EVENTS | VIEW CALENDAR
The big picture
The drug discovery and development world, along with diagnostic research and development, has a tendency to produce alphabet soup, and oncology research is no different, with some of the notable acronyms being SNP, CNV, miRNA and FISH, along with full words like gene expression, sequencing and resequencing.
Some arrive and may dominate the spotlight for a while, but in the end, each plays important role, and all of them share one things in common: they increasingly revolve around genetic technologies and molecular approaches.
It makes sense, of course, as the Cancer Genome Project based at the Wellcome Trust Sanger Institute in the United Kingdom notes that "All cancers occur due to abnormalities in DNA sequence," and thus, "the identification of genes that are mutated and hence drive oncogenesis has been a central aim of cancer research since the advent of recombinant DNA technology."
The Cancer Genome Project's Genomics of Drug Sensitivity initiative, for example, is a five-year effort launched in late 2008 involving investigators at the Sanger Institute and other researchers, notably those at the Massachusetts General Hospital Cancer Center, and they plan to look at how some 1,000 genetically characterized cancer cell lines respond to treatment with 400 anti-cancer treatments, alone and in combination. Along with drug sensitivity information, the team is providing genetic data on the cancer cell lines tested, including information on mutations, copy number variants (CNVs) and gene expression patterns.
Findings from Genomics of Drug Sensitivity studies looking at the effects of 18 anti-cancer drugs on 350 genetically characterized cancer samples are already being made available to other researchers because of the promising insights offered by the various gene-oriented technologies. Many treatment-related genetic patterns are already coming to the fore, including activating mutations in the BRAF gene in melanoma that correspond to BRAF-targeting treatment response.
"It is very encouraging that we are able to clearly identify drug–gene interactions that are known to have clinical impact at an early stage in the study, " says Ultan McDermott, co-project leader and a medical oncologist and human genetics researcher at the Sanger Institute. "It suggests that we will discover many novel interactions even before we have the full complement of cancer cell lines and drugs screened."
But there is no hard-and-fast rule about which technology is "best" because each has its own strengths and weaknesses, and the research culture of any given pharma or biotech might work better with some approaches than others, even aside from issues like cost or time.
For example, FISH (fluorescence in situ hybridization) was shown in a February 2004 journal article in Urology to be 92 percent effective in detecting bladder cancer markers when testing urine specimens, compared with a sensitivity of only 64 percent for traditional cytology screenings.
But now, six years later, is FISH as relevant?
"FISH, it seems to me, is one of the technologies that will be replaced by newer ones with higher resolution; for example, arrays," notes Dr. Ulrich Schwoerer, head of global marketing for 454 Life Sciences Corp., the sequencing subsidiary of Roche.
But much also depends on the area of focus and not simply which technologies might be fading in prominence.
"Broadly speaking, if you are looking at things like disease proclivity or disease risk and changes that might be inherited, those things are more concerned with the single nucleotide polymorphisms [SNPs] and mutations—things that are passed from generation to generation," notes Dr. Nandan Padukone, president and CEO of Nuvera Biosciences, a company focused on developing novel molecular diagnostics that make a significant impact on cancer care. "But if you're looking at more complex issues, like responsiveness to therapy or lifestyles that lead to cancer, you need to look at things like gene expression, methylation or proteomics, for example."
Smaller companies, he points out, will often have to focus on just one platform because of staff and budget limitations, although as they grow or need better characterization of the oncology data, they will invest in other platforms.
"Then they have to think, do we go after SNPs or CNV or gene expression or interactions," Padukone says. "But size is a key, as is the market demand for what you have, and that drives how many different platforms you need. If you're looking at a couple biomarkers, your work might go fine just on one platform, versus the more extreme side at a larger company where you need two or more platforms, multiplexing and are dealing with 50 biomarkers at once. The endpoint and the marketplace determine a lot of this for you, and your endpoint as well—pharmas may go one direction on approaches while a diagnostic company in a similar oncology area may need to go another."
Also at issue is how comprehensive a given approach or technology is. Short-read technologies are are very effective in identifying SNPs, Schwoerer notes, but you have to have long reads in order to detect any of the larger mutations, like insertions, deletions, inversions and translocations.
"With cancer, the whole genome is messed up—parts are duplicated, other parts are relocated, others again deleted," Schwoerer says. "Also, the three-dimensional structure often looks differently. It is very important to get the full picture in order to understand this disease better."
"Also, you have to de novo sequence genomes, especially when it comes to disease, but also in general for human genome sequencing," he adds. "Multiple recent publications have shown that you miss parts of the genome if you only resequence. In order to effectively de novo sequence, you have to have long reads; short reads are not capable to de novo sequence genomes."
Most of the short-read technologies currently do between 75 and 100 base pairs, Schwoerer says, and although some of the companies competing with 454 claim longer reads than that, he maintains that the error rate can often become a real problem.
454 Sequencing, he says, is already at a mode of 500 base pairs, with Q20 quality at the 400th base. His company is about to launch a platform that can read up to 800 base pairs, and IBM is working with Roche and 454 to develop a DND Transistor technology that promises reads in the several thousands of base pairs.
"Ultimately, sequencing is the technology with the highest resolution and is geared to become the gold standard for many genomic research areas," Schwoerer says. "The blessing of bioinformatics as well as the challenge is to deal with all this genomic data; with longer reads, the assembly becomes easier and bioinformaticians and researchers can focus more on the biological questions. There are plenty of challenges left when you get the whole picture, and we believe that this is where the really interesting discoveries are."