March 15, 2006 (Vol. 26, No. 6)
Harnessing Useful Diagnostic and Therapeutic Information
Biomarkers are increasingly recognized for their potential not only as biologic harbingers that provide early warnings of disease but also biochemical monitors of patient responses to therapies. However, the field is hampered by a lack of reliable, quantifiable, easily measured biomarkers that correlate well with biochemical measures of disease progression.
Companies are making strides to close the gap between bench and bedside. New advances in the field include identifying and quantifying biomarkers within tissue samples, designing image analysis tools that model human thought, and generating panels of biomarkers that provide a specific molecular signature.
Immunohistochemistry provides a powerful means to obtain fundamental insights into biological processes and disease states. Detection of protein biomarkers in tissue can be a vexing problem due to concentration variability and subjective scoring methods done one sample at a time by a pathologist. A new system developed by HistoRx (www.historx.com), called AQUA (Automated Quantitative Analysis), provides rapid quantitative readings of tissue microarrays and other samples, according to the company.
The AQUA system allows high-throughput, high-resolution analyses in a quantitative and reproducible fashion, states Paul Stroobant, Ph.D., vp head of research. It is unique because it isnt morphology-based but is linked to a fluorescent microscope that detects expression of multiple biomarker proteins by measuring the intensity of antibody-coupled fluorophores in molecularly specified subcellular compartments, such as the plasma membrane, cytoplasm, and nucleus.
Dr. Stroobant reports that a key capability of the technology is its Rapid Exponential Subtraction Algorithm. We view a multicolor slide, take two pictures in and out of focus, and then subtract the out-of-focus version to get a crystal clear picture similar to a confocal image, which allows rigorous quantification. We generate standard curves for quantification by calibrating the fluorescence signals with the quantities of antigens expressed by various cell lines and determined by standard ELISA.
Using the AQUA technology, HistoRx cofounder David Rimm found a disturbing pattern in conventional histological testing. He determined that test results for quantifying some tissue antigens with conventional histological assays can vary widely depending on the dilution of the detecting biomarker antibody used. This could have a profound effect on the interpretation of studies using biomarkers as predictors of outcome or as hallmarks of response to therapy, says Dr. Stroobant.
Although scientists can automate image acquisition from biological samples, interpreting those images is another matter. Noisy or low contrast data, especially when combined with complex images, require tedious human scrutiny, one image at a time. Definiens (www.definiens.com) says it is changing that paradigm with their Cognition Network, a technology designed to model human thought.
We are developing image-analysis tools in which we embed the decision-making processes into software that scientists use to analyze tissue, reports Zachary W. Pitluk, business manager, life sciences North America. Customers provide their own biomarkers and our software helps them assess morphological changes in tissues. This allows validating the distribution of biomarkers as well. One of the unique features of our software is that it is a scalable platform that represents a totally different approach to image analysis.
Cognition Network can extract, represent, and store knowledge from a complex input, such as images, by maintaining a hierarchy of structures in which multiple changes are represented as a pattern. Defining a pattern gives it power and can faithfully mimic human cognition of the observed phenotype.
Some companies outsource various aspects of biomarker discovery and analysis. Indiana Centers for Applied Protein Sciences (INCAPS; www.indianacaps.com) is a CRO with expertise in protein biomarker collaborations. Capital costs for instrumentation and methodology development can be considerable, notes James R. Ludwig, Ph.D., president and CEO. An alternative approach is to outsource to a company with these capabilities.
We began INCAPS with the idea to provide needed services, develop consortia for basic science problems, and validate new technologies. We also decided not to hold intellectual property. Its really a win-win situation. The company developing biomarkers gets to try the latest technology without the expense of buying it, the manufacturer gets their instrumentation independently validated, and we get to help customers.
INCAPS utilizes both untargeted and targeted approaches for protein biomarker discovery/validation. Some researchers are looking for broad protein changes. We are able to identify all proteins in a certain concentration range and determine which ones change. Since the range of protein concentrations in plasma is 1011, we cannot see them all in this way. High-abundance proteins can mask important low-abundance proteins. That is why for other clients we provide a targeted approach in which the sample is enriched and evaluated by more selective mass spectrometry approaches.
Dr. Ludwig points out that finding a biomarker is just the tip of the iceberg. Once identified, biomarkers must be validated and shown to correlate to the disease. Just finding a marker doesnt mean that it will translate well to the clinic. A researcher must ask many questions. Is the biomarker specific? Is it sensitive? Is it detectable in an appropriate way? Can it follow the course of disease or treatment?
Molecular Signatures
A single biomarker may provide clues on a disease-specific change, but a biomarker panel provides a more sensitive and compelling picture of a disease process, according to Daniel H. Farkas, Ph.D., vp for clinical diagnostics, ChondroGene (www.chondrogene.com).
Because biomarkers reflect biological activity, they provide information that may be useful diagnostically and therapeutically, potentially enabling a more rational approach to healthcare management. ChondroGene is applying what we term the Sentinel Principle, based on observations of our co-founder and chief scientist, C.C. Liew, Ph.D. Dr. Liew showed that clues in blood mirror what is happening systemically. As blood circulates, interaction occurs between blood and tissue, and detectable gene expression changes result. By profiling these changes in blood, we generate unique molecular signatures that reflect disease activity.
However, a critical requirement for generating such profiles of unique blood-based transcript combinations is how samples are generated. Appropriate sample handling is important since RNA is extremely labile. The idea of garbage in, garbage out is relevant in laboratory medicine. An advance in handling RNA clinically has been made by PreAnalytiX (www.preanalytix.com). The company developed and markets blood collection tubes that stabilize RNA immediately, enabling RNA-based biomarker diagnostics.
Some companies are focusing more of their business on biomarker discovery and analysis. PerkinElmer (www.perkinelmer.com) recently launched a new molecular medicine business unit, specializing in biomarker solutions for personalized medicine.
The company developed a platform called Bioxpression that includes sample preparation reagents, robotic instrumentation for processing/detection, and analytical software. The Bioxpression Biomarker platform combines three powerful technologies, explains Mary Lopez, Ph.D., strategic collaborations leader, molecular medicine.
First, it provides automated high-throughput carrier protein enrichment using novel membrane chemistries that were developed in association with Sartorius (www.sartorius.com). This reduces the complexity of samples and amplifies low-abundance proteins. Second, it analyzes samples using very high resolution Maldi-Orthogonal TOF mass spectrometry. With this instrumentation, pulsing ions orthogonally into the time-of-flight detection regions eliminates the influence of ionization conditions on mass accuracy. Third, the mass spectrometric data generated by the platform are analyzed with PG 600 www.nonlinear.com). This process generates biomarker fingerprints for discovery and screening applications.
PerkinElmer applied its platform to search for potential biomarkers related to Alzheimers disease. They processed more than 300 serum samples, identified four potential biomarkers, and created a fingerprint that successfully classified Alzheimers disease samples from controls.
Serum Biomarkers
Whatman/Schleicher & Schuell (www.arraying.com) offers a serum biomarker chip for the high-throughput comparative analysis of known serum biomarkers. We have developed a single antibody capture chip that can be used to identify biomarkers in serum, explains Michael Harvey, Ph.D., principal scientist R&D. It addresses the need for high-throughput biomarker screening. It is built on our FAST Slide dual pad platform in which each slide has two identical arrays of antibodies printed in triplicate on proprietary membranes. A fluorescent scanner is used for detection.
Serum proteins are directly labeled with two haptens, using a new chemistry method that targets multiple amino acids, not just the traditional lysines. This allows efficient and uniform labeling of complex mixtures. The two haptens are detected with binders coupled to two different fluorescent dyes. We spot the slide with 120 antibodies to classical cancer biomarkers. Antibodies provide a high degree of specificity and allow semi-quantitative comparisons to be made.
The first pad on the slide is probed with a mixture of two different protein samples, each labeled with a different hapten. The other pad is probed with the same two protein samples only with the haptens reversed. The ratio between the signal intensity of each spot corresponds to the protein concentration. By making comparisons with control and experimental samples we generate a pattern-recognition profile. This is ideal for risk stratification, disease prognosis, drug eligibility, and monitoring of therapy.
Whole Genome Data Extension
Biomarkers for toxicology are becoming key development tools allowing companies to screen candidates early in the discovery/development process. Use of microarray technology of gene-expression profiles produced by toxicants and stressors helps not only identify biomarkers, but also may provide critical mechanistic information.
Iconix Pharmaceuticals (www.iconixpharm.com) DrugMatrix toxicogenomic database of more than 630 compounds derived from gene-expression profiles will be extended to include whole genome data using Affymetrix (www.affymetrix.com) chips. This will provide additional data for hundreds of compound treatments in liver, kidney, heart, and cultured primary rat hepatocytes, according to the company. The database now includes data generated on the Affymetrix GeneChip platform.
