Strategies for bringing drug delivery tools into discovery

Abstract

The past decade has yielded a significant body of literature discussing approaches for development and discovery collaboration in the pharmaceutical industry. As a result, collaborations between discovery groups and development scientists have increased considerably. The productivity of pharma companies to deliver new drugs to the market, however, has not increased and development costs continue to rise. Inability to predict clinical and toxicological response underlies the high attrition rate of leads at every step of drug development. A partial solution to this high attrition rate could be provided by better preclinical pharmacokinetics measurements that inform PD response based on key pathways that drive disease progression and therapeutic response. A critical link between these key pharmacology, pharmacokinetics and toxicology studies is the formulation. The challenges in pre-clinical formulation development include limited availability of compounds, rapid turn-around requirements and the frequent un-optimized physical properties of the lead compounds. Despite these challenges, this paper illustrates some successes resulting from close collaboration between formulation scientists and discovery teams. This close collaboration has resulted in development of formulations that meet biopharmaceutical needs from early stage preclinical in vivo model development through toxicity testing and development risk assessment of pre-clinical drug candidates.

Introduction

Early assessment of in vivo efficacy and toxicological assessment of potential drug candidates depends upon effective delivery to the desired therapeutic target. While this usually can be accomplished with simple formulation strategies, sometimes creative technological solutions are needed in order to drive drug absorption to the target in order to answer key questions. This paper shares some examples where formulations have made an impact during discovery with regards to decisions around target validation, efficacy, and safety.

R&D expenditure in the last few years has increased by 80%, while productivity decreased by 43% (Mark Crawford, 2010). This poor return on investment and reduced productivity is the subject of investigation and scrutiny in the Pharma Industry. Despite the increase in sophisticated sales forecasts and market analyses to enhance predictions of success, the rate of producing blockbusters has not improved over the last 20 years (Munos, 2009). It is obvious that there are underlying causes that are either not fully understood or are not being addressed properly. In a review by DiMasi et al. (2003), a major factor for many of these failures was determined to be the high attrition rate in drug development. The taxonomy of the risks showed that efficacy and toxicity are the major cause of attrition (Ismail and John, 2004, Fearn, 2000, FDA, 2004) (Fig. 1). In addition, the pursuit of therapeutic targets that have notoriously unpredictive animal efficacy models (i.e., CNS, oncology) does not favor success in Phase II or III clinical trials (Booth et al., 2003, Roberds et al., 2001).

A survey of these failures and successes calls for measures to rethink the strategy, goals, and efficiency of drug discovery. In older models, medicinal chemistry worked closely with drug metabolism and pharmacokinetics (DMPK) to understand the absorption, distribution, metabolism and excretion (ADME) of the candidate without a thorough understanding of how the compound was delivered in the animals under study. Most of the approaches for delivery focused on solubilization and generally used DMSO as the vehicle. As highlighted in several references addressing formulation support during discovery (Bailey et al., 1996, Railkar et al., 1996, Venkatesh and Lipper, 2000, Chaubal, 2004, Saxena et al., 2009, Maas et al., 2007), use of dimethyl sulfoxide (DMSO) may provide an easy formulation; however, these types of formulations do not reflect those used in development stages and thus provide an inflated absorption outcome. To address this issue, it has been recognized (Bailey et al., 1996, Railkar et al., 1996, Venkatesh and Lipper, 2000, Chaubal, 2004, Saxena et al., 2009, Maas et al., 2007) that the pharmaceutical industry has to evolve such that closer collaboration between functional areas, as illustrated in Fig. 2, is realized. This would lead to co-location of the team members that are dedicated to bringing molecules through discovery into development. This will allow the teams to understand and solve the problems during discovery allowing for disciplined decisions to select quality candidates. It is clear that the team members have to understand the development space to minimize the risks and liabilities of the candidate before progressing into development. In other words, drug discovery would need to be conducted without borders to allow dedicated development scientists to facilitate medicinal chemistry, pharmacology and biology. Efforts will need to be enhanced for delivering not just Phase I compounds but also a safe and effective commercializable drug that is differentiated from existing therapies. Therefore, studies during the discovery phase needs to include: (1) efficacy in animal models using appropriate formulations to ensure exposure and pharmacodynamic (PD) effects, (2) physical and biopharmaceutical properties of active pharmaceutical ingredient (API) that are amenable to downstream development, (3) compounds that meet ADME requirements and (4) are de-risked around toxicological concerns.

To understand the gaps and where there are opportunities to collaborate, we need to dissect the drug discovery process. It basically includes three typical phases (Fig. 3): target validation, lead identification and lead optimization. These phases all have their individual complexities and can have many feedback loops to each other and downstream to decisions early stages of development. Thus, it is important to aim for high compound quality early on because challenges live with the drug development program throughout its lifetime, including later stages of development.

In the target validation phase, one or several targets are studied to determine their role in the mechanism of a given disease state. Recent publication of the FDA “Critical Path” initiative (FDA, 2004) stresses the importance of translational research and the development of tools such as biomarkers and appropriate animal models for efficacy and toxicological testing which may potentially impact the selection of appropriate candidate for clinical development. The challenge for biomarkers is to allow earlier, more robust drug safety and efficacy measurement. Because of the multifaceted nature of optimal in vivo models and/or biomarker development, the research operating plans for the selection and applications of these models should have input from each functional area including formulations, such that a holistic approach can be taken to make the appropriate decisions. At the lead identification stage, discovery teams face a different challenge in using un-optimized leads for developing in vivo efficacy models to determine the specificity and selectivity of the lead. Finally, in lead optimization, a lead candidate emerges via SAR studies which use profiling to inform on efficacy, PK, toxicology and differentiation from comparators (as needed) to arrive at a single or small set of development candidates.

In the overall discovery process, whether these are proof-of-concept pre-clinical studies, biomarker development or mechanism-based toxicity studies, the conclusive evaluation is through an in vivo study. As such, formulation can influence API release rate, the PK profile, oral absorption and hence the PD effect. Hence, involvement of development scientists in this space can have a profound contribution to decision-making to enable the “fail fast/fail cheap” paradigm, reducing costs and downstream resources. Consequently, use of appropriate formulations and/or route of administration can be useful in the development of appropriate efficacy and toxicity models and thus enable studies previously inaccessible using conventional approaches.