May 15, 2009 (Vol. 29, No. 10)

Diseases that Have Defied Description Can Now Be Described and Explained

Array comparative genomic hybridization (CGH) is illuminating DNA copy number variations (CNV) throughout the genome to diagnose disease etiology and to help tailor treatment regimens for patients to improve health outcomes.

In January 2007, the College of American Pathologists and American Society of Clinical Oncology (CAP/ASCO) published guideline recommendations for testing human epidermal growth factor receptor 2 (HER2) in the diagnostic workup  for all invasive breast cancers. About 15 to 20% of breast cancers are HER2-positive. As a predictive marker, HER2-positive tumors are associated with higher rates of breast cancer recurrence and mortality. Furthermore, HER2 status can help predict the success of using certain systemic or endocrine therapies.

The CAP/ASCO recommendations were made following a study of previously published HER2 test data. The study found that patients with early-stage breast cancer and metastatic cancer benefited from trastuzumab (Herceptin) HER2-targeted therapy, but that interlaboratory variability in HER2 test results ran as high as 20%. Thus, some patients never received Herceptin, though they would have benefited from it, and others who got it would not have benefited and have been put at higher risk for cardiac problems.

In normal (nondividing) cells one HER2 gene is on each of two copies of chromosome 17 and instructs a cell to manufacture HER2 protein. HER2 protein receptors on the cell’s surface program cell growth and division. In HER2-positive breast cancer, the cell has more copies of the HER2 gene and is amplified. Greater numbers of HER2 receptors on the cell’s surface send more messages for it to grow and divide more aggressively.

The two most widely used methods in the clinic for HER2 testing are immunohistochemistry (IHC) and fluorescent in situ hybridization (FISH). Test results from both IHC and FISH have an impact on the treatment regimen for women with breast cancer.

CNV distribution across the human genome, taken from the database of genomics variation.
(Centre for Applied Genomics, Hospitals for Sick Children, University of Toronto)


IHC, a protein assay, assesses the amount of HER2 protein receptors on a cancer cell’s surface. In HER2-positive tumor cells, HER2 protein production is greater compared with normal cells.

A sample of breast cancer tissue is exposed to an antibody against HER2 protein receptors. The antibody reacts causing a color change in the sample. Degree of color change is evaluated; the darker the stain, the greater the amount of HER2 protein. Percent of stained cells and their color intensity determine the test score. The scoring scale for an IHC test ranges from 0 to 3+. Scores of 0 and 1+ are negative for HER2; a score of 2+ is borderline or equivocal; a score of 3+ is positive for HER2.

Potential problems with IHC testing include damaging the sample during preparation, or subjectively judging the degree of color staining, which can adversely affect results.

Fluorescent In Situ Hybridization

FISH, also a visualization method, is used for locating and mapping chromosomes, genes, and gene segments. It can be performed on dividing, nondividing, and formalin-fixed cells, and uses single-stranded DNA probe labeled with a fluorescent dye. The probe binds to its complementary DNA site, and its fluorescent tag depicts its location. Different FISH probes can be used for different types of gene interrogation such as finding genes, determining chromosome number, and examining chromosomal abnormalities.

One FISH procedure measures the ratio of genes to chromosome 17 in a 60-cell sample. In a normal cell, the HER2 gene per chromosome 17 ratio is 1:1. HER2 gene per chromosome 17 ratios ranging from 1.8 to 2.2 are borderline. Tumor samples with an average HER2 gene per chromosome 17 ratio greater than or equal to 2:1 are reported as HER2 positive. Another FISH procedure measures the average number copy number of HER2 genes per cell. If the average HER2 gene copies per cell are greater than four but fewer than six, results are borderline. Like the IHC test, FISH scoring is subjective.

Shelly Gunn, M.D., Ph.D., medical director, Combimatrix Molecular Diagnostics, states that, “the CAP/ASCO directives are meant to deal mainly with inconsistencies of HER2 testing due to preanalytical variables such as formalin-fixation times.” However, she does not think that the directives solve the inherent subjectivity of IHC and FISH. “Both tests rely totally on visualization of results, either by protein immunostaining or by fluorescent DNA probes, and have unreliable, inter-observer reproducibility.”

To address the CAP/ASCO directives, array CGH is gaining use in determining HER2 copy number and in helping direct the course of breast cancer treatment.

CGH was developed in 1992 to analyze cytogenetic changes in solid tumors. Though not initially designed as an array, it was an alternative to tumor karyotyping, which can be hampered by insufficient numbers of high-quality metaphase cells to study. CGH can use fixed tissue or fresh samples to map DNA CNVs as a function of chromosome location in the genome.

CNVs are DNA segment gains and losses ranging from 1 kb to several Mb observed by comparing two or more reference genomes. These quantitative structural variants can include one or more genes. In some cases, they can affect gene function profoundly; in other cases, they display no overt phenotype.

To design an array CGH for high- throughput testing, oligonucleotide microarrays are designed to span the entire genome of a species based on the reference sequence for that species. For human genomic testing, arrays are made using bacterial artificial chromosomes (BACs) that contain referenced regions of the human genome.

Test genomic DNA and a control are labeled differentially with fluorescent tags and are hybridized competitively to that microarray. The array is scanned and fluorescent ratios indicate relative CNVs of the given oligo probe for the test subject versus the control. Each oligo probe is mapped to the reference genome. These fluorescent ratios are processed through an algorithm to get CNVs.

Global Genomic View

In addition to performing in high throughput, other advantages according to Dr. Gunn are that, “it is an objective and reproducible test that not only measures HER2 gene copy number, but gives a global genomic view of that patient’s tumor. This allows the clinician to design personalized treatment programs to give the patient an individualized prognosis based on the total profile of genomic aberrations present in the tumor.

“For example, many of the HER2-positive breast cancer genomes analyzed show co-amplification of the MYC gene on 8q24, which confers a less favorable prognosis and predicts decreased response to Herceptin.”

Array CGH can also complement IHC and FISH for subtyping newly diagnosed breast cancer. “IHC and FISH can be used in the original diagnostic workup of the patient to get a rough idea of key prognostic factors such as assessment of ER/PR (estrogen receptor per progesterone receptor), and HER2 status, which are critical for predicting response to systemic therapies,” Dr. Gunn explains.

“Using array CGH during the post-operative workup is best because fresh tissue is available, and the array can be used to confirm HER2 status and further subtype the tumor.”

There are some disadvantages to array CGH. It works best on fresh tissue, which may be difficult to get, and the technology is not yet widely available. As Dr. Gunn states, “it costs a little more than FISH analysis for HER2 gene status, but delivers a global view of the genome, which would be impossible to obtain using individual FISH probes.”

Though array CGH is seen as a rapid and cost-effective way of examining a whole genome of a subject for CNVs, Paul Kassner, Ph.D., principal scientist, lead discovery at Amgen, states that, “it is useful for finding the range of CNVs in the normal population that do not overtly cause disease, but that can influence phenotypes and impact disease. For example, extra copies of gene CC3L1 can reduce a person’s susceptibility to HIV infection and progression to AIDS.”

Some disadvantages that Dr. Kassner finds in working with mouse models are that arrays designed based on the reference genome for a given species can often be incomplete, thus, making array CGH results incomplete. “Array CGH is also unable to elucidate structure differences in the genome that do not result in copy number changes and is not able to tell us about translocations or inversions,” explains Dr. Kassner, who gave a presentation at the San Diego meeting.

Hypothesis-Free Data Gathering

“Chromosome microarray analysis is showing clinical utility in a proportion of individuals with autism,” states Stephen W. Scherer, Ph.D., director of the Centre for Applied Genomics, Hospital for Sick Children, University of Toronto. “In autism about 5 to 10% of idiopathic cases are found to carry de novo CNVs, and some genes involved have also been implicated using other experimental approaches. Tests are available for some genes and loci.” Though, he cautions, that when analyzing results, autism is a complex disorder so any data interpretation needs to be viewed in that context.

“Notwithstanding, for some of CNV regions being identified, the general lessons of medical genetics seem to apply and diagnoses are being made.”

Dr. Scherer, another speaker at the Cambridge Healthtech conference, says that enabling technologies like array CGH allow genome-wide hypothesis-free data gathering. “From these comprehensive genomic datasets one can focus in and generate informed hypotheses based on knowledge of complete data. It is just this type of algorithm of analysis that has led to such rapid progress in this field.”

New Genetic Paradigm

“The concept of copy number variation as the etiology of human disease is increasingly being driven by technology and revealing a completely new part of genetics,” states Charles Strom, M.D., Ph.D., medical director of the genetic testing center of Quest Diagnostics Nichols Institute.

Quest Diagnostics’ ClariSure™ array CGH postnatal test was approved for patient testing in New York in March 2009. It uses a single 10 mL blood sample to detect CNVs implicated in neuropsychiatric conditions including mental retardation, birth defects, autism spectrum, and developmental disorders, which conventional laboratory tests may not detect.

Mental retardation is defined as having an intelligence quotient of less than 70, and affects from 1 to 3% of the population. Initial investigations using karyotyping found an etiology in about 3.7% of patients tested. FISH with probes specific to subtelomeric regions can detect chromosomal aberrations in another 4% of individuals, though for other chromosomal regions additional probes are required. Array CGH can detect cytogenetic abnormalities throughout the genome in a single assay in 5 to 11% of individuals with mental retardation.

Dr. Strom considers the technological advances in genomics today analogous to the invention of the telescope and the impact it had on driving advances in astronomy. “We are still collecting and sorting out information. Then it will take the clinic and time to better understand the full etiologic implications of these neuropsychiatric conditions. With these tools, diseases that have defied description will be able to be described and explained.”

Next articlePfizer to Extend Use of Patient Selection Technology from Monogram