In response to the realization that many products have failed due to formulation issues, pharmaceutical companies are putting much more emphasis on achieving optimal product formulations earlier in the development process. Using traditional manual experimental methods there is often only time to evaluate a limited number of formulations, sometimes throwing a drug development project into crisis mode.
To respond to this need, Symyx Technologies (Santa Clara, CA) has introduced the CORE(x) system of integrated Renaissance software, robotics, and analytical tools to enable a number of pharmaceutical sciences core workflows with high throughput, automated testing of drug candidates for formulation and preformulation solubility and stability assessment.
This integrated workflow increases formulation discovery productivity by 100x by making it possible to perform more experiments in the same amount of time required by traditional methods. A wide range of experiments can be run to gain an understanding of the product's solution behavior. An estimated annual throughput of 60,000 experiments or more can be achieved using this workflow. This translates to over 25 compounds per year that can be thoroughly evaluated.
Within preclinical development, the emphasis is on fully characterizing the new chemical entities coming out of lead optimization and turning those "leads" into "drugs". In general, development scientists identify crystallizable forms with salt selection and polymorph studies, determine solubility profiles, partition coefficients, pKa, determine the stability of the compounds and identify what decomposition products are formed, and perform initial formulation studies for excipient compatibility in order to improve solubility and stability.
Solubility profiles are important to understand the bioavailability of the drug as well as to develop the appropriate processing conditions for scale up and manufacturing. Scientists are interested in the solubility as a function of solvent composition, pH, and temperature. To develop a pharmaceutically acceptable drug, chemical and physical stabilities must be explored and addressed.
A nonoptimal formulation is often sufficient during the early development process but can often create major problems as clinical trials ensue. The earlier the active pharmaceutical is stabilized with an effective formulation, the more quickly it can move through the drug pipeline to the clinic. Many products have failed or been delayed due to formulation issues such as poor stability, loss of activity, and inability to meet specifications.
As the drug development pipeline increases, dosage form demands grow, and time pressures increase, the demand for additional preformulation and formulation resources makes it essential to automate the experimental process.
A workflow that generates and analyzes a large volume of data provides the level of understanding that makes it possible to move beyond intuition and guesswork in overcoming formulation problems. An integrated program of experiments can provide understanding of the product and its behavior in various environments and with various excipients.
Accelerating Speed of Pharmaceutical Development Experiments
Modular tools have been developed to address these challenges by accelerating the speed with which key preformulation and formulation experiments can be performed and by collecting these data into a common database that can be used by scientists to build comprehensive reports.
This workflow measures solubility in arrays of user-defined solvents, viscous liquids, and solvent mixtures, at ambient and higher temperatures, partitioning of the species partitioning as a function of pH (log D), partitioning between water and hydrophobic solvents (log P), and solubility in various pH buffers or bodily fluid mimic solutions.
For example, it is often desirable to have a complete profile of solubility vs. pH at ambient and physiological temperatures. The solubility of the formulation at different pHs affects how the drug will be absorbed into the body.
Software can be used to design tens to thousands of experiments. After the libraries have been designed, the formulations are rapidly prepared and visual, chromatographic, or spectroscopic tools are used to monitor the stability of the library of samples.
The system also provides a platform for determining the protonization of a molecule through measurement of pKa. Stability measurements can be performed by starting with a known concentration of an array of proposed formulations and measuring its concentration as well as the appearance of decomposition products over time.
The first step in these workflows is to use Renaissance Library Studio software to design library arrays for synthesis and screening. The software provides an environment for creating, visualizing, and planning sets of experiments and provides capabilities to map out arbitrary x-y matrices, including ternary and higher phase diagrams.
Using Library Studio, users retrieve chemicals and reagents from an archived database, electronically create stock solutions, and build libraries using a drag and drop interface.
The libraries are then stored in a database along with the recipe information for access by the Renaissance Impressionist software used to create and process the library in the laboratory. The user specifies the robotic programs and routines using a common set of drop-down actions and subroutines.
In a typical protocol, solubility is measured versus solvent composition and temperature with 12 solvent compositions and 8 replicates of each composition. The protocol begins with adding the stir bars to the plate; then the powder dispenser adds drug to each well according to the Library Studio recipe.
Subsequently, the protocol dispenses solvents according to the design, and the entire library is then stirred and equilibrated at temperature using parameters specified in the experimental design.
The protocol then filters the supernatant from sealed source vessel to sealed vessel at a controlled temperature and performs serial dilutions (1:10, 1:100, and 1:1000) to create a series of dilution daughter plates. Using Renaissance Epoch software the 1:100 plate is then run through the HPLC and the chromatograms and concentration information are stored into the database.
Based on the peak signals from each well of this initial plate, Epoch determines whether to run selected wells. The final step generates plots and reports using the composition, processing, and HPLC data that is all available in the common database. Querying and reporting are done using Renaissance PolyView software.
An integrated electrode pH meter sheathed in a piercing tip that allows it to analyze sealed samples determines solubility pH profiles. The system can be used with two-point, three-point, or n-point calibration. Each pH measurement and the calibration data is saved to the database, allowing one to collect a series of pH measurements or look for when the system comes into equilibrium.
Recent protocols make it possible to robotically adjust wells to a target pH. This works by entering a target pH and threshold value into the design. The data acquisition software then takes an initial measurement of the well, and depending on the difference between the pH measurement and the target, aliquots of HCl and NaOH solutions are added to adjust the pH.
In one study, the robot was used to mix USP buffers into vials containing powder dispensed Naproxen or Ketoconazole. The resulting pH of each well was measured. The vials were then filtered, sampled, and analyzed by HPLC. The solubility was determined by HPLC measurements and was plotted against the measured pH to show the solubility pH profile.
Determining Partition Coefficients
Protocols have also been developed to determine the partition coefficient of drug compounds. In one example, the partitioning of Naproxen in water and octanol was determined as a function of pH. The first plate contained solutions for calibration curves and 50/50 mixtures of water and octanol. The second plate contained separated layers from the 50/50 mixtures that were captured from plate one. The drug was dispensed in solution and then the solvent was removed.
The desired solvent was added to the dry samples, either water, octanol, or a 50/50 mixture, according to the experimental design. Each well was then equilibrated with stirring, and aliquots of the calibration wells were transferred for analysis using the pH sensor, HPLC, and UV/Vis. The layers of the 50/50 mixtures were then separated robotically and transferred to plate two. Each well in plate two was then diluted and aliquots were transferred for pH, HPLC, and UV/Vis analysis.
Performing Forced Degradation Studies
By treating drug compounds to various chemical or environmental exposures and tracking the change in drug concentration over time, the CORE(x) system can be used to perform stress testing. In one example, amoxicillin solutions of various concentrations were followed over time. Both acidic and alkaline conditions lead to the hydrolysis of the drug compound.
Each well contained 1 mg of amoxicillin in 800 L total volume and were maintained at room temperature. The data showed a strong dependence on solution pH with the lowest degradation rate being observed in samples with a pH of about 6.0 and a rapid increase in degradation as the pH was lowered below 4.0 or raised above 8.0.
The trends found in this study are consistent with literature results that indicate the most stable solutions of amoxicillin have a pH of 6.2 .
In another study, a cephalexin library was subjected to a forced degradation study in which the effect of elevated temperature and chemical environment was measured. pH was increased from left to right across the library array while the results showed the concentration of cephalexin plotted against time for each of the formulations. Each sample was maintained at 80C.
The results showed that cephalexin degrades more quickly at high pH, with no cephalexin at all detectable after 4 hours with a pH greater than 7. Exposure to the radical generator AIBN, which was evaluated in two rows, did not appreciably change the degradation behavior of Cephalexin either in the non-pH adjusted solution or in the pH range explored. Exposure to HOOH, on the other hand, led to an increase in degradation.
The capability of performing forced degradation studies at different temperatures makes it possible to predict room temperature stability of an array of different formulations in order to identify a formulation that will provide high stability while maintaining high solubility. Using the robots, one can lay out replicate formulation libraries, equilibrate them, filter them, and measure their room temperature pH and solubility.
Subsequently, each of these plates can be heated to a different temperature, and aliquots can be pulled over time. For each temperature, a degradation rate constant can be determined. The Arrhenius relationship can then be used to extrapolate the predicted room temperature stability based on the three rate constants.
Dispensing High Viscosity Fluids
Higher viscosity fluids can be dispensed with special disposable tips that have a larger bore and require no cleaning between samples. Examples includes solubilizers such as lecithin, acacia, poloxamer, and Tweens; surfactants such as sodium lauryl sulphate and sorbitan monooleate; oils such as cottonseed oil, peanut oil, and Migliol oil; and other fluids such as Labrafil, other monoglycerides, and diglycerides.
In a typical experiment, Naproxen was dispensed as a powder and the solvents were dispensed robotically with the disposable tips. The samples were stirred and equilibrated at room temperature over a weekend. Samples were then centrifuged for one hour and an aliquot of the supernatant was extracted and diluted. The results showed that cottonseed oil and sorbitan monooleate create saturated solutions of the compound.
The data for sorbitan monooleate showed more scatter than the other systems due to incomplete separation with centrifugation. In recent work with corporate partners we have included the filtration of these viscous systems in the workflow to effect better separation of the supernatant from the solids.
Recently, we used this approach in a collaboration with a major pharmaceutical company in an effort to find a stable and highly soluble parenteral formulation for a commercial drug that is being sold as a freeze-dried formulation. The company was interested in developing a parenteral formulation for ease of manufacturing, patient compliance, and life cycle management reasons.
Working with the partner company, Symyx developed a workflow using the CORE(x) system that allowed two researchers to evaluate over 5,000 potential formulations with different cosolvents, pHs, and complexing agents in a period of five months. A number of promising formulations were identified and are currently undergoing further study.
High throughput methodology addresses a critical bottleneck in the preclinical drug development process by enabling high throughput, automated testing of drug candidates for formulation and preformulation solubility assessment. The use of a high throughput workflow increases the number of high quality, well-controlled, and validated experiments that a research group can perform without increases in staffing.
Such an approach increases the probability of development success by making it possible to restart stalled compounds, develop less expensive formulations, and increase patient compliance by improving formulations. The end result is an increase in both research productivity and value creation for preclinical drug development research groups.