Strategies for Early-Stage Drug Development

This article is part of a special issue on API Development, Formulation, Synthesis and Manufacturing.

Pharmaceutical companies face the ongoing challenge of how to best allocate research and development (R&D) resources to make the drug-development process work in the most time efficient and cost-effective way. Industry estimates put the average time to develop a drug at 10–15 years at an average cost of $1.2–1.3 billion (1). As a result, pharmaceutical companies have to consider new ways to make the process work better in an overall economic environment of declining budgets and tighter resource allocation. Such is the case for Bristol-Myers Squibb (New York), which implemented a "fit-for-purpose" model for how the company develops and produces clinical- trial materials for early-stage drug development.

(ELLIOT ELLIOT, GETTY IMAGES)

Resource allocation

The crux of the fit-for-purpose model comes out of a concept of "cost-disciplined science," says Mark Powell, senior vice-president of pharmaceutical development at Bristol-Myers Squibb. "Cost-disciplined science is a development strategy that we are applying throughout R&D," he says. "But in the specific context of drug development, this concept is based on maximizing overall productivity by selectively allocating resources to compounds based on where they are in the drug-development continuum. Among pharmaceutical companies as a whole, we know that there is a high level of attrition of drug candidates as they proceed through the drug-development process. So how do we allocate our resources in the best way to focus on the right drug candidates, and what is the best way to apply our resources so that the most promising candidates have the best chance to succeed?"

The fit-for-purpose model at work

To address that issue, Bristol-Myers Squibb implemented a fit-for-purpose model for early-stage compounds. Under this approach, the company reduces the level of process optimization early in development for both the drug substance and drug product (i.e., formulation). Historically, the company's chemical development department invested significant resources in synthetic process optimization prior to producing early active pharmaceutical ingredients (APIs) batched for clinical studies. In the new fit-for-purpose model, process optimization is generally delayed until a compound proceeds beyond Phase I and approaches or achieves proof of concept. Using this approach, the company liberates resources that can be applied to additional early-stage compounds or can be applied later in development after a compound has been shown to be a viable drug candidate from a clinical perspective. Often the company will elect to outsource the manufacturing of the smaller quantities of the API needed at this early stage (i.e., Phase I) using the discovery-chemistry route that was developed internally. If a molecular target is of particular interest or priority to the company, process optimization can begin internally in parallel to the outsourced early API manufacturing work, but that is reserved for only the highest priority programs, Powell says.

Bristol-Myers Squibb applies a similar philosophy in formulation development by using a de minimis approach to the development of dosage forms that will be used in early clinical testing. Historically, tablet or capsule formulations were developed prior to the start of Phase I clinical trials. Given the wide dosing ranges employed in these early clinical studies, the formulation development work was significant and time-consuming as well as API-consuming. Approximately 70% of all the early API manufactured was used for this formulation-development work. The new fit-for-purpose formulation approach has significantly changed that.

The company predominantly uses a "powder-in-a-bottle" formulation approach in which it puts the API directly in a bottle and makes a simple solution or suspension for drug administration in Phase I clinical studies. Applying this approach requires several additional factors to be taken into consideration for a Phase I clinical trial, says Powell. For example, a pharmacist has to be physically present at the clinical-trial site to make up the final dosing solution. The on-site pharmacist requirement can be an additional cost and also may limit clinical-site locations for a given trial.

By minimizing formulation development, vital resources are freed to support additional early-stage projects, and/or focus on the highest priority projects. Additionally, early clinical projects proceed at a much faster rate, thereby reaching critical decision points earlier. Bristol-Myers Squibb may elect to do more advanced formulation development for a given early-stage compound if it feels that it is necessary, but that decision is made on a case-by-case basis and is reserved for the highest priority programs or compounds.

Metrics

Powell explains that the fit-for-purpose approach involves purposeful trade-offs in resourcing support levels for early- development stage compounds. "The advantage is that by limiting our upfront investment, we are now effectively supporting an early clinical pipeline which has increased more than 50% in the last three years with the same headcount and budget as before," he says. "Additionally, early compounds are proceeding to critical decision points faster then they ever have before. The trade-off is that for those selected compounds that make it successfully past proof of concept, there could be a significant amount of concentrated API and formulation work necessary to progress the programs and support timely progress in late-stage clinical development. But we are looking at our resource allocation across the entire product portfolio and feel these trade-offs are worth it."

Bristol-Myers Squibb has developed specific metrics to quantify the cost and time benefits in this fit-for-purpose model. The company has reduced the amount of API that it produces for early-stage clinical-trial materials by approximately 50%. The company also uses a metric to assess the time it takes between declaration of an exploratory candidate nomination (ECN), the term the company uses to identify an early-stage drug candidate that has moved from the discovery phase to early development, and the start of the first good laboratory practice (GLP) toxicology studies. The rate-limiting aspects of this metric are normally the availability of GLP API. Using that metric of time from ECN to first GLP toxicology studies, Bristol-Myers Squibb has been able to reduce this average time from a historical base of approximately six months to essentially zero time, meaning that GLP toxicology studies can start immediately upon ECN declaration. That is a time savings that directly allows projects to reach critical decision points sooner than ever before and faster than industry benchmarks. A direct result of working more efficiently is that the Pharmaceutical Development group can now support a clinical development pipeline with 64 active compounds with the same number of people that previously supported 40 compounds.

Adapting to change

Powell explains that the fit-for-purpose model was adapted by Bristol-Myers Squibb three years ago and was born out of the need facing many companies of how to do more with less. "In evaluating our strategy in drug development, I assumed that our resource allocation for early drug development would be tied into the company's overall R&D organic growth projections, and we needed to adopt a model that would fit into those constraints," says Powell. "We needed to accept some trade-offs in implementing the model and get comfortable with the idea that not every compound would be supported at the same level early in development."

Although delaying formulation optimization is part of the fit-for-purpose model, the company still deploys new formulation technologies for specific compounds where necessary in early development, notes Powell. Examples are technologies that enhance the bioavailability of poorly water-soluble molecules. As a consequence of the increased number and diversity of new molecules coming from combinatorial chemistry and high-throughput screening in drug discovery, solubility has emerged as an important problem that often has to be solved before dose selection in clinical studies. Solid-based systems, such as coprocessed solid dispersions and liquid-based systems, are both amenable to the model. Coprocessed solid dispersions can be manufactured using spray drying at small scales without extensive optimization and be delivered in either simple powder-in-a-bottle or powder-in-capsule formulations. Lipid/surfactant solutions are amenable to the bottle-based approach as well as hard-gelatin capsule formulations. If the drug candidate proceeds to commercialization, both approaches can be readily extended to larger-scale manufacturing—spray drying or hot-melt extrusion for the solid-based approach, or soft-gelatin capsules for the liquid-based approach. Powell concludes that the selection of the solubilization approach mirrors the fit-for-purpose model for poorly soluble compounds: the choice is based on a balance of cost versus technical considerations that are appropriate to the candidate in early development.

Powell sums up the fit-for-purpose model: "It is all about how to best apply money, people, and technology to reach the key development decision points for every compound as fast and as efficiently as possible. We think this is an overall model that best addresses those issues."

Reference

1. PhRMA, Pharmaceutical Industry Profile (Pharmaceutical Research and Manufacturers of America, Washington, DC, 2010).