| |
HIV-1 Can Persist in Aged Memory CD4+ T Lymphocytes With Minimal Signs of Evolution After 8.3 Years of Effective Highly Active Antiretroviral Therapy
|
| |
| |
[Basic Science]
JAIDS Journal of Acquired Immune Deficiency Syndromes:Volume 50(4)April 2009pp 345-353
Nottet, Hans S L M PhD*; van Dijk, Sabine J MSc*; Fanoy, Ewout B MSc*; Goedegebuure, Irma W*; de Jong, Dorien*; Vrisekoop, Nienke PhD; Baarle, Debbie van PhD; Boltz, Valerie MSc; Palmer, Sarah PhD; Borleffs, Jan C C MD, PhD; Boucher, Charles A B MD, PhD*
From the *Department of Medical Microbiology; and Department of Immunology, University Medical Center Utrecht, Utrecht, The Netherlands; HIV Drug Resistance Program, National Cancer Institute, National Institutes of Health, Frederick, MD; and Department of Internal Medicine, University Medical Center Utrecht, Utrecht, The Netherlands.
"In conclusion, during long-term suppressive HAART, HIV is able to persist in terminally differentiated CD4+ T cells as proviral DNA of which the evolution is restricted to a minimum.....One suggested mechanism which could contribute to the latent reservoir is a low level of ongoing viral replication.11,48 Viral replication under selective pressure of HAART will ultimately lead to emergence of drug-resistant viruses. However, no major drug resistance mutations in protease did develop in our patients. Our study thus extends the findings of a previous study reporting a lack of drug resistance development in protease in PBMCs after 2 years of successful HAART by a period of almost 7 years"
Abstract
Background: Patients on long-term highly active antiretroviral therapy (HAART) were studied to determine persistence, drug resistance development, and evolution of HIV-1 proviral DNA.
Methods: Peripheral blood mononuclear cells were obtained by large volume blood drawn (500 mL) from 8 clinically successfully treated patients who had received uninterrupted HAART for up to 8.9 years. HIV-1 load was determined by Taqman real-time polymerase chain reaction. Drug resistance mutations were determined by sequencing and ultrasensitive, allele-specific, reverse transcriptase (RT)-polymerase chain reaction.
Results: HIV-1 DNA load was significantly higher in aged memory (CD45RO+ CD57+) when compared with memory (CD45RO+ CD57-) and naive (CD27+ CD45RO-) CD4+ T cells after HAART. Sequencing revealed no major drug resistance mutations in protease in all patients and appearance of resistance mutations in RT in just 1 patient. In 1 of 5 patients with undetectable viremia during treatment, RT M184 substitutions were detected. Phylogenetic analysis showed short genetic distances between patient sequences.
Conclusions: During long-term HAART, HIV-1 is able to persist in terminally differentiated CD4+ T cells as proviral DNA. Viral evolution was restricted, and in 80% of the patients with undetectable viremia, no sign of viral replication could be detected.
INTRODUCTION
The introduction of highly active antiretroviral therapy (HAART) has led to a significant decline in morbidity and mortality as a result of AIDS.1,2 Successful therapy reduces plasma HIV-1 RNA load below the detection level (50 copies/mL) in most patients.3-5 However, total eradication of the virus is impossible with current HAART regimens due to persistence in various cellular reservoirs.6-8 Resting memory CD4+ T cells form a very important reservoir, but HIV-1 DNA has also been detected in naive CD4+ T cells, monocytes, and macrophages.9,10 Replication competent viruses could be isolated from CD4+ T cells even after 8 years of successful HAART.11-13 However, the exact T-cell subsets in which HIV-1 resides over such a long period still have to be identified.
One potential mechanism by which HIV-1 can persist for long periods in CD4+ T cells is establishment of latency.14,15 HIV-1 preferentially infects activated CD4+ T cells, and although most of them die as a result of the cytopathic effects of HIV-1 infection, some infected effector cells return to a resting memory state carrying integrated proviral DNA.16,17 Memory cells are designed to survive for longer periods and in this way provide a stable reservoir with a half-life of more than 43 months.11,17,18 However, the half-life of the viral reservoir is longer than the estimated half-life of memory CD4+ T cells, suggesting that there is another mechanism that preserves the reservoir.11 A very low level of ongoing replication of a magnitude below the customary detection level might be possibly responsible for replenishment of the reservoir.14,19 The detection of HIV-1 in monocytes and double long terminal repeat HIV DNA in naive CD4+ T cells are indeed suggestive for ongoing replication.10-12,17
If maintenance of a stable viral reservoir during effective HAART is the result of ongoing viral replication, conditions are created that favor selection of HIV-1 variants harboring drug-associated mutations in protease and reverse transcriptase (RT). Early studies have reported the absence and the appearance of these HIV-1 variants in patients after 1 year of HAART.6,8,19-23 More recently, however, it has been shown that clinically successful HAART may result in an arrest in viral replication and evolution. For instance, genotypic analysis of protease and RT in plasma of patients on HAART showed low diversity in sequences, a lack of temporal relationship and absence of new resistance mutations.24-26 In addition, HAART also significantly limits the evolution of HIV-1 in resting CD4+ T cells.27 Analysis of protease sequences in peripheral blood mononuclear cells (PBMCs) of drug-naive patients who were treated for 2 years with successful HAART showed an absence of both clinically relevant evolution and appearance of drug resistance mutations.28 So far, no study on the appearance of drug resistance has been performed for a period up to 9 years. Follow up over such a long period might be necessary to rule out viral evolution and emergence of drug resistance at a very low pace.
To gain insight into the mechanisms that maintain the latent reservoir, we investigated levels of proviral DNA in naive, memory, and terminally differentiated memory CD4+ T cells of patients who had received uninterrupted long-term HAART. Furthermore, evolution of and drug resistance mutations in protease and RT sequences were determined in plasma samples and PBMCs obtained before and after an average of 8.3 years of HAART.
DISCUSSION
Although HAART can control HIV infection in most patients, viral persistence in a latent reservoir seems to be the biggest obstacle to virus eradication.15,38-44 It is plausible to hypothesize that during effective HAART, integrated HIV-1 DNA is, due to its presence in the cellular genome, intimately linked to development and differentiation of CD4+ T cells. In conformity with this hypothesis, we found that HIV-1 DNA load in terminally differentiated cells was on average 40 times higher when compared with memory cells. In the scenario where HAART completely blocks infection, the amount of HIV DNA gradually moves from the memory to the terminally differentiated memory CD4+ T-cell pool. These cells have passed through the maximum number of cell divisions to achieve terminal differentiation, and with each cell division, the amount of HIV DNA doubles. In addition, terminally differentiated memory CD4+ T cells are expanded during HIV infection.45 Furthermore, replenishment of the naive and the memory CD4+ T-cell pool by thymic production of new uninfected CD4+ T cells occurs.46,47 These events, all together, might explain the relative high proviral DNA load in the aged memory cells. Interestingly, in a mixed group consisting for 86% of untreated and unsuccessfully treated patients, terminally differentiated memory CD4+ T cells did contain on average 10-fold fewer copies of HIV-1 DNA compared with memory CD4+ T cells.9 This finding supports our hypothesis that under effective HAART, proviral DNA, either belonging to replication competent or incompetent HIV, is becoming part of restored T-cell development and thus accumulates in aged memory cells.
One suggested mechanism which could contribute to the latent reservoir is a low level of ongoing viral replication.11,48 Viral replication under selective pressure of HAART will ultimately lead to emergence of drug-resistant viruses. However, no major drug resistance mutations in protease did develop in our patients. Our study thus extends the findings of a previous study reporting a lack of drug resistance development in protease in PBMCs after 2 years of successful HAART by a period of almost 7 years..28 The appearance of the L210W and K219Q substitutions in RT, as observed in one of our patients, has been reported to arise during treatment with zidovudine.49,50 Because acquisition of these drug resistance mutations most likely occurred in the first year of treatment of this patient, it is tempting to assume that the drug-resistant genotype did persist for up to 7 years in PBMC. Indeed, CD4+ T cells can serve as a permanent archive for wild-type and drug-resistant variants that arise during treatment.40,51,52 In addition, in patients who had failed on therapy before, only resistance mutations that were attributable to prior nonsuppressive therapy were found, which were already present before commencement of suppressive HAART.24,25 Our sequence analysis, which was performed at the population level, did not provide any evidence of viral replication during long-term successful HAART. In support of this finding are genotypic studies performed at the clonal level, which also did not detect any resistance mutations in plasma HIV-1 sequences obtained from treatment-naive patients who have been on successful HAART for up to 5 years.24,25 However, our results of the allele-specific RT-PCR assay did detect the appearance of lamivudine-resistant HIV in 1 of 5 patients who never experienced detectable viremia during treatment. Thus, it can be concluded that during 8.3 years of suppressive HAART, signs of viral replication could be detected in 20% of our patients. In addition, it should be noticed that this study deals with the blood compartment and that absence of drug-resistant virus in blood samples does not rule out the occurrence of viral replication in tissue compartments such as brain and lymph nodes.
Phylogenetic reconstruction of protease and RT DNA sequences revealed presence of a temporal structure between sequences within our patients. Recently, significant viral genetic evolution has been reported in proviral HIV envelope sequences in patients on successful HAART for 5 years.53 Thus it seems that HAART can select HIV variants in cellular reservoirs under conditions that do not allow viral replication. When a patient starts treatment with HAART, viral replication is forestalled and only the integrated form of HIV-1 DNA will remain. With the disappearance of all unintegrated forms, much variation disappears. Selection of HIV variants on the long term might be the consequence of linkage of HIV to CD4+ T-cell development. As cellular activation triggers virus-induced cytotoxic events, even during effective HAART, proviral DNA sequences most likely disappear from naive and memory cells that are programmed for further cell divisions. Memory CD4+ T cells infected with HIV at the end stage of T-cell development and division go through a limited number of virus-induced cytotoxic events and thus contribute to the relatively high HIV DNA load in the terminally differentiated memory CD4+ T cells.
In conclusion, during long-term suppressive HAART, HIV is able to persist in terminally differentiated CD4+ T cells as proviral DNA of which the evolution is restricted to a minimum.
RESULTS
Characteristics of the Study Population
The clinical characteristics of the studied patients who were treated with HAART for up to 8.9 years are shown in Table 1. The mean ± SD plasma HIV-1 RNA level was 5.1 ± 0.5 log10 copies per milliliter at the start of HAART. Under HAART, the viral load declined to below the detection limit of 50 copies per milliliter and remained undetectable for an average of 8.3 years in patients 1-5, 7, and 8. The CD4+ T-cell counts of the patients increased significantly from an average of 293 ± 199 to 803 ± 277 cells per microliter during therapy (P = 0.001, paired t test), suggesting restoration of the immune system (Table 1).
Persistence of HIV-1 in Aged Memory Cells (CD45RO+CD57+) TOP
The most important reported cellular reservoir in which HIV-1 resides is the pool of CD4+ memory cells. However, the distribution of proviral HIV-1 DNA over the cellular compartments in this pool is not fully appreciated yet. Therefore, highly purified CD4+ T cells were obtained after 8.3 years of HAART by magnetic cell sorting (Fig. 1B). These cells were stained with monoclonal antibodies directed against CD27 and CD45RO to sort the naive CD4+ T cells (CD27+CD45RO-) by FACS (Fig. 1C). Cells were also stained with monoclonal antibodies directed against CD45RO and CD57 to sort the memory CD4+ T cells (CD45RO+CD57-) and terminally differentiated memory CD4+ T cells (CD45RO+CD57+) by FACS (Fig. 1D). CD57 has been described as a marker for terminally differentiated T cells, a subset of the memory pool.34,35 Levels of HIV-1 DNA were quantified in each cell population and did show considerable variation among the patients, especially in the naive T-cell population (Table 2). The mean viral HIV-1 DNA load was 3.0 ± 1.3 log10 per 106 CD4+ T cells in naive CD4+ T cells, 2.9 ± 0.8 log10 per 106 CD4+ T cells in memory CD4+ T cells, and 4.5 ± 1.2 log10 per 106 CD4+ T cells in terminally differentiated memory CD4+ T cells (Table 2). HIV-1 DNA load was significantly higher in the terminally differentiated memory CD4+ T cells as compared with the naive (P = 0.008, paired t test) and memory CD4+ T cells (P = 0.009, paired t test). No statistically significant difference (P = 0.8, paired t test) in HIV-1 DNA load was observed between naive and memory CD4+ T cells (Table 2)..
Drug Resistance Mutations in RT and Protease
Appearance of drug resistance mutations and viral evolution during HAART was monitored by sequencing the protease and RT genes. Sequences obtained from plasma and PBMC before therapy were therefore compared with sequences from PBMC obtained after an average of 8.3 years of HAART. The sequences were screened for known drug resistance mutations published by the International AIDS Society.32
Sequences obtained from plasma and PBMC before therapy showed minor drug resistance mutations in the protease gene, such as the L63P, L63A, V77I, and I93L mutations, in all patients (Table 3). Although these mutations are associated with resistance to protease inhibitors, they do not confer resistance by themselves and generally emerge later than major mutations. Indeed, the L63P, L63A, V77I, and I93L mutations are frequently observed polymorphisms in HIV-1 subtype B.36 These mutations could also be detected in PBMC obtained after 8.3 years of HAART in 8 of 9 patients (Table 3). Finally, however, no major resistance mutations in protease did appear after 8.3 years of HAART in all patients suggesting that drug resistance development to protease inhibitors did not occur, at least as could be determined by population sequencing.
Sequences obtained from plasma and PBMC before therapy showed no major drug resistance mutations in RT (Table 4). In addition, the appearance of amino acid substitutions in RT conferring drug resistance mutations were not detected in PBMC obtained after 8.3 years of HAART in 6 of 8 patients. In patient 6 and 9, however, both the wild type and the K219Q drug-resistant phenotype could be detected. In addition, in patient 6, also the L210W drug-resistant phenotype could be detected, whereas in patient 9, also the K103N drug-resistant phenotype could be detected. Drug resistance development in patient 9 could be attributed to adherence problems and long-term virological failure.
To detect extreme low levels of amino acid substitutions during HAART, an allele-specific PCR for codon 184, which is critical for resistance to lamivudine, was performed on plasma and PBMC. Codon 184 was selected because all patients have been treated continuously with lamivudine, and M184 substitutions appear rapid during lamivudine treatment conferring high-level resistance to this drug.37 Both the M184V and M184I substitutions could be detected in PBMC obtained after 8.3 years of HAART in patient 5 at a frequency of, respectively, 5.3% and 9.7%, in patient 6 at a frequency of 3.4% and 2.9%, and in patient 9 at a frequency of 4.1% and 3.6%. In contrast, these M184 substitutions were absent in patients 1, 2, 3, and 8 and were not able to be determined in the 2 remaining patients. Finally, the wild-type M184 phenotype was present at a frequency of 99% in clinical specimens of each patient obtained before HAART.
Evolution of RT and Protease After 8.3 Years of HAART
Amino acid sequences of protease (Table 3) and RT (Table 4) were nearly identical within each individual patient. Therefore, phylogenetic analysis was performed on the underlying population sequences to determine genetic distances and presence of temporal structures between plasma and PBMC samples obtained for each individual patient. Phylogenetic reconstruction of protease and RT DNA sequences revealed small genetic distances between sequences obtained from individual patients suggesting absence of viral evolution (Fig. 2). However, in 5 of 6 patients who could be completely analyzed, the genetic distance between plasma and PBMC-derived sequences obtained before HAART is smaller than the distance between sequences obtained before and after 8.3 years of HAART. The presence of this temporal structure between the sequences within these patients might be indicative for the occurrence of viral evolution, albeit minimal, during effective HAART. However, proviral DNA sequences obtained before and after HAART differed less than 0.5% in 4 of 7 patients indicating that the occurrence of changes in proviral DNA sequences within PBMC is restricted during long-term effective HAART (Fig. 2).
PATIENTS AND METHODS
Patients and Specimen Collection
Nine patients, who had previously participated in the CHEESE study,29 were enrolled in this study. All patients were therapy naive before onset of HAART. To determine the efficacy of HAART on viral replication and production, plasma viral load was determined on a monthly basis in the first year of treatment and thereafter once in every 3 months. Eight patients did maintain undetectable levels of plasma viremia (<50 copies/mL) after initiation of HAART until the end of the study period, with the exception of patient 5 who did have detectable viral load during 2 transient episodes of HIV-1 viremia (blip) in the first year of HAART. In the first blip, 2 subsequent measurements of approximately 1000 copies per milliliter were detected, whereas the second blip was characterized by one such measurement. These patients were treated with clinically effective HAART for on average 8.3 years (Table 1). Patient 9, however, did experience virological failure for 9 months due to treatment interruption after 5.7 years HAART. In addition, HIV-1 viremia of low magnitude (90 copies/mL) could be detected 8.9 years after starting HAART (Table 1). Written and signed informed consent was collected from all patients, and the study protocol was approved by the Medical Ethics Committee of the University Medical Center in Utrecht.
Blood samples have been collected and stored since 1997. After an average of 8.3 years of HAART, all patients donated up to 500 mL blood. PBMCs were isolated from whole blood using Ficoll-Hypaque gradient and cryopreserved.
Quantification of Viral Load
Cellular DNA was extracted from thawed PBMCs using a Qiamp DNA blood mini kit (Qiagen, Venlo, The Netherlands) according to the manufacturer's protocol. Viral load of cellular DNA was determined by Taqman real-time polymerase chain reaction (PCR) for gag using Taqmix (Applied Biosystems, Foster City, CA), 7.5 μM primers gag forward (5'-TGGGTAAAAGTAGTAGA AGAGAAGGCTTT-3') and gag reverse (5'-TGTGTTTAGCATGGTGGTTAAGTC TTG-3'), and 5 μM probe 6-FAM-TACCCATGTTTTCAGCATTATCAGAAGGAGCCAC-TAMRA. Amplification was performed during 45 cycles at 95ºC for 15 seconds and 60ºC for 1 minute. For quantitative analysis, gag PCR was performed simultaneously on U1 cells to construct a standard curve. Copy numbers per CD4+ T cell were computed using flow cytometric cell counts of the proportion of CD4+ T cells in a blood sample drawn simultaneously with the sample from which DNA was isolated from.
RNA in plasma was isolated according to the method of Boom et al.30 Plasma HIV-1 RNA load was measured by Roche's Ultrasensitive assay, which has a detection limit of 50 copies per milliliter.
Flow Cytometry and Cell Sorting
Cryopreserved PBMCs were thawed and enriched for CD4+ T cells by magnetic bead separation according to manufacturer's instructions (Miltenyi Biotec Inc, Sunnyvale, CA). We have chosen to start with the isolation of CD4+ T cells from PBMCs because these cells express the primary HIV receptor and thus contribute predominantly, if not only, to the viral reservoir present in PBMCs. To get rid of the monocytes, which do express very low amounts of CD4, forward and side scatters were used to select the lymphocytes from this population of mononuclear cells (Fig. 1A). To purify CD45RO-CD27+ (naive), CD45RO+CD57- (memory) and CD45RO+CD57+ (terminally differentiated memory) CD4+ T lymphocytes, the CD4+ enriched fraction was stained with fluorescein isothiocyanate (FITC)-conjugated monoclonal antibody (MoAb) directed to CD57 (Becton Dickinson (BD), San Jose, CA), peridinin chlorophyll protein (PERCP)-conjugated MoAb directed to CD27 (BD), PERCP-conjugated MoAb directed to CD4 (BD) and allophycocyanin (APC)-conjugated MoAb directed to CD45RO (BD Pharmingen). The specified cell fractions were isolated by fluorescence-activated cell sorting (FACS) on a FACSAria high speed cell sorter (BD). Because it might be expected that during effective HAART, proliferating cells could also serve as viral reservoir, both resting and proliferating CD4+ T cells were included in the studied memory and naive T-cell subsets to obtain overall numbers of the viral load. Postsort analysis of memory and naive CD4+ T cells revealed a mean purity of 95%.
Sequencing
Both RT and protease genes were bulk amplified as described before.31 The RT gene was amplified by PCR using primers RT18 (5'-GGAAACCAAAAATGATAGGGGGAATT GGAG-3') and RT21 (5'-CTGTATT TCTGCTATTAAGTCTTTTGATGGG-3'). A nested PCR was then performed on the first-round product using primers RT19 (5'-GGACATAAAGCTATAGGTACAG-3') and RT20 (5'-CTGCCAGTTCTAGCTC TGTGCTTC-3').31 The protease gene was amplified using primers 5'clea-1 (5'-GATGACAGAAACCTTGTTGGTCC-3') and 3'prot-1 (5'-GCAAATACTGGA GTATTGTATGGATTTTCAGG-3'). For a nested PCR, the primers 5'clea-2 (5'-AAATGATGACAGCATGTCAGGG-3') and 3'prot-2 (5'-AATGCTTTTATTTTTTCTT CTGTCAATGGC-3') were used.31 RNA was reverse transcribed and bulk amplified in a 1-step RT-PCR using the same primers as used for DNA amplification. After purification of the product, a sequence PCR was performed using the BigDye Terminator v1.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA) with primers RT19, RT20, BRT (5'-GGGATGGAAAGGATCACC-3'), and BRRT (5'-GGTGATCCTTTCCATCC-3') for the sequencing of RT and primers PR2 (5'-CTTTTGGGCCATCCATTC-3') and PR5 (5'-AGCC AACAGCCCCACCAG-3') for the sequencing of protease..
The population sequences were analyzed with SeqScape 2.1.1 (Applied Biosystems) and screened for known drug resistance mutations published by the International AIDS Society.32
Allele-Specific PCR Assay for Codon 184 of RT
To detect low-frequency selection of lamivudine-resistant variants, we analyzed 20 patient samples with an allele-specific RT-PCR assay that quantifies variants encoding 184I or 184V to a frequency of 0.1%. A region of HIV-1 RT from bases 2195 to 2818 of the HIV-1 genome was amplified and quantified using real-time PCR from a RT amplicon of each patient virus using primers 2195F (AAACAATGGCCATTGACAGAAGA) and 2818R (CCAAAGGAATGGAGGTTCTTTCTG). This DNA product, encompassing codons 22-229 of RT, was diluted to approximately 107 copies per reaction, and used as template for a second round of PCR, which was performed using primers that discriminate between the 184M, I, or V allele. Second-round PCR product was detected using SYBR green fluorescence.33 To confirm amplification specificity, PCR products were subjected to thermal denaturation analysis.
Statistical Analysis and Phylogenetic Trees
Differences between viral load before therapy and after 8.3 years of HAART and between the sorted fractions were tested for significance by the paired t test. The P values less than 0.05 were regarded as statistically significant.
A phylogenetic tree was inferred from nucleotide sequences of the protease and RT gene through the use of Molecular Evolutionary Genetics Analysis program, version 2.1. The Kimura 2 model was used to construct a neighbour-joining tree.
|
|
| |
| |
|
|
|
