Analytical ART interruption does not irreversibly change pre-interruption levels of cellular HIV

Abstract

Objective:

The impact of short-term analytical antiretroviral therapy (ART) interruptions on the levels of cellular HIV and of residual activation after subsequent ART-mediated plasma HIV viral load re-suppression remains under active investigation.

Design:

Peripheral blood mononuclear cells (PBMC) from 23 ART-suppressed, chronically HIV-1-infected subjects were evaluated at the initiation of an analytical treatment interruption (ATI), during ATI, and following plasma re-suppression of HIV with ART.

Methods:

T cell activation was measured by flow cytometry. Total cellular HIV DNA, and episomal 2-long terminal repeat (2-LTR) circles were measured by droplet digital PCR (ddPCR). Cellular HIV multiply spliced RNA (tat/rev), unspliced (gag), and poly(A) tailed transcripts [poly(A)] were measured by reverse transcriptase-ddPCR. Analyses were performed using R version 2.5.1, or JMP Pro 11.

Results:

ATI (median ATI duration = 4 weeks) resulted in a rise of plasma HIV RNA (median=72900 copies/ml), decrease in CD4⁺ T cells/mm³ (median = 511.5 cells/mm³; P=0.0001), increase in T cell activation, and increase in cellular HIV DNA and RNA. Mean fluorescence intensity of CD38 on CD4⁺HLA-DR⁺ T cells at baseline was positively associated with total HIV DNA levels during ATI (pol: P=0.03, Rho=0.44). Upon ART resumption, plasma HIV re-suppression occurred after a median of 13 weeks and resulted in restoration of pre-ATI CD4⁺ T cells/mm³, T cell activation, and levels of cellular HIV DNA and RNA.

Conclusion:

Monitored viremia and immune activation during an ATI in ART-suppressed chronic HIV-infected subjects does not change the amount of persistent cellular HIV RNA or total HIV DNA after ART-mediated re-suppression.

Introduction

HIV cure-directed studies aim to permanently cease HIV replication after cessation of antiretroviral therapy (ART). The ability to block viral HIV replication after interruption of ART requires a strategy that can eradicate or control persistent measures of HIV before or during ART interruption. HIV cure-directed clinical strategies have tested therapy intensification^[1–7], purging of HIV-1 from the latent reservoir in the presence of ART (e.g. using T cell activating cytokines^[8–11], histone deacetylase inhibitors^[12–14], etc.), cytokine immunotherapy^[15], or gene therapy^{[16, 17]}. A limitation of all current HIV-cure-directed clinical research is the inability to accurately predict whether changes detected in persistent HIV or immune modulation while on suppressive ART are directly associated with control of HIV absent ART. Therefore, ART interruption and a closely monitored period of observation for the return of viremia is commonly included, as an “analytical” ART interruption (ATI) period. The role of an ATI step in determining the ultimate success of a cure-directed strategy is often counter-balanced by the concern for the potential risk of reversing the gains of long-term ART suppression upon return of viral replication. Prolonged ATI can have deleterious consequences^[18]. Therefore, there is concern that the rise of plasma HIV viral load (RNA) during an ATI, however limited, will result in a rise in residual immune activation and a higher amount of persistent cellular HIV after subsequent ART-mediated plasma HIV RNA re-suppression.

We previously reported the results of a longitudinal randomized, open-label, single-center study investigating the immune, viral, and safety outcomes of ATIs in subjects with chronically suppressed HIV-1 infection as compared to equal follow-up of patients on continuous ART and including a final therapy interruption in both arms^[19]. In the current study, we used activation data and cryopreserved PBMC from the above-mentioned parent study to investigate whether an ATI-mediated viremic episode, during which cellular HIV RNA and total DNA levels and immune activation rose, would irreversibly change the pre-ATI levels of cellular HIV RNA and total DNA load when measured following plasma HIV re-suppression with resumption of ART.

Materials and Methods

Study participants

Cryopreserved PBMC samples for the current study were derived from peripheral blood collected from 23 ART-suppressed, chronically HIV-1 infected subjects participating in a treatment interruption study (parent cohort of 42 subjects).

Patient demographics and parent study were previously described^{[19, 20]}. Briefly, in the parent study subjects were randomized in a continuous therapy/single interruption arm undergoing 40 weeks on continuous ART (Phase I) followed by an open-ended ATI (Phase II), and a repeated interruptions arm undergoing three sequential interruptions [Phase I: ATI number 1 (2 weeks), ATI number 2 (4 weeks), ATI number 3 (6 weeks)] followed by an open-ended ATI (Phase II). Recruitment characteristics in the parent study were as follows: subjects at least 18 years old, receiving ART (≥three drugs), with more than 400 CD4⁺ T cells/μl of blood with a history of nadir CD4⁺ T cell count of at least 100 cells/μl, and plasma HIV-1 RNA of less than 50 copies/ml at the time of ART interruption. All participants of the parent study had a history of ART-mediated suppression of less than 500 copies/ml for more than 6 months. As previously described, in the parent study criteria for ART re-initiation in open-ended ATI were: a viral load greater than 500000 copies/ml, or a viral load greater than 30000 copies/ml for three consecutive time points, or a drop in CD4⁺ T cell count more than 45%^[19].

In the current study, PBMC were evaluated at the initiation of a monitored ATI while on ART (baseline, time point 1), during ATI-associated HIV viremia (time point 2) and following re-suppression of HIV RNA after the resumption of ART (time points 3 and 4). Selection of the monitored ATI used in the current study was based on the following criteria: a) presence of a viremic episode (i.e. plasma HIV-RNA >50 copies/ml), b) available samples for all time points studied. The first ATI that met these criteria was used for analysis. Informed consent for the parent study was obtained according to the Human Experimentation Guidelines of the US Department of Health and Human Services and of the authors’ institutions. The study protocol was approved by the Institutional Review Boards of the Wistar Institute and Philadelphia FIGHT.

Assessment of T cell activation phenotypes by flow cytometry

T cell activation data collected during the parent study (parent cohort of 42 subjects) were available for statistical analysis in the current study. In the parent study, assessment of T cell activation [frequency (%) and mean fluorescence intensity (MFI) of CD38, HLA-DR, CD28, CD95 and TNFRII expression on CD3⁺, CD4⁺, CD8⁺ T cells] was performed using whole blood flow cytometry at the time of blood collection as previously described^[21–23] by using the following directly conjugated anti-cell surface antigen antibodies: 1) CD3-phycoerythrin (PE), CD38-PE, CD28-fluorescein isothiocyanate (FITC), CD4-allophycocyanin (APC), CD8-APC, isotypes: mouse IgG1-PE, mouse IgG2a-PE, mouse IgG1-FITC, mouse IgG2a-FITC, mouse IgG2b-APC, mouse IgG2a-TriColor (TC) (Pharmingen, San Diego, CA,USA); 2) CD95-FITC, CD4-FITC, HLA-DR-APC, CD4-TC, CD8-TC, isotypes: mouse IgG1-APC (CalTag, Burlingame, CA, USA); and 3) tumor necrosis factor II (TNFRII)-PE (R&D Systems, Minneapolis, MN).

Briefly, 100 μl of whole blood were first incubated for 10 min at room temperature (RT) with 10 μl of 10% mouse serum (Sigma-Aldrich, St. Louis, MO), and then stained for 20 min at RT with the appropriate monoclonal antibody combinations. The cells were then lysed for 10min at RT with 3 ml of FACS Lysis solution (BD Biosciences) and centrifuged for 5 min at 1200 rpm. Cells were subsequently washed twice with 3 ml of FACS washing buffer (1× PBS, 2.5% heat inactivated FBS, 0.1% BSA, 0.02% NaN₃) at 1500 rpm, and re-suspended in 200 μl of FACS washing buffer. Samples were analyzed on a Becton Dickinson FACScalibur flow cytometer using the CellQuest software package for acquisition and analysis. Live cell gates were set manually during acquisition of 10000 events for each staining. Detection thresholds were set according to isotype-matched negative controls. Results were expressed as mean MFI and % positive.

Assessment of cellular HIV DNA, and 2-LTR circular HIV DNA

DNA was extracted in the current study from cryo-preserved PBMC (5–15×10⁶ cells/vial) for each time point using the AllPrep DNA/RNA minikit (Qiagen, Valencia, CA). Extracted DNA was then used to quantify cellular HIV DNA [total HIV DNA (polymerase, pol) and the 2-long terminal repeat (2-LTR) junction] by droplet digital PCR (ddPCR) as previously described^[24–26]. Briefly, 1000 ng of DNA per replicate were digested with BSAJ1 enzyme (New England BioLabs, Ipswich, MA) before ddPCR quantification. Total HIV DNA (pol) and 2-LTR PCRs were performed with a duplex assay using HEX (pol) and FAM (2-LTR) probes, respectively, with the following cycling conditions: 10 min of initial enzyme activation at 95°C, 40 cycles consisting of a 30-s denaturation at 94°C followed by a 60°C extension for 60 s, and a final inactivation of 10 min at 98°C. The host cell RNase P/MRP 30-kDa-subunit gene (RPP30) was used as cellular normalizer, as it appears in 2 copies/cell. The digested DNA was diluted 10-fold, and 100 ng of the digested DNA/replicate were used for quantification of RPP30 by ddPCR (probe HEX) and cycled with the same parameters as for the pol, 2-LTR duplex. Copy numbers were calculated as the mean of replicate PCR measurements and normalized to 1×10⁶ PBMC and 1×10⁶ CD4⁺ T cells as determined by RPP30 (total cell count) and flow cytometry (CD4⁺ T cells %)^[26].

Assessment of cellular HIV RNA

Cellular HIV RNA was extracted in the current study from cryo-preserved PBMC (5–15×10⁶ cells/vial) for each time point using the AllPrep DNA/RNA minikit (Qiagen) according the manufacturer’s protocol. DNA contamination was avoided by addition of a DNase (RNase-free DNase Set, Qiagen) step to the protocol. Extracted RNA (500 ng) was then reverse transcribed into 20 μl of cDNA (iScript Advanced cDNA synthesis kit; Bio-Rad, Hercules, CA) according to the manufacturer’s protocol, and the cDNA product (8 μl; approximately 300 ng) was added to the ddPCR mixture. PCRs for assessment of unspliced HIV RNA (usRNA) (gag) and multiply spliced HIV RNA (msRNA) (tat/rev) were performed as a duplex with HEX (unspliced gag) and FAM (multiply spliced tat/rev) probes, respectively, using primers and probes as previously described^{[25, 26]}. Levels of all fully elongated and correctly processed HIV mRNA molecules (referred to as polyA) were also measured as previously described^[26–28], with the following cycling conditions for this PCR: 10 min at 95°C, 60 cycles consisting of a 30-s denaturation at 94°C followed by a 58°C extension for 60 s, and a final 10 min at 98°C. Copy numbers were calculated as the means of replicate PCR measurements and normalized to total RNA (1000 ng of total RNA) as determined by A260/A280 absorbance ratio using a NanoDrop 2000 spectrophotometer (Thermo Scientific, Waltham, MA)^[29].

Statistical analysis

Clinical markers (i.e. plasma HIV-1 RNA, CD4⁺ T cells/mm³, CD4%) and T cell activation data collected at the time of the parent study using whole blood, as well as cellular DNA and RNA data collected in the current study using cryopreserved PBMC from the parent study were used for statistical analysis. Data are presented as medians with 25^th and 75^th quartiles. For analysis and graphing purposes, plasma HIV RNA below 50 copies/ml was considered as equal to 50 copies/ml (threshold of detection). Differences between time points were assessed by Wilcoxon signed-rank test based on data distribution. Adjusted P values that were less than 0.1 are reported, based on the approach of Benjamini and Yekutieli (BY). Primary analysis focused on the comparison between baseline (time point 1) and post ATI (time points 3 and 4) levels of study variables. Correlations between T cell activation and viral measures were assessed using Spearman’s rank tests. Correlations were considered meaningful for Rho values above 0.3 with P values lower than 0.05. Analyses were performed using R version 2.5.1 (R Core Team, R Foundation for Statistical Computing, Vienna, Austria), or JMP Pro 11 (SAS Institute, Cary, NC, USA).

Results

Study subjects and analytical ART interruption (ATI) duration

A total of 23 ART-suppressed chronically HIV-1 infected subjects were studied. Study subjects characteristics are shown in Table 1. Ten of 23 patients were in the continuous therapy/single interruption arm of the parent study and the remaining thirteen patients were in the repeated interruptions arm. No difference was found at baseline values of all study variables between subjects from the continuous therapy/single interruption arm and subjects from the repeated interruptions arm.

Table 1.

Subjects and analytical therapy interruption characteristics.

Subject ID	Age	Sex	Ethnicity	Arm assignment in the parent study	Parent study ATI used in the current study	Baseline: Plasma HIV RNA (copies/ml) (Time point 1: on ART)	ATI: Plasma HIV RNA (copies/ml) (Time point 2: off ART)	ATI: Weeks off ART up to time point 2	ATI: Additional weeks off ART after time point 2	Post ATI: Weeks on ART up to time point 3	Post ATI: Weeks on ART after time point 3
S-04	38	M	White	Repeated interruptions	Phase I ATI number 2	<50	72900	4	0	12	3
S-07	59	M	White	Repeated interruptions	Phase I ATI number 2	<50	317984	4	0	12	3
S-14	29	M	White	Continuous therapy/single interruption	Phase II Open-ended ATI	<50	55430	16	44	29	N/A
S-19	39	M	Black	Repeated interruptions	Phase II Open-ended ATI	<50	28981	12	7	17	N/A
S-20	23	M	Black	Continuous therapy/single interruption	Phase II Open-ended ATI	<50	8970	28	21	7	2
S-21	39	M	White	Continuous therapy/single interruption	Phase II Open-ended ATI	<50	37372	9	14	15	N/A
S-22	49	M	White	Repeated interruptions	Phase I ATI number 2	<50	750000	4	0	18	5
S-23	47	M	White	Repeated interruptions	Phase I ATI number 2	<50	556372	4	0	16	3
S-29	45	M	Black	Continuous therapy/single interruption	Phase II Open-ended ATI	<50	71012	4	5	26	N/A
S-30	45	M	White	Repeated interruptions	Phase I ATI number 3	<50	517473	4	2	21	5
S-31	45	M	White	Repeated interruptions	Phase II Open-ended ATI	<50	32558	8	38	12	N/A
S-32	41	M	White	Continuous therapy/single interruption	Phase II Open-ended ATI	<50	18034	9	26	13	N/A
S-33	49	F	Black	Continuous therapy/single interruption	Phase II Open-ended ATI	<50	73066	4	1	17	N/A
S-36	40	M	White	Continuous therapy/single interruption	Phase II Open-ended ATI	<50	363774	5	3	13	N/A
S-40	55	M	White	Continuous therapy/single interruption	Phase II Open-ended ATI	<50	94489	5	3	30	N/A
S-41	54	M	White	Repeated interruptions	Phase II Open-ended ATI	<50	124747	8	38	8	N/A
S-42	41	M	White	Repeated interruptions	Phase I ATI number 2	<50	16083	4	0	8	3
S-44	43	M	White	Repeated interruptions	Phase I ATI number 2	<50	127193	4	0	12	1
S-45	39	M	White	Repeated interruptions	Phase I ATI number 2	<50	67248	4	0	8	4
S-54	42	M	White	Repeated interruptions	Phase I ATI number 2	<50	116243	4	0	18	1
S-56	41	M	White	Repeated interruptions	Phase I ATI number 2	<50	31377	4	0	13	N/A
S-60	50	M	White	Continuous therapy/single interruption	Phase II Open-ended ATI	<50	95361	5	45	7	N/A
S-62	48	M	Black	Continuous therapy/single interruption	Phase II Open-ended ATI	<50	50715	9	15	15	N/A

ATI, analytical treatment interruption; ART, antiretroviral therapy; M, male; F, Female; N/A: not available

Clinical markers and T cell activation data collected in the parent study as well as cellular DNA and RNA measured in the current study in cryopreserved PBMC from the parent study were evaluated at baseline on ART (time point 1), during a monitored ATI (time point 2) of a median duration of 4 weeks, (IQR=4–9), and after plasma re-suppression of HIV RNA with resumption of ART (time point 3: median time on ART=13 weeks, IQR=12–18). The median peak plasma HIV RNA during ATI was 72900 copies/ml (IQR=32558–127193), while the HIV RNA at all time points on ART was <50 copies/ml. Ten of 23 patients had samples available for measurement of cellular DNA for an additional time point on ART (time point 4: median time after time point 3=3 weeks, IQR=2–4).

ATI-associated changes in plasma HIV viral load and T cell activation return to pre-TI levels after ART re-initiation

We first assessed the impact of ATI and subsequent resumption of ART on clinical markers (i.e. plasma HIV-1 RNA, CD4⁺ T cells/mm³, CD4%) and markers of T cell activation using data collected in the parent study for the time points of the current study. Results are summarized in Fig. 1, and Supplementary Table 1. As expected, ATI resulted in a rise in plasma HIV RNA and a decrease in CD4⁺ T cells/mm³ (P=0.0001), and CD4% (P=0.0004). Markers of activation (e..g. HLA-DR and CD38 co-expression and MFI of CD38 on T cells have been associated with disease progression^[30–32]. As a result, changes in these markers were evaluated in this study along with other markers of T cell activation. An increase from baseline in markers of T cell activation (e.g. CD8⁺HLA-DR⁺ %: P=0.03; CD8⁺CD38⁺ %: P=0.005; CD8⁺CD38⁺HLA-DR⁺ %: P=0.0005; MFI of CD38 on CD8⁺HLA-DR⁺ T cells: P<0.0001; CD4⁺HLA-DR⁺ %: P=0.004; CD4⁺CD38⁺HLA-DR⁺ %: P=0.001; MFI of CD38 on CD4⁺HLA-DR⁺ T cells: P=0.004) were noted during ATI-mediated viremia. The resumption of ART resulted in re-suppression of plasma HIV RNA and restored CD4⁺ T cells/mm³ levels and frequencies of activated T cells to pre-ATI levels. No change was observed for CD4⁺CD38⁺ % at any time point. Return to pre-ATI levels were further re-confirmed by a later time point on ART in 10 subjects (time point 4).

Fig. 1. T cell activation markers increase during an analytical ART interruption but return to pre-interruption levels following ART-mediated viral suppression.

(a) Plasma HIV RNA (log₁₀ HIV-1 RNA copies/ml) at baseline of analytical ART interruption (ATI) while on ART (time point 1), during ATI (time point 2), and following ART-mediated HIV RNA re-suppression (time points 3 and 4). (b) CD4⁺ T cells/mm³ at baseline of ATI while on ART (time point 1), during ATI (time point 2), and following ART-mediated HIV RNA re-suppression (time points 3 and 4). (c) CD8⁺HLA-DR⁺ T cells %, CD8⁺CD38⁺ T cells % and CD8⁺CD38⁺HLA-DR⁺ T cells %, CD4⁺HLA-DR⁺ T cells percentage (%), CD4⁺CD38⁺ T cells %, CD4⁺CD38⁺HLA-DR⁺ T cells % at baseline of ATI while on ART (time point 1), during ATI (time point 2), and following ART-mediated HIV RNA re-suppression (time points 3 and 4). Data are shown per patient during follow-up with inter-quartile box plots with median and outliers, and significant P values. Different symbols/colors were used per patient as indicated in the side of the Figure.

Increases in cellular HIV RNA and DNA during ATI return to pre-ATI levels after re-initiation of ART

We assessed the effect of ATI and ART resumption on cellular HIV levels. First, ATI-mediated viremia (time point 2) resulted in an increase from baseline in cellular HIV DNA (pol: P<0.0001) and RNA [tat/rev: P<0.0001; gag: P=0.0004; poly(A): P<0.0001] levels. Second, as observed with T cell activation changes, subsequent resumption of ART and plasma HIV RNA re-suppression (time points 3 and 4) returned levels of cellular HIV DNA (pol) and RNA [tat/rev, gag, poly(A)] to pre-ATI levels (Fig. 2, Supplementary Table 2).

Fig. 2. Rise of cellular HIV DNA and RNA during ATI followed by return to pre-interruption levels upon ART-mediated viral suppression.

(a) cellular HIV DNA [total HIV DNA (pol copies) and episomal 2-long terminal repeat (2-LTR) circles] at baseline of ATI while on ART (time point 1), during ATI (time point 2), and following ART-mediated HIV RNA re-suppression (time points 3 and 4). (b) cellular HIV multiply spliced RNA (tat/rev), unspliced (gag), and poly(A) tailed transcripts [poly(A)] at baseline of ATI while on ART (time point 1), during ATI (time point 2), and following ART-mediated HIV RNA re-suppression (time point 3). Data are shown per patient during follow-up with inter-quartile box plots with median and outliers, and significant P values. Different symbols/colors were used per patient as indicated in the side of the Figure. Log10 values of each variable were used in the plots to better show the spread of the values (particularly at baseline and post ATI).

Positive association between pre-ATI levels of T cell activation and cellular HIV levels during ATI

We also assessed the association between markers of T cell activation and cellular HIV measures during follow-up. Although no significant association was observed following adjustments for multiple testing, unadjusted data showed an association between T cell activation and HIV measures. First, pre-ATI levels (time point 1) of activated CD4⁺ T cells (i.e. MFI of CD38 on CD4⁺HLA-DR⁺ T cells) were positively associated with cellular HIV DNA during ATI (time point 2, pol: P=0.03, Rho=0.44) and after ART-mediated plasma HIV RNA re-suppression (time point 3, P=0.03, Rho=0.47) (Fig. 3). Interestingly, no association was observed between pre-ATI levels (time point 1) of T cell activation and plasma HIV RNA during ATI (time point 2). Second, as expected, positive associations between T cell activation and cellular HIV measures during ATI were detected [e.g. MFI of CD38 on CD8⁺HLA-DR⁺ T cells and tat/rev (P=0.01, Rho=0.52), gag (P=0.009, Rho=0.54), and poly(A) (P=0.009, Rho=0.54) respectively, Fig. 3].

Fig. 3. Correlation between T cell activation and cellular HIV levels.

All panels show Spearman’s Rank correlations. (a) Correlation of mean fluorescent intensity (MFI) of CD38 on CD4⁺HLA-DR⁺ T cells at baseline of ATI while on ART with cellular HIV pol during ATI. (b) Correlation of MFI of CD38 on CD4⁺HLA-DR⁺ T cells at baseline of ATI while on ART with cellular HIV pol after ART-mediated viral re-suppression. (c) Correlation of MFI of CD38 on CD8⁺HLA-DR⁺ T cells during ATI with cellular HIV DNA (upper left panel: 2-LTR circles) and cellular HIV RNA [upper right panel: tat/rev, bottom panels: gag, and poly(A)] during ATI. Data are shown as regression lines with number of patients, Rho and P values. Note that although Spearman’s Rank correlations were performed, regression lines were used for graphic purposes only.

Discussion

In this study we assessed the impact of analytical therapy interruption on cellular HIV DNA and RNA levels and on residual activation before ATI and after ART-re-initiation. We found no evidence that a monitored period of therapy interruption with associated viremia altered the levels of plasma and cellular HIV measures after resuming ART and achieving re-suppression.

Although proviral HIV DNA has been used as a general measure of cellular reservoir size, other measures such as 2-LTR circles and cellular RNA provide additional information about viral dynamics and presence of transcriptionally active cells^[24]. Cellular HIV-1 DNA levels can represent an important virological correlate of the natural history of the disease as they are correlated with the extent of immunosuppression in naive patients^{[33, 34]}, and can be informative on suppressive outcomes after ART treatment^[35]. Therefore, a lack of change from pre-ATI levels of HIV DNA after re-suppression with ART as well as the lack of difference on HIV DNA baseline levels between subjects with a history of multiple ATIs (i.e. repeated interruptions arm) and subjects with no history of ATIs (i.e. continuous therapy/single interruption arm) provides re-assurance that monitored viremic episodes are unlikely to significantly reverse the benefits of ART or result in a larger latent reservoir after ATI in participants who initiated ART during chronic infection.

Our data in 23 subjects confirm and extend previous studies ^{[33, 36, 37]} in smaller number of subjects also showing a similar increase of cellular HIV DNA and RNA measures after viremia and therapy interruption, followed by a decline after resumption of therapy. Furthermore, it is important to note that our data are derived from a majority of short-term ATIs (most <6 weeks) in contrast to open-ended therapy interruptions which may allow for months of viremia and a different outcome after restarting ART^[18]. Moreover, our study now shows that the assessment of the different cellular viral RNAs produced in HIV-infected cells [msRNA (tat/rev), usRNA (gag), and poly(A) tailed transcripts] informs whether stages of viral transcription have been altered due to the ATI. Although cellular multiply spliced and unspliced HIV RNA increased during ATI, it was of interest to note that no rise in transcriptionally active cells remained with sustained suppression after ART. We interpret the latter to indicate no long-term impact of a monitored short ATI on host or viral factors in circulation after ART-mediated plasma HIV RNA re-suppression. Future studies will need to assess if tissue level of infected cells is equally unchanged after ART re-suppression following a short-term ATI as suggested by our data.

In addition, our data further support the hypothesis that residual activation on ART is not irreversibly impacted by ATI as evidenced by a return in the frequency of activated T cells to pre-ATI levels after resumption of ART ^[37], as well as the lack of difference on T cell activation baseline levels between subjects with or without a history of multiple ATIs (i.e. repeated interruptions arm vs. continuous therapy/single interruption arm). Our data though, do not address cases in which T cell activation may be increased by causes independent of viral rebound during or after an ATI (i.e. onset of comorbidity). Of interest, although unadjusted analysis did detect a positive association between pre-ATI levels of T cell activation markers and cellular HIV levels during ATI no association was observed between T cell activation on ART and the magnitude of the ATI viral rebound. This finding further supports the interpretation that plasma HIV RNA is a more direct indicator of the “magnitude of ATI viral rebound” than cellular HIV levels.

This study has limitations. First, cellular HIV DNA and RNA levels were assessed in total PBMC and results were expressed per 10⁶ PBMC or per 10⁶ CD4⁺ T cells based on flow cytometric assessment of % CD4⁺ T cells. While determinations on isolated CD4⁺ T cells would have provided greater sensitivity by removing the potential impact of cellular shifts in PBMC from affecting comparisons between time-points, we did observe a rise in cellular viral measures at a time of decreasing CD4⁺ T cell count (ATI) to support our ability to measure rises within CD4⁺ T cell subsets in PBMC. Second, no direct measurement of integrated DNA was performed. This measurement could further inform whether the observed changes in total DNA and 2-LTR circles observed during this study are not due to changes in the linear or circular unintegrated DNA that is more transient than genomic DNA. Yet no change in the 2-LTR circles was found in our study as a result of the ATI. Furthermore, although not exactly reflecting the amount of integrated DNA, subtraction of 2-LTR from total DNA showed similar results as total DNA suggesting that the observed changes are due to changes in integrated DNA. Future studies assessing changes in integrated DNA could further clarify this. Third, cellular HIV content was assessed by PCR-based assays only; no direct measurement of replication competent virus or viral sequence diversity was performed. Although recent data suggest that levels of replication competent virus on ART remain unchanged after ART re-initiation following an ATI as suggested by our data^[37], this needs to be validated in a larger cohort.

Irrespective of viral rebound or immune activation status, all subjects undergoing an ATI re-suppressed after re-starting ART. It is of interest to note that drug resistance data for 14 of 23 patients participating in this study were previously reported^{[23, 38]} to document a lack of drug resistance emergence predicting failure to re-suppress after resumption of ART. Joining our data with others^{[33, 36, 37]}, clinical evidence to date shows a lack of viral and host adverse effects (resistance, increases in persistent activation/viral levels) following completion of monitored ATIs. As future clinical study designs move forward to test additional cure-directed concepts, the information reported here is expected to further support that the inclusion of monitored ATIs is scientifically warranted.