| |
Transmitted HIV Drug Resistance Persists & HAART May Be Suboptimal
|
| |
| |
HIV-1 resistant strains acquired at the time of primary infection massively fuel the cellular reservoir and persist for lengthy periods of time
AIDS: Volume 20(2) 9 January 2006 p 159-170
Ghosn, Jadea; Pellegrin, Isabelleb; Goujard, Cecilec; Deveau, Christianed; Viard, Jean-Paule; Galimand, Juliea; Harzic, Martinef; Tamalet, Catherineg; Meyer, Laurenced; Rouzioux, Christinea; Chaix, Marie-Laurea; for the French PRIMO Cohort Study Group (ANRS CO 06)
From the aLaboratoire de Virologie, CHU Necker Enfants Malades, Universite Rene Descartes Paris, France
".....The presence of HIV drug-resistant variants poses significant challenge to effective therapeutic intervention at any stage of infection, including PHI. We not only confirm that acquired resistant viruses establish themselves as the dominant viral population at primary infection, but also, focusing on PBMC HIV DNA, we show that these resistant variants are also present in the cellular reservoir in all 44 patients at this early time-point in HIV infection. Moreover, resistance mutations remain detectable not only in cell-free virions but also persist in cell-associated HIV for at least 24 months (up to 48 months for two patients). Resistance mutations persisted in a drug-free environment in circulating and in intracellular HIV strains in five of five patients...."
".....In this context, classical empirical antiretroviral triple-combination might be suboptimal. Our results strengthen the accuracy of recent 2004 French guidelines which recommend performing genotypic resistance tests prospectively in patients with PHI, with results available soon after PHI confirmation. Thus, treatment failures such as occurred in patients would not be expected to occur with current practice guidelines in use, as if early HAART is initiated at the time of PHI, its virologic efficacy would be closely monitored to allow a rapid modification of antiretroviral regimen if resistance mutations were detected, in order not to impair the already limited therapeutic options...."
ABSTRACT
Objective: Characterization of the early establishment of the viral reservoir in patients acquiring resistant strains at primary HIV-1 infection (PHI), and longitudinal analysis of resistance mutations in circulating virions and intracellular HIV strains.
Patients and methods: Drug-resistance was compared between HIV RNA and peripheral blood mononuclear cell (PBMC)-HIV DNA at the time of PHI in 44 patients enrolled in the Primo Cohort and harbouring plasma HIV-1 resistant to at least one antiretroviral drug. Longitudinal monitoring of viral load and resistance genotype was performed in plasma-HIV RNA and PBMC HIV DNA for at least 24 months in a subset of 10 patients. Phylogenetic analysis of HIV DNA protease gene clones was used to explore the diversity of quasi-species at baseline.
Results: Baseline resistance profile was identical in paired HIV RNA and PBMC HIV DNA for all 44 patients.
All resistance-associated mutations persisted in plasma and PBMC over 2 years in the five untreated patients.
Of the five patients started on empirical HAART, two achieved undetectable HIV RNA at month 6, with long-term persistence of archived drug-resistance mutations in PBMC HIV DNA.
Virological failure was observed in the other three patients, resulting in the accumulation of additional drug-resistance mutations in HIV RNA and HIV DNA for two of them.
Phylogenetic analysis of HIV DNA clones showed highly homogenous and exclusively resistant quasi-species in the cellular reservoir at baseline.
Conclusion: HIV resistant strains acquired at the time of PHI massively fuel the cellular reservoir, and their prolonged persistence is supported by the early expansion of a dominant homogenous and resistant viral population. Results in treated patients showed that classical empirical triple-combination may be suboptimal.
INTRODUCTION
Increasing numbers of HIV-1 infected patients have a history of multiple treatment failures [1-3]. Resistance mutations in key genes may confer impaired fitness and reduced replicative capacity [4-6], and viruses harbouring multiple resistance mutations would have a substantially reduced transmission capacity [7]. Despite this, sexual transmission of such viral variants occurs [8-10]. In France, the frequency of transmitted resistant strains at the time of primary HIV infection (PHI) is stable and reaches 12% [11]. It could be expected that resistant HIV-1 strains acquired at the time of PHI would not persist over time in a drug-free environment, being overwhelmed by fitter wild-type strains, as seen during chronic disease in patients who discontinue a failing regimen [12,13]. However, there is recent evidence that detectable resistance may persist in plasma over a long time after PHI without drug-selective pressure [14-18]. The clinical implications are of serious concern since multidrug resistance (MDR) can result in treatment failure and clinical progression of PHI patients with MDR virus [16,19,20], especially within the first year following PHI [21]. Our objective was to characterize the early establishment of the viral cellular reservoir, and to analyse the temporal evolution of resistance mutations acquired at PHI not only in circulating recently produced HIV particles, but also in strains stored in the cellular reservoir, in treated and in untreated patients. Therefore, we compared resistance mutational pattern between HIV RNA and HIV DNA sampled at the time of PHI in patients enrolled in a national cohort of primary HIV-1 infection in France. Then, we monitored HIV resistance-associated mutations present in plasma HIV RNA, and in HIV DNA in peripheral blood mononuclear cells (PBMC), in a subset of patients, in the absence or in the presence of early HAART.
Study population
Our study comprised patients presenting with PHI, enrolled in the multicentre prospective French Primo Cohort (ANRS CO 06) [19,22-24]. The ethics committee of Cochin Hospital approved the study and all patients gave their written informed consent. From 1996 to December 2004, 518 PHI patients were enrolled in the cohort, with a median HIV RNA of 5.05 log10 copies/ml (range, 2.3-8.33 log10) and a median CD4 cell count of 521/ml (range, 69-1542/ml) at inclusion. This observational cohort does not impose guidelines for systematic treatment of patients presenting with PHI, the decision of initiating HAART or not relying on the primary care physician in each clinical setting. Seventy percent (364/518) of enrolled patients were started on antiretroviral therapy at confirmation of diagnosis of PHI, with a significant decrease from 92% in 1996 to 56% in 2001 (P < 10-6) [23], and to 33% in 2004 [25].
For all patients enrolled in the Primo Cohort, plasma samples were collected at inclusion and stored for retrospective genotypic resistance testing for epidemiological purposes [26]. Therefore, when initiated during PHI, HAART began before the results of genotypic resistance testing were available.
Primary HIV-infection was identified as previously described [23]. Briefly, primary HIV-infection was assessed by: (i) a negative or indeterminate HIV ELISA associated with a positive antigenemia or plasma HIV RNA; (ii) a western blot profile compatible with ongoing seroconversion (incomplete western blot with absence of antibodies to Pol proteins); or (iii) an initially negative test for HIV antibody followed by positive HIV serology within 6 months.
Patients enrolled in the Primo Cohort were selected for the present study if: (i) they harboured plasma HIV-1 strains resistant to at least one antiretroviral drug at time of PHI; and (ii) they were not co-included in structured treatment interruption trials [27] or in immune-based therapy trials [28,29]. Longitudinal monitoring of resistance mutational pattern was performed in a subset of patients with a minimum of 2 years of follow up if stored blood plasma and PBMC samples were available. Genotypic resistance tests were performed retrospectively on paired stored samples of plasma and PBMC collected at baseline, at month 6 (M6), M12, M24 and then every 12 months when available.
Discussion
The presence of HIV drug-resistant variants poses significant challenge to effective therapeutic intervention at any stage of infection, including PHI. We not only confirm that acquired resistant viruses establish themselves as the dominant viral population at primary infection, but also, focusing on PBMC HIV DNA, we show that these resistant variants are also present in the cellular reservoir in all 44 patients at this early time-point in HIV infection. Moreover, resistance mutations remain detectable not only in cell-free virions but also persist in cell-associated HIV for at least 24 months (up to 48 months for two patients).
Resistance mutations persisted in a drug-free environment in circulating and in intracellular HIV strains in five of five patients (32-36). This absence of genotypic change is supported by phylogenetic analysis of HIV DNA protease gene clones at baseline, showing the early expansion of a dominant homogenous [35-37] and resistant population. Indeed, the absence of detectable intracellular wild-type quasispecies (i.e., the absence of competition) in our PHI patients is consistent with the persistence of resistant viruses during follow-up.
This is supported by the detection of genotypic resistance in PBMC HIV DNA as soon as PHI by whole population-based sequencing methods in all 44 patients, suggesting that the cell-associated initial viral pool is mainly homogenous and resistant in those patients [38]. Despite the usual dynamic processes affecting the pool of infected cells [39], HIV DNA isolated from circulating PBMC far from PHI still exhibited the same resistance mutational pattern than at the time of PHI, indicating that resistant viruses entering the cellular reservoir early in infection had not been replaced. Interestingly, phylogenetic analysis of the RT gene showed that, for some patients, HIV RNA in plasma did not cluster with paired HIV DNA extracted from circulating blood cells at the same time-point. This finding suggested that circulating virions recently produced in plasma might arise from sources other than circulating infected blood cells [40,41]. However, HIV RNA sequence sampled far from PHI still exhibited the same resistance mutational pattern as at the time of PHI, thus suggesting that other reservoirs (such as anatomical reservoirs [42]) had been also fuelled with the same resistant strains at the time of PHI.
Interestingly, if a high number of resistance-associated mutations within the viral genome at time of PHI makes the occurrence of back mutations at all relevant loci less likely [43,44], it would have been expected that viruses harbouring only one single mutation, such as K103N, would have reverted at a relatively early time [45]. However, in patient 36, single mutation K103N conferring little fitness impairment [4,15,46], persisted for 24 months in both HIV RNA and archived HIV DNA without any drug selective pressure. In contrast, mutation M184V with deleterious impact on viral fitness [47,48] shifted to wild-type amino acid residue in plasma for patient 40 as soon as lamivudine-selective pressure was discontinued. Reversion of 184V acquired at the time of PHI has already been described in plasma HIV RNA [44,49]. However, even if no longer detectable in plasma, we showed that this resistance mutation persisted in the cellular reservoir, which makes rapid re-emergence possible in plasma if lamivudine was to be resumed. Moreover, in patient 7, a mixture of amino acid residue M/I was detected at position 184 of the RT gene in HIV DNA, but only wild-type amino acid residue was detected in paired plasma HIV RNA. This suggests that strains harbouring a wild-type amino acid residue at position 184 of the RT gene overwhelmed the mutated strains and were the dominant population in plasma in the absence of drug-selective pressure. Nevertheless, resistant strains harbouring M184V mutations would quickly emerge in plasma if lamivudine was to be started in patient 7.
There is evidence that transmitted drug-resistant HIV can lead to suboptimal response to first-line therapy in newly infected patients [16,19,50-52]. Intriguingly, patients 37 and 38 who harboured viral strains that were resistant to at least two drugs from their triple combination achieved undetectable plasma HIV RNA as soon as M6, which was sustained until M24 and M48, respectively. These findings suggest a T-cell mediated control of HIV replication [53-55], with a specific recognition of mutated epitopes [56-58]. Other factors such as neutralizing antibodies [59], specific HLA types or host genetic factors [60], and chemokine receptor polymorphisms might have played a role. Of note, we cannot exclude that these two patients would have achieved undetectable plasma viral load even without any therapeutic intervention. Indeed, plasma viral load had been spontaneously below 400 copies/ml without any antiretroviral treatment between the M6 and M24 following seroconversion in 4% of patients included in the Seroco study [61,62].
Three out of five treated patients had a slow decrease in HIV RNA, with a prolonged viral replication under drug selective pressure, thus promoting the accumulation and storage of additional drug-resistance mutations selected by the ongoing treatment in two patients, and jeopardizing the already limited therapeutic options [63]. Moreover, persisting viral replication despite HAART might increase secondary sexual transmission of drug-resistant variants [17,64].
In conclusion, transmitted HIV resistant strains ensure the dissemination and establishment of the cellular reservoir at the earliest time-point in HIV infection, with the same transmitted resistant strain. They are durably detectable in blood plasma and are archived in PBMC for lengthy periods of time. At the time patients 32-41 were included in the Primo Cohort, French guidelines recommended performing restrospective genotypic resistance test for an epidemiological purpose in patients with PHI [26]. In this context, classical empirical antiretroviral triple-combination might be suboptimal. Our results strengthen the accuracy of recent 2004 French guidelines which recommend performing genotypic resistance tests prospectively in patients with PHI [25], with results available soon after PHI confirmation. Thus, treatment failures such as occurred in patients 39-41 would not be expected to occur with current practice guidelines in use, as if early HAART is initiated at the time of PHI, its virologic efficacy would be closely monitored to allow a rapid modification of antiretroviral regimen if resistance mutations were detected, in order not to impair the already limited therapeutic options.
Results
Patients characteristics at baseline
Forty-four patients (80% men) enrolled in the Primo cohort harboured an HIV-1 strain that was resistant to at least one antiretroviral-drug at the time of PHI, and were eligible for the present study. Median time between HIV infection and inclusion in the cohort was 46 days (range, 24-183 days). At the time of PHI, median CD4 cell count, plasma HIV RNA and PBMC HIV DNA loads were 507/ml (range, 83-1509/ml), 4.61 log10 copies/ml (range, 1.77-6.73 log10 copies/ml) and 3.21 log10 copies/106 cells (range, 1.70-4.41 log10 copies/106 cells) respectively. Resistance to NRTI, NNRTI, and PI was detected in plasma HIV RNA in 27/44 patients, 13/44 patients and 23/44 patients, respectively. Eleven and five patients harboured plasma HIV-1 strains with resistance to two and all three antiretroviral drug classes, respectively. Mutation M184V in the RT gene was detected in HIV RNA of 11/44 patients.
For the 44 patients, all resistance mutations detected at the time of PHI in plasma HIV RNA were also detected in HIV DNA extracted from paired PBMC samples (Fig. 1a and b). However, in patient 7, a mixture of amino acid residue M/I was detected at position 184 of the RT gene in PBMC HIV DNA, whereas only wild-type amino acid was detected at that specific position in HIV RNA extracted from paired plasma sample.
Baseline characteristics of the subset of 10 patients with longitudinal follow up
Those 10 patients were included in the Primo cohort between March 1999 and February 2002. Median time between HIV infection and inclusion in the cohort was 57 days (range, 28-183 days), and median time between HIV diagnosis and inclusion was 32 days (range, 12-150 days). All patients presented with mild to severe symptomatic PHI. Baseline median plasma HIV RNA, PBMC HIV DNA and CD4 cell count values were 4.30 log10copies/ml (range, 2.4-5.7 log10copies/ml), 2.92 log10 copies/106 PBMC (range, 1.8-3.7 log10 copies/106 PBMC) and 562/ml (range, 300-1508/ml), respectively. For each patient, we found that baseline genotypic resistance profile was identical in HIV RNA from blood plasma and in HIV DNA from PBMC. Longitudinal monitoring of viral genotype in HIV RNA and PBMC HIV DNA was performed on samples from all ten patients for a median follow up period of 24 months (range, 24-48 months).
Longitudinal assessment of viral genotype in the five untreated patients (patients 32-36)
Results are shown in Table 1. At baseline, median HIV RNA load was 3.73 log10 copies/ml (range, 2.45-4.62 log10 copies/ml), median PBMC HIV DNA load was 2.4 log10 copies/106 PBMC (range, 1.84-3.65 log10 copies/106), and median CD4 cell count was 605/ml (range, 585-1508/ml). At inclusion, three patients were identified with resistance to one antiretroviral class (36, NNRTI; 34 and 35; NRTI), one patient to two antiretroviral classes (33; NRTI and PI) and one patient (32) to three antiretroviral classes (Table 1). The final evaluation took place at M24 for patients 32, 34, 35, 36, and M48 for patient 33.
As shown in Table 1, all HIV resistance-associated mutations persisted in HIV RNA and in PBMC HIV DNA throughout the follow up, except for mutation 215Y with a stepwise back-mutation at that specific position, leading to a shift to 215C in plasma HIV RNA at M12 then in PBMC HIV DNA at M24 for patient 35, and in both plasma HIV RNA and PBMC HIV DNA at M24 in patients 32 and 33.
Finally, none of the untreated patients experienced rapid progression to AIDS during follow up.
Longitudinal assessment of viral genotype in the five patients on HAART (patients 37-41)
Results are presented in Table 2. At baseline, median HIV RNA was 4.33 log10 copies/ml (range, 4.06-5.11 log10 copies/ml), PBMC HIV DNA load was 2.98 log10 copies/106 PBMC (range, 2.89-3.72 log10 copies/106 PBMC), and median CD4 cell count was 461/ml (range, 300-593/ml). One patient was identified with resistance to one antiretroviral class (41, PI), two with resistance to two antiretroviral classes (39 and 40, NRTI and PI), and two (37, 38) with resistance to three antiretroviral classes (Table 2). Patients were started on HAART as soon as PHI was confirmed, that is on the day of inclusion in the Primo Cohort (D0). At baseline (D0), HAART consisted of two NRTI in association with efavirenz (n = 3) or with nelfinavir (n = 2). The final evaluation took place at M24 in patients 37 and 39, M36 in patients 40 and 41, and M48 in patient 38.
Patient 37 was started on zidovudine, lamivudine and efavirenz, while he harboured an HIV strain with mutations associated with resistance to both zidovudine (215F) and efavirenz (181C). Patient 38 who was receiving zidovudine, lamivudine and nelfinavir, was infected with an HIV mutated strain that harboured mutations associated with resistance to all three drugs (zidovudine: 67N/70R/151M/215V/219Q; lamivudine: 184V, nelfinavir: 36I/46I/82A). These two patients achieved sustained undetectable plasma HIV RNA beginning from M6, with retention of all archived drug resistance-associated mutations in PBMC HIV DNA throughout median 24 months of follow up (Table 2).
In the other three patients, viral replication persisted under drug-selective pressure, with accumulation of additional drug-resistance mutations selected by the ongoing treatment in two out of three patients (40, +70R, 101E, 190S; 41, +184V). These additional mutations were first detected in plasma (M6), then in PBMC HIV DNA at M24 for patient 40 and M36 for patient 41.
Patient 40, whose virus bore mutation 184V at time of PHI, discontinued his failing regimen (zidovudine, lamivudine and efavirenz) at M18. A shift to wild-type amino acid residue occurred at positions 70 and 184 at M24 in plasma HIV RNA, but with a retention of both resistance mutations in PBMC HIV DNA.
Phylogenetic analysis of HIV RT gene
To analyse the temporal evolution of HIV RT gene in plasma and in PBMC, a total of 70 sequences of HIV RNA and HIV DNA RT gene from study patients were analysed, along with 10 RT sequences of patients from the surrounding community presenting with PHI without transmitted drug resistance (Fig. 2). All sequences showed the expected patient-specific clustering. Two features were evidenced. For a given patient, HIV RNA and HIV DNA sequences clustered together at the same time of sampling, with high bootstrap values (32-34). Conversely for another patient (35), HIV RNA and HIV DNA sequences from paired samples at D0, M6 and M12 did not cluster together at a given time-point.
Based on pairwise evolutionary distances, the temporal analysis of RT sequences showed limited evolution between the first and the last available sequence for each patient. Genetic evolution was extremely limited for untreated patients, with less than 1% variability in HIV RNA and ranging from 0.2 to 1.6% in HIV DNA (Table 3), most likely reflecting an absence of selective (e.g., drug) pressure. Genetic evolution was a little higher in treated patients 40 and 41 who failed their first-line empirical regimen, with a variability ranging from 1.5 to 2.8% in HIV RNA and from 0.5 to 3% in HIV DNA (Table 2).
Phylogenetic analysis of HIV DNA protease gene clones from baseline PBMC
We analyzed the sequences of HIV DNA protease gene clones from PBMC samples at baseline in two patients (40,41) (Fig. 3). In both cases, the most striking finding was the high homogeneity of the quasispecies integrated in PBMC at an early time-point during PHI. Based on pairwise evolutionary distances, the intravariability between quasispecies was 0.49% (range, 0-1) in patient 40 and 0.50% (range, 0-1.4) in patient 41.
Further analysis of resistance mutations strengthened the homogeneity of strains integrated in PBMC early during PHI, showing exclusively homogenous and resistant quasispecies, with all quasispecies harbouring the same mutational pattern for each patient. Finally, no wild-type quasispecies were detected in PBMC sampled at baseline.
|
|
| |
| |
|
|
|
