Biomarker technologies

NGP talks to three industry experts about the challenges of using biomarkers in drug development. With GlaxoSmithKline, Eurofins Medinet, and NextGen Sciences

What challenges are pharmaceutical companies likely to encounter when using biomarkers in drug development?
Edwin Janssen. While the pharmaceutical R&D has shifted focus towards biomarker-mediated drug development, the regulatory authorities are still hesitant to accept biomarker-based evidence for approval of drugs. The acceptance of a biomarker's clinical utility, therefore, is highly depending on acceptance by the scientific and clinical community to support a claim. Needless to say that the efforts put into qualification of a biomarker can be a cash burn. From a regulatory perspective, the clinical development teams should be aware that the applied biomarker requires rigorous analytical assay validation in order to understand the analytical limitations. Unfortunately, this awareness is sometimes lacking, but can severely obstruct the drug development programme in terms of time and additional costs. In case a global study is conducted, standardisation of the applied biomarker assay becomes critically important. Only the production of consistent biomarker data enables critical decision-making and requires that same procedures are applied among different test facilities.

Michael Pisano. The role of biomarkers spans all aspects of drug discovery and development and the pharmaceutical industry is looking towards biomarkers as one key solution to the drug development problem. When one thinks about molecular markers one thinks about genomic markers and protein markers. While genomic data can be gathered very quickly with lots of data and patient testing, protein markers will more likely be more informative, but it has been said that working with proteins is difficult. One reason proteins are often overlooked as good candidate markers is that a bottleneck exists in protein biomarker development, namely, the ability to develop assays in a timeframe acceptable to move viable candidates forward in line with the therapeutic compound. Another challenge is high quality sample collections, but I will not address this in this forum.

Aiden Flynn. There are a number of issues with biomarkers in drug development. In some cases, biomarkers may have been collected as part of a study that was designed for another purpose. So the analysis relating to the biomarker is exploratory or hypothesis generating. In this case, the study can involve the evaluation of a large number of biomarkers and it can be problematic to efficiently identify biomarkers that are clinically meaningful. Whilst these exploratory analyses are a much needed part of finding new biomarkers, they can be costly, resource intensive, time consuming and yet have a low probability of success.

There is a lack of understanding of the range of potential clinical biomarker applications and also uncertainty around the evaluation of the utility of biomarkers in those applications. In addition, it is not clear how the use of biomarkers will affect the development program, whether there is the need to develop a diagnostic and how drug regulators will view the use of the biomarker. With this added uncertainty, developers may choose not to invest or commit to the co-development of biomarkers. Lastly, there is the issue of poor biomarker data reproducibility, quality and standards. The lack of defined biomarker data standards can prevent the integration of biomarker data with other clinical data and poor quality and reproducibility compounds the difficulties with identifying and implementing biomarkers in clinical studies.

How can these challenges be overcome?
AF. One of the key areas that can help overcome some of these challenges is a clear biomarker strategy to be used in the early planning and integration of biomarkers in the clinical development plan. Where possible, biomarkers should not be an add-on to a study. Rather, study designs should be scrutinised to ensure they provide a reasonable chance of success for biomarker research. In addition, biomarkers should be considered as an integral part of the entire development program. However, we should accept that there may also be scenarios where biomarkers have little or no impact, such as the use of biomarkers to explain variability in drug response when there is no evidence of unexpected variability. We should try to focus our efforts relating to use of biomarkers by using existing knowledge of disease, drug and current biomarker technology to identify realistic opportunities where biomarkers can make a difference. With regards to data reproducibility, quality and standards there is no easy solution other than to say that the entire biomarker research community needs to work together to ensure that biomarker data is fit for purpose.

EJ. A good understanding of how the biomarker assay will be used during drug development is pivotal to allow critical decision-making. This requires good coordination and communication between the translational and clinical development teams to clearly define the purpose of the biomarker assay, early in the drug development process. Proof that a biomarker reflects a clinical benefit will take costly and lengthy clinical trials. Working together in consortia enables to share costs in the development and qualification of biomarkers. A qualified biomarker, eventually, will facilitate drug development across the industry. Good examples of the outcome of such an endeavor are the recent qualification of novel human kidney safety biomarkers, led by the Critical Path Institute, and the demonstration of adiponectin as a predictive biomarker for type 2 diabetes, as demonstrated by the Biomarker Consortium. In the Netherlands, we have the TI-Pharma and CTMM initiative, for example.

MP. The ability to discover putative protein biomarkers is always increasing with improved discovery platforms in all therapeutic areas so there is no issue in finding new putative markers. The issue is having the biomarkers and assays ready for a clinical study. This requires the biomarker program to be an integral part from early discovery and the ability to rapidly develop assays to allow putative biomarkers to be validated in the timeline required. NextGen Sciences' mass spectrometry-based platform significantly shortens the assay development time relative to more conventional platforms such as immunoassays. Biomarker projects are supported by NextGen Sciences' proprietary biomarkerlibrary, which contains catalogues of proteins from a variety of biological fluids, tissues and cell lines from pre-clinical models to human. This knowledge accelerates the development of robust, accurate and precise multiplexed assays suitable during all phases of clinical development.

What specific methods and tools can companies use to increase the successful application of biomarkers in the evaluation of drug therapies?
EJ. Within our laboratory we do see an increased demand for analysis of cellular drug responses. Although challenging in terms of analytical and biological variations, cellular drug responses are key in providing early proof of concept. PBMCs, or whole blood assays, can be good surrogate sources to assess in vivo target exposure in patients, establish proof of concept, and ensure optimal clinical development. Proteomics and biomarker imaging technologies are important in drug development as well. To what extent proteomics and imaging tools will be successfully integrated in early or late stage drug development programs remains to be seen. In terms of sensitivity, reproducibility, throughput and cost-efficiency, ligand-binding assay formats do still find their greatest utility in monitoring drug safety and efficacy, particularly during late stage development. Promising approaches also include the use of miRNA-based biomarkers, which embody novel diagnostic and prognostic biomarkers. Analysis of disease-related miRNAs from the blood may represent a promising approach for monitoring of disease progression and drug efficacy.

AF. It is hard to specify methods or tools that can generally improve the successful use of biomarkers across all the potential clinical applications. One approach is the use of modeling, simulation and decision trees to estimate the probability of success for the range of development options and study designs where biomarkers are being considered. This approach helps project teams focus on what they are trying to achieve and to understand the options and implications. It is also important to have access to a biomarker knowledge base that includes disease and drug specific information, scientific and regulatory developments in biomarker research and case studies describing biomarker applications. Such a tool would be a very valuable resource.

MP. There is no specific method or tool that will increase the successful application. Instead it is the application of a systematic approach. In other words, two things need to happen; first, as new technologies emerge there is often the ability to detect what could not be detected previously and second is to look at signatures whether that be multiple proteins or combinations of proteins, genes and metabolites or images with protein signatures. As complex as biology is it should not be expected that single entities will paint a complete picture.

How do you see the use of biomarkers in the pharmaceutical industry developing over the next five to 10 years?
MP. The pharmaceutical industry will integrate biomarkers to help understand the mechanism of a disease, to aid the development of therapeutics, as diagnostics and in personalised medicine. The integration of biomarkers through the different phases of drug development can yield safer drugs with enhanced therapeutic efficacy in a cost-effective manner. Pre-clinical and clinical applications of biomarkers include safety, patient selection in a drug trial, measuring drug efficacy, monitoring response and adverse events and studying alternative indications for a drug in development. Biomarkers are critical in helping to reduce costs in the drug discovery and development process. Biomarkers can be used as a way to rescue drugs that have failed in development and even those that have been withdrawn from the market.

EJ. Because of the ongoing advances in biomarker research and qualification processes, many more biomarkers are likely to become accepted. Accordingly, these biomarker assays need to be validated for their application in clinical trials. The recent qualification of novel kidney safety biomarkers already paves the way for central laboratories to encourage sponsors and investigators to analyse these injury response biomarkers to demonstrate drug safety. The recent validation of circulating tumour cell measurements has also shown its suitability for routine assessment of drug efficacy in metastatic (breast) cancer patients. Epigenetic factors, such as histone deacetylation or promotor region methylation can be expected to become important in oncology, and be used on a more routine basis as well.

Central laboratories are likely to become outsourcing partners for successfully integrating biomarker assays in drug development programs in the years to come. The support by the central laboratory of many clinical studies with a broad variety of biomarker applications will become valuable in advising sponsors on the type of biomarkers and the analytical limitations of these assays.

AF. A lot of biomarker research in recent years has been driven by the ability of biomarker platforms to generate increasingly large volumes of biomarker data. However, we are still developing the capabilities to perform the best analysis of these data. These analysis methods need to go beyond the traditional statistical approaches used in clinical trials. They need to account for the biological relationships in the biomarker data as well as the statistical associations with the clinical data. Over the next few years statisticians, bioinformaticians, computer scientists and others will be working together more to develop methods for use in clinical biomarker research. These methods will open up many opportunities such the use of a combination of biomarkers, clinical information and environmental variables for predictive modeling. When we develop these tools, we're going to see a more integrated, informed and pragmatic approach to the use of biomarkers.

Aiden Flynn has spent the past eight years at GlaxoSmithKline where he is a Director of the group providing statistical support for clinical biomarker studies. Prior to joining GSK, he worked as a lecturer in the Department of Oncology at University College London, where he obtained his PhD in Biophysics.

Edwin Janssen (PhD) joined Schering-Plough (former Organon), following several years of postdoctoral experiences. At Schering-Plough, he participated in the discovery, development and implementation of biomarkers in lead optimisation programs and proof of concept studies. Since 2009, Edwin has been the Head of the Biomarker and Biopharmaceutical Group of Eurofins Medinet.

Michael Pisano is President and CEO of NextGen Sciences, Inc. He has more than 18 years in the pharmaceutical industry with senior positions held in RPR Pharmaceuticals, Aventis Pharmaceuticals and Genomic Solutions. Pisano was co-founder and President of Proteomic Research Services, Inc. and is now Chief Scientific Officer of NextGen Group and President and CEO of NextGen Sciences, Inc.