Background: The drug maraviroc differs from other HIV-1 drugs as it binds the host cell coreceptor CCR5 as opposed to a virus protein. Lack of sensitivity to such entry-inhibitor drugs occurs primarily through emergence of pre-existing viruses that use the CXCR4 coreceptor, while genuine CCR5-antagonist resistance occurs through virus use of the drug-bound CCR5 coreceptor. The aim of this project is to understand the R5 resistance pathway in a comprehensive way based on envelope sequence data. Methodology: We use population genetic, coevolutionary, recombination, positive selection and structural analysis to analyse the envelope protein sequences from 20 patients from the clinical trials MOTIVATE 1 and 2. All these patients have exclusively CCR5-using (R5) virus. All patients have two groups of sequences that have been sequenced before and after treatment with maraviroc/placebo-arm. Three patients under placebo-arm treatment are used as control. Results: We find the resistant viral population to be phylogenetically distinct from those susceptible to the drug and associated with a clear genetic bottleneck in each patient, consistent with de novo emergence of resistance. We find limited overlap in the envelope sites or point mutations involved. Particularly, we find: (1) complex coevolutionary networks (indicating epistatic interactions in the context of protein structure) in both the sensitive and resistant populations. (2) A subset of coevolving sites are under positive selection indicating their importance for drug resistance. (3) CD4 binding sites formed a unique coevolutionary network that is independent of the V3 loop. (4) The coevolutionary network formed between the V3 loop and other regions of gp120 and gp41 intersects with sites involved in the N-linked glycosylation motif. (5) Key residues in the signal peptide involved in Env function and expression level coevolve with the potential N-linked glycosylation sequence motifs (NXS/TX). (6) Finally, there is no statistically significant change of the V3 loop structural disorder between sensitive and resistant R5 virus. This indicates changes in other protein regions are required for achieving the binding specificity required for resistant R5 virus, which is consistent with the results of the above coevolutionary analysis. Conclusions: Coevolution between V3 loop, N-linked glycosylation motifs and the signal peptide are key to V3/gp120/gp41 folding into particular conformations that can utilize drug-bound CCR5 for cell entry. Collectively our results demonstrate that although each viral population’s resistance pathway is context dependent, the evolutionary trajectories of the viruses are significantly constrained accounting for the low levels of resistant R5 virus.
CROI Conference
