MATERIALS AND METHODS

Target DNA for microarray

BAC/PAC clones used in this study are shown in table 1. We selected subtelomeric clones based on the second generation set of human chromosome specific subtelomeric probes reported by Knight et al.⁸ Five clones (GS-124I4 for 16p, GS-48O23 and GS-325I23 for 19q, GS-82O2 for 20p, and GS-202M24 for Xq) that showed signals appearing at sites different from those originally reported were excluded from this study. Three clones (GS-99K24, CITD-3018K1 for 22q, and GS-225F6 for XqYq) were unavailable. New clones (RP11-359B7 for 19q, RP11-238J15 for 20p, RP3-402G11 for 22q, and RP11-431N21 for XqYq) were obtained from online databases, the human telomere mapping and sequencing project (http://www.wistar.upenn.edu/Riethman/) and the human genome browser (http://genome.ucsc.edu/cgi-bin/hgGateway?Db=hg11). Although RP11-431N21 mapped to Xq/Yq showed unequal signal intensities between Xq and Yq (FISH signals of Xq were twice as intense as those of Yq), we included it in our microarray because no other suitable clones could be obtained. In addition, we added two X chromosome specific clones (RP11-267O9 on Xp22.32 and RP4-617A9 on Xp22.3) as internal controls. Clones were purchased from either Incyte Genomics (Polo Alto, CA, USA) or Invitrogen (Carlsbad, CA, USA). DNA was extracted using QIAgen Plasmid Midi kits (QIAgen, Hilden, Germany) from 200 ml bacterial culture with appropriate antibiotics according to the manufacturer’s protocol.

Degenerate oligonucleotide primed PCR

DOP-PCR was performed to amplify target clone DNA using three different 5′-aminolabelled PCR primers, DOP-A (5’-CCGACTCGAGNNNNNNATGTGG-3’), DOP-B (5’-CCGACTCGAGNNNNNNTTCACC-3’), and DOP-C: (5’-CCGACTCGAGNNNNNNAACAGG-3’). The sequence of DOP-A is the same as that of 6MW by Telenius et al.⁹ Two other primers were randomly designed and confirmed as capable of amplifying clone DNAs as effectively as DOP-A. Primers were purchased from Hokkaido System Science Co Ltd (Sapporo, Japan). DOP-PCR was performed in a 100 μl cocktail containing 50 ng BAC/PAC DNA, 1 μmol/l of a DOP primer, 3 mmol/l MgCl₂, 200 mmol/l of each dNTP, 1× ExTaq buffer (Takara Bio, Ohtsu, Japan), and 0.5 U ExTaq^™HS enzyme (Takara Bio). PCR was cycled 30 times at 94°C for 30 seconds, at linear ramping of 37–72°C over 10 minutes, and at 72°C for 1 minute. Successfully amplified PCR products, usually 0.2–2 kb in size, were used for subsequent experiments. Using DOP-A, DOP-B, or DOP-C, each PCR product (100 μl) was mixed with the same volume (300 μl) of 4×print buffer (1 mol/l sodium phosphate/0.001% N-lauroyl sarcosine, pH 8.5) and concentrated to a final volume of 30 μl using Microcon 100 (Millipore, Bedford, MA, USA). The products were mixed with DMSO and ddH₂O to a final concentration of 1×print buffer/20% DMSO and were ready for prints.

Subject DNA

Genomic DNA from 69 cases of idiopathic MR, whose karyotype was normal by G banding chromosome analysis at the 400 band level, was analysed. MCA was seen in all subjects except one, but recognisable syndromes were ruled out by at least one of the co-authors. We also used DNA from five patients with WHS (WHS1–5), who had been diagnosed as having various deletions for distal 4p, as positive controls to examine the reliability of the microarray CGH system described below (table 2), and that from normal controls as reference DNA for the system.

Microarray CGH

Each product was spotted in triplicate onto CodeLink^™ activated slides (Amersham Bioscience, Buckinghamshire, UK) using Stampman (Nippon Laser and Electronics Laboratory, Nagoya, Japan) with a centre to centre distance of 1 mm between adjacent spots. The same triplicate spots were printed twice to block A and block B on the same slide. The slides were then pretreated,¹⁰ denatured, and stored in a desiccator until use. Subjects’ DNA was labelled with Cy3, and control DNA with Cy5 (CGH 1). Reverse labelling (patient DNA with Cy5, and control DNA with Cy3) was also performed (CGH 2). BamHI digested genomic DNA (1 μg) was labelled with Cy5 or Cy3 dCTP by random prime labelling. Differentially labelled DNA was mixed with 67.5 μg CotI DNA (Invitrogen, Tokyo, Japan), and suspended with 60 μl hybridisation buffer (50% formamide, 10% dextran sulfate, 0.1% Tween 20, 2×SSC, 10 mmol/l Tris-HCl; pH 7.4) with 60 μg yeast tRNA (Roche, Basel, Switzerland). After prehybridisation with herring sperm DNA and human CotI DNA using an in situ frame (Eppendorf, Hamburg, Germany) surrounding each block on a slide in a hybridisation chamber (Genetix, New Milton, UK), labelled probe mixtures of CGH1 and CGH2 were simultaneously applied to blocks A and B, respectively. Slides were incubated at 37°C for 72 hours with gentle shaking (5 rpm), and then washed twice with solution A (1×PBS with 0.05% Tween 20) for 10 minutes at room temperature, twice with solution B (2×SSC with 50% formamide) at 43°C for 15 minutes, and finally with solution A at room temperature for 10 minutes. Slides were scanned with GenePix 4000B (Axon Instruments Inc, Union City, CA, USA) and analysed with GenePix Pro 4.0 software (Axon Instruments Inc).

The intensity ratio between patient and control DNA was calculated from the data of the single slide experiment, using the ratio of means formula. Average normalised inter-locus fluorescence ratio (ANILFR)¹¹ was calculated for each CGH. We had previously performed five sets of CGH experiments with two different normal DNA samples in order to evaluate the thresholds for copy number gain and loss, which showed that the average ANILFR of each clone ranged from 0.86 to 1.18, with an SD of 0.06. Based on this, the normal ANILFR range was defined as numbers within 2SD. This configuration was verified by the fact that the ANILFR of all clones excluding unique X chromosomal clones were within 2SD (fig 1A). If the mean signal ratio on each spot of clone was more than 2SD out of range, FISH analysis was subsequently performed.

FISH analysis

Subtelomeric BAC/PAC DNA labelled with either SpectrumOrange^™ dUTP or SpectrumGreen^™ dUTP (Vysis, Downers Grove, IL, USA) by nick translation was used for FISH analysis. Probes were denatured and hybridised to metaphase chromosomes of patients at 37°C for 16 hours and washed. Photoimages were captured using a monochrome CCD camera (Carl Zeiss, Jena, Germany) on an Axioplan 2 fluorescence microscope (Carl Zeiss) with appropriate filters, and were analysed with ISIS software (version 3.4.5; MetaSystems, Altlussheim, Germany). All five WHS cases were analysed using all subtelomeric clones to confirm the results of the microarray analysis. In MR patients, specific subtelomeric probes were used for FISH analysis where a deletion or a duplication was suspected.

RESULTS

Patients with WHS

Five cases of WHS were analysed as positive controls by microarray CGH and FISH using all subtelomeric probes. As expected, using microarray CGH with the 4p subtelomeric clone, GS-36P21, the signal ratios for the patients were beyond the 2SD range (fig 1B–F), and all were confirmed to have a deletion at the GS-36P21 site. In addition, other subtelomeric aberrations were suspected by microarray analysis—that is, a 10q duplication in WHS1 and a 5q duplication in WHS2 (fig 1B, C). FISH using respective subtelomeric clones confirmed these aberrations. The parents of WHS1 and the mother of WHS2 were normal by FISH studies, but the father of WHS2 had a balanced translocation between 4p and 5q. High resolution G banding analysis indicated his karyotype to be 46,XY,t(4;5)(p15.3;q35.1). Thus, the karyotypes of WHS1 and 2 were interpreted as 46,XY.ish der(4)t(4;10)(p16;q26)de novo and 46,XY,der(4)t(4;5)(p15.3;q35.1)pat, respectively (table 2). In order to verify the accuracy of these micorarray CGH data, all subtelomeric regions were investigated by FISH in the five WHS cases, which confirmed the CGH data and supported the reliability of the method (table 2).

Idiopathic MR patients

We analysed 69 cases of idiopathic MR with or without MCA by microarray CGH. Subtelomeric rearrangements were detected in the following four cases (5.8%): N02R245, N03R003, N03R017, and N03R088 (table 2).

N02R245

A 2 year old girl was the second child of non-consanguineous healthy parents without any family history of MR or congenital anomalies. Her birth at 38 gestational weeks was uneventful, with a birthweight of 3130 g (+0.3 SD), length of 50 cm (+1SD) and Apgar score of 9 at 5 minutes. She showed moderate developmental delay. She could control her head at 5 months of age, and sit at 10 months, but could neither walk nor speak any meaningful words at 19 months. At the age of 21 months, her height was 80 cm (25th percentile), weight 9.6 kg (25th percentile), and head circumference 49 cm (90th percentile). Minor facial anomalies noted included epicanthal fold, depressed nose, low nasal bridge, and thin upper lip. A brain MRI study at 10 months did not reveal any abnormalities. A chromosomal gain of a 12p subtelomeric region was detected by microarray CGH (fig 2A). FISH analysis on metaphase chromosomes using the 12p subtelomeric clone GS-124K20 revealed that a signal on one homologous chromosome 12 was twice as large as that on the other (fig 3A), and duplicated signals and a separate single signal were detected on interphase cells (fig 3A), suggesting that tandem duplication of a 12p subtelomeric region had occurred. A similar finding was observed in her mother, whose phenotype was normal, but it was not observed in her father. High resolution G banding analysis of the patient and her mother did not reveal any differences in chromosome 12 (fig 4A). In order to see whether the 12p duplication was a normal variant, FISH analysis using GS-124K20 as the probe was performed on 50 normal Japanese controls, none of which showed the duplication.

N03R003

An 8 year old boy had severe MR, seizures, microcephaly, dysmorphic face, and other anomalies. Detailed clinical information has been described elsewhere as patient 2.¹² Chromosomal deletion of a 9q subtelomeric region was detected by micorarray CGH (fig 2B), and was confirmed by FISH using a 9q subtelomeric probe as described previously¹² (fig 3B). As FISH analysis confirmed that neither of his parents had such a deletion, it was considered a de novo mutation. G banding chromosome analysis could not identify the deletion in the patient (fig 4B).

N03R017

A 15 month old boy was the first child of non-consanguineous healthy parents without any family history of MR or congenital anomaly. His birth at 39 gestational weeks was uneventful. His birth weight was 3046 g (+0.25 SD), birth length 52 cm (+1.5 SD), and Apgar score 9 at 5 minutes. He presented with severe developmental delay. He could control his head at 5 months of age, and sit at 12 months, but could neither walk nor speak any meaningful words at 14.5 months old. At 1 year of age, his height was 75 cm (50th percentile), and weight 7.8 kg (3rd percentile). Minor craniofacial anomalies were microcephaly, flat occiput, epicanthal folds, depressed nasal root, long philtrum, low set ears, and micrognathia. Infantile leukaemia was observed at the age of 10 months, and chemotherapy was performed. A subtelomeric chromosomal deletion of 5p and a gain of 20p were detected by microarray CGH (fig 2C), and were confirmed by FISH (fig 3C). FISH analysis of his parents revealed that his phenotypically normal father showed a balanced translocation between 5p and 20p, but no morphological difference between homologous chromosomes 5 and 20 was observed by G banding analysis (fig 4C).

N03R088

A 6 year old girl was born at 41 gestational weeks uneventfully. Her birth weight was 3020 g (−1 SD), birth length 47 cm (−1.5 SD), and Apgar score 9/10 at 1 and 5 minutes. She was the second child of non-consanguineous healthy parents. She showed general hypotonia and severe developmental delay. She could control her head at 6 months of age, roll over at 9 months, and sit at 4 years, but could neither walk nor speak any meaningful words at 6 years of age. She had febrile seizures that occurred at 1 year of age, but never developed epilepsy. Craniofacial abnormalities were noted: flat occiput, square shaped face, hypertelorism, low set ears, downturned corners of the mouth, cleft soft palate, and micrognathia. Brain MRI revealed mild dilatation of the lateral ventricles, cavum septum pellucidi, and cavum vergae. A sacral dimple was identified, but MRI could not reveal any gross abnormalities. Microarray CGH analysis detected a subtelomeric chromosomal deletion of 10q and a gain of 12p (fig 2D), which were subsequently verified by FISH (fig 3D). High resolution G banding analysis revealed that her karyotype was 46,XX,der(10)t(10;12)(q26.3;p13.1) (fig 4D). A parental origin of the unbalanced translocation was not confirmed because parental materials were unavailable.

DISCUSSION

Molecular studies focusing on cryptic subtelomeric rearrangements in idiopathic MR patients have been performed mainly by FISH using subtelomere specific probes, and genotyping using hypervariable genetic markers.^1,4–6 Microarray CGH, a high throughput technology, was developed to identify changes in copy number of chromosomal segments or genes not only in somatic events (neoplasms), but also in germline derived events (constitutional chromosomal changes).^13–18 Recently, Veltman et al showed the efficiency of subtelomere specific microarray CGH in detecting known and novel cytogenetic abnormalities.⁷ We have developed subtelomere specific microarray CGH for MR. Its accuracy has been confirmed using five cases of WHS as positive controls, and we applied it to analysis of 69 patients with idiopathic MR. All the WHS cases showed a 4p subtelomeric deletion, and two were found to have either a 10q or a 5q duplication by microarray CGH. Four (5.8 %) of 69 cases of idiopathic MR were found to have subtelomeric rearrangements, the frequency being consistent with the data from recent reports.^1,4–6 The accuracy of our microarray CGH was verified by FISH using all subtelomeric probes in the five WHS cases; the genomic rearrangements were not overlooked by our microarray CGH in any of them. ANILFR of clones with either a deletion or a duplication, and in those without changes of the copy number were 0.628, 1.433, and 0.995, respectively, in all the patients. The two former ratios were beyond the 2SD range, while all other clones were within the range (figs 1 and 2). We used three different DOP-PCR primers, DOP-A (the original primer developed by Telenius et al and named as 6MW⁹), DOP-B, and DOP-C (both originally designed by us). DOP-PCR products using these primers should represent original clones more faithfully than those using a single DOP primer. Similarly, Fiegler et al¹⁰ adopted DOP-PCR using three different primers designed to amplify human specific DNA instead of contaminated bacterial DNA, and showed efficient detection of gains and deletions in a cancer cell line. Although our primers, DOP-B and DOP-C, were not designed for human specific DNA, DOP-PCR using the three primers was suitable for generating DNA for microarray. In addition, using the two simultaneous, independent hybridisations, CGH1 and CGH2, on one glass slide can eliminate conflicting results due to different conditions in each hybridisation, and detect genomic gains and deletions easily, because the data complement each other.

As well as the common 4p deletion, additional chromosomal abnormalities were detected in two (WHS1 and 2) of five WHS patients. Both showed a typical WHS phenotype, suggesting that the WHS critical region could be the “master” regulatory sequence on clinical manifestations with a strong pleiotropic effect,¹⁹ or that the effects of 10q26 duplication and 5q35.3 duplication could be minimal.

Cryptic subtelomeric rearrangements were detected in 4 of the 69 cases of idiopathic MR (table 2). A 9q34.3 terminal deletion was found in N03R003. This, together with the 4p terminal deletions identified in three WHS (WHS3, 4, and 5), might have originated from hidden unbalanced translocations that our microarray system could not detect. If so, breakpoints might exist distal to the subtelomeric clones that were used in microarray CGH. However, as each subtelomeric clone is located within 500 kb of the chromosomal end, a phenotypic effect of a hidden duplication derived from the unbalanced translocation may be subtle. A tandem duplication (ish dup(12)(p13.33p13.33)mat) detected in N02R245 may be a normal variant, because her phenotypically normal mother had the same abnormality, but we could not detect this duplication among 50 normal Japanese controls. Further FISH analysis using other BAC clones mapped at 12p13.33 showed that the duplicated region was mapped to a region 0–530 kb proximal to the 12p telomeric end according to the UCSC genome browser (data not shown), but the orientation (inverted or directed) of the tandem duplication was uncertain. Two unbalanced translocations, ish der(5)t(5;20)(p15.33;p13)pat in N03R017 and der(10)t(10;12)(q26.3;p13.1) in N03R088, were detected. The origin of the latter remained unknown because parental materials were not available. The 5p subtelomeric deletion in N03R017 merits comment; a BAC clone covering CTNND2 encoding delta-2 catenin associated with the critical deleted region of cri du chat syndrome (CdCS)²⁰ was not deleted in this patient (Okamoto, in preparation), supporting the absence of any CdCS phenotypes and indicating that the 5p deletion was less than 1.1 Mb distance from the 5p telomere. In N03R088, who had a 10q26.3 deletion, a triangular face was not observed as previously reported.²¹ A five item checklist was proposed by de Vries et al to improve the detection rate of subtelomeric rearrangements in MR subjects.²² Retrospective scoring using the checklist revealed that the mean score in our 69 cases was 5.8, and scores in N02R245, N03R003, N03R017, and N03R088, who had subtelomeric rearrangements, were 2, 6, 6, and 6, respectively. These figures support the previous findings, which showed that a score of 3 or higher had a sensitivity of 100% with rather low specificity of 27% and was therefore recommended as the cut off point.²² Several types of chromosomal aberrations were recently reported to be related to the olfactory receptor (OR) gene clusters at 8p23 and 4p16. A common inversion polymorphism between two of these clusters at 8p23 exists in the normal population, and mothers heterozygous for the inversion may have a risk of bearing children with a complex rearrangement, inv dup del(8p).^23,24 A similar inversion between two OR clusters at 4p16 has also been found.²⁵ Mothers with double heterozygous inversions at 8p23 and 4p16 tend to have an increased risk of t(4;8)(p16;p23) in their offspring,²⁵ and children with der(4)t(4;8)(p16;p23) may suffer from WHS owing to distal 4p deletions.^26–28 However no such translocation was found in our cases. To detect all submicroscopic chromosomal abnormalities, construction of a high resolution genomic microarray covering the whole genome is required.^13,14,16 Buckley et al reported a full coverage microarray for chromosome 22 facilitating successful detection of genomic aberrations in acral melanoma, dermatofibrosarcoma, DiGeorge syndrome, and neurofibromatosis type 2.¹⁷ We are now developing a high resolution microarray system that covers the whole human genome and are planning to identify genes responsible for genetic disorders including MR and MR related syndromes as research bases. In the near future, such microarrays targeting the whole genome including subtelomeric regions may be used in clinical diagnosis.