Aiming at Cancer - Bioscience Technology Online

Bioscience Technology

100 Enterprise Drive
Rockaway, NJ, 07866

Aiming at Cancer

By Tanuja Koppal

The rate of cancer incidence continues to remain stable while the death rate from cancer continues to decline, states the 2005 Cancer Trends Progress Report Update, published by the National Cancer Institute (NCI). This progress can be largely attributed to advances in areas such as genomics, proteomics and informatics that have made possible early detection, diagnosis and treatment of disease. On the horizon are technologies that promise to match patients with the drug they are most likely to respond to, paving the path toward personalized medicine.

Cancer as a genetic disease
A significant effort is underway to understand the genetic basis of cancer and its effect on downstream cellular pathways. In December 2005 the NCI and the National Human Genome Research Institute launched a three-year pilot project called The Cancer Genome Atlas to “test the feasibility of using large-scale genome analysis technologies to determine all of the important genomic changes involved in cancer.” This pilot project aims to complete the genomic analysis of three cancer types — lung, brain (glioblastoma) and ovarian cancer — to identify specific alterations in genes and differentiate the various cancer sub-types.

Longitudinal BLI analysis of EL1-Luc/EL1-TAg (F116) mice showing pancreatic tumor progression. The results of this experiment indicate that tumor onset is rapid and stochastic and that the rate of growth is variable within a cohort of this transgenic line. (Image courtesy Caliper Life Sciences, Inc.)

“We are moving oncology to an era where we are targeting the therapy to very specific changes in gene or protein,” says Rick Lesniewski PhD, director of cancer research at Abbott Laboratories in Chicago, IL. Studying the genetic basis of different cancers, looking at the genetic profiles of tumors after treatment with various drugs, and correlating the tumor response to changes seen at the DNA level is what the Abbott group is engaged in. “We believe that cancer is a genetic disease and the tools that we have been emphasizing are those that help us understand cancer at a genetic level.”

The group finds promise in using Comparative Genomic Hybridization (CGH), a microarray-based technology that measures genome wide changes in DNA copy number. “It’s a technology that is applicable at multiple levels of our research,” says Lesniewski. “Very early on when we are looking at cells in culture, in more advanced stages of drug development when we are looking at animals, and finally, we are now beginning to assess whether it can have applications in human tumors from patients.” The hope is that individual genetic profiles obtained using CGH will lead to the discovery of biomarkers that can help match a patient with a treatment to which they are most likely to respond.

There are numerous advantages to using CGH, says Dimitri Semizarov, PhD research investigator and tumor genomics group leader at Abbott. These include wide genomic coverage, high resolution (~25 Kb) and quantitative data. However, CGH is very labor intensive, involves complex data analysis and currently cannot be used on archived samples. “For routine patient monitoring we think an assay like florescence in situ hybridization (FISH) will be used, but CGH will provide the fundamental information that leads us to specific probes,” says Lesniewski.

Proteomics ramping up

Reduction in ERK phosphorylation in human PBMCs. Read-out is quantitative and reproducible.Click to enlarge

With an eye towards developing targeted therapies, increasing attention is being paid to protein families like proteases, caspases and kinases that play a critical role in cancer. Proteomics is also playing a major role in biomarker discovery. “Working with kinase targets allows you to design logical biomarker strategies to understand how a drug candidate works in cancer and which endpoints you can measure to develop an assay that can be used in the clinic,” says Angela Romanelli, PhD, translational oncology research group leader at the Serono Research Institute in Boston, MA. Serono currently has a number of small molecule-based kinase inhibitors in their oncology pipeline.

Romanelli’s group mostly uses florescence activated cell sorting (FACS) based analysis to measure phosphorylation states of target proteins at specific time points. “FACS-based approaches sound pretty mundane and routine, but they have become quite sophisticated, allowing you to measure many more parameters simultaneously and in parallel,” says Steve Arkinstall, PhD vice president, U.S. Research at Serono. “It’s a beautiful technology for developing biomarkers in oncology since you can measure multiple phosphorylation states on multiple proteins in different pathways at the same time.”

Flow cytometric assay for detection of target modulation in human PBMCs (Peripheral blood mononuclear cells): Activated ERK is assessed in fixed PBMCs using an antibody selectively recognizing phosphorylated/activated ERK. Antibody is directly to conjugated to a fluorophore to allow rapid detection flow cytometry assay. (Graphs courtesy Serono Inc.) Click to enlarge.

Miraculins Inc., a company based in Manitoba, Canada, utilizes a combination of mass spectrometric and protein chemistry techniques to purify and identify clinically relevant biomarkers for cancer. Their approach is top-down proteomics, in which a large number of patient samples, such as blood serum, plasma and urine, from diseased and healthy individuals, are analyzed. They use Ciphergen’s ProteinChip Series 4000 instrument, which is based on surface enhanced laser desorption ionization (SELDI) technology, for routine analysis.

“With the most highly expressed proteins we are able to get low nanogram or high picogram levels that are able to give significant protein peaks,” says Douglas Barker, PhD, research & development manager at Miraculins. The newer ProteinChip instrument also has more automated features. “With the old instrument you had to load one chip at a time. Now we can lay down 140 to150 chips and let the machine run overnight.”

The increasing use of proteomic technologies has led the NCI to award $35.5 million over five years to five multi-disciplinary academic teams to establish a collaborative network for “Clinical Proteomic Technology Assessment for Cancer.” The teams are funded to evaluate the full range of commercially available proteomic technology platforms and software relevant to help standardize experimental protocols and methodologies, making it easy for researchers to compare data and analyze results by reducing intra-platform and inter-laboratory variability.

How predictable are animal models?
Animal models have been used extensively in cancer research and the most commonly used model has been the human tumor cell xenograft in immunodeficient mice. “That model has served us well over the years in terms of identifying cytotoxic agents,” says Robert Abraham, Ph.D., vice president of oncology discovery research at Wyeth Pharmaceuticals in Pearl River, NY. “But recently we have become increasingly interested in targeting the tumor micro-environment and, unfortunately, this model does not recapitulate the tumor microenvironment anywhere near to the extent that we would like.” Transgenic mice are proving to be a good alternative, since tumors in these models form in a normal host tissue. However, the tumors tend to be very homogeneous when compared to human cancers.

CGH can be used to identify gene amplification/deletions that correlate with the sensitivity to a drug. In this example, an amplification of a long region on the long arm of chromosome 18 is present only in sensitive cell lines. (Graph courtesy Abbott Laboratories.)Click to enlarge

"What is important in an anti-cancer animal model is that you are inhibiting the same pathway that is activated in a human tumor cell,” says Arkinstall. Along those lines Abraham feels that using miniature pig models may not seem outrageous in the future. “The pig genome closely resembles the human genome, and during development as well, a lot of the organ systems (in the pig) recapitulate human systems.”

Improvisations in in vivo imaging are further enhancing the use of animal models by enabling scientists to visualize biological processes in real-time, in a living animal. The Xenogen IVIS Spectrum (Caliper Life Sciences Inc., Hopkinton, MA) allows mouse embryos to be pre-programmed or injected with molecularly modified cells that emit light when a gene of interest is expressed. “Its this (the light producing transgenic animals) that the FDA is getting very excited about because they feel that to the extent that they can start to make animals with the disease — as opposed to injecting cell lines with disease — you have a much better chance of emulating what’s going on in the body,” says Kevin Hrusovsky, president and CEO of Caliper. He says that the technology is extremely quantitative and its use can be extended beyond rodents and primates to humans in the clinic.

Equally noteworthy are advancements being made in in vitro cell culture systems using sophisticated technology that is easily implemented, highly predictive and physiologically relevant. Most current data is generated using cells grown as monolayers in Petri dishes, where they don’t interact freely with one another, or using cells grown in fetal calf serum or under 20% oxygen conditions, which are physiologically irrelevant. “There is considerable interest in the derivation of new systems in which the cells can interact freely in an artificial cell matrix in vitro under more realistic cell conditions,” says Abraham. “That can make a huge difference in the biology that you read out and I will place my money on the fact that it will make a huge difference in drug responses as well.” Companies like BD Biosciences, Cambrex and others are now providing specialized extracellular matrices, cell culture media and freshly isolated, well-characterized human cell lines that make in vitro systems more physiologically relevant.

Garbage in, garbage out
Despite all the advancements in science and technology a lot of oncology research remains very qualitative. Robotics and automation can automate sample processing and data analysis but when it comes to actual collection of samples and data entry there is still a lot of personnel involvement.

“What we have found is paying attention to details and doing good statistics, having good sample processing methodologies are absolutely vital to getting good quality data,” says Barker. “People are realizing that what they need is more high quality samples and controls, better experimental design, and inclusion/exclusion criteria for their patients. Or else for any instrument out there it’s going to be garbage in, garbage out.”