Mapping Out New Strategies

Research into the use of biomarkers as indicators of disease, as well as of the potential for response to treatment, is a fast-developing area, and Christina Coughlin is right in the thick of it. As Associate Director in Discovery Translational Medicine at Wyeth Research, she leads research in early biomarker discovery efforts in translational medicine in the oncology pipeline.In this capacity, she is responsible for biomarker discovery in the early stage proof of concept, getting involved in the early discovery process.

Coughlin and her team devise biomarker strategies and early medical concepts within the discovery stages of drug development. Translational scientists at Wyeth work on five main areas of biomarkers – target validation, target compound interaction, pharmacodynamics, disease and patient selection.While these are exciting times for the field, Coughlin recognizes that there are a number of challenges to moving forward. The biggest of these involves the availability of preclinical models to understand the activity of compounds in the preclinical setting that recapitulate human disease. Several animal models are being investigated in the preclinical setting, including panels of human tumor xenograft models, genetically engineered murine models (transgenic models), as well as companion animals, or pets (dogs, cats) with spontaneous cancers being treated at veterinary centers.

Coughlin is optimistic on this front: “Early studies suggest that newer animal models such as transgenics and companion animals may more accurately predict the compound’s efficacy in the clinic; however, more studies are needed, including a rigorous bedside back to the bench approach, to fully understand and assess the predictability of the various animal models. We haven’t figured out yet, in a preclinical setting, what the optimal model is to understand compound sensitivity and resistance. If we get down to the basics of patient selection, then this is what you need to understand to build a hypothesis of which patients need which drugs,” she says.

No less important is the tumor and surrogate tissue sampling within trials and the availability of patient samples. “As of right now, there are many studies ongoing to understand the biologic underpinnings to metastasis and drug resistance in oncology. These studies will help translational scientists and clinical research scientists to identify the optimal tumor tissue sample to study in the clinic with respect to patient selection approaches.”

“In the preclinical setting – in developing new cell lines and new animal models, how can we use these modeling approaches to understand compound resistance? When you get to the clinics, do you need to sample in the phase I/II setting? Can you use that old paraffin block that exists everywhere? These are some big questions that we’re struggling with.” On this topic, several sessions at the recent American Society for Clinical Oncology meeting addressed the question of biologic sampling within clinical trials: which sample is most representative of the disease state, the original biopsy in a cancer patient or a recent biopsy of a metastatic lesion?

She is confident that the company has good technology to tackle these kinds of challenges. “The problem is not the technologies; we are able to measure gene expression, protein levels, phosphorylation events in both fresh and paraffin embedded or preserved tissues. It’s a matter of weeding through the various technologic approaches, assessing the predictability and feasibility of each, and building and testing hypotheses in the preclinical setting, because once you get into the clinic it is often too late to discover the biomarkers that accurately predict patient selection for the compound.”

Exploring all avenues

The company is taking ‘diverse approaches’ in the preclinical setting and assessing cell line models, panel xenographs and transgenic models. “We aim to approach the clinic with multiple hypotheses and multiple algorithms for patient selection,” explains Coughlin. “During the first few sets of clinical trials using high content biomarkers, we’ve been using paraffin blocks due to the uncertain outcomes; the initial diagnostic samples hold a tremendous amount of information. With the next few compounds we’ll go back and try and work through the sampling. After this stage, we need to build academic partnerships. This is part of our fundamental strategy.”Coughlin has a few ideas to kick-start the biomarker discovery approach in the early preclinical setting. The first of these is to start early to enable rigorous target validation, even before the compounds are identified. She recalls how with some of the newer targets (now a few years from the clinic), work started very early in order to understand the pathways and to determine what tumors were dependent on that process.

“When a qualified compound is in hand, it’s important to begin to understand the sensitivity and resistance to it,” she says. “We still are not sure what the optimal animal model is to model compound resistance in the preclinical setting. This is critical, because the molecular underpinnings of the tumor that determine whether or not something is resistant can help develop the hypothesis of a biomarker early in the preclinical setting and, as well lead you toward rational combination therapies. We can then use animal models to validate that.”

The need for dialogue

An important dialogue that needs to improve is the interactions and collaborations among the pharmaceutical industry, academia and the regulatory agencies. “At some level, in cancer research, we’re behind where we could be, because we have two groups of very smart people in academia and the pharmaceutical industry who could probably cure cancer, but we can’t always talk openly and collaborate with each other.” Some issues that separate these groups are confidentiality and the issues surrounding intellectual property. “Figuring out how these groups can collaborate is going to be a significant step in this arena. Wyeth supports collaboration with academic investigators and sponsors the Translational Medicine Research Collaboration. These types of collaborative efforts need to continue and be supported from both sides.”

At the moment, existing regulatory systems are not well designed to address the development of high content patient selection biomarkers, such as the complexity of genomic data. According to Coughlin, one way forward is to begin the dialogue regarding the process of biomarker development with the FDA and EMEA, the regulatory agencies, earlier in the clinical development stages. “Applying biomarkers for internal decision-making in the clinic is a very different outcome than biomarkers that might one day be included in the drug label.”

The development of such patient selection biomarkers for compounds will be critical going forward, and early discussions with both diagnostic companies as well as regulatory agencies will be needed for successful implementation in the clinic. “Ultimately, this is a step forward in the treatment of cancer, being able to match a patient’s tumor with the right drug or right combination based on biology as opposed to empiric evidence alone. Our plan at Wyeth is to be at the forefront of implementing novel biomarker approaches to enhance the predictability in the clinic toward the goal of improving the drug discovery process.”

About Christina Coughlin

Christina joined Wyeth Research in 2006 as Associate Director, Discovery Translational Medicine. She leads the early biomarker discover efforts in translation medicine in the oncology pipeline. Prior to joining Wyeth, Christina was attending oncologist at the Children’s Hospital of Philadelphia and research associate in the Abramson Family Cancer Research Institute at the University of Pennsylvania School of Medicine.