In Western Europe and North America, 6% to 13% of drug-naïve HIV-positive persons are infected with strains carrying resistance-associated mutations (RAMs) when tested by routine genotyping methods.1-9 Although discordant data have been reported,8,10 these patients appear to have an increased risk of poor virologic responses on starting highly active antiretroviral therapy (HAART).7,11-13 Bulk sequencing of plasma viral RNA fails to detect variants present within the quasispecies at a frequency below 20% to 30%.14 Although transmitted mutants have been shown to persist long term in the absence of drug pressure, a reversion occurs over time at rates that appear to be faster for some mutations (eg, the reverse transcriptase [RT] mutation M184V), possibly reflecting the reduced fitness of the mutant.15 As a consequence of the progressive reversion, routine genotyping cannot detect transmitted drug resistance (TDR) once the frequency of the transmitted mutant has decayed below the assay detection threshold.
Ultrasensitive techniques have recently been developed that, in drug-naïve patients, significantly increase the detection of resistant mutants relative to bulk genotyping.7,16-19 Evidence from clinical validation studies is not consistent, however, because both a negative impact of low-frequency resistant mutants on virologic suppression20 and lack of a discernible effect have been described in patient starting first-line HAART.7,21 These studies have used different testing methodologies, applied different sensitivity cutoffs for interpretation, and tested heterogeneous populations in terms of duration of infection and composition of the first-line HAART regimen. In this study, we used sensitive real-time polymerase chain reaction (PCR) to assess the effect of high- and low-frequency TDR on virologic responses to first-line non-nucleoside RT inhibitors (NNRTIs)-based HAART. We elected to test stored plasma samples collected immediately before starting HAART in a population that started therapy in the absence of a baseline genotypic resistance test, and set the mutation-specific interpretation cutoffs at 0.3% to 0.9% to gain confidence in our ability to specifically detect TDR rather than naturally occurring variants.19
Patients were indentified from the Royal Free Hospital clinic database. They were diagnosed with established HIV infection from 1996 onward, started HAART with efavirenz or nevirapine plus two or more nucleoside/nucleotide RT inhibitors (NRTIs) in 1998-2007 without a prior genotypic resistance test, had 24 or more weeks of follow up, and either experienced virologic failure or maintained consistent virologic suppression below 50 copies/mL without treatment changes. Virologic failure was defined as primary when patients did not achieve a viral load (VL) below 50 copies/mL within 24 weeks of starting HAART and secondary when patients achieved a VL below 50 copies/mL within 24 weeks but experienced confirmed virologic rebound within 48 weeks.
With the Royal Free Hospital Ethics Committee and Centers for Disease Control and Prevention Institutional Review Board approvals, plasma samples collected before starting HAART and, where available, additional samples collected at HIV diagnosis, were blinded to treatment outcomes and tested by sensitive real-time PCR targeting the RT mutations K65R, K103N, Y181C, M184V, and G190A as previously described.19,20 A detailed description of the methodology is given in reference 19 and can be freely accessed (www.plosone.org/article/info:doi/10.1371/journal.pone.0000638). Briefly, RT (nucleotides 58-777) was amplified by reverse transcription PCR (RT-PCR) and the template used in real-time PCR. To verify samples positive for RAMs, RT sequencing (nucleotides 133-708) was performed on the products of the primary RT-PCR reaction and the mutation-positive amplicons derived from the real-time PCR. The latter were evaluated for insertions and deletions to verify that the sequence was intact and for the presence of additional RAMs not targeted by the real-time PCR. Valid results were obtained in 93 of 100 tested patients (18 cases, 75 control subjects). A qualitative determination of whether the sample had RAMs was based on the difference in the mean (duplicate measurements) threshold amplification cycles using established interpretation cutoffs (0.3%-0.9%) below naturally occurring frequencies for each mutation.19,20 Plasma samples collected at virologic failure underwent bulk genotyping using the ViroSeq system (Celera Diagnostics, Alameda, CA) modified to include a nested PCR step to enable testing at VL below 1000 copies/mL.
Baseline characteristics were compared using χ2 or Fisher exact tests for categorical variables and t tests or Kruskal-Wallis tests for continuous variables. The association between virologic failure and TDR was tested by Fisher exact test. The analyses were performed using STATA Version 10.0 (STATA, College Station, TX).
Pre-treatment Resistance Profiles and Virologic Responses
From a population of 100 patients, real-time PCR was successful in 93 persons. These comprised 18 (19%) patients with virologic failure and 75 (81%) with virologic suppression while receiving first-line HAART, most commonly with zidovudine, lamivudine, and efavirenz (n = 37 [39.8%]). Baseline characteristics were generally balanced (Table 1). The median year of HIV diagnosis was 2000 (interquartile range [IQR], 1999-2003) and patients started HAART after median 1.0 (IQR, 0.0-3.4) year with a median CD4 count of 218 (IQR, 131-296) cells/mm3. In each patient, resistance testing was performed on a sample collected before starting HAART and, where available, on an additional sample collected at the time of diagnosis (Table 2). The median time between resistance testing and starting HAART was 0.0 (IQR, 0.0-1.7) years (Table 1). The median time between diagnosis and resistance testing was 1.0 (IQR, 0.0-3.0) years.
After unblinding of the data set, seven of 18 (38.9%) patients with virologic failure and zero of 75 patients with virologic success showed NNRTI RAMs in the pre-HAART sample (one G190A; six K103N). In three cases, one with G190A and two with K103N, the RAMs were strongly positive by real-time PCR (G190Ahigh, K103Nhigh) and were also detected by bulk genotyping. In the other four cases, K103N was present at low frequency and was only detected by real-time PCR (K103Nlow). The mutations K65R, Y181C, and M184V were not detected in any of the samples. In four patients without RAMs in the pre-HAART sample, no RAMs were detected in an earlier sample collected at the time of HIV diagnosis. None of the patients with resistance showed additional RAMs in the RT products amplified from the primary RT-PCR or the mutation-positive amplicons derived from the real-time PCR.
Patients with pre-HAART resistance were all males diagnosed HIV-positive in 1998-2007 and infected with subtype B (n = 6) or C (n = 1) (Table 2). Their CD4 count at the time of resistance testing ranged between 184 and 400 cells/mm3. The detection of bulk pre-HAART resistance in these patients was significantly associated with virologic failure (three of 18 versus zero of 75; P = 0.006). A significant association was also found between detection of pre-HAART K103Nlow and virologic failure (four of 18 versus zero of 75; P = 0.001). Combining bulk and low-frequency resistance increased the strength of the association (seven of 18 versus zero of 75; P < 0.0001). Logistic regression analysis was not performed as a result of the lack of baseline resistance in the control group.
Resistance Profiles at Virologic Failure
At Week 24, the seven patients with primary virologic failure showed a median VL of 3.6 (range, 2.3-4.5) log10 copies/mL; they either changed treatment immediately or continued the same regimen for up to 53 weeks without achieving virologic suppression. The 11 patients with secondary virologic failure experienced confirmed VL rebound after a median of 36 (range, 14-48) weeks with a median VL of 3.2 (range, 2.4-4.8) log10 copies/mL. At the time of failure, bulk genotyping yielded a result in 16 of 18 patients, but failed in two patients with a VL of less than 1000 copies/mL (Table 2). The patient with G190Ahigh, the two patients with K103Nhigh, and two of four patients with K103Nlow showed a bulk genotype that included the pre-HAART mutations plus additional RT mutations. One other patient with K103Nlow showed K103R when the VL was 2.1 log10 copies/mL. The K103R mutation, which does not confer NNRTI resistance in isolation, was also detected in the pre-HAART sample by bulk genotyping. This patient maintained a persistent low-level VL up to Week 53 while receiving tenofovir, emtricitabine, and efavirenz without emergence of K103N or other mutations by bulk genotyping. The fourth patient with K103Nlow showed V106M+F227L plus additional RT mutations at Week 24. Among the 11 patients without TDR, nine had a genotype at failure and four showed RAMs consistent with the treatment regimen (Table 2).
NNRTI-based regimens are very effective in suppressing HIV-1 replication but are vulnerable to rapid loss of activity in the presence of drug resistance.22 The K103N mutation, which confers high-level resistance to nevirapine and efavirenz, is highly prevalent in NNRTI-experienced patients and is commonly found by bulk genotyping in patients with TDR.4 In this study, low-frequency K103N mutants were as prevalent as bulk-detectable variants in samples collected from drug-naïve patients before starting HAART. This finding is noteworthy given that most patients had been diagnosed in the years immediately after the introduction of NNRTIs in clinical practice. Low-frequency K103N was detected in patients diagnosed as early as 1998. Although the dates of HIV seroconversion were not known, the patients had established infection at the time of diagnosis and generally low CD4 counts at the time of resistance testing. Taken together, these observations indicate that transmission of NNRTI resistance started early after the introduction of NNRTIs in clinical practice. They also demonstrate that although K103N mutants can persist for several years after transmission, possibly reflecting preserved viral fitness, their frequency may decline over time and eventually fall below the detection threshold of bulk genotypic assays. Despite being present at low frequency, however, the mutants retain a persistent negative impact on virologic responses to nevirapine- or efavirenz-based HAART.
Interestingly, there was no evidence of low-frequency K65R, Y181C, M184V, and G190A in this population. It should be acknowledged that we did not screen for all clinically relevant mutations (eg, Y181V or G190S). However, the mutations we targeted are those found most commonly in treatment-experienced patients. Furthermore, we detected single NNRTI-resistant mutants with no evidence of additional, linked RAMs in the RT products amplified from the primary RT-PCR and the mutation-positive amplicons derived from the real-time PCR.
As described previously,19 we used qualitative real-time PCR point-mutation assays (with direct sequencing of resistance mutation-specific PCR products) that are able to detect mutants at levels as low as 0.001% to 0.2% when testing mixtures of cloned virus sequences.19 However, screening of archived wild-type virus samples from the preantiretroviral drug era (1982-1985) allowed us to measure background reactivity and to define mutation-specific interpretation cutoffs ranging between 0.3% and 0.9%.19 Applying these cutoffs gives us confidence in our ability to specifically detect TDR rather than naturally occurring variants, increasing specificity. The finding that K103N impacts on virologic responses to NNRTI-based HAART when present at a frequency above the interpretative cutoff of 0.9% identifies a clinically relevant breakpoint. Larger studies are required to refine the quantitative relationship between frequency of the mutant and magnitude of the impact on virologic responses.
Using the same methodology, it was recently reported that detection of low-frequency K103N, Y181C, or M184V significantly reduced virologic responses in drug-naïve patients starting efavirenz in combination with lamivudine plus abacavir or zidovudine.20 Although we were unable to quantify the relationship between TDR and virologic failure as a result of lack of baseline RAMs in the control subjects, taken together, the data indicate a significant impact of both high- and low-frequency TDR mutants on virologic outcomes of first-line HAART. Other studies found no association between detection of low-frequency resistant mutants in drug-naïve patients and either viral load decline between 1 and 6 months of therapy7 or virologic failure.21 Low numbers of patients with TDR might confound analyses of treatment outcomes involving low-frequency resistant mutants. An additional potential confounder is the sensitivity of the PCR assay, which can have a cutoff well below 0.1%, thus blurring the demarcation between natural sequence variation and TDR. This is an important distinction when considering that TDR would seed resistance in a high proportion of the viruses archived. It is plausible that the clinical significance of resistant mutants is at least partly related to the frequency expressed within the quasispecies. Consequently, the time elapsed between resistance testing and HAART initiation may play a confounding role, because the effect on responses may decline with decaying prevalence of the mutants.15,23 The type of resistance mutations detected is likely to further influence outcomes. It is interesting that studies reporting a high prevalence of low-frequency M184V generally failed to find a significant impact of low-frequency mutants on virologic responses.7,21 Although this observation may reflect the reduced fitness and hypersusceptibility effects of M184V,24 available studies are insufficient to draw firm conclusions as a result of inclusion of patients showing a very low frequency of M184V and receiving first-line therapy with ritonavir-boosted protease inhibitors. The composition and genetic barrier of the first-line regimen can be proposed as a key confounder, because it may be anticipated that low-frequency mutations in RT have a low overall impact on virologic responses to regimens with a high genetic barrier to resistance.25
Under selective drug pressure, low-frequency mutants would be expected to acquire a replication advantage and become detectable by bulk genotyping. We found a fair association between the pre-HAART detection of low-frequency K103N and the failure genotype. In one case, bulk genotyping at very low viral load may have biased the detection of mutants, whereas in one other patient, the dominant mutations may have evolved during 10 weeks of virologic failure.26 Prospective studies are required to investigate the kinetics of transmitted resistant mutants in the course of treatment failure.
Our study is limited by the small number of observations and studies focusing on very recent infections with potentially greater chances of TDR might allow for stronger analyses. In addition, pre-HAART viral load in patients who experienced virologic failure tended to be higher than in patients who maintained suppression, albeit the difference did not reach statistical significance. It is possible that the higher viral load in the failures increased the likelihood of detecting low-frequency mutants. Nonetheless, our data demonstrate that real-time PCR doubles the rates of NNRTI TDR in this U.K. cohort, provide evidence of a significant impact of NNRTI TDR on responses to first-line NNRTI-based HAART, and give support to the controversial concept that transmitted NNRTI-resistant mutants affect virologic outcomes even when present at low frequencies within the quasispecies.20 Because current U.K. guidelines recommend NNRTI-based regimens as the favored first-line combination,27 larger studies should be designed to confirm the benefit of introducing sensitive screening methods for detecting transmitted NNRTI resistance in routine practice. Based on our observation that patients acquire single NNRTI-resistant mutants, the impact of low-frequency variants on responses to second generation NNRTIs should be investigated.
1. Masquelier B, Bhaskaran K, Pillay D, et al. Prevalence of transmitted HIV-1 drug resistance and the role of resistance algorithms: data from seroconverters in the CASCADE collaboration from 1987 to 2003. J Acquir Immune Defic Syndr
2. Wensing AM, van de Vijver DA, Angarano G, et al. Prevalence of drug-resistant HIV-1 variants in untreated individuals in Europe: implications for clinical management. J Infect Dis
3. Booth CL, Garcia-Diaz AM, Youle MS, et al. Prevalence and predictors of transmitted antiretroviral resistance in newly diagnosed HIV-1 infection. J Antimicrob Chem
4. Geretti AM. Epidemiology of antiretroviral drug resistance in drug-naive persons. Curr Opin Infect Di
5. SPREAD programme. Transmission of drug-resistant HIV-1 in Europe remains limited to single classes. AIDS
6. Vercauteren J, Derdelinckx I, Sasse A, et al. Prevalence and epidemiology of HIV type 1 drug resistance among newly diagnosed therapy-naive patients in Belgium from 2003 to 2006. AIDS Res Hum Retroviruses
7. Peuchant O, Thiébaut R, Capdepont S, et al. Transmission of HIV-1 minority-resistant variants and response to first-line antiretroviral therapy. AIDS
8. Bannister WP, Cozzi-Lepri A, Clotet B, et al. Transmitted drug resistant HIV-1 and association with virologic and CD4 cell count response to combination antiretroviral therapy in the EuroSIDA Study. J Acquir Immune Defic Syndr
9. Weinstock HS, Zaidi I, Heneine W, et al. The epidemiology of antiretroviral drug resistance among drug-naive HIV-1-infected persons in 10 US cities. J Infect Dis
10. Pillay D, Bhaskaran K, Jurriaans S, et al. The impact of transmitted drug resistance
on the natural history of HIV infection and response to first-line therapy. AIDS
11. Little SJ, Holte S, Routy JP, et al. Antiretroviral-drug resistance among patients recently infected with HIV. N Engl J Med
12. Violin M, Cozzi-Lepri A, Velleca R, et al. Risk of failure in patients with 215 HIV-1 revertants starting their first thymidine analog-containing highly active antiretroviral therapy. AIDS
13. Kuritzkes DR, Lalama CM, Ribaudo HJ, et al. Preexisting resistance to nonnucleoside reverse-transcriptase inhibitors predicts virologic failure of an efavirenz
-based regimen in treatment-naive HIV-1-infected subjects. J Infect Dis
14. Gunthard HF, Wong JK, Ignacio CC, et al. Comparative performance of high-density oligonucleotide sequencing and dideoxynucleotide sequencing of HIV type 1 pol from clinical samples. AIDS Res Hum Retroviruses
15. Bezemer D, de Ronde A, Prins M, et al. Evolution of transmitted HIV-1 with drug-resistance mutations in the absence of therapy: effects on CD4+ T-cell count and HIV-1 RNA load. Antivir Ther
16. Metzner KJ, Rauch P, Walter H, et al. Detection of minor populations of drug-resistant HIV-1 in acute seroconverters. AIDS
17. Palmer S, Kearney M, Maldarelli F, et al. Multiple, linked human immunodeficiency virus type 1 drug resistance mutations in treatment-experienced patients are missed by standard genotype analysis. J Clin Microbiol
18. Johnson JA, Li JF, Morris L, et al. Emergence of drug-resistant HIV-1 after intrapartum administration of single-dose nevirapine
is substantially underestimated. J Infect Dis
19. Johnson JA, Li J-F, Wei X, et al. Simple PCR assays improve the sensitivity of HIV-1 subtype B drug resistance testing and allow linking of resistance mutations. PLoS ONE
20. Johnson JA, Li J-F, Wei X, et al. Minority HIV-1 drug resistance mutations are present in antiretroviral treatment-naive populations and associate with reduced treatment efficacy. PLoS Med
21. Metzner KJ, Rauch P, Von Wyl V, et al. Prevalence of minority quasispecies of drug-resistant HIV-1 in patients with primary HIV-1 infection in Zurich in the years 2002-2006. Antivir Ther
22. Riddler SA, Haubrich R, DiRienzo AG, et al. Class-sparing regimens for initial treatment of HIV-1 infection. N Engl J Med
23. Lockman S, Shapiro RL, Smeaton LM, et al. Response to antiretroviral therapy after a single, peripartum dose of nevirapine
. N Engl J Med
24. Wainberg MA. The impact of the M184V substitution on drug resistance and viral fitness. Expert Rev Anti Infect Ther
25. Dunn D, Geretti AM, Green H, et al. Population trends in the prevalence and patterns of protease resistance related to exposure to unboosted and boosted protease inhibitors. Antivir Ther
26. Kolber MA. Development of drug resistance mutations in patients on highly active antiretroviral therapy: does competitive advantage drive evolution. AIDS Rev
27. Gazzard BG, BHIVA Treatment Guidelines Writing Group. British HIV Association guidelines for the treatment of HIV-1-infected adults with antiretroviral therapy 2008. HIV Med