The prediction of a drug’s likelihood of inducing an adverse drug reaction (ADR) is extremely difficult especially if this ADR event occurs via an immune based mechanism, also known as idiosyncratic ADR. A common idiosyncratic ADR pathway occurs when a drug binds to the human leukocyte antigen (HLA) protein which is an immune-receptor protein involved in the signaling pathway for T-cell activation. This pathway is extremely difficult to predict for a multitude of reasons. First, there are currently over 16,000 HLA allele sequences available in the IMGT/HLA database (read full text here). Second, these variants occur at different frequencies in different populations, and lastly a drug may induce idiosyncratic ADRs by binding to multiple HLA-variants. For example, in Han Chinese the drug carbamazepine (used to treat seizures) will induce Steven-Johnson Syndrome (SJS) in patients who have the HLA-B*15:02 variant; however, in Northern Europeans carbamazepine induces SJS by binding to the HLA-A*31:01 variant. Due to these factors, it is extremely difficult to predict ADR events and these events are typically not observed until a drug is administered to the population.
In 2012 two research groups (Illing et al. and Ostrov et al.) successfully identified the binding mode of abacavir (used to treat HIV) with the HLA-B*57:01 variant and three co-binding peptides by solving three X-ray crystals (PDB: 3VRI, 3VRJ, and 3UPR). Using these newly solved crystals we developed a virtual screening model using molecular docking to forecast drug-HLA interactions with the HLA-B*57:01 varaint. Typically, when molecular docking has been used to study drug-HLA interactions there is a limited focus on the role of co-binding peptide; however, our model specifically explored the detailed interactions between HLA, drug, and peptide in a two-tiered docking protocol. First, docking was conducted in the absence of peptide and then docking was performed in the presence of peptide. The findings of our model can be found here as a recent publication in the Journal of Cheminformatics.
Van Den Driessche, G.; Fourches, D.; Adverse drug reactions triggered by the common HLA-B*57:01 variant: A molecular docking study. 2017, 9, 1-13.