Despite a decrease in plasma viral RNA to below the level of detection after starting highly active antiretroviral therapy (HAART), virus persists for a very long period in what is called the latent reservoir. 1–6 Our current knowledge of this reservoir (i.e., how it originates, how it is maintained and eventually renewed) is still very limited. An important stable long-term viral reservoir in patients on HAART is thought to be composed of resting memory CD4+ T cells carrying replication-competent viral genomes. 1,2 The proportion of these cells is supposed to be low, but as demonstrated recently, they persist for many years even in the absence of active virus replication. 7 Finzi et al. 3,4 showed that a latently infected CD4+ T-cell compartment becomes established very early in infection, but the factors that are involved in the maintenance and eventual replenishment of this compartment are still largely unknown. Moreover, much debate continues about the importance of residual viral replication as a mechanism of replenishment of the latent reservoir during HAART. 4,8–10
If not fully suppressive, any antiretroviral treatment used today will result in the development of resistance. We and others have shown that in the majority of patients with drug-resistant virus, treatment interruption results in the reemergence of drug-susceptible HIV-1. 11–13 These observations have evoked large interest in structured treatment interruptions as a way to reduce the amount of resistant virus to very low levels, thereby possibly increasing the chance of durable viral suppression on subsequently resumed therapy. 13–16 The results of these structured treatment interruption studies are still controversial but a fast reemergence of resistant variants under the selective pressure of the new antiretroviral regimen has been demonstrated already. 17,18 If and to what extent resistant virus enters the latent reservoir, and for how long it persists in this reservoir, are still unknown. A better insight into the kinetics of the latent virus reservoir and its composition with regard to wild-type and resistant virus might help to improve the strategies for successful treatment of heavily exposed individuals.
To study the persistence of proviral sequences carrying drug-resistant mutations, we selected 11 patients who were on fully suppressive HAART for several years but who had been exposed to suboptimal therapy previously. The main objective was to see whether many years of selection for drug-resistant virus followed by a long period of suppression of replication would lead to the disappearance of the wild-type drug-sensitive virus and replacement by mutant drug-resistant virus in the reservoir. Also, we wanted to examine the value of sequencing of the cellular provirus as a way to obtain information about previous drug resistance. We used limiting dilution polymerase chain reaction (PCR) followed by sequencing of the single-copy PCR products to genotype the reverse transcriptase (RT) and protease gene of different proviral variants.
MATERIALS AND METHODS
Eleven patients were selected from the patient cohort of the AIDS Reference Center of the University Hospital in Ghent, Belgium. Patients from this cohort are followed intensively. Enrollment was based on the following criteria: the patients had received suboptimal treatment before HAART, they carried drug-resistant virus as revealed by sequencing at the time of HAART initiation, and they had currently been on HAART for >4 years. Viral load determinations during HAART were performed at least every 2–3 months. Viral load remained undetectable (<50 copies/mL) during the whole period with exception of 1 occasional positive result (206 copies/mL) after 48 months of HAART in patient 1. Viral load returned to <50 copies/mL in a sample taken from this patient 1 month later. Before HAART, the patients were treated with either zidovudine (AZT), zalcitabine (ddC), didanosine (ddI), or lamivudine (3TC) as monotherapy or in combination. One patient participated in a nonnucleoside reverse transcriptase inhibitor (NNRTI) trial and received the drug loviride. Suboptimal therapy was given for a mean period of 46 months. Most patients received a total of 2 or 3 drugs. During the pre-HAART treatment, plasma was collected every 3–6 months for viral load determination. In all patients, the viral load remained detectable during the whole pre-HAART treatment period.
At the time the peripheral blood mononuclear cell (PBMC) samples were taken, all patients were on HAART for a mean period of 59 months (range 54–68) with a combination of 2 nucleoside reverse transcriptase inhibitors (NRTIs) and 1 or 2 protease inhibitors (PIs). All patients were white and infected with subtype B virus. Their mean age was 43 years (range 32–54).
HIV RNA Quantification
Plasma samples were obtained at each visit and stored at −70°C. HIV RNA quantification was performed using the Ultrasensitive Cobas Amplicor HIV-1 Monitor Test (Roche Molecular Systems, Branchburg, NJ) with a detection limit of 50 copies/mL.
DNA Extraction and Limiting Dilution PCR
DNA was extracted from freshly isolated PBMCs using the QIAamp Blood Kit (QIAGEN GmbH, Hilden, Germany). DNA samples were diluted 10-fold and 5 μL of this dilution was added to each of at least 40 identical PCR mixes containing the outer primer set (sense 5′-ATGATGCAGAGAGGCAATTT-3′; antisense 5′-TTCTGTATGTCATTGACAGTCCAGC-3′) to amplify an approximately 1200-bp fragment spanning the protease and the first 240 amino acids of the RT gene. Amplification was performed for a total of 35 cycles (20 seconds at 94°C, 20 seconds at 50°C, and 1 minute at 72°C), after which 2 μL of the amplified products were transferred to 48 μL of 2 reaction mixes containing either a primer set to amplify the protease gene (sense 5′-AGAGCCAACAGCCCCACCA-3′; antisense 5′-GGGCCATCCATTCCTGGCTT-3′) or a primer set to amplify the first part of the RT gene (sense 5′- CCAAAAGTTAAACAATGGCCATTGACAGA-3′; antisense 5′-AGTTCATAACCCATCCAAAG-3′). Nested PCR amplification was performed for 30 cycles (20 seconds at 94°C, 20 seconds at 57°C, and 30 seconds at 72°C). Positive PCR products were selected for sequence analysis only if less than one-third of the replicate reactions were found positive. DNA samples for which more than one-third of the reactions were positive were diluted 10-fold further, and the PCR reactions were repeated until a dilution was found for which no more than one-third of the reactions were positive. Both positive and negative controls were included in all PCR assays to assess the sensitivity of the reaction and to detect possible contamination. The lower limit of detection of the PCR assay was equivalent to 1 copy per reaction. All positive PCR products were sequenced.
Direct sequencing of both sense and antisense strands of the inner PCR products was done with the dRhodamine Terminator Cycle Sequencing Ready Reaction kit (Applied Bio-systems, Foster City, CA). The sequencing reaction was performed with the same primers as the ones used in the inner PCR reactions, but to obtain a full sequence of both strands of the RT gene fragment, 2 additional sequencing reactions were run with internal primers (sense 5′- GGGNGAYGCATATTTTTCARTWCC-3′; antisense 5′- CCTGGTGTYTCATTRTTTRYACTT-3′). Sequencing reaction products were analyzed on an ABI 310 Genetic Analyzer (Applied Biosystems). A minimum of 11 (range 11–18) different PCR products were sequenced for each patient sample. The principle of limiting dilution sequencing is based on the mathematical calculation that if no more than one-third of replicate PCR reactions are positive, the likelihood that the PCR products are the result of the amplification of only 1 molecule is ~70%. Limiting dilution sequencing allows us to obtain an accurate profile of the distribution of different variants in a single sample. 19 The likelihood of comparing single provirus sequencing data is further enhanced by withdrawing all sequencing products for whom visual inspection of the electropherograms revealed nucleotide mixtures at ≥1 positions. Sequences containing stop codons or frame shifts indicating defective virus were also withdrawn. Sequencing results were only used in the analysis if the sequence of both strands was available and fully concordant.
Genotypic Analysis of Plasma Viral RNA
RT-PCR of the RT and protease genes was performed on stored plasma viral RNA using the Titan One Tube RT-PCR System (Roche Molecular Systems). Direct sequencing of the PCR product was done as described for the proviral DNA samples. Amino acid substitutions were identified by comparison of the plasma RNA sequences with a consensus HIV-1 subtype B sequence.
Nucleotide sequences were assembled using the BioEdit package (www.mbio.ncsu.edu/BioEdit). Phylogenetic analyses and neighbor-joining tree reconstructions were performed using programs from version 3.6 of the PHYLIP package (http://evolution.genetics.Washington.edu/phylip), with a maximum likelihood distance matrix and a transition to transversion ratio of 2.0. Approximate confidence limits for individual branches were assigned by bootstrap resampling with 1000 replicates. Tree diagrams were plotted with Treeview v1.4 (http://taxonomy.zoology.gla.ac.uk/rod/treeview).
Nucleotide Sequence Accession Numbers
The nucleotide sequences reported in this paper have been submitted to GenBank and were given accession numbers AY356748 to AY357066.
Response to HAART
Table 1 summarizes the treatment history and response to HAART for the patients enrolled in the study. Patients are ordered according to the time on suboptimal therapy. All patients showed a rapid decline in plasma viral load to below the levels of quantification (<50 copies/mL) after initiation of HAART. Values all remained undetectable on repeated measurements during the whole treatment period. HAART also resulted in an important increase in CD4 count (mean CD4 rise: 575, range 203–837).
Proviral DNA Sequencing
Table 2 summarizes the results of the sequencing analysis of the RT gene of proviral DNA variants isolated from PBMCs collected after an average HAART period of 59 months and the results of the sequencing analysis of the RT gene of plasma virus isolated from consecutive blood samples during the pre-HAART period. Baseline plasma samples were not always available (missing for patients 6, 8, 9, 10, and 11). The proviral DNA was shown to be constituted of a mixture of wild-type proviruses and drug-resistant proviruses in 9 of the 11 patients studied. Only in patient 1 were no drug-resistant proviral variants detected. This patient had been on suboptimal therapy for the shortest period (11 months). In patient 9, only variants with resistant mutations were found. This patient had been on suboptimal therapy for a long period (5 years), indicating a possible association between the time on suboptimal therapy and the amount of resistant proviral variants. This assumption is further strengthened by the observed overall correlation between the time on suboptimal therapy and the relative amount of mutant sequences between the proviral variants (logistic regression coefficient r2= 0.6308; P = 0.004) (Fig. 1).
The number of different proviral variants that were detected in 1 sample varied from 3 to 8. Proviral variants with resistant mutations were heterogeneous in all patients; variants with either different numbers of mutations or different combinations of mutations were detected in the same patient. Results of sequencing analysis performed retrospectively on stored plasma samples revealed that most of these variants had been circulating in the plasma transiently during the period of suboptimal therapy (Table 2). Proviral variants with additional mutations as compared with the variants found in plasma were seen in only 1 patient (patient 5). This patient carried a variant with an additional M184I mutation and a variant with an additional T69N mutation. Proviral variants with PI-associated primary mutations were not observed (data not shown).
After aligning the whole nucleotide sequence of about 1000 base pairs (protease gene and part of the RT gene) of all proviral and viral variants, a phylogenetic tree was constructed for each patient. The codons associated with resistance were removed from the alignment before the phylogenetic analysis was performed so that the tree topology was not determined by the resistance mutations. All trees showed a pronounced intermingling of viral and proviral sequences. No indications for a separate evolution within the provirus population were found. Despite removal of codons associated with resistance mutations before the analysis, variants with the same resistance pattern always clustered together. Two representative trees are shown in Figure 2. One tree is constructed from the results of patient 3. This patient had been on suboptimal therapy for 20 months. The proportion of wild-type variants in the provirus is high. Drug-resistant viral and proviral variants cluster together but bootstrap support for clustering was low (<50%). The second tree is constructed from the results of patient 7. This patient had been on suboptimal therapy for 50 months. The proportion of wild-type variants in the provirus is low. An intermingling of viral and proviral variants can be observed, and there is no evidence for a separate evolution of the provirus population.
Although advances in HIV treatment have reduced the morbidity and mortality rates among HIV-infected individuals, all currently prescribed antiretroviral drugs fail to eliminate the latent reservoir and it is clear that, with the current treatment strategies, eradication of the virus will never be possible. Therapy has to be taken for life, and this is complicated due to the adverse effects of the drugs and due to the emergence of drug resistance. HIV-infected individuals in whom drug regimens have repeatedly failed often harbor virus with multiple drug resistance–associated mutations. Although it has been shown that stopping therapy or switching from one class of drugs to another leads to the disappearance of the resistant strains from the plasma in the majority of cases, the question of whether this also will enable recycling of these drugs in the future is not yet clearly answered. 11,12 Although viral latency under HAART is the subject of several studies, the mechanisms of HIV persistence and reservoir establishment remain largely unknown. 2–5,7,8,20,21
We studied the variability of the RT and protease gene in the provirus of patients who were under long-term HAART but who had a history of suboptimal therapy in the past. Our results confirm the observations of others that cells containing HIV-1 provirus remain detectable for periods extending several years. 7,10 Proviral sequences with a fully wild-type RT gene were found in 10 of the 11 patients despite the fact that in all these patients drug-resistant mutants have been favored by the selective conditions for many years. For the 1 patient in whom no wild-type proviral variants were detected, no pretreatment plasma samples were available so we cannot exclude the presence of the 70R mutation as a polymorphism already before starting medication.
In accordance with the findings in HIV-1–infected children, our results show that viruses in the latent reservoir are diverse and reflect selection by the pre-HAART regimens. 21 From the results of this study we have arguments to support the observation also made recently by Strain et al. 10 that the maintenance of the cellular reservoir is a dynamic process. New variants that are able to replicate for a certain period enter the reservoir to be conserved for longer periods. With a few exceptions, all mutant virus variants that were found in the plasma during the process of gradually building up resistance were still detectable several years later in the provirus. From the correlation that we observed between the period on suboptimal therapy before HAART and the proportion of mutant proviral sequences in the PBMCs, we can conclude that the quantity that a certain variant occupies within the reservoir will depend in part on the period that this variant has been able to replicate. However, the slow fading out of the oldest variants—in these cases the wild-type variants—might also contribute to the observed correlation.
We were not able to find indications for a further virus evolution under HAART. Only in patient 5, two observations might reflect some evolution: the detection of a 184I-carrying proviral variant and a 69N variant in provirus but not in plasma. The patient was on a combination of stavudine (d4T) + 3TC + ritonavir (RTV) + saquinavir (SQV) and selection of a 69N by this combination is possible, although it is more likely that this mutant arose during the pre-HAART bitherapy with AZT and ddI but was missed in plasma. The 184I mutation is known to be a 3TC-resistant transient intermediate stage between the wild-type 184M and the 3TC-resistant 184V. 22 Because 3TC is a component of the HAART regimen and the patient has never taken 3TC before, the chance is high that this variant arose during the HAART period. However, we cannot exclude the occurrence of 184I as a natural polymorphism. In this regard it is important to note that the 184I mutation was detected in a proviral variant with an otherwise completely wild-type background. 3TC was a component of HAART in 9 of the 11 patients, but no other patients showed proviral sequences with mutations at codon position 184.
PIs were a component of the HAART regimen in all patients but no additional PI-associated mutations compared with the secondary mutations already present in the plasma virus before HAART initiation were detected (results not shown). An additional argument against further evolution of proviral sequences is the fact that phylogenetic analysis revealed intensive intermingling of proviral and viral variants in all patients.
Currently, plasma is the only compartment used routinely for drug resistance testing and studies that address the role of the cellular reservoir with regard to emerging drug resistance and conservation of drug resistance are limited. 21 Our results show that infected PBMCs of patients under HAART contain a heterogeneous mixture of different viral variants. Because of this heterogeneity, population-based sequencing of provirus will presumably only detect major variants and will not provide valuable information about the resistance potential. Limiting dilution sequencing, conversely, was shown to allow detection of archived viral resistance, but the method is time consuming and expensive and therefore not suitable for large-scale use. It also remains to be examined to what extent the archived viruses remain replication competent. In this study, 10 sequences with stop codons, frame shifts, or hyper-mutations were found on the total of 173 sequences that were analyzed and they were removed from the analysis. However, our observations are limited to the HIV-1 protease and part of the RT gene and we cannot exclude the occurrence of mutations resulting in defective virus elsewhere in the genome. Despite the fact that we have no evidence for the replication competence of the archived proviral sequences, we consider the lack of any selective pressure that might be induced by in vitro culture as an important advantage, allowing a better estimate of the quantitative distribution of the different proviral variants present.
The results described here have important clinical implications because they confirm the long-term persistence of any drug-resistant virus once it has arisen, thereby permanently jeopardizing certain treatment options. In the patients studied here, HAART was initiated at a time when resistance testing was not performed. It is important to notice that, with the current knowledge, several of the drug combinations used at that time would no longer be prescribed in these patients, considering the observed genotypic resistance patterns. However, even with a suspected “less active” HAART, all patients showed a long-term virologic and immunologic response to the treatment, indicating that drugs to which resistance is predicted can still have therapeutic value in a combination regimen. Besides the fact that resistance is seldom an all-or-nothing phenomenon and low-grade resistance can be overcome by high drug concentrations, the results of this study point to another possible explanation for this observation, the fact that despite the detection of fully resistant virus in plasma, the majority of infected cells might still contain wild-type virus. As part of a combination regimen, drugs to which resistance has been developed can still add to the activity of the combination by preventing the replication of this latent wild-type, drug-sensitive, virus pool.
The authors thank M. Cnockaert and C. De Boever for technical assistance.
1. Chun TW, Finzi D, Margolick J, et al. In vivo fate of HIV-1-infected T cells: quantitative analysis of the transition to stable latency. Nat Med
2. Chun TW, Stuyver L, Mizell SB, et al. Presence of an inducible HIV-1 latent reservoir during highly active antiretroviral therapy. Proc Natl Acad Sci U S A
3. Finzi D, Hermankova M, Pierson T, et al. Identification of a reservoir for HIV-1 in patients on highly active antiretroviral therapy. Science
4. Finzi D, Blankson J, Siliciano JD, et al. Latent infection of CD4+ T cells provides a mechanism for lifelong persistence of HIV-1, even in patients on effective combination therapy. Nat Med
5. Persaud D, Pierson T, Ruff C, et al. A stable latent reservoir for HIV-1 in resting CD4(+) T lymphocytes in infected children. J Clin Invest
6. Wong JK, Hezareh M, Gunthard HF, et al. Recovery of replication-competent HIV despite prolonged suppression of plasma viremia. Science
7. Siliciano JD, Kajdas J, Finzi D, et al. Long-term follow-up studies confirm the stability of the latent reservoir for HIV-1 in resting CD4(+) T cells. Nat Med
8. Ramratnam B, Mittler JE, Zhang L, et al. The decay of the latent reservoir of replication-competent HIV-1 is inversely correlated with the extent of residual viral replication during prolonged anti-retroviral therapy. Nat Med
9. Blankson JN, Persaud D, Siliciano RF. The challenge of viral reservoirs in HIV-1 infection. Annu Rev Med
10. Strain MC, Gunthard HF, Havlir DV, et al. Heterogeneous clearance rates of long-lived lymphocytes infected with HIV: intrinsic stability predicts lifelong persistence. Proc Natl Acad Sci U S A
11. Devereux HL, Youle M, Johnson MA, et al. Rapid decline in detectability of HIV-1 drug resistance mutations after stopping therapy. AIDS
12. Verhofstede C, Wanzeele F, Van Der Gucht B, et al. Interruption of reverse transcriptase inhibitors or a switch from reverse transcriptase to pro-tease inhibitors resulted in a fast reappearance of virus strains with a reverse transcriptase inhibitor-sensitive genotype. AIDS
13. Deeks SG, Wrin T, Liegler T, et al. Virologic and immunologic consequences of discontinuing combination antiretroviral-drug therapy in HIV-infected patients with detectable viremia. N Engl J Med
14. Hance AJ, Lemiale V, Izopet J, et al. Changes in human immunodeficiency virus type 1 populations after treatment interruption in patients failing antiretroviral therapy. J Virol
15. Izopet J, Massip P, Souyris C, et al. Shift in HIV resistance genotype after treatment interruption and short-term antiviral effect following a new salvage regimen. AIDS
16. Miller V, Sabin C, Hertogs K, et al. Virological and immunological effects of treatment interruptions in HIV-1 infected patients with treatment failure. AIDS
17. Delaugerre C, Valantin MA, Mouroux M, et al. Re-occurrence of HIV-1 drug mutations after treatment re-initiation following interruption in patients with multiple treatment failure. AIDS
18. Izopet J, Souyris C, Hance A, et al. Evolution of human immunodeficiency virus type 1 populations after resumption of therapy following treatment interruption and shift in resistance genotype. J Infect Dis
19. Becker-Pergola G, Mellquist JL, Eshleman JR, et al. Improved detection of human immunodeficiency virus type 1 variants by analysis of replicate amplification reactions: relevance to studies of human immunodeficiency virus type 1 vertical transmission. Mol Diagn
20. Pierson T, McArthur J, Siliciano RF. Reservoirs for HIV-1: mechanisms for viral persistence in the presence of antiviral immune responses and antiretroviral therapy. Annu Rev Immunol
21. Ruff CT, Ray SC, Kwon P, et al. Persistence of wild-type virus and lack of temporal structure in the latent reservoir for human immunodeficiency virus type 1 in pediatric patients with extensive antiretroviral exposure. J Virol.
22. Frost SD, Nijhuis M, Schuurman R, et al. Evolution of lamivudine resistance in human immunodeficiency virus type 1-infected individuals: the relative roles of drift and selection. J Virol
Keywords:© 2004 Lippincott Williams & Wilkins, Inc.
drug resistance; provirus; latent reservoir; persistence of drug resistance