RESUMEN
OBJECTIVE: Exposure to abacavir is associated with T-cell-mediated hypersensitivity reactions in individuals carrying human leukocyte antigen (HLA)-B57 : 01. To activate T cells, abacavir interacts directly with endogenous HLA-B57 : 01 and HLA-B57 : 01 expressed on the surface of antigen presenting cells. We have investigated whether chemical modification of abacavir can produce a molecule with antiviral activity that does not bind to HLA-B57 : 01 and activate T cells. DESIGN: An interdisciplinary laboratory study using samples from human donors expressing HLA-B57 : 01. Researchers were blinded to the analogue structures and modelling data. METHODS: Sixteen 6-amino substituted abacavir analogues were synthesized. Computational docking studies were completed to predict capacity for analogue binding within HLA-B57 : 01. Abacavir-responsive CD8 clones were generated to study the association between HLA-B57 : 01 analogue binding and T-cell activation. Antiviral activity and the direct inhibitory effect of analogues on proliferation were assessed. RESULTS: Major histocompatibility complex class I-restricted CD8 clones proliferated and secreted IFNγ following abacavir binding to surface and endogenous HLA-B57 : 01. Several analogues retained antiviral activity and showed no overt inhibitory effect on proliferation, but displayed highly divergent antigen-driven T-cell responses. For example, abacavir and N-propyl abacavir were equally potent at activating clones, whereas the closely related analogues N-isopropyl and N-methyl isopropyl abacavir were devoid of T-cell activity. Docking abacavir analogues to HLA-B57 : 01 revealed a quantitative relationship between drug-protein binding and the T-cell response. CONCLUSION: These studies demonstrate that the unwanted T-cell activity of abacavir can be eliminated whilst maintaining the favourable antiviral profile. The in-silico model provides a tool to aid the design of safer antiviral agents that may not require a personalized medicines approach to therapy.