The clock on patent expiry starts ticking long before the regulatory authorities grant marketing approval. Therefore, any strategy that can accelerate a drug development pathway to market is likely to improve its commercial impact … should it be approved. However, the identification of a suitable formulation for orally delivered small molecule drugs is becoming ever more challenging. Roughly two thirds of all oral new chemical entities in the development pipeline can be classified as poorly soluble, with permeability issues also affecting a fifth of drugs. As a result, absorption may be both insufficient and inconsistent when such drugs are administered along with or without food.
The traditional method for categorising molecules according to their solubility and permeability properties is the Biopharmaceutics Classification System (BCS). This categorises a drug into one of four classes and represents a good, broad guideline for identifying a successful formulation; however, it is not definitive for those drugs that fall into Class II, as these have problems with solubility, for which the solutions are rarely straightforward. The classification system therefore has a limit to its usefulness when formulating drugs in this quadrant.
More recently, an alternative, the Developability Classification System (DCS), has been introduced. This is more effective at helping to identify formulation alternatives by splitting the problematic Class II quadrant into two subgroups. If the drug has a limited dissolution rate, it will fall into Class IIa; whereas if the issue is limited solubility, then it will be defined as Class IIb. The boundary between the two is represented by the solubility limited absorbable dose (SLAD), which is the dose above which absorption is limited by solubility. Above the SLAD, merely making smaller particles to speed up dissolution will not suffice.
DCS evaluation uses both a higher volume of gastrointestinal (GI) fluid — 500 mL instead of 250 mL — as well as fasting state intestinal fluid instead of buffer. It also considers the fact that permeation from the GI tract will allow more of the drug to dissolve to replace the active pharmaceutical ingredient (API) that was absorbed.
Creating the optimal dosage form requires experience and expertise, and organisations that have significant experience in overcoming solubility and bioavailability challenges can clearly play a role in assisting companies, from Big Pharma to small biotech, in formulation development. Speed and cost are both very important; but, in many cases, the supply of API is limited in these early stages of development, so careful use of available supplies is vital.
Accelerated screening
Parallel screening technology, such as Catalent’s OptiForm Solution Suite platform, can be an effective way to speed up the process. With its foundation in the DCS, it allows potential formulations to be screened rapidly by looking at several different proven industry technologies simultaneously. Although the protocol is aimed at candidate molecules that are at the start of formulation development, there is the potential to customise the process if some of the information is already available. First, data about the molecule are collected in the “assess” phase; then, the “enhance” stage looks to find ways to improve the solvated form and evaluate the potential of existing formulation strategies; finally, the “deliver” phase will provide up to four candidate formulations, for which 2-week stability studies have been performed. These suggested formulations are usually available within 12 weeks, with evaluations having been made on the basis of hard data, not subjective opinions.
Case study: Trio Medicines
The OptiForm platform was successfully applied to a product being developed by Trio Medicines. The starting point in the drug development programme, coded TML-001, had already undergone clinical evaluation, and the authorities had already granted a licence for a first-in-human (FIH) study with no further toxicology work on the basis of those earlier trials.
This study indicated that its bioavailability was poor; therefore, an acetylated prodrug, TML-002, was created. An FIH trial indicated that although the bioavailability was better than the parent compound, it was still unacceptably low. The dosage form that was used in that FIH study consisted of the unmodified API, enclosed within a capsule. The drug fell into that problematic BCS Class II, and data showed that it would be in DCS Class IIa, wherein dissolution rate is the problem rather than intrinsic solubility.
The first step was to collect information about its compatibility with a range of excipients. It was screened across a standard set of excipients to identify any broad incompatibilities with major functional groups. This carefully selected group of excipients is deliberately chosen to reduce screening time and cost. Using an even wider selection of excipients might give more granular data; however, the cost would rapidly become prohibitive. Using the same panel of excipients for each project does allow an early assessment to be achieved by comparison with previous projects for which the same dataset has been collated. Powder X-ray diffraction experiments showed that TML-002 was in an easy-to-handle crystalline form: it was not hygroscopic or micronised, and had a melting point in the normal range. Other data on particle size and solid state were collected using a range of analytical techniques, including optical microscopy, dynamic vapour sorption, thermogravimetric analysis and differential scanning calorimetry.
Solubility data were amassed, too, and showed a pH-dependent solubility profile that was not unexpected (given its chemical structure). It was, however, shown to be a little unstable in lipid solution. The solubility in fasted state intestinal fluid was comparable with that in a standard buffer at pH 6.5, which is predictive of the presence of a food effect. This confirmed that the drug fell into Class IIa of the DCS for a 100 mg dose, as the SLAD was calculated at 367 mg; but, for the lower 50 mg dose, the drug actually fell into the far less problematic Class I.
This result implies that permeability may be the problem, not solubility. However, something is missing in this assumption — a realisation that permeability calculations do not include important transporter systems that may be having an effect, such as efflux. It may well be that these transporter systems lie at the root of the bioavailability problems, and intrinsic solubility should not be an issue.
Data in both the solid and solution states were collected during a period of 2 weeks at 40 °C. TML-002 proved somewhat unstable, particularly when exposed to light in the solid state, as well as in acid conditions. However, the results did not cause concern regarding in vivo stability. Furthermore, data collected at high humidity showed no changes in crystalline habit.
Formulation selection
The data collected in this first phase of the process were then used to evaluate the different formulation technologies that might be applied. There were three main alternatives: particle size reduction, lipid-based formulation and an amorphous solid dispersion created via hot melt extrusion technology. Pharmaceutically acceptable excipients were used below the maximum daily intake level (when appropriate). Excipient selection is a critical part of product development, not just in terms of solubilisation, but in ensuring the reliability of supply and stability. A number of potential formulations were made, with drug loadings ranging from 2.5–10%.
For the particle size reduction, a simple air jet mill could be used to micronise the API, and still smaller particles could be created through cryomilling. The API can also be comilled alongside surfactants or other suitable excipients to improve its physical properties, whether this is to improve powder flow, solubility or reduce particle size still further. With the lipid solution, adequate solubility was achieved in four different excipients, allowing multiple candidate formulations to be made. To evaluate solid dispersions using hot melt extrusion technology, the excipients and API were melted together and the resulting mixtures reviewed for evidence of crystallinity.
Two solid dispersion candidate formulations were selected and made using a twin-screw extruder, but there was some difficulty in processing the samples and, as such, it was felt that this was not the preferred process for solid dispersion manufacture. In further testing, a small amount of decomposition was seen with the lipid formulations, which may mean that containing it within an oxidation-protective softgel capsule would be preferable. Although the stability was good with the solid dispersions, there was greater degradation, which may have been a result of thermal action during its manufacture. The most stable was the micronised API.
In all, four candidates were presented to Trio for further evaluation at the end of the screening protocol. Two were lipids, one was a solid dispersion powder and the fourth was the result of particle size reduction. A risk ranking table was created to guide selection. This indicated that the riskiest option to choose would be the solid dispersion, particularly in light of its challenging processability. At the other end of the risk scale was the micronised API, which was the simplest to manufacture, with good chemical and physical stability, and a good fit for the DCS criteria.
The solid dispersion was dismissed as an option after a crossover study in healthy humans showed that it offered no improvement compared with the original formulation. In contrast, the other three all performed much better. These represent a good starting point, and it may be possible that small adjustments and developments may offer further improvements to bioavailability.
It is self-evident that an optimised formulation that offers good bioavailability will maximise the chances of success in clinical trials. With challenging APIs, there is no substitute for a careful study of the available options, and beginning this work as early in the development cycle as possible will prevent the planned start dates of these trials being delayed.
