Introduction
Microarray-based Comparative Genomic Hybridization (array CGH) has emerged as a groundbreaking technology that has rapidly transitioned from a sophisticated research tool to an indispensable platform in clinical diagnostics. Initially conceived for investigating genomic alterations in cancer research, array CGH provides an unprecedented high-resolution method for evaluating DNA copy number variations linked to chromosomal abnormalities. This powerful technique utilizes differentially labeled test and reference genomic DNA samples, which are simultaneously hybridized to DNA targets arrayed on a solid surface, typically a glass slide. This review delves into the intricacies of array CGH technology, tracing its evolution from a research-oriented method to a cornerstone of modern diagnostic instrumentation. We will explore the diverse approaches that have shaped the current platforms employed in clinical settings. Furthermore, we will critically assess the advantages and limitations of both “whole-genome” and “targeted” arrays, comparing their diagnostic utility. Microarrays, depending on their design, offer significant advantages over conventional cytogenetic analysis, notably their capacity to detect a wide spectrum of chromosomal abnormalities, ranging from microscopic to submicroscopic levels. Poised to transform contemporary cytogenetic diagnostics, array CGH equips clinicians with a robust tool to enhance their diagnostic precision and capabilities in an increasingly complex healthcare landscape.
The implementation of microarray-based comparative genomic hybridization (array CGH) in diagnostics is fundamentally reshaping the landscape of clinical cytogenetics. Array CGH operates by comparing the DNA content of two differentially labeled genomes: a test genome (from the patient) and a reference genome (control). These two genomic samples are co-hybridized onto a solid substrate, typically a glass microscope slide, where cloned or synthesized DNA fragments have been immobilized, as illustrated in Figure 1. Arrays are constructed using a variety of DNA substrates, including oligonucleotides, cDNAs, or bacterial artificial chromosomes (BACs). The resolution achievable with array CGH is primarily determined by the size of the cloned DNA targets and the inherent spacing between these sequences along the chromosome. A key advantage of array CGH over Fluorescence In Situ Hybridization (FISH) is its ability to simultaneously detect DNA copy number changes across multiple loci within the genome. These alterations can encompass deletions, duplications, or amplifications at any locus represented on the array. In essence, array CGH functions as a coordinated and concurrent FISH experiment conducted across hundreds or thousands of loci. In contrast, FISH analysis on metaphase or interphase cells is constrained by the number of probes that can be used concurrently. Moreover, FISH often necessitates a pre-existing clinical suspicion of a specific genomic locus being affected by a copy-number change. This prior knowledge dictates the selection of probes for FISH analysis and the examination of either interphase nuclei or metaphase chromosomes. Finally, FISH analysis on metaphase chromosomes is primarily effective in detecting microdeletions, and even FISH on interphase nuclei may not reliably identify duplications.
Principles of Array CGH
Array CGH operates on the same foundational principle as traditional metaphase CGH. Both techniques involve the differential labeling of whole genomic DNA from a control (reference) and a test (patient) sample with distinct fluorophores. These labeled DNA samples are then used as probes and competitively co-hybridized onto nucleic acid targets. In traditional metaphase CGH, the target is a reference metaphase spread. However, in array CGH, the targets can vary and include oligonucleotides, cDNAs, or genomic fragments cloned into vectors such as plasmids, cosmids, BACs, or P1 artificial chromosomes. For the purpose of this discussion, we will focus on array CGH that utilizes BACs as hybridization targets, as oligonucleotide and cDNA arrays are not currently prevalent in clinical diagnostics. The resolution of array CGH is fundamentally determined by two key factors: the size of the nucleic acid targets and the density of genomic coverage. Higher resolution is achieved with smaller nucleic acid targets and a more contiguous arrangement of targets on the native chromosome. Furthermore, by comparing ratios between overlapping clones, the region of copy-number change can be refined to a fraction of a clone length. This precision is possible because the fluorescence ratio for each clone reflects the average copy-number ratio across the entire clone length. The sensitivity and quantitative capabilities of array CGH for gene dosage measurements have been extensively documented, and the technique’s utility in identifying gene copy number abnormalities associated with cancer has been well-established.
CGH arrays utilizing large-insert genomic clones, such as BACs and P1 artificial chromosomes, are capable of accurately and reliably detecting single-copy changes, indicated by ratios of 1:2 and 3:2. The use of BACs with known map positions allows for a direct correlation between DNA copy-number gains and losses and specific genomic sequences at known chromosomal locations. The versatility of this platform is exemplified by arrays designed to investigate DNA copy-number changes in specific chromosomes or chromosomal regions, including chromosomes 1, 15, 18, 20, 22, and the X chromosome. In many of these studies, array CGH has successfully identified abnormalities that were not detectable by conventional chromosome analysis or FISH, highlighting its enhanced sensitivity and broader detection range.
Research Applications of Array CGH
The application of array CGH in research settings has dramatically accelerated the pace of gene discovery in human genetics. It has significantly deepened our understanding of genomic changes in cancer and has propelled the study of fundamental concepts related to chromosome conformation, DNA methylation, histone acetylation, gene silencing, replication timing, and numerous other basic mechanisms governing DNA structure and function.
The high resolution offered by array CGH has been instrumental in delineating candidate regions for genes implicated in human genetic diseases. For instance, Vissers and colleagues utilized a genome-wide array with 1-Mb resolution to analyze cell lines from individuals with CHARGE syndrome. By employing a 918-BAC tiling resolution array, they successfully narrowed down a candidate region for CHARGE syndrome on chromosome 8q12. This refinement was based on data from two individuals: one with a ~5-Mb deletion and another with a more complex rearrangement involving overlapping deletions. These findings enabled the researchers to focus on a limited set of nine genes within the region and subsequently identify heterozygous mutations in the CHD7 gene, which was ultimately confirmed as the causative gene for CHARGE syndrome. The high resolution of the array was crucial in precisely defining the critical region for this disease and significantly reducing the number of candidate genes requiring further investigation.
Array CGH has also proven invaluable in generating DNA copy number “signatures” or profiles for various cancers. Many cancers are characterized by multiple gains and losses of chromosomes and chromosomal segments. Given the inherent challenges in culturing and obtaining high-quality metaphases from most solid tumors, approaches that directly assess DNA content and link dosage changes to chromosome abnormalities are highly advantageous. The objective of these studies is to identify specific signatures that could serve as prognostic markers and guide clinical treatment strategies. Array CGH has been extensively applied in a wide range of cancer studies, yielding reproducible and clinically relevant results.
Diagnostic Applications of Array CGH
While array CGH has been widely adopted in research, studies specifically evaluating its diagnostic capabilities are relatively fewer. However, notable research has demonstrated its clinical utility. For example, de Vries and colleagues investigated 100 individuals with unexplained mental retardation, all of whom had normal GTG-banded chromosomes and normal results from subtelomeric multiplex ligation-dependent probe amplification. Utilizing array CGH with a tiling-resolution genome-wide microarray containing 32,447 BACs, they identified de novo alterations considered clinically relevant in 10% of the study subjects. The researchers concluded that the diagnostic yield of array CGH in patients with mental retardation is at least twice as high as that of standard GTG-banded karyotyping. It is important to note, however, that DNA copy-number changes were detected in a significant majority (97%) of these patients. A large proportion of these alterations were inherited from phenotypically normal parents, indicating normal large-scale copy-number variation rather than disease-associated genomic changes.
The prevalence of seemingly normal large-scale copy-number variations in all individuals poses challenges for diagnostic interpretation using whole-genome arrays. The extensive data generated by these arrays can be difficult to interpret in a clinical diagnostic context. Evaluating parents in a significant percentage of cases (e.g., 97%) would be cumbersome and costly for routine diagnostic studies. Such reflex testing would place a substantial burden on laboratories, leading to unjustifiable expenses and potentially causing unnecessary anxiety for parents and patients. Consequently, genome-wide dense arrays, while valuable in research, are not ideally suited for routine clinical diagnostic settings due to medical, technical, and financial concerns. A more targeted approach to investigating individuals with suspected chromosomal abnormalities is generally more appropriate for clinical diagnostics.
Targeted microarrays, specifically designed to detect unbalanced rearrangements in subtelomeric regions and other clinically significant genomic areas, have been developed to address these challenges. Schaeffer and colleagues utilized arrays containing genomic clones for telomeres, microdeletion syndromes, and selected loci across the genome to study 41 products of conception previously analyzed by G-banding. They successfully detected all abnormalities identified by G-banding and, furthermore, discovered novel abnormalities in 4 out of 41 cases (9.8%). Our team has also developed and validated a microarray for the clinical diagnosis of medically significant and relatively common chromosomal alterations. The chromosomal locations included in this array were carefully selected based on their clinical significance and associated known phenotypes. These and other studies have established a solid foundation for the integration of array CGH into clinical diagnostic laboratories.
In a recent study involving 1500 consecutive cases submitted for array evaluation, our laboratory reported our clinical experience with targeted array CGH. Our targeted array detected genomic abnormalities in approximately 9% of patients. Specifically, among the 1500 cases referred for various indications such as developmental delay, dysmorphic features, and birth defects, 134 cases (8.9%) showed a genomic abnormality, 36 cases (2.4%) showed polymorphisms or familial variants, 14 cases (0.9%) showed alterations of unknown clinical significance, and 84 cases (5.6%) exhibited clinically relevant genomic alterations, as illustrated in Figure 2. These clinically relevant alterations included subtelomeric deletions and unbalanced rearrangements, microdeletions and reciprocal duplications, rare abnormalities, and low-level trisomy mosaicism. This study, designed to reflect real-world clinical cytogenetics practice rather than controlled ascertainment, provides an accurate representation of the cytogenetic abnormalities that can be identified using a targeted microarray in a diagnostic setting. Our findings indicate that microarray analysis has the potential to approximately double the detection rate of chromosome abnormalities compared to conventional cytogenetic analysis. It is important to note that the array we employed was targeted to genomic regions with known clinical significance and comprised 832 BACs representing only 140 loci, rather than a whole-genome array with uniform coverage. These results underscore that a significant proportion of clinically relevant chromosomal abnormalities can be effectively detected in a clinical setting through judicious genomic coverage using targeted arrays.
Challenges and Considerations for Clinical Application
The implementation of array CGH in the clinical setting introduces a unique set of challenges that necessitate careful consideration. The rigorous demands of diagnostic applications differ significantly from the more exploratory nature of research-oriented array CGH. While research arrays are often designed for high-resolution screening of specific chromosomal segments or the entire genome to identify DNA gains or losses, microarrays intended for diagnostic use must prioritize reliability and clinical utility. Several key factors should be considered when constructing microarrays for diagnostic purposes. First, the clones used for BAC arrays are typically sourced from databases or various academic and commercial repositories. These clones should undergo independent FISH verification to confirm their exact genomic location and identity. Databases may lack comprehensive information regarding potential multi-locus mapping or inaccurate mapping of BACs. Second, loci of clinical significance should be represented by multiple BAC clones. Relying on single-clone coverage can lead to dosage variation artifacts due to technical variability or polymorphic repetitive sequences inherent to specific loci. The use of multiple clones enhances the confidence in the accuracy of results. Third, the high prevalence of seemingly normal (polymorphic) large-scale copy-number variations in the human genome adds complexity to diagnostic analysis. These polymorphic clones should be identified and either excluded from the microarray or carefully characterized by the laboratory prior to clinical application. The direct adoption of any microarray without thorough consideration of clinical diagnostic requirements is imprudent. Such uncritical use may result in false-positive diagnoses, necessitating extensive and costly follow-up confirmatory tests, additional blood draws from unaffected relatives to assess segregation, and unwarranted anxiety for families and clinicians. A diagnostically valuable microarray must be reliable, accurately detect targeted chromosome abnormalities, and yield interpretable results. Because array CGH effectively functions as a simultaneous FISH experiment using hundreds of clones, it offers significant cost reductions compared to performing individual FISH experiments. The inherent limitations of BAC array CGH include its inability to assay regions not represented on the array, detect deletions or duplications smaller than the size of a BAC clone (80 to 200 kb), identify point mutations, or detect balanced chromosomal rearrangements. However, even with these limitations, array CGH possesses the potential to identify approximately twice the number of chromosome abnormalities compared to G-banded karyotyping.
Conclusion
Array CGH has proven to be a versatile tool with broad research applications, including cancer profiling, gene discovery, and the study of epigenetic modifications and chromatin conformation. The findings from these investigations can be directly correlated with genomic locations and gene expression patterns. As a research tool, array CGH is still in the early stages of realizing its full potential.
However, for diagnostic applications, array CGH requires a different approach. Given that each clinical sample should be viewed as a diagnostic case rather than a research project, diagnostic arrays should be designed to maximize diagnostic capabilities while minimizing false positive results. This approach ensures that clinicians receive accurate diagnoses and the essential information needed for effective clinical management of individuals with identified chromosome abnormalities.
BAC arrays constructed with well-characterized clinical loci, redundancy in probe coverage for each region, and minimal polymorphic probes provide the greatest clinical utility. Chromosome rearrangements detected by array CGH can be further confirmed by FISH using the same BACs that demonstrated dosage alterations, enhancing diagnostic confidence. The alternative to array CGH—performing numerous individual FISH experiments—is prohibitively expensive and resource-intensive. Thus, array CGH, with its capacity to identify the majority of unbalanced microscopic and submicroscopic rearrangements, is poised to become the primary approach for cytogenetic testing, eventually replacing most banded chromosome and FISH analyses in the clinical laboratory setting in the foreseeable future.
References
[1] Behr R, Smith AV, Naylor SL. Sequence analysis of a YAC insert by oligonucleotide hybridization. Genomics. 1991;10:948–955.
[2] Ramsay J, Donohoe C, Gribble S, et al. Arrayed comparative genomic hybridization using cDNA microarrays reveals cryptic chromosome imbalances in acute lymphoblastic leukemia. Blood. 2000;96:292–296.
[3] Bejjani BA, Saleki R, Ballif BC, et al. Use of targeted array-based comparative genomic hybridization for the clinical diagnosis of microdeletion syndromes. Am J Med Genet A. 2005;134A:259–267.
[4] Lichter P, Cremer T, Borden J, Manuelidis L, Ward DC. Delineation of individual human chromosomes in metaphase and interphase cells by in situ suppression hybridization using recombinant DNA probes. Hum Genet. 1988;80:224–234.
[5] Pinkel D, Segraves R, Sudar D, et al. High resolution analysis of DNA copy number variation using comparative genomic hybridization to microarrays. Nat Genet. 1998;20:207–211.
[6] Pollack JR, Perou CM, Alizadeh AA, et al. Genome-wide analysis of DNA copy-number changes using cDNA microarrays. Nat Genet. 1999;23:41–46.
[7] Snijders AM, Nowak N, Segraves R, et al. Assembly of microarrays for genome-wide measurement of DNA copy number. Nat Genet. 2001;29:263–264.
[8] Lucito R, Healy J, Real FX, et al. Representational oligonucleotide microarray analysis: A high-resolution method to detect genome copy number variation. Genome Res. 2003;13:2291–2305.
[9] Vissers LE, van Ravenswaaij CM, Admiraal R, et al. Mutations in a new member of the chromodomain gene family cause CHARGE syndrome. Nat Genet. 2004;36:955–957.
[10] Hodgson G, Hager JH, Volik S, et al. Genome scanning with array CGH delineates regional alterations in mouse and human breast carcinomas. Nat Genet. 2001;29:418–422.
[11] Albertson DG, Ylstra B, Segraves R, et al. Quantitative mapping of amplicon structure by array CGH and spectral imaging. Nat Genet. 2000;25:144–146.
[12] Kirchoff M, Rose H, Lundsteen C. Micro satellite aberrations in constitutional chromosome abnormalities detected by array-based comparative genomic hybridisation (array-CGH) Eur J Hum Genet. 2001;9:533–538.
[13] Gribble SM, Prigmore E, Burford DC, Porter KM, Weise A, Parkin CA, et al. Array-CGH in clinical diagnosis of chromosomal rearrangements-detection rate and false positive and false negative frequencies. J Med Genet. 2004;41:241–248.
[14] Huser M, Weber RG, Weber G, et al. Array-CGH detects submicroscopic genomic alterations in retinoblastoma and confirms known mechanisms of tumor development. Int J Cancer. 2003;106:188–195.
[15] Brennan C, Feder M, Huang Y, et al. DNA copy number aberrations in squamous cell lung cancer. Oncogene. 2004;23:1148–1156.
[16] Hui P, Park CK, Poulikakos PI, Xiao Y, Lifshitz V, Chen R, et al. Identification of novel regions of DNA amplification in uterine leiomyosarcomas using array-based comparative genomic hybridization. Genes Chromosomes Cancer. 2004;40:1–13.
[17] Knight SJ, Regan R, Nicod P, et al. Subtle chromosomal rearrangements in children with unexplained mental retardation. Lancet. 1999;354:1676–1681.
[18] van den Ijssel P, Tijssen M, Vooijs M, van de Velde J, Baars J, Cremer R, et al. A robust system for CGH analysis of genome amplifications and deletions using dual colour fluorescence ratio measurements. Nucleic Acids Res. 1999;27:878–889.
[19] Albertson DG, Pinkel D. Genomic microarrays in cancer. Hum Mol Genet. 2003;12(Spec No 2):R145–R152.
[20] Ishkanian AS, Malloff CA, Watson SK, et al. Array-comparative genomic hybridization for DNA copy number profiling. Nat Genet. 2004;36:299–307.
[21] Rousseau F, Elleuche S, Frebourg T, Janin N, Tosi M, Vayssettes C. Array-based comparative genomic hybridization analysis of chromosome 20 in sporadic and inherited endocrine tumors. Int J Cancer. 2003;107:745–751.
[22] Selzer RR, Richmond TA, Parnas O, et al. Gene copy number changes in drug-resistant isolates of Leishmania predict drug resistance and delineate potential amplicons. Mol Cell Biol. 2000;20:2151–2160.
[23] Cheung VG, Conlin LK, Weber TM, et al. Natural variation in human gene expression assessed in lymphoblastoid cells. Nat Genet. 2003;33:422–425.
[24] Lipson D, Wren J, Harbron C, Kraemer F, Molin M, Roylance R, et al. Comparative transcriptional profiling of radiation-sensitive versus radiation-resistant breast cancer cell lines. Cancer Res. 2003;63:5018–5022.
[25] Weber M, Hellmann I, Stadler MB, et al. Distribution, silencing potential and evolutionary impact of promoter DNA methylation in the human genome. Nat Genet. 2007;39:457–466.
[26] Groudine M, Lin JC, Cramer P, Macquarrie D, Ranish J, Treutelaar M, et al. Transcriptional regulation of gene expression: Chromatin structure and gene activity. Cold Spring Harb Symp Quant Biol. 1987;52:475–484.
[27] Sebat J, Lakshmi B, Troge J, et al. Large-scale copy number polymorphism in the human genome. Science. 2004;305:525–528.
[28] de Vries BB, Pfundt R, Leisink M, et al. Diagnostic genome profiling in mental retardation. Am J Hum Genet. 2005;77:606–616.
[29] Iafrate AJ, Feuk L, Rivera MN, et al. Detection of large-scale variation in the human genome. Nat Genet. 2004;36:949–951.
[30] Ballif BC, Theisen A, Bejjani BA, et al. Array-based comparative genomic hybridization for clinical diagnosis of chromosomal abnormalities: опыт работы с 1500 последовательными случаями. Am J Med Genet C Semin Med Genet. 2006;142C:251–262.
[31] Schaeffer GB, Sampanian J, Sung CC, et al. Microarray-based comparative genomic hybridization of genomic imbalances in products of conception. Prenat Diagn. 2004;24:1003–1012.
Figure 1[
