Yesterday I had the privilege of attending a workshop at the NIH on the National Cancer Institute Clinical Development of Small Molecules:
This one-day workshop will provide specialized training and information to NCI-supported investigators who plan to undertake clinical development of novel concepts and who are directly involved with implementing translational clinical research. Individuals will benefit from the opportunity for direct interaction with FDA’s Center for Drug Evaluation and Research and NCI’s Developmental Therapeutics Program senior staff.
I am grateful to Dan Zaharevitz for the invitation to the workshop yesterday and to a visit of his screening labs later today. Dan has a vision that the drug discovery and development process could benefit tremendously through openness at all stages by facilitating communication between all parties involved from discovery through clinical testing. For example, if discussions about formulation, ease of scale-up and toxicity occurred early on, perhaps a more efficient overall process would result. Our selection of Ugi products that crystallize from the reaction mixture and that are obtainable from cheap commercially available starting materials is an attempt at anticipating the scale-up and cost factors down the road. Of course, this is what Open Notebook Science is all about - getting feedback from those with expertise at the earliest possible moment. The scale-up consideration I mentioned is an obvious factor from the perspective of an organic chemist. But there are other parameters that could benefit immensely from input from the drug development community. Formulation is a big one. So far I haven't placed any restriction or requirements on functional groups in our Ugi virtual libraries. Most are not going to be very soluble in water and I don't have a good feeling about how important that is. At the workshop, it was repeated several times that a water soluble compound is preferable but it is certainly not a deal breaker. For IV delivery liposomal formulations are an option, as are some other FDA approved co-solvent systems. Right now we're shipping powders for our assays. It is my understanding that these are usually taken up in DMSO. But we don't have to do that - I have some experience with the preparation of liposomes and it would not be too much of a hassle to ship our compounds formulated in that way. There are issues of stability that have to be considered but these are manageable. Now another approach would be creating compounds that are water soluble from the start. We could do that by introducing tertiary amines as a second functionality on our starting materials but that would severely reduce the size of our virtual libraries. Another strategy would involve using boc-protected amino acids, which we know work well in the Ugi reaction. The problem there is that they would have to be deprotected and that would reduce the convenience of the preparation. Then there is the issue of optimizing by metabolites. Perhaps we should not be docking only the Ugi products but also their likely metabolic products. An expert in pharmacodynamics could certainly provide valuable input here. Anticipating toxicity is another consideration. These are just examples of the types of conversations that we could be having well before any compounds get to in vivo or clinical trials. And that is what Open Notebook Science is all about. Instead of waiting for our paper on the synthesis of new inhibitors of a certain enzyme to appear in print, give us some feedback while the experiment is being done. There are definitely some obstacles to overcome to achieving this type of transparency and collaboration but technology is not the bottleneck. Even finding collaborators is no longer the key issue, with people like Dan Zaharevitz, Rajarshi Guha, Egon Willighagen, Gus Rosania, Philip Rosenthal, Tsu-Soo Tan, Cameron Neylon, Antony Williams, Kevin Owens, Peter Murray-Rust, and others stepping up to contribute what they can. I think the main issue now is convincing a funding organization that this is a model worthy of support. Maybe NSF will be one of them.