A number of companies offer high-throughput crystallisation services, and typically charge in the region of £100,000 (€145,000) to carry out the process. Accentus is commercialising a new in silico predictive crystallisation technology that should be available for a little over half that price.
Called CrystalGEM, the system takes the form of a combination of computational models with actual experimental data, a marriage that significantly increases its predictive power compared to earlier computational approaches.
Dr George Tranter, director of UK company Chiralabs that developed the system and licensed it to Accentus, told In-PharmaTechnologist.com that one of the problems with crystallisation experiments is that there are an almost infinite number of solvent combinations that can be used to make crystals. And this means that thousands of experiments need to be undertaken - under many different experimental conditions - in order to identify the desired one.
CrystalGEM is based on the analysis of thousands of real crystallisation systems using single or mixed combinations of over 80 commercially available solvents, coupled to a proprietary pattern recognition algorithm.
Put simply, the system filters out the thousands of experiments that might feasibly be done to screen crystal forms down to a few dozen, with significant savings in time and also the amount of compound needed.
In experiments using indomethacin, the system has been showed to have an 86 per cent hit rate in predicting the experimentally derived crystallisation, and 87 per cent reliability, said Tranter.
It uses a standard 96-well microplate and requires just 1mg of sample per well, and reduces the total time for conducting the screen from 3 months to about three weeks, according to David Hipkiss, Accentus' general manager. The total requirement of around 50mg per screen matches the amount of a compound that typically comes out of medicinal chemistry, and is significantly less than the half kilo or so needed using conventional high throughput crystallisation screening.
Other benefits include minimising operator exposure to a new compound, which can be a significant factor with high-potency drugs, for example.
"Importantly, the approach highlights areas of solvent space that have the highest chance of becoming a system suitable for scale-up and manufacture," said Tranter.
This makes the requirement for high-cost, high throughput crystallisation screening obsolete, he added.
Companies offering high-throughput crystallisation screening services to the drug industry include Avantium of the Netherlands, Switzerland's Solvias and SSCI of the US.
Meanwhile, Accentus is interested in exploring the possibilities of using CrystalGEM to identify polymorphs of drugs that could be made commercially and circumvent existing patents. In this scenario, the company would expect a royalty-bearing licensing type of arrangement, said Hipkiss.