Population and single genome kinetics driving the evolution of multiple linked multiclass drug resistance mutations in the viral protease and reverse transcriptase of HIV-1 subtype C in children receiving early protease inhibitor based combination therapy

• Main Author: Lange, C. M.
• Published: University College London (University of London) 2015
• Subjects: 616
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.668495

This thesis examines the evolution of HIV-1 subtype C multiple linked multi-class antiretroviral resistance mutations in the viral protease (PR) and reverse transcriptase (RT) genes of vertically infected children. Emergence of PI resistance on the backdrop of pre-existing non-nucleoside reverse transcriptase inhibitor (NNRTI) resistance could compromise long-term treatment options in such children. We characterised multi-class drug resistance using single genome sequencing (SGS) in children with viraemia while receiving PI-based ART. We applied SGS of HIV-1 protease (PR) and reverse transcriptase (RT) to longitudinal samples from a cohort of the Children with HIV Early Antiretroviral Therapy (CHER) trial with viral loads >1000c/ml after 40 weeks of early ART. Bulk sequencing revealed NVP-selected resistance in 50% of these children while SGS revealed NVP-selected resistance in 70%. Two children had baseline NRTI and PI mutations, suggesting previous maternal ART. Linked multi-class drug resistance following PI-based ART was detected by SGS in 2/10 children. In one child, the majority species contained M184V in RT linked to L10F, M46I/L, I54V and V82A in PR and a triple-class drug resistant variant with these mutations linked to the NNRTI mutation V108I. In the second child, the majority species contained M184V and V82A linked within viral genomes. I correlated nucleotide variation of PR-RT with the number of single genomes obtained at each time point and ART status and used maximum likelihood trees, recombination analysis, positive selection analysis and co-evolution analysis to describe the evolution of PR-RT of the viral populations. Six children who received early ART for 40/96 weeks only or received continuous ART for the duration of the CHER trial had clusters of identical sequences from baseline and week 40 of ART. These sequences did not harbour known drug resistance mutations. Therefore one could hypothesize viral replication from a persisting viral reservoir that was established from infection that occurred prior to the initiation of ART. The rooted ML trees of 2 children who developed drug resistance during ART had clusters of identical sequences harbouring common drug resistance mutations from multiple time points which is characteristic of the selection of drug resistant viral populations that cause virological failure during ART. When drug resistant viral populations developed during treatment failure, M184V single mutated viruses were selected from multiple wildtype viral populations but only one population became the major contributor to drug resistant viraemia in both children. Triple-class drug resistant sequences that had common DRMs (M184V, V108I in RT and M46I in PR) did not cluster together. I found no evidence of recombination or coevolving sites in PR-RT for any of these children. I used a luciferase based single replication cycle assay to examine drug susceptibility and replication capacity (RC) conferred by multi-class drug resistant PR-RT from the 2 children who developed such drug resistant variants. I tested the susceptibility of pseudoviruses to the components of early ART (AZT, 3TC and LPV), the components of second-line therapy for these children (Abacavir (ABC), Didanosine (ddI), Efavirenz (EFV) and (NVP)), the PIs Nelfinavir (NFV) and Saquinavir (SQV), which are also approved for use in children and Darunavir (DRV), which has been identified as a PI option needed in paediatric co-formulation. Pseudoviruses with known PI resistance conferring mutations showed reduced susceptibility to all PIs except DRV. Those with known NNRTI resistance conferring mutations showed reduced susceptibility to EFV and NVP. M184V mutated pseudoviruses conferred high-level resistance to 3TC. In one child, a combination or one of the RT mutations V35T, E36D, T39R, S48T, T165I, K173A, D177E, T200A, Q207D, R211K, V245Q, E248N, D250N, A272P, K277R, E291D, I293V, T296N may be associated with high-level ABC and ddI resistance when genetically linked with M184V. Population sequence analysis was used to characterize the viral gag genes that encoded matrix, capsid, nucleocapsid, p6, and spacer peptides 1 and 2 along with PR-RT as a single amplicon. I determined the presence of compensatory PI-resistance mutations in gag, drug resistance mutations in PR and RT and other amino acid changes that occurred during ART. To determining the polymorphic nature of these sites, I compared them to a position-specific scoring matrix for gag that was derived from HIV-1 subtype C sequences from children from Sub-Saharan Africa. P453L in the p1/p6 cleavage site of Gag emerged in the viral population of one child during PI-based ART. It was the only amino acid change in Gag that emerged among all children in the study cohort that has been characterised as a compensatory mutation that is selected by and enhances PI-resistance. This project is the first to identify multi-class drug resistance mutations in PR and RT that were linked on the same genome as well as characterise their development during early PI-based ART in children. Triple class drug resistant viruses detected in the minority species of the viral population of one child demonstrated significant levels of resistance to LPV, SQV, NFV, 3TC and NVP, and established that such variants could compromise future ART regimes if they became the dominant species of the viral population. I note that the small convenience sample (n = 10) chosen for this project limited the power of this study so that findings could not be generalized.