Drugs and light: A computational approach to predict phototoxicity

Drugs exposure, either to visible (Vis) or ultraviolet (UV) light, along the pharmaceutical chain is inevitable; from their manufacture until their dispensation or even after administration. Thus, it is crucial to investigate their photophysics and the photoreactive paths activated upon photon absorption for predicting either an eventual loss of potency of the active pharmaceutical products or a possible production of photochemical reactive species. Photosafety recommendations are outlined in the International Conference on Harmonization (ICH) S10 guidance. The ICH S10 guidance (and the associated ICH M3 guidance) suggests the characterization of the UV visible absorption spectrum as the initial assessment because it can obviate any further photosafety evaluation. First attempts to predict photostability and phototoxicity of drugs in-silico, were based on the calculation of the HOMO-LUMO energy gap. However, the predictive power of this indicator is limited for certain classes of drugs, such as for instance, Non-Steroidal Anti-Inflammatory Drugs (NSAIDs) and therefore they are sometimes complemented with indicators of light absorption intensity, i.e. the molar extinction coe cient. As a first approach to predict photostability, we have modelled the absorption spectra of a selected group of NSAIDs, such as aspirin (in gas phase as well as in solvent) and ibuprofen, indomethacin, carprofen and suprofen (in gas phase). Multistate second order perturbation theory on state average complete active space self-consistent field wavefunctions MS-CASPT2//SA-CASSCF and time dependent density functional theory (TD-DFT) were the computational protocols for this purpose. We explored the most probable photophysical deactivation mechanism of the excited molecules with the semi-classical dynamics simulations, performed with the surface-hopping algorithm incorporating spin orbit coupling. Higher chances of producing phototoxic species are expected the longer the drug remains excited, so the keys behind photostability of a drug is the accessibility of S0, although reactive processes may also take place from a hot ground state. Therefore, our goal with these results is to generate a model that would allow us to predict photostability with the information of the di erent deactivation mechanisms and new indicators and at the same time, improve the assessment of the photophysical properties of drugs.