Advances in Drug Discovery

GE Healthcare’s Kevin Daish and Joseph P. Brown of LifeSpan BioSciences talk latest developments.

Kevin Daish is the Global Marketing Director for Gene Expression at GE Healthcare. As such, he is responsible for the marketing of the CodeLink Microarray Gene Expression system. Kevin has been involved in microarrays for many years including companies such as Expression BioSystems and MWG Ltd where he held various positions ranging from product management through to Managing Director.

Joseph P. Brown, PhD is co-founder, President, CEO and Chairman of LifeSpan BioSciences, having previously held executive roles at Pathogenesis Corp., Bristol-Myers Squibb and Genetic Systems Corp. From 1974-1983 he was at the Fred Hutchinson Cancer Research Center, where he made the first monoclonal antibodies to cancer. Dr Brown has extensive software development experience in addition to his background in immunology and biochemistry, and is an expert in data mining algorithms and their applications to genomics and proteomics.

NGP. Molecular pathology is predicted to be a fast-growing area in major clinical laboratories as new technology enables access to gene-specific information. What kind of impact will this have on early drug discovery?

KD. One area of molecular pathology that is seeing an increase in growth is array Comparative Genomic Hybridisation – aCGH. The advantages of this method over conventional FISH or CGH is that it is quicker and can detect single copy number changes in DNA. In addition, you can look for many abnormalities in one experiment. In early drug discovery, its use will enable potential drug candidates to pass through to full development or be with drawn quicker if it is known there will be potential problems. This way development time could be made shorter with less problematic candidates.

NGP. “Drug discovery today is data-rich but information-poor.” To what extent do you agree with this assessment? What technologies are evolving to ensure those in the discovery process have the best information available to them?

JB. Going from data to wisdom is the real issue. There’s a huge amount of data out there – hundreds of databases related to human genome project and associated areas – and so there has been a great need for tools to not only generate data but to organize it and cross reference the data, and of course to cross-reference different types of data.

We’ve had experience generating and providing molecular pathology data to our clients, but also they want that linked to DNA sequences, protein sequences, scientific publications, patent applications and so on, and we’ve done a lot of that and provided it in the form of subscription databases. A lot of that work turns out to have been done manually. It’s PhD-level curators sitting down, reviewing the data, reading the abstracts, extracting the key information and creating the content.

Clearly, the core technology is the ability to access data over the internet. There’s also visualization technologies that allow you to view ‘heat maps’ or graphical displays of data, but a lot of it involves adding ‘human content’ to link the data together. In this way, the database contains not only a nice view of what’s in GenBank, but actual editorial and curatorial input to provide a higher quality product.

KD. Technologies today generate a lot of data but it needs identification and analysis to put any meaning to the results. What is needed is software that gives you the statistically relevant information with a easy to use interface. In addition, it has to be able to interrogate results from a number of platforms and correlate them to give researchers the meaningful results they are looking for.

NGP. High-throughput technologies such as microarray and imaging technologies are having a huge impact on the drug discovery process. What advances are these technologies enabling in terms of drug discovery?

JB. You mention two types of technology here. Microarrays (or gene chips) are allowing people to look at tens of thousands of gene or other DNA sequences (and now sometimes protein sequences) in parallel. Imaging technologies are used for this too, but the more interesting applications of imaging technologies are looking directly at biological specimens – in some cases for high-throughput cellular analysis, and in other cases (which is what we specialize in) for looking at tissues. We look at tissue sections through a microscope and analyze the structure by looking at the morphology, or by looking at the patterns of staining with various antibodies.

Together, those approaches are first of all allowing scientists to look at tens of thousands of genes at once, and second, to look at many more specimens at a time. We also look at what we call tissue microarrays, which can hold up to several thousand specimens per slide (although they typically range from tens to hundreds per slide). You can generate a vast amount of data – hundreds of genes, thousands of antibodies, thousands of tissue specimens – which creates a great need for automated analysis, and for automated generation of viewing of databases.

NGP. Target discovery and validation is crucial before entering a costly and time-consuming drug development track. How can better validation enable drug companies improve the efficiency and profitability of their drug discovery and development programs?

KD. Target discovery and validation involves proving that DNA, RNA or a protein molecule is directly involved in a disease process and can be used as a suitable target for development of a new therapeutic drug. It is therefore essential to prove that such systems are implicated in the disease process so that the drug candidates are truly evaluated for their worth and are focused on that particular disease. This dramatically reduces the cost for the discovery process and also the number of failures.

JB. Since the human genome project has been essentially completed, every drug company has access to every single one of the 25,000-plus human genes. These are all now potential candidates – in other words, there’s no such thing any more as gene discovery, they’re already all out there in the database. The question then becomes entirely one of validation. Of the many possible leads that you will find associated with any disease, how do you narrow down the number that you can reasonably take forward into drug discovery and then into development? Companies are asking many more questions earlier on in the process – about the nature of the protein produced by that gene, its localization within cells, its localization within tissues, how its associated with disease, how specific that association is, and so on. This means that when they do have a compound and test it in animals (and ultimately in humans), the attrition rate is lower later on in the process – by which time it’s gotten very expensive.

Obviously, a failure in phase III clinical trials or a failure after market release can be incredibly expensive when compared to dropping a compound early on. A lot of the work we do is to enable companies to make that cut earlier on and move ahead with fewer targets but with greater confidence.

NGP. What is the importance of biomarker identification?

JB. There are two major types of biomarkers, as I understand it. The first is proteins or other molecules that are typically found in blood or other biological fluids such as tears, saliva and urine. These can be used to monitor what’s going on throughout the whole body. The other class of biomarkers are those associated with tissues that you would look at by taking tissue specimens on a slide and staining them with an antibody.

We focus more on the second type, and these are already proving very important – both in drug development and also in the use of drugs – for selecting, on the one hand, patient populations for whom its appropriate to treat with the drug; or if you are a patient, for selecting the most appropriate drug for your treatment. This is very much the case in cancer now, where there are new drugs being released to the market every year. These tend to be more specific, molecular-targeted drugs than in the past, and demand a more specific diagnostic procedure to select patients whom it is appropriate to treat.

According to some experts, gene-expression analysis represents the next frontier and the next major area of opportunity. What technologies are coming together to provide the tools to answer the tough questions? What benefits will this bring?

KD. Microarray technology will have a large impact on the future of the drug discovery process. Many companies are investing in databases of expression data with new drug candidates at varying concentrations, different tissues, and linked phenotype. They can then use this data to predict which future candidates may cause a subsequent problem. It will enable drug candidates to progress quicker through to development or be halted if they are shown to be a potential problem at the gene expression level. This ultimately saves time and development costs.

JB. I would argue that gene expression analysis has been important for the last decade; it’s why we founded the company. We believed that with the advent of the Human Genome Project there would be many more genes to study in the drug development process. One of the tools that is very important is antibodies, and one of our goals (and a more general goal of the industry) is to have antibodies available to every protein encoded by the human genome, and then you’re in a position to develop specific immuno-assays to look at those proteins as markers to help in the drug discovery process or in the clinical use of drugs. There are other gene expression technologies, but I think antibodies are proving to be a very useful tool of longstanding value because of their great specificity.