Biomarkers: Biologists’ Best Friends
Tue, 2005-11-08 07:25
ChondroGene laboratory technologist extracting RNA from whole blood samples. The RNA extracted from these samples is used in the development and testing of the company's blood-based colon cancer test.
Biomarkers have become the workhorses of basic research, drug discovery, preclinical drug development, clinical testing, and at their most refined, human diagnostics. Under their broadest definition, biomarkers include any biological or chemical indicator of an underlying process. Physicians often refer to medical images as “biomarkers” as well. Mario Ehlers, MD, PhD, CMO of Pacific Biometrics (Seattle, WA), defines biomarkers from a purely drug-development perspective, as “any biological indicator that might be useful in understanding a drug’s action or safety.” Ehlers includes conventional biochemical tests, pharmacogenomic and proteomic assays, and imaging techniques like magnetic resonance imaging and x-rays. For purposes of this article, biomarkers are defined as chemical surrogates for any underlying biological event.
Promoting a newly-discovered biological indicator to biomarker status involves first determining its relevance to the drug target. At this stage developers begin thinking in terms of specificity, sensitivity, and statistical relevance. Once the science is out of the way, researchers validate the protocol for robustness. In other words, how good is the test as a test? Can it be run in a format appropriate for the project at hand? Does it provide results rapidly and reliably? Finally, the biomarker must compare favorably with well-established clinical or laboratory parameters known to be relevant to the disease or biological effect of interest. For example, a cartilage degeneration biomarker for osteoarthritis should support and confirm clinical findings and x-rays.
New drugs would experience nearly impossible approval hurdles were it not for novel, specific, and sometimes custom biomarker tests employed during clinical trials to assess drug safety and efficacy. Biomarker use during clinical trials is one area where the U.S. Food and Drug Administration permits drug sponsors and their clinical partners wide scientific latitude and creativity.
Accumulating enough data to move a biomarker into development as approved human diagnostics can take many years. For example, bone turnover biomarkers have been used for fifteen years to monitor the efficacy of osteoporosis drugs, and have been partly responsible for development of new arthritis drugs. But clinicians still do not use these assays routinely in clinical practice. C-reactive protein, a biomarker for inflammation useful in diagnosing coronary events, has been used for more than a decade and has only recently gained acceptance among cardiologists.
"One hopes that over the course of clinical development, the new biomarker will provide information that is more dynamic, sensitive, and specific for disease progression and modification by the drug of interest," notes Dr. Ehlers. "If you reach the point where the biomarker is the single best measure of efficacy, it becomes a de facto companion diagnostic."
Ehlers referred to the FDA's Critical Path initiative, in which the agency called for the use of prospectively identified biomarkers to help drug developers raise the risk-benefit ratio during clinical trials. The FDA initiative also sets guidelines for development of diagnostic products based on safety, efficacy, and manufacturability.
Turning biomarkers into approved diagnostics or clinical chemistry products is far from straightforward, since diagnostic products must pass FDA approval in a process similar to the one for drugs. However, FDA's open-minded attitude towards experimental biomarkers used to monitor pharmaceutical efficacy has been a boon both to diagnostics and basic science. "Central labs must comply with certain standards for qualifying the assay and validating its performance, but beyond that the agency is very receptive to the use of new assay, imaging, and other testing procedures," notes Dr. Ehlers.
Joseph F. Clark, Ph.D., Associate Professor in the Department of Neurology at the University of Cincinnati, studies bilirubin oxidation molecules, or BOXes, biomarkers and causative agents involved in vasospasm. It took over 4 years to come up with a structure of the small molecule that takes about 10 seconds to draw, and another 5 years to demonstrate its role in the spinal fluid of the stroke patients.
Sugars are in many ways the forgotten macromolecule. Proteins, RNA, and DNA have long served as biomarkers and therapies but sugars, because of their difficult chemistry, have attracted relatively little development-stage interest from drug and diagnostic companies. Glycominds (Redwood City, CA) is an exception to "glycol-phobia." The company develops sugar-based biomarkers for important inflammatory diseases
Glycominds can distinguish between these diseases by analyzing glycan antibody biomarkers specific for each disorder. The multiple sclerosis biomarker panel, for example, tests for eight glycan antibodies by attracting these proteins to sugars anchored to microwells. Glycominds offers a commercial test for inflammatory bowel disease based on three glycan antibody markers, and plans approved tests for other inflammatory diseases.
"For multiple sclerosis this is a real breakthrough because it not only can diagnose the disorder, but predict the next flareup," says Steve Lehrer, President. Patients and their physicians can manage the disease much more effectively by withholding the strongest medications for times when they know the next attack will come. A glycan antibody biomarker panel could replace much more costly, less versatile spinal taps and magnetic resonance imaging.
Researchers are approaching genetic profiles as biomarker panels from a variety of angles. ChondroGene's (Toronto, ON) Sentinel approach, developed by co-founder and chief scientist Choong-Chin Liew, PhD, is based on an old idea: circulating blood reflects health or disease status. Interactions between white blood cells and diseased tissues induce differential gene expression in the cells, which serve as markers for such maladies as cancer, central nervous system disorders, heart disease, and arthritis. Bruno Maruzzo, Director of Corporate Development, claims cell-cell "communication" can also differentiate between colon, prostate, and bladder cancers.
During an early-stage Sentinel project hundreds of genes will be found to deviate from normal expression patterns. ChondroGene scientists apply quantitative polymerase chain reaction to reduce that number to between two and ten marker genes that provide the best sensitivity and specificity. "We don't know what we're looking for when we begin an investigation," Maruzzo admits, "but eventually the data speaks for itself. The genes just pop out."
Sentinel has helped ChondroGene identify marker panels for diseases that have stumped other diagnostics developers. In January, 2005, the company published results on tests for distinguishing schizophrenia from bipolar disorder.
Chondrogene's goal of manufacturing approved diagnostics is on track to succeed. The company will introduce a $100 research-only colon cancer prototype test next year, which the company hopes will replace many colonoscopies. Development-stage products are currently offered to monitor disease progression, drug effectiveness and toxicity, cancer recurrence, and identifying responder populations for drug trials.
click to enlarge
ChondroGene's Sentinel Principle is used to detect diseases from blood.
Most biomarkers are secondary to biological processes, meaning they do not cause disease directly, but are rather byproducts of disease. Elevated levels of bilirubin, for example, are present in liver disease. Bilirubin is also believed to cause mental retardation in infants, but this effect is distinct from its role as a liver disease marker.
In the late 1990s Joseph S. Clark, PhD, Associate Professor of Neurology at the University of Cincinnati (Cincinnati, OH), discovered a set of biomarkers known as BOXes (bilirubin oxidation products), which are not only biomarkers for hemorrhagic disorders, but are believed to exacerbate conditions such as stroke, post-surgical complications, and compartment syndrome (compression of nerves and blood vessels). BOXes, which arise from oxidative stress, damage bilirubin, induce vasoconstriction and neurotoxicity, and inhibit numerous desirable phosphatase pathways. "BOXes are not just diagnostics, but prognostics," says Prof. Clark. "BOXes cause things to happen."
"One of the pitfalls of biomarkers is they will tell you if something has already happened, but don't suggest what to do next," Clark notes. "BOXes are the kind of biomarker that could very well point to possible therapies. They are effectors as well as markers."
Skip the protein
Metabolon (Research Triangle Park, NC) uses mass spectrometry to develop biomarker patterns based on metabolomics. Where genomics and proteomics examine gene and protein expression, metabolomics seeks to quantify and correlate all small-molecule biochemicals in a sample. With proteins and genes removed, Metabolon can focus on small molecules like glucose, cholesterol, glutathione, and many others. "The best known biomarkers are small molecules," observes CEO John Ryals, PhD. "Most of the metabolic disease tests newborns undergo are based on small molecule biomarkers."
Metabolon has recently received its third grant, from the National Institute of Environmental Health Sciences (NIEHS), to study markers for amyotrophic lateral sclerosis (ALS). The other two awards were funded by the National Institute of Neurological Disorders and Stroke and the ALS Association. Since there is no effective treatment for ALS, biomarkers will help distinguish it from other neurodegenerative diseases such as Alzheimer's and Parkinson's.
Dr. Ryals believes that for many diseases, small-molecule biomarkers are superior to high-tech gene and protein panels. "There's a certain dogma out there that diagnostics must be based on molecular biology," he says, "but I'm not sure that idea will hold up because of the multiple effects and pathways for what we think of as single diseases."
Biomarkers are just one tool in the effort to make biology relevant. While developers and end-users continue to seek "silver bullet" biomarkers, most realize that the elucidation of complex biological events will probably require multiplexing new molecular biology tools with more traditional small-molecule biomarkers and, when applicable, traditional pathology and anatomy. A big-picture view of disease is only possible through an eclectic view of chemical and non-chemical markers.
Absent this approach, the danger will always remain that biomarkers will model an academic definition of disease perfectly, but provide little insight or comfort for real-life prognosis. The last thing bioscience needs is this twist on an old adage: The biomarker profile was favorable after drug treatment, but the patient died anyway.