By Thomas Baer
February 18, 2004
| The age of personalized medicine based on advances in genetic profiling of individual patients and patient groups is rapidly dawning. The full implications of this trend for drug research and development are only beginning to be understood, but clearly one challenge for pharmaceutical and biotechnology companies will be to ensure that suitable diagnostic tests are available for new drugs as they are developed.
Fortunately, the new science of molecular pathology is quickly coalescing to meet this challenge, most notably in gene-expression-based diagnostic tests for cancer patients. The other good news is that many of the tools being used are already familiar to drug researchers in the preclinical area. They include quantitative gene-expression-analysis platforms, specialized reagents to extract and amplify RNA, and tumor-cell-specific gene signatures.
For the first time, a combination of these technologies is enabling clinical labs to develop and validate gene-expression-based tests and to perform these tests within the setting of clinical development or post-trial therapy. Each technology has significantly contributed to the ability to develop diagnostic tests that separate patients based on the gene-expression characteristics of their individual tumors, paving the way for more customized, personalized therapies.
Someone diagnosed with cancer has immediate questions. How far has it spread? How can I get rid of the cancer as fast as possible? Am I going to die from this? Current medical practice and diagnostic tests don't provide adequate answers. Physicians and their patients must make critical therapeutic decisions with only limited information about the likely outcomes of these decisions.
A good example of the limitations of current clinical diagnostic testing is the reliance on measures such as estrogen receptor (ER) and lymph node status. Considered the gold standard in determining whether a woman should be prescribed anti-estrogen hormone therapy, such as tamoxifen, or a significantly more toxic regimen of chemotherapy, physicians must rely on ER and lymph node status as key inputs to their therapeutic decisions.
ER-positive, node-negative patients are expected to have good outcomes from anti-estrogen therapy. Yet approximately 30 percent of these women experience a recurrence of their cancer within 10 years. If diagnostic tests could differentiate the responding patients from the nonresponding patients, who are now classed together, physicians could more aggressively treat early-stage tumors in appropriate patients, potentially saving thousands of lives each year.
The clinical diagnostic community faces the challenge of classifying patients more distinctly for all cancer types. With approximately 80 percent of solid tumor cancers located in four major organs — breast, colon, lung, and prostate — concentrating on new approaches to clinical diagnostics that identify drug-responsive patients in these four areas will have a meaningful impact not only on therapeutic decisions and patient outcomes, but also on drug development itself.
Gene-expression-analysis tools, including microarrays, quantitative PCR (Q-PCR), and other newer technologies, enable the simultaneous identification of expressed and nonexpressed genes for as few as one gene, to as many as 44,000 gene transcripts found on a whole-human-genome microarray. These measurement tools provide the clinical pathologist an opportunity to develop detailed gene-expression information on cancerous and noncancerous tumors using a single platform.
Molecular pathologists and clinicians can use the gene profiles of tumors captured today as benchmark data in the future to understand the progression of cancer over time, as well as the efficacy of emerging drugs that treat it.
Gene-expression data are only as accurate as the source material, however. Normal tumor biopsies contain a variety of tissue types. Mixing cancer and noncancer cell types can distort gene-expression results. But the new technology of laser-capture microdissection, often referred to as LCM, has allowed researchers, and now clinical pathologists, to select individual cells or groups of cells in the most precise way possible.
The result is pure populations of selected cells for analysis, rather than heterogeneous cells from tissue scrapes or homogenized biopsy samples. In the precise identification of finely differentiated classes of cancer patients, a pure cell population from complex biopsy samples is essential as an analytical starting point.
Clinical laboratories must be properly outfitted, however, to deal with RNA-based diagnostic tests, since virtually all clinical biopsy samples are now fixed in formalin and embedded in paraffin (FFPE) immediately after excision. This simple and reliable method prevents degradation of the tissue and facilitates easy transport of the biopsy material to the lab.
Prior to methods and products developed recently, researchers could not unlock genetic information from formalin-preserved samples due to the molecular bonds that cross-link RNA and DNA in the process. The new techniques and reagents reverse the cross-linking process caused by the formalin and allow extraction, purification, and amplification of messenger RNA from the samples.
This approach can be applied to archived biopsy samples, as well as to new biopsies. Due to the sensitivities inherent in working with microarrays, researchers assumed that RNA from FFPE samples was so degraded by the preservation process that gene-expression analysis on chips was impossible. Custom microarrays developed to read gene expression in RNA from FFPE samples are now available, however — another key milestone in enabling clinical labs to develop RNA-based diagnostic tests without the need to change sample-collection protocols in hospitals and other biopsy-collection sites.
Next Frontier: Gene Expression
Tumor-cell-specific gene signatures are being discovered that correlate with disease progression and tumor response to drugs. Using laser-capture microdissection to secure pure cell populations, along with whole-genome microarrays for signature discovery, key genes are being identified that, when expressed or not expressed, can indicate the genetic nature of the cancer. New patient biopsy samples can be matched to groups of patients with known outcomes. These gene profiles will provide the answers to questions that were previously unanswerable.
Challenges remain for clinical diagnostic laboratories in developing tests based on these new technologies. Unlike the preclinical research market, the use of microarrays in clinical laboratories has not been widespread because microarrays couldn't handle RNA from FFPE samples until very recently. With the new products and techniques available today, clinical labs will need to install equipment and train their personnel to capture pure cell populations and to perform gene-expression analysis on microarrays for certain types of RNA-based tests.
In addition, clinical labs will be extracting, purifying, and amplifying RNA for the first time, a capability that has been mastered in the life science research setting but has not been a focus in clinical labs due to the previous technical inability to unlock FFPE samples.
Nevertheless, molecular diagnostic testing will be a fast-growing area in major clinical laboratories as new technology enables access to gene-specific information. Tests are already available for DNA markers and protein expression. Gene-expression analysis is the next frontier and the next major area of opportunity for clinical development. Multiple technologies are coming together to provide the tools to answer the tough questions. And millions of cancer patients worldwide will surely benefit.
Thomas Baer is chairman and CEO of Arcturus Bioscience, in Mountain View, Calif. He may be reached at email@example.com.
ILLUSTRATION BY MARK A. GABRENYA; IMAGE COURTESY ARCTURUS BIOSCIENCE