CHROMOSOMAL STRUCTURAL ALTERATIONS AND GENOMIC IMBALANCES: METHODOLOGICAL STRATEGIES FOR DETECTION

Chromosomal alterations in cancer have been detected by
classical cytogenetics methods, mainly banding karyotyping
(G-, R-, or Q-banding). Solid tumors frequently exhibit numerous
changes in chromosome numbers, including gain of
whole-genome complement and gains and losses of specific
chromosomes. Tumors also have structural intrachromosomal
and interchromosomal rearrangements, which change the copy
numbers of genes when deletions, duplications, or amplifications
occur, and affect the transcription of genes when positioning
changes are introduced by insertions, inversions, and
translocations. Although conventional cytogenetic methods
were fundamental for important discoveries on molecular
mechanisms of hematological diseases, they failed to provide
similar contribution on solid tumors. Typically, rearrangements
in solid tumors are numerous and complex, and the resolution
of 5 to 10 megabase (Mb) of the banding karyotype is not
satisfactory for identifying the spectrum of genomic changes
responsible for most of the specific biological characteristics of
the cancer cell.
The development of molecular cytogenetic strategies such
as multiplex fluorescence in situ hybridization (M-FISH) 2 and
spectral karyotyping (SKY) 3 have facilitated the identification
of extranumerary chromosomes and increased the accuracy of
identification of chromosomal origins of complexly rearranged
chromosomal (Fig. 6.1). M-FISH and SKY, which paint genomic
material from each of the 24 human chromosomes in
specific fluorescent colors, are technologies especially tailored
to uncover interchromosomal rearrangements and have been
successful in revealing subtle karyotypic alterations, which
would be otherwise overlooked.
For detection of genomic imbalances, the cornerstone technology
was the comparative genomic hybridization (CGH),
which was introduced in the early 1990s. 4 As proposed initially,
CGH involved hybridization of differentially labeled
DNA from two genomes, the genome to be tested and a normal
genome used as a reference against a normal metaphase
template. This approach is called metaphase CGH (mCGH)
or chromosomal-based CGH (cCGH). Measuring the fluorescent
intensity that dominates each given chromosome region
in the template allows the identification of regions in the tested
genome carrying normal copy numbers or gains and losses relative
to the normal reference. Although mCGH proved to be
useful for detection of genomic imbalances in solid tumors,
the analysis is performed in metaphases and the high level of
chromatin condensation at that cell cycle stage also limit the
resolution of genomic changes to 5 to 10 Mb. Importantly,
several studies have shown that the expression of genes located
in chromosomal regions of gains or losses varies consistently
with the DNA copy number. 5–10 The availability of genomic
resources and technological advances has fostered a major improvement
in the last decade, represented by the shift from
cCGH to microarray-based platforms. In their first generation,
called matrix-CGH and array CGH, 11,12 these arrays included
only few hundreds of DNA clones. Soon after, two microarray
platforms were launched using bacterial artificial chromosome
(BAC) clones as DNA probes. These instruments, the tiling
BAC array 13 and the 1-Mb resolution whole-genome BAC
array 14 were able to mine the entire genome for copy number
variants. Despite the fact that the BAC arrays cannot reliably
detect aberrations smaller than the BAC insert (100 to 200
kilobase [kb]), these new tools granted a significantly higher
detection rate for copy number abnormalities than any of the
metaphase-based cytogenetic techniques.
More recently, oligonucleotide-based arrays have emerged
as the platform of choice for genome-wide analysis of copy
number alterations because of their high-throughput and highresolution
characteristics. Some of the commercially available
oligonucleotide arrays have probes specifically developed for
detection of copy number variation (NimbleGen Systems and
Agilent Technologies), whereas others were developed as genotyping
arrays designed to identify single-nucleotide polymorphisms
(SNPs) and later modified to uncover copy number
variations (Affymetrix and Illumina). The NimbleGen CGH
microarrays contain 45- to 85-mer oligonucleotide probes that
are directly synthesized on a silica surface using light- directed
photochemistry. A whole-genome tiling array is available with
2.1 million probes (HG18 WG Tiling 2.1M CGH v2.0D)
and custom tiling arrays are also available. The Agilent Human
Genome CGH Microarray (G2519A) contains 60-mer oligonucleotide
probes printed onto glass slides through an industrial
inkjet printing process. This microarray includes 40,000 probes
spanning the human genome with an average spatial resolution
of 75 kb, including coding and noncoding sequences, and
has an emphasis on the most common cancer-related genes.
Agilent arrays may also be customized from more than 8 million
predesigned and validated probes. With the NimbleGen
Systems and Agilent Technologies platforms, the test and reference
genomes are labeled with different fluorophores (usually
Cy3 and Cy5), and cohybridized to the same array, similarly
to the mCGH technology. The signal intensity ratio of the test
sample versus the reference sample is calculated for each probe
across the entire genome.
The Affymetrix GeneChip arrays contain 25-mer SNPbased
oligonucleotide markers or probes directly synthesized
on the array surface. The Genome-Wide Human SNP Array
6.0 features 1.8 million genetic markers designed to uniformly
cover the entire genome, including approximately half in SNPs
and half in non-SNP probes for the detection of copy number
variation. For the evaluation of copy number changes, the test
genome is labeled and hybridized to the array and the signal
intensity from the probes is computationally compared with a
control set (HapMap individuals). The Illumina BeadChip arrays
are made from silica beads that are self-assembled on silica
slide microwells and each bead is covered with specific 50-mer
SNP-based, oligonucleotide probes. The HumanCNV370-
quad DNA Analysis Beadchip platform covers approximately
380,000 SNP and non–SNP-based probes. The test specimen
is hybridized with the array and the copy number variations
are determined by computationally comparing the signal intensity
from probes with a control set provided by the platform
manufacturer.
These high-resolution platforms have been successfully
used to identify copy number changes in lung cancer and
other solid tumors. However, two major characteristics of the
solid tumors, the largely abnormal number of chromosomes
and the intratumor heterogeneity, make copy number analyses
difficult in these platforms. Current array CGH platforms
were designed under the assumption that the natural ploidy
state of the test DNA specimen is diploid, which is rarely the
case in solid tumors. Therefore, the detection of a single-copy
gain may represent a gain if the specimen is diploid or actually
represent a loss if the specimen is tetraploid, a condition that is
most common in solid tumors. Additionally, tumors are often
mixtures of distinct types of cells, each of them potentially carrying
different copy number changes. Because the DNA from
the test specimen is extracted from the cell mixture, the results
reflect an average change across the different cell types.
Changes occurring in cell-specific compartments are likely to
be diluted and remain undetected.
For the genome-wide high-resolution arrays, another perceived
limitation is the detection of copy number variations
that may not be involved in the disease. Recent studies have
shown that normal, healthy individuals carry a large number
of copy number variations detected by more than one consecutive
probe in BAC and oligonucleotide arrays. 15,16 Thus,
a more detailed characterization of the variation in the normal
genome is necessary before an accurate detection of pathogenic
copy number aberrations can be reached in tumors.
Chromosomal abnormalities detected by conventional
and molecular technologies and genomic imbalances detected
by mCGH or array platforms have been validated by independent
laboratory approaches, such as fluorescence in situ
hybridization (FISH) or polymerase chain reaction (PCR)–
based techniques. FISH is a high-resolution technique able to
identify specific regions involved in rearrangements and define
them accurately (Fig. 6.2). FISH, as opposed to the PCR-based
techniques, has among its critical advantages the ability to investigate
the target phenomenon in single cells and to preserve
the original tissue architecture. However, FISH is not a highthroughput
technology and is unable to answer genome-wide
questions. Nevertheless, the development of FISH methods
has significantly improved the accuracy of solid tumor cytogenetics.
Ultimately, it is the combination of multiple technical
approaches that provides the most powerful strategy for understanding
the molecular pathways underlying the lung tumor
development.