A 32-year-old, mentally retarded male was referred to our centre for further clinical genetic analysis. He was born to non-consanguineous parents after 42 weeks’ gestation with a birth weight of 3500 g. He had a healthy older brother. In the neonatal period he was hypotonic and at 8 weeks of age he underwent surgery because of an inguinal hernia with removal of an atrophic right testis. His motor development was severely delayed with sitting at 3.5 years and walking at 5 years of age. Speech was poorly developed, characterised by the usage of only a few words. During infancy an optic nerve hypoplasia was diagnosed, and during childhood he frequently suffered from luxations of the patellae, which required surgery. At the age of 32 years his height is 163 cm (−3 SDS) and head circumference 52.5 cm (⩽2.5 SDS). He has a narrow receding forehead, widened inner canthal distance of 3.5 cm (90th centile), normal outer canthal distance of 8.5 cm (25th centile), telecanthus, short and down slanting palpebral fissures, epicanthal folds, ptosis, long, straight eyelashes, high nasal bridge, low set large ears, flat philtrum, small mouth with high, narrow palate and retrognathia. The thorax is broad with increased internipple distance and slight gynaecomastia. A recent renal ultrasound revealed multiple cysts in the left, dystrophic kidney and two uncomplicated cysts in the enlarged, right kidney. The patient has a normally sized phallus with absent right testis and small left testis. His hands show a simian crease right and tapering fingers with broad proximal interphalangeal joints. He shows sandal gaps on both flat feet with clinodactyly of the fourth and fifth toes (fig 1). He has a wide, unsteady gait and he is moderately to severely mentally handicapped with normal pleasant behaviour.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 The proband at the age of 7 years (A) and 32 years (B–F). Note a narrow receding forehead, widened inner canthal distance, telecanthus, short and down slanting palpebral fissures, epicanthal folds, ptosis, long eyelashes, high nasal bridge, low set large ears, flat philtrum, small mouth with high, narrow palate, retrognathia and a broad thorax with increased internipple distance and slight gynaecomastia. Reproduced with proxy consent by parents.

After routine cytogenetic analysis on cultured peripheral blood lymphocytes, a normal male karyotype was found and subsequent analysis of the subtelomeric regions by multiplex ligase-dependent probe amplification (MLPA) on DNA extracted from uncultured peripheral blood cells did not reveal any abnormalities as well. The patient’s DNA was then subjected to array based comparative genomic hybridisation (CGH) analysis using our tiling path array that encompasses 32 447 bacterial artificial chromosome (BAC) clones covering the entire human genome as previously described.1 The genome profile (fig 2A) revealed a significant loss within the short arm of chromosome 2. A total of 40 BACs spanning a region of 3.9 Mb from Mb-positions 57.7 (2p16.1) to 61.6 (2p15) (fig 2B) had a significantly lowered test over reference log2 intensity ratio with RP11-426N8 and RP11-470L12 being the telomeric and centromeric breakpoint clones, respectively. The loss was validated and confirmed by region-specific MLPA analysis with probes designed in VRK2 and USP34 (fig 2C). According to ISCN 2005 nomenclature, the combined karyotype is described as follows: 46,XY.arr cgh 2p16.1p15(RP11-426N8→RP11-470L12)x1. Please note that the aberrant region detected by array CGH is indicated from pter to qter in the karyotype, whereas a similar deletion detected by fluorescence in situ hybridisation (FISH) would be indicated with the breakpoint more proximal to the centromere specified first—that is, 46,XY.ish del(2)(p15p16.1)(FANCL-,REL-). Next, four BAC clones from the aberrant region were selected and labelled for validation by FISH on metaphase spreads from our patient. All four probes were hybridised individually and on separate slides. With each probe, only one signal was observed in 20/30 cells, whereas the remainder of the cells revealed a normal pattern with two signals, one on each 2p (data not shown). We conclude from this that the microdeletion was present in mosaic in our patient with both normal cells and abnormal cells. Parental samples were subsequently analysed by routine chromosome analysis and by FISH analysis with the aforementioned probes. In both parents, a normal karyotype and a normal FISH pattern was observed. The microdeletion in our patient has occurred de novo, which is to be expected considering the fact that it appeared to be present in mosaic. In addition to other reports describing the detection of even low level mosaics by array CGH analysis,2 3 this is yet another example demonstrating the strength of array CGH not only to reliably detect imbalances smaller than 10 Mb, but also the ability to detect unbalanced chromosome abnormalities present in mosaic. Convincing data are accumulating that seem to confirm that aberrant cells may be underrepresented in the PHA stimulated, cultured T lymphocytes that are used for routine cytogenetic analysis. The aberrant cells are more difficult to stimulate to divide and are therefore more likely to be detected with array CGH analysis using DNA isolated directly from whole blood samples.3

• Download figure
• Open in new tab
• Download powerpoint

Figure 2 A–C. Detection of a microdeletion within the short arm of chromosome 2 by tiling resolution genome wide array comparative genomic hybridisation (CGH) analysis confirmed by multiplex ligase dependent probe amplification (MLPA). The genome profile of the proband (A) after array CGH analysis (hybridisation against a sex matched reference DNA pool). Each dot represents the test over reference log2 intensity ratio (Y axis) of a certain bacterial artificial chromosome (BAC) clone. These values are ordered along the X axis according to the location of each BAC on the genome from chromosome 1 to chromosome Y and from pter to qter. The threshold for a single copy gain or loss is set at 0.3 and −0.3, respectively. Multiple BACs of chromosome 2 show a significantly lowered intensity ratio. This is further depicted in detail in the plot of chromosome 2 (B) where multiple BACs show a copy number loss from 57.7 Mb (2p16.1) to 61.6 Mb (2p15). The loss was confirmed by region specific MLPA analysis with synthetic probes designed in VRK2 (2p16.1) and USP34 (2p15). Graph (C) depicts the relative amount of product per probe (Y axis) for each of these region specific MLPA probes as well as five control probes (X axis). The results from the patient’s DNA are shown together with the results form both parents (duplicate samples) as well as six control samples. A decreased value of the specific MLPA probes in our patient confirms the loss in 2p15p16.1. D. Schematic representation of the commonly deleted region in 2p15p16.1. Screenshot obtained from http://genome.ucsc.edu/cgi-bin/hgTracks, UCSC Genome Browser on Human March 2006 Assembly. The upper part of the figure shows the idiogram of chromosome 2 in which the aberrant region is indicated by a red square. This region is enlarged below the idiogram with from top to lower part of the figure the Mb position, the chromosome bands, the FISH-mapped clones (of which the majority are present on the 1 Mb BAC array), known genes, segmental duplications and putative structural variations as previously reported by various groups. The five and seven BACs deleted in the respective two patients of Rajcan-Separovic et al4 are indicated by blue squares, the distal two ones are indicated with dashed lines because these were deleted in only one of the two patient. The deletion detected in our patient is indicated by the two vertical red lines and equals the common deleted region as it falls within the aforementioned deleted regions of the other two patients. The green rectangle indicates the apparently variable region.

The 3.9 Mb loss in 2p15p16.1 detected in our patient falls within the deleted regions with respective approximate sizes of 4.5 and 5.7 Mb that were present in two patients recently described by Rajcan-Separovic et al.4 All three patients share a number of malformations and rather specific dysmorphic features, including microcephaly and optic nerve hypoplasia, as well as moderate to severe intellectual disability (table 1).

Considering the phenotype of our patient and the two patients described by Rajcan-Separovic et al,4 the presence of normal cells in our patient does not seem to have a tempering effect on the severity of the phenotype. However, it remains difficult to determine to what extent the presence of normal cells will have an effect on the phenotype, in particular because the percentages of normal and abnormal cells may vary from one tissue to another. Although Rajcan-Separovic and colleagues4 used a 1 Mb resolution array while we applied a tiling resolution array, it is clear that the distal breakpoints in all three patients differ significantly (fig 2D), whereas the proximal breakpoint seems consistent. The mutual region of overlap that is lost in all three patients contains a total of 15 protein-coding genes. Six genes have previously been reported (VRK2, FANCL, BCL11A, PAPOLG, REL, and PEX13), three genes have recently been identified (KIAA1841, AHSA2, and USP34), and six genes are encoding a hypothetical protein (hCG_15200, LOC730134, LOC647038, LOC730142, FLJ32312, and LOC339804). In addition, the region also contains six pseudogenes and three unknown candidates.

Copy number variable regions (CNVRs) have recently been identified on chromosome 2 from base pairs 58306194 to 61252101 by both Whole Genome TilePath (WGTP) array CGH and by comparative intensity analysis using the Affymetrix 500 K EA platform in the so-called HapMap collection.5 In 270 healthy individuals, only one gain was detected in this region in a single individual; no losses were observed. This apparently variable region contains FANCL, BCL11A, PAPOLG, REL and PEX13 as well as FLJ32312 and KIAA1841, and at least for FANCL and PEX13, a gain is not expected to cause a clinical phenotype. In addition, McCarroll et al already had identified a 2 kb deletion in this region (no genes known) around 59.5 Mb.6 Although it is clear that copy number variation in this region does rarely occur in healthy individuals, one should not exclude the possible clinical significance of a deletion leading to haploinsufficiency of one or more genes in this region. Another possibility could be that haploinsufficiency of the aforementioned seven genes is not the (only) cause of the phenotype of the three patients, but that haploinsufficiency of the remaining three genes (VRK2, AHSA2 and USP34) outside the CNVR, but inside the common deleted 2p15p16.1 region (fig 2D), contributes to the clinical phenotype observed. As yet, there is no evidence for the presence of segmental duplications flanking the deleted region in 2p15p16.1, not even up to 5 Mb outside the deletion breakpoints. Therefore, the local genomic architecture and the mechanism by which this recurrent microdeletion arises are yet to be determined.