New mechanisms of resistance in virological failure to protease inhibitors: selection of non-described protease, Gag and Gp41 mutations Free

Abstract

Introduction

Selection of drug-resistant HIV during viral replication under treatment pressure is a major reason for therapy failure. While the introduction of PIs has ensured robust combination therapy, virological failure (VF) still occurs in clinical practice. PIs are known to suppress proteolytic cleavage of HIV Gag and Pol polyproteins that include essential structural and enzymatic components of the virus.¹ This prevents the conversion of HIV particles into their mature infectious form. Major PI resistance mutations in the protease gene lead to decreased PI binding.^2–5 In parallel, compensatory mutations outside the substrate-binding pocket of protease can at least partially restore viral replication.^6,7

The viruses of most patients on PIs who experience VF do not harbour any resistance associated-mutations in the protease gene. Studies of mutations inside or outside gag gene cleavage sites (CSs) have contributed to the understanding of new PI resistance mechanisms,^8–12 but the impact of some identified mutations has not been clearly defined.¹³ Moreover, it has been suggested that Gp41 mutations could also have implications for resistance to PIs. Indeed, approximately half the inhibitory potential of PIs occurs at the entry step, probably reflecting the interactions between the uncleaved Gag polyprotein and the cytoplasmic tail (CT) of the envelope transmembrane glycoprotein Gp41.¹⁴

To further understanding of VF in patients failing a first-line PI-based regimen, we used ultradeep sequencing (UDS) to study the protease, gag and gp41 regions.

Patients and methods

Patients

All naive patients initiating an antiretroviral treatment based on a boosted darunavir-, atazanavir- or lopinavir-containing regimen and experiencing VF without any transmitted drug resistance (TDR) mutation detected by Sanger sequencing at baseline on protease and reverse transcriptase genes according to the 2009 consensus TDR list were selected.¹⁵ VF was defined as two consecutive plasma viral loads >50 copies/mL after a plasma viral load <50 copies/mL or as a viral load decrease not satisfying French monitoring guidelines after treatment initiation.¹⁶ Plasma concentrations of all ART drugs were determined using UPLC coupled with tandem mass spectrometry (Waters Corp., Milford, MA, USA).

Patients were monitored by the Department of Infectious Diseases at Pitié-Salpêtrière (from January 2011 to January 2016) and Bichat-Claude Bernard hospitals (from January 2012 to March 2015) (Paris, France). Data were obtained from the existing electronic database or medical records. Patients were informed that their demographic and clinical data would be recorded during follow-up and could be used for retrospective studies, and gave their consent. This study was approved by the Agence Nationale de Recherches sur le SIDA et les hépatites virales (ANRS) AC11 Ethics Committee.

UDS

UDS of protease, gag and gp41 genes was performed in plasma before initiation of treatment and at VF. After RNA extraction (Qiagen^® QIAamp viral RNA mini, Venlo, The Netherlands), a reverse transcription into DNA was performed (Transcriptor^®, Roche, Basel, Switzerland) and the DNA was then amplified by PCR. A nested PCR was carried out with a high-fidelity Taq polymerase (Q5 High-Fidelity DNA Polymerase, New England Biolabs, Ipswich, MA, USA for the protease gene; Superscript III^® one-step RT-PCR with platinum Taq high fidelity, Invitrogen, Carlsbad, CA, USA for the gag and gp41 genes). UDS was in accordance with the manufacturer’s recommendations (Nextera^® kit, Illumina^®, San Diego, CA, USA) with a cut-off of 1% at baseline. At failure, only emergent mutations present in more than 20% of the total viral population, detected in at least one of two samples and whose proportion increased at least 2.5-fold between treatment initiation and VF were taken into account. Geneious^® 9.0 (Brisbane, Australia) was used to analyse UDS results.¹⁷

Results

During the period under study, 409 patients began antiretroviral treatment with reverse transcriptase inhibitors in combination with boosted darunavir (318 patients), atazanavir (72 patients) or lopinavir (19 patients).

Of the total number of patients, 32 experienced VF without any TDR mutation detected by Sanger sequencing of protease and reverse transcriptase genes (7.8%) and were retrospectively included in the study. Among this group, 53.1% were infected with HIV-1 recombinant CRF02_AG. The median time between baseline and VF was 495 days and the median nadir CD4 cell count was 220 cells/mm³. Antiretroviral plasma concentrations in samples taken at VF were consistent with adequate compliance for all patients. Patient characteristics and treatments received are presented in Table 1.

Before initiation of treatment, samples from 26, 27 and 27 patients were analysed for protease, gag and gp41 genes, respectively. At failure, samples from 21, 13 and 14 patients were analysed for protease, gag and gp41 genes, respectively. Before initiation of treatment, the median number of reads was 118443, 195405 and 197368 for protease, gag and gp41 sequences, respectively. At VF the median number of reads was 289786, 360697 and 316552 for protease, gag and gp41 sequences, respectively.

At VF, five of the 32 patients displayed viruses harbouring emergent mutations in the protease gene that have not been previously described at VF: 64M (one patient infected with a CRF02_AG subtype), 64M/70R in combination with mutation 15V (two patients infected with a CRF02_AG subtype), 79A (one patient infected with a CRF06_cpx subtype), and 79A in combination with mutation 15V (one patient infected with a CRF02_AG subtype). All five patients were treated with a combination of two NRTIs and boosted darunavir q24h (Table 2).

Two patients were found to have viruses containing the mutation R286K in Gag, outside the CS. Both patients were infected with a CRF02_AG virus and treated with a combination of two NRTIs and boosted darunavir q24h. At baseline, these two patients’ viruses also harboured CS Gag mutations in >90% of the total viral population: one patient’s virus harboured the mutation L449P and the other patient’s virus harboured the mutations Y132F and L449P. In addition, 13 other patients’ viruses harboured mutations of CS before treatment began (Table 2). Most of these mutations were present in more than 90% of total viral population: Y132F (p17/p24), S373P and S373T (p17/p24), K436R (p7/p1), L449P, S451N and P453L (p1/p6).

In the gp41 region, two patients’ viruses had selected a mutation inside the CT, V321I (one patient with subtype A and one with subtype B viruses; both patients were treated with a combination of two NRTIs and boosted darunavir q24h), and eight had selected mutations outside the CT: K106R (two patients with CRF02_AG viruses; one patient was treated with boosted darunavir q24h and the other with boosted atazanavir), E109D (one patient with CRF02_AG and one with subtype B viruses; both patients were treated with two NRTIs and boosted atazanavir), T165S (one patient with CRF02_AG and one with subtype B viruses; both patients were treated with two NRTIs and boosted darunavir q24h); and K172R (one patient with CRF02_AG and one with subtype A viruses; both patients were treated with two NRTIs and boosted darunavir q24h) (Table 2).

Sanger sequencing of integrase and reverse transcriptase genes was performed at VF for 17/32 patients whose samples were still available. No mutation was observed in the integrase gene. For the reverse transcriptase gene, one mutation (M184I) was detected at failure in patient 12.

Discussion

In 14 patients failing a boosted PI first-line regimen we identified the emergence of protease, Gag and gp41 mutations that have not previously been described or clearly associated with PI resistance. Eleven were treated with boosted darunavir once daily.

These results highlight the emergence of a pattern of protease mutations (15V, 64M, 70R) among CRF02_AG viruses under darunavir pressure. These mutations may affect the virological response of HIV-1 non-B subtype to PIs, and when they are detected a more rigorous follow-up might be considered for these patients. The mutation 15V has not previously been associated with darunavir resistance but has already been associated with resistance to saquinavir. The mutations 64M and 70R have already been identified in CRF02_AG HIV as polymorphisms, but not as emergent mutations at failure.

We did not identify the emergence of the previously described CS Gag mutation at VF. However, one mutation emerged outside CS (R286K). This mutation has already been associated with a significantly higher rate of VF under a PI-based regimen and contributes to PI resistance.¹⁸ Before treatment commenced, mutations identified in CS were already present for many patients in >90% of the total viral population.^9,10 Even if these mutations were not associated with VF of darunavir, they might have influenced the selection of treatment-associated substitutions.

To the best of our knowledge, data on emerging mutations in the gp41 gene in cases of VF of a PI regimen are reported here for the first time. It would be interesting to study further the mutation V321I located on CT which has been identified in two viruses’ gp41 gene at VF.¹⁴ This mutation could be linked to some interactions between the uncleaved Gag polyprotein and the CT of the Env protein, and may have a potential impact on virological response to PIs.¹⁴

Our study is limited by its sample size, mainly as a result of the strict criteria applied to patient selection. Moreover, Sanger sequencing of the reverse transcriptase gene could not be performed for every case. Nevertheless, the sequencing was performed on protease, gag and gp41 genes by a very sensitive technology, at baseline and at VF, in treatment-compliant patients.

In conclusion, this study shows selection of some protease, Gag and Gp41 mutations that have not previously been described and could be related to VF. Site-directed mutagenesis and phenotypic studies should be conducted to improve understanding of these mutations and better characterize their impact on susceptibility to PIs.

Acknowledgements

We thank Isabelle Malet for technical assistance.

Funding

This work was supported by the Agence Nationale de Recherche sur le SIDA et les hépatites virales (ANRS). This study was also funded by Janssen-Cilag and BMS. The study sponsors agreed with the study design, but otherwise had no role in data collection and analysis, decision to publish or preparation of the manuscript.

Transparency declarations

None to declare.

References