In silico screening of drug candidates for thermoresponsive liposome formulations

Liposomal formulations can be advantageous in many scenarios such as targeted delivery to reduce the systemic toxicity of highly potent active pharmaceutical ingredients (APIs), to increase drug bioavailability by prolonging systemic circulation, to protect labile APIs from degradation in the gastrointestinal tract, or to improve skin permeation in dermal delivery. However, not all APIs are suitable for encapsulation in liposomes. Some of the issues are too high permeability of the API across the lipid bilayer, which may lead to premature leakage, too low permeability, which may hinder the drug release process, or too strong membrane affinity, which may reduce the overall efficacy of drug release from liposomes. Since the most reliable way to test API encapsulation and release from liposomes so far has been experimental, an in silico model capable of predicting API transport across the lipid bilayer might accelerate formulation development. In this work, we demonstrate a new in silico approach to compute the temperature-dependent permeability of a set of compounds across the bilayer of virtual liposomes constructed by molecular dynamics simulation. To validate this approach, we have conducted a series of experiments confirming the model predictions using a homologous series of fluorescent dyes. Based on the performance of individual molecules, we have defined a set of selection criteria for identifying compatible APIs for stable encapsulation and thermally controlled release from liposomes. To further demonstrate the in silico-based methodology, we have screened the DrugBank database, identified potent drugs suitable for liposome encapsulation and successfully carried out the loading and thermal release of one of them – the antimicrobial compound cycloserine.