Crystal structure prediction is increasingly becoming significant in the field of organic materials, as emphasized by Gregory Beran from the University of California Riverside. Beran’s research merges hybrid density functional theory modeling with intramolecular energy correction to forecast intricate crystal structures, including the pharmaceutical compound axitinib. This method also differentiates between salt and co-crystals within multi-component crystals, tackling a difficult aspect of organic crystal modeling.
Axitinib, a targeted therapy for advanced renal cell carcinoma, inhibits tyrosine kinases that promote tumor growth. Nevertheless, its development encountered obstacles due to polymorphism, as axitinib can crystallize in several forms with differing solubility and bioavailability characteristics. Initially, form IV was targeted until forms XXV and eventually the FDA-approved form XLI surfaced, illustrating the variation in polymorphic stability.
Forecasting polymorph stabilities through density functional theory (DFT) has traditionally posed challenges, partly due to density-driven delocalization errors impacting salt-co-crystal predictions. Beran’s combination of DFT and intramolecular energy correction has produced the first 0K crystal structure prediction for axitinib that closely aligns with experimental findings. He optimized pre-existing crystal structures and projected variants and multi-component crystals that incorporate acids such as fumaric, suberic, and trans-cinnamic.
By employing periodic planewave DFT and exchange-hole dipole moment dispersion correction, Beran adjusted atomic positions and lattice parameters, enhancing energies with generalized-gradient approximation and hybrid functionals. Intramolecular corrections and Orca software further improved energy assessments.
The results confirm that the majority of low-energy predicted structures for axitinib are experimentally acknowledged, with form XLI recognized as the global minimum at 0K. This approach also distinguished between axitinib salts with fumaric acid and co-crystals with suberic or trans-cinnamic acids.
Sarah (Sally) Price at University College London remarks on the study’s promise in forecasting molecule crystallization as salts or co-crystals, paving the way for novel experimental modeling opportunities. Although crystal structure prediction is not flawless, Beran is confident that it is on the verge of substantially affecting future advancements.