Utility of Oligonucleotide Array–Based Comparative Genomic Hybridization for Detection of Target Gene Deletions

Materials and Methods

Results

In the patient with citrin deficiency (case 1), array CGH showed a heterozygous deletion within the SLC25A13 gene on chromosome 7q21.3 that involved multiple consecutive oligonucleotide probes (Fig. 1A⇓ ). The hybridization data were consistent with a deletion of approximately 4.6 kb that included parts of introns 2 and 3 as well as all of exon 3. PCR amplification across the putative breakpoints (Fig. 1B⇓ ) and sequencing of the isolated PCR product identified the breakpoints within the SLC25A13 gene at c.70–862 and c.212 + 3527 (Fig. 1C⇓ ), which correspond to a 4 532-bp deletion. This heterozygous deletion was also confirmed in the proband’s asymptomatic sister and her father by direct DNA sequencing of the junction fragment and by qPCR (data not shown). Previous direct sequencing of the patient’s DNA had revealed a novel heterozygous alteration, c.848 + 3 A>C (IVS8 + 3A>C), which SpliceView software (see http://bioinfo.itb.cnr.it/oriel/splice-view.html) had predicted to produce a loss of a splice donor site. We confirmed the aberrant splicing of the c.848 + 3 A>C mutation by cDNA sequencing; in addition, finding these mutations in the proband’s mother confirmed the trans configuration of these mutations.

• Download figure
• Open in new tab

Figure 1.

Heterozygous intragenic deletion of SLC25A13 in patient 1.

(A), Array CGH results. The left panel shows a whole-chromosome view of data from chromosome 7. Note the clustered oligonucleotide coverage at 9 genes of interest. Results for oligonucleotides showing the typical copy number (log₂ ratio = 0 ± 0.25) are shown in black, whereas those in green or red represent log ratios outside this range indicating copy number loss or gain, respectively. The boxed region at 7q21.3 containing SLC25A13 was the only location where a copy number loss was detected with >5 consecutive oligonucleotides. The middle panel is an expansion of this region, and the right panel is a further blowup of the small boxed region, indicating a heterozygous deletion by 11 of 13 consecutive oligonucleotide probes within the region indicated by the sequence coordinates. (B), Primer positions for PCR amplification of the deletion junctions. Exon 3 (E3) is shown as a box with the beginning and ending cDNA nucleotide positions indicated. (C), Sequencing results show the breakpoint junctions to be at c.70−862 in intron 2 and c.212+3527 in intron 3, a 4532-bp deletion.

In the patient with progressive intrahepatic cholestasis (case 2), array CGH detected a homozygous deletion at the ABCB11 locus on chromosome 2q31.1 that involved >50 oligonucleotide probes (Fig. 2A⇓ ). Sequence analysis of the junction region revealed breakpoints consistent with the oligonucleotide array CGH data that indicated deletion of sequence from c.1309+1048 to c.2076−610 (Fig. 2B⇓ ). This 10.5-kb deletion contains part of intron 12, all of exons 13–17, and part of intron 17 (Fig. 2C⇓ ). Unfortunately, parental samples for this patient were not available for comparison.

• Download figure
• Open in new tab

Figure 2.

Homozygous intragenic deletion of the ABCB11 gene in patient 2.

(A), Array CGH results showing a whole–chromosome 2 view on the left and a blowup of the ABCB11 gene region at 2q31.1 on the right. The green dots indicate a region with log₂ ratios for hybridization far below the −1 value expected for a 1:2 copy number ratio; >50 oligonucleotide probes indicate the presence of a homozygous deletion within this approximate sequence region. (B), The sequencing results for the junction region show the deletion to be from c.1309+1048 in intron 12 to c.2076−610 in intron 17, a 10.5-kb deletion. (C), Schematic of the breakpoints and deleted exons in the ABCB11 gene.

A large deletion of part of one X chromosome (from Xp11.4–21.2) was detected with the gene-based array in the infant girl with hyperammonemia (case 3). This approximately 7.4-Mb deletion extended from the telomeric side of DMD to within exon 4 of the OTC gene on the centromeric side (data not shown). Because the oligonucleotide array was targeted only to specific genes of interest, we also analyzed the DNA with an Agilent 244K whole-genome oligonucleotide array to identify the breakpoints (Fig. 3⇓ ). The deletion breakpoints were determined to be at approximately X:30.7 Mb and X:38.1 Mb. Analysis of DNA from parental blood samples in case 3 (Fig. 3⇓ ) did not detect any deletions, suggesting germ line mosaicism or a de novo mutation.

• Download figure
• Open in new tab

Figure 3.

OTC deletions.

An enlarged view of CGH data from an Agilent 244K array analysis for the OTC-DMD region of the X chromosome, showing the extent of the deletion in patient 3 (left panel), the lack of deletion in this region in both parents of patient 3 (center panels), and the extent of the deletion in patient 4 (right panel). Note that the chromosomal breakpoints in patients 3 and 4 are different. Patient 4 shows a discontinuous pattern at the distal breakpoint; whether this pattern is due to a complex deletion mechanism or to the coincidental occurrence of a polymorphic copy number variation in this region is not known.

We did not detect CPS1 deletions in the patient of case 4, who also presented with hyperammonemia; but instead we found a deletion of the OTC gene. An Agilent 244K array confirmed an approximately 9-Mb deletion encompassing a region of the X chromosome similar to that of patient 3 (Fig. 3⇑ ). The proximal breakpoint was centromeric to the OTC gene at approximately X:38.4 Mb. The distal region of the deletion was more complex, with an approximately 0.2-Mb deleted region followed by a 0.24-Mb nondeleted region and a large 8.84-Mb deletion. We were not able to obtain parental DNA samples for this patient to evaluate the inheritance of these regions; therefore, we do not know whether this pattern is due to a complex rearrangement mechanism or a small polymorphic copy number variant in this region.

Several known genes are in the X-chromosome region between the DMD and OTC genes, including: (a) CYBB [cytochrome b-245, beta polypeptide (chronic granulomatous disease)], which encodes the cytochrome b β chain involved in superoxide production and mutations of which cause chronic granulomatous disease; (b) XK [X-linked Kx blood group (McLeod syndrome)], which is the X-linked Kx blood group of the kell precursor and mutations of which have been associated with McLeod syndrome, an X-linked, recessive disorder characterized by abnormalities in the neuromuscular and hematopoietic systems; and (c) RPGR (retinitis pigmentosa GTPase regulator), the X-linked retinitis pigmentosa GTPase regulator isoform B. Deletion of these genes, however, would not explain these patients’ clinical presentations with a proximal urea cycle defect.

Discussion

Oligonucleotide array–based CGH has now been routinely used to detect large chromosomal deletions (9)(11). Most of these assays monitor regional changes that may affect multiple genes, although large rearrangements impacting a single gene have been investigated in a few cases. Examples of the latter application include the study of large rearrangements in MECP2 [methyl CpG binding protein 2 (Rett syndrome)], the gene responsible for Rett syndrome and X-linked mental retardation (12), and an investigation of copy number changes in the DMD gene, which are responsible for Duchenne muscular dystrophy (8). We hypothesized that similar approaches could be used to simultaneously screen for intragenic deletions within a group of genes relevant to specific related disorders. We were particularly interested in the genes involved in mitochondrial and metabolism-related disorders to complement the gene-sequencing analyses in our laboratory.

The use of array CGH for gene-based analyses has several advantages over alternative methods, such as MLPA, qPCR, and fluorescence in situ hybridization analysis. In our experience, the initial setups of the latter 3 assays are time consuming, and their performance requires costly validation. In addition, maintaining proficiency with these tests in a low-throughput clinical-testing environment is problematic. Furthermore, these techniques, which are typically designed to analyze 1 or 2 probes at a time, usually require multiple measurements. Such hurdles may have inhibited the rapid introduction of deletion or duplication assays into the clinical-testing sphere as new disease genes have been discovered. The availability of well-characterized oligonucleotides covering the whole human genome from established databases reduces the time to create and validate probes when one compares the time to do the same with MLPA- or qPCR-based assays. Unlike qPCR or MLPA, which evaluates a single region in an exon, this highly multiplexed technology allows us to detect deletions in flanking introns as well as partial exon deletions.

It is advantageous to have oligonucleotide probes on the same array for genes involved in the same pathway or for similar diseases. For example, the patient in case 4 was originally suspected of having a CPS1 deficiency, but a deletion involving the OTC gene was discovered. Similarly, citrin deficiency often presents with an increased citrulline concentration, but it must be distinguished from disease caused by mutations in ASS1 (argininosuccinate synthetase) and ASL (argininosuccinate lyase). Furthermore, the molecular etiology of intrahepatic cholestasis may involve mutations in ATP8B1 (PFIC1), ABCB11 (PFIC2), or ABCB4 [PFIC3; ATP-binding cassette, sub-family B (MDR/TAP), member 4]. Thus, it is convenient to be able to analyze all 3 of these genes simultaneously.

Large genes may have an inherently higher risk for intragenic deletions. Both the ABCB11 (PFIC2) and SLC25A13 (citrin) genes are large. The ABCB11 gene consists of 27 coding exons spanning 108 kb. The citrin gene consists of 18 exons spanning 201 kb; introns 2 and 3 are particularly large (19.5 kb and 42.2 kb, respectively). A heterozygous 4.6-kb deletion in the citrin gene, consisting of a single exon and flanking intronic sequences, and a homozygous 10-kb deletion in the ABCB11 gene were readily detectable on the oligonucleotide-based microarray assay. On the other hand, OTC is not a large gene. A recent review reported 341 different point mutations (13). The majority of these mutations were private mutations, and, interestingly, mutations were not detected in 80% of patients. Conversely, 6%–20% of OTC mutations have been reported to be large deletions involving all or part of the gene (14)(15)(16). Therefore, the possibility of deletions should be considered in clinically and biochemically confirmed OTC-deficient patients. This consideration is particularly important in female patients, because heterozygous intragenic deletions cannot be detected with the traditional sequencing approach. Our results demonstrated that 2 disease-manifesting female patients in fact had large deletions in this region that included all or part of the OTC gene, suggesting that this region on the X chromosome may be particularly prone to deletions.

In summary, our experience suggests that oligonucleotide-based assays offer a valuable tool for clinical analysis of intragenic deletions, duplications, and rearrangements. In particular, we have demonstrated that targeted oligonucleotide arrays can be of great utility when used in conjunction with direct DNA sequencing in the context of autosomal recessive diseases when only one heterozygous mutant allele is detected. We are currently further evaluating extension of this oligonucleotide array CGH approach to the simultaneous analysis of copy number mutations in nuclear genes, mitochondrial DNA depletion, and single deletions in the mitochondrial genome leading to heteroplasmy.