Biomarkers Roundtable

European Pharmaceutical Review invited three individuals to discuss current ideas and issues surrounding biomarkers and their possibilities.

How do you feel the subject of Biomarkers is currently viewed within the majority of the Pharmaceutical Industry?

Wing: Biomarkers have been used for decades by biomedical scientists to make decisions during drug development and so it is not surprising that the majority viewpoint is that biomarker analysis is a useful if not critical tool for the development of new medicines. The elevated prominence given more recently to biomarkers by the regulatory authorities, the Critical path Institute, academia and pharma has raised industry’s expectation though some have questioned whether “biomarkers” are just being re-packaged by changing the vocabulary, though the growing list of drugs whose success can be attributed to biomarker analysis is counteracting this sentiment.

Uttenthal: My own company is devoted to diagnostic innovation by means of immunochemical assays for new biomarkers, so I cannot pretend to speak for the pharmaceutical industry. However, recent trends indicate that the concept of “biomarkers” is acquiring a life of its own, so that all sorts of specialised meanings are being applied to this very broad term. It is becoming one of today’s buzz words, with the danger that its basic meaning may be forgotten.

The FDA (Guidance for Industry: Pharmacogenomic Data Submissions) cites the definition of a biomarker as “a characteristic that is objectively measured and evaluated as an indicator of normal biologic processes, pathogenic processes, or pharmacologic responses to a therapeutic intervention.” I believe this definition to be very serviceable. It follows that the whole of western medicine for the last few centuries (and going back to Hippocrates) has been based on biomarkers, and all diagnostic procedures and the whole of the diagnostic industry is, and always has been, entirely based on biomarkers.

What is now happening is that the use of the term is being narrowed down to certain specific applications and classes of biomarker molecule. Now that cardiac markers have become established, the focus in the diagnostic area is shifting to damage markers for other vital organs, such as the kidney. At the pharmaceutical interface with the clinical world, there is a great deal of interest in biomarkers of pharmacogenomics and certain types or subtypes of cancer. The aim is towards classifying patients according to genetically determined metabolism of a particular drug, the early diagnosis of cancers and the response of subtypes of cancer to particular, usually expensive, new treatments. Within the pharmaceutical industry itself, attention is centering on biomarkers of safety and efficacy as facilitators of the drug development process. These will often be relevant to laboratory animals and may or may not translate into useful clinical biomarkers. With these different focuses in the three different areas mentioned, there is a danger that misunderstandings will arise if spokesmen for these areas were to be understood to imply that their particular approach to biomarkers is paradigmatic for the biomarker field as a whole.

Patterson: Biomarkers have always been used in drug development but for many years predominantly in late stage development. The trend now is to employ them earlier in the pipeline. This is possible because pathways interdicted by targeted therapies can in some cases be directly measured.

What currently available technologies do you feel are most effective in delivering results when conducting research in this field?

Wing: Biomarkers are so diverse that the chance of success is maximised by having access to a full range of technologies, which may differ during the life cycle of a given marker. For instance genomic and proteomic based platforms may be appropriate for biomarker discovery i.e. large microarrays, but once candidate biomarkers or pathways have been identified, multiplexing technologies such as luminex may be more suitable to evaluate a more manageable number of exploratory biomarkers to maximise the information gained from the limited samples that are likely to be available at the early phase of drug development. Access to pathway analysis software for biomarker selection and analysis will also be a valuable tool at this early stage. As the value of the biomarker(s) is confirmed through the drug development process then “technical” validation of the assay becomes more important and so platforms amenable to robust validation, high throughput and automation are preferred.

Uttenthal: The conventional answers tend to emphasise genomic and proteomic approaches, DNA or antibody microarrays and so on, but I am skeptical about this in relation to diagnostic markers. It should be remembered that I am speaking of clinically relevant diagnostic markers, not the other types of marker mentioned. Biomarkers coming into actual diagnostic use have been developed and validated by classical techniques, and I think that this will continue to be pattern for quite a few years to come. Genomics and proteomics will play their part in the discovery process for new biomarkers, but it is amazing how many biomarkers turn out to have been missed when such results are analysed retrospectively. One needs to know what one is looking for before starting, and this may be too narrowly or incorrectly defined. Also, the phenotype does not always reflect the genotype very precisely, and very important quantitative aspects can get left out of these screening approaches. So I tend to back good old-fashioned detailed knowledge of pathophysiology and cell biology as being most likely to lead to the identification of the most valuable diagnostic biomarkers for quite some time.

Patterson: There are a range of technologies that one must have access to if the biomarker strategy is to interrogate the impact of your targeted therapeutic on specific pathways. The challenge is that in some cases, research-based technologies are being introduced into the clinical setting and as such the preanalytical as well as analytical performance characteristics of the assay have to be well understood so that robust data can be generated upon which decisions on the program can be made.

With further research into the subject of Biomarkers how likely is it that the concept of “Personalised medicine” will become a reality?

Wing: There is widespread acceptance that the blockbuster model of drug development is unsustainable since around 50% of patients treated with a typical drug receive no benefit or worse an adverse drug reaction. Personalised medicine is a reality for a growing list of drugs, delivering tangible results for patients, the drug companies and health care providers. This is best exemplified by the reversal of regulatory decision not to approve Pantumumab in Europe until patient stratification data was presented to the regulators which distinguished responders on the basis of their wild-type KRAS genotype.

Uttenthal: “Personalised medicine” has become another buzz word with a specialised meaning in the pharmaceutical world, but in its basic meaning it defines one of the essential features of western medical practice. Personalised medicine has always been the aim and should already be the reality. But what the question refers to is the specific aid to the physicians that some biomarkers will provide, to reveal in advance certain specific tumor cell vulnerabilities, specific patterns of drug metabolism and specific disease susceptibilities in certain patients. This is already a reality with respect to a few isolated tests in these areas, and the number such tests should grow and will grow, to aid the physician’s desire to give the best possible treatment to the individual patient.

Patterson: That depends upon ones definition of ‘Personalized Medicine’. In its strictest sense it seems to imply a very large number of available therapies from which one can choose based upon objective criteria (biomarkers). However, the reality is that biomarkers may stratify the population into likely responders or non-responders (i.e., two strata). This is beginning to happen and I think will continue to come into play as we gain deeper understanding of the disease processes and how the pharmacologic agents impact that.

How far away do you think that the industry is currently from realising this potential?

Wing: Whilst most genetic biomarkers used to personalise medicines have been identified in late clinical development or even post marketing, these examples will stimulate the co-development of companion diagnostic biomarkers with the drug to enable the selection of responsive patients on to the trial, thereby reducing the size and cost of the trial and bringing the drug to market faster. Clearly co-development of such a diagnostic will depend on a thorough understanding of the human disease at the molecular basis, perhaps assisted by comparing the genetic differences of non-responders and responders identified in previous trials. It is anticipated that major breakthroughs will happen in the next 5-10 years for many diseases however the complexity of human diseases makes it likely that multiple genetic biomarkers will need to be identified for some drugs and that this complexity may be the rate limiting factor for the full potential of personalised medicine to be realized.

Uttenthal: Quite far, actually. The limits are imposed by the rates of discovery, development and validation, by the cost of the tests, the market penetration of the analytical platforms, and all the usual factors that govern the acceptance of new diagnostic, prognostic or theranostic markers. Public health and even ethical issues will arise in some cases, which will act as brakes to implementation. Ethical issues will arise particularly in the case of markers of disease susceptibilities

Patterson: There are only a few examples now, but I believe those will increase over the next few years to the stage (5-10 years from now) where in oncology, for example, it will be likely that all drugs will come with a companion diagnostic.

What future changes do you think that the industry will need to make in order to fully realise the potential which Biomarker research can unlock?

Wing: Currently pharma does not typically develop diagnostic biomarkers to identify patients that should either be enrolled onto a clinical trial or post marketing should receive/avoid the drug. Instead this process is typically under taken by niche service providers during late clinical development or post marketing. To encourage the earlier development of a companion diagnostics pharma will either have to undertake this activity themselves or outsource this work to service providers. The development of a companion diagnostic at such an early phase will be a risky venture if undertaken by the niche provider at their expense though the potential reward of owning a blockbuster diagnostic kit that health care providers will use to test all patients likely to receive the drug may encourage a shared risk and reward financial model.

Uttenthal: When considering the future, my mind turns to the old television series “Star Trek”, which was unusual in presenting a vision of future medical practice. This vision included the almost instantaneous diagnosis of any condition by computerised analysis of what must have been fast-reacting multiplex tests for no doubt a huge number of biomarkers. That vision may well be the ultimate goal of the current biomarker drive. However, there is a long way to go and a number of impediments to progress. Among these is also the very significant one of coordination of efforts within the pharmaceutical and diagnostic industries. A few years ago it was quite fashionable to talk about synergy between therapeutics and diagnostics, whereby a new diagnostic test would provide the indication for a new, expensive treatment, or new therapies would require a new type of monitoring. There were a few large pharmaceutical companies which, for historical reasons, also had well-developed diagnostics divisions, and it was expected that these companies would be well placed to exploit this synergy. In fact, the trend has been quite the reverse. These companies have mostly sold off the diagnostics side. One reason is perhaps the very different business models that apply to the two sides: high risk, high investment and high profit in pharmaceuticals, much lower risks, investments and profits in diagnostics. So there are potent business factors to impede the realisation of the synergic effect, which is no longer much talked about. However, one change the pharmaceutical industry will have to make is to consider (seriously, this time) involving the biomarker boys of the diagnostic side at a very early stage. Coordination has been much desired in theory, but impossible in practice. This has resulted in potentially valuable therapeutic products being launched, but failing because of the lack of a good test to select the patients that are most likely to respond to treatment.Activated protein C for sepsis treatment may be a case in point. Equally, interesting biomarker tests have been available for years in other fields, while development of the appropriate treatment has lagged far behind. Here I am thinking, for example, of mannan-binding lectin, a molecule of the innate immune system for which replacement therapy is only now under way.

Patterson: Drug development is a global process and for the value of biomarker research applied to personalized medicine to be realized there will need to be global harmonization of how regulatory authorities handle companion diagnostic data. I am confident the science will move forward at a rapid pace, but how this data can be assimilated into the drug development process, including of course regulatory review, will be key to the success of the use of biomarkers for personalized medicine. As for biomarkers being applied into the earliest stages of drug development, I think that trend will continue.

Is there anything else that you would like to add which you think would be relevant to our readers regarding this area of study and the possibilities it holds for the Pharmaceutical Industry?

Wing: More pre competitive collaborations between pharma and the regulators to qualify better safety biomarkers to reduce the rate of drug attrition due to off target organ toxicity. Such a qualification process is on-going with North American pharma who have formed working groups with the Critical Path Institute with a view to submitting data to the regulatory authorities (FDA and EMEA). This has recently resulted in the qualification of seven rat urine biomarkers suitable to detect drug-induced nephrotoxicity. These collaborations might want to consider using proprietary compounds which failed in late clinical development having raised no safety concerns during the preclinical phases of development and only consider markers where there is good evidence of their translation in man, ie more emphasis on reverse translation from man to the bench.

Research and investment into better PK/PD software models to assist with the integration of in vitro man versus toxicology species potency data plus in vivo exposure/PD data to asset with FiH studies where there is uncertainty over the relevance of the preclinical safety package due for instance to significant differences in the relative potency of the drug with man, ie Guideline on strategies to identify and mitigate risks for first in human clinical trials (EMEA 2007) EMEA/CHMP/SWP/28367/07 www.emea.europa.eu/pdfs/human/swp/ 2836707enfin.pdf

Patterson: Continued advances in technology will drive the further use of biomarkers in drug development. These are coming at a very rapid pace however, and as mentioned earlier, the key is successful integration of these technologies into the clinical setting. This requires a good understanding of both the preanalytical and analytical performance characteristics of the assay – it may also require more specialized sampling of tissue from patients.