Several drugs are available to control HIV, but they do not completely eliminate the virus from the body. Instead, these treatments stop the virus from multiplying, but unless a person is treated very soon after infection, inactive HIV can hide inside cells, and the infection is not completely cleared. Once treatment stops, the inactive virus starts to multiply, reaching pre-treatment levels within weeks. This means that infected individuals must continue treatment for life, or the virus will return and may cause disease. To prevent this, scientists are trying to find a way to eliminate HIV from the body, permanently curing the disease.

Testing HIV drugs is difficult because there is no simple way to determine if all the inactive virus has been removed. One way to test whether a person is cured is to stop treatment, and see if the virus comes back. An alternative method is to measure the amount of inactive HIV present in blood cells. However, Pinkevych et al. have now shown that levels of inactive HIV in the blood may not be a good predictor of whether HIV levels will rebound after treatment.

In the experiments, Pinkevych et al. infected macaques with a monkey version of HIV called simian immunodeficiency virus, or SIV for short. The virus was genetically modified to have a ‘genetic barcode’ that allowed researchers to track thousands of individual strains of virus simultaneously . The macaques were then treated with a combination of drugs starting either 4 or 27 days after infection, and then the treatment was stopped to check how rapidly strains of the virus would re-emerge.

Although the virus rebounded in both groups, during treatment the inactive virus was often undetectable in the blood from the group treated four days after infection , suggesting that the virus may be hiding elsewhere. In the group treated 27 days after infection, more blood cells had inactive virus and reactivation was more frequent. However, the amount of inactive virus in the blood did not directly predict the frequency of reactivation: a 100-fold increase in viral levels only led to reactivation being twice as frequent. This suggests that the amount of inactive virus in the blood is not a good read-out for whether the virus will come back.

These experiments suggest that inactive virus hides in cells quickly, since treatment just four days after infection failed to eliminate the virus completely, and that it must hide in cells other than blood cells. Evidence from humans suggests something similar may occur in HIV infection. Further studies may help identify which cells harbor the inactive virus and contribute to infections re-emerging after treatment stops.

Current treatment for HIV infection requires an individual to undergo lifelong daily cART administration, as interruptions in dosing usually result in the rapid recrudescence of virus. This viral recrudescence is thought to result from periodic activation of viral production from cells, including latently infected cells (Pinkevych et al., 2015; Fennessey et al., 2017), allowing the virus to spread to uninfected target cells no longer protected from infection by cART. A variety of approaches are currently being studied to reduce the RCR or the frequency of activation leading to virus production, in order to produce anti-retroviral free remission in treated patients (Deeks et al., 2016; Davenport et al., 2019). A variety of assays have been developed to measure HIV DNA (total or integrated virus, intactness of viral genomes, or replication competence in outgrowth assays), levels of HIV-RNA production (spliced or unspliced), or the ability to drive viral production through in vitro or in vivo stimulation of cells (Eriksson et al., 2013; Procopio et al., 2015; Finzi et al., 1999; Lorenzi et al., 2016; Metcalf Pate et al., 2015; Barton and Palmer, 2016; Bruner et al., 2015). These measures usually demonstrate a very wide range of values between individuals, varying between 100-fold and 1,000-fold for different individuals, and different measures are often poorly correlated (Eriksson et al., 2013). Moreover, although these are often described as measures of the ‘viral reservoir’, their contribution to the actual post-ART viral recrudescence is unclear.

While complete eradication of the HIV RCR would prevent viral recrudescence, greatly decreasing the frequency of cells producing infectious virus may provide prolonged remission and the possibility of functional ‘cure’ (Davenport et al., 2019; Cromer et al., 2017; Hill et al., 2016). In human studies, time-to-detectable viremia can be used as a rough estimate of the size of the RCR in a cohort of individuals (Pinkevych et al., 2015) and has been used to compare the size of the RCR between groups (Bar et al., 2016; Colby et al., 2018; Li et al., 2015). However, this approach has very limited statistical power, due in part to variability of the exact time first detectable viremia occurs and wide confidence intervals within individual groups. It has been estimated that groups of >100 individuals may be required to convincingly demonstrate even quite substantial changes in the reservoir size with treatment (Pinkevych et al., 2015; Moore et al., 2019). Therefore, the efficacy of these interventions is often quantified by surrogate measures such as the levels of various forms of HIV DNA, cell-associated HIV RNA, or the ability of latent cells to be reactivated ex-vivo (Rasmussen et al., 2014; Søgaard et al., 2015; Elliott et al., 2014). Understanding how these surrogate measures relate to the frequency of HIV reactivation and the resulting duration of viral rebound free remission after treatment interruption would greatly facilitate the accurate evaluation of novel therapies to reduce the RCR (Eriksson et al., 2013; Sharaf and Li, 2017).

The frequency of individual infected cells able to activate and produce sufficient viral progeny to allow for a spreading infection is postulated to be determined by a number of factors, including; (i) the number of infected cells harbouring HIV DNA, (ii) the fraction of that DNA representing replication-competent proviruses, (iii) the viral integration site and level of cellular activation and epigenetic and transcriptional states of individual proviruses (which together may help determine the per cell frequency of reactivation of virus; Hill, 2017; Yukl et al., 2018), and (iv) the level of immune control of viral reactivation. A number of studies have reported that levels of integrated HIV DNA are a poor predictor of time-to-recrudescence of virus. Williams et al. (2014) investigated a cohort of 154 individuals who initiated treatment in the chronic phase of infection and underwent treatment interruption, and found that integrated HIV DNA was not a significant predictor of time-to-recrudescence (defined as time to reach 50 copies/ml of virus in plasma; Williams et al., 2014). In a study by Calin et al. (2016) , patients selected for having very low HIV DNA levels failed to show a delayed time-to-rebound viremia. A recent study by Colby et al. (2018) showing time-to-recrudescence of virus in patients with very low HIV DNA who were treated within weeks of infection (in Fiebig stage 1) suggests relatively little change in the frequency of HIV reactivation from latency compared to cohorts treated in chronic infection. However, other studies have suggested that early ART treatment in combination with lower levels of HIV DNA at interruption may be associated with delayed time-to-HIV-relapse (Li et al., 2015; Williams et al., 2014; Assoumou et al., 2015; Wen et al., 2018; Steingrover et al., 2008). Therefore, the relationship between surrogate measures of HIV ‘reservoir’ (in peripheral blood) and the later time-to-viral-rebound remains unclear. Moreover, in studies of the efficacy of candidate latency reducing strategies, it is unclear if changes measured in peripheral blood will be associated with meaningful changes in post-treatment time to viral rebound (Petravic et al., 2017).

In this study, we used a macaque model of AIDS virus infection, employing the barcoded SIVmac239M virus to investigate the dynamics of SIV rebound after treatment interruption, and assess what host factors might correlate with or predict cellular activation and subsequent viral rebound. The use of the barcoded virus allows estimation of the frequency of SIV reactivation from latency in individual animals (Fennessey et al., 2017). We studied viral reactivation rates in macaques in which cART was initiated at different times after infection (between day 4 and day 27), allowing for differing extents of viral dissemination reflected by different levels of viral DNA (vDNA) (spanning two logs) in PBMC. SIV DNA levels decayed under treatment and when cART was stopped after 43–68 weeks of treatment, we were able to estimate the frequency of SIV reactivation following treatment interruption. While SIV DNA level in PBMC just before treatment interruption was a predictor of SIV reactivation, this relationship appeared extremely ‘flat’. That is, an ≈100 fold increase in SIV DNA levels was associated with only a two-fold difference in reactivation rates during a analytic treatment interruption (ATI). We also investigated a variety of other factors that could plausibly affect the relationship between vDNA levels and the frequency of reactivation from latency. This work suggests that the rebound competent reservoir may be saturable and can be seeded early in the course of infection Subsequent increases in viral DNA due to a delay in therapy only marginally alters the dynamics of post-ATI rebound.

A major question in HIV ‘cure’ research is whether biomarkers in peripheral blood correlate with and can be used to predict the duration of HIV remission after treatment interruption (Margolis and Deeks, 2019). We used variable timing of the initiation of cART, allowing different extents of viral dissemination before starting suppressive treatment, to produce SIV reservoirs of different sizes. The levels of SIV DNA at the time of ATI varied by >100 fold between different animals, from <3.2 copies/million PBMC in animals treated beginning on day 4, to >1000 copies in some animals starting treatment on day 10. The frequency of successful SIV reactivation after ATI also varied by ≈8-fold (from 0.4 to 3.10 reactivations per day). This is remarkable when one considers the different levels of exposure to infection. The peak viral load varied ≈350 fold between animals starting treatment on day 4 (geometric mean = 1.5×10⁵ copies ml⁻¹, CI of geo. mean (4.9 × 10⁴, 4.4 × 10⁵), n = 6) and animals treated on day 27 (geometric mean = 5.2×10⁷ copies ml⁻¹, CI of geo. mean (1.8 × 10⁷, 1.5 × 10⁸), n = 5). Similarly, the ‘area under the curve’ of viral exposure varied by >4,000 fold (geometric means 5.2 × 10⁴, CI of geo. mean (1.8 × 10⁴, 1.5 × 10⁵) n = 6, vs 2.2 × 10⁸, CI of geo. mean (1.3 × 10⁸, 3.8 × 10⁸), n = 5). However, although the prolonged viral replication led to an increase in SIV CA-DNA (Archin et al., 2012), the additional exposure to virus between day 4 and 27 was associated with only approximately double the frequency of reactivation. This suggests that most of the additional infected cells, reflected in increased SIV CA-DNA generated between day 4 and day 27 did not significantly contribute to rebound viremia. The results suggest that key anatomic and/or cellular compartments that represent the major contributors to the rebound viremia may be limited and saturable as early as d4 after a high dose intravenous inoculation. The fact that very large changes in SIV CA-DNA are associated with only small changes in the frequency of reactivation perhaps goes some way to explaining why HIV DNA is a relatively poor predictor of time-to-reactivation after treatment interruption in many studies (Li et al., 2015; Williams et al., 2014; Calin et al., 2016; Assoumou et al., 2015).

One potential explanation for our observations that this SIV DNA that accumulates later in infection is not as replication competent as DNA integrated earlier in infection. Sequencing studies suggest that only 2.4% of HIV proviral DNA is intact in patients starting cART in chronic infection (Hiener et al., 2017; Bruner et al., 2016; Bruner et al., 2019). Thus, for example, if SIV proviral DNA integrated at day 4 was 100% replication competent, but DNA integrated after this were 99% defective, this might explain the observed relationship between reactivation frequency and SIV DNA copies. However, near full length viral sequencing from animals treated on day 10 and day 27 revealed that the vast majority (>80%) of sequences were intact. Studies of SIV_mac251 DNA in animals treated in chronic infection (95 weeks post-infection) for 36 weeks revealed that 28% remained intact (Bender et al., 2019). Thus, it appears that at least over the first year or two of infection, SIV DNA remains intact at much higher rates than HIV. Taken together, it is clear that the observed differences in reactivation frequency between animals treated on different days cannot be explained by different proportions of replication competent virus.

Previous work has suggested that extremely early treatment can either delay or prevent post-ATI SIV viral recrudescence (Whitney et al., 2014; Whitney et al., 2018; Okoye et al., 2018). In a study by Whitney et al, intrarectal inoculation of 500 TCID50 of SIVmac251 and 6 months of cART begun on day three post-infection (when animals did not have detectable plasma virus) resulted in delayed viral rebound. However, animals treated for 6 months beginning on days 7, 10 and 14 post infection showed similar rebound times (Whitney et al., 2014). Okoye et al. (2018) infected macaques with 500 IU of SIVmac239, and found that animals treated for nearly 2 years beginning on day 4–5 post infection (with viral loads of 60–1100 copies/ml at treatment initiation) failed to rebound after treatment interruption (Okoye et al., 2018). By contrast, our study involved intravenous inoculation of 2.2 × 10⁵ IU of SIVmac239M viral stock, and animals treated on day four had maximal viral loads of between 3.8 × 10⁴ and 9.1 × 10⁵ (geometric mean of 1.5 × 10⁵) copies/ml of plasma virus. Taken together, these data suggests that extremely early treatment (with viral loads at treatment <1000–10,000 copies/ml) may severely restrict the establishment of the RCR. However, once plasma viremia passes around 10,000 copies/ml the subsequent post-ATI frequency of reactivation is established. Extending the duration of infection prior to starting cART (from day 4 to day 27) may greatly increase the level of proviral DNA (Archin et al., 2012) and cell-associated RNA in PBMC, but does comparatively little to increase the frequency of reactivation.

Although animal models of infection provide the ability to control factors such as time and duration of treatment, it is not clear whether these same lessons are applicable to HIV infection. The RCR of SIV may differ from that in HIV, either because of the cells targeted for infection, or because of differences in infected cell behaviour. For example, cellular entry using alternative receptors such as CXCR6 and GPR15, or differences in host restriction factors may lead to differences in infected cell phenotypes in SIV (Riddick et al., 2016; Reynolds et al., 2011). Similarly, although coinfection of cells with multiple copies of HIV is thought to be low (Josefsson et al., 2011), rates of SIV coinfection of cells were not determined in our study and could play a role in determining the frequency of reactivation per DNA copy observed. However, recent studies have suggested that integrations sites and the propensity for clonal expansion of latent virus are similar between SIV and HIV (Ferris et al., 2019). The initial inoculum of virus in our studies is several orders of magnitude higher than that estimated in HIV (Keele et al., 2008), and thus this may disproportionately contribute to the early establishment and potential saturation of the SIV RCR. However, since viral loads increase around 10-fold per day in early infection (Figure 1a,c,e), it is unlikely that this initial inoculum virus equates to >0.1% of virus present even at day four post-infection (the timing of earliest treatment). Thus, it seems unlikely that the higher inoculum in SIV infection could be a major contributor to the SIV reservoir measured later during treatment.

Our finding of early ‘filling’ of the RCR appear consistent with the more limited observations of reactivation in early HIV infection. In particular, although some studies have shown delayed post-ATI rebound in HIV patients treated early in infection and with small reservoirs (Li et al., 2015; Steingrover et al., 2008), the size of the reservoir appears only weakly associated with the time to recrudescence of HIV after treatment interruption (Li et al., 2015; Williams et al., 2014). A number of recent studies have shown that even in individuals treated extremely early and/or with a very small reservoir, the time to viral recrudescence can be very similar to that observed in individuals treated beginning in chronic infection and with a presumably larger reservoir (Calin et al., 2016; Assoumou et al., 2015). So while the case study of the Mississippi baby suggests that extremely early treatment can significantly delay viral rebound (Luzuriaga et al., 2015), Colby et al have recently shown that individuals treated very early during Fiebig stage one have only slightly longer average time to detection of plasma viral rebound than individuals starting treatment in chronic infection (median 26 vs. 14 days), despite considerably lower HIV DNA levels (Colby et al., 2018). This strongly supports our observations with SIV, in that the frequency of post-ATI viral reactivation is not only established extremely early in infection, but once established does not increase much from early primary infection to chronic infection.

A significant limitation of our study is that we were not able to perform quantitative viral outgrowth assays (QVOA) (Finzi et al., 1999) on our macaque samples, due to limitations in available cell numbers in rhesus macaques. Comparison of HIV sequences between circulating provirus and ex vivo-inducible virus show that there is a significant sequence overlap between these two compartments (Lu et al., 2019). However, a recent study showed that despite the overlap between provirus and QVOA, there was no overlap of either compartment with post-ATI rebound virus (Lu et al., 2019). This was taken to suggest that rebound virus originates from a different pool of cells to that reflected in circulating provirus. Combined with our SIV results, it is clear that most current measures of the HIV reservoir are in fact measuring total CA-HIV or even presumptively intact proviruses from PBMCs (Bullen et al., 2014) but not the major contributor to post-ATI rebound viremia. Further work is clearly required to focus attention on what must be a small proportion of the total latently infected cells (compared to total provirus in PBMC), that appear to be the major drivers of rebound.

The observation that DNA levels do not predict reactivation frequency between individuals does not answer the question of whether it might be a useful predictor of the impact of anti-latency therapies within an individual. Thus, for example, if the RCR declines (or is reduced by treatment) in proportion to the total HIV DNA, then monitoring total DNA might still be useful to assess the impact of anti-latency treatments on reactivation frequency within an individual patient over time. The stability of the HIV reservoir has been measured in humans by measuring the decay of both HIV DNA and ex vivo inducible HIV, which were both found to decay with a half-life of >4 years (Besson et al., 2014; Siliciano et al., 2003). Our study suggests that, after 6 months of therapy, SIV DNA decays with a half-life of around 345 days and SIV CA-RNA decays faster (half-life = 86 days, p<0.016, LME model). The frequency of reactivation declines slightly faster than SIV-DNA with a half-life of ≈216 days (although not significantly different, p=0.53, LME model). Thus, reactivation frequency appears to decay at the same or faster rate than DNA levels in PBMC.

It is important to understand the mechanisms of latency formation, maintenance and reactivation when designing interventions aimed to induce long-term remission. Our work suggests that the RCR can be established very early following infection and provides some explanation as to why treating early may only have a small effect on subsequent time-to-recrudescence of virus (even if total HIV DNA levels continue to increase). A major impediment in HIV remission studies is finding a proxy measure that can predict the time-to-recrudescence of virus, without patients undergoing the risks of treatment interruption (Julg et al., 2019). This may be challenging, as our studies suggest that total PBMC SIV CA-DNA is not the main determinant of post-treatment rebound. Our work also suggests that efforts to develop optimal interventions to prevent HIV reactivation following treatment interruption will require a more thorough understanding of viral latency.

Source data files have been provided for the figures.

1. 1. Archin NM
   2. Vaidya NK
   3. Kuruc JD
   4. Liberty AL
   5. Wiegand A
   6. Kearney MF
   7. Cohen MS
   8. Coffin JM
   9. Bosch RJ
   10. Gay CL
   11. Eron JJ
   12. Margolis DM
   13. Perelson AS

   (2012) Immediate antiviral therapy appears to restrict resting CD4+ cell HIV-1 infection without accelerating the decay of latent infection

   PNAS 109:9523–9528.

   https://doi.org/10.1073/pnas.1120248109

   • PubMed
   • Google Scholar
2. 1. Assoumou L
   2. Weiss L
   3. Piketty C
   4. Burgard M
   5. Melard A
   6. Girard PM
   7. Rouzioux C
   8. Costagliola D
   9. ANRS 116 SALTO study group

   (2015) A low HIV-DNA level in peripheral blood mononuclear cells at antiretroviral treatment interruption predicts a higher probability of maintaining viral control

   Aids 29:2003–2007.

   https://doi.org/10.1097/QAD.0000000000000734

   • PubMed
   • Google Scholar
3. 1. Bar KJ
   2. Sneller MC
   3. Harrison LJ
   4. Justement JS
   5. Overton ET
   6. Petrone ME
   7. Salantes DB
   8. Seamon CA
   9. Scheinfeld B
   10. Kwan RW
   11. Learn GH
   12. Proschan MA
   13. Kreider EF
   14. Blazkova J
   15. Bardsley M
   16. Refsland EW
   17. Messer M
   18. Clarridge KE
   19. Tustin NB
   20. Madden PJ
   21. Oden K
   22. O'Dell SJ
   23. Jarocki B
   24. Shiakolas AR
   25. Tressler RL
   26. Doria-Rose NA
   27. Bailer RT
   28. Ledgerwood JE
   29. Capparelli EV
   30. Lynch RM
   31. Graham BS
   32. Moir S
   33. Koup RA
   34. Mascola JR
   35. Hoxie JA
   36. Fauci AS
   37. Tebas P
   38. Chun TW

   (2016) Effect of HIV antibody VRC01 on viral rebound after treatment interruption

   New England Journal of Medicine 375:2037–2050.

   https://doi.org/10.1056/NEJMoa1608243

   • PubMed
   • Google Scholar
4. 1. Barton KM
   2. Palmer SE

   (2016) How to define the latent reservoir: tools of the trade

   Current HIV/AIDS Reports 13:77–84.

   https://doi.org/10.1007/s11904-016-0304-1

   • PubMed
   • Google Scholar
5. 1. Bender AM
   2. Simonetti FR
   3. Kumar MR
   4. Fray EJ
   5. Bruner KM
   6. Timmons AE
   7. Tai KY
   8. Jenike KM
   9. Antar AAR
   10. Liu PT
   11. Ho YC
   12. Raugi DN
   13. Seydi M
   14. Gottlieb GS
   15. Okoye AA
   16. Del Prete GQ
   17. Picker LJ
   18. Mankowski JL
   19. Lifson JD
   20. Siliciano JD
   21. Laird GM
   22. Barouch DH
   23. Clements JE
   24. Siliciano RF

   (2019) The landscape of persistent viral genomes in ART-Treated SIV, SHIV, and HIV-2 infections

   Cell Host & Microbe 26:73–85.

   https://doi.org/10.1016/j.chom.2019.06.005

   • PubMed
   • Google Scholar
6. 1. Besson GJ
   2. Lalama CM
   3. Bosch RJ
   4. Gandhi RT
   5. Bedison MA
   6. Aga E
   7. Riddler SA
   8. McMahon DK
   9. Hong F
   10. Mellors JW

   (2014) HIV-1 DNA decay dynamics in blood during more than a decade of suppressive antiretroviral therapy

   Clinical Infectious Diseases 59:1312–1321.

   https://doi.org/10.1093/cid/ciu585

   • PubMed
   • Google Scholar
7. 1. Bruner KM
   2. Hosmane NN
   3. Siliciano RF

   (2015) Towards an HIV-1 cure: measuring the latent reservoir

   Trends in Microbiology 23:192–203.

   https://doi.org/10.1016/j.tim.2015.01.013

   • PubMed
   • Google Scholar
8. 1. Bruner KM
   2. Murray AJ
   3. Pollack RA
   4. Soliman MG
   5. Laskey SB
   6. Capoferri AA
   7. Lai J
   8. Strain MC
   9. Lada SM
   10. Hoh R
   11. Ho YC
   12. Richman DD
   13. Deeks SG
   14. Siliciano JD
   15. Siliciano RF

   (2016) Defective proviruses rapidly accumulate during acute HIV-1 infection

   Nature Medicine 22:1043–1049.

   https://doi.org/10.1038/nm.4156

   • PubMed
   • Google Scholar
9. 1. Bruner KM
   2. Wang Z
   3. Simonetti FR
   4. Bender AM
   5. Kwon KJ
   6. Sengupta S
   7. Fray EJ
   8. Beg SA
   9. Antar AAR
   10. Jenike KM
   11. Bertagnolli LN
   12. Capoferri AA
   13. Kufera JT
   14. Timmons A
   15. Nobles C
   16. Gregg J
   17. Wada N
   18. Ho YC
   19. Zhang H
   20. Margolick JB
   21. Blankson JN
   22. Deeks SG
   23. Bushman FD
   24. Siliciano JD
   25. Laird GM
   26. Siliciano RF

   (2019) A quantitative approach for measuring the reservoir of latent HIV-1 proviruses

   Nature 566:120–125.

   https://doi.org/10.1038/s41586-019-0898-8

   • PubMed
   • Google Scholar
10. 1. Bullen CK
    2. Laird GM
    3. Durand CM
    4. Siliciano JD
    5. Siliciano RF

    (2014) New ex vivo approaches distinguish effective and ineffective single agents for reversing HIV-1 latency in vivo

    Nature Medicine 20:425–429.

    https://doi.org/10.1038/nm.3489

    • PubMed
    • Google Scholar
11. 1. Calin R
    2. Hamimi C
    3. Lambert-Niclot S
    4. Carcelain G
    5. Bellet J
    6. Assoumou L
    7. Tubiana R
    8. Calvez V
    9. Dudoit Y
    10. Costagliola D
    11. Autran B
    12. Katlama C
    13. ULTRASTOP Study Group

    (2016) Treatment interruption in chronically HIV-infected patients with an ultralow HIV reservoir

    Aids 30:761–769.

    https://doi.org/10.1097/QAD.0000000000000987

    • PubMed
    • Google Scholar
12. 1. Colby DJ
    2. Trautmann L
    3. Pinyakorn S
    4. Leyre L
    5. Pagliuzza A
    6. Kroon E
    7. Rolland M
    8. Takata H
    9. Buranapraditkun S
    10. Intasan J
    11. Chomchey N
    12. Muir R
    13. Haddad EK
    14. Tovanabutra S
    15. Ubolyam S
    16. Bolton DL
    17. Fullmer BA
    18. Gorelick RJ
    19. Fox L
    20. Crowell TA
    21. Trichavaroj R
    22. O'Connell R
    23. Chomont N
    24. Kim JH
    25. Michael NL
    26. Robb ML
    27. Phanuphak N
    28. Ananworanich J
    29. RV411 study group

    (2018) Rapid HIV RNA rebound after antiretroviral treatment interruption in persons durably suppressed in Fiebig I acute HIV infection

    Nature Medicine 24:923–926.

    https://doi.org/10.1038/s41591-018-0026-6

    • PubMed
    • Google Scholar
13. 1. Cromer D
    2. Pinkevych M
    3. Rasmussen TA
    4. Lewin SR
    5. Kent SJ
    6. Davenport MP

    (2017) Modeling of antilatency treatment in HIV: what is the optimal duration of antiretroviral Therapy-Free HIV remission?

    Journal of Virology 91:e01395-17.

    https://doi.org/10.1128/JVI.01395-17

    • PubMed
    • Google Scholar
14. 1. Davenport MP
    2. Khoury DS
    3. Cromer D
    4. Lewin SR
    5. Kelleher AD
    6. Kent SJ

    (2019) Functional cure of HIV: the scale of the challenge

    Nature Reviews Immunology 19:45–54.

    https://doi.org/10.1038/s41577-018-0085-4

    • PubMed
    • Google Scholar
15. 1. Deeks SG
    2. Lewin SR
    3. Ross AL
    4. Ananworanich J
    5. Benkirane M
    6. Cannon P
    7. Chomont N
    8. Douek D
    9. Lifson JD
    10. Lo YR
    11. Kuritzkes D
    12. Margolis D
    13. Mellors J
    14. Persaud D
    15. Tucker JD
    16. Barre-Sinoussi F
    17. Alter G
    18. Auerbach J
    19. Autran B
    20. Barouch DH
    21. Behrens G
    22. Cavazzana M
    23. Chen Z
    24. Cohen ÉA
    25. Corbelli GM
    26. Eholié S
    27. Eyal N
    28. Fidler S
    29. Garcia L
    30. Grossman C
    31. Henderson G
    32. Henrich TJ
    33. Jefferys R
    34. Kiem HP
    35. McCune J
    36. Moodley K
    37. Newman PA
    38. Nijhuis M
    39. Nsubuga MS
    40. Ott M
    41. Palmer S
    42. Richman D
    43. Saez-Cirion A
    44. Sharp M
    45. Siliciano J
    46. Silvestri G
    47. Singh J
    48. Spire B
    49. Taylor J
    50. Tolstrup M
    51. Valente S
    52. van Lunzen J
    53. Walensky R
    54. Wilson I
    55. Zack J
    56. International AIDS Society Towards a Cure Working Group

    (2016) International AIDS society global scientific strategy: towards an HIV cure 2016

    Nature Medicine 22:839–850.

    https://doi.org/10.1038/nm.4108

    • PubMed
    • Google Scholar
16. 1. Elliott JH
    2. Wightman F
    3. Solomon A
    4. Ghneim K
    5. Ahlers J
    6. Cameron MJ
    7. Smith MZ
    8. Spelman T
    9. McMahon J
    10. Velayudham P
    11. Brown G
    12. Roney J
    13. Watson J
    14. Prince MH
    15. Hoy JF
    16. Chomont N
    17. Fromentin R
    18. Procopio FA
    19. Zeidan J
    20. Palmer S
    21. Odevall L
    22. Johnstone RW
    23. Martin BP
    24. Sinclair E
    25. Deeks SG
    26. Hazuda DJ
    27. Cameron PU
    28. Sékaly RP
    29. Lewin SR

    (2014) Activation of HIV transcription with short-course vorinostat in HIV-infected patients on suppressive antiretroviral therapy

    PLOS Pathogens 10:e1004473.

    https://doi.org/10.1371/journal.ppat.1004473

    • PubMed
    • Google Scholar
17. 1. Eriksson S
    2. Graf EH
    3. Dahl V
    4. Strain MC
    5. Yukl SA
    6. Lysenko ES
    7. Bosch RJ
    8. Lai J
    9. Chioma S
    10. Emad F
    11. Abdel-Mohsen M
    12. Hoh R
    13. Hecht F
    14. Hunt P
    15. Somsouk M
    16. Wong J
    17. Johnston R
    18. Siliciano RF
    19. Richman DD
    20. O'Doherty U
    21. Palmer S
    22. Deeks SG
    23. Siliciano JD

    (2013) Comparative analysis of measures of viral reservoirs in HIV-1 eradication studies

    PLOS Pathogens 9:e1003174.

    https://doi.org/10.1371/journal.ppat.1003174

    • PubMed
    • Google Scholar
18. 1. Fennessey CM
    2. Pinkevych M
    3. Immonen TT
    4. Reynaldi A
    5. Venturi V
    6. Nadella P
    7. Reid C
    8. Newman L
    9. Lipkey L
    10. Oswald K
    11. Bosche WJ
    12. Trivett MT
    13. Ohlen C
    14. Ott DE
    15. Estes JD
    16. Del Prete GQ
    17. Lifson JD
    18. Davenport MP
    19. Keele BF

    (2017) Genetically-barcoded SIV facilitates enumeration of rebound variants and estimation of reactivation rates in nonhuman primates following interruption of suppressive antiretroviral therapy

    PLOS Pathogens 13:e1006359.

    https://doi.org/10.1371/journal.ppat.1006359

    • PubMed
    • Google Scholar
19. 1. Ferris AL
    2. Wells DW
    3. Guo S
    4. Del Prete GQ
    5. Swanstrom AE
    6. Coffin JM
    7. Wu X
    8. Lifson JD
    9. Hughes SH

    (2019) Clonal expansion of SIV-infected cells in macaques on antiretroviral therapy is similar to that of HIV-infected cells in humans

    PLOS Pathogens 15:e1007869.

    https://doi.org/10.1371/journal.ppat.1007869

    • PubMed
    • Google Scholar
20. 1. Finzi D
    2. Blankson J
    3. Siliciano JD
    4. Margolick JB
    5. Chadwick K
    6. Pierson T
    7. Smith K
    8. Lisziewicz J
    9. Lori F
    10. Flexner C
    11. Quinn TC
    12. Chaisson RE
    13. Rosenberg E
    14. Walker B
    15. Gange S
    16. Gallant J
    17. Siliciano RF

    (1999) Latent infection of CD4+ T cells provides a mechanism for lifelong persistence of HIV-1, even in patients on effective combination therapy

    Nature Medicine 5:512–517.

    https://doi.org/10.1038/8394

    • PubMed
    • Google Scholar
21. 1. Hansen SG
    2. Ford JC
    3. Lewis MS
    4. Ventura AB
    5. Hughes CM
    6. Coyne-Johnson L
    7. Whizin N
    8. Oswald K
    9. Shoemaker R
    10. Swanson T
    11. Legasse AW
    12. Chiuchiolo MJ
    13. Parks CL
    14. Axthelm MK
    15. Nelson JA
    16. Jarvis MA
    17. Piatak M
    18. Lifson JD
    19. Picker LJ

    (2011) Profound early control of highly pathogenic SIV by an effector memory T-cell vaccine

    Nature 473:523–527.

    https://doi.org/10.1038/nature10003

    • PubMed
    • Google Scholar
22. 1. Hiener B
    2. Horsburgh BA
    3. Eden JS
    4. Barton K
    5. Schlub TE
    6. Lee E
    7. von Stockenstrom S
    8. Odevall L
    9. Milush JM
    10. Liegler T
    11. Sinclair E
    12. Hoh R
    13. Boritz EA
    14. Douek D
    15. Fromentin R
    16. Chomont N
    17. Deeks SG
    18. Hecht FM
    19. Palmer S

    (2017) Identification of genetically intact HIV-1 proviruses in specific CD4⁺ T Cells from Effectively Treated Participants

    Cell Reports 21:813–822.

    https://doi.org/10.1016/j.celrep.2017.09.081

    • PubMed
    • Google Scholar
23. 1. Hill AL
    2. Rosenbloom DI
    3. Goldstein E
    4. Hanhauser E
    5. Kuritzkes DR
    6. Siliciano RF
    7. Henrich TJ

    (2016) Real-Time predictions of reservoir size and rebound time during antiretroviral therapy interruption trials for HIV

    PLOS Pathogens 12:e1005535.

    https://doi.org/10.1371/journal.ppat.1005535

    • PubMed
    • Google Scholar
24. 1. Hill AL

    (2017) Mathematical models of HIV latency

    Current Topics in Microbiology and Immunology 417:131–156.

    https://doi.org/10.1007/82_2017_77

    • Google Scholar
25. 1. Josefsson L
    2. King MS
    3. Makitalo B
    4. Brännström J
    5. Shao W
    6. Maldarelli F
    7. Kearney MF
    8. Hu WS
    9. Chen J
    10. Gaines H
    11. Mellors JW
    12. Albert J
    13. Coffin JM
    14. Palmer SE

    (2011) Majority of CD4+ T cells from peripheral blood of HIV-1-infected individuals contain only one HIV DNA molecule

    PNAS 108:11199–11204.

    https://doi.org/10.1073/pnas.1107729108

    • PubMed
    • Google Scholar
26. 1. Julg B
    2. Dee L
    3. Ananworanich J
    4. Barouch DH
    5. Bar K
    6. Caskey M
    7. Colby DJ
    8. Dawson L
    9. Dong KL
    10. Dubé K
    11. Eron J
    12. Frater J
    13. Gandhi RT
    14. Geleziunas R
    15. Goulder P
    16. Hanna GJ
    17. Jefferys R
    18. Johnston R
    19. Kuritzkes D
    20. Li JZ
    21. Likhitwonnawut U
    22. van Lunzen J
    23. Martinez-Picado J
    24. Miller V
    25. Montaner LJ
    26. Nixon DF
    27. Palm D
    28. Pantaleo G
    29. Peay H
    30. Persaud D
    31. Salzwedel J
    32. Salzwedel K
    33. Schacker T
    34. Sheikh V
    35. Søgaard OS
    36. Spudich S
    37. Stephenson K
    38. Sugarman J
    39. Taylor J
    40. Tebas P
    41. Tiemessen CT
    42. Tressler R
    43. Weiss CD
    44. Zheng L
    45. Robb ML
    46. Michael NL
    47. Mellors JW
    48. Deeks SG
    49. Walker BD

    (2019) Recommendations for analytical antiretroviral treatment interruptions in HIV research trials-report of a consensus meeting

    The Lancet HIV 6:e259–e268.

    https://doi.org/10.1016/S2352-3018(19)30052-9

    • PubMed
    • Google Scholar
27. 1. Keele BF
    2. Giorgi EE
    3. Salazar-Gonzalez JF
    4. Decker JM
    5. Pham KT
    6. Salazar MG
    7. Sun C
    8. Grayson T
    9. Wang S
    10. Li H
    11. Wei X
    12. Jiang C
    13. Kirchherr JL
    14. Gao F
    15. Anderson JA
    16. Ping LH
    17. Swanstrom R
    18. Tomaras GD
    19. Blattner WA
    20. Goepfert PA
    21. Kilby JM
    22. Saag MS
    23. Delwart EL
    24. Busch MP
    25. Cohen MS
    26. Montefiori DC
    27. Haynes BF
    28. Gaschen B
    29. Athreya GS
    30. Lee HY
    31. Wood N
    32. Seoighe C
    33. Perelson AS
    34. Bhattacharya T
    35. Korber BT
    36. Hahn BH
    37. Shaw GM

    (2008) Identification and characterization of transmitted and early founder virus envelopes in primary HIV-1 infection

    PNAS 105:7552–7557.

    https://doi.org/10.1073/pnas.0802203105

    • PubMed
    • Google Scholar
28. 1. Li JZ
    2. Etemad B
    3. Ahmed H
    4. Aga E
    5. Bosch RJ
    6. Mellors JW

    (2015) The size of the expressed HIV reservoir predicts timing of viral rebound after treatment interruption

    AIDS 30:343–353.

    https://doi.org/10.1097/01.aids.0000499516.66930.89

    • Google Scholar
29. 1. Li H
    2. Wang S
    3. Kong R
    4. Ding W
    5. Lee FH
    6. Parker Z
    7. Kim E
    8. Learn GH
    9. Hahn P
    10. Policicchio B
    11. Brocca-Cofano E
    12. Deleage C
    13. Hao X
    14. Chuang GY
    15. Gorman J
    16. Gardner M
    17. Lewis MG
    18. Hatziioannou T
    19. Santra S
    20. Apetrei C
    21. Pandrea I
    22. Alam SM
    23. Liao HX
    24. Shen X
    25. Tomaras GD
    26. Farzan M
    27. Chertova E
    28. Keele BF
    29. Estes JD
    30. Lifson JD
    31. Doms RW
    32. Montefiori DC
    33. Haynes BF
    34. Sodroski JG
    35. Kwong PD
    36. Hahn BH
    37. Shaw GM

    (2016) Envelope residue 375 substitutions in simian-human immunodeficiency viruses enhance CD4 binding and replication in rhesus macaques

    PNAS 113:E3413–E3422.

    https://doi.org/10.1073/pnas.1606636113

    • PubMed
    • Google Scholar
30. 1. Long S
    2. Fennessey CM
    3. Newman L
    4. Reid C
    5. O'Brien SP
    6. Li Y
    7. Del Prete GQ
    8. Lifson JD
    9. Gorelick RJ
    10. Keele BF

    (2019) Evaluating the intactness of persistent viral genomes in SIV-infected rhesus macaques initiating antiretroviral therapy within one year of infection

    Journal of Virology JVI.01308-19.

    https://doi.org/10.1128/JVI.01308-19

    • PubMed
    • Google Scholar
31. 1. Lopker MJ
    2. Del Prete GQ
    3. Estes JD
    4. Li H
    5. Reid C
    6. Newman L
    7. Lipkey L
    8. Camus C
    9. Easlick JL
    10. Wang S
    11. Decker JM
    12. Bar KJ
    13. Learn G
    14. Pal R
    15. Weiss DE
    16. Hahn BH
    17. Lifson JD
    18. Shaw GM
    19. Keele BF

    (2016) Derivation and characterization of pathogenic transmitted/Founder molecular clones from simian immunodeficiency virus SIVsmE660 and SIVmac251 following mucosal infection

    Journal of Virology 90:8435–8453.

    https://doi.org/10.1128/JVI.00718-16

    • PubMed
    • Google Scholar
32. 1. Lorenzi JC
    2. Cohen YZ
    3. Cohn LB
    4. Kreider EF
    5. Barton JP
    6. Learn GH
    7. Oliveira T
    8. Lavine CL
    9. Horwitz JA
    10. Settler A
    11. Jankovic M
    12. Seaman MS
    13. Chakraborty AK
    14. Hahn BH
    15. Caskey M
    16. Nussenzweig MC

    (2016) Paired quantitative and qualitative assessment of the replication-competent HIV-1 reservoir and comparison with integrated proviral DNA

    PNAS 113:E7908–E7916.

    https://doi.org/10.1073/pnas.1617789113

    • PubMed
    • Google Scholar
33. Conference

    1. Lu CL
    2. Pai J
    3. Nogueira L
    4. Mendoza P
    5. Oliviera TY
    6. Barton J

    (2019)

    Relationship between intact HIV-1 proviruses and plasma rebound viruses

    CROI.

    • Google Scholar
34. 1. Luzuriaga K
    2. Gay H
    3. Ziemniak C
    4. Sanborn KB
    5. Somasundaran M
    6. Rainwater-Lovett K
    7. Mellors JW
    8. Rosenbloom D
    9. Persaud D

    (2015) Viremic relapse after HIV-1 remission in a perinatally infected child

    New England Journal of Medicine 372:786–788.

    https://doi.org/10.1056/NEJMc1413931

    • PubMed
    • Google Scholar
35. 1. Margolis DM
    2. Deeks SG

    (2019) How unavoidable are analytical treatment interruptions in HIV Cure-Related studies?

    The Journal of Infectious Diseases 220:S24–S26.

    https://doi.org/10.1093/infdis/jiz222

    • PubMed
    • Google Scholar
36. 1. Metcalf Pate KA
    2. Pohlmeyer CW
    3. Walker-Sperling VE
    4. Foote JB
    5. Najarro KM
    6. Cryer CG
    7. Salgado M
    8. Gama L
    9. Engle EL
    10. Shirk EN
    11. Queen SE
    12. Chioma S
    13. Vermillion MS
    14. Bullock B
    15. Li M
    16. Lyons CE
    17. Adams RJ
    18. Zink MC
    19. Clements JE
    20. Mankowski JL
    21. Blankson JN

    (2015) A murine viral outgrowth assay to detect residual HIV type 1 in patients with undetectable viral loads

    Journal of Infectious Diseases 212:1387–1396.

    https://doi.org/10.1093/infdis/jiv230

    • PubMed
    • Google Scholar
37. 1. Moore CL
    2. Stöhr W
    3. Crook AM
    4. Richert L
    5. Leliévre JD
    6. Pantaleo G
    7. García F
    8. Vella S
    9. Lévy Y
    10. Thiébaut R
    11. McCormack S

    (2019) Multi-arm, multi-stage randomised controlled trials for evaluating therapeutic HIV cure interventions

    The Lancet HIV 6:e334–e340.

    https://doi.org/10.1016/S2352-3018(19)30082-7

    • PubMed
    • Google Scholar
38. 1. Okoye AA
    2. Hansen SG
    3. Vaidya M
    4. Fukazawa Y
    5. Park H
    6. Duell DM
    7. Lum R
    8. Hughes CM
    9. Ventura AB
    10. Ainslie E
    11. Ford JC
    12. Morrow D
    13. Gilbride RM
    14. Legasse AW
    15. Hesselgesser J
    16. Geleziunas R
    17. Li Y
    18. Oswald K
    19. Shoemaker R
    20. Fast R
    21. Bosche WJ
    22. Borate BR
    23. Edlefsen PT
    24. Axthelm MK
    25. Picker LJ
    26. Lifson JD

    (2018) Early antiretroviral therapy limits SIV reservoir establishment to delay or prevent post-treatment viral rebound

    Nature Medicine 24:1430–1440.

    https://doi.org/10.1038/s41591-018-0130-7

    • PubMed
    • Google Scholar
39. 1. Petravic J
    2. Rasmussen TA
    3. Lewin SR
    4. Kent SJ
    5. Davenport MP

    (2017) Relationship between measures of HIV reactivation and decline of the latent reservoir under Latency-Reversing agents

    Journal of Virology 91:e02092.

    https://doi.org/10.1128/JVI.02092-16

    • PubMed
    • Google Scholar
40. 1. Pinkevych M
    2. Cromer D
    3. Tolstrup M
    4. Grimm AJ
    5. Cooper DA
    6. Lewin SR
    7. Søgaard OS
    8. Rasmussen TA
    9. Kent SJ
    10. Kelleher AD
    11. Davenport MP

    (2015) HIV reactivation from latency after treatment interruption occurs on average every 5-8 days--implications for HIV remission

    PLOS Pathogens 11:e1005000.

    https://doi.org/10.1371/journal.ppat.1005000

    • PubMed
    • Google Scholar
41. 1. Procopio FA
    2. Fromentin R
    3. Kulpa DA
    4. Brehm JH
    5. Bebin AG
    6. Strain MC
    7. Richman DD
    8. O'Doherty U
    9. Palmer S
    10. Hecht FM
    11. Hoh R
    12. Barnard RJ
    13. Miller MD
    14. Hazuda DJ
    15. Deeks SG
    16. Sékaly RP
    17. Chomont N

    (2015) A novel assay to measure the magnitude of the inducible viral reservoir in HIV-infected individuals

    EBioMedicine 2:874–883.

    https://doi.org/10.1016/j.ebiom.2015.06.019

    • PubMed
    • Google Scholar
42. 1. Rasmussen TA
    2. Tolstrup M
    3. Brinkmann CR
    4. Olesen R
    5. Erikstrup C
    6. Solomon A
    7. Winckelmann A
    8. Palmer S
    9. Dinarello C
    10. Buzon M
    11. Lichterfeld M
    12. Lewin SR
    13. Østergaard L
    14. Søgaard OS

    (2014) Panobinostat, a histone deacetylase inhibitor, for latent-virus reactivation in HIV-infected patients on suppressive antiretroviral therapy: a phase 1/2, single group, clinical trial

    The Lancet HIV 1:e13–e21.

    https://doi.org/10.1016/S2352-3018(14)70014-1

    • PubMed
    • Google Scholar
43. 1. Reynolds MR
    2. Sacha JB
    3. Weiler AM
    4. Borchardt GJ
    5. Glidden CE
    6. Sheppard NC
    7. Norante FA
    8. Castrovinci PA
    9. Harris JJ
    10. Robertson HT
    11. Friedrich TC
    12. McDermott AB
    13. Wilson NA
    14. Allison DB
    15. Koff WC
    16. Johnson WE
    17. Watkins DI

    (2011) The TRIM5{alpha} genotype of rhesus macaques affects acquisition of simian immunodeficiency virus SIVsmE660 infection after repeated limiting-dose intrarectal challenge

    Journal of Virology 85:9637–9640.

    https://doi.org/10.1128/JVI.05074-11

    • PubMed
    • Google Scholar
44. 1. Riddick NE
    2. Wu F
    3. Matsuda K
    4. Whitted S
    5. Ourmanov I
    6. Goldstein S
    7. Goeken RM
    8. Plishka RJ
    9. Buckler-White A
    10. Brenchley JM
    11. Hirsch VM

    (2016) Simian immunodeficiency virus SIVagm efficiently utilizes Non-CCR5 entry pathways in african green monkey lymphocytes: potential role for GPR15 and CXCR6 as viral coreceptors

    Journal of Virology 90:2316–2331.

    https://doi.org/10.1128/JVI.02529-15

    • Google Scholar
45. 1. Sharaf RR
    2. Li JZ

    (2017) The alphabet soup of HIV reservoir markers

    Current HIV/AIDS Reports 14:72–81.

    https://doi.org/10.1007/s11904-017-0355-y

    • PubMed
    • Google Scholar
46. 1. Siliciano JD
    2. Kajdas J
    3. Finzi D
    4. Quinn TC
    5. Chadwick K
    6. Margolick JB
    7. Kovacs C
    8. Gange SJ
    9. Siliciano RF

    (2003) Long-term follow-up studies confirm the stability of the latent reservoir for HIV-1 in resting CD4+ T cells

    Nature Medicine 9:727–728.

    https://doi.org/10.1038/nm880

    • PubMed
    • Google Scholar
47. 1. Søgaard OS
    2. Graversen ME
    3. Leth S
    4. Olesen R
    5. Brinkmann CR
    6. Nissen SK
    7. Kjaer AS
    8. Schleimann MH
    9. Denton PW
    10. Hey-Cunningham WJ
    11. Koelsch KK
    12. Pantaleo G
    13. Krogsgaard K
    14. Sommerfelt M
    15. Fromentin R
    16. Chomont N
    17. Rasmussen TA
    18. Østergaard L
    19. Tolstrup M

    (2015) The depsipeptide romidepsin reverses HIV-1 latency in vivo

    PLOS Pathogens 11:e1005142.

    https://doi.org/10.1371/journal.ppat.1005142

    • PubMed
    • Google Scholar
48. 1. Steingrover R
    2. Pogány K
    3. Fernandez Garcia E
    4. Jurriaans S
    5. Brinkman K
    6. Schuitemaker H
    7. Miedema F
    8. Lange JM
    9. Prins JM

    (2008) HIV-1 viral rebound dynamics after a single treatment interruption depends on time of initiation of highly active antiretroviral therapy

    Aids 22:1583–1588.

    https://doi.org/10.1097/QAD.0b013e328305bd77

    • PubMed
    • Google Scholar
49. 1. Wen Y
    2. Bar KJ
    3. Li JZ

    (2018) Lessons learned from HIV antiretroviral treatment interruption trials

    Current Opinion in HIV and AIDS 13:416–421.

    https://doi.org/10.1097/COH.0000000000000484

    • PubMed
    • Google Scholar
50. 1. Whitney JB
    2. Hill AL
    3. Sanisetty S
    4. Penaloza-MacMaster P
    5. Liu J
    6. Shetty M
    7. Parenteau L
    8. Cabral C
    9. Shields J
    10. Blackmore S
    11. Smith JY
    12. Brinkman AL
    13. Peter LE
    14. Mathew SI
    15. Smith KM
    16. Borducchi EN
    17. Rosenbloom DI
    18. Lewis MG
    19. Hattersley J
    20. Li B
    21. Hesselgesser J
    22. Geleziunas R
    23. Robb ML
    24. Kim JH
    25. Michael NL
    26. Barouch DH

    (2014) Rapid seeding of the viral reservoir prior to SIV viraemia in rhesus monkeys

    Nature 512:74–77.

    https://doi.org/10.1038/nature13594

    • PubMed
    • Google Scholar
51. 1. Whitney JB
    2. Lim SY
    3. Osuna CE
    4. Kublin JL
    5. Chen E
    6. Yoon G
    7. Liu PT
    8. Abbink P
    9. Borducci EN
    10. Hill A
    11. Lewis MG
    12. Geleziunas R
    13. Robb ML
    14. Michael NL
    15. Barouch DH

    (2018) Prevention of SIVmac251 reservoir seeding in rhesus monkeys by early antiretroviral therapy

    Nature Communications 9:5429.

    https://doi.org/10.1038/s41467-018-07881-9

    • PubMed
    • Google Scholar
52. 1. Williams JP
    2. Hurst J
    3. Stöhr W
    4. Robinson N
    5. Brown H
    6. Fisher M
    7. Kinloch S
    8. Cooper D
    9. Schechter M
    10. Tambussi G
    11. Fidler S
    12. Carrington M
    13. Babiker A
    14. Weber J
    15. Koelsch KK
    16. Kelleher AD
    17. Phillips RE
    18. Frater J
    19. SPARTACTrial Investigators

    (2014) HIV-1 DNA predicts disease progression and post-treatment virological control

    eLife 3:e03821.

    https://doi.org/10.7554/eLife.03821

    • PubMed
    • Google Scholar
53. 1. Yukl SA
    2. Kaiser P
    3. Kim P
    4. Telwatte S
    5. Joshi SK
    6. Vu M
    7. Lampiris H
    8. Wong JK

    (2018) HIV latency in isolated patient CD4⁺ T cells may be due to blocks in HIV transcriptional elongation, completion, and splicing

    Science Translational Medicine 10:eaap9927.

    https://doi.org/10.1126/scitranslmed.aap9927

    • PubMed
    • Google Scholar

Acceptance summary:

HIV eradication is the ultimate goal of AIDS research and some strategies are evaluated in clinical trials. One caveat of these studies is that currently no validated biomarkers that accurately predict the efficacy of these strategies in prolonging virus-free remission after treatment interruption are available. The present study addresses this important issue by evaluating various immune and virologic biomarkers as potential correlates of time to viral rebound during treatment interruption using barcoded SIVmac239 constructs in the macaque model. The study shows that very early initiation of treatment (4 dpi) reduces the frequency of reactivation, while initiation at 10 or 27 dpi hardly made a difference. This very early treatment strongly reduced the levels of proviral DNA; surprisingly, however, a 100-fold difference in the proviral copy numbers in blood PBMCs was only associated with a 2-fold difference in reactivation frequencies from latency. Thus, the data indicate that peripheral blood measures of the reservoir are insufficient to predict reactivation from latency. Although the results are largely negative they are clearly of significant interest for everybody interested in the cure of HIV/AIDS.

Decision letter after peer review:

Thank you for submitting your article "SIV reactivation from latency is established early and not predicted by standard measures of reservoir" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, including Frank Kirchhoff as the Reviewing Editor and Reviewer #1, and the evaluation has been overseen by a Reviewing Editor and Neil Ferguson as the Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Adam Mitchell Spivak (Reviewer #2).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

Although HIV eradication is the ultimate goal of AIDS research and some strategies are currently evaluated in clinical trials, there are no validated biomarkers that accurately predict the efficacy of these strategies in prolonging virus-free remission after treatment interruption. Davenport and colleagues address this important issue by evaluating various immune and virologic biomarkers as potential correlates of time to viral rebound during treatment interruption using barcoded SIVmac239 constructs in the macaque model. Using a previously established mathematical model, the authors estimate the number of reactivation events in each animal after ART cessation that contributed to rebound viremia. These rebound rates were compared to measures of SIV DNA and SIV CA-RNA as well as CD8 responses to in vitro SIV peptide pulsing experiments. The authors show that very early treatment initiation (4 dpi) reduces the frequency of reactivation, while initiation at 10 or 27 dpi hardly made a difference. This very early treatment reduced the levels of proviral DNA by about two orders of magnitude but was only associated with a 2-fold difference in reactivation frequencies from latency. Altogether, no stringent relationship between the modelled SIV reactivation rate and any of the virologic and immunologic parameters investigated could be established. Thus, the authors conclude that peripheral blood measures of the reservoir are insufficient to predict reactivation from latency.

All reviewers agree that the study is well performed and addresses a very important and timely question. Limitations are that findings are largely negative and that (as discussed by the authors) no viral outgrowth assays could be performed since available cell numbers are limited in the SIV/macaque model. While the results are clearly of significant interest several concerns should be addressed.

Essential revisions:

1) The manuscript focuses almost exclusively on the measurement of total proviral DNA, and does not directly take into account that in humans this measurement vastly over-estimates the reservoir size due to the accumulation of defective proviruses that functionally may not contribute to the reservoir at all. In the discussion of this work, the authors dismiss this phenomenon by pointing to an upcoming manuscript evaluating this in the SIV model and provide a very high frequency of intact proviruses (80%, section). No data or methods accompany this, and it is markedly different than what is seen in humans (2% of total proviruses are thought to be genetically intact) – which raises some concerns about the nature of this marked difference and the applicability of this model to the human HIV reservoir. This should be addressed in more detail or (better) substantiated by experimental data.

2) The viral inoculum that gives rise to primary infection here is many orders of magnitude larger than what is thought to cause primary infection in humans (at least via sexual transmission), and this may affect the kinetics of reservoir establishment – making this a less-than-ideal model of human reservoirs. How much of the concept of 'early / rapid filling' of the reservoir shown here is an artefact of the non-human primate SIV infection model, in which many thousands of intact virions are used simultaneously for primary infection? This quantity of virus is many logs higher than what is estimated to cause infection in humans in vivo. Studies of the "sexual transmission bottleneck" estimate that a single founder virus or clone establishes infection in humans. Therefore the 'filling' of the reactivation-prone reservoir early in infection in macaques may not be modelling what is occurring in humans. This caveat needs to be discussed.

3) The authors conclude that peripheral blood measures of the viral reservoirs are insufficient to predict reactivation from latency. However, although there is not good quantitative correlation, the predictive value for the combination of treatment duration and DNA levels of interruption seems pretty good (R²=0.88). Thus, the authors should consider mentioning this in the Abstract and revising the Title.

4) It would be interesting to know whether proviral copy numbers per individual cell differ between the 4 dpi and 10 or 27 dpi treatment initiation settings. Although these viruses have evolved mechanisms to prevent super/coinfections these do occur and seem more likely in an environment where many cells are already infected. It would be interesting to know whether the same proviral copy numbers for the 4 dpi treatment corresponds to a larger number of latently infected cells than the later time points since this could potentially help to explain some findings.

5) Subsection “Measuring the frequency of SIV reactivation from latency”: what evidence exists to support the hypothesis that all proviruses contribute equally to HIV reactivation? The opposite hypothesis, that most proviruses have little-to-no contribution to viral recrudescence and a small minority are capable of reactivation and viral production, has much evidence to support it, including heterogeneity of the latent reservoir with respect to proviral integration site, accumulation of defective proviruses containing hypermutations / insertions / deletions, epigenetic phenomena contributing to transcription threshold, unique half-lives of T cell subsets and potential differences arising from local tissue microenvironments (tissue vs. peripheral blood) all strongly suggest that proviruses cannot be assumed to all possess the same potential for proviral reactivation.

6) Subsection “Measuring the frequency of SIV reactivation from latency”: doesn't this suggest an accumulation of defective proviruses over time, that increases the longer viral replication is allowed to proceed? The initial viral infection would be with a large population of replication-competent virions. Subsequent replication in vivo will give rise to defective proviruses in accordance with the error-prone nature of SIV RT. It is well known that the reservoir is established early in primary infection. Therefore, if ART is started within days of primary infection, the proportion of intact proviruses will be much higher than if ART is delayed and replication persists. It looks as though this hypothesis is addressed to some degree in the Discussion section however this is not included in this manuscript. The estimate of 80% of proviruses as genetically intact is considerably higher than what is observed in human infection in vivo (40-fold higher, per the Siliciano IPDA Nature paper). Do the authors think this marked difference between the SIV model and human HIV infection with regard to the proportion of intact proviruses may affect their results or make these findings challenging to apply to human infection?

Essential revisions:

1) The manuscript focuses almost exclusively on the measurement of total proviral DNA, and does not directly take into account that in humans this measurement vastly over-estimates the reservoir size due to the accumulation of defective proviruses that functionally may not contribute to the reservoir at all. In the Discussion section of this work, the authors dismiss this phenomenon by pointing to an upcoming manuscript evaluating this in the SIV model and provide a very high frequency of intact proviruses (80%, section). No data or methods accompany this, and it is markedly different than what is seen in humans (2% of total proviruses are thought to be genetically intact) – which raises some concerns about the nature of this marked difference and the applicability of this model to the human HIV reservoir. This should be addressed in more detail or (better) substantiated by experimental data.

We thank the reviewers for this suggestion. The issue of replication competence is clearly one that needs better explanation, as it is touched upon by several of the comments (including Essential revision 6). We have now added a new methods section, and a new Results section and figure (new Figure 5) in order to explain the approach in more detail. We note that this data is reproduced from a publication that is now in press (Long et al., in press), as is explained in the manuscript:

Subsection “Single genome amplification and sequencing of full-length SIV genome”:

“Genomic DNA was purified from PBMC from SIV-infected rhesus macaques using QIAamp DNA Mini Kit (Qiagen). […] SGA PCR products were directly sequenced using BigDye Terminator Sanger Sequencing (Life Technologies) with the primers previously described (Lopker et al., 2016).”

Discussion section:

“However, near full length viral sequencing from animals treated on day 10 and day 27 revealed that the vast majority (>80%) of sequences were intact. […] Taken together, it is clear that the observed differences in reactivation frequency between animals treated on different days cannot be explained by different proportions of replication competent virus.”

In addition, we now address the differences in replication competence in HIV (point 6 below), citing recent work on the proportion of intact viral DNA in HIV and SIV.

2) The viral inoculum that gives rise to primary infection here is many orders of magnitude larger than what is thought to cause primary infection in humans (at least via sexual transmission), and this may affect the kinetics of reservoir establishment – making this a less-than-ideal model of human reservoirs. How much of the concept of 'early / rapid filling' of the reservoir shown here is an artefact of the non-human primate SIV infection model, in which many thousands of intact virions are used simultaneously for primary infection? This quantity of virus is many logs higher than what is estimated to cause infection in humans in vivo. Studies of the "sexual transmission bottleneck" estimate that a single founder virus or clone establishes infection in humans. Therefore the 'filling' of the reactivation-prone reservoir early in infection in macaques may not be modelling what is occurring in humans. This caveat needs to be discussed.

We thank the reviewers for suggesting further discussion of this point, as it is one that we feel is sometimes misunderstood. The monkeys are infected with a large dose of virus, and by a different route (intravenous), compared to sexual transmission of HIV. Therefore, it is worth asking ‘does this early virus drive earlier reservoir formation’? Considering the animals treated on day 4, for example, how much of the virus present at ART initiation might be the result of the initial high inoculum? The viral load increases around 10 fold per day in early infection (growth rate of the virus is around 2 day^-1). Thus, broadly speaking we expect the viral load on day 4 to be around 10,000 fold higher than the viral load at infection. Even if we are off by several orders of magnitude, it is still clear that the initial inoculum does not play a big role in the total population of virus treated on day 4.

Another way to frame this is to ask, ‘do we have evidence for early establishment of the reservoir in HIV?’ We do not think that the exact timing of reservoir establishment in high-dose infected macaques will be precisely recapitulated in humans, but the idea of establishing the rebound competent reservoir early in the infection process is valid and has already been demonstrated in a limited cohort of infected humans. We would argue that the studies of Colby and colleagues (Colby et al., 2018) indicate that not only is the reservoir established extremely early in HIV, but that – as we find in SIV – the frequency of reactivation is almost as high as would be observed following treatment in chronic infection.

We think this is an important point to clarify, and have added a paragraph to the Discussion section:

“The initial inoculum of virus in our studies is several orders of magnitude higher than that estimated in HIV (Keele et al., 2008), and thus this may disproportionately contribute to the early establishment and potential saturation of the SIV RCR. However, since viral loads increase around 10-fold per day in early infection (Figure 1A,C,E), it is unlikely that this initial inoculum virus equates to >0.1% of virus present even at day 4 post-infection (the timing of earliest treatment). Thus, it seems unlikely that the higher inoculum in SIV infection could be a major contributor to the SIV reservoir measured later during treatment.”

3) The authors conclude that peripheral blood measures of the viral reservoirs are insufficient to predict reactivation from latency. However, although there is not good quantitative correlation, the predictive value for the combination of treatment duration and DNA levels of interruption seems pretty good (R²=0.88). Thus, the authors should consider mentioning this in the Abstract and revising the Title.

We thank the reviewers for pointing this out and have altered the Abstract and Title to reflect this. This is a good point, and we have changed the Title to “Predictors of SIV recrudescence following antiretroviral treatment interruption”.

4) It would be interesting to know whether proviral copy numbers per individual cell differ between the 4 dpi and 10 or 27 dpi treatment initiation settings. Although these viruses have evolved mechanisms to prevent super/coinfections these do occur and seem more likely in an environment where many cells are already infected. It would be interesting to know whether the same proviral copy numbers for the 4 dpi treatment corresponds to a larger number of latently infected cells than the later time points since this could potentially help to explain some findings.

This is an interesting question, which could potentially play some role in the relationship between proviral copy number and reactivation frequency. Unfortunately we do not have experimental data on the rate of coinfection in these animals, and to our knowledge this has not been studied in SIV (although studies in HIV suggest the rate is relatively low (Josejsson et al., 2011)).

Interestingly, it is not entirely clear that coinfection would change the relationship between DNA copies and frequency of reactivation in our system. That is, we detect the number and size ratio of different barcode sizes. If a cell harbours two different barcodes and both ‘wake up’ independently, then we would see and count the second barcode (unless the coinfection was with identical barcodes). Moreover, it is not clear whether the limitation on viral reactivation is ‘infected cells’ (and the probability of a cell waking up), or ‘DNA copies’ (and the probability of a DNA copy being integrated in a site where it might reactivate). Coinfection would largely be an issue in the former case, and much less so in the latter.

This seems a little too complex to go into in our study, however we have added a line in the Discussion section to state:

“Similarly, although coinfection of cells with multiple copies of HIV is thought to be low (Josejsson et al., 2011), rates of SIV coinfection of cells were not determined in our study and could play a role in determining the frequency of reactivation per DNA copy observed.”

5) Subsection “Measuring the frequency of SIV reactivation from latency”: what evidence exists to support the hypothesis that all proviruses contribute equally to HIV reactivation? The opposite hypothesis, that most proviruses have little-to-no contribution to viral recrudescence and a small minority are capable of reactivation and viral production, has much evidence to support it, including heterogeneity of the latent reservoir with respect to proviral integration site, accumulation of defective proviruses containing hypermutations / insertions / deletions, epigenetic phenomena contributing to transcription threshold, unique half-lives of T cell subsets and potential differences arising from local tissue microenvironments (tissue vs. peripheral blood) all strongly suggest that proviruses cannot be assumed to all possess the same potential for proviral reactivation.

We agree this was poorly worded and did not convey our intent. We intended to discuss whether the DNA measured in different animals contributed equally to reactivation – not to indicate that all DNA copies contributed equally. We have tried to reword this section to discuss whether DNA on average made a similar contribution (subsection “SIV DNA in PBMC and frequency of reactivation”).

“Although individual SIV genomes may vary greatly in their probability of reactivation (for example due to replication competence, integration site, or cell phenotype, activation state, and epigenetic or transcriptional blockades), if SIV DNA measured in the animals treated on different days had on average a similar probability of contributing to SIV reactivation from latency, then a doubling of DNA would equate to a doubling of reactivation (and we should expect a 1:1 correlation in this log:log plot).”

6) Subsection “Measuring the frequency of SIV reactivation from latency”: doesn't this suggest an accumulation of defective proviruses over time, that increases the longer viral replication is allowed to proceed? The initial viral infection would be with a large population of replication-competent virions. Subsequent replication in vivo will give rise to defective proviruses in accordance with the error-prone nature of SIV RT. It is well known that the reservoir is established early in primary infection. Therefore, if ART is started within days of primary infection, the proportion of intact proviruses will be much higher than if ART is delayed and replication persists. It looks as though this hypothesis is addressed to some degree in the Discussion section however this is not included in this manuscript. The estimate of 80% of proviruses as genetically intact is considerably higher than what is observed in human infection in vivo (40-fold higher, per the Siliciano IPDA Nature paper). Do the authors think this marked difference between the SIV model and human HIV infection with regard to the proportion of intact proviruses may affect their results or make these findings challenging to apply to human infection?

We agree that the issue of replication competence needed further explanation, and have added a section on our sequencing method, as well as expanding our discussion of replication competence in both SIV and HIV (the changes are detailed in the response to Q1 above).

Author details

1. Mykola Pinkevych

   Infection Analytics Program, Kirby Institute for Infection and Immunity, UNSW Australia, Sydney, Australia

   Contribution

   Resources, Data curation, Software, Formal analysis, Validation, Investigation, Visualization, Writing—original draft, Writing—review and editing

   Competing interests

   No competing interests declared

2. Christine M Fennessey

   AIDS and Cancer Virus Program, Frederick National Laboratory for Cancer Research, Frederick, United States

   Contribution

   Data curation, Validation, Investigation, Writing—original draft, Writing—review and editing

   Competing interests

   No competing interests declared

3. Deborah Cromer

   Infection Analytics Program, Kirby Institute for Infection and Immunity, UNSW Australia, Sydney, Australia

   Contribution

   Software, Formal analysis, Investigation, Visualization, Writing—original draft, Writing—review and editing

   Competing interests

   No competing interests declared

4. Carolyn Reid

   AIDS and Cancer Virus Program, Frederick National Laboratory for Cancer Research, Frederick, United States

   Contribution

   Data curation, Validation, Investigation, Writing—original draft, Writing—review and editing

   Competing interests

   No competing interests declared

5. Charles M Trubey

   AIDS and Cancer Virus Program, Frederick National Laboratory for Cancer Research, Frederick, United States

   Contribution

   Data curation, Validation, Investigation, Writing—original draft, Writing—review and editing

   Competing interests

   No competing interests declared

6. Jeffrey D Lifson

   AIDS and Cancer Virus Program, Frederick National Laboratory for Cancer Research, Frederick, United States

   Contribution

   Conceptualization, Supervision, Methodology, Writing—original draft, Project administration, Writing—review and editing

   Competing interests

   No competing interests declared

7. Brandon F Keele

   AIDS and Cancer Virus Program, Frederick National Laboratory for Cancer Research, Frederick, United States

   Contribution

   Conceptualization, Supervision, Funding acquisition, Validation, Investigation, Methodology, Writing—original draft, Project administration, Writing—review and editing

   For correspondence

   keelebf@mail.nih.gov

   Competing interests

   No competing interests declared

8. Miles P Davenport

   Infection Analytics Program, Kirby Institute for Infection and Immunity, UNSW Australia, Sydney, Australia

   Contribution

   Conceptualization, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Methodology, Writing—original draft, Project administration, Writing—review and editing

   For correspondence

   M.Davenport@unsw.edu.au

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-4751-1831

• 952

  Page views

• 152

  Downloads

• 12

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.