Acquired and transmitted drug resistance in HIV-1 subtype C : implications of novel mutations on replication capacity, cleavage and drug susceptibility.

Abstract

Introduction Large scale roll-out of combination antiretroviral therapy (cART) has been successful in improving the quality of life of HIV-1 infected individuals in South Africa (SA). However the development and transmission of drug resistance threatens the future success and longevity of cART in the country. Studies have shown that resistance to Protease inhibitors (PI’s), in the absence of mutations in Protease (PR), is increasing in SA. Whilst some studies attribute this to poor treatment adherence, others have shown that mutations in Gag contribute to PI resistance. The majority of these studies however have been conducted on HIV-1 subtype B, despite HIV-1 subtype C being the most prevalent subtype globally. Given that Gag is highly polymorphic between subtypes, studies focusing on HIV-1 subtype C are required. Despite the high rate of virologic failure of patients on PI inclusive treatment regimens, no transmitted drug resistance (TDR) studies have identified PI associated TDR mutations. This could be due to the high fitness cost associated with PR mutations which would result in rapid reversion or low frequency of mutations within the viral quasispecies. Most TDR studies in SA, as in other resource limited settings, have used recently infected cohorts to measure TDR. It is however unlikely that rapidly reverting mutations would be detected in recent infection. Furthermore, these studies have all used Sanger sequencing which only detects mutations at frequencies >15-20%. With recent studies showing that low frequency mutations present at frequencies as low as 1% impact treatment outcomes, the elucidation of these mutations using deep-sequencing techniques is necessary. For a true measure of TDR, studies employing acute infection cohorts and deep-sequencing techniques are required. The current study aimed to identify mutations in Gag-Protease associated with PI resistance/exposure, and to determine their impact on replication capacity and drug susceptibility. The prevalence of low frequency TDR mutations in an HIV-1 subtype C acute infection cohort was also investigated. Methods A cohort of 80 HIV-1 subtype C infected participants failing a PI inclusive treatment regimen (i.e. PCS cohort) from 2009–2013 in Durban, South Africa was used to assess the role of Gag in PI resistance. Gag mutations were divided into three groups: PI exposure associated Gag mutations; resistance associated Gag mutations (rGag) and novel Gag mutations (nGag). Frequencies of each of these mutations were compared amongst: 80 PCS cohort sequences, 2,481 HIV-1 subtype B treatment naïve sequences, 954 HIV-1 subtype C treatment naïve sequences and 54 HIV-1 subtype C sequences from acutely infected individuals, in order to identify PI associated mutations and natural polymorphisms. Next, recombinant viruses for all 80 participants were generated by co-transfection of a CEM derived T-cell line (i.e. GXR cells) with an NL43-deleted-gag-protease (NL43Δgag-protease) backbone and patient derived Gag-Protease amplicons. Thereafter, the replication capacity of each virus was assessed using a replication assay that employed a green fluorescent protein reporter cell line and flow cytometry. Associations between replication capacity and Gag-Protease mutations were established. Eighteen viruses with mutations of interest were then selected for use in drug susceptibility assays, where the impact of mutations on susceptibility to lopinavir (LPV) and darunavir (DRV) was assessed in a luciferase based assay. Lastly, the impact of novel Gag mutations on replication capacity and drug susceptibility was validated by generating site-directed mutant viruses with mutations of interest and using these mutant viruses in replication capacity and drug susceptibility assays. Furthermore the cleavage profile of each site-directed mutant virus was established by western blotting. Samples available from 47 HIV-1 subtype C acutely infected individuals collected from 2007-2014 in Durban, South Africa, was used to assess low frequency TDR mutations in HIV-1 subtype C acute infection. Firstly the RT and PR region of each virus was genotyped using the Viroseq HIV-1 genotyping system in order to identify the prevalence of TDR in the cohort. Thereafter 14 participant samples were selected, based on the availability of plasma at one week after onset of plasma viremia (OPV), for sequencing by ultra-deep pyrosequencing (UDPS). This served to identify low frequency mutations. Comparisons in TDR prevalence was made between Sanger sequencing and UDPS. Thereafter, the impact of low frequency TDR mutations on treatment outcomes was assessed by comparing time to virologic suppression for two participants with low frequency mutations to that of four participants without low frequency mutations. Results Protease resistance associated mutations (RAMs) occurred in 34/80 (42.5%) participants, whilst Gag mutations associated with PI resistance in subtype B were detected in 67/80 (84%) participants. Overall, 12 Gag mutations associated with PI exposure (i.e. E12K, V35I, G62R, V370A/M, S373P/Q/T, A374P, T375N, I376V, G381S, I389T, I401T and H219Q), eight rGag mutations (i.e. R76K, Y79F, V128I, A431V, K436R, L449F, R452K and P453L) and four nGag mutations (i.e. Q69K, S111C/I, T239A/S and I256V) were identified in the PCS cohort. The

E12K, V370A/M, T375N, G381S, R76K and Y79F mutations all occurred as natural polymorphism in HIV-1 subtype C. The A431V, K436R, L449F, R452K, P453L, Q69K, S111C/I, T239A/S and I256V mutations were all associated with PI resistance/exposure. Interestingly all viruses with PR RAMs harboured rGag and nGag mutations, however rGag and nGag mutations were also found to occur without PR RAMs. Protease RAMs were associated with significantly reduced replication capacity. The K335R and A431V mutations were the only Gag mutations associated with significantly reduced replication capacity. Viruses with PR RAMs were associated with significantly reduced susceptibility to LPV (>15 FC in IC50) and DRV (>6 FC in IC50). Furthermore, the following combinations of rGag and nGag mutations were found to confer reduced susceptibility to LPV and DRV in the absence of PR RAMs: R76K+Y79F+K436R+L449P+I256V (5.2 fold increase in IC50 for DRV), R76K+R453L (23.88 fold increase in IC50 for LPV and a 6.73 fold increase in IC50 for DRV) and R76K+K436R+Q69K+S111C (7.40 fold increase in IC50 for LPV). Analysis of recombinant viruses showed that the Q69K nGag mutation rescued replication capacity of all viruses harbouring A431V+PR RAMs. This was validated by SDM, where Q69K rescued the replication capacity of site-directed mutant viruses harbouring A431V+V82A. The Q69K mutation was also associated with increasing polyprotein cleavage when found in conjunction with A431V+V82A. With regards to TDR, we demonstrated a prevalence of 57% of TDR mutations with UDPS and 2.2% with Sanger sequencing. Sanger sequencing identified the K103N non-nucleoside reverse transcriptase inhibitor (NNRTI)-associated TDR mutation. In addition to K103N (frequency: >99%), the following low frequency mutations were detected by UDPS: the K65R (1-1.5%) and D67N (3.88%) nucleotide reverse transcriptase inhibitor (NRTI)-associated TDR mutations, the F53L (17.6%) and M46L (6.3%) Protease inhibitor (PI)-associated TDR mutations, and the T97A (2.90%) integrase strand transfer inhibitor (InSTI)-associated TDR mutations. Participants with low frequency TDR mutations took 40 days longer to achieve viral suppression than participants without low frequency TDR mutations, when placed on fixed dose combination antiretroviral therapy.