Biomarkers Come of Age - Pharmaceutical Executive


Biomarkers Come of Age
In the past five years, biomarkers have become an essential part of pharmaceutical R&D. Seven industry experts explain how it happened—and what comes next.

Pharmaceutical Executive

In early development at Eli Lilly, for example, many different markers are pursued simultaneously, including DNA, RNA, protein, and metabolites; the goal is to think in terms of systems biology when deciphering the data.

"All these technologies are relevant," says BMS' Dracopoli. The mix of technologies chosen to create a panel depends on the biological endpoint being investigated. "Panels of biomarkers are required when biologically heterogeneous endpoints are pursued," he says.

There is wide agreement that panels of biomarkers will become the norm. "Multiple biomarkers are probably a better reflection of the disease state," said Mitch Dowsett, PhD, Royal Marsden Hospital, London, at the American Society of Clinical Oncology's 2005 annual meeting. They also provide a better view of the "multi-dimensional quality of clinical response," as Douglas Throckmorton, MD, acting deputy director of FDA's Center for Drug Evaluation and Research, put it at the 2005 Biomarker World Congress.

Biomarker discovery Pharmaceutical companies generally pursue biomarkers in relation to pathway analysis and in conjunction with hypothesis-driven, deductive, knowledge-based target discovery. However, work on biomarkers presents a scientific opportunity to rediscover biology by better understanding individual diseases and their associated pathways.

Lilly's Edmonds gives an example. When his company was developing Xigris (drotrecogin), a new treatment for sepsis, the project team went to the literature and selected six classical markers of inflammation. When they were tested as a possible biomarker set, though, they failed.

A second set of biomarkers was chosen, using an inductive, data-driven, "pure discovery" approach. Not only were the markers in this second set different—but they had not previously been considered as a combination for sepsis. The second set was put into validation studies and brought forward.

Toxicology There is currently a lack of biomarkers for important toxicity endpoints and for toxicology in general, although several companies mentioned that major breakthroughs will be forthcoming this year.

At Lilly, biomarkers are being used to predict hepatotoxicity from preclinical information. The company is working with FDA to develop biomarker standards that will enable the company to evaluate hepatotoxicity in the clinic.

An important goal involves using biomarkers to discover ways to identify toxicity as early as possible in animal studies. For example, Novartis researchers who conducted animal studies for a number of compounds targeting inflammation found that particular gene expression signatures (RNA) in animal gastric tissue could predict gastrointestinal toxicity as early as day one.

Once they had identified the signature, they found that it also occurred in white blood cells, indicating that blood testing could be used as surrogate for GI tract toxicity for this class of compounds. (Traditional methods of determining animal toxicity require at least one full month to be predictive for humans.)

In a broader application, Novartis is also showing that organ-specific gene expression signatures can be used to detect the effects of compounds on various organs.

Drug response Although it appears that many companies are working on biomarkers in the area of drug response, few have commented on specific examples. Two exceptions are Novartis and BMS.

Working on a nervous system indication in which there was a large placebo effect, Novartis evaluated genotypes and identified 12 single nucleotide polymorphisms (SNPs) in six genes that permitted researchers to stratify patients to maximize the difference in response after four and 12 weeks of treatment. (For more on SNPs, see "Snip, Snip, Snip")

Fifty percent of patients did not have the SNP variation; their response rate was similar to the placebo group (31 percent compared to 27 percent). When patients had the defined changes in the SNPs, the drug-treated group demonstrated a markedly higher rate of response (62 percent) compared with the placebo group (23 percent).

These genes were either associated with the drug target or were mechanistically plausible. This was the first time that this combination of biomarkers was demonstrated for this type of disease; it was validated in a second independent study according to the new guidance on pharmacogenomics data submission.


blog comments powered by Disqus

Source: Pharmaceutical Executive,
Click here