February 1, 2014 (Vol. 34, No. 3)
Kate Marusina Ph.D.
Biomarkers defining specific phenotypes are becoming increasingly important for developing new drugs for specific patient subpopulations. The value of a new biomarker is measured by its ability to reduce risk.
Ideally, the biomarker should be developed in parallel with the new drug, as nearly 50% of the projected development costs can be saved by shutting down a development program before it enters Phase II. A meaningful risk-benefit analysis of a biomarker requires estimates of its cost and accuracy, as well as the consequences of decisions that it will enable.
For the biomarker to be of value, the cost of its development has to be less than the projected costs of development from Phase II onwards, discounted to present time. While multiple competing business considerations affect a pharmaceutical company’s decision to proceed with a biomarker program, the skyrocketing market for biomarker discovery underscores the pharmaceutical industry’s hope that biomarkers will bolster the success rates of pipeline products.
“Imaging biomarkers have been largely underutilized in drug development,” says Kevin Cox, Ph.D., CEO of London-based Imanova. “But we believe that molecular imaging has the power to assist in successful translation of molecules by reducing the risk of several specific causes of failure in Phase II clinical studies. Imaging biomarkers, or i-biomarkers, are especially valuable in giving confidence of tissue delivery, determination of target engagement, and the evaluation of a drug’s pharmacodynamic effects.”
While imaging is routinely used in clinical diagnostics for cancer, its acceptance in drug development has been slow. “This is a highly specialized area of knowledge,” Dr. Cox observes. “Designing imaging experiments to answer the right questions is not trivial. Combined with the perceived high costs and dearth of well-equipped facilities, this has slowed down the adoption of imaging as an integral step in drug development.”
Imanova presents an innovative and highly integrated solution in reducing the barriers for use of molecular imaging. Located in the former GlaxoSmithKline imaging center, Imanova’s staff applies the knowledge needed for translational application of imaging science.
“Another historical barrier for use of molecular imaging has been the lack of versatile PET tracers for key therapeutic targets,” remarks Dr. Cox. Together with its pharmaceutical clients, Imanova develops proprietary tracers that can answer critical questions about target engagement directly after drug administration. A structured approach for i-biomarker development takes the novel tracer from the candidate pool to clinical validation.
Uniquely, Imanova utilizes in silico biomathematical modeling to predict a candidate with ideal physicohemical characteristics. “The i-biomarker development pipeline adheres to a strict quality system,” continues Dr. Cox. “We not only provide candidate selection and labeling, but also rigorous preclinical evaluation in several species, combined with blood chemistry or other physiological measurements.”
The resulting biomarker provides quantitative information to make informed go/no-go decisions. Imanova hopes to develop an open innovation approach to i-biomarker research, and to encourage pharmaceutical companies to collaborate on tracer development.
“By collaborating in this pre-competitive space, a pharma-academic consortium can de-risk i-biomarker development programs and generate new tools to eliminate costs associated with futile activities downstream,” concludes Dr. Cox. “Most tracers need to be utilized early in the drug development process. Used at the right time, imaging biomarkers are able to inform the design of Phase II studies, including dose ranging and possibly patient selection, saving many months in development and millions of dollars in costs.”
Answers from Big Data
“Clinical bioinformatics is the application of a data-driven, high-tech approach in clinical setting,” says Jerome Wojcik, Ph.D., CEO of Quartz Bio, a clinical bioinformatics service provider located in Plan-Les-Ouates, Switzerland. “We use clinical bioinformatics to adapt treatment to patients, that is, to identify cohorts that respond to the drug in a predictable manner,” says Dr. Wojcik.
Pharmaceutical partners supply Quartz Bio with data collected in a course of clinical trials. The data (which may include information from protein and RNA expression, genotyping, molecular diagnostics, and flow cytometry studies) often exists in silos within a pharma company. To make sense of the data, Quartz Bio integrates heterogeneously formatted data, analyzes it for consistency, and identifies gaps and outliers.
Dr. Wojcik’s team dedicates over 40% of the overall analysis time to the biomarker data management. This key step is crucial for the quality of the overall analysis. According to Quartz Bio, all the data-management processes are documented, auditable, and reproducible.
Once the “Big Data” horde is adequately cleaned up, the team applies adaptive statistical methods to generate multiple hypotheses linking the drug action with subpopulations of patients. “Our challenge is to generate reliable hypotheses on a fairly small statistical patient sample, for example, a thousand patients, but using millions of biomarker datapoints,” continues Dr. Wojcik. “We do not rely on statistics alone. Graphical visualization adapted to the objectives of the study is necessary for interpretation of results.”
In a recent project, Quartz Bio analyzed multiple oncology biomarkers, such as gene expression, circulating tumor cells, and immunohistochemistry, to identify patient cohorts that would most likely benefit from a novel treatment. Biomarker analysis revealed a subpopulation whose survival rate increased significantly over the population average, bringing a potential application of personalized medicine closer to reality.
In another example, Quartz Bio correlated the genetic and pharmacokinetic data to categorize the patients into fast and slow metabolizers of an investigational drug. In slow metabolizers, the compound persisted in the bloodstream long enough to create a toxic effect; in fast responders, the drug was excreted before it could reach the target.
Such stratification analysis assists in adapting therapies to the biological characteristics of a patient, as measured by biomarkers. “Precision medicine is a fast-growing field that in the near future will have tremendous impact on drug development, insurance reimbursement, and disease management,” concludes Dr. Wojcik.
miRNA as a Blood-Based Biomarker
The Heidelberg-based Comprehensive Biomarker Center (CBC) specializes in innovative biomarker discovery and validation, with a specific focus on microRNA (miRNA). miRNAs regulate genes post-transcriptionally in what seems to be virtually all cellular pathways.
“Traditionally, miRNAs were isolated from solid tissues, such as tumor samples,” says Markus Beier, Ph.D., vice president of strategy and IP. “CBC developed a unique pipeline for discovery of miRNA biomarkers in blood cells.” Because microRNAs are remarkably stable in blood and because their blood levels do not fluctuate under normal physiological conditions, these molecules hold great promise as biomarkers of pathogenic processes.
When answering clinical questions, CBC narrows a set of 2,000 currently known miRNA biomarkers to a biomarker signature consisting of approximately 2–50 molecules, with the goal of further reducing the number of miRNAs in assay and clinical validation. The company, which holds several patent applications covering biomarker signatures for disease diagnostics, focuses its current clinical validation efforts on acute myocardial infarction (AMI), heart failure, prostate cancer, and multiple sclerosis.
“AMI is a very challenging indication, because time is of the essence,” states Dr. Beier. “Cardiac troponins are the gold standard for AMI diagnostics, but due to their high sensitivity, they also pick up other cardiovascular diseases. Our miRNA signature is designed to complement this test and seems to reflect processes that precede myocardial necrosis.”
CBC and its clinical partner, the department of cardiology at the University of Heidelberg, compared 20 patients who had AMI with 20 patients who reported chest pain but did not have AMI. Out of 121 differentially expressed miRNAs, a signature of 20 allows for diagnosing AMI with 93% accuracy. In addition, a first-ever kinetic study showed that changes in the miRNA are the most evident during early stages of AMI, and the differences with the control diminish 24 hours post AMI.
To underscore the companion value of miRNA profiling in blood cells, the company notes that its biomarkers can provide information about disease processes originating from the immune system. For example, CBC developed a set of 48 miRNAs that allow for the discrimination of multiple sclerosis (MS) patients from normal controls with an accuracy of 96%. The company also refined a set of miRNA candidates for monitoring MS therapy. “Finding reliable blood-based biomarkers means improving diagnosis of the disease, and rapid selection of personalized treatments,” concludes Dr. Beier.
Mining Published Works
“Could drug research be supported by mining the texts of knowledge repositories, such as PubMed? We believe that semantic technologies bring new e-biomaker discovery opportunities, in a manner which is 50% more cost-efficient than traditional molecular biomarker discovery process,” asserts Paul Walti, Ph.D., CEO of InfoCodex, a software provider based in Buchs, Switzerland.
InfoCodex software distinguishes itself from more traditional software, which is based on natural language processing, by recognizing and “comprehending” the actual content of a large number of unstructured documents. By analyzing seemingly unrelated documents, publications, and reports, InfoCodex can categorize unstructured information and correlate small, seemingly unrelated facts.
The software combines a very large thesaurus organized in a complex taxonomy of about 10,000 concepts. It applies information theory to transform documents into mathematical models, conduct unsupervised semantic clustering, and match multilingual documents according to meaning.
“Our software is able to determine the meaning of unknown words and correlate them with words in the InfoCodex Linguistic Database, providing a cross-language content recognition,” says Dr. Walti.
As opposed to natural language processing, which recognizes relationships between facts only if they are already explicitly stated in the document, the InfoCodex semantic engine can find hidden correlations distributed over groups of documents. “Our engine is almost like a supreme human super-reader, except that no human team, however specialized, is able to create and maintain the overview of all publications, simply because of their sheer number and the rate of accumulation,” continues Dr. Walti.
In a pilot experiment with Merck, InfoCodex took on the task of discovering new e-biomarkers from over 120,000 PubMed abstracts, clinical trial summaries, and internal Merck documents. Without any involvement of subject matter experts, InfoCodex was able to identify over 10,000 potential biomarker/phenotype candidates, further narrowed down to just 1,000 of specific genes or proteins.
Next, the Merck and Thomson Reuter scientists further parsed the cohort to about 20 novel high-quality candidates prioritized by confidence scores. Some of these have since been validated in biological experiments.
It should be noted that none of the 20 abstracts contained the term “biomarker.” Despite the constraints of this first pilot experiment, the ability of automated data mining to identify new e-biomarker candidates has the potential to impact pharmaceutical research.