The current report is a follow-up study of 5 animals that received prolonged tenofovir monotherapy. Despite the emergence of K65R viral mutants with 5-fold reduced in vitro susceptibility, all five animals had eventually reached undetectable viremia [10, 11, 13, 15]. Although we and others have previously described low or undetectable viremia after the emergence of K65R viral mutants in tenofovir-treated animals [12, 26], the uniqueness of the present cohort of 5 animals resides in the unprecedented extensive period of tenofovir therapy (8 to 14 years) and survival. These animals were infected with virulent virus, and animals infected with these same viruses (wild-type or K65R) but not receiving any antiviral drug therapy were never able to spontaneously suppress viremia to low or undetectable levels and generally developed symptoms of AIDS within 2–24 months [10, 11, 13, 15].
It is remarkable that during tenofovir treatment lasting 8 to 14 years, 4 of the 5 tenofovir-treated animals continued to have undetectable viremia with occasional viral blips. Similarly to observations in humans , the reasons for viral blips were not clear but may reflect a transient increase in viral production (e.g., due to immune activation) and/or minor fluctuations in antiviral effector mechanisms. Overall, the sustained suppression of virus replication in the 4 animals testifies to the strength of antiviral immune responses that were documented previously via CD8+ depletion experiments [10, 11]. In other SIV studies, including many vaccine studies, animals that initially controlled replication of virulent virus often showed an increase in viremia after prolonged follow-up [28–31]. As explained in more detail elsewhere, it is plausible that, whereas as no single factor may be sufficient, a synergistic combination of (i) effective antiviral immune responses, preserved by early initiation of treatment, and sustained by ongoing low-level replication of K65R virus, (ii) a minor effect of K65R and compensatory mutations on viral replication fitness or diversity, and (iii) some residual drug efficacy against K65R mutants was responsible for a steady-state situation without viral rebound in these tenofovir-treated animals [9–11]. The demonstration that a combination of tenofovir and antiviral immune responses can suppress K65R SIV replication in macaques for many years is also consistent with the lack of viral rebound in treatment-experienced patients who develop K65R viral mutants during tenofovir treatment, and the observations that viremia in persons with detectable K65R mutants can be suppressed by tenofovir-containing regimens [32–34].
An exception was animal 33091, which despite having nearly 3 years of undetectable plasma viremia, slowly lost control of virus replication and demonstrated a slow disease progression while still on tenofovir treatment. It is unknown whether the transient CD8+ depletion that was performed previously in this animal at 2 weeks of SIV infection, when antiviral immune responses were in their early stages and probably most vulnerable, may have had a negative impact on the strength and breadth of the antiviral CD4+ or CD8+ cell repertoire, predisposing the eventual outgrowth of viral immune escape variants. In contrast, three of the four animals that demonstrated sustained suppression of viremia during tenofovir treatment had been depleted of CD8+ cells during a later stage of infection (≥ 39 weeks), when viremia was already suppressed by strong antiviral immune responses, and when the immune perturbation caused by CD8+ depletion may have had only a transient impact on antiviral immune responses.
For 3 of the 4 animals with sustained suppression of viremia during tenofovir treatment, the treatment was eventually stopped. Despite some transient periods of detectable low viremia, no rapid viral rebound was observed during the 10 months of observation. These animals that controlled viremia after tenofovir withdrawal also had low levels of viral DNA in blood and lymphoid tissues, stable CD4+ cell counts and a broad variety of cellular and humoral antiviral immune responses. These observations recapitulate the features of long-term non-progressor (LTNP’s), who naturally suppress virus replication for prolonged periods of time without antiretroviral drug intervention (reviewed in ).
The absence of a rapid rebound in all three animals can theoretically be due to low replication fitness of the virus, to strong antiviral immune responses, or to a combination of both. Previous experiments demonstrated that SIV and RT-SHIV isolates having the K65R mutation in combination with compensatory mutations have high replication fitness and virulence, and generally do not revert back to wild-type sequence following tenofovir withdrawal [10, 11, 13, 16, 17]. Also in the current study, withdrawal of tenofovir treatment did not lead to a detectable reversion from K65R to wild-type virus. Accordingly, the absence of a viral rebound likely represents an effective immune-mediated control of virus replication, rather than a major replication-attenuated phenotype. The development and maintenance of the cellular and humoral immune responses we observed must have been promoted by ongoing low-level replication of K65R viral mutants during the prolonged period of tenofovir treatment which created a balanced antigen expression/immune response steady-state.
Previously, depletion experiments revealed a major role for CD8+ cell-mediated immune responses in suppressing virus replication during tenofovir treatment in all 5 animals [10, 11]. Because the cM-T807 antibody depletes both the CD8 + CD3+ T cells and the CD8 + CD3- NK cells, the relative contribution of these two cell subsets could not be established in those studies. As NK cells are also effector cells of ADCVI , they may play some role considering that the sera of all animals had high ADCVI activity, even against SIV strains that were poorly neutralized in the absence of effector cells. Even though animal 33091’s serum samples exhibited similarly high ADCVI activity in vitro as the other animals, it is unclear whether a reduced function of ADCVI effector cells in vivo contributed to its poorer control of virus replication. Additional CD8+ cell depletion experiments during the tenofovir-free period were not feasible because all five animals had developed antibodies against the cM-T808 antibody and other available CD8+ depleting antibodies (data not shown).
All SIV-infected animals had SIV-specific CD4+ and CD8+ cell-mediated responses. Although there was much variability in the functionality and tissue distribution among the different animals, the controlling animals had a trend towards a more multifunctional response that was located particularly in the intestinal tissues and lymph nodes, suggestive of an immunological control of virus replication. In contrast, animal 33091’s response was more mono-functional, absent in the intestinal tract, and residing predominantly in the peripheral blood, suggestive of an antigen-exposure driven immune response. These patterns are consistent with observations in other studies .
A closer look at the historical data of this cohort and other studies suggests several trends in the evolution of the immunological control of virus replication. Firstly, as shown in Figure 1, the viral rebound or blips shortly after tenofovir interruption in some animals in the current study were generally smaller than those observed with short-term treatment interruption (7 to 9 weeks) on these same animals earlier during the course of infection. In other animal studies, treatment interruptions also resulted in more rapid or higher viral rebound [10, 38–54]. These observations suggest that the antiviral immune responses in these K65R virus-infected animals strengthened – rather than weakened- during the consecutive years of tenofovir treatment. Although this observation of strengthening immune responses provides hope, the time frame during which this was observed in the current study also highlights a research dilemma: future studies aimed at a functional cure, whether in animal models or human cohorts, may require a duration of more than 5 to 10 years, and therefore, a long-term investment in funding by research agencies.
Secondly, the optimism of the current results needs to be balanced with caution: although the frequency of low plasma RNA levels during the 10 months off tenofovir was similar to that during the prior period of tenofovir treatment, the magnitude of viremia was at times higher than what was observed during the preceding 5 years on tenofovir therapy. This is consistent with our previous observations that despite strong antiviral immune responses, there is still some benefit of continued tenofovir treatment, either by a residual direct antiviral effect and/or by immunomodulatory effects (reviewed in ). Although in the current study each transient increase of viral RNA was followed by a period of regained control, it is possible that during a much longer drug-free observation period such blips may become higher or more frequent and may require reinitiation of antiretroviral therapy to prevent disease progression.
It has to be re-emphasized that this cohort of animals had unique circumstances, namely prolonged tenofovir monotherapy, started relatively early in infection, in the presence of K65R viral mutants resulting in the generation and maintenance of effective antiviral immune responses and creation of a relative steady-state balance. Therefore, these results may not apply to HIV-infected patients who are started on ART late during infection, effectively control viremia and do not develop drug-resistant viral mutants. Such patients will typically show a rapid viral rebound upon drug withdrawal, possibly indicating insufficient antigenic exposure to generate, restore or maintain antiviral immune responses [55–58]. Consistent with our observation in monkeys, a minority of patients (5 out of 32) started on ART during primary HIV infection had good immunologic control of viremia after ART interruption, and these controllers had a trend toward earlier initiation and a longer duration of ART in comparison to noncontrollers or transient controllers .
The cohort of animals described here shares similarities with HIV-infected humans who, despite the presence of multi-drug resistant viral mutants, have low viremia associated with strong antiviral immune responses [60, 61]. A difference, however, is that in the HIV-infected patients, interruption of nucleoside reverse transcriptase RT inhibitors led to an immediate and persistent increase in viremia and a reduction in CD4+ cell counts [62–64]. Potential reasons for this difference include (i) later initiation of ART, (ii) a much shorter duration of treatment, (iii) persistent detectable viremia (> 400–1000 copies/ml) while on ART (indicating less effective antiviral immune responses even prior to drug withdrawal), and (iv) regimens without tenofovir, which in addition to its antiviral activity has shown unique immunomodulatory properties in animal models [65, 66].
For all these reasons, our results do not support withdrawing ART in HIV-infected individuals.