Treatment of chronic HIV-1 infection with highly active antiretroviral therapy (HAART) allows a drastic reduction of plasma HIV-1 RNA load, responsible for immune restoration. However, virus eradication is not a realistic goal because viral reservoirs established in body and cell compartments are either not susceptible to the action of antiretroviral drugs or characterized by long half-life and slow turnover . Measurement of HIV-1 DNA in peripheral blood mononuclear cells (PBMC) gives access to the viral reservoir, since cell-associated HIV-1 DNA reflects both actively and latently infected cells .
Here, the evolution of the HIV-1 DNA load and of the CD4 cell count was assessed in HIV-1-infected patients selected for maximally successful viral suppression on HAART in the long term, which allows for maximal immune restoration . These parameters are relevant to the issue of treatment discontinuation or maintenance in patients experiencing virological and immunological success after several years on HAART.
Patients with HIV-1 infection followed at a single institution (Necker Hospital) were retrospectively selected on the basis of (i) undetectable plasma HIV-1 RNA load (< 400 or < 50 copies/ml) within 6 months of starting their first HAART regimen between May 1996 and April 1997; (ii) undetectable plasma HIV-1 RNA since month 6 of HAART, without either replication (`blips') > 400 copies/ml or treatment interruption; and (iii) follow-up for more than 48 months. Sampling once a year for HIV-1 DNA measurement was carried out in patients who gave agreement. All patients were prospectively evaluated clinically and for CD4 cell count and plasma HIV-1 RNA [median interval between visits, 3 months; median number of measurements, 20, interquartile range (IQR), 17–21]. Informed consent was obtained from study participants.
From May 1996 to April 1997, 217 subjects started HAART. Among them, 13 died, 19 were lost to follow-up and 41 (18.9% in an intent-to-treat analysis) met the very stringent criteria set for this study of permanent plasma HIV-1 RNA suppression; of these, 25 were tested for HIV-1 DNA.
Quantification of HIV-1 RNA in plasma
Blood was sampled in ethylenediaminetetraacetic acid (EDTA) and plasma was frozen at −80°C within 6 hours of collection. Plasma HIV-1 RNA was measured with the Amplicor HIV Monitor assay (Roche Diagnostic Systems, Neuilly, France). From September 1997, samples were systematically tested using the ultrasensitive procedure (detection limit 50 copies/ml).
Quantification of HIV-1 DNA in peripheral blood mononuclear cells
Blood samples collected in EDTA were separated into plasma and cells by Ficoll–Hypaque density gradient centrifugation (Eurobio, Les Ulis, France) and dry pellets of 106 PBMC were stored at −80°C. Total cell DNA was extracted from pellets by adsorption on silica membranes (Qiagen, Hilden, Germany) and stored at −80°C. HIV-1 DNA was quantified by real-time polymerase chain reaction, using 5′ nuclease assay in the long terminal repeat (LTR) gene . The reaction was performed with ABI PRISM 7000 (Applied Biosystems, Courtaboeuf, France). Briefly, 1 μg DNA was amplified with the sense primer NEC 152 (GCCT CAATAAAGCTTGCCTTGA) and the reverse primer NEC 131 (GGCGCCACTGCTAGAGATTTT) in the presence of a dually (FAM and TAMRA) labelled NEC LTR probe (AAGTAGTGTGTGCCCGTCTG TTRTKTGACT). The first PCR cycle allowing fluorescence detection permitted to quantify HIV-1 DNA by reference to a standard curve (fivefold dilutions of 8E5 cell DNA). All samples from each patient were tested in the same assay. Results were expressed as number of DNA copies/106 PBMC.
CD4 cell count
Absolute CD4 cell count was determined by flow cytometry on a FACScan (Becton-Dickinson, San Jose, California, USA) equipped with the CellQuest software.
A chi-square or a Mann–Whitney U test was used for categorical or continuous variables, respectively, in comparison of two groups. Trajectories of CD4 cell counts and log10-transformed HIV-1 DNA levels were described by using the Loess procedure, which draws a representative smooth curve through data by using robust local regressions . For each marker, a piecewise linear regression analysis was performed with one or two change points to estimate and test mean slopes at different periods after HAART initiation. Time at which changes occurred were determined according to the inflexion points observed in the Loess curves. All tests were two-sided. Statistical analysis was performed by use of Statview (Abacus Concepts, Berkeley, California, USA) and SAS v.8.1 (SAS Institute, Cary, North Carolina, USA).
Table 1 shows the characteristics of the 41 patients on HAART for a median of 60.4 months (range, 50–71) who entered the study. Two died at month 61 of HAART, one of terminal liver failure and one of lymphoma, but still had undetectable plasma HIV RNA loads. All others were clinically well, except for the lipodystrophy syndrome, present in 58.5%.
All studied patients started with a regimen including two nucleoside analogues and one protease inhibitor, indinavir in 93%. For intolerance reasons, some patients switched to another protease inhibitor or a non-nucleosidic reverse transcriptase inhibitor, and some changed nucleoside analogues. At time of analysis, 48% of patients were still on their first-line regimen, and 87% were taking a protease inhibitor.
The 25 patients studied for HIV-1 DNA were slightly younger and had moderately lower CD4 cell count nadir, plasma viral load zenith and pre-HAART plasma RNA level than the 16 patients who were not sampled for HIV-1 DNA. Otherwise, they did not differ in terms of medical and therapeutic history.
Loess estimates of CD4 cell count on HAART (Fig. 1) showed a marked increase during the first 18 months, much lower afterwards. Using a piecewise linear regression analysis, the following estimations of the mean CD4 cell count slopes were obtained: 168 × 106 cells/l per year (SD, 20) during the first 18 months and subsequently 38 × 106 cells/l per year (SD, 9). CD4 cell count slopes differed significantly from zero (P < 10−4 at each period) and the increase was significantly lower after than before the 18th month (P < 10−4). The CD4 cell count increase was similar for patients in the upper and lower halves of pre-HAART CD4 cell count values (median, 135 × 106 cells/l), before (P = 0.30) as well as after (P = 0.24) the 18th month.
The proportion of patients reaching a CD4 cell count constantly ≥ 400 × 106/l until the end of follow-up was analysed (patients with at least two measurements during the period were considered, and all values ≥ 400 × 106/l): 30 patients out of 41 (73.2%) reached this endpoint during follow-up; 11 (52.4%) and 19 (95 %) of the patients with baseline CD4 cell counts below and above the median, respectively (P < 0.01). A CD4 cell count of ≥ 400 × 106/l was achieved in 10 of the 41 patients (24.4%) between months 6 and 18 of HAART: two (9.5%) with baseline CD4 cell counts below the median and eight (40%) with counts above the median (P = 0.04). Between months 30 and 42, 23 patients out of 41 (56.1%) met this endpoint [five (23.8%) with baseline CD4 cell counts below the median and 18 (90%) with counts above the median; P < 0.001], and between months 54 and 66, 23 patients out of 33 (69.7%) fulfilled this criterion [8 out of 17 (47%) with baseline CD4 cell counts below the median and 15 out of 16 (94%) with counts above the median; P < 0.01]. Thirty-six of the patients (87.5%) who had CD4 cell counts ≥ 400 × 106/l between months 6 and 18 kept this status at 3 years; 38 (93.8%) of the patients who reached ≥ 400 × 106/l at year 3 still had CD4 cell counts ≥ 400 × 106/l at year 5.
Evolution of HIV-1 DNA during therapy
The Loess curve of HIV-1 DNA decrease on HAART (Fig. 1) showed mild inflexions after 1 year and after 3 years, and this overall pattern applied for all patients. Introducing two change points in the regression model, the following estimations were obtained: mean decrease of 0.48 log10 copies/106 PBMC (SD, 0.19) during the first year, 0.18 log10 copies/106 PBMC per year (SD, 0.08) during the 2nd and 3rd years, and 0.01 log10 copies/106 PBMC per year (SD, 0.07) afterwards. The slopes for the first two periods were significantly different from zero (P = 0.012 and P = 0.025, respectively), while that for the third period was not (P = 0.94). However, the difference between the first two slopes did not reach statistical significance (P = 0.20), which may be because of a lack of statistical power linked to the small size of the study group. The evolution of HIV DNA was not different for patients with baseline CD4 cell counts above or below median value.
Since HAART has no direct effect on integrated virus and is only virustatic, virus is stored in resting CD4 T cells, and a low level of viraemia persists even in patients with undetectable plasma HIV RNA load . Therefore, any beneficial impact of the long-term maintenance of HAART on the HIV-1 reservoir is elusive. The measurement of cell-associated HIV-1 DNA can be used as a tool to study the effects of HAART on the cell-associated reservoir.
The evolution of HIV-1 DNA has been studied in the short and mid term in patients taking HAART. In studies of chronic infection, HIV-1 DNA decreased by approximately 0.4 log10 copies/106 PBMC during the first year [6,7] and reached a plateau after approximately 80 weeks [2,8]. In HAART-treated primary infection, a more pronounced decrease of HIV-1 DNA has been described: 0.78 log10 copies/106 PBMC after 1 year , 1.0 log10 copies/106 PBMC after 18 months  (and even 0.9 log10 copies/106 PBMC at 48 weeks under the effect of both HAART and interferon-α ), with a slower decay rate after 3 months, linked to residual replication . At HAART interruption, HIV-1 DNA in PBMC has been shown to rise rapidly  and decrease more rapidly at treatment resumption than at first HAART prescription. This indicates that, under these circumstances, a pool of recently infected cells contributes to the level of HIV-1 DNA, which is only partly the case in chronic infection. The HIV-1 DNA level, assessed by a test measuring both integrated and unintegrated DNA (as in the present study), has also been shown to be a prognostic marker of HIV-1 infection, independent of the CD4 cell count and plasma RNA load [14,15]. Therefore, HIV-1 DNA in PBMC will reflect not only recently infected cells but also the solidly established stock of virus in the body, present since a very early step in the natural history of infection, and only partially reduced by HAART in the short and mid term, particularly in chronic infection.
The present study, although limited in its conclusions by the small number of patients, stresses the levelling off, with time, of the effects of HAART. Extending previous data [2,4,6,8], non-parametric estimation and regression analysis of HIV-1 DNA over a 5-year follow-up showed a homogeneous pattern in the patients studied: an initial decay of 0.5 log10 copies/106 PBMC during the first year was followed by a milder decrease of < 0.2 log10 copies/106 PBMC per year during the 2nd and 3rd years, without any further diminution afterwards. Maintenance of therapy beyond 3 years is necessary to avoid the replenishment of the compartment of cells actively producing virus, but any additional benefit in terms of reduction of the viral reservoir seems highly unlikely.
As previously shown , there was only a very slight gain in CD4 cells after 18 months of treatment. Interestingly, the absolute increase in the CD4 cell count was not different in the patients with CD4 baseline values below or above the median value of the population, and there was no marked further benefit in the long term in either group, but the level of the CD4 curve was different. The majority of patients had reached a CD4 cell count constantly ≥ 400 × 106/l by month 30, and nearly all of those with a baseline CD4 cell count above the median (135 × 106/l) had attained that level at the end of follow-up. For patients who have reached this level of CD4 cell counts, a reasonably safe level both for the protection against opportunistic infections and for considering treatment interruption, the immunological benefit of maintaining HAART appears debatable.
It is more and more difficult to imagine anti-HIV treatments as life-long prescriptions, given the side effects described in the long term, such as lipodystrophy (found here in nearly 60% of patients), metabolic disturbances, a possibly increased cardiovascular risk, mitochondrial toxicity and altered quality of life. In other words, the inconvenience of a very-long-term treatment may outweigh the benefit of maintaining the CD4 cell count at a high level, considering that treatment beyond 2 to 4 years will not result in a significant further reduction of the HIV-1 DNA load. For patients with high CD4 cell values (e.g. > 400 × 106/l) after this time on HAART, would it be reasonable to consider stopping therapy when the level of HIV-1 DNA reaches its lowest plateau and then wait for the patients to meet again the criteria for treatment initiation? Trials should be designed to compare this long-term strategy of prolonged periods on and off therapy with the standard attitude of maintaining HAART and with structured treatment interruptions with a shorter time scale.
In summary, the data presented here show that HIV-1 DNA does not seem influenced by HAART after the third year and confirm that the CD4 cell count gain is less apparent after 18 months on treatment. Based on these observations, we question the benefits of a life-long treatment for HIV infection.
We thank the patients for their participation in the study and the HIV quantification group (AC11) of the ANRS.
1. Blankson JN, Persaud D, Siliciano R. The challenge of viral reservoirs in HIV-1 infection.Annu Rev Med
2. Furtado MR, Callaway DS, Phair JP, Kunstman KJ, Stanton JL, Macken CA, et al
. Persistence of HIV-1 transcription in peripheral blood mononuclear cells in patients receiving potent antiretroviral therapy.N Engl J Med
3. Kaufmann R, Bloch M, Zaunders JJ, Smith D, Cooper DA. Long-term immunological response in HIV-1–infected subjects receiving potent antiretroviral therapy.AIDS
4. Burgard M, Chaix ML, Ngo N, Leruez M, Rouzioux C, and the French Perinatal HIV Infection Study Group. Real-time HIV-1 PCR for the diagnosis of infection in newborn of African mothers.Eighth Conference on Retroviruses and Opportunistic Infections.
Chicago, February 2001 [abstract 243].
5. Diggle PJ, Liang KY, Zeger SL. Fitting smooth curves to longitudinal data.
In Analysis of Longitudinal Data.
Edited by Atkinson AC, Copas JB, Pierce DA, Schervish MJ, Titterington DM. Oxford: Oxford University Press; 1994:41–47.
6. Izopet J, Salama G, Pasquier C, Sandres K, Marchou B, Massip P, et al
. Decay of HIV-1 DNA in patients receiving suppressive antiretroviral therapy.J Acquir Immune Defic Syndr Hum Retrovirol
7. Burgard M, Izopet J, Dumon B, Tamalet C, Descamps D, Ruffault A, et al
. HIV RNA and HIV DNA in peripheral blood mononuclear cells are consistent markers for estimating viral load in patients undergoing long-term potent treatment.AIDS Res Hum Retroviruses
8. Saitoh A, Hsia K, Fenton T, Powell CA, Christopherson C, Fletcher CV, et al.Persistence of human immunodeficiency virus (HIV) type 1 DNA in peripheral blood despite prolonged suppression of plasma HIV-1 RNA in children.J Infect Dis
9. Ngo-Giang-Huong N, Deveau C, Da Silva I, Pellegrin I, Venet A, Harzic M, et al.Proviral HIV-1 DNA in subjects followed since primary infection who suppress plasma viral load after one year of highly active antiretroviral therapy.AIDS
10. Garrigue I, Pellegrin I, Hoen B, Dumon B, Harzic M, Schrive MH, et al
. Cell-associated HIV-1 DNA quantitation after HAART-treated primary infection in patients with persistently undetectable plasma HIV-1 RNA.AIDS
11. Emilie D, Burgard M, Lascoux-Combe C, Laughlin M, Krzysiek R, Pignon C, et al
. Early control of HIV replication in primary HIV-1 infection treated with antiretroviral drugs and pegylated IFNγ: results from the Primoferon (ANRS 086) Study.AIDS
12. Yerly S, Perneger TV, Vora S, Hirschel B, Perrin L. Decay of cell-associated HIV-1 DNA correlates with residual replication in patients treated during acute HIV-1 infection.AIDS
13. Jubault V, Burgard M, Le Corfec E, Costagliola D, Rouzioux C, Viard JP. High rebound of plasma and cellular HIV load after discontinuation of triple combination therapy.AIDS
14. Rouzioux C. Early HIV-1 DNA level predicts disease progression and death independently of HIV-RNA level and CD4 count.First International AIDS Society Conference on HIV Pathogenesis and Treatment.
Buenos Aires, July 2001[abstract A23].
15. Kostrikis LG, Touloumi G, Karanicolas R, Pantazis N, Anastassopoulou C, Karafoulidou A, et al.Quantitation of human immunodeficiency virus type 1 DNA forms with the second template switch in pripheral blood cells predicts disease progression independently of plasma RNA load.J Virol
2002; 76: 10099–10108.
16. Tarwater PM, Margolick JB, Jin J, Phair JP, Detels R, Rinaldo C, et al.Increase and plateau of CD4 T-cell counts in the 3½ years after initiation of potent antiretroviral therapy.J Acquir Immune Defic Syndr