(686b) Process Development for a Co-Crystal of Telaprevir (VX-950) and 4-Hydroxybenzoic Acid

Kline, B., Vertex Pharmaceuticals
Blythe, T., Vertex Pharmaceuticals
Borsje, E., Vertex Pharmaceuticals
Dokou, E., Vertex Pharmaceuticals
Gasparac, R., Vertex Pharmaceuticals
Hurrey, M. L., Vertex Pharmaceuticals
Israni, M., Vertex Pharmaceuticals
Johnston, M., Vertex Pharmaceuticals
Jurkauskas, V., Vertex Pharmaceuticals
Knezic, D., Vertex Pharmaceuticals
Mudunuri, P., Vertex Pharmaceuticals
Pesha, Y., Vertex Pharmaceuticals
Reda, K., Vertex Pharmaceuticals
Phenix, B., Vertex Pharmaceuticals
Van Kirk, C., Vertex Pharmaceuticals

When development work was initiated, the existing process for generation of a co-crystal of telaprevir (VX-950) and 4-hydroxybenzoic acid (4-HBA) was uncontrolled precipitation from a solvent mixture.  Anecdotally, there was the existence of a “magic” co-crystal batch that had superior dissolution performance and had been generated from the uncontrolled process.  For a time, technical work was driven by this knowledge and there was an ill-fated attempt to reproduce this batch.  However, when it was determined that the precipitation led to co-crystal with variable properties and performance, focus shifted to obtaining a more fundamental understanding of the crystallization.  The objective for developing the process was then to design a controlled crystallization that would lead to co-crystal material with reproducible physical properties and dissolution behavior.  There is a significant difference in solubility of the API and the coformer which led to difficulty in choosing an appropriate solvent system.  This presentation will review the candidate solvent systems and modes of crystallization that were attempted.  After choosing dichloromethane and methyl tert-butyl ether as the solvent system, an equilibrium phase diagram was generated for the solvent composition at the seed point.  This phase diagram indicated that the process was at risk to produce co-crystal with excess VX-950.  By slightly changing the parameters in the crystallization, it was found that only co-crystal would be formed.  Further work involved optimization of the solvent composition at the end-point to improve yield and minimize agglomeration.  Development and understanding of the phase diagram was essential to reliably producing VX-950/4-HBA co-crystal.