A Highly Intensified ART Regimen Induces Long-Term Viral Suppression and Restriction of the Viral Reservoir in a Simian AIDS Model: 'functional cure?'
Download the PDF here|
"......A short course of highly intensified ART following the previous treatment resulted, upon therapy suspension, in a remarkably spontaneous control of the infection, that may pave the way to a persistent suppression of viremia in the absence of ART.....If the results of the present study should prove reproducible in humans, H-iART could represent a useful tool for improving the viro-immunological conditions of HIV-infected individuals and a useful addition to experimental anti-reservoir strategies.......the results obtained with the present macaque model suggest that a short cycle of H-iART could be used for improving the efficacy of our previous anti-reservoir treatment based on auranofin and strengthen the idea that an arrest in disease progression may be obtained during the chronic phase of the disease."
"In the pilot study presented above, we unexpectedly found that H-iART profoundly impacted on viral DNA. First, there was a late viral DNA decay to levels below the assay detection limit which was associated with the addition of MRV to the drug cocktail (Fig. 5A). In addition, the CD4/CD8 ratio, the decrease of which is a marker of the viral reservoir and/or ongoing viral replication , , significantly increased during treatment (Fig. 5B). Of note, viral DNA in PBMCs also fell below the assay detection limit in all macaques included in the group treated with H-iART ab-initio (median treatment duration = 125 days, range from 45 to 174 days), i.e. no viral DNA copies were detectable in six out of six repeats with a threshold sensitivity of 2 copies/5*105 cells."
"Finally, given the aforementioned effects of H-iART, we tested whether this therapeutic regimen might be adopted to improve the effect of a previous anti-reservoir strategy based on the anti-memory drug auranofin in combination with antiretrovirals . Upon interruption of this anti-reservoir treatment, SIVmac251-infected macaques experience an acute infection-like condition, i.e. an initial viral load peak followed by rapid containment of viral load . The peak, which is rapidly reached upon virus re-appearance in plasma, is associated with the reconstitution of the viral reservoir, as shown by the previously published independent association between the area under the curve (AUC) describing the initial peak of viral load and the eventual viral load set point ( see also Fig. 8A). From this association, it follows that decreasing the AUC at peak artificially through a cycle of H-iART should limit the reconstitution of the viral reservoir and may result in spontaneous control of viral load following H-iART suspension."
"The experiment was attempted in two macaques. A first macaque (P252) was treated with a one-month cycle of H-iART at viral load rebound, after the suspension of the aforementioned auranofin/antiretroviral regimen. Another macaque (P177) was treated with auranofin in addition to H-iART as a follow-up of the treatment presented in the pilot study. Eventually, following therapy suspension, P177 was subjected to a short H-iART cycle at viral rebound, similar to that administered to P252. In both cases, the short H-iART cycle promptly abated viral load to levels below the assay detection limit, thus efficiently decreasing the initial AUC (Fig. 8 A-C)."
"Moreover, we could not detect viral DNA in lymph node and rectal tissue biopsies (detection limit: 2 copies/5*105 cells, three repeats per sample) in all the macaques of the pilot study tested (Table 2). Lymph node viral DNA was also below the assay detection limit in one of three macaques from those treated with H-iART ab-initio, while viral DNA was below the limit of detection in rectal biopsies of all the macaques of the same group (Table 2). The results were further validated by excluding the presence of PCR inhibitors using spiked DNA for selected samples (Table S2)."
"Results showed that H-iART decreased the memory CD4+ T-cell numbers over time (Fig. 6A,B), while it carried out no significant effect on the naïve T-cell subpopulation (Fig. 6C). This result is in accordance with the in-vitro inhibitory effect of MRV on the proliferation of sorted memory T-cell subpopulations (Fig. S7). MRV significantly decreased the numbers of activated (HLA-DR+) CD4+ TEM cells (Fig. 6D). This effect is in line with decreased levels of immune activation already observed in humans treated with this drug , . In conclusion, MRV decreased the number of memory T-cells as well as TEM cell-activation. Since these two parameters are linked to the magnitude of the viral reservoir and ongoing viral replication , , this effect is in good agreement with the aforementioned three-phase decay of viral DNA induced by MRV (Fig. 5C)."
Iart Luca Shytaj1#, Sandro Norelli1#, Barbara Chirullo1#, Alessandro Della Corte1, Matt Collins2, Jake Yalley-Ogunro2, Jack Greenhouse2, Nunzio Iraci3, Edward P. Acosta4, Maria Letizia Barreca3, Mark G. Lewis2, Andrea Savarino1*
1 Department of Infectious, Parasitic and Immune-mediated Diseases, Istituto Superiore di Sanita, Viale Regina Elena, Rome, Italy, 2 BIOQUAL, Inc., Rockville, Maryland, United States of America, 3 Dipartimento di Chimica e Tecnologia del Farmaco, Facolta di Farmacia, Universita di Perugia, Perugia, Italy, 4 The University of Alabama at Birmingham, Division of Clinical Pharmacology, Birmingham, Alabama, United States of America
Stably suppressed viremia during ART is essential for establishing reliable simian models for HIV/AIDS. We tested the efficacy of a multidrug ART (highly intensified ART) in a wide range of viremic conditions (103-107 viral RNA copies/mL) in SIVmac251-infected rhesus macaques, and its impact on the viral reservoir. Eleven macaques in the pre-AIDS stage of the disease were treated with a multidrug combination (highly intensified ART) consisting of two nucleosidic/nucleotidic reverse transcriptase inhibitors (emtricitabine and tenofovir), an integrase inhibitor (raltegravir), a protease inhibitor (ritonavir-boosted darunavir) and the CCR5 blocker maraviroc. All animals stably displayed viral loads below the limit of detection of the assay (i.e. <40 RNA copies/mL) after starting highly intensified ART. By increasing the sensitivity of the assay to 3 RNA copies/mL, viral load was still below the limit of detection in all subjects tested. Importantly, viral DNA resulted below the assay detection limit (<2 copies of DNA/5*105 cells) in PBMCs and rectal biopsies of all animals at the end of the follow-up, and in lymph node biopsies from the majority of the study subjects. Moreover, highly intensified ART decreased central/transitional memory, effector memory and activated (HLA-DR+) effector memory CD4+ T-cells in vivo, in line with the role of these subsets as the main cell subpopulations harbouring the virus. Finally, treatment with highly intensified ART at viral load rebound following suspension of a previous anti-reservoir therapy eventually improved the spontaneous containment of viral load following suspension of the second therapeutic cycle, thus leading to a persistent suppression of viremia in the absence of ART. In conclusion, we show, for the first time, complete suppression of viral load by highly intensified ART and a likely associated restriction of the viral reservoir in the macaque AIDS model, making it a useful platform for testing potential cures for AIDS.
Novel research aimed at finding a cure for AIDS requires animal models responding to human antiretroviral drugs. However, there have been few antiretrovirals cross-active against the simian viruses. In this study, we expanded the arsenal of drugs active against the simian retrovirus SIVmac251 and showed that this virus is inhibited by the protease inhibitor, darunavir, and the CCR5 blocker, maraviroc. Administration of these two drugs in combination with the reverse transcriptase inhibitors, tenofovir and emtricitabine, and the integrase inhibitor, raltegravir, resulted in prolonged plasma viral loads below assay detection limits, and, surprisingly, restricted the viral reservoir, a marker of which is viral DNA. We then decided to employ this multidrug regimen (termed "highly intensified ART") in order to increase the potency of a previous strategy based on the gold drug auranofin, which recently proved able to restrict the viral reservoir in vivo. A short course of highly intensified ART following the previous treatment resulted, upon therapy suspension, in a remarkably spontaneous control of the infection, that may pave the way to a persistent suppression of viremia in the absence of ART. These results corroborate the robustness of the macaque AIDS model as a vanguard for potentially future treatments for HIV in humans.
The study of persistence of viral sanctuaries during antiretroviral therapy (ART) and the possibility for their therapeutic targeting is crucial for eradication of HIV-1. Animal models for lentiviral persistence during therapy are therefore needed. The creation of such animal models requires knowledge of the response of animal lentiviruses to antiretroviral drugs adopted in treatment of humans with HIV-1. Finding cross-active drugs has been a difficult task because non-HIV-1 lentiviruses often mimic drug resistance mutations found in HIV-1. This mimicry has been shown for the viral protease  and for the portion of reverse transcriptase (RT) that binds the non-nucleosidic reverse transcriptase inhibitors (NNRTIs) .
One of the current models is based on macaques infected with a molecularly engineered simian immunodeficiency virus (SIVmac239) expressing HIV-1 RT, in order to overcome drug resistance mimicry of the primate lentiviruses to NNRTIs . Another model (SIV-based) has been developed for neurotropic infection, a condition often occurring in late-stage AIDS . In this case, in order to by-pass the different response to antiretrovirals, the authors used a drug combination which is not adopted in humans. However, in both of these animal models, low-level viremia persisted and viral RNA was consistently detectable in anatomical sanctuaries , .
A model recently developed by our group is based on a polyclonal virus, such as SIVmac251, mimicking, at least in part, the genetic diversity of HIV-1 naturally inoculated in human subjects . It was recently shown that SIVmac251 responds to combined ART consisting of two nucleosidic/nucleotidic reverse transcriptase inhibitors (NRTIs), i.e. tenofovir and emtricitabine, and the integrase inhibitor raltegravir , . In this treatment model, the virus persists during ART, and viral load rebounds following treatment suspension in a time frame remarkably similar to that observed in humans after treatment interruption .
Recent research has added more credit to the macaque AIDS model, showing that, similarly to humans , , rhesus macaques (Macaca mulatta) harbour a central memory CD4+ T-cell reservoir, which plays a pivotal role in AIDS pathogenesis , . Important insight has been derived from comparisons between rhesus macaques and sooty mangabays (Cercocebus atys) which, unlike M. mulatta, do not progress to AIDS . M. mulatta, but not C. atys, shows up-regulation of the lentiviral co-receptor CCR5 in activated central memory T-cells, thus rendering this T-cell pool highly permissive to infection . Conversely, the reduction of the long-lived memory T-cells (CD95+CD28+), including central memory T-cells, by the gold-based compound auranofin in intensified ART (iART)-treated rhesus macaques resulted in decreased levels of viral DNA and delayed progression of the infection upon therapy suspension . Therefore, a model mimicking the effects of suppressive ART in humans is of fundamental importance also for the study of the dynamics of this viral reservoir.
One major limitation of current models for HIV persistence during therapy is their large discrepancy from conditions observed in humans. So far, due to financial and temporal constraints, animals have been chosen from homogeneous cohorts in terms of timing, type and route of the inocula, and have been treated in the early phases of chronic infection - or during acute infection . Instead, at therapy initiation, HIV-infected humans are usually characterized by different timings and routes of disease acquisition and different levels of progression of the infection. In order to obtain a robust animal model for HIV persistence during therapy, the drug regimens should display similar efficacies as compared to those employed for human treatment, and reproducible control of heterogeneous viral loads in wide cohorts of subjects with different characteristics and previous treatment histories.
Here, we report a highly intensified ART (H-iART) regimen for the simian model, reproducibly capable of decreasing viral load to levels below assay detection limits in SIVmac251-infected macaques starting from a wide range of baseline viremic conditions, and overcoming previous treatment failures. We also report an unexpectedly impressive restriction of viral DNA in peripheral blood mononuclear cells, obtained by means of a pharmacological strategy entirely based on antiretroviral drugs.
SIVmac251 is susceptible to darunavir (DRV) and maraviroc (MRV)
The first part of this study was aimed at obtaining long-term viral suppression in a group of macaques (n = 4) in order to develop a suitable platform for testing experimental eradication strategies. We first analyzed the susceptibility of SIVmac251 to the protease inhibitor darunavir (DRV) and the CCR5 blocker maraviroc (MRV) in order to expand the arsenal of antiretroviral options available for the macaque AIDS model. DRV was chosen because of its well documented ability to inhibit several drug-resistant HIV-1 isolates as well as HIV-2, a virus closely related to SIVmac251 , , . Moreover, the choice of this drug was supported by preliminary bioinformatic and molecular modeling analyses showing the potential interactions of DRV with the SIVmac251 protease [Text S2 and Fig. S2]. MRV, a CCR5 antagonist, was chosen on the basis of the important role of CCR5 as a SIVmac251 co-receptor  and due to the antilentiviral activity previously demonstrated by one experimental CCR5 blocker in macaques . Moreover, retrospective analysis of one previous in-vivo experiment supported the anti-SIVmac251 effect of this drug [Text S3 and Fig. S3]. Results from tissue culture experiments showed that both DRV and MRV inhibited SIVmac251 replication in the nanomolar range, with EC50 values well below the trough concentrations detected in macaques treated with these drugs and described below in the text. (Fig. 1).
DRV improves the virological response of SIVmac251-infected macaques to ART
A group of macaques [n = 4] displaying signs of immune deterioration (eighteen months post-inoculation) was treated with a regimen of tenofovir, emtricitabine and raltegravir (Fig. 2). These macaques were derived from viral titration experiments and selected among those maintaining stable plasma viral loads (Fig. 2A). The selected animals displayed viral load set points between 103 and 105 viral RNA copies/mL. As our study was aimed at obtaining a model mimicking the conditions found in HIV-1-infected individuals under ART, such baseline values were chosen in order to reflect the average viral loads at which treatment is started in humans. The CD4 counts displayed by the macaques enrolled in this "pilot" study were significantly lower than values observed in uninfected controls (Fig. S4), suggesting that they were unlikely to be long-term non-progressors or elite controllers.
The three-drug regimen proved insufficient to maintain control of viral load in three of the four animals treated (Fig. 2A). DRV (375 mg bid), boosted with ritonavir (50 mg bid), henceforth referred to as DRV/r, was added to the treatment in an attempt to obtain a more stable control of viral load. DRV/r significantly improved control of viral load, inasmuch as viral RNA in plasma was maintained at a significantly lower level as compared to the pre-therapy values (Fig. 2A). No similarly decreasing trend of viral load was observed in an untreated control group of macaques [n = 2] showing non-significant differences in baseline viral loads as compared to the treatment group (two tailed t-test: P = 0.803; Fig. 2A). We conclude that the iART regimen adopted improves control of viral load in SIVmac251-infected macaques.
A H-iART regimen induces a prolonged control of residual viremia
To increase the chances for long-term control of SIVmac251 replication, we explored the in-vivo efficacy of the CCR5 inhibitor MRV. This drug (100 mg BID) was eventually added to the drug cocktail in the aforementioned group of macaques (Fig. 2). After MRV was started, all macaques stably maintained viral loads below the limit of detection of the assay (i.e. 40 copies RNA/mL; Fig. 2A). There were also significant increases in the absolute numbers of CD4+ T-lymphocytes (Fig. 2B). Henceforth, this multidrug combination will be referred to as highly intensified ART (H-iART).
MRV exerts antiretroviral effects in vivo
In order to further support the contribution of MRV to the antiretroviral effects observed, we treated two macaques with MRV (ritonavir boosted, MRV/r) in monotherapy (Fig. 3). In line with its CCR5-blocking ability, MRV decreased the viral loads in two drug-naïve macaques with dynamics similar to those previously shown by an investigational CCR5 blocker . When the other H-iART drugs were added to MRV, a quick abatement of viral load to levels below the assay detection limit could be demonstrated (Fig. 3).
H-iART suppresses viremia in a broad range of viremic conditions
Prior to treatment with antiretrovirals, approximately one third of the experimental infections of macaques with SIVmac251 results in viral set points comparable to those displayed by the macaques described in the previous paragraphs (Fig. S5). To check whether H-iART might reproducibly control viral replication in SIVmac251 infected macaques characterized by higher viral loads, five animals with viral set points ranging from 103 to 107 viral RNA copies/mL of plasma were treated with H-iART, and the viral decay dynamics were compared with those of macaques treated with iART. Results clearly showed that H-iART induced a significantly more rapid decay in viral load than did iART (Fig. 4A). In line with the efficacy of H-iART, CD4+ T-cells increased in all study macaques (Fig. S6). We conclude that MRV-containing H-iART is superior to iART in abating viremia load in a group of macaques with a wide array of baseline viral loads.
The extent of suppression of viral replication is dependent on baseline viral loads and drug dosage in H-iART-treated macaques
We then analyzed the viral load decay dynamics in macaques treated with H-iART ab-initio. SIVmac251-infected macaques responded to administration of H-iART with a two phase exponential decay, as described in humans treated with suppressive ART  (Fig. 4). Similarly to the average treatment outcomes in humans , the level of viral load suppression depended on the baseline viral loads, with macaques starting from higher viral loads showing viral blips or residual, though markedly decreased (>3 Logs), viral replication (Fig. 4D-F).
We increased the DRV and MRV dosage in macaques 4887, BD64 and BD69, i.e. those starting from higher baseline viral loads (>105) and showing incomplete control of viral replication or major blips. Results showed that the improved drug regimen led to viral loads consistently below the assay detection limit in animals BD64 and BD69 (Fig. 4D,E). The increased drug dosage was also able to decrease the amplitude of the remaining sporadic blips (Fig. 4E). The resulting blips were lower than 103 copies of viral RNA/mL, thus mimicking those observed in humans under ART . Nevertheless, one animal (4887) experienced a further viremic episode. Analysis of the cerebrospinal fluid (CSF) of this animal showed a viral load approximately one order of magnitude higher than that in plasma, while CSF samples were below the assay detection limit (i.e. 40 copies/mL) in the macaques showing stable control of viral replication (data not shown). This evidence suggested that the central nervous system (CNS) was a likely major source for the rebounding virus in macaque 4887. According to previously published studies: 1) virus levels in the CSF during the advanced stages of the disease are mostly due to CNS sources , and 2) the protease inhibitors (i.e., the only drug class in our cocktail acting at a post-translational level, and hence on chronically infected cells) are extruded from the CNS by P-glycoprotein (P-gp) molecules in the blood-brain barrier . We thus intensified the P-gp blockade by increasing, from 50 to 100 mg bid, the dosage of ritonavir, which is a well-known P-gp inhibitor . The viral load decreased in both plasma and CSF, with a more rapid decay kinetic in plasma, in which viral RNA eventually fell to levels below the assay detection limit (Fig. 4F). This result is in good agreement with the hypothesis of the CNS as a major source for the rebounding virus.
We conclude that macaques starting from high viral loads respond to H-iART similarly to HIV-infected humans and that viral loads can be abated to levels below the assay detection limit by adjusting the drug dosages and boosting procedures.
A highly sensitive viral load detection assay shows profound suppression of viral replication by H-iART
To check the presence of low-level viremia in SIVmac251-infected macaques under H-iART, we lowered the detection limit to 3 copies of viral RNA/mL and re-measured viral loads in some selected pooled serum samples. We found no evidence for low-level viral replication in plasma of all of the macaques tested (Table 1). Of note, viral RNA was below the assay detection limit in the plasma samples taken from macaque 4887 before its last viremic episode, supporting the hypothesis that H-iART was able to completely control viral replication in the periphery, despite the presence of a major CNS reservoir (Fig. 4F). Analyses conducted on lymph node biopsies (inguinal) showed that four out of six macaques analyzed had levels of cell-associated RNA below the limit of detection of the assay (i.e. 2 copies/5*105 cells/mL) (Table 2). The presence of cell-associated RNA in lymph nodes was independent of baseline viremia at treatment initiation (Logit analysis P = 0.801), thus supporting the idea that the suppressive efficacy of H-iART is not confined only to those macaques starting from moderate viral loads. In addition, cell associated RNA measured in samples taken from rectal biopsies was below the assay detection limit in all animals analyzed, supporting the idea of full suppression of peripheral viral replication (Table 2). This was rather surprising, because other antiretroviral regimens adopted in macaques proved unable to completely control viral RNA in anatomical sanctuaries , .
H-iART impacts on viral DNA in PBMCs, lymph nodes and rectum
In the pilot study presented above, we unexpectedly found that H-iART profoundly impacted on viral DNA. First, there was a late viral DNA decay to levels below the assay detection limit which was associated with the addition of MRV to the drug cocktail (Fig. 5A). In addition, the CD4/CD8 ratio, the decrease of which is a marker of the viral reservoir and/or ongoing viral replication , , significantly increased during treatment (Fig. 5B). Of note, viral DNA in PBMCs also fell below the assay detection limit in all macaques included in the group treated with H-iART ab-initio (median treatment duration = 125 days, range from 45 to 174 days), i.e. no viral DNA copies were detectable in six out of six repeats with a threshold sensitivity of 2 copies/5*105 cells.
Viral DNA decay dynamics during H-iART
The dynamics of the viral DNA decay during H-iART were studied in those animals to which all H-iART drugs were administered simultaneously and for which viral DNA measurements were available.
The levels of viral DNA in PBMCs during time were consistent with a three-phase decay, with the first two phases paralleling the two-phase decay of viremia, and a third, slower phase occurring after viremia had fallen to levels below the assay detection limit (Fig. 5C). This last phase of the viral decay has been ascribed to the latently infected T-cell numbers . This result was noteworthy, because no such decreasing trends in viral DNA had been observed in animals treated with iART (i.e. without MRV) .
H-iART impacts on the memory T-cell pool
In line with the reportedly stimulating effect of the major CCR5 ligand RANTES on T-cell proliferation  some studies suggested that MRV, by acting as an antagonist of this cytokine, might alter the T-cell dynamics in vivo . To study these phenomena, the CD4+ T-cell subpopulations were analyzed by six-color flow-cytometry at different time points following addition of MRV to the therapeutic regimen (Fig. 6). To avoid biasing the result with the possible effects of a detectable viral load on the T-cell subpopulations, these tests were conducted on PBMCs from macaques P157, P185 and P188 which already displayed a viral load below the assay detection limit when MRV was added (Fig. 2). Results showed that H-iART decreased the memory CD4+ T-cell numbers over time (Fig. 6A,B), while it carried out no significant effect on the naïve T-cell subpopulation (Fig. 6C). This result is in accordance with the in-vitro inhibitory effect of MRV on the proliferation of sorted memory T-cell subpopulations (Fig. S7). MRV significantly decreased the numbers of activated (HLA-DR+) CD4+ TEM cells (Fig. 6D). This effect is in line with decreased levels of immune activation already observed in humans treated with this drug , . In conclusion, MRV decreased the number of memory T-cells as well as TEM cell-activation. Since these two parameters are linked to the magnitude of the viral reservoir and ongoing viral replication , , this effect is in good agreement with the aforementioned three-phase decay of viral DNA induced by MRV (Fig. 5C).
MRV impacts on the viral set point following therapy suspension
The results so far obtained were in line with a recently issued report which suggested that MRV decreased the magnitude of the viral reservoir in HIV-1-infected individuals . This study, which was unable to provide conclusive evidence, did not show an impact of MRV on the viral set point following therapy suspension, a parameter stringently associated with the extent of the viral reservoir , , . To test this hypothesis, we analyzed the difference in the pre and post-therapy viral set points in those macaques from our cohort that had received MRV and that had undergone therapy suspension (for treatment details see Figs. 2, 4 and Text S3). Results show that treatment with MRV is associated with a reduction of the viral set point post-therapy (Fig. 7A), and that the extent in the viral set point decrease depends on the total exposure to the drug (Fig. 7B). These results are suggestive of an independent effect of MRV on the viral set point following therapy suspension and add credit to the hypothesis that MRV may contribute to an anti-reservoir effect of H-iART.
H-iART improves the spontaneous control of viral load following a previous anti-reservoir strategy
Finally, given the aforementioned effects of H-iART, we tested whether this therapeutic regimen might be adopted to improve the effect of a previous anti-reservoir strategy based on the anti-memory drug auranofin in combination with antiretrovirals . Upon interruption of this anti-reservoir treatment, SIVmac251-infected macaques experience an acute infection-like condition, i.e. an initial viral load peak followed by rapid containment of viral load . The peak, which is rapidly reached upon virus re-appearance in plasma, is associated with the reconstitution of the viral reservoir, as shown by the previously published independent association between the area under the curve (AUC) describing the initial peak of viral load and the eventual viral load set point ( see also Fig. 8A). From this association, it follows that decreasing the AUC at peak artificially through a cycle of H-iART should limit the reconstitution of the viral reservoir and may result in spontaneous control of viral load following H-iART suspension.
The experiment was attempted in two macaques. A first macaque (P252) was treated with a one-month cycle of H-iART at viral load rebound, after the suspension of the aforementioned auranofin/antiretroviral regimen. Another macaque (P177) was treated with auranofin in addition to H-iART as a follow-up of the treatment presented in the pilot study. Eventually, following therapy suspension, P177 was subjected to a short H-iART cycle at viral rebound, similar to that administered to P252. In both cases, the short H-iART cycle promptly abated viral load to levels below the assay detection limit, thus efficiently decreasing the initial AUC (Fig. 8 A-C).
The macaques showed exceptionally low viral set points after the short cycle of H-iART was suspended, in line with the expected values calculated on the basis of our AUC/viral set point correlation curve (Fig. 8A). Both macaques periodically displayed viral load peaks that subsequently decreased to low-level viremia (<500 copies of viral RNA/mL) or to levels below the assay detection limits. The CD4 slope was non-significant during the follow-up period (P = 0.7079 for P252 and P = 0.2319 for P177; Fig. 8D,E), in line with the previous observation that the CD4 slope following therapy suspension identifies the impact of a treatment on the viral reservoir . Conversely, CD4 counts had shown significantly decreasing trends in both macaques before all treatments were started (P<0.0001 for P252 and P = 0.0039 for P177; Fig. 8D,E), thus supporting the concept that the therapies adopted significantly impacted on the natural course of the disease.
Consistently with its exceptional reduction of the AUC at peak, macaque P177 showed a remarkable degree of spontaneous control of viral load during six months of follow-up, which was not yet considerable as, but seemingly close to a drug-free remission of the disease (Fig. 8C). In this macaque, viral load was maintained at levels below the assay detection limit during the periods between peaks (detection limit: 40 RNA copies/mL) and, when the RNA detection limit was further lowered to 3 copies/mL, no evidence of residual viremia was found (see Table 1). This control of viral replication could hardly be ascribed to cell-mediated responses, in that a moderate increase in the number of IFN-γ positive spots could be detected only at viral rebound but not during the viral set point (Fig. S8), thus suggesting that H-iART induced a true containment of the viral reservoir reconstitution, similarly to other experimental strategies restricting the formation of the viral reservoir during acute infection , , . We conclude that a short course of H-iART, in line with the highly suppressive effect of this therapeutic regimen on SIVmac251, may prevent the viral reservoir reconstitution following suspension of a previous anti-reservoir therapy and result in a drug-free spontaneous control of viral load.
Some investigators recently questioned the robustness of primate models, citing the difficulty of obtaining, with the cross-active drug options available, full viral suppression in sanctuaries and viral loads below the assay detection limits for prolonged periods , . The results reported in the present article do not support this argument.
1) Since a good animal model should mirror full viral suppression in humans, we checked viral loads in plasma for prolonged periods and analyzed the presence of viral nucleic acids in anatomical sanctuaries. The level of abatement of viral nucleic acids that we found in the present study in peripheral blood and anatomical sanctuaries of the majority of the macaques tested provide the maximum degree of viral suppression so far observed in antiretroviral treated primates. The level of reproducibility of these results is shown by the fact that they were obtained in a heterogeneous group of macaques, likely mirroring a wide number of possible disease conditions in humans. This is the first report, to our knowledge, of a therapy capable of stably controlling viral replication to levels below the assay detection limits also in macaques in the advanced stage of the disease, since the studies so far published have been able to report control of SIV replication only during acute infection  or in the early chronic phase of the disease -. Apart from mimicking the clinical conditions of a significant portion of HIV-infected individuals who are diagnosed in the chronic or pre-AIDS stages of the disease, this 'late' treatment allows excluding those macaques able to spontaneously control the infection, a phenomenon which usually occurs soon after the acute infection phase . For the macaques enrolled in this study, the average plasma viral load at the time of therapy initiation was of 4.8±1.1 Log10 RNA copies/mL (mean ± SD). This value is lower than those reported in some articles during chronic SIVmac infection of macaques , , but similar to those published in other articles , . As in this study we have not included macaques with viral loads during chronic infection higher than 6.8 Log10 RNA copies/mL or with the rapid progressor phenotype, the effect of our H-iART regimen on this more aggressive course of SIV infections remain to be ascertained.
Of note, persistence of the virus at low level in the lymph nodes of a minority of H-iART treated macaques provides another similarity of our macaque model with clinical conditions observed in humans infected with HIV-1, as this anatomical sanctuary has recently been shown to be a major site for ongoing viral replication in humans . Studies of drug penetration in this anatomical compartment will be necessary to overcome this limitation in both macaques and humans.
2) As in any well respected science, the results are in good agreement with mathematical models (Fig. 9), and are mathematically predictable (as an example, see Fig. 8A). In this regard, important insight into the necessity for a multidrug regimen to control viral loads in macaques can be derived from a mathematical model developed by Rong and Perelson  and based on experimental observations . This model suggests that a superior drug efficacy is required in simian AIDS to control viral replication (Fig. 9) because of the viral burst size, (i.e. the average number of virions produced by a single productively infected cell in a day). The viral burst size was shown to be higher in SIV infection as compared to HIV-1 infection , where a lower drug efficacy is expected to be sufficient to maintain viral control (Fig. 9A-C). Also a drug acting on the proliferation rate of activated T-cells, such as MRV (which antagonizes the proliferative effect of RANTES, see ref 25 and Fig. S7), appears to be important for containment of the viral blips (Fig. 9D). These simulations also show that the decreased proliferation rates may impact on the viral reservoir size (half-life: Å200 days, see Text S4, S3, S2, S1 and Fig. 9D), which shows a half-life of the same order of magnitude as that calculated by analyzing the dynamics of the viral DNA decay during H-iART (Fig. 5).
3) According to the idea that a good animal model should represent a vanguard for future treatments to be tested in humans, our quest for increased drug efficacy in the macaque AIDS model allowed identifying unexpected benefits of H-iART on the immune system. Apart from the possible impact of H-iART on the viral reservoir (a concept supported by recent data in humans ), reduction by MRV of the memory T-cell subpopulation may restrict one major source for viral spread and ongoing viral replication. A decrease in the memory T-cell size is a logical expectation of the anti-proliferative effect exerted by MRV through CCR5 inhibition (Fig. S7), as antigen-driven proliferation contributes to maintenance of the size of this T-cell subpopulation . It is well known that memory T-cells are a preferential target of HIV-1 replication , and that their decrease may affect the overall viral dynamics in vivo. In this regard, the MRV-induced decrease in the memory T-cell size is not only unlikely to be dangerous but, rather, likely to be beneficial. This hypothesis is supported by results showing that the pool of TCM cells is a correlate of anergy towards the viral antigens in Macaca mulatta but not in Cercocebus atys, which is naturally resistant to CD4+ T-cell loss and full-blown AIDS . In addition, the results obtained with the present macaque model suggest that a short cycle of H-iART could be used for improving the efficacy of our previous anti-reservoir treatment based on auranofin and strengthen the idea that an arrest in disease progression may be obtained during the chronic phase of the disease. Although the data on the combined effect of the two subsequent treatment cycles are derived from a limited number of macaques, the result obtained is corroborated by the fact that no similar trend was observed in the same animals prior to starting therapy ,  or in historical controls that had not received H-iART at rebound . Of note, although certain major histocompatibility complex (MHC) class I alleles, including Mamu-A*01 and Mamu-B17* are associated with slow disease progression in SIV infected macaques , , independently, the presence of these alleles is not predictive for disease outcome , and none of our macaques presented the protective alleles in association (Table S1). Instead, P177, which, following our therapies, remarkably controlled viral load, presented the HLA Mamu-B*01 allele, that is associated with aggressive simian lentivirus infection . In line with this genotype, P177 showed a significant immune deterioration before our treatments were initiated (Fig. 8C).
Finally, recent analyses [reviewed in 49] re-evaluated the necessity of wide numbers of subjects as a support for breakthrough findings, such as, in this case, the obtainment of a condition close to a persistent suppression of viremia in the absence of ART.
If the results of the present study should prove reproducible in humans, H-iART could represent a useful tool for improving the viro-immunological conditions of HIV-infected individuals and a useful addition to experimental anti-reservoir strategies.