Journal of Antimicrobial Chemotherapy Advance Access published March 8, 2015
J Antimicrob Chemother doi:10.1093/jac/dkv048
Highly frequent HIV-1 minority resistant variants at baseline of the ANRS 139 TRIO trial had a limited impact on virological response Charlotte Charpentier1–3*, Guinevere Q. Lee4,5, Christophe Rodriguez6,7, Benoit Visseaux1–3, Alexandre Storto3, Catherine Fagard8,9, Jean-Michel Molina10, Christine Katlama11, Yazdan Yazdanpanah1,2,12, P. Richard Harrigan4,5 and Diane Descamps1–3 1
*Corresponding author. Hoˆpital Bichat-Claude Bernard, Laboratoire de Virologie, 46 Rue Henri Huchard, 75018 Paris, France. Tel: +33-1-40-25-61-50; Fax: +33-1-40-25-67-69; E-mail: [email protected]
Received 7 November 2014; returned 22 December 2014; revised 5 February 2015; accepted 6 February 2015 Objectives: To assess the prevalence of minority resistant variants (MRVs) at baseline and their impact on the virological response. The ANRS 139 TRIO trial evaluated the combination of raltegravir, etravirine and darunavir, plus an optimized background therapy, in 87% of cases. Patients were highly experienced and harboured multiresistant viruses, but were naive to the three drugs, and showed a high level of virological suppression. Methods: Ultra-deep sequencing of reverse transcriptase, protease and integrase regions was performed at the trial baseline, and sequences were interpreted according to the ANRS algorithm. MRVs were assessed using MiSeq and 454 technologies (limit of detection 1%). Results: At baseline, minority variants with at least one NRTI, one NNRTI, one PI, one major PI or an integrase inhibitor resistance-associated mutation were present in 46%, 45%, 68%, 24% and 13% of patients, respectively. When minority variants are taken into account, the prevalence of resistance to etravirine, darunavir and raltegravir at baseline was 29%, 40% and 9%, respectively. No difference was observed in the prevalence of MRVs between patients with virological failure and those with virological success, except a trend for patients exhibiting baseline etravirine MRVs (50% versus 26%, P ¼ 0.09). Conclusions: We have shown a high level of MRVs at baseline in highly pre-treated patients harbouring multiresistant viruses. However, these MRVs were not associated with an increased risk of virological failure, except for a trend for etravirine MRVs. Keywords: HIV, quasispecies, ultra-deep sequencing
Introduction The newly developed ultra-deep sequencing (UDS) technologies have allowed the detection of viral quasispecies down to a proportion of 1%, compared with a threshold of 20% obtained by population sequencing, which is the current standard genotyping technology used during the follow-up of HIV-infected patients. Viral variants representing ,20% of quasispecies, and therefore not detected by population sequencing, are considered to be minority variants.
Several studies, conducted in antiretroviral-naive patients, have shown a 2- or 3-fold increase in the prevalence of transmitted drug resistance when detecting minority variants using the UDS technologies.1 – 5 Regarding their impact on virological response, a large meta-analysis and also a recent European study have shown that the presence, at baseline, of minority variants exhibiting resistance-associated mutations (RAMs) to NNRTIs was associated with a 2.5–3-fold increased risk of virological failure (VF) in patients initiating a first-line regimen that contained an NNRTI, independently of adherence level.6,7 Thus, for the NNRTI drug class, it is
# The Author 2015. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please e-mail: [email protected]
1 of 7
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Francisco on March 24, 2015
INSERM, IAME, UMR 1137, F-75018 Paris, France; 2Universite´ Paris Diderot, IAME, UMR 1137, Sorbonne Paris Cite´, F-75018 Paris, France; AP-HP, Hoˆpital Bichat-Claude Bernard, Laboratoire de Virologie, F-75018 Paris, France; 4BC Centre for Excellence in HIV/AIDS, Vancouver, BC, Canada; 5University of British Columbia, Vancouver, BC, Canada; 6Department of Virology, Hoˆpital Henri Mondor, Universite´ Paris-Est, Cre´teil, France; 7INSERM U955 Team 18, Cre´teil, France; 8INSERM, ISPED, Centre INSERM U897—Epidemiologie-Biostatistique, F-33000 Bordeaux, France; 9Universite´ Bordeaux, ISPED, Centre INSERM U897—Epidemiologie-Biostatistique, F-33000 Bordeaux, France; 10AP-HP, Hoˆpital Saint-Louis, Service de Maladies Infectieuses et Tropicales, INSERM U941, Universite´ Paris Diderot, Paris, France; 11AP-HP, Hoˆpital de la Pitie´-Salpe´trie`re, Service de Maladies Infectieuses et Tropicales, F-75013 Paris, France; 12AP-HP, Hoˆpital Bichat, Service de Maladies Infectieuses et Tropicales, F-75018 Paris, France 3
Charpentier et al.
Patients and methods A total of 103 HIV-1-infected patients, receiving ART for a median duration of 13 years (IQR 11 – 15), were included in the ANRS 139 TRIO trial.8 Written informed consent was obtained from all patients before inclusion in the ANRS 139 TRIO trial.10 UDS was performed at baseline on the reverse transcriptase (RT), protease and integrase regions. Briefly, HIV RNA was extracted from 1 mL of both plasma samples and the control 8E5/LAV subclone using automated extraction methods with a NucliSENS easyMAG (bioMe´rieux). One-step RT– PCR was performed in triplicate. A second-round PCR amplification was then performed using primers as previously described.11 The PCR product was purified using AMPure (Agencourt) according to the manufacturer’s protocol, then quantified with the Quant-iT PicoGreen dsDNA Assay Kit (Invitogren) and DTX 800 Multimode Detector (Beckman Coulter). Each amplification, diluted to a concentration of 2×1012 DNA molecules, was tagmented by the Nextera XT transposome (Illumina) before combining. This combined set of PCR products then underwent emulsion PCR and deep sequencing was performed with MiSeq (Illumina, San Diego, CA, USA) and the GS20 technology (454 Life Sciences, Roche, Branford, CT, USA) platform.11,12 UDS generated read-lengths of .250 bp of data in each direction. Too-short sequences and truncated reads were excluded from the analysis. UDS generated a mean of .3000 sequences/sample. The quality of sequences generated was tested by means of a modified statistical test based on the binomial law13 to eliminate sequences that were too rare and/or of poor quality, probably as a result of sequencing errors.14 This latter analysis combined with the analysis of the 8E5 subclone control sequences showed different thresholds of detection between the two UDS technologies, and we decided to choose the limit of detection of 1% for both technologies. According to these criteria, UDS results were available in 57, 84, 83 and 87 patients for the NRTI, NNRTI, PI and INI drug classes, respectively. The lower number of UDS results we obtained for the RT region, especially the region that assessed NRTI RAMs, was the result of a higher rate of amplification failure observed with a specific amplicon from the former GS20 technology. Paired GS20-454 and MiSeq methods were performed only in RT and integrase regions and results were obtained in n ¼ 79 RT and n ¼ 59 integrase samples. UDS results were interpreted using the ANRS resistance algorithm (www.hivfrenchresistance.org). In our analyses, viruses exhibiting either possible resistance or full resistance to a drug were all considered resistant. The major PI resistance mutations were identified using the 2014 IAS-USA drug resistance mutation list.15
2 of 7
Mutational load was calculated at baseline based on the percentage of viral quasispecies by UDS and plasma VL. Viral quasispecies present at a proportion ≥20% were considered to be major quasispecies and quasispecies present at a proportion ,20% were considered to be minor quasispecies. Comparisons between the proportions of patients harbouring MRVs with VF and those with virological success (VS) were assessed using Fisher’s exact test (and StatEL software).
Results Prevalence of minority variants harbouring drug resistance mutations at baseline in the ANRS 139 TRIO trial Among the 103 patients included in the ANRS 139 TRIO trial, baseline median VL was 4.2 log10 copies/mL (IQR 3.6 – 4.6) and UDS results were available for 57, 84, 83 and 87 samples for the NRTI, NNRTI, PI and INI drug classes, respectively. The overall concordance of the observed percentages of minority variants between the two UDS methods was good: r 2 ¼ 0.886 in RT sequences and r 2 ¼ 0.940 in integrase sequences. Using the 20% threshold to detect viral variants, we found that 68%, 39% and 100% of the patients displayed plasma viruses with at least one NRTI, NNRTI and major PI RAM, respectively. The median number of RAMs was 6, 1 and 4 for NRTI, NNRTI and major PI drug classes, respectively. Regarding the INI drug class, three patients (3%) displayed viruses with INI RAMs: two had the E157Q natural polymorphism and one had the L74I secondary mutation. Using the 1% threshold, 68%, 57% and 100% of the patients displayed plasma viruses with at least one NRTI, NNRTI and major PI RAM, respectively. The median number of RAMs was 7, 2 and 4 for NRTI, NNRTI and major PI drug classes, respectively. Minority variants exhibiting at least one NRTI, NNRTI, major PI and all PI RAM were detected in 26 (46%), 38 (45%), 20 (24%) and 57 (69%) of the patients, respectively. The most prevalent RAM minority variants were T215F/Y, V179I/L/T and M46I/L for the NRTI, NNRTI and PI drug classes, respectively (Figure 1). Overall, median mutational load was 531 copies/mL (IQR 170–1969). By drug class, median mutational load was 511 (IQR 197–1666), 649 (IQR 259–1977) and 713 (IQR 311–1000) copies/mL for NRTI, NNRTI and major PI RAMs, respectively (Figure 1). Regarding the INI drug class, 13 INI RAMs were detected in minority viral variants issued from 11 patients (13%) (Figure 1). Key INI RAMs were detected in eight cases, including the Q148K in five cases, always in a proportion below 2%. The five remaining INI RAMs detected were secondary mutations. The median mutational load was 145 copies/mL (IQR 119–422). The complete UDS results for all four regions assessed were available in 57 patients; among these, 48 (84%) displayed minority variants with at least one RAM. Fourteen (29%), 23 (48%) and 11 (23%) patients displayed viruses harbouring at least one RAM for one, two and three different drug classes, respectively. None of the patients exhibited minority viruses harbouring RAMs to the four drug classes tested. Regarding the three main drugs included in the ANRS 139 TRIO trial, the prevalences of minority variants with at least one RAM to these drugs are depicted in Table 1. Minority variants exhibiting at least one etravirine RAM were detected in 27 patients (32%). The most prevalent etravirine RAMs detected as being minority
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Francisco on March 24, 2015
now well described that the presence of minority resistant variants (MRVs) at baseline can jeopardize the virological response. However, in antiretroviral-treated patients with VF, few data are available on the prevalence of MRVs or the impact of virological response to a subsequent antiretroviral regimen. A 96-week Phase II non-comparative trial, the Agence nationale de recherches sur le SIDA et les he´patites virales (ANRS) 139 TRIO trial, evaluated the integrase inhibitor (INI) raltegravir, the NNRTI etravirine and the PI darunavir, boosted with ritonavir combination in 103 highly experienced patients harbouring multiresistant viruses. Eighty-seven percent of patients in the trial also received an NRTI and/or an enfuvirtide-containing optimized background therapy. A high level of virological suppression was observed: 93% of patients had a plasma viral load (VL) of ,50 copies/mL by week 48 and 88% by week 96.8,9 The aim of this virological substudy of the ANRS 139 TRIO trial was to assess the prevalence of MRVs at baseline and their impact on virological response.
JAC
Minority quasispecies in the ANRS 139 TRIO trial
0I V9
K7 0
S 55
F N1
S1
48
53
K
S Q1
H
43 Y1
L7
8E T7 V8 4P 2A /S /T I8 4V L9 0M
4L
Q5
I5
0L I5
8V
L
7V I4
I/
G4
M
46
2I V3
0N D3
43
1 0
0
Y1
2
1
C
3
2
40
3
G1
4
4M
4
K
INI RAMs 5
38
N
R
D6 7
K6 5
M 41
T6 9
(d)
5
Figure 1. Distribution of HIV-1 minority variants exhibiting RAMs for: (a) NRTIs; (b) NNRTIs; (c) PIs; and (d) INIs. The data are expressed according to resistant VL: proportion of minority variants/HIV-1 plasma VL. Regarding the PI drug class, only the major PI RAMs, as defined by the IAS-USA list, are shown.15 Table 1. Description of the baseline minority variants with a drug resistance mutation for the three major drugs used in the ANRS 139 TRIO trial
Patients with minority variants harbouring DRMs, n (%) Proportion of minority variants harbouring DRMs, median % (IQR) Resistant VL, median copies/mL (IQR)
Etravirine (n ¼84)
Darunavir (n ¼83)
Raltegravir (n¼87)
27 (32) 4.4 (2.0 –7.9) 560 (195 –1644)
22 (27) 6.2 (2.9–10.4) 713 (342–1268)
6 (7) 1.7 (1.3 –2.6) 128 (108 –161)
DRMs, drug resistance mutations.
variants were V179I/L/T (n ¼ 15), K101P/R (n ¼ 7), V106I (n ¼ 7) and E138A/G/K/Q/R (n ¼ 7). The median proportion of minority variants exhibiting an etravirine RAM was 4.4% (IQR 2.0–7.9) and the median mutational load was 560 copies/mL (IQR 195–1644). Minority variants exhibiting at least one darunavir RAM were detected in 22 patients (27%) with a median proportion of 6.2% (IQR 2.9–10.4). The most prevalent darunavir RAMs detected as minority variants were V11I (n¼12), I54L (n¼4) and L89V (n¼4). The median mutational load was 713 copies/mL (IQR¼342–1268). Minority variants exhibiting at least one raltegravir RAM were detected in six patients (7%). Among them, five exhibited Q148Kmutated minority viruses with a proportion between 1% and 2%. The median mutational load was 128 copies/mL (IQR¼108–161).
Prevalence of minority variants interpreted as resistant at baseline in the ANRS 139 TRIO trial The prevalence of drug resistance according to the level of detection of viral variants is depicted in Figure 2. For the NRTI drug class and using the 20% threshold to detect viral variants, 89%, 95% and 77% of patients displayed viruses resistant to
lamivudine/emtricitabine, abacavir and tenofovir, respectively. Using the 1% threshold to detect viral variants, the prevalence of resistance to lamivudine/emtricitabine, abacavir and tenofovir increased to 93%, 96% and 82%, respectively. Regarding the NNRTI drug class, 46%, 17%, 50% and 38% of patients displayed viruses resistant to efavirenz, etravirine, nevirapine and rilpivirine, respectively, with a 20% threshold. These prevalences increased to 54%, 29%, 56% and 46% when using the 1% threshold. For the PI drug class, 23% and 33% of patients displayed viruses resistant to darunavir and tipranavir, respectively, with the 20% threshold; the prevalences then increased to 40% and 36%, respectively, when using the 1% threshold. For the INI drug class, the 20% threshold showed that 2%, 2% and 0% of patients displayed viruses resistant to raltegravir, elvitegravir and dolutegravir, respectively. The prevalence of resistance to raltegravir, elvitegravir and dolutegravir increased to 9%, 11% and 2% when using the 1% threshold.
MRVs and virological response During the follow-up period, until week 96, 19 of the 103 patients exhibited VF: 8 at week 24; 4 between weeks 24 and 48; and
3 of 7
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Francisco on March 24, 2015
Major PI RAMs
K1 /R 03 N V1 /S E1 38 06A A/ /I G/ K/ V1 Q/R 79 I/ L/ Y1 T 81 C/ I Y1 G1 88 L 90 A/ E/ S/ T
0 0I
1
0
1P
1
K1 0
2
L1 0
3
2
A9
3
E/ R L7 4I /V V7 5M Y1 1 M 5F 18 4I T2 /V 15 L 2 C/ D/ 10W F/ H/ L/ V K2 /Y 19 E/ Q
4
D/ N
4
8G
NNRTI RAMs 5
(c) HIV-1 VL (log10 copies/mL)
(b)
E1
NRTI RAMs 5
L
HIV-1 VL (log10 copies/mL)
(a)
Proportion of patients with MRVs (%)
Charpentier et al.
100 90 80 70 60 50 40 30 20 10 0
Proportion of variant >20% Proportion of variant >5% Proportion of variant >1%
EFV
NVP
ETR
DRV
TPV
RAL
PI
EVG
DTG 3TC/FTC TDF
INI
ABC
NRTI
Figure 2. Prevalence of antiretroviral drug resistance at baseline in the ANRS 139 TRIO trial according to the proportion of the resistant variants (thresholds: 20%, 5% and 1%). 3TC/FTC, lamivudine/emtricitabine; ABC, abacavir; DRV, darunavir; DTG, dolutegravir; EFV, efavirenz; ETR, etravirine; EVG, elvitegravir; NVP, nevirapine; RAL, raltegravir; RPV, rilpivirine; TDF, tenofovir; TPV, tipranavir.
Proportion of patients (%)
60 50
P = 0.09 P = 0.03
6/12
5/12 40 30
P = 0.8 6/14
26/67
P = 0.5 4/14
18/70
14/67
20
P = 0.3
9/70
P=1
8/71
10 0/14
0 Etravirine
Darunavir
0/14
2/71
Raltegravir
Patients with VF exhibiting baseline resistant viruses (minority and majority proportions) Patients with VF exhibiting baseline resistant viruses present in majority proportions Patients with VS exhibiting baseline resistant viruses (minority and majority proportions) Patients with VS exhibiting baseline resistant viruses present in majority proportions Figure 3. Proportion of patients harbouring resistant viruses at baseline in the ANRS 139 TRIO trial and who had VS (grey bars) or VF (black bars). Viral quasispecies present at a proportion ≥20% were considered as major quasispecies and quasispecies present at a proportion ,20% were considered as minor quasispecies.
7 between weeks 48 and 96, including two patients whose data were missing.9 In our analyses, we excluded the two latter patients, who had no assessable virological response. The patients with VF more frequently displayed etravirineresistant variants at a .20% proportion at baseline compared with those with VS [n ¼ 5/12 (42%) versus n ¼ 9/70 (13%); P ¼ 0.03; Figure 3]. When MRVs were included in the analyses, we observed a trend towards a higher proportion of etravirine MRVs in patients experiencing VF than in those experiencing VS [n¼ 6/12 (50%) versus n ¼ 18/70 (26%) in the VF and VS groups, respectively; P ¼ 0.09]. The median mutational load of minority etravirine RAMs did not differ between patients with VF and those with VS [473 (IQR ¼ 177 – 1786) versus 1195 (IQR 486 – 1514), respectively; P¼0.37]. The proportion of patients harbouring darunavir MRVs was similar between patients with VF and those with VS, when only major variants were analysed [i.e. 4/14 (29%) versus 14/67 (21%) in the VF and VS groups, respectively; P ¼ 0.5] or when
4 of 7
both major and minority variants were analysed [i.e. 6/14 (43%) versus 26/67 (39%) in the VF and VS groups, respectively; P ¼ 0.8]. No analysis regarding median mutational load for minority darunavir RAMs was performed as a minority of darunavir RAMs were observed in only two patients with VF. Among the eight patients harbouring raltegravir-resistant viruses at baseline, as well as majority or minority variants, all experienced VS during the trial’s follow-up period. Dual-class-resistant viruses resistant to both etravirine and darunavir were found in three patients using the 20% threshold, and in nine other patients using the 1% threshold. Among these 12 patients, 5 (42%) experienced VF, including three that displayed MRVs at baseline.
Longitudinal analysis of MRVs UDS results were available during the follow-up for 10 patients who had a detectable plasma VL measured at least once during
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Francisco on March 24, 2015
NNRTI
RPV
(b)
100%
100%
100%
100%
100%
90% 90%
70%
80%
100%
59.5%
100%
70%
60%
60% 50%
30.7% 21.4% 20.8%
50%
40%
40%
1.3% 8.8%
30%
30%
10.1%
20% 20%
0%
Pt C
W40 W48
W0
Pt E 100% 100%
W0
Pt G
100%
(c)
W48
Pt B
1.9%
Pt C
W96
Pt H
4.5%
Pt E
(d)
Pt F
100%
Pt H
100%
90%
G190S Y181I V179I E138G E138A K103N K101P L100I V90I
1.9%
W40
W24 W32
T215Y M184V V75M K70R M41L
W20 W24 W32
W0 W32
W0 W96
0%
W0 W40
2.5%
W24 W32 W40 W48 W72
1.6%
W0 W20 W32 W40 W72
10%
Pt J
100% 100% 100%
80% 70%
100%
96%
90%
60%
80%
50% 2.5%
30%
60%
37.2%
31.0%
5.9% 12.5%
50%
22.3%
40%
23.4% 8.1%
Pt E
4.3%
Pt G
Pt H
2.6% 3.2%
Pt I
W0 W20 W24 W32 W40
Pt D
20%
W0 W96
Pt C
30%
9.8%
W0 W96
Pt B
2.6% 32.6% 19.6% 12.2% 19.0% 26.0% 26.0%
W0 W48
W0 W20 W32 W40 W72
0%
34.6% 29.3%
W24 W32 W40 W48 W72
10%
22.1%
W0 W60 W72
20%
70%
3.6%
40%
W0 W32
Proportion of variants (%)
9.2% 11.2%
10%
W0 W32
Proportion of variants (%)
100% 80%
Minority quasispecies in the ANRS 139 TRIO trial
(a)
L8 I8 9V L 4V I5476V I5 M I5 4L I470V L3 V V3 3F V1 2I 1I
8.5%
10%
1.1%
0% W0 W20
W32W40 W72
Pt B
1.1% 1.6%
W24 W32 W40 W48
Pt E
1.3%
1.0% 4.8%
W72
W0 W40
Pt F
W0 W20 W24 W32
W40
S153F Q148K L74I T66K T66I
Pt J
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Francisco on March 24, 2015
5 of 7
Figure 4. Longitudinal analyses of viral quasispecies in patients with detectable plasma viraemia during the follow-up of the ANRS 139 TRIO trial. (a) All NRTI RAMs. (b) Etravirine RAMs. (c) Darunavir RAMs. (d) All INI RAMs. Open circles correspond to viral variants not detected by UDS (i.e. below 1%). Patients A to E: patients with VF according to the ANRS 139 TRIO trial criteria. Patients F and G displayed a viral blip. Patients H and I exhibited a detectable VL at the last visit (week 96) with no further control. Patient J stopped antiretroviral drug therapy at week 22. Pt, patient; W, week. This figure appears in colour in the online version of JAC and in black and white in the print version of JAC.
JAC
Pt J
1.2%
Charpentier et al.
Discussion In this virological substudy of the ANRS 139 TRIO trial, which included highly pretreated patients harbouring multiresistant viruses, we have shown a high level of MRVs at baseline. However their presence was not associated with an increased risk of VF, except for a trend in patients exhibiting etravirine MRVs at baseline. It has been well described that the use of UDS technologies increases, by 2 –3-fold, the rate of transmitted RAMs detected in antiretroviral-naive patients.1 – 5 The use of UDS technologies also results in the detection of a higher prevalence of RAMs at the time of VF.16 – 18 In the ANRS 139 TRIO trial, which enrolled patients treated with antiretroviral drugs over a median period of 13 years and who harboured multiresistant viruses, regarding the NNRTI drug class, the use of a 1% threshold as the limit of detection showed a 1.5-fold increase in the prevalence of NNRTI RAMs. Although the ANRS 139 TRIO trial was a salvage therapy trial taking into account the activity of three new drugs, one limitation of our study is that baseline NRTI resistance could be assessed in only slightly more than half of the cases. Regarding the three main antiretroviral drugs used in the ANRS 139 TRIO trial, at inclusion, 32%, 27% and 7% of patients displayed minority variants with at least one etravirine, darunavir and raltegravir RAM, respectively. This differential prevalence, which depends on drug class, is clearly associated with the use of previous drug classes. Similar findings were observed for the mutational load, with a lower level of INI-mutational load than NRTI-, NNRTIand PI-mutational loads. The studied patients had never been exposed to INIs, and raltegravir was the first compound in this drug class to be licensed. This occurred at the same time as the patients were included in the trial; thus, no major raltegravirresistant viruses should have been circulating at this time. Other studies also reported the presence of minority variants harbouring raltegravir resistance mutations in patients never exposed to INIs.19,20 Regarding the interpretation of drug resistance, we have shown an increase in the prevalence of resistance from 17% to
6 of 7
29% for etravirine, from 23% to 40% for darunavir and from 2% to 9% for raltegravir when decreasing the threshold of viralvariant detection from 20%, the limit of standard genotyping, to 1%, the limit obtained with UDS technologies. Of note, 17% and 23% of the patients displayed viruses resistant to etravirine and darunavir as the major viral population at baseline. Despite the high prevalence of MRVs at baseline, there was a high level of virological suppression in the ANRS 139 TRIO trial, with 88% of patients having a VL ,50 copies/mL at week 96.9 No difference could be observed in the prevalence of MRVs between patients with VF and those with VS, except for a trend in patients exhibiting etravirne MRVs. However, we can hypothesize that the low number of patients in VF observed in the ANRS 139 TRIO trial (n ¼ 17) limited the ability to observe a possible impact of minority variants on virological response and this low sample size could represent a limitation of the study. The impact of baseline MRVs on the optimized background therapy was not assessable, as the prevalence of baseline resistance by population sequencing to lamivudine/emtricitabine, abacavir and tenofovir, the most common drugs used in the optimized background therapy, was already as high as 89%, 95% and 77%, respectively. In our study, the presence of dual-class resistant viruses at baseline did not seem to impact to a greater extent the virological response compared with the presence of single-class resistant viruses. A recent study, based on 132 highly experienced patients harbouring multiresistant viruses of whom 28 experienced VF of their salvage therapy, showed that genotypic sensitivity scores based on UDS genotyping discriminated salvage ART outcomes better than those relying on population sequencing.21 In five patients experiencing VF, we had UDS results available at different timepoints during the follow-up; these showed the emergence of NNRTI and PI RAMs in two patients. We also showed the emergence of raltegravir RAMs, in a small proportion, i.e. in two patients. Thus, this longitudinal analysis showed that the dynamics of viral quasispecies evolution was dependent on the presence of specific selection pressures. In conclusion, this study showed that a high proportion of highly pretreated patients with multiresistant viruses present as a major viral population also displayed MRVs. However, the proportion of patients with baseline MRVs was not higher in patients experiencing VF when compared with those virologically suppressed, except for a trend for etravirine MRVs. Since the newest compounds licensed have been in the INI drug class, it would also be interesting to assess the impact of MRVs on the virological response to an elvitegravir- or dolutegravir-based regimen. Further analyses to assess the prevalence and impact of MRVs in the context of highly experienced patients harbouring multiresistant viruses are needed to determine what role, if any, the level of detection of MRVs can play in resistance assessments in this specific population.
Acknowledgements This work was presented in part at the Conference on Retroviruses and Opportunistic Infections, Boston, MA, USA, 2014 (Abstract 605). We thank the patients who participated in this trial and all the studies’ investigators, collaborators and participating centres. We thank: M. J. Commoy and S. Couffin-Cadiergues (ANRS); members of the Steering Committee and of the Data Safety Monitoring Board; and members of the study team (Inserm Bordeaux U897). We also thank Merck Sharp
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Francisco on March 24, 2015
the trial. Among the 10 patients, 5 experienced VF, according to the ANRS 139 TRIO trial criteria, 2 displayed a viral blip, 2 exhibited detectable VL at the last visit (week 96) with no further control and the other patient stopped antiretroviral drug therapy at week 22. UDS results were available at the trial’s baseline for all patients but one. Among the 10 patients assessed, UDS results for NRTI, NNRTI, PI and INI drug classes were available for 4, 7, 9 and 10 patients, respectively. The longitudinal evolution of quasispecies harbouring RAMs is depicted in Figure 4. Regarding the five patients with VF, the emergence of raltegravir RAM was detected in two patients among the five assessed [L74I (8.5%) and Q148K (1.6%)], and the emergence of etravirine RAMs was observed in the two assessable patients (two major RAMs and one minor variant). The emergence of eight darunavir RAMs was observed in two of the four patients assessed (two major RAMs and six minor variants). Interestingly, we observed the disappearance of darunavir RAMs during the follow-up in all four patients (n¼ 8; six major and two minor variants). Finally, during the follow-up, NRTI RAMs were no longer detected in two of the patients assessed (n ¼4: three major and one minor variant).
Minority quasispecies in the ANRS 139 TRIO trial
JAC
and Dohme-Chibret for providing raltegravir, and Tibotec (a division of Janssen-Cilag) for providing etravirine.
treatment-experienced patients infected with multidrug-resistant HIV: results of the ANRS 139 TRIO trial. Clin Infect Dis 2009; 49: 1441– 9.
Funding
9 Fagard C, Colin C, Charpentier C et al. Long-term efficacy and safety of raltegravir, etravirine, and darunavir/ritonavir in treatment-experienced patients: week 96 results from the ANRS 139 TRIO trial. J Acquir Immune Defic Syndr 2012; 59: 489–93.
The research leading to these results has received funding from the French National Agency for Research on AIDS and Viral Hepatitis (ANRS).
Transparency declarations None to declare.
11 Wang C, Mitsuya Y, Gharizadeh B, Ronaghi M, Shafer RW. Characterization of mutation spectra with ultra-deep pyrosequencing: application to HIV-1 drug resistance. Genome Res 2007; 17: 1195– 201. 12 Knapp DJ, McGovern RA, Poon AF et al. ‘Deep’ sequencing accuracy and reproducibility using Roche/454 technology for inferring co-receptor usage in HIV-1. PLoS One 2014; 9: e99508.
1 Johnson JA, Li JF, Wei X et al. Minority HIV-1 drug resistance mutations are present in antiretroviral treatment-naı¨ve populations and associate with reduced treatment efficacy. PLoS Med 2008; 5: e158.
13 Campbell PJ, Pleasance ED, Stephens PJ et al. Subclonal phylogenetic structures in cancer revealed by ultra-deep sequencing. Proc Natl Acad Sci USA 2008; 105: 13081– 6.
2 Lataillade M, Chiarella J, Yang R et al. Prevalence and clinical significance of HIV drug resistance mutations by ultra-deep sequencing in antiretroviral-naı¨ve subjects in the CASTLE study. PLoS One 2010; 5: e10952.
14 Rodriguez C, Chevaliez S, Bensadoun P, Pawlotsky JM. Characterization of the dynamics of hepatitis B virus resistance to adefovir by ultra-deep pyrosequencing. Hepatology 2013; 58: 890– 901.
3 Xiaobai Z, Xi C, Tian H et al. Prevalence of WHO transmitted drug resistance mutations by deep sequencing in antiretroviral-naı¨ve subjects in Hunan Province, China. PLoS One 2014; 9: e98740. 4 Cunningham ES, Bibby D, Lythgou K et al. Enhanced surveillance of HIV-1 transmitted drug resistance and transmission clusters in recently infected UK MSM. In: Abstracts of the International Workshop on Antiviral Drug Resistance: Meeting the Global Challenge, Berlin, 2014. Abstract 2. 5 Todesco E, Wirden M, Charpentier C et al. Transmitted drug resistance in MSM is five times more frequent with ultra deep sequencing than with classical sequencing. In: Abstracts of the International Workshop on Antiviral Drug Resistance: Meeting the Global Challenge, Berlin, 2014. Abstract 28. 6 Li JZ, Paredes R, Ribaudo HJ et al. Low-frequency HIV-1 drug resistance mutations and risk of NNRTI-based antiretroviral treatment failure: a systematic review and pooled analysis. JAMA 2011; 305: 1327 –35. 7 Cozzi-Lepri A, Noguera-Julian M, Di Giallonardo F et al. Low-frequency drug-resistant HIV-1 and risk of virological failure to first-line NNRTIbased ART: a multicohort European case– control study using centralized ultrasensitive 454 pyrosequencing. J Antimicrob Chemother 2015; 70: 930–40. 8 Yazdanpanah Y, Fagard C, Descamps D et al. High rate of virologic suppression with raltegravir plus etravirine and darunavir/ritonavir among
15 Wensing AM, Calvez V, Gu¨nthard HF et al. 2014 Update of the drug resistance mutations in HIV-1. Top Antivir Med 2014; 22: 642–50. 16 Lataillade M, Chiarella J, Yang R et al. Virologic failures on initial boosted-PI regimen infrequently possess low-level variants with major PI resistance mutations by ultra-deep sequencing. PLoS One 2012; 7: e30118. 17 Fisher R, van Zyl GU, Travers SA et al. Deep sequencing reveals minor protease resistance mutations in patients failing a protease inhibitor regimen. J Virol 2012; 86: 6231 –7. 18 Todesco E, Rodriguez C, Morand-Joubert L et al. Improved detection of resistance at failure to a tenofovir, emtricitabine and efavirenz regimen by ultradeep sequencing. J Antimicrob Chemother 2015; doi:10.1093/jac/ dku557. 19 Charpentier C, Laureillard D, Piketty C et al. High frequency of integrase Q148R minority variants in HIV-infected patients naive of integrase inhibitors. AIDS 2010; 24: 867– 73. 20 Liu J, Miller MD, Danovich RM et al. Analysis of low-frequency mutations associated with drug resistance to raltegravir before antiretroviral treatment. Antimicrob Agents Chemother 2011; 55: 1114– 9. 21 Pou C, Noguera-Julian M, Pe´rez-A´lvarez S et al. Improved prediction of salvage antiretroviral therapy outcomes using ultrasensitive HIV-1 drug resistance testing. Clin Infect Dis 2014; 59: 578– 88.
7 of 7
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Francisco on March 24, 2015
References
10 Charpentier C, Roquebert B, Colin C et al. Resistance analyses in highly experienced patients failing raltegravir, etravirine and darunavir/ritonavir regimen. AIDS 2010; 24: 2651 –6.
J Antimicrob Chemother doi:10.1093/jac/dkv048
Highly frequent HIV-1 minority resistant variants at baseline of the ANRS 139 TRIO trial had a limited impact on virological response Charlotte Charpentier1–3*, Guinevere Q. Lee4,5, Christophe Rodriguez6,7, Benoit Visseaux1–3, Alexandre Storto3, Catherine Fagard8,9, Jean-Michel Molina10, Christine Katlama11, Yazdan Yazdanpanah1,2,12, P. Richard Harrigan4,5 and Diane Descamps1–3 1
*Corresponding author. Hoˆpital Bichat-Claude Bernard, Laboratoire de Virologie, 46 Rue Henri Huchard, 75018 Paris, France. Tel: +33-1-40-25-61-50; Fax: +33-1-40-25-67-69; E-mail: [email protected]
Received 7 November 2014; returned 22 December 2014; revised 5 February 2015; accepted 6 February 2015 Objectives: To assess the prevalence of minority resistant variants (MRVs) at baseline and their impact on the virological response. The ANRS 139 TRIO trial evaluated the combination of raltegravir, etravirine and darunavir, plus an optimized background therapy, in 87% of cases. Patients were highly experienced and harboured multiresistant viruses, but were naive to the three drugs, and showed a high level of virological suppression. Methods: Ultra-deep sequencing of reverse transcriptase, protease and integrase regions was performed at the trial baseline, and sequences were interpreted according to the ANRS algorithm. MRVs were assessed using MiSeq and 454 technologies (limit of detection 1%). Results: At baseline, minority variants with at least one NRTI, one NNRTI, one PI, one major PI or an integrase inhibitor resistance-associated mutation were present in 46%, 45%, 68%, 24% and 13% of patients, respectively. When minority variants are taken into account, the prevalence of resistance to etravirine, darunavir and raltegravir at baseline was 29%, 40% and 9%, respectively. No difference was observed in the prevalence of MRVs between patients with virological failure and those with virological success, except a trend for patients exhibiting baseline etravirine MRVs (50% versus 26%, P ¼ 0.09). Conclusions: We have shown a high level of MRVs at baseline in highly pre-treated patients harbouring multiresistant viruses. However, these MRVs were not associated with an increased risk of virological failure, except for a trend for etravirine MRVs. Keywords: HIV, quasispecies, ultra-deep sequencing
Introduction The newly developed ultra-deep sequencing (UDS) technologies have allowed the detection of viral quasispecies down to a proportion of 1%, compared with a threshold of 20% obtained by population sequencing, which is the current standard genotyping technology used during the follow-up of HIV-infected patients. Viral variants representing ,20% of quasispecies, and therefore not detected by population sequencing, are considered to be minority variants.
Several studies, conducted in antiretroviral-naive patients, have shown a 2- or 3-fold increase in the prevalence of transmitted drug resistance when detecting minority variants using the UDS technologies.1 – 5 Regarding their impact on virological response, a large meta-analysis and also a recent European study have shown that the presence, at baseline, of minority variants exhibiting resistance-associated mutations (RAMs) to NNRTIs was associated with a 2.5–3-fold increased risk of virological failure (VF) in patients initiating a first-line regimen that contained an NNRTI, independently of adherence level.6,7 Thus, for the NNRTI drug class, it is
# The Author 2015. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please e-mail: [email protected]
1 of 7
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Francisco on March 24, 2015
INSERM, IAME, UMR 1137, F-75018 Paris, France; 2Universite´ Paris Diderot, IAME, UMR 1137, Sorbonne Paris Cite´, F-75018 Paris, France; AP-HP, Hoˆpital Bichat-Claude Bernard, Laboratoire de Virologie, F-75018 Paris, France; 4BC Centre for Excellence in HIV/AIDS, Vancouver, BC, Canada; 5University of British Columbia, Vancouver, BC, Canada; 6Department of Virology, Hoˆpital Henri Mondor, Universite´ Paris-Est, Cre´teil, France; 7INSERM U955 Team 18, Cre´teil, France; 8INSERM, ISPED, Centre INSERM U897—Epidemiologie-Biostatistique, F-33000 Bordeaux, France; 9Universite´ Bordeaux, ISPED, Centre INSERM U897—Epidemiologie-Biostatistique, F-33000 Bordeaux, France; 10AP-HP, Hoˆpital Saint-Louis, Service de Maladies Infectieuses et Tropicales, INSERM U941, Universite´ Paris Diderot, Paris, France; 11AP-HP, Hoˆpital de la Pitie´-Salpe´trie`re, Service de Maladies Infectieuses et Tropicales, F-75013 Paris, France; 12AP-HP, Hoˆpital Bichat, Service de Maladies Infectieuses et Tropicales, F-75018 Paris, France 3
Charpentier et al.
Patients and methods A total of 103 HIV-1-infected patients, receiving ART for a median duration of 13 years (IQR 11 – 15), were included in the ANRS 139 TRIO trial.8 Written informed consent was obtained from all patients before inclusion in the ANRS 139 TRIO trial.10 UDS was performed at baseline on the reverse transcriptase (RT), protease and integrase regions. Briefly, HIV RNA was extracted from 1 mL of both plasma samples and the control 8E5/LAV subclone using automated extraction methods with a NucliSENS easyMAG (bioMe´rieux). One-step RT– PCR was performed in triplicate. A second-round PCR amplification was then performed using primers as previously described.11 The PCR product was purified using AMPure (Agencourt) according to the manufacturer’s protocol, then quantified with the Quant-iT PicoGreen dsDNA Assay Kit (Invitogren) and DTX 800 Multimode Detector (Beckman Coulter). Each amplification, diluted to a concentration of 2×1012 DNA molecules, was tagmented by the Nextera XT transposome (Illumina) before combining. This combined set of PCR products then underwent emulsion PCR and deep sequencing was performed with MiSeq (Illumina, San Diego, CA, USA) and the GS20 technology (454 Life Sciences, Roche, Branford, CT, USA) platform.11,12 UDS generated read-lengths of .250 bp of data in each direction. Too-short sequences and truncated reads were excluded from the analysis. UDS generated a mean of .3000 sequences/sample. The quality of sequences generated was tested by means of a modified statistical test based on the binomial law13 to eliminate sequences that were too rare and/or of poor quality, probably as a result of sequencing errors.14 This latter analysis combined with the analysis of the 8E5 subclone control sequences showed different thresholds of detection between the two UDS technologies, and we decided to choose the limit of detection of 1% for both technologies. According to these criteria, UDS results were available in 57, 84, 83 and 87 patients for the NRTI, NNRTI, PI and INI drug classes, respectively. The lower number of UDS results we obtained for the RT region, especially the region that assessed NRTI RAMs, was the result of a higher rate of amplification failure observed with a specific amplicon from the former GS20 technology. Paired GS20-454 and MiSeq methods were performed only in RT and integrase regions and results were obtained in n ¼ 79 RT and n ¼ 59 integrase samples. UDS results were interpreted using the ANRS resistance algorithm (www.hivfrenchresistance.org). In our analyses, viruses exhibiting either possible resistance or full resistance to a drug were all considered resistant. The major PI resistance mutations were identified using the 2014 IAS-USA drug resistance mutation list.15
2 of 7
Mutational load was calculated at baseline based on the percentage of viral quasispecies by UDS and plasma VL. Viral quasispecies present at a proportion ≥20% were considered to be major quasispecies and quasispecies present at a proportion ,20% were considered to be minor quasispecies. Comparisons between the proportions of patients harbouring MRVs with VF and those with virological success (VS) were assessed using Fisher’s exact test (and StatEL software).
Results Prevalence of minority variants harbouring drug resistance mutations at baseline in the ANRS 139 TRIO trial Among the 103 patients included in the ANRS 139 TRIO trial, baseline median VL was 4.2 log10 copies/mL (IQR 3.6 – 4.6) and UDS results were available for 57, 84, 83 and 87 samples for the NRTI, NNRTI, PI and INI drug classes, respectively. The overall concordance of the observed percentages of minority variants between the two UDS methods was good: r 2 ¼ 0.886 in RT sequences and r 2 ¼ 0.940 in integrase sequences. Using the 20% threshold to detect viral variants, we found that 68%, 39% and 100% of the patients displayed plasma viruses with at least one NRTI, NNRTI and major PI RAM, respectively. The median number of RAMs was 6, 1 and 4 for NRTI, NNRTI and major PI drug classes, respectively. Regarding the INI drug class, three patients (3%) displayed viruses with INI RAMs: two had the E157Q natural polymorphism and one had the L74I secondary mutation. Using the 1% threshold, 68%, 57% and 100% of the patients displayed plasma viruses with at least one NRTI, NNRTI and major PI RAM, respectively. The median number of RAMs was 7, 2 and 4 for NRTI, NNRTI and major PI drug classes, respectively. Minority variants exhibiting at least one NRTI, NNRTI, major PI and all PI RAM were detected in 26 (46%), 38 (45%), 20 (24%) and 57 (69%) of the patients, respectively. The most prevalent RAM minority variants were T215F/Y, V179I/L/T and M46I/L for the NRTI, NNRTI and PI drug classes, respectively (Figure 1). Overall, median mutational load was 531 copies/mL (IQR 170–1969). By drug class, median mutational load was 511 (IQR 197–1666), 649 (IQR 259–1977) and 713 (IQR 311–1000) copies/mL for NRTI, NNRTI and major PI RAMs, respectively (Figure 1). Regarding the INI drug class, 13 INI RAMs were detected in minority viral variants issued from 11 patients (13%) (Figure 1). Key INI RAMs were detected in eight cases, including the Q148K in five cases, always in a proportion below 2%. The five remaining INI RAMs detected were secondary mutations. The median mutational load was 145 copies/mL (IQR 119–422). The complete UDS results for all four regions assessed were available in 57 patients; among these, 48 (84%) displayed minority variants with at least one RAM. Fourteen (29%), 23 (48%) and 11 (23%) patients displayed viruses harbouring at least one RAM for one, two and three different drug classes, respectively. None of the patients exhibited minority viruses harbouring RAMs to the four drug classes tested. Regarding the three main drugs included in the ANRS 139 TRIO trial, the prevalences of minority variants with at least one RAM to these drugs are depicted in Table 1. Minority variants exhibiting at least one etravirine RAM were detected in 27 patients (32%). The most prevalent etravirine RAMs detected as being minority
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Francisco on March 24, 2015
now well described that the presence of minority resistant variants (MRVs) at baseline can jeopardize the virological response. However, in antiretroviral-treated patients with VF, few data are available on the prevalence of MRVs or the impact of virological response to a subsequent antiretroviral regimen. A 96-week Phase II non-comparative trial, the Agence nationale de recherches sur le SIDA et les he´patites virales (ANRS) 139 TRIO trial, evaluated the integrase inhibitor (INI) raltegravir, the NNRTI etravirine and the PI darunavir, boosted with ritonavir combination in 103 highly experienced patients harbouring multiresistant viruses. Eighty-seven percent of patients in the trial also received an NRTI and/or an enfuvirtide-containing optimized background therapy. A high level of virological suppression was observed: 93% of patients had a plasma viral load (VL) of ,50 copies/mL by week 48 and 88% by week 96.8,9 The aim of this virological substudy of the ANRS 139 TRIO trial was to assess the prevalence of MRVs at baseline and their impact on virological response.
JAC
Minority quasispecies in the ANRS 139 TRIO trial
0I V9
K7 0
S 55
F N1
S1
48
53
K
S Q1
H
43 Y1
L7
8E T7 V8 4P 2A /S /T I8 4V L9 0M
4L
Q5
I5
0L I5
8V
L
7V I4
I/
G4
M
46
2I V3
0N D3
43
1 0
0
Y1
2
1
C
3
2
40
3
G1
4
4M
4
K
INI RAMs 5
38
N
R
D6 7
K6 5
M 41
T6 9
(d)
5
Figure 1. Distribution of HIV-1 minority variants exhibiting RAMs for: (a) NRTIs; (b) NNRTIs; (c) PIs; and (d) INIs. The data are expressed according to resistant VL: proportion of minority variants/HIV-1 plasma VL. Regarding the PI drug class, only the major PI RAMs, as defined by the IAS-USA list, are shown.15 Table 1. Description of the baseline minority variants with a drug resistance mutation for the three major drugs used in the ANRS 139 TRIO trial
Patients with minority variants harbouring DRMs, n (%) Proportion of minority variants harbouring DRMs, median % (IQR) Resistant VL, median copies/mL (IQR)
Etravirine (n ¼84)
Darunavir (n ¼83)
Raltegravir (n¼87)
27 (32) 4.4 (2.0 –7.9) 560 (195 –1644)
22 (27) 6.2 (2.9–10.4) 713 (342–1268)
6 (7) 1.7 (1.3 –2.6) 128 (108 –161)
DRMs, drug resistance mutations.
variants were V179I/L/T (n ¼ 15), K101P/R (n ¼ 7), V106I (n ¼ 7) and E138A/G/K/Q/R (n ¼ 7). The median proportion of minority variants exhibiting an etravirine RAM was 4.4% (IQR 2.0–7.9) and the median mutational load was 560 copies/mL (IQR 195–1644). Minority variants exhibiting at least one darunavir RAM were detected in 22 patients (27%) with a median proportion of 6.2% (IQR 2.9–10.4). The most prevalent darunavir RAMs detected as minority variants were V11I (n¼12), I54L (n¼4) and L89V (n¼4). The median mutational load was 713 copies/mL (IQR¼342–1268). Minority variants exhibiting at least one raltegravir RAM were detected in six patients (7%). Among them, five exhibited Q148Kmutated minority viruses with a proportion between 1% and 2%. The median mutational load was 128 copies/mL (IQR¼108–161).
Prevalence of minority variants interpreted as resistant at baseline in the ANRS 139 TRIO trial The prevalence of drug resistance according to the level of detection of viral variants is depicted in Figure 2. For the NRTI drug class and using the 20% threshold to detect viral variants, 89%, 95% and 77% of patients displayed viruses resistant to
lamivudine/emtricitabine, abacavir and tenofovir, respectively. Using the 1% threshold to detect viral variants, the prevalence of resistance to lamivudine/emtricitabine, abacavir and tenofovir increased to 93%, 96% and 82%, respectively. Regarding the NNRTI drug class, 46%, 17%, 50% and 38% of patients displayed viruses resistant to efavirenz, etravirine, nevirapine and rilpivirine, respectively, with a 20% threshold. These prevalences increased to 54%, 29%, 56% and 46% when using the 1% threshold. For the PI drug class, 23% and 33% of patients displayed viruses resistant to darunavir and tipranavir, respectively, with the 20% threshold; the prevalences then increased to 40% and 36%, respectively, when using the 1% threshold. For the INI drug class, the 20% threshold showed that 2%, 2% and 0% of patients displayed viruses resistant to raltegravir, elvitegravir and dolutegravir, respectively. The prevalence of resistance to raltegravir, elvitegravir and dolutegravir increased to 9%, 11% and 2% when using the 1% threshold.
MRVs and virological response During the follow-up period, until week 96, 19 of the 103 patients exhibited VF: 8 at week 24; 4 between weeks 24 and 48; and
3 of 7
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Francisco on March 24, 2015
Major PI RAMs
K1 /R 03 N V1 /S E1 38 06A A/ /I G/ K/ V1 Q/R 79 I/ L/ Y1 T 81 C/ I Y1 G1 88 L 90 A/ E/ S/ T
0 0I
1
0
1P
1
K1 0
2
L1 0
3
2
A9
3
E/ R L7 4I /V V7 5M Y1 1 M 5F 18 4I T2 /V 15 L 2 C/ D/ 10W F/ H/ L/ V K2 /Y 19 E/ Q
4
D/ N
4
8G
NNRTI RAMs 5
(c) HIV-1 VL (log10 copies/mL)
(b)
E1
NRTI RAMs 5
L
HIV-1 VL (log10 copies/mL)
(a)
Proportion of patients with MRVs (%)
Charpentier et al.
100 90 80 70 60 50 40 30 20 10 0
Proportion of variant >20% Proportion of variant >5% Proportion of variant >1%
EFV
NVP
ETR
DRV
TPV
RAL
PI
EVG
DTG 3TC/FTC TDF
INI
ABC
NRTI
Figure 2. Prevalence of antiretroviral drug resistance at baseline in the ANRS 139 TRIO trial according to the proportion of the resistant variants (thresholds: 20%, 5% and 1%). 3TC/FTC, lamivudine/emtricitabine; ABC, abacavir; DRV, darunavir; DTG, dolutegravir; EFV, efavirenz; ETR, etravirine; EVG, elvitegravir; NVP, nevirapine; RAL, raltegravir; RPV, rilpivirine; TDF, tenofovir; TPV, tipranavir.
Proportion of patients (%)
60 50
P = 0.09 P = 0.03
6/12
5/12 40 30
P = 0.8 6/14
26/67
P = 0.5 4/14
18/70
14/67
20
P = 0.3
9/70
P=1
8/71
10 0/14
0 Etravirine
Darunavir
0/14
2/71
Raltegravir
Patients with VF exhibiting baseline resistant viruses (minority and majority proportions) Patients with VF exhibiting baseline resistant viruses present in majority proportions Patients with VS exhibiting baseline resistant viruses (minority and majority proportions) Patients with VS exhibiting baseline resistant viruses present in majority proportions Figure 3. Proportion of patients harbouring resistant viruses at baseline in the ANRS 139 TRIO trial and who had VS (grey bars) or VF (black bars). Viral quasispecies present at a proportion ≥20% were considered as major quasispecies and quasispecies present at a proportion ,20% were considered as minor quasispecies.
7 between weeks 48 and 96, including two patients whose data were missing.9 In our analyses, we excluded the two latter patients, who had no assessable virological response. The patients with VF more frequently displayed etravirineresistant variants at a .20% proportion at baseline compared with those with VS [n ¼ 5/12 (42%) versus n ¼ 9/70 (13%); P ¼ 0.03; Figure 3]. When MRVs were included in the analyses, we observed a trend towards a higher proportion of etravirine MRVs in patients experiencing VF than in those experiencing VS [n¼ 6/12 (50%) versus n ¼ 18/70 (26%) in the VF and VS groups, respectively; P ¼ 0.09]. The median mutational load of minority etravirine RAMs did not differ between patients with VF and those with VS [473 (IQR ¼ 177 – 1786) versus 1195 (IQR 486 – 1514), respectively; P¼0.37]. The proportion of patients harbouring darunavir MRVs was similar between patients with VF and those with VS, when only major variants were analysed [i.e. 4/14 (29%) versus 14/67 (21%) in the VF and VS groups, respectively; P ¼ 0.5] or when
4 of 7
both major and minority variants were analysed [i.e. 6/14 (43%) versus 26/67 (39%) in the VF and VS groups, respectively; P ¼ 0.8]. No analysis regarding median mutational load for minority darunavir RAMs was performed as a minority of darunavir RAMs were observed in only two patients with VF. Among the eight patients harbouring raltegravir-resistant viruses at baseline, as well as majority or minority variants, all experienced VS during the trial’s follow-up period. Dual-class-resistant viruses resistant to both etravirine and darunavir were found in three patients using the 20% threshold, and in nine other patients using the 1% threshold. Among these 12 patients, 5 (42%) experienced VF, including three that displayed MRVs at baseline.
Longitudinal analysis of MRVs UDS results were available during the follow-up for 10 patients who had a detectable plasma VL measured at least once during
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Francisco on March 24, 2015
NNRTI
RPV
(b)
100%
100%
100%
100%
100%
90% 90%
70%
80%
100%
59.5%
100%
70%
60%
60% 50%
30.7% 21.4% 20.8%
50%
40%
40%
1.3% 8.8%
30%
30%
10.1%
20% 20%
0%
Pt C
W40 W48
W0
Pt E 100% 100%
W0
Pt G
100%
(c)
W48
Pt B
1.9%
Pt C
W96
Pt H
4.5%
Pt E
(d)
Pt F
100%
Pt H
100%
90%
G190S Y181I V179I E138G E138A K103N K101P L100I V90I
1.9%
W40
W24 W32
T215Y M184V V75M K70R M41L
W20 W24 W32
W0 W32
W0 W96
0%
W0 W40
2.5%
W24 W32 W40 W48 W72
1.6%
W0 W20 W32 W40 W72
10%
Pt J
100% 100% 100%
80% 70%
100%
96%
90%
60%
80%
50% 2.5%
30%
60%
37.2%
31.0%
5.9% 12.5%
50%
22.3%
40%
23.4% 8.1%
Pt E
4.3%
Pt G
Pt H
2.6% 3.2%
Pt I
W0 W20 W24 W32 W40
Pt D
20%
W0 W96
Pt C
30%
9.8%
W0 W96
Pt B
2.6% 32.6% 19.6% 12.2% 19.0% 26.0% 26.0%
W0 W48
W0 W20 W32 W40 W72
0%
34.6% 29.3%
W24 W32 W40 W48 W72
10%
22.1%
W0 W60 W72
20%
70%
3.6%
40%
W0 W32
Proportion of variants (%)
9.2% 11.2%
10%
W0 W32
Proportion of variants (%)
100% 80%
Minority quasispecies in the ANRS 139 TRIO trial
(a)
L8 I8 9V L 4V I5476V I5 M I5 4L I470V L3 V V3 3F V1 2I 1I
8.5%
10%
1.1%
0% W0 W20
W32W40 W72
Pt B
1.1% 1.6%
W24 W32 W40 W48
Pt E
1.3%
1.0% 4.8%
W72
W0 W40
Pt F
W0 W20 W24 W32
W40
S153F Q148K L74I T66K T66I
Pt J
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Francisco on March 24, 2015
5 of 7
Figure 4. Longitudinal analyses of viral quasispecies in patients with detectable plasma viraemia during the follow-up of the ANRS 139 TRIO trial. (a) All NRTI RAMs. (b) Etravirine RAMs. (c) Darunavir RAMs. (d) All INI RAMs. Open circles correspond to viral variants not detected by UDS (i.e. below 1%). Patients A to E: patients with VF according to the ANRS 139 TRIO trial criteria. Patients F and G displayed a viral blip. Patients H and I exhibited a detectable VL at the last visit (week 96) with no further control. Patient J stopped antiretroviral drug therapy at week 22. Pt, patient; W, week. This figure appears in colour in the online version of JAC and in black and white in the print version of JAC.
JAC
Pt J
1.2%
Charpentier et al.
Discussion In this virological substudy of the ANRS 139 TRIO trial, which included highly pretreated patients harbouring multiresistant viruses, we have shown a high level of MRVs at baseline. However their presence was not associated with an increased risk of VF, except for a trend in patients exhibiting etravirine MRVs at baseline. It has been well described that the use of UDS technologies increases, by 2 –3-fold, the rate of transmitted RAMs detected in antiretroviral-naive patients.1 – 5 The use of UDS technologies also results in the detection of a higher prevalence of RAMs at the time of VF.16 – 18 In the ANRS 139 TRIO trial, which enrolled patients treated with antiretroviral drugs over a median period of 13 years and who harboured multiresistant viruses, regarding the NNRTI drug class, the use of a 1% threshold as the limit of detection showed a 1.5-fold increase in the prevalence of NNRTI RAMs. Although the ANRS 139 TRIO trial was a salvage therapy trial taking into account the activity of three new drugs, one limitation of our study is that baseline NRTI resistance could be assessed in only slightly more than half of the cases. Regarding the three main antiretroviral drugs used in the ANRS 139 TRIO trial, at inclusion, 32%, 27% and 7% of patients displayed minority variants with at least one etravirine, darunavir and raltegravir RAM, respectively. This differential prevalence, which depends on drug class, is clearly associated with the use of previous drug classes. Similar findings were observed for the mutational load, with a lower level of INI-mutational load than NRTI-, NNRTIand PI-mutational loads. The studied patients had never been exposed to INIs, and raltegravir was the first compound in this drug class to be licensed. This occurred at the same time as the patients were included in the trial; thus, no major raltegravirresistant viruses should have been circulating at this time. Other studies also reported the presence of minority variants harbouring raltegravir resistance mutations in patients never exposed to INIs.19,20 Regarding the interpretation of drug resistance, we have shown an increase in the prevalence of resistance from 17% to
6 of 7
29% for etravirine, from 23% to 40% for darunavir and from 2% to 9% for raltegravir when decreasing the threshold of viralvariant detection from 20%, the limit of standard genotyping, to 1%, the limit obtained with UDS technologies. Of note, 17% and 23% of the patients displayed viruses resistant to etravirine and darunavir as the major viral population at baseline. Despite the high prevalence of MRVs at baseline, there was a high level of virological suppression in the ANRS 139 TRIO trial, with 88% of patients having a VL ,50 copies/mL at week 96.9 No difference could be observed in the prevalence of MRVs between patients with VF and those with VS, except for a trend in patients exhibiting etravirne MRVs. However, we can hypothesize that the low number of patients in VF observed in the ANRS 139 TRIO trial (n ¼ 17) limited the ability to observe a possible impact of minority variants on virological response and this low sample size could represent a limitation of the study. The impact of baseline MRVs on the optimized background therapy was not assessable, as the prevalence of baseline resistance by population sequencing to lamivudine/emtricitabine, abacavir and tenofovir, the most common drugs used in the optimized background therapy, was already as high as 89%, 95% and 77%, respectively. In our study, the presence of dual-class resistant viruses at baseline did not seem to impact to a greater extent the virological response compared with the presence of single-class resistant viruses. A recent study, based on 132 highly experienced patients harbouring multiresistant viruses of whom 28 experienced VF of their salvage therapy, showed that genotypic sensitivity scores based on UDS genotyping discriminated salvage ART outcomes better than those relying on population sequencing.21 In five patients experiencing VF, we had UDS results available at different timepoints during the follow-up; these showed the emergence of NNRTI and PI RAMs in two patients. We also showed the emergence of raltegravir RAMs, in a small proportion, i.e. in two patients. Thus, this longitudinal analysis showed that the dynamics of viral quasispecies evolution was dependent on the presence of specific selection pressures. In conclusion, this study showed that a high proportion of highly pretreated patients with multiresistant viruses present as a major viral population also displayed MRVs. However, the proportion of patients with baseline MRVs was not higher in patients experiencing VF when compared with those virologically suppressed, except for a trend for etravirine MRVs. Since the newest compounds licensed have been in the INI drug class, it would also be interesting to assess the impact of MRVs on the virological response to an elvitegravir- or dolutegravir-based regimen. Further analyses to assess the prevalence and impact of MRVs in the context of highly experienced patients harbouring multiresistant viruses are needed to determine what role, if any, the level of detection of MRVs can play in resistance assessments in this specific population.
Acknowledgements This work was presented in part at the Conference on Retroviruses and Opportunistic Infections, Boston, MA, USA, 2014 (Abstract 605). We thank the patients who participated in this trial and all the studies’ investigators, collaborators and participating centres. We thank: M. J. Commoy and S. Couffin-Cadiergues (ANRS); members of the Steering Committee and of the Data Safety Monitoring Board; and members of the study team (Inserm Bordeaux U897). We also thank Merck Sharp
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Francisco on March 24, 2015
the trial. Among the 10 patients, 5 experienced VF, according to the ANRS 139 TRIO trial criteria, 2 displayed a viral blip, 2 exhibited detectable VL at the last visit (week 96) with no further control and the other patient stopped antiretroviral drug therapy at week 22. UDS results were available at the trial’s baseline for all patients but one. Among the 10 patients assessed, UDS results for NRTI, NNRTI, PI and INI drug classes were available for 4, 7, 9 and 10 patients, respectively. The longitudinal evolution of quasispecies harbouring RAMs is depicted in Figure 4. Regarding the five patients with VF, the emergence of raltegravir RAM was detected in two patients among the five assessed [L74I (8.5%) and Q148K (1.6%)], and the emergence of etravirine RAMs was observed in the two assessable patients (two major RAMs and one minor variant). The emergence of eight darunavir RAMs was observed in two of the four patients assessed (two major RAMs and six minor variants). Interestingly, we observed the disappearance of darunavir RAMs during the follow-up in all four patients (n¼ 8; six major and two minor variants). Finally, during the follow-up, NRTI RAMs were no longer detected in two of the patients assessed (n ¼4: three major and one minor variant).
Minority quasispecies in the ANRS 139 TRIO trial
JAC
and Dohme-Chibret for providing raltegravir, and Tibotec (a division of Janssen-Cilag) for providing etravirine.
treatment-experienced patients infected with multidrug-resistant HIV: results of the ANRS 139 TRIO trial. Clin Infect Dis 2009; 49: 1441– 9.
Funding
9 Fagard C, Colin C, Charpentier C et al. Long-term efficacy and safety of raltegravir, etravirine, and darunavir/ritonavir in treatment-experienced patients: week 96 results from the ANRS 139 TRIO trial. J Acquir Immune Defic Syndr 2012; 59: 489–93.
The research leading to these results has received funding from the French National Agency for Research on AIDS and Viral Hepatitis (ANRS).
Transparency declarations None to declare.
11 Wang C, Mitsuya Y, Gharizadeh B, Ronaghi M, Shafer RW. Characterization of mutation spectra with ultra-deep pyrosequencing: application to HIV-1 drug resistance. Genome Res 2007; 17: 1195– 201. 12 Knapp DJ, McGovern RA, Poon AF et al. ‘Deep’ sequencing accuracy and reproducibility using Roche/454 technology for inferring co-receptor usage in HIV-1. PLoS One 2014; 9: e99508.
1 Johnson JA, Li JF, Wei X et al. Minority HIV-1 drug resistance mutations are present in antiretroviral treatment-naı¨ve populations and associate with reduced treatment efficacy. PLoS Med 2008; 5: e158.
13 Campbell PJ, Pleasance ED, Stephens PJ et al. Subclonal phylogenetic structures in cancer revealed by ultra-deep sequencing. Proc Natl Acad Sci USA 2008; 105: 13081– 6.
2 Lataillade M, Chiarella J, Yang R et al. Prevalence and clinical significance of HIV drug resistance mutations by ultra-deep sequencing in antiretroviral-naı¨ve subjects in the CASTLE study. PLoS One 2010; 5: e10952.
14 Rodriguez C, Chevaliez S, Bensadoun P, Pawlotsky JM. Characterization of the dynamics of hepatitis B virus resistance to adefovir by ultra-deep pyrosequencing. Hepatology 2013; 58: 890– 901.
3 Xiaobai Z, Xi C, Tian H et al. Prevalence of WHO transmitted drug resistance mutations by deep sequencing in antiretroviral-naı¨ve subjects in Hunan Province, China. PLoS One 2014; 9: e98740. 4 Cunningham ES, Bibby D, Lythgou K et al. Enhanced surveillance of HIV-1 transmitted drug resistance and transmission clusters in recently infected UK MSM. In: Abstracts of the International Workshop on Antiviral Drug Resistance: Meeting the Global Challenge, Berlin, 2014. Abstract 2. 5 Todesco E, Wirden M, Charpentier C et al. Transmitted drug resistance in MSM is five times more frequent with ultra deep sequencing than with classical sequencing. In: Abstracts of the International Workshop on Antiviral Drug Resistance: Meeting the Global Challenge, Berlin, 2014. Abstract 28. 6 Li JZ, Paredes R, Ribaudo HJ et al. Low-frequency HIV-1 drug resistance mutations and risk of NNRTI-based antiretroviral treatment failure: a systematic review and pooled analysis. JAMA 2011; 305: 1327 –35. 7 Cozzi-Lepri A, Noguera-Julian M, Di Giallonardo F et al. Low-frequency drug-resistant HIV-1 and risk of virological failure to first-line NNRTIbased ART: a multicohort European case– control study using centralized ultrasensitive 454 pyrosequencing. J Antimicrob Chemother 2015; 70: 930–40. 8 Yazdanpanah Y, Fagard C, Descamps D et al. High rate of virologic suppression with raltegravir plus etravirine and darunavir/ritonavir among
15 Wensing AM, Calvez V, Gu¨nthard HF et al. 2014 Update of the drug resistance mutations in HIV-1. Top Antivir Med 2014; 22: 642–50. 16 Lataillade M, Chiarella J, Yang R et al. Virologic failures on initial boosted-PI regimen infrequently possess low-level variants with major PI resistance mutations by ultra-deep sequencing. PLoS One 2012; 7: e30118. 17 Fisher R, van Zyl GU, Travers SA et al. Deep sequencing reveals minor protease resistance mutations in patients failing a protease inhibitor regimen. J Virol 2012; 86: 6231 –7. 18 Todesco E, Rodriguez C, Morand-Joubert L et al. Improved detection of resistance at failure to a tenofovir, emtricitabine and efavirenz regimen by ultradeep sequencing. J Antimicrob Chemother 2015; doi:10.1093/jac/ dku557. 19 Charpentier C, Laureillard D, Piketty C et al. High frequency of integrase Q148R minority variants in HIV-infected patients naive of integrase inhibitors. AIDS 2010; 24: 867– 73. 20 Liu J, Miller MD, Danovich RM et al. Analysis of low-frequency mutations associated with drug resistance to raltegravir before antiretroviral treatment. Antimicrob Agents Chemother 2011; 55: 1114– 9. 21 Pou C, Noguera-Julian M, Pe´rez-A´lvarez S et al. Improved prediction of salvage antiretroviral therapy outcomes using ultrasensitive HIV-1 drug resistance testing. Clin Infect Dis 2014; 59: 578– 88.
7 of 7
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Francisco on March 24, 2015
References
10 Charpentier C, Roquebert B, Colin C et al. Resistance analyses in highly experienced patients failing raltegravir, etravirine and darunavir/ritonavir regimen. AIDS 2010; 24: 2651 –6.
