Send to printer »

Feature Articles : Apr 15, 2010 (Vol. 30, No. 8)

Strategies to Improve Biomarker Successes

Collaboration Is Essential to Advance Field and Improve Drug Development and Patient Care
  • Elizabeth Lipp

An oft-heard complaint in R&D is that technology can’t keep up with the concepts that scientists are trying to prove or disprove. “In the case of biomarkers, having the technology is simply not the problem,” says Daniel Chelsky, Ph.D, CSO of Caprion Proteomics. “In the last five years, mass spec has improved a bit, but only by about an order of magnitude in sensitivity. The technology is there, but most labs are not using it to its full potential.”

According to Dr. Chelsky, we will be hearing a lot more about mass spec-based proteomics because the field has matured to the point where it can now have a significant impact on drug discovery, from both the perspective of discovering new markers and validating candidate markers.

At “Biomarker World Congress” to be held in Philadelphia next month, presenters will talk about advances in the field, and also consider the perks and pitfalls as biomarker validation finds its way to standardization.

Caprion has implemented three new tools geared to improve the success rate of biomarker identification. “Our main focus is on the use of multiple reaction monitoring,” says Dr. Chelsky. “MRM allows you to multiplex up to 100 peptide candidates at a time, getting a very quantitative and convincing assessment and verification of candidate markers with speed, accuracy, reliability, and without the expense associated with discovery. This is why people are looking at MRM as the next step in mass spec-based proteomics and biomarker validation.”

This approach can also be used for specific proteins. “Recently, a company told us that the sensitivity of its ELISA was not sufficient. In two weeks, we were able to develop an MRM assay using an antibody pull-down that could quantify their protein at 1 ng/mL in plasma. We can go much lower with small proteins and peptides, even without antibody pull-downs. The other advantage of MRM is its specificity that antibodies alone can’t match.”

And the lower abundance proteins are where the most interesting findings are. Dr. Chelsky notes that researchers have become more selective. As a result, Caprion has developed a way to examine secreted proteins by isolating golgi and secretory vesicles, breaking them open, and comparing the contents under different conditions.

“We’ll look at diseased versus healthy, or treated versus untreated, and what we see in the ‘secretome’ will mimic what we later find in the blood. The advantage is that, when we isolate proteins straight from cells or tissues, they are highly concentrated and you know where they came from. We’ve been particularly successful with cancer because we have been able to get fresh human tumor tissue, especially in metabolic disease where people have been quite willing to give up some of their visceral and subcutaneous fat.”

Depletion columns mark another difference. These columns can deplete the high and moderately abundant proteins in plasma, removing between 95% and 99% of the total protein, resulting in a 20-fold minimum improvement in sensitivity. “But proteins that are not normally present in plasma are also found after using these  columns. These tend to be the ones that are most interesting and now there is a way of detecting them,” Dr. Chelsky adds.

Consortium

The Biomarkers Consortium is a partnership of the Foundation for the National Institutes of Health (FNIH) that brings together the resources of NIH, FDA, the pharmaceutical industry, nonprofit organizations, and academic medical research centers to develop and qualify biomarkers that support more effective therapeutic development and patient care.

The Consortium “helps qualify biomarkers and validate related analytical platforms in a precompetitive context,” says David Wholley, director. “There is a clear need to normalize and standardize our approach to biomarkers and underlying technologies across a therapeutic class or a common platform, such as the use of specific imaging modalities for diagnosis or to assess response to treatment. And, consensus across all stakeholders involved in these efforts is critical to their success.”

The consortium now has eight launched projects and many others in development. At the meeting, Wholley will talk about a collaboration with the Predictive Safety Test Consortium (PSTC) of the Critical Path Institute to provide clinical qualification of multiple biomarkers of renal toxicity that PSTC had previously worked on with industry to qualify for use in preclinical phases of drug development.

Wholley will also talk about the state of the field, and the role his consortium plays in it. The FNIH has been developing these kinds of biomedical research partnerships in support of the mission of the NIH for the past 12 years, but the Biomarkers Consortium is unique for providing an ongoing mechanism to develop a variety of projects of interest across multiple therapeutic areas.

“Now, more than ever, there is a clear need for all participants in the biomedical research field to collaborate in the interest of moving the science forward, given the complexity of the research and current economic realities,” Wholley explains. “When managed appropriately, these kinds of public-private partnerships are a particularly effective mechanism for driving collaborations.”

Early Development

Shaking up the drug pipeline means rethinking the way drugs are currently moving through the pipeline. “Confident, data-based decision making in early development of novel medicines is a critical capability in the pharmaceutical industry,” says Hans Winkler, Ph.D., senior director and global head, oncology biomarkers at Ortho Biotech Oncology Research and Development, a division of Janssen Pharmaceutica.

“The way we are looking at handling early clinical trials now is that we have to study safety first and do those escalations as rapidly as possible in early cohorts where we are working on low doses and will collect only peripheral, noninvasive biomarker samples and not have mandatory tumor collection. Traditionally, we used to escalate dose until we established a maximal tolerated dose (MTD).”

However, Dr. Winkler notes, with new targeted therapies, “we may not ever see a MTD, but we might actually have to rely on biological activity and we’re thinking it might be a biologically effective dose that will be the Phase II dose.”

Once his group figures out a biologically effective dose, it will go into an expansion phase that will focus on specific tumor types, where the researchers believe that the target plays a controlling role. “We will test predictive markers to identify some populations, and in that focused population we will examine antitumor activity, looking at tumor cells pre- and post-treatment to make sure that we are giving the optimal dose.”

The biggest challenge is getting tumor samples in early trials. “It’s not feasible to collect serial tumor biopsies, and the quality of biopsies is always an issue,” explains Dr. Winkler. “Logistics are still difficult, especially when it comes to larger trials. If you collect samples, you sometimes lose half of the samples in some way. We need to collect more like 80 to 90 percent usable samples.”

For pharmacodynamic analysis, Dr. Winkler reports, “we need an early start and enough support to actually develop the assays to the degree that they are acceptable for clinical applications ahead of time. For predictive markers, we need to work up the assays to a point where we can actually test them in Phase II, we need to sort out the regulatory process, we need to have a strong hypothesis to be tested when we get into Phase II, and we need more predictive models to allow us to discover these predictive markers ahead of time, rather than in the clinic.”

Alcoholism

Willard Freeman, Ph.D., from the College of Medicine at Pennsylvania State University, who conducted an alcoholism study with Kent Vrana, also from Penn State, and Kathy Grant from the Oregon National Primate Research Center, will talk about primate-based research at the meeting. “Right now there is no definitive clinical test to determine whether someone has an ongoing alcohol abuse problem, but we are looking to change that,” Dr. Freeman says.

While there are social implications to these findings, Dr. Freeman emphasizes that it’s about making a definitive diagnosis. “Our goal is to create a tool to provide a starting point for treating problem drinking.”

In the study, plasma protein biomarker discovery and validation were undertaken using an alcohol self-administering non-human primate model system. The goal was to develop a diagnostic that accurately classifies subjects into nondrinking, non-abusive drinking, and abusive drinking categories.

Plasma protein profiling was performed at Rules-Based Medicine (RBM). Initial biomarker discovery was conducted by multiplex Luminex analysis of 90 known plasma cytokines, growth factors, and other proteins samples. As a result, a 17-plasma protein panel was determined that correctly classified abusive drinking with 100 percent sensitivity and also differentiated any level of drinking from alcohol abstinence with 88 percent accuracy,” Dr. Freeman says.

The results of this well-controlled monkey model should be validated in humans next, he adds. “The panel provided by RMB represents proteins from the entire body. Being able to stratify the monkeys into three groups is an important part of the study, since differentiating socially acceptable drinking from excessive is critical.”

Metabolites

Metabolites constitute a broad range of biochemicals, including clinically accepted biomarkers like glucose or cholesterol, and serve as the core of Metanomics Health’s mass spec profiling platform, which the company uses to analyze a endogeneous and xenobiotic metabolites from body fluids and tissues, such as amino acids, lipids, carbohydrates, and many more, notes Tim Boelke, Ph.D., managing director.

“The science has evolved to a point where many of these small potential biomarkers can be simultaneously measured and provide a snapshot of biochemical homeostasis and physiological status, or its response to disease or drug treatment.”

Kristina Busch, Ph.D., business development manager, says that the company’s mass spec-based platform combines a broad profiling approach with targeted methods to provide the greatest coverage of nearly all metabolite classes. “We can also find low-abundance hormones such as catecholamines and steroids, as well as different classes of lipids and highly polar species. We’re gearing toward prospective and retrospective biomarker studies in high medical need areas.”

Since metabolism is sensitive to changes induced by drugs and diseases and also confounding factors like nutrition, lifestyle, and environment, robust study design practices and sampling are important to a successful biomarker identification campaign.

“This is not unlike all biomarker discovery methodologies like transcriptomics and proteomics. But, because of the sensitivity inherent in measuring biological response to stimuli, our methods can be applied to all kinds of problems, from understanding the metabolic effects of different nutritional and exercise regimes to how you best feed a cell culture for a recombinant-protein bioreactor run,” explains Dr. Boelke.

Assay Validation

Cirion has extensive experience performing PK and immunogenicity assays to support preclinical and clinical trials.

“You can analyze a biomarker in a variety of ways,” explains Lorella DiDonato, Ph.D., vp of R&D. “We focus on making sure that we have a diversity of technologies to ensure that the most appropriate technology is available to perform the analysis for each biomarker.”

A major challenge, according to Dr. DiDonato, is when the sponsor does not have a clear idea of the intended purpose of the biomarker analysis—whether it is exploratory in nature or part of a clinical endpoint.

“With PK assays, the intended purpose  is, in a majority of cases, a primary endpoint. At an early stage in the drug development program, the analysis of several biomarkers is often exploratory in nature since it is unclear which of the biomarkers is the key marker that will support efficacy or safety of the drug in development.

“At this stage, it is appropriate to qualify the assays and analyze the samples to evaluate the levels present in healthy versus patient populations, as well as how the levels are affected pre- and post-administration of the drug. Once a biomarker is qualified, full validation, which will be performed exactly as a PK assay, is required. 

“If there are commercially available kits to perform the analysis in an exploratory analysis, we will not challenge the methodology of the kits but will evaluate in a feasibility testing whether the kits are performing according to manufacturer’s specifications. Should the results concur with the kit specifications, we will then proceed with the analysis and discuss with the client whether the results are informative. 

“At this point, the list of biomarkers to analyze is reduced to just several biomarkers. We will then challenge the methodology of the kits to ensure that the assay is enhanced in terms of robustness, precision, and accuracy. We will also prepare quality controls and not use the controls of the kits. In addition, we will perform stability evaluations and not rely on the manufacturer’s stability data.”

Dr. DiDonato will talk about one instance when a commercial kit was available to perform a biomarker assay validation, but it didn’t perform as expected.

“We ended up developing an assay from scratch similar to that of a PK assay. Often  the perception is that a biomarker assay, especially when a commercial kit is available, should not require the same time and cost as a PK assay. Therefore, the challenge we face is: how do you present to the sponsor that the development and validation of a biomarker assay will cost as much as if it were a PK assay, which is a primary endpoint?” concludes Dr. DiDonato.