Biomarkers, typically genetic or biochemical parameters, are used as characteristic indicators to measure outcomes. They have assumed increased importance in assessing drug toxicity effects in compound screens, in diagnostic assay development, and in clinical evaluations to monitor therapy and prognosis.
A number of presentations highlighted important advances in this field at Cambridge Healthtech's recent "Biomarker World Congress," which was held in Philadelphia.
"Biomarker identification has progressed significantly with the advanced understanding of the human genome," said David K. Bol, Ph.D., vp pharmaceutical development, at Avalon Pharmaceuticals (www.avalonrx.com). "Many companies are embracing gene-expression technologies to identify markers of exposure, response, and predisposition to specific drugs by studying the global gene-expression effects of drugs in cellular and animal models, as well as in patients."
Avalon offers a suite of technologies known as AvalonRx, such as microarray systems, quantitative RTPCR techniques, and advanced bioinformatics tools, which can be applied throughout the drug discovery and development process. These tools enable one to identify genetic biomarkers as an outcome of a drug's activity and relate drugs to each other using those biomarkers. The markers are then reduced to an efficient practice of qPCR detection and followed through drug discovery and development, driving decisions in lead optimization and generating valuable information on patient responses to drugs in clinical trials.
The company successfully identified a set of gene biomarkers to determine novel modulators of the beta-catenin pathway in drug discovery. siRNA was used to distinguish a set of nine biomarker genes that reported on the down regulation of beta-catenin in cancer cells. These nine gene markers were then used in high-throughput screening with qPCR to find lead compounds that affected B-catenin protein function.
The company also identified a panel of 32 specific gene biomarkers for tracking the activity of Avalon's lead anticancer clinical candidate, AVN944, which is in Phase I trials.
Eric A. G. Blomme, Ph.D., project leader of cell and molecular toxicology at Abbott Laboratories (www.abbott.com), discussed the company's approach of using genetic biomarkers as tools to investigate compound toxicity. The company validated its gene-expression profiling technology by using two reference databases containing gene-expression profiles from rat hepatocyte cultures and rat tissues exposed to a wide variety of reference compoundsone internal and one commercial (DrugMatrix from Iconix Pharmaceuticals;www.iconixpharm.com). These data are used to develop predictive gene expression-based models that allow one to differentiate between toxic and nontoxic compounds.
Abbott, in collaboration with Iconix, also identified a panel of gene expression-based signatures (typically composed of 20-50 genes) in several tissues, such as kidney, heart, or spleen. These signature genes are up- or down-regulated in a similar fashion by test compounds that have similar modes of action or similar toxic effects.
The overall impact of this gene-expression profiling technology is significant, according to Dr. Blomme. One can identify a variety of toxicity issues using these sensitive genomic-based biomarkers in short-term toxicologic studies (3-5 days). Also, toxic changes can be detected earlier compared to traditional toxicological tools, such as histopathology. Thus, one can run short-term studies early in the discovery process to generate toxicological data on compounds, resulting in time-savings, cost-savings, and an improved selection of development candidates.
Expression Profiling Technologies
James M. Dixon, bioinformatics post doctoral fellow, drug discovery at Johnson & Johnson (www.jnj.com), presented a seminar on novel biomarker expression profiling technologies. In his talk, Dr. Dixon discussed the pros and cons of using Spotfire (www.spotfire.com) solutions, BioTrove's (www.biotrove.com) OpenArray, and novel RT-PCR screening technologies to find genes of interest for biomarker discovery.
Of particular attraction was the nanofluidic technology results obtained with BioTrove's OpenArray format, based on through-hole technology. The OpenArray system is composed of a microscope slide-sized plate with 3,072 through-holes, equivalent to eight 384-well plates. The through-holes are further arranged in 48 sub-arrays of 64 holes. Sample reactions are run in these holes in a 33-nL reaction volume. The hydrophobic surface chemistry coupled with the hydrophilic hole interior effectively maintains sample integrity and prevents cross-contamination among holes.
Results presented with TNF-alpha demonstrated highly reproducible biomarker amplification that was comparable to microplate experimentation with the attendant advantage of a shorter run. Dr. Dixon also demonstrated a novel RT-PCR cell-based assay for mechanism of action screening of oncology targets.
Presented results demonstrate that biomarker discovery could be facilitated by automation, miniaturization, and integration of existing technologies.
Identification of Clincal Applications
Kristen Antonellis, senior product specialist at Gene Logic (www.genelogic.com, discussed the use of the ASCENTA System to identify and qualify a biomarker for diseases, with a case study of Wegener's granulomatosis. This disease involves the inflammation of blood vessels and has no definitive diagnostic test. The ANCA (antineutrophil cytoplasmic antibodies) test that is used to detect the disease is often inconclusive due to false negative results. Biopsy of an affected organ, such as lung or kidney, is usually necessary for confirmation.
The purpose of the study was to assess the current state of diagnostic markers for Wegener's granulomatosis and identify potential biomarkers using the ASCENTA System, Gene Logic's web-based gene-expression profiling tool. This system houses gene-expression profiles from over 8,700 human, rat, and mouse samples, curated into over 1,600 sample sets, and holds detailed descriptions of disease pathologies from diseased and normal sample sets relevant to key therapeutic areas.
Antonellis described the use of this system for identification of Arg 1 and other putative biomarkers. Biomarker selection criteria were further applied to evaluate and narrow down the candidates to Arg 1. Tissue distribution data across normal and disease tissues favored Arg1 over other enzyme markers. Analysis at the molecular level showed that cytokines known to induce arginase in other systems were also involved in the Wegener's disease process by upregulating receptors involved in binding of these cytokines. The presence and upregulation of relevant transcription factors needed for arginase expression qualified Arg 1 as a definitive blood biomarker for this disease.
Targeted Mass Spectrometry
"A major challenge in biomarker research is how to qualify the potential discovery candidates that result from genomics and mass spectrometry-based discovery experiments," said Tina Settineri, Ph.D., director, Q TRAP proteomics product line at Applied Biosystems (www.appliedbiosystems.com).
"Discovery is typically performed with a small sample or pooled sample population. It is time consuming to find a candidate biomarker with typical discovery workflows. Inconsistent results across duplicate experiments and the biological relevancy of the biomarker are some of the issues that need to be dealt with for discovery qualification. Applied Biosystems enables biomarker discovery qualification with the 4000 Q TRAP System and the MIDAS workflow (which utilizes MRM-Initiated Detection and Sequencing). The system is based on targeted mass spectrometry," she explained.
The power of the 4000 Q TRAP System was discussed by John Hevko, senior field applications specialist. He presented findings that mapped the development of protein biomarker MRM panels for cardiovascular disease, based on quantitative screening of plasma samples on their system.
The targeted workflow is enabled by the MRM triple-quadruple system (Q1,Q2,Q3) and is a well-validated technology. The system can be configured to detect signal from specific peptide molecules in a sequential fashion to generate potential peptide profiles, which are surrogates for the protein candidates of interest. The profiles are developed using the MIDAS workflow, which allows the researcher to predict, develop, and refine unique MRM profiles without any protein or peptide standards as part of downstream method/assay development. These assays enable high-throughput, quantitative profiling of clinical samples.
"Identification of protein biomarkers is important because a drug directly acts on proteins," commented Jerry Feitelson, Ph.D., manager of strategic marketing at Beckman Coulter (www.beckmancoulter.com). "Protein biomarkers are generally more predictive compared to genetic or transcriptional biomarkers." They can be used to characterize candidate drugs and are extremely valuable in clinical applications to diagnose disease and monitor therapies.
Protein biomarker research requires analysis of complex biological mixtures that vary in concentration. The biomarker of interest is more often than not present in very low quantities in these samples, relative to the thousands of other proteins present.
"Partitioning and fractionation are two key steps for sample analysis to enrich candidate biomarker proteins,"explained Hans Dewald, product manager for protein enrichment products. "Passage of human sample biofluids, such as plasma, serum, or CSF, through Beckman Coulter's ProteomeLab IgY chemistry partitions 12 of the highly abundant proteins, enriching medium- to low-abundant proteins.
"As opposed to depletion strategies, partitioning allows the captured proteins to be eluted off for additional investigation. We use polyclonal IgY antibodies for the partitioning process because they generate higher avidity capture, have higher specificity, and can be used cross species. This approach can handle up to 250-L sample load, resulting in 1-2 mg of enriched protein of interest per cycle."
Once the samples are partitioned, they can be further enriched through proteome fractionation in two dimensions. The Beckman Coulter ProteomeLab PF 2D is a liquid-fractionation system that allows high mass loads to be fractionated by pH and hydrophobicity. This enables the subsequent identification of proteins at much lower abundance levels, while preserving post-translational modifications, both of which are likely biomarker candidates.
The PF 2D generates liquid fractions of the proteome, as well as pH/hydrophobicity maps, that allow users to visually identify regions of differential protein expression. "Our customers have successfully applied this technology to protein biomarkers, eight orders of magnitude deeper than human serum albumin," said Dr. Feitelson.