August 1, 2011 (Vol. 31, No. 14)
Angelo DePalma Ph.D. Writer GEN
Technique that Augments Karotyping Now Being Used in Prenatal Testing, Cancer, and Autism
Array-based comparative genomic hybridization (aCGH) has become a standard method for detecting copy-number variations (CNVs) and other genetic defects associated with genetic diseases. Once primarily the domain of postnatal testing for developmental disorders, the technique is branching into prenatal (and preconception) testing and cancer.
aCGH involves labeling test and reference genomes with differently colored dyes, mixing them, then hybridizing the samples to a microarray of probes covering a genomic area of interest. The scope may range from a few hundred genes to the entire genome.
aCGH differs from standard CGH in its resolution. Because conventional CGH hybridizes samples directly to chromosomes, resolution is limited to five megabases or so. Arrays employ not whole chromosomes but probes consisting of 50 to 75 bases. The densest commercially available arrays today consist of between two and three million probes. More common are 180,000-spot slides, which provide whole-genome analysis at a resolution of 5–10 kilobases.
“The advantage of microarrays is that they allow viewing of the entire genome as well as regions of special interest,” said James Clough, vp for clinical and genomic solutions at Oxford Gene Technology (OGT).
Founded by Edwin Southern, Ph.D., the Oxford professor of Southern blot fame who also invented microarrays, OGT began to provide microarray services about seven years ago. “It soon became clear that array CGH had a role to play in the investigation of constitutional disorders,” added Clough.
At the time karyotyping was king, and oligonucleotide arrays were thought to be too noisy. But eventually, researchers found clinical relevance beneath signals originally ascribed to noise. But what eventually won over researchers was the clinical yield: just 5% to 8% for karyotyping, up to 17% on BAC arrays (an early embodiment of oligo arrays), and up to 22% for aCGH.
The most pronounced limitation of aCGH is the inability of the native technique to detect balanced translocations in which precisely the same quantity of genetic material switches from one chromosome to another, and vice versa. Probes can only detect the relative amount of material corresponding to a genetic region, but not its actual location. Balanced translocations are associated with many solid and hematopoietic cancers.
Another limitation is that aCGH cannot distinguish triploidy, tetraploidy, or any other abnormality where every chromosome in the cell is detected in multiple copies, generally termed polyploidy.
“These rearrangements can only be detected using conventional chromosome testing techniques like karyotyping,” said Gary Harton, head of molecular research at Reprogenetics. “However, aCGH can be used to detect the inheritance of unbalanced translocations in embryos from couples that carry these chromosome rearrangements.”
At the “Association of Genetic Technologies” annual meeting, PerkinElmer described a work-around for balanced translocations, thereby opening up new areas of investigation for aCGH.
PerkinElmer’s approach uses gene amplification of known translocations followed by aCGH analysis. It involves polymerase and primers for the affected chromosomes.
For example, the Philadelphia translocation in chronic myeloid leukemia involves chromosomes 9 and 22. As the primer for chromosome 9 and polymerase amplify the gene, they encounter the sequence from chromosome 22 and amplify that as well. Since only the primer for chromosome 9 is present, one would not expect an amplified signal for chromosome 22. Separately, the primer for the breakpoint on chromosome 22 amplifies the genetic material from chromosome 9.
“Now, with that amplified pool in hand, when we look at the array we see chromosome 22 has a signal where you wouldn’t expect one,” said Christopher Williams, global market segment leader for oncology.
Further analysis would reveal where the DNA broke apart, which holds significance for both prognosis and treatment.
PerkinElmer is offering a service based on this approach and a new product, OncoChip, which targets more than 1,800 loci associated with hematologic cancers. Onco- Chip provides baseline coverage of one oligo probe every 35 kb, and one probe every 500–5,000 bases in targeted regions.
Pre- and Postnatal Testing
Many firms use aCGH for postnatal diagnostics. Almost all, like CombiMatrix, also employ FISH testing and full chromosome analysis as needed.
In 2010 the American College of Medical Genetics and American Pediatrics Association designated aCGH as a first-tier method for postnatal testing. aCGH provides 20% greater detection of abnormalities compared with standard karyotyping, according to Dan Forsche, senior vp of marketing.
aCGH is making inroads quickly into prenatal testing as well. A 4,000-subject study on this application is under way. Results will be reported early next year at the meeting of the Society of Material and Fetal Medicine. By then, Forsche predicts, aCGH will have earned first-tier status for prenatal testing as well.
Cancer screening via aCGH is still considered experimental, or basic research, but the possibilities are as exciting as with pre- and postnatal testing. CombiMatrix employs a 180K array for hematologic cancers, which according to Forsche “picks up things that FISH cannot detect.”
The company also uses a tumor array containing about 100 oncogene probes, which may be used on both fresh and preserved tissue. Eventually, researchers hope to provide prognosis and advice on potential treatments based on the results.
CombiMatrix is also looking into sequencing, which will become more significant as the price falls from the current $8,000 to $5,000. A number of companies and research groups engage in targeted sequencing, which examines regions of the genome identified as trouble spots by aCGH.
An article published in Neuron described an aCGH technique for detecting the CNVs that underlie autism spectrum disorder (ASD). The study, undertaken on 1,000 families with one autistic child and one unaffected sibling, was conducted by Michael Wigler, Ph.D., of Cold Spring Harbor in collaboration with a group from Columbia University.
Since molecular biology began to dominate the study of autism, experts believed that inheritable genetic mutations were the principal cause of ASD. In unveiling a unified theory of this disorder, Dr. Wigler demonstrated that these mutations accounted for only about 25% of ASD cases.
The remainder arise from de novo mutations that did not appear in either parent and must have arisen spontaneously. Dr. Wigler identified the minimum number of involved CNVs at between 250 and 300.
Commenting on his work, Dr. Wigler noted, “The causes of autism when fully fleshed out are likely to be very diverse, some of which may be treatable much more readily than others. However, the diversity of causes implies that an effective future treatment for one form of ASD may be specific only for a narrow subset of those affected.”
Applying sophisticated mathematical analysis tools to aCGH, researchers resolved genomic irregularities—differences between affected and unaffected siblings—at much greater resolution than previously.
De novo CNVs were found in about 8% of children with ASD, compared with just 2% of unaffected siblings. The study design was biased toward discovering de novo CNVs, as ASD is very likely to be inherited in families with multiple affected children.
If Dr. Wigler’s theory holds, more than half of all ASD cases arise from rare, de novo CNVs. Most of these CNVs, in fact, were observed only once. An accompanying paper identified the locations of these CNVs as regions of the genome previously implicated in studies of autism and cognitive disability.
Interestingly, girls were more likely to have a higher frequency of mutations than boys (11.7% vs. 7.4%), with more mutations involved (15.5 vs. 2.0), which would imply a higher incidence of ASD in females than males. Dr. Wigler hypothesized that females are somehow resistant to this phenotype, and that much larger genetic defects are required to confer the disease on girls than boys.
aCGH is showing promise in preconception testing as well. In 2010 physicians employed the technique to screen human eggs for genetic defects that increase the risk of miscarriage. As a result, two healthy babies were delivered in Germany, and one in Italy, of mothers who had experienced difficulty conceiving artificially and had a history of miscarriage. Several other post-aCGH pregnancies are currently under study, and a large-scale clinical trial is scheduled for 2012.
One might ask how an analytical technique can scan the entire genome and still provide a viable egg. Before an egg is fertilized it releases half of its 46 chromosomes to make room for the 23 that the sperm will provide. These chromosomes, held within a structure known as a “polar body,” are exact duplicates of those remaining in the egg. aCGH analyzes these castoff genes.
The technique has several advantages over conventional prenatal testing. It does not involve the egg itself, either before or after fertilization, nor does it touch the embryo, thereby nearly eliminating ethical or religious objections. And since it selects an egg deemed viable, the temptation to implant multiple embryos, as is commonly done today during in vitro fertilization, is lessened.
Not Only for Somatic Cells
Ambry Genetics has extensive experience with aCGH, other array types, and next-generation sequencing. The company has performed commercial aCGH for several years, mostly for detecting deletions and amplifications implicated in common genetic diseases. About two years ago Ambry began running research tests for academic labs and pharmaceutical companies, mostly on cancer tissue samples.
Around this time the company realized that aCGH analysis of stem cell lines could provide insights into how a cell line behaved, particularly when used therapeutically. “No arrays existed at the time that targeted regions in the genome implicated in stem cell differentiation and proliferation,” said Aaron Elliot, Ph.D., senior scientist at Ambry.
Working with Roche NimbleGen, Dr. Elliot designed an aCGH microarray that covered the genomic backbone but also targeted specific regions of interest for stem cells at a resolution of 15 kilobases. “We can detect the deletion or duplication of a single exon,” Dr. Elliot explained.
This level of sensitivity is critical, for example, in assuring that implanted stem cells are not lacking an obvious tumor suppressor gene or overexpressing a tumor promoter. “You would never be able to find defects that small through karyotyping.” The lower detection limits of karyotyping, said Dr. Elliot, is about five megabases.
Ambry recently reached a collaborative agreement with Cell Line Genetics (CLG) that combines Ambry’s experience in microarrays with CLG’s knowledge of preclinical cell-line characterization. The goal is the introduction of additional microarray-based products for stem cell analysis.
When working with stem cell lines it is not unusual for an embryonic or induced pluripotent stem cell to begin with a normal karyotype, then develop abnormalities over time. “Cytogenetics can pick up emerging cell populations much earlier than array CGH, albeit at much lower resolution,” says Julie Johnson, director of laboratory operations.
Johnson notes that some authors mistakenly claim that aCGH replaces karyotyping. “Actually, the techniques go hand-in-hand. Cytogenetics is better for detecting low-level, emerging cell populations and balance translocations, while array CGH offers higherresolution detection of chromosomal abnormalities and greater sensitivity.”
Developing a specialized array is an iterative process that takes about six months. Companies like Ambry typically draw up a list of sequences they believe are relevant, and submit them to an array fabricator such as Roche NimbleGen, Agilent Technologies, or Oxford Gene Technology.
The deeper the coverage of genes of interest, the higher the resolution. Probes may be deleted in subsequent versions of the array if testing shows they do not perform as expected.