in fact reveal a recessive disorder in the proband due to a mutation on the non-deleted allele that is not detectable by the array. In this sense, as it was suggested by others [Vermeesch et al., 2012] inherited imbalances may well turn out to be susceptibility loci contributing to pathogenicity. The screening for submicroscopic aberrations has been subsequently tested in 540 samples by using KaryoArray1 v3.0 and demonstrated 104 pathogenic cases, 17 VOUS, 247 individuals with benign CNVs and 172 without any CNV detected (Table II), with a total of 697 common CNVs. Therefore, in 172 out of 540 patients studied at this resolution level no CNV were detected. These CNVs range in size from 2.5 to 91.33 Mb. On average, each patient carried 1.29 CNVs when they were studied with KaryoArray1v3.0, far away from previous versions of the array. KaryoArray1 revealed 104 pathogenic CNVs out of 540 patients included in this group (19%). Regarding the VOUS cases (3%), CNVs range in size from 17.5 to 909 Kb. All of them affected region that could be considered with uncertain clinical significance by size (less than 1 Mb), although all of them are within pathogenic regions. For these cases parental analysis must be mandatory. Unfortunately, clinical examination of parents in some cases has not been possible. In addition, KaryoArray1 has been crucial in the description of a new microdeletion syndrome [Palomares et al., 2011; Supplemental data, Fig. 1], and the evaluation of a complex chromosome 8 rearrangement [Rodrı´guez et al., 2011].

KaryoArray1 Versus Other Platforms and Other Techniques Figure 3 summarizes some genomic aberrations detected by KaryoArray1 and the comparison in silico with other commercial Agilent-based platforms, such as 44, 60, and 244 K as well as 60KISCA Consortium. Interestingly, KaryoArray1 seems to be critical for diagnostic en cases shown in Figure 3a,c,f and was helpful to define breakpoints in Figure 3b,d when these results were compared to other commercial platforms, mainly 44, 60 K and sometimes 244 K or 60K ISCA. Similarly this was the case for other chromosomal regions such as PLP1, FOXL2, and NSD1genes or 8p23.3 region (Fig. 4). Finally, the array was able to detect two mosaicisms that conventional karyotype (a) and MLPA (b) analysis cannot detect properly, such as an extra isochromosome in blood lymphocytes (a) and a distal monosomy 2qter (b), respectively (Fig. 5).

DISCUSSION CMA is a valuable clinical diagnostic assay for patients with ID and other genetic conditions, such as MCA or ASD. This methodology became to be worldwide accepted as first-tier for clinical diagnosis in this group of patients [Edelmann and Hirschhorn, 2009; Vissers et al., 2010; Miller et al., 2010]. Diagnosis of ID, ASD and/or MCA represents the largest proportion of tests for our laboratory, due to

1954

AMERICAN JOURNAL OF MEDICAL GENETICS PART A

FIG. 1. KaryoArray1 was able to detect additional genomic imbalance events associated with known genetic disorders. Some examples of the aCGH-plots are shown: cases 3, 19, 26, 33, 42 and 45 (a–f).

VALLESPI´N ET AL.

1955

FIG. 2. Algorithm for the assessment of clinical relevance for a particular CNV. Based on Buysse et al. [2009] and Miller et al. [2010]. CNV; copy number variant; VOUS, variant of uncertain significance; MR, intellectual disability; MCA, multiple congenital anomalies.

its high prevalence in the population. Taken together these three groups of conditions affect 1 in every 20 individuals [Shevell et al., 2008; Newschaffer et al., 2007]. As expected, CMA analysis offered more advantages than karyotype, FISH, MLPA, all together, even in conditions when this technology seems to be weaker, such as detecting low grade of mosaicism (Fig. 5). In terms of comparison with other techniques, the implementation of whole-genome microarray allows the detection of copy-number changes of 100 kb throughout the genome, thus

TABLE II. CNVs Detection Rates by KaryoArray1 v3.0; 8  60K Total CNVs Causal VOUS Benign None CNV detected CNV detected/sample

Number of samples: 540 697 19.44% (105) 2.96% (16) 45.74% (247) 31.85% (172) 1.29

CNV, copy number variant; VOUS, variant of uncertain significance

increasing the diagnostic yields from 11% to 20% [reviewed in Miller et al., 2010]. In addition, analysis of CMA in 13.926 subjects with ID or MCA, most of which had a normal conventional karyotype reported a rate of 10% (7–35%) of pathogenic genomic rearrangements [Sagoo et al., 2009]. Similarly, a study of 36.325 patients with ID estimated that arrays with an average spacing of a probe density of 30–70 kb can detect a pathogenic disturbance in 19% of patients [Hochstenbach et al., 2009]. The resolution of molecular karyotyping is a priori dependant on the platform. In this sense, there is still much debate about the optimal resolution for genomic arrays; increasing the resolution has the advantage that even smaller pathogenic CNVs can be detected; however as a consequence, the number of benign CNVs detected increases exponentially. In other words, and the real situation, while the analytical validity increases, there is a drop in the clinical validity of the information obtained [Vermeesch et al., 2012]. Thus, an ideal CMA platform for clinical routine should cover the entire genome with the capability to detect all known genomic aberration and provide sufficient probe density for the discovery of novel pathogenic chromosomal rearrangements with high sensitivity and specificity while maintaining acceptable number of benign CNVs or VOUS. In addition, it should be cost-effective to be introduced in a clinical setting.

1956

AMERICAN JOURNAL OF MEDICAL GENETICS PART A

FIG. 3. Virtual comparison among KaryoArray1 and several Agilent-based commercial aCGH platforms (such as 44, 60, and 244 K, as well as 60 K ISCA Consortium) for detecting some genomic aberrations found in our clinical setting. a: atypical-9q34.2 microdeletion syndrome; b: 16p11.2 microdeletion syndrome; c: a monosomy/trisomy 10pter. d: Pitt–Hopkins microdeletion syndrome. e: 22q11.2 distal microdeletion syndrome. f: atypical HNPP deletion.

In our hands, we showed the utility and cost-effective features of a customized oligonucleotide aCGH platform in a clinical setting using a 60 K platform that supports significantly with data many of the recommendations for aCGH-platforms in a recent paper [Vermeesch et al., 2012], that can be considered as the European Consensus Document, which recommends using a resolution of 200 kb for clinical setting. Genomic backbone coverage below 200 kb may increase the number of CNVs of unknown significance to the extent that their clinical validity becomes too low. In fact, in our array the average density of the total probe coverage for the backbone is 175 Kb, and 7 Kb (3–27 Kb) in known enrichment pathogenic regions.

Chromosomal Microarray Analysis using KaryoArray Interestingly, KaryoArray1 made also possible a fine mapping within target-regions and a better resolution in silico than other commercially available Agilent-based array platforms, such as 44, 60, and 244 K or even than ISCA-60K (Figs. 3 and 4). Indeed,

average resolution of KaryoArray1 is approximately 7 Kb. This resolution is, for regions of interests in MCA/MR/ASD patients, equal and/or higher than commercial Agilent-based 244 K aCGH platform. Thus, with a “sufficient” coverage of oligoprobes, all the current commercial platforms of arrays would be also able to provide a sufficient sensitivity and specificity for the clinical tests of a-CGH, another issue is the cost per patient. Although it is expected that higher resolution level of aCGH can reveal novel and more submicroscopic rearrangements [Friedman et al., 2006; Shaffer et al., 2007; Toruner et al., 2007; Lybaek et al., 2008], its interpretation may also become increasingly challenging in a clinical setting. Our array was able to detect cryptic rearrangements, submicroscopic anomalies and even single-exon deletions (Fig. 3d). This kind of resolution in such target-regions may have great importance in cases in which the inclusion or exclusion of certain genes may predict clinical features of the condition (e.g., dosage-sensitive genes). KaryoArray1 yields 20% of pathogenic cases in non-strictly selected patients with ID. This yield is situated at the top high of studies reviewed above, and supports a relevant role of

VALLESPI´N ET AL.

1957

FIG. 4. Virtual comparison between KaryoArray1 and several Agilent-based commercial aCGH platforms (such as 44, 60, and 244 K, as well as 60 K ISCA Consortium) for some genes and/or genomic regions. a: 17q21.31 microdeletion syndrome; b: PLP1 gene; c: NSD1/FGFR4 genes for Sotos syndrome; d: FOXL2 gene for BPES syndrome; and e: 8p23.2 region.

KaryoArray1 in the clinical setting of this type of patients. Interestingly, a relevant number of pathogenic CNVs identified were less than 1 Mb in length, which is an important data to be considered when evaluating a possible cut-off for a positive result. Although at this point, we cannot consider a specific size cut-off across the genome. Indeed, as far as we know, the smallest known recurrent microdeletion syndromes described are at 17q21.31 region (OMIM 610443) [Shaw-Smith et al., 2006], 16p11.2 (OMIM 611913) [Ballif et al., 2007], or in the 16p12 region [Girirajan et al., 2010], and all are of 200–500 kb range. However, there are many more alterations not flanked by LCRs that are smaller than 100 kb, such as those that affect MBD5, SHANK3, or MECP2 genes. In our custom designed tool, average coverage within the above-mentioned genes is at 1– 3 kb/probe. In addition, KaryoArray1 was able to maintain an acceptable balance between the number of benign CNVs or VOUS (variants of uncertain significance, 3%), as well as it was able to detect all known genomic aberrations. In addition, it provides sufficient probe density for discovery of a novel pathogenic chromosomal rearrangement within the backbone [Palomares et al., 2011], with a high sensitivity and specificity (see Results Section) and at a relative low cost. In fact, the expert recommends that the array-CGH in a

clinical use should be designed to detect imbalances of approximately 20–50 kb in the regions of interests (e.g., within known OMIM genes) and to detect imbalances of 100–250 Kb in other regions of the genome (the so-called backbone) [Vermeesch et al., 2012].

Re-diagnosing Patients With KaryoArray1 An unexpected, but significant remarkable aspect of our tool validation process was the identification of additional cryptic submicroscopic imbalances in known pathogenic samples previously established by karyotype and MLPA/FISH. In fact, the detection of 17.5% of additional genomic aberrations (most of them were interstitials) further support the clinical value of this tool and also confirms the analytical superiority of CMA compared to current gold standard procedures [Miller et al., 2010]. On the other hand, a higher frequency of complex rearrangements in previous well-defined known pathogenic samples (by means of a-CGH) has been also established by others: 13.13% [Toruner et al., 2007], 40% [Xiang et al., 2008] 3.12% [Yu et al., 2009], and 11.1% [Xiang et al., 2010], but no significant comment were made on them.

1958

AMERICAN JOURNAL OF MEDICAL GENETICS PART A

FIG. 5. KaryoArray1 is able to detect a low grade of mosaicism. a: 10% of mosaicism at p-arm of chromosome 12 detected in a blood sample. Blood karyotype and MLPA failed to detect it. b: 30% of mosaicism at 2q37 qter detected by KaryoArray1. Subtelomere MLPA (Kit SALSA P036) failed to detect it, and specific 2q telomere MLPA (kit SALSA P264, MRC-Holland) was uncertain.

These findings may have additional implications on the interpretation of results regarding the complexity of the extra rearrangement and for genetic counseling. It also suggests that trisomies or submicroscopic deletions/duplications may co-exist with additional rearrangements, which in some way might contribute to the abnormal phenotypes. Indeed, a recent multicenter study of 1.500 consecutive cases using aCGH reported that approximately 20% of the abnormal cases involved two or more DNA segments in one or more chromosomes [Xiang et al., 2010]. A critical point is to find out whether or not these unexpected new genomic aberrations have phenotypic relevance. Indeed, in case #19 a known deletion at 17p11.2 associated with HNPP disease (hereditary neuropathy with liability to pressure palsies, OMIM#162500) co-exists with a double del/dup at 15q13.2–q13.3, involving CHRNA7 gene. Microduplications of CHRNA7 has been associated with developmental delay/intellectual disability, muscular hypotonia, and a variety of neuropsychiatric disorders [Szafranski et al., 2010], and it has been considered a common risk

factor for many neurobehavioral disorders [Cooper et al., 2011]. Interestingly, clinical phenotype in this patient was more complex than a classical HNPP and included ID. In the same way, in patient #42 the array detected a 9.7 Mb subtelomeric deletion at 13q arm and an additional deletion of 550 Kb at 16p11.2, a region associated with ASD and with a range of behavioral anomalies and cognitive, developmental, and speech delay [Ballif et al., 2007]. In regions such as these, copy number changes may unmask recessive alleles or work in conjunction with various genetic modifiers, perhaps even other CNVs, to produce or give a more severe clinical phenotype [Girirajan et al., 2010]. Regarding chromosomal region16p11.2, in case #62 the array revealed besides a known duplication of SHOX, a paternal inherited duplication at 16p11.2, which may be further responsible of the abnormal behavior (aggressivity) in the proband, his brother, his father, and grand-mother. Similarly, in other cases such as #19, 26, 58, or 63, the additional findings may have significant importance for deciphering the mechanisms by which genomic rearrangement originated,

VALLESPI´N ET AL. unveiling a more complex structural genomic disorder. Finally, reanalyzing case #58 using SNP-array (Human 610 Quad BeadChip, Illumina.San Diego, CA) and different STRs (Short Tandem Repeats) revealed beside the distal trisomy 14qter, an additional paternal segmental iso-disomy within chromosomal regions 14q32.13-qter. These regions contain several imprinted genes (paternally-or maternally-expressed). In summary, we found that extra genomic imbalances may coexist with trisomies and other classical micro-deletion/duplication syndromes, suggesting that CMA may be also indicated in these individuals, at least when karyotype/MLPA results and clinical phenotype do not match completely. In addition, our data have shown that a well design customized aCGH tool, such as KaryoArray1, may provide a good balance among prize/resolution, an excellent sensitivity and specificity for introducing CMA in clinical practice.

REFERENCES Altman DG, Bland JM. 1994. Diagnostic tests. Brit Med J 308:1552. Ballif BC, Hornor SA, Jenkins E, Madan-Khetarpal S, Surti U, Jackson KE, Asamoah A, Brock PL, Gowans GC, Conway RL, Graham JM Jr, Medne L, Zackai EH, Shaikh TH, Geoghegan J, Selzer RR, Eis PS, Bejjani BA, Shaffer LG. 2007. Discovery of a previously unrecognized microdeletion syndrome of 16p11.2–p12.2. Nat Genet 39:1071–1073. Buysse K, Delle Chiaie B, Van Coster R, Loeys B, De Paepe A, Mortier G, Speleman F, Menten B. 2009. Challenges for CNV interpretation in clinical molecular karyotyping: Lessons learned from a 1001 sample experience. Eur J Med Genet 52:398–403. Cooper GM, Coe BP, Girirajan S, Rosenfeld JA, Vu TH, Baker C, Williams C, Stalker H, Hamid R, Hannig V, Abdel-Hamid H, Bader P, McCracken E, Niyazov D, Leppig K, Thiese H, Hummel M, Alexander N, Gorski J, Kussmann J, Shashi V, Johnson K, Rehder C, Ballif BC, Shaffer LG, Eichler EE. 2011. A copy number variation morbidity map of developmental delay. Nat Genet 43:838–846. de Ravel TJ, Devriendt K, Fryns JP, Vermeesch JR. 2007. What’s new in karyotyping? The move towards array comparative genomic hybridisation (CGH). Eur J Pediatr 166:637–643. Edelmann L, Hirschhorn K. 2009. Clinical utility of array CGH for the detection of chromosomal imbalances associated with mental retardation and multiple congenital anomalies. Ann NY Acad Sci 1151:157–166. Friedman JM, Baross A, Delaney AD, Ally A, Arbour L, Armstrong L, Asano J, Bailey DK, Barber S, Birch P, Brown-John M, Cao M, Chan S, Charest DL, Farnoud N, Fernandes N, Flibotte S, Go A, Gibson WT, Holt RA, Jones SJ, Kennedy GC, Krzywinski M, Langlois S, Li HI, McGillivray BC, Nayar T, Pugh TJ, Rajcan-Separovic E, Schein JE, Schnerch A, Siddiqui A, Van Allen MI, Wilson G, Yong SL, Zahir F, Eydoux P, Marra MA. 2006. Oligonucleotide microarray analysis of genomic imbalance in children with mental retardation. Am J Hum Genet 79:500–513. Girirajan S, Rosenfeld JA, Cooper GM, Antonacci F, Siswara P, Itsara A, Vives L, Walsh T, McCarthy SE, Baker C, Mefford HC, Kidd JM, Browning SR, Browning BL, Dickel DE, Levy DL, Ballif BC, Platky K, Farber DM, Gowans GC, Wetherbee JJ, Asamoah A, Weaver DD, Mark PR, Dickerson J, Garg BP, Ellingwood SA, Smith R, Banks VC, Smith W, McDonald MT, Hoo JJ, French BN, Hudson C, Johnson JP, Ozmore JR, Moeschler JB, Surti U, Escobar LF, El-Khechen D, Gorski JL, Kussmann J, Salbert B, Lacassie Y, Biser A, McDonald-McGinn DM, Zackai EH, Deardorff MA, Shaikh TH, Haan E, Friend KL, Fichera M, Romano C, Ge´cz J, DeLisi LE, Sebat J, King MC, Shaffer LG, Eichler EE. A recurrent

1959 16p12.1 microdeletion supports a two-hit model for severe developmental delay. Nat Genet 2010. 42:203–209. Harris CP, Lu XY, Narayan G, Singh B, Murty VV, Rao PH. 2003. Comprehensive molecular cytogenetic characterization of cervical cancer cell lines. Genes Chromosomes Cancer 36:233–241. Hochstenbach R, van Binsbergen E, Engelen J, Nieuwint A, Polstra A, Poddighe P, Ruivenkamp C, Sikkema-Raddatz B, Smeets D, Poot M. 2009. Array analysis and karyotyping: Workflow consequences based on a retrospective study of 36,325 patients with idiopathic developmental delay in the Netherlands. Eur J Med Genet 52:161–169. Itsara A, Cooper GM, Baker C, Girirajan S, Li J, Absher D, Krauss RM, Myers RM, Ridker PM, Chasman DI, Mefford H, Ying P, Nickerson DA, Eichler EE. 2009. Population analysis of large copy number variants and hotspots of human genetic disease. Am J Hum Genet 84:148–161. Lau CC, Harris CP, Lu XY, Perlaky L, Gogineni S, Chintagumpala M, Hicks J, Johnson ME, Davino NA, Huvos AG, Meyers PA, Healy JH, Gorlick R, Rao PH. 2004. Frequent amplification and rearrangement of chromosomal bands 6p12–p21 and 17p11.2 in osteosarcoma. Genes Chromosomes Cancer 39:11–21. Lu XY, Harris CP, Cooley L, Margolin J, Steuber PC, Sheldon M, Rao PH, Lau CC. 2002. The utility of spectral karyotyping in the cytogenetic analysis of newly diagnosed pediatric acute lymphoblastic leukemia. Leukemia 16:2222–2227. Lybaek H, Meza-Zepeda LA, Kresse SH, Høysaeter T, Steen VM, Houge G. 2008. Array-CGH fine mapping of minor and cryptic HR-CGH detected genomic imbalances in 80 out of 590 patients with abnormal development. Eur J Hum Genet 16:1318–1328. Miller DT, Adam MP, Aradhya S, Biesecker LG, Brothman AR, Carter NP, Church DM, Crolla JA, Eichler EE, Epstein CJ, Faucett WA, Feuk L, Friedman JM, Hamosh A, Jackson L, Kaminsky EB, Kok K, Krantz ID, Kuhn RM, Lee C, Ostell JM, Rosenberg Scherer SW, Spinner NB, Stavropoulos DJ, Tepperberg JH, Thorland EC, Vermeesch JR, Waggoner DJ, Watson MS, Martin CL, Ledbetter DH. 2010. Consensus statement: Chromosomal microarray is a first-tier clinical diagnostic test for individuals with developmental disabilities or congenital anomalies. Am J Hum Genet 86:749–764. Newschaffer CJ, Croen LA, Daniels J, Giarelli E, Grether JK, Levy SE, Mandell DS, Miller LA, Pinto-Martin J, Reaven J, Reynolds AM, Rice CE, Schendel D, Windham GC. 2007. The epidemiology of autism spectrum disorders. Annu Rev Public Health 28:235–258. Palomares M, Delicado A, Mansilla E, de Torres ML, Vallespı´n E, Fernandez L, Martinez-Glez V, Garcı´a-Min˜aur S, Nevado J, Simarro FS, Ruiz-Perez VL, Lynch SA, Sharkey FH, Thuresson AC, Annere´n G, Belligni EF, Martı´nez-Ferna´ndez ML, Bermejo E, Nowakowska B, Kutkowska-Kazmierczak A, Bocian E, Obersztyn E, Martı´nez-Frı´as ML, Hennekam RC, Lapunzina P. 2011. Characterization of an 8q21.11 microdeletion syndrome associated with mental retardation and a recognizable phenotype. Am J Hum Genet 89:295–301. Pinkel D, Albertson DG. 2005. Comparative genomic hybridization. Annu Rev Genomics Hum Genet 6:331–354. Rao PH, Houldsworth J, Dyomina K, Parsa NZ, Cigudosa JC, Louie DC, Popplewell L, Offit K, Jhanwar SC, Chaganti RS. 1998. Chromosomal and gene amplification in diffuse large B-cell lymphoma. Blood 92:234– 240. Ravnan JB, Tepperberg JH, Papenhausen P, Lamb AN, Hedrick J, Eash D, Ledbetter DH, Martin CL. 2006. Subtelomere FISH analysis of 11 688 cases: An evaluation of the frequency and pattern of subtelomere rearrangements in individuals with developmental disabilities. J Med Genet 43:478–489. Rodrı´guez L, Nevado J, Vallespin E, Palomares M, Golmayo L, Bonaglia MC, Delicado A, Abarca E. 2011. Molecular characterization of an

1960 atypical inv dup del 8q. Proposal of a mechanism of formation. Am J Med Genet A 155A:915–919. Sagoo GS, Butterworth AS, Sanderson S, Shaw-Smith C, Higgins JP, Burton H. 2009. Array CGH in patients with learning disability (mental retardation) and congenital anomalies: Updated systematic review and meta-analysis of 19 studies and 13,926 subjects. Genet Med 11:139–146. Schouten JP, McElgunn CJ, Waaijer R, Zwijnenburg D, Diepvens F, Pals G. 2002. Relative quantification of 40 nucleic acid sequences by multiplex ligation-dependent probe amplification. Nucl Acids Res 30:e57. Shaffer LG, Beaudet AL, Brothman AR, Hirsch B, Levy B, Martin CL, Mascarello JT, Rao KW. Working Group of the Laboratory Quality Assurance Committee of the American College of Medical Genetics. 2007. Microarray analysis for constitutional cytogenetic abnormalities. Genet Med 9:654–662. Shaffer LG, Lupski JR. 2000. Molecular mechanisms for constitutional chromosomal rearrangements in humans. Annu Rev Genet 34:297–329. Shaw-Smith C, Pittman AM, Willatt L, Martin H, Rickman L, Gribble S, Curley R, Cumming S, Dunn C, Kalaitzopoulos D, Porter K, Prigmore E, Krepischi-Santos AC, Varela MC, Koiffmann CP, Lees AJ, Rosenberg C, Firth HV, de Silva R, Carter NP. 2006. Microdeletion encompassing MAPT at chromosome 17q21.3 is associated with developmental delay and learning disability. Nat Genet 38:1032–1037. Shevell MI, Bejjani BA, Srour M, Rorem EA, Hall N, Shaffer LG. 2008. Array comparative genomic hybridization in global developmental delay. Am J Med Genet B 147B:1101–1108. aCGH Spanish Consensus Document 2012. Consenso para la implementacio´n de arrays (CGH y SNP-arrays) en la Gene´tica Clı´nica. www. Institutoroche.es. Speicher MR, Gwyn BS, Ward DC. 1996. Karyotyping human chromosomes by combinatorial multi-fluor FISH. Nat Genet 12:368–375. Speicher MR, Carter NP. 2005. The new cytogenetics: Blurring the boundaries with molecular biology. Nat Rev Genet 6:782–792. Stankiewicz P, Beaudet AL. 2007. Use of array CGH in the evaluation of dysmorphology, malformations, developmental delay, and idiopathic mental retardation. Curr Opin Gen Devel 17:182–192.

AMERICAN JOURNAL OF MEDICAL GENETICS PART A Szafranski P, Schaaf CP, Person RE, Gibson IB, Xia Z, Mahadevan S, Wiszniewska J, Bacino CA, Lalani S, Potocki L, Kang SH, Patel A, Cheung SW, Probst FJ, Graham BH, Shinawi M, Beaudet AL, Stankiewicz P. 2010. Structures and molecular mechanisms for common 15q13.3 microduplications involving CHRNA7: Benign or pathological? Hum Mutat 31:840–850. Toruner GA, Streck DL, Schwalb MN, Dermody JJ. 2007. An oligonucleotide based array-cgh system for detection of genome wide copy number changes including subtelomeric regions for genetic evaluation of mental retardation. Am J Med Genet A 143A:824–829. Trask BJ. 1991. Fluorescence in situ hybridization: Applications in cytogenetics and gene mapping. Trends Genet 7:149–154. Vermeesch JR, Fiegler H, de Leeuw N, Szuhai K, Schoumans J, Ciccone R, Speleman F, Rauch A, Clayton-Smith J, Van Ravenswaaij C, Sanlaville D, Patsalis PC, Firth H, Devriendt K, Zuffardi O. 2007. Guidelines for molecular karyotyping in constitutional genetic diagnosis. Eur J Hum Genet 15:1105–1114. Vermeesch JR, Brady PD, Sanlaville D, Kok K, Hastings RJ. 2012. Genomewide arrays: Quality criteria and platforms to be used in routine diagnostics. Hum Mutat 33:906–915. Vissers L, de Vries B, Joris AVeltman. 2010. Genomic microarrays in mental retardation: From copy number variation to gene, from research to diagnosis. J Med Genet 47:289–297. Xiang B, Zhu H, Shen Y, Miller DT, Lu K, Hu X, Andersson HC, Narumanchi TM, Wang Y, Martinez JE, Wu BL, Li P, Li MM, Chen TJ, Fan YS. 2010. Genome-wide oligonucleotide array comparative genomic hybridization for etiological diagnosis of mental retardation. A multicenter experience of 1499 clinical cases. J Mol Diagn 12:204–212. Xiang B, Li A, Valentin D, Nowak NJ, Zhao H, Li P. 2008. Analytical and clinical validity of whole-genome oligonucleotide array comparative genomic hybridization for pediatric patients with mental retardation and developmental delay. Am J Med Genet A 146A:1942–1954. Yu S, Bittel DC, Kibiryeva N, Zwick DL, Cooley LD. 2009. Validation of the Agilent 244K oligonucleotide array-based comparative genomic hybridization platform for clinical cytogenetic diagnosis. Am J Clin Pathol 132: 349–360.

Suggest Documents