Early treatment regimens achieve sustained virologic remission in infant macaques infected with SIV at birth

Early antiretroviral therapy (ART) in HIV-infected infants generally fails to achieve a sustained state of ART-free virologic remission, even after years of treatment. Our studies show that viral reservoir seeding is different in neonatal macaques intravenously exposed to SIV at birth, in contrast to adults. Furthermore, one month of ART including an integrase inhibitor, initiated at day 3, but not day 4 or 5 post infection, efficiently and rapidly suppresses viremia to undetectable levels. Intervention initiated at day 3 post infection and continued for 9 months achieves a sustained virologic remission in 4 of 5 infants. Collectively, an early intervention strategy within a key timeframe and regimen may result in viral remission or successful post-exposure prophylaxis for neonatal SIV infection, which may be clinically relevant for optimizing treatment strategies for HIV-infected or exposed infants.

: The plasma viral load data in Figure 1B is not complete (only shows 6 of 14 animals), and interpretation is limited. It would be helpful to know the plasma viral load of the animals that were sacrificed, and that is not easily discerned from data shown on days 1 and 3.
Plasma VL at day 1 is shown for 6 animals, 3 positive and 3 negative. Which of these are from the animals euthanized and correspond to the data shown from PBMC? What subsequent data is available for the 3 negative? Do the viral loads at day 3 represent the 8 remaining animals? It is difficult to count the number of symbols shown for Day 3; are 8 remaining animals represented in 1B? How do specificplasma VL compare to those in tissues? Is there a threshold of plasma virus levels that correlate with detection of total DNA and/or integrated DNA (suggesting sensitivity of detection methods?) Why do the y axis for the integrated DNA in panels I-M start at 10-2?
The legend/text state that 9 control animals were used for comparison, yet only 3 are represented in panels E, F, G. How many are represented in the various time points of panel D, as error bars are only shown for the day 9 time point?
Reviewer #2 (Remarks to the Author): In this manuscript Wang et al. assessed the impact of very early antiretroviral initiation on the viral reservoir in perinatally SIV-infected rhesus macaque infants. Specifically, the study investigated SIV DNA and RNA levels in RM infants infected i.v. with SIVmac251 within 6 hours after birth and treated or not with ART initiated between day 3 and day 5. The impact of early ART on the viral reservoir was estimated by measuring total or integrated SIV DNA in cells isolated from blood, lymph nodes and gut and/or by analytical treatment interruption. The main result of this study is the long-term control of SIV replication following treatment interruption in 4 out of 5 rhesus macaque infants initiated on ART 3 days post infection and maintained on ART for 9 months. This absence of viral rebound was observed for up to 18 months even following experimental CD8+ T cell depletion.
The manuscript needs serious English editing. It also suffers from the choice of graphical representation-at best the graphs are difficult to read but they also can be misleading. The assays used to evaluate reservoir seeding are not optimal and the methods are insufficiently described. The data are not appropriately described and interpreted. The significance of the results is questionable notably because they are not very novel. It is well documented that early ART restricts the latent reservoir establishment in adults and infants. Delayed viral rebound or sustained virologic control following ART interruption has been reported in children treated early post infection. The Mississipi baby treated 31 hours after birth maintained undetectable viremia for 27 months in absence of ART. Preliminary results from the IMPAACT P1115 study showed that two years after starting ART within 48 hours of birth, two thirds of infants had nondetectable cell-associated HIV DNA and almost 90% tested negative for HIV antibody. ATI will be performed in eligible study participants.
Please see specific comments and detailed review below: 1-To interrogate SIV dynamics or the viral reservoir seeding, the accuracy and sensitivity of the assays used is critical. In this manuscript, some key information regarding the methods is missing such as the limit of detection of each assay and the number of cells assessed (known limiting factor in pediatric studies). The type of cells assessed is also unclear to this reviewer (PBMC and lymphocytes/CD3+ isolated from tissues or tissue mononuclear cells?). If I understand correctly, dilutions of a plasmid in DNA from uninfected cells was used as a standard for the integrated SIV DNA quantification which bypasses the Alu-gag PCR and is thus not appropriate. The manuscript reports unintegrated SIV DNA which has not been measured but extrapolated from the measurement of total and integrated SIV DNA (total SIV DNA in that case is not shown). (Figure 1). The number of animals euthanized or initiated on ART for each time point is not indicated on the study design schematic nor in its legend. The LOD of the assays is missing for total SIV RNA and integrated DNA. It is also unclear if PVL was assessed for all living animals at each time point. Using individual symbols for each animal would greatly improve the figure. I count 9 symbols of the PVL at day 3 but according to the study design, only 8 animals were still alive by then.

2-Untreated SIV infection
3-Early short-term ART ( Figure 2): It is unclear what comparisons are the p values representing although the legend states that comparison were made with d3 p.i. Surprisingly, SIV RNA levels appear to be higher in the group of animals receiving ART at day 5 p.i. than in the untreated control group. Were those all assessed at the same time point (day 28 p.i.)? The results indicate that in untreated infection, viremia plateaud at 7 day p.i. but the graph shows a peak at day 9 p.i. 4-Early "long-term" ART and ATI ( Figure 3): A control group initiated on ART later and going through ATI is missing. In panel A, it is difficult to distinguish the PVL of each animal at day 3 p.i. One RM out of 5 experienced a viral rebound following ATI. The rebound was detected 2 months post ATI. Was the PVL assessed during these 2 months? The absence of symbol in panel A does not allow the reader to get that information. The antibody used to assess CD8 expression following experimental depletion is not indicated.
One restatement of the conclusions of this work that could be drawn from this observation is that the authors have not induced HIV remissions, but simply observed successful post-exposure prophylaxis, and that in the setting of new infection of the infant immune system, that potent ART that includes an INSTI within 3 days, but not later, of viral exposure is sufficient to block all persistent infection in 80% of exposed infant animals. There may have been viral integration events that led to productive infection before day 3, but few enough so that no latent, persistent infections were established that could lead to viral rebound upon later ATI.
I view this work as suggestive, but preliminary. A larger cohort should be studied. Longer ATI and more tissue sampling should be pursued. Cell infection transfer experiments, to test if infection can be passed by cells from one animal to another should be done. These points should be discussed by the authors.
However, the findings of this work are potentially important and should be disseminated to stimulate the field.
Current US guidelines (Dec. 2021) state that: A newborn's ARV regimen should be determined based on maternal and infant factors that influence the risk of perinatal transmission of HIV (AII). The uses of ARV regimens in newborns include the following: • ARV Prophylaxis: The administration of one or more ARV drugs to a newborn without documented HIV infection to reduce the risk of perinatal acquisition of HIV.
• Presumptive HIV Therapy: The administration of a three-drug ARV regimen to newborns who are at highest risk of perinatal acquisition of HIV. Presumptive HIV therapy is intended to be preliminary treatment for a newborn who is later documented to have HIV, but it also serves as prophylaxis against HIV acquisition for those newborns who are exposed to HIV in utero, during the birthing process, or during breastfeeding and who do not acquire HIV.
• HIV Therapy: The administration of a three-drug ARV regimen at treatment doses (called antiretroviral therapy [ART]) to newborns with documented HIV infection (see Diagnosis of HIV Infection in Infants and Children).
These findings suggest that such presumptive HIV therapy or therapy given for HIV infection documented at birth should include an INSTI if possible. Further support for the findings in this manuscript could lead to the study of ATIs in newborns treated with INSTIs.
Such studies are underway but it is unclear if they are sufficiently enrolled or powered to answer the question anytim soon. Therefore, even these preliminary findings are of interest.

Minor points:
Line 59: Protease inhibitors for "proteinase inhibitors" Line 59-60: "prevent cells from viral genome integration" is awkward. Change to "prevent viral genome integration in cells" ?

Response:
We appreciate the reviewer's concerns. To address the concerns regarding the early virion production from unintegrated viral DNA and the sensitivity of the qPCR assay, we performed additional in vitro experiments, in combination with viral parameters measured by nested (Alu) qPCR. Briefly, PBMCs isolated from one-month-old infant macaques (due to limited blood collected from newborns or neonates even at animal necropsy) were first activated for 3 days by CD4+ T cell activation kit (Miltenyi) in the presence of human IL-2 (10%, Hemagen), and then infected with SIVmac251 (3×10^7 cell/mL with 100 TCID50 SIV) for 2 hours in CO2 incubator, followed by 3 washes and seeding in 24-well plate at 1×10^6 cell/well/mL, without or with raltegravir (RTG) at final 1µM concentration. Cells and cell supernatants were harvested to measure CA SIV RNA/DNA and supernatant viral load. Please see the revised Method section in the revised manuscript.

Tulane University
As shown in Suppl Fig.1 We would prefer to add even more detail in Methods but have space limitations. However, these assays are based on prior  SIV DNA and standard) or SIV U5/Alu 1-4 fragments (for proviral DNA) for 20 cycles, followed by second-step qPCR using same U5 primer/probe set for 40 cycles with 1/10 volume of 1 st amplicon. Internal standard curve was used to calculate the initial copies of both total SIV DNA and proviral DNA (an example of STD cure attached here, very stable from batch to batch).
Using this assay, we have measured hundreds of samples from SIV-infected animal cohorts (Ziani W, 2021 a,b,c ; Wang X, 2021). We performed nested qPCR in parallel to quantify both total viral DNA and proviral DNA in each sample (yielding comparable data), along with standard. Thus, nested qPCR assay can generate sensitive and comparable data of both total SIV DNA and proviral DNA. Also, the results shown in Supple Fig. 1 suggest that Alu qPCR is able to essentially reflect the sensitivity and status of SIV genome integration in RTG-or untreated-PBMCs, thus inferring lymphocyte samples isolated from tissues. Basically, nested qPCR (two-rounds PCR amplification) can increase sensitivity and specificity of target gene, however, we admit that the possibility that proviral DNA may not be fully detected due to viral reservoir distribution in various tissues in the body and likely lack of chromosomal Alu islands situated near the integrated SIV genome, we discussed this in the revised DISCUSSION. Alu islands are widely dispersed within the chromosomal DNA (accounting for 6-13% of genomic DNA), which are likely lacked or far from SIV LTR region, albeit very few such case existed. qPCR targeting SIV gag is used to determine plasma viral load with a limit of detection of 83 copies per mL (core service). Together, early treatment, building on viral genome integration assessed by our assay, achieved sustained virologic remission in the neonates infected with SIV, which were largely consistent with the onset of proviral reservoirs in typical systemic and lymphoid tissues of very early SIV infection. We realize that more animal cohorts/sampling are needed, in combination with intrauterine/postnatal infection, one or more ARV regimens as well as optimal period of treatment and so on. We have carefully considered the interpretations of our results and added more information and detail in Method section, please see in our revised manuscript.

Despite the limitations regarding their interpretation of integrated provirus, the paper does detail novel and important
data regarding the efficacy of early ART in a pediatric population.

Response:
We thank reviewer's positive comments, we have carefully read through the text and made appropriate interpretations and caveats to our results.
4. Revise all discussion regarding the evidence for lack of reservoir establishment, with the inclusion of assay sensitivity and interpretation of results reported with these limitations.

Response:
We thank reviewer's suggestion, we discussed proviral reservoir establishment in neonates, especially assay sensitivity and interpretation of proviral reservoir measurement in vitro and in vivo studies, please see DISCUSSION in our revised manuscript. Figure 1B is not complete (only shows 6 of 14 animals), and interpretation is limited. It would be helpful to know the plasma viral load of the animals that were sacrificed, and that is not easily discerned from data shown on days 1 and 3.

Response:
We apologize for mot making it clearer. Whole blood request and neonatal anesthesia is strictly controlled in terms of IACUC (12mL/Kg/Month in total) and limited volume of experimental blood available because of newborns weight (~ 400g, only 1.2cc EDTA blood available in general), physiologic weight loss after birth as well as mandatory bleeding for CBC. It is not feasible to consistently collect 1.2cc whole blood from newborn infants and neonates in the first few days of birth and larger volumes are limited to collection at necropsy. Therefore, whole blood is collected by staggered request from individual newborn/neonates. We added the grouped animals, the day(s) post SIV inoculation, plasma viral load from blood sampling as well as text in revised manuscript as Suppl. initiated at 3-, 4-or 5dpi, and one group without treatment control (n=3), all these animals are euthanized at 28dpi.
Additional data (not samples) from our previous study with age-matched SIV-infected infant macaques (n=9) are used to compare routine plasma viral load and the percentages of peripheral CD4+ T cells ( Fig. 2D and Figs. 3A-3B). We went through the data and revised our text and Figure Legend. As for Fig.2D, which is summarized from Figs and sensitivity of the assay especially for integrated viral DNA measurement, with or without raltegravir (RTG) treatment.
As shown in Suppl Fig.1, CA integrated viral DNA was consistently not detected in PBMCs in presence of RTG (viral genome integration blockade), while CA SIV RNA/DNA (unintegrated, due to undetectable integrated vDNA) as well as supernatant viral load were still detectable in the early SIV infection. These data strongly suggest that our assay can essentially reflect the status of SIV genome integration, by comparing the results in the presence or absence of RTG treatment, and thus inferring this application to measure the lymphocyte samples isolated from tissues ex vivo.
In this study, we utilized the neonatal NHP model intravenously infected with SIV, facilitating to precisely time the very early days of infection (compared with uncertain dpi by repeated SIV/SHIV challenges to the infant macaques by oral route or HIV+ children in the clinic) and collected complete tissue sets at necropsy for examination of viral reservoirs. As shown in Fig.2  In this study, we evaluated the dynamics and tissue distribution of viral reservoirs and plasma viral load in several newborn macaques at very early timed-stages of SIV infection, which provided critical clues for early treatment interventions, as indicated by outcomes of viral remission achieved by early treatment for 9 months and further confirmed by subsequent CD8+ cell depletion in vivo. We consider this a de facto success to achieve viral remission in 4/5 infant macaques by early treatment (all 5 infants infected with identical SIV inoculum at birth and treated at 3dpi for 9 months). These 4 cured infants are still alive with ART-free viral remission, and currently resisting repeated SIV challenges by the rectal route. Considering sensitivity, rapidity, and feasibility of qPCR, we spent several years to develop and optimize nested Alu qPCR for quantification of the CA integrated SIV DNA, based on previous published reports using conserved human/rhesus Alu-1 and -2 (please see refs by Nishimura Y, 2009;Mavigner M, 2016;Whitney JB, 2014 and2018, etc). To strengthen proviral SIV DNA assay, rhesus specific repetitive Alu DNA, which are widely dispersed within the chromosomal DNA in rhesus macaques, is targeted by additional Alu-3 and -4 primers. Our optimized Alu qPCR thus includes Alu 1-4 outward primers to target multiple Alu islands in rhesus species. As for the accuracy and sensitivity of the assays concerned, please see Answer#2.
It is true that we use plasmids as standard containing SIV gene as standard with known copies, which is indispensable to perform nested qPCR (unlike the RNA quantification by Droplet digital PCR, STD not required standard. Outward Alu 1-4 primers used in qPCR indeed increase the sensitivity and detection rate of proviral SIV DNA in rhesus samples. To yield comparable data in the quantification of total SIV DNA and proviral DNA, nested qPCR is performed in parallel to quantify both total SIV DNA and proviral DNA for each sample. Type of cells are lymphocytes that are freshly isolated from systemic and lymphoid tissues at biopsies or necropsies, 10^7 lymphocytes per sample are generally collected for nested qPCR assay. Briefly, we performed nested qPCR to quantify total viral DNA, proviral DNA, along with standard. First-step routine PCR is used to pre-amplify total SIV DNA and standard (by LTR U5 primer pair) or integrated SIV DNA (by U5/Alu 1-4 primer pairs) for 20 cycles (optimized cycle for the 1 st PCR in our assay). Second qPCR is further performed with 1/10 volume of 1 st amplicon targeting SIV LTR U5 with same primer/probe set for 40 cycles. Internal standard curve is used to calculate the initial copies of both total and integrated SIV DNA. Please note that it is essentially difficult to quantify gene fragments of SIV/Alu by qPCR in the first round PCR due to unpredictable size of gene fragments generated (tens of bp to several kb) and thereby amplification efficiency is variable from sample to sample.
Thus, it is feasible and reasonable to use plasmid standard in the quantification of both CA total SIV DNA and proviral DNA at 2nd qPCR using same primer/probe set, irrelevant to Alu repetitive sequences. We agree with reviewer's comments about unintegrated SIV DNA, now we changed wording to indicate total SIV DNA in the revised figures.
Although our assay is far more sensitive to measure proviral DNA, nested qPCR increases sensitivity and specificity of gene targes, as also indicated by Ct (e.g., cycle threshold of 10copies, from ~36 at 1 st to ~16 at 2 nd PCR). However, status of viral genome integration by our assay essentially matches the actual scenario: undetectable proviral DNA in PBMCs infected with SIV in presence of RTG, and achievable ART-free viral remission in infants when early cART is initiated prior to proviral reservoir seeding. We have thoroughly modified our text as suggested in the revised manuscript. PVL is measured in plasma from euthanized animals or staggered blood samples. The LOD of assay is added in the revised manuscript. Also please see the Answer #5 and #6 above in this letter including animals used and corresponding PVL at different timepoints (Suppl. Table 1). (Figure 2)

Response:
We thank the reviewer for these comments and for not making comparisons clearer among the groups. We revised figures and legend in the manuscript. We did see levels of CA SIV RNA are higher at 28dpi by late treatment.
Conceivably, cART (FTC/TFV/DTG) suppresses reverse transcription of viral RNA and viral integration, late treatment may result in accumulation of intracellular SIV RNA which are transcribed from both unintegrated and proviral DNA (e.g., before treatment at 5dpi), while CA SIV RNA in untreated control may be relatively exhausted by virus packaging/release, contributing to viral peak and thereafter. All samples examined are freshly collected from euthanized animal cohorts at 28dpi (Figs. 2E-2F). PVL data in age-matched SIV-infected infant controls come from our previous studies. PVL in SIVinfected neonatal macaques is generally monitored at 7, 14 and 21dpi, viremia actually reaches the peak at 7-9 dpi, we have corrected the description.

Early "long-term" ART and ATI (Figure 3): A control group initiated on ART later and going through ATI is missing.
In All 5 infant animals are weekly monitored by PVL to see if there is viral rebound after ATI. Anti-CD8 Ab clone used to assess CD8+ cells is also added in the revised manuscript. with one-month-old infant PBMC infected with SIV in vitro, in presence or absence of integrase inhibitor (raltegravir/RTG), followed by CA SIV RNA/DNA and viral load measurement, which are expected to address the efficiency of RTG on the blockade of viral genome integration, ability of early unintegrated SIV DNA in the virion production as well as sensitivity of qPCR assay, please see the Suppl. Fig.1 in the revised manuscript. Besides typical systemic and mucosal lymphoid tissues presented, other tissues, including spleen, thymus, jejunum, colon and liver, were actually examined in our study, proviral DNA was not detected in these tissues of very early SIV-infected neonates. We agree with reviewer's constructive advice that more studies are needed, including animal numbers, comprehensive tissue sampling, optimal period of early treatment, prolonged ATI, cell transfer in vivo to repeated challenge, etc, please see our discussion in the revised manuscript.

One restatement of the conclusions of this work that could be drawn from this observation is that the authors have not induced HIV remissions, but simply observed successful post-exposure prophylaxis, and that in the setting of new infection of the infant immune system, that potent ART that includes an INSTI within 3 days, but not later, of viral
After CD8+ cell depletion in vivo, the 4 infant animals with ART-free viral remission are still alive and healthy, further study with repeated SIV challenge by rectal route is underway. Finally, even if this is only "post exposure prophylaxis" it is clearly different from adults and remains translationally relevant to clinical pediatric HIV care, thus we consider these findings highly significant.
14. However, the findings of this work are potentially important and should be disseminated to stimulate the field.

Current US guidelines (Dec. 2021) state that: A newborn's ARV regimen should be determined based on maternal and infant factors that influence the risk of perinatal transmission of HIV (AII). The uses of ARV regimens in newborns include the following:
• ARV Prophylaxis: The administration of one or more ARV drugs to a newborn without documented HIV infection to reduce the risk of perinatal acquisition of HIV.
• Presumptive HIV Therapy: The administration of a three-drug ARV regimen to newborns who are at highest risk of perinatal acquisition of HIV. Presumptive HIV therapy is intended to be preliminary treatment for a newborn who is later documented to have HIV, but it also serves as prophylaxis against HIV acquisition for those newborns who are exposed to HIV in utero, during the birthing process, or during breastfeeding and who do not acquire HIV. macaques of one-month-old and pregnant animals. All these animals will be utilized to investigate the ARV regimens (one or more drugs), fetal and maternal factors, intrauterine infection, INSTI-based prior or after viral integration, optimal timepoint and period of early treatment and outcomes (immune development and viral rebound) etc.

• HIV Therapy: The administration of a three-drug ARV regimen at treatment doses (called antiretroviral therapy [ART]) to newborns with documented HIV infection (see Diagnosis of HIV Infection in Infants and Children
was initiated within 72h of infection the outcome might be closer to a successful PEP than a cure or remission and this should be discussed.
Please see additional comments below: Regarding the novel ex vivo experiment presented in supplementing Fig 1, not detecting integration before 5dpi in pre-stimulated PBMC infected in vitro in absence of RTG treatment is surprising and suggest that the integrated assay might be less sensitive than the total DNA one. The response to Reviewer 1 does not provide the sensitivity of the integrated assay nor a comparison of the LOD of this assay with that of the total DNA assay. Similarly, in the in vivo study, it cannot be excluded that integration is not detected before 3dpi in tissues and 5 dpi in PBMC because of a lack of sensitivity of the Alu gag PCR assay. Because the absence of detection of integrated DNA does not necessarily means that there is no integration, the authors should dampen their interpretations and conclusions. This won't take away from the main finding of this study that shows long-term control of SIV replication following treatment interruption in 4 out of 5 rhesus macaque infants initiated on ART 3 days post infection and maintained on ART for 9 months.
I believe there is an error in supplemental table 1 as according to Fig1A no animals were sacrificed 4dpi (I think the animals sampled at 4dpi were sacrificed at 28dpi). This should be corrected. Further, the sentence "The results showed that plasma viral load was detectable in ~50% neonates at 1dpi and subsequent time points thereafter" should be reformulated. After 1dpi all tested samples were positive (not 50%). In the results section, only the first group of animals is introduced "Newborn macaques were intravenously inoculated with identical doses of SIVmac251 within 6 hours after birth, and then euthanized at day 1, 2, 3, 5 and 7 post SIV inoculation for complete tissue collections (Fig. 1A)" before presenting the PVL results of figure 1B that includes animals from the other group of animals (necropsy at 28 dpi after ART). This is confusing. The authors should make clear that PVL were obtained from both groups in absence of ART. The new color coding of the symbols in Figure 1 is not explained in the legend or on the figure. I assume that all animals sacrificed at day 28 have a black symbol and the other ones got a color. I am not sure why the same color was used on different shapes for different animals. This could be made much easier for the readers. At least a key animal ID (as presented in supplemental Fig1)/symbol should be added to this figure.
Comment 8-It has to be noted that while all adult macaques experienced viral rebound following ATI in the Whitney paper, some adult macaques maintained undetectable PVL in the Okoye paper (Nat Med, 2019). While it is clear that the reservoir is seeded early, some differences are seen between models and while the pediatric model might be different, it seems premature to draw definitive conclusions.
Comment 9-Cell type (how they have been isolated), cell number per PCR ("generally 10M") and integration assay sensitivity are still obscure.

Response:
We appreciate reviewer's comments and admit that the Alu qPCR assay may not accurately reflect the bona fide seeding and size of viral reservoirs, and that the data based on this assay are not adequate to fully assess levels of potential provirus, so we cautiously interpreted the results. We also realize that protective mechanisms cannot be explained based on these assays, and more animal studies may provide more insight into the relationship of early provirus seeding and cure strategies. To this end, we recently requested newborn macaques for the comparison in the outcomes of treatment regimens with one or more antiretroviral drugs. Although these studies are underway, our primary study showed that plasma viral load was maintained high levels in infants on single DTG initiated at 2dpi (2dpi: 1.8×10 4 ± 8128; 7dpi: 7.4×10 4 ± 25172; 14dpi: 5.8×10 6 ± 3748497, n=3), at least partly demonstrating the contribution of linear viral DNA in virus production under integrase inhibitor treatment. We expect that our further studies and findings could answer these key questions.
Collectively, the data does suggest that there are few proviral copies in the early days of infection, but this is to be expected. Without a clear limit of detection for the Alu PCR, it is not possible to make the claim that virions are a result unintegrated viral DNA. Indeed, the early dynamics will be difficult to determine in vivo, as only small samples of PBMC can be analyzed.

Response:
We agree with the reviewer's comments, we sincerely revised description in the manuscript.
Other concerns raised in the initial review have been satisfactorily addressed with the edits.

Response:
We appreciate reviewer for the time and comments, and the opportunity to improve our manuscript.

Reviewer #2
The

Response:
We appreciate reviewer's comments. As mentioned above, we cautiously interpretate the Results and counterpart conclusion, and discuss the aspects of PEP, remission, or cure, please see in our revisions.
We also absolutely agree with the suggestions, further in vivo studies are needed to fully address the proviruses seeding, early treatment regimens and outcomes in a cure or remission for pediatric HIV infection. Therefore, we recently requested newborn macaques for the different animal groups, treated by one or more antiretroviral drugs administered at the specific early timepoint. We expect that the findings could answer these questions in near future.
Regarding the novel ex vivo experiment presented in supplementing Fig 1,   Response: We thank reviewer to point out and feel sorry to make confused. We corrected the typo in the supplemental table 1 and other descriptions as suggested. To track the viral parameters in individual animal, colored symbols with different shapes are coded. In Figure 1A, different symbol shapes represented plasma viral load of individual animals which were euthanized at day 1, 2, 3, 5 or 7 post SIV infection (colored solid symbols), or from staggered blood samples in infant animals that were sacrificed at 28dpi without treatment (opened black symbols). We added necessary notes and information in the figure legend.
Comment 8-It has to be noted that while all adult macaques experienced viral rebound following ATI in the Whitney paper, some adult macaques maintained undetectable PVL in the Okoye paper (Nat Med, 2019). While it is clear that the reservoir is seeded early, some differences are seen between models and while the pediatric model might be different, it seems premature to draw definitive conclusions.

Response:
We thank reviewer's concerns. We also noticed the obviously discrepant findings regarding early treatment and viral rebound off-ART in adult animals, as shown that viral rebound is not observed in 4/16 (ART initiation at 4-5dpi) or 3/17 animals (ART initiation at 6dpi) (Okoye, 2019), essentially contradictory to the reports (100% viral rebound in animals on ART initiation at 3dpi, Whitney, 2018) and clinical setting. It is interesting that total viral DNA is nearly undetectable in both PBMC and lymph nodederived lymphocytes in all SIV-infected adult on early ART throughout the studies (Okoye, 2019), yet only small animal populations ultimately do not show viral rebound after ATI. It is also curious to know the status of proviral DNA in the animal cohorts, unfortunately, proviruses have not been examined. The possible reason may lie in vaccination relevance and sensitivity of assays, despite the reason is still intrigued. We discussed these concerns in revised manuscript.
Comment 9-Cell type (how they have been isolated), cell number per PCR ("generally 10M") and integration assay sensitivity are still obscure.

Response:
We added more detail in Methods in revised manuscript.

REVIEWERS' COMMENTS
Reviewer #1 (Remarks to the Author): The revised manuscript has satisfactorily addressed the concerns noted in the second review.
Reviewer #2 (Remarks to the Author): Most comments from the previous review have been addressed. Please see remaining comments below 1-The first sentence of the abstract sounds like early ART leads to rapid viral rebound. ATI is followed by viral rebound despite years of treatment 2-Introduction. It is largely admitted that current ART regimens block new cycle of virus replication (with or without integrase inhibitor). Why would the pediatric treatment regimens "permit new or continual proviral seeding"?
3-Results. The new paragraph describing assay validation is poorly written. Supplementary Fig 1A is not referenced in the text and a slope of -2.3 is not good for a standard curve (-3.2?). As mentioned in the previous review, not detecting integration before 5dpi in pre-stimulated PBMC infected in vitro in absence of RTG treatment is surprising and suggest that the integrated assay might be less sensitive than the total DNA one.
4-As the sensitivity of the integration quantification assay is still unclear, I would suggest removing or rephrasing the following sentences: -"Our studies showed that in contrast to adults, skewing of viral integration was observed in neonatal macaques…" (Abstract page 2)