Second-phase half-lives were calculated using linear regression on HIV RNA log10 copies/ml from the second time point (day 15 on average), until the first point where HIV RNA fell below the 50 copies/ml limit. The median second-phase half-life for raltegravir individuals was 15.5 days [interquartile range (IQR), 14], whereas it was 18.3 days (IQR, 10) for those on efavirenz, but these half-lives were not significantly different (P = 0.2). The slightly numerically faster second-phase dynamics for the raltegravir group may be because some of these individuals had a first-phase decline that was prolonged beyond day 15. This appears likely given the extended first-phase decline exhibited under monotherapy. Including only those individuals on raltegravir who had detectable HIV RNA at days 15 and 29 (n = 42), gave a median second-phase half-life of 19.6 days, which also was not significantly different to the efavirenz group (P = 0.6).
Although the rate of decline during the second phase did not differ between the raltegravir and efavirenz regimens, the point at which this decline initiated did differ. The significantly lower viral levels in plasma from day 15 onwards for the raltegravir group meant that the first phase of viral decay was more extensive, thereby reducing the viral level at which the second phase commenced (Fig. 1). Estimation of the size of the viral compartment contributing to second-phase viral decay was calculated by extrapolating the regression line of this second phase back to baseline. The ratio of this baseline second-phase value to baseline HIV RNA (‘M ratio’), has been listed for six individuals in Perelson et al.  (Table 1, pM*0/cV0). In that study combination therapy with nelfinavir, zidovudine and lamividine resulted in M ratios with a mean of 0.04 and median of 0.03. The lower the M ratio the more extensive was first phase decline. In the present study, M ratios for the efavirenz group were lower but not statistically different to values reported in the literature, with the higher literature values perhaps arising from a slightly different method where a viral load curve was fitted to the data in that study  (P = 0.08). The M ratio was 70% lower for individuals taking raltegravir than efavirenz (median 0.0042 versus 0.014; P < 0.0001), implying that raltegravir reduces viral production from the second-phase source by this amount over and above standard regimens.
The decrease in the second-phase viral production with the addition of raltegravir necessitates rethinking current theories of the source of virus responsible for the second phase. Current hypotheses are that second-phase viral load arises through production by long-lived infected cells , release from follicular dendritic cells  or decreased cytotoxic lymphocyte response . None of these theories can explain the effects produced by raltegravir, if it is acting as a pure integrase inhibitor, since they suggest that second-phase viral load originates from sources that an integrase inhibitor would not affect: previously infected cells or virus already produced but bound to dendritic cells.
Hypotheses for second-phase virus
If an integrase inhibitor is to influence viral levels in the second phase, the virus at that stage must arise from (i) cells that are either newly infected and where the integrase inhibitor provides an additional benefit above that of reverse transcriptase and protease inhibitors, (ii) cells that have been previously infected but where the proviral DNA has yet to be integrated into the host cell genome, or (iii) a sanctuary site where the integrase inhibitor has improved penetration. To investigate the viability of these hypotheses as explanations of the reduced second phase under an integrase inhibitor, three mathematical models were developed and applied to the median data derived from the monotherapy and combination trials with raltegravir and efavirenz. Model 1 simulated the first hypothesis above, model 2 simulated the second, and model 3 simulated the third (supplementary text). A cartoon of the effects of raltegravir in these models is displayed in Fig. 2.
In model 1, new productive infection (I), which generates virus (V) in the second phase, originates from long-lived infected cells (M). These latter cells decay at the slow rate observed in this phase, thereby providing this same dynamic to newly infected cells and second-phase virus (Fig. 3a,b). The integrase inhibitor provides an additional effect to reverse transcriptase and protease inhibitors in their inhibition of new productive infection from long-lived infected cells, decreasing second-phase viral levels more (solid lines Fig. 3a,b) than with efavirenz (dashed lines).
The results achieved with the integrase inhibitor in model 1 could theoretically be achieved through more potent reverse transcriptase or protease inhibitors. By comparison, model 2 simulates a process of infection that is not susceptible to traditional drugs, where virus from the second phase arises from latently infected cells that contain proviral DNA in a full-length unintegrated form, which has been estimated to form the vast majority of latent infection . In this model, second-phase virus originates from latently infected cells with unintegrated viral DNA (D) that are activated and converted to productively infected cells (I). Increased potency of traditional drugs will reduce the replenishment of the latent pool D but will not stop its conversion to productive infection. Model 2 also provides a consistent explanation for the effect achieved by the integrase inhibitor (Fig. 3c,d).
The third model, where the differential effects of the integrase inhibitor are achieved through better penetration of a sanctuary site, was less successful in duplicating the data (Fig. 3e,f). Although it achieved lower second-phase viral load, the slope of this second phase was significantly faster for the integrase inhibitor (solid line Fig. 3e) than for the efavirenz arm (dashed line Fig. 3e). This was at odds with the results described above, where there was no significant difference. The second-phase virus in this model traffics from the sanctuary site, where it is produced by productively infected cells. In order to achieve this differential effect for the integrase inhibitor, the efavirenz arm must have extremely low efficacy, 6%, in order for the additional efficacy provided by the integrase inhibitor to significantly perturb the decay rate of productively infected cells in the sanctuary site (green solid line versus green dashed line Fig. 3f).
Raltegravir significantly extended the first phase of viral decay and reduced the plasma HIV RNA level at which the second phase commenced (Fig. 1). It enabled individuals to suppress HIV RNA to below detection limits significantly faster than current regimens (Table 1). The 50 copies/ml assay limit did not allow investigation of any effects this may have on subsequent viral levels. By comparing M ratios for individuals enrolled on the raltegravir arm with those on the efavirenz arm, raltegravir achieved a 70% reduction of second-phase viral levels.
The effect of raltegravir on the second phase of viral decay was unexpected given current hypotheses for the processes underlying this phase [6,11,12] and the mechanism by which the drug is thought to work. We proposed three hypotheses to explain the observed effects and compared simulations of mathematical models for these hypotheses to median data for ART with or without an integrase inhibitor. Models 1 and 2 successfully reproduced the clinical data (Fig. 3).
The effect of adding an integrase inhibitor to model 1 is to provide an additional effect to that of the reverse transcriptase inhibitors (and any protease inhibitors that act in the model through decreasing infectivity), thereby decreasing new infection from virus associated with long-lived infected cells. Hence under these assumptions, the integrase inhibitor does not target new infection mechanisms but reduces the success of transmission from long-lived infected cells to productively infected cells. Model 2, where the second phase is assumed to arise from activation of latently infected cells that have HIV DNA in a predominantly unintegrated form, provides a mechanism that is completely resistant to any additional benefit from improvements in reverse transcriptase inhibition. For this second model, the integrase inhibitor provides a protective mechanism that cannot be replaced by reverse transcriptase or protease inhibitors, no matter their potency, since for reverse transcriptase and protease inhibitors new virions will be produced, although these will be mostly noninfectious in the presence of protease inhibitors . Under each of these models, there may be an increase in aborted HIV DNA integration and a resulting rise in extrachromosomal viral DNA in circular forms containing one or two long terminal repeats (2LTR circles), as observed in vitro . In those experiments, cell viability was not affected, and clinical trials with raltegravir have not shown toxicities higher than with current antiretroviral drugs [3,4]. Therefore, any resulting increases in unintegrated HIV DNA would not be expected to be detrimental to these cells. Although not conclusive, the simulations of model 3 suggest that second-phase virus does not originate from a sanctuary site where raltegravir is more successful in penetrating than standard antiretroviral drugs.
Availability of protease inhibitors in the 1990s and their combination with reverse transcriptase inhibitors evoked hopes that this new potent ART would at last allow clearance of HIV infection. Estimates were made on the length of treatment needed before this could be achieved . Although plasma levels of HIV RNA fall below detection limits within several months of ART, it was soon realized that latent infection provided a lower bound on the rate at which HIV infection could be removed from the body. This component is established early in infection  and consists of both unintegrated and integrated HIV DNA even under ART . The high levels under ART of unintegrated HIV DNA in resting CD4 T cells , and proviral sequence evolution in individuals with undetectable HIV RNA , suggest that there is persistent viral replication even in the presence of seemingly successful ART. Regardless of whether ongoing infection or the slow turnover of resting CD4 T cells is primarily responsible, the latent pool decreases slowly under standard ART, with a half-life of anywhere between 6 months [16,17] and 44 months . This leads to estimates of between 8 years [16,18] and more than 60 years for clearance of infection. The rate of decrease is dependent on residual viral replication , and hence drugs like raltegravir that increase the effectiveness of ART, especially for viral components such as unintegrated HIV DNA that are immune to current drugs, may improve this rate of decay.
There is considerable uncertainty as to the prevalence of unintegrated HIV DNA in resting CD4 T cells during ART, and the impact this component makes to renewing the integrated pool. In-vitro experiments have established that HIV can enter resting CD4 T cells with high efficiency, but that productive infection does not ensue. The cause of this failure seems dependent on the cell culture conditions, as infection of resting cells can fail during reverse transcription [19,20] or through the lability of the full-length unintegrated HIV DNA transcript , or it can lead to stable infection that is transcriptionally inactive without cellular activation . The in-vivo situation under ART is equally unclear. Total HIV DNA exceeds integrated HIV DNA in resting CD4 T cells by 100-fold, yet does not seem to contribute significantly to replication competent virus . The presence of unintegrated HIV DNA under ART, however, indicates either that there is substantial ongoing infection, if this molecule is labile, or that it is long lived. If it is the latter, then integrase inhibitors can provide an effective block in new productive infection upon activation of this latent component, as hypothesized in model 2. Regardless, integrated HIV DNA must undergo some replenishment since it decays slowly under ART but rebounds within weeks upon cessation of therapy . It is difficult to reconcile these very different dynamics if there is no compensatory loss in this compartment, and hence continual replacement. Detailed analysis of integrated and unintegrated HIV DNA in resting CD4 T cells, in the presence and absence of an integrase inhibitor, would provide valuable information on this matter.
We thank the patients and investigators who participated in this study.
Sponsorship: The National Centre in HIV Epidemiology and Clinical Research is funded by the Australian Commonwealth Department of Health and Ageing and is affiliated with the Faculty of Medicine at the University of New South Wales. AK is supported by a Practitioner Fellowship from the NHMRC.
Note: The P004 study is registered at ClinicalTrials.gov under identifier NCT00100048.
Note: Supplementary material can be accessed online.
1. DeJesus E, Berger D, Markowitz M, Cohen C, Hawkins T, Ruane P, et al
. Antiviral activity, pharmacokinetics, and dose response of the HIV-1 integrase inhibitor GS-9137 (JTK-303) in treatment-naive and treatment-experienced patients. J Acquir Immune Defic Syndr 2006; 43:1–5.
2. Hazuda DJ, Felock P, Witmer M, Wolfe A, Stillmock K, Grobler JA, et al
. Inhibitors of strand transfer that prevent integration and inhibit HIV-1 replication in cells. Science 2000; 287:646–650.
3. Markowitz M, Morales-Ramirez JO, Nguyen BY, Kovacs CM, Steigbigel RT, Cooper DA, et al
. Antiretroviral Activity, pharmacokinetics, and tolerability of MK-0518, a novel inhibitor of HIV-1 integrase, dosed as monotherapy for 10 days in treatment-naive HIV-1-infected individuals. J Acquir Immune Defic Syndr 2006; 43:509–515.
4. Grinsztejn B, Nguyen BY, Katlama C, Gatell JM, Lazzarin A, Vittecoq D, et al
. Safety and efficacy of the HIV-1 integrase inhibitor raltegravir (MK-0518) in treatment-experienced patients with multidrug-resistant virus: a phase II randomised controlled trial. Lancet 2007; 369:1261–1269.
5. Plipat N, Ruan PK, Fenton T, Yogev R. Rapid Human immunodeficiency virus decay in highly active antiretroviral therapy (HAART)-experienced children after starting mega-HAART. J Virol 2004; 78:11272–11275.
6. Perelson AS, Essunger P, Cao Y, Vesanen M, Hurley A, Saksela K, et al
. Decay characteristics of HIV-1-infected compartments during combination therapy. Nature 1997; 387:188–191.
7. Kuritzkes DR, Ribaudo HJ, Squires KE, Koletar SL, Santana J, Riddler SA, et al
. Plasma HIV-1 RNA dynamics in antiretroviral-naive subjects receiving either triple-nucleoside or efavirenz-containing regimens: ACTG A5166s. J Infect Dis 2007; 195:1169–1176.
8. Palmisano L, Giuliano M, Nicastri E, Pirillo MF, Andreotti M, Galluzzo CM, et al
. Residual viraemia in subjects with chronic HIV infection and viral load < 50 copies/ml: the impact of highly active antiretroviral therapy. AIDS 2005; 19:1843–1847.
9. Ho DD, Neumann AU, Perelson AS, Chen W, Leonard JM, Markowitz M. Rapid turnover of plasma virions and CD4 lymphocytes in HIV-1 infection. Nature 1995; 373:123–126.
10. Wei X, Ghosh SK, Taylor ME, Johnson VA, Emini EA, Deutsch P, et al
. Viral dynamics in human immunodeficiency virus type 1 infection. Nature 1995; 373:117–122.
11. Hlavacek WS, Stilianakis NI, Notermans DW, Danner SA, Perelson AS. Influence of follicular dendritic cells on decay of HIV during antiretroviral therapy. Proc Natl Acad Sci USA 2000; 97:10966–10971.
12. Arnaout RA, Nowak MA, Wodarz D. HIV-1 dynamics revisited: biphasic decay by cytotoxic T lymphocyte killing? Proc R Soc Lond, Ser B Biol Sci 2000; 267:1347–1354.
13. Chun TW, Stuyver L, Mizell SB, Ehler LA, Mican JA, Baseler M, et al
. Presence of an inducible HIV-1 latent reservoir during highly active antiretroviral therapy. Proc Natl Acad Sci USA 1997; 94:13193–13197.
14. Ashorn P, McQuade TJ, Thaisrivongs S, Tomasselli AG, Tarpley WG, Moss B. An inhibitor of the protease blocks maturation of human and simian immunodeficiency viruses and spread of infection. Proc Natl Acad Sci USA 1990; 87:7472–7476.
15. Chun TW, Engel D, Berrey MM, Shea T, Corey L, Fauci AS. Early establishment of a pool of latently infected, resting CD4(+) T cells during primary HIV-1 infection. Proc Natl Acad Sci USA 1998; 95:8869–8873.
16. Zhang L, Ramratnam B, Tenner-Racz K, He Y, Vesanen M, Lewin S, et al
. Quantifying residual HIV-1 replication in patients receiving combination antiretroviral therapy. N Engl J Med 1999; 340:1605–1613.
17. Ramratnam B, Mittler JE, Zhang L, Boden D, Hurley A, Fang F, 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 antiretroviral therapy. Nat Med 2000; 6:82–85.
18. Finzi D, Blankson J, Siliciano JD, Margolick JB, Chadwick K, Pierson T, 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 1999; 5:512–517.
19. Zack JA. The role of the cell cycle in HIV-1 infection. Adv Exp Med Biol 1995; 374:27–31.
20. Zhou Y, Zhang H, Siliciano JD, Siliciano RF. Kinetics of human immunodeficiency virus type 1 decay following entry into resting CD4+ T cells. J Virol 2005; 79:2199–2210.
21. Spina CA, Guatelli JC, Richman DD. Establishment of a stable, inducible form of human immunodeficiency virus type 1 DNA in quiescent CD4 lymphocytes in vitro. J Virol 1995; 69:2977–2988.
Keywords:© 2007 Lippincott Williams & Wilkins, Inc.
antiviral therapy; HIV-1; integrase inhibitors; mathematical models; raltegravir; viral load