Secondary Logo

Journal Logo


HIV latency can be established in proliferating and nonproliferating resting CD4+ T cells in vitro

implications for latency reversal

Moso, Michael A.a,b,c; Anderson, Jenny L.a; Adikari, Samanthaa; Gray, Lachlan R.b,c; Khoury, Georgesa,d; Chang, Judy J.a; Jacobson, Jonathan C.a,d; Ellett, Anne M.c; Cheng, Wan-Jungc; Saleh, Suhaa; Zaunders, John J.e,f; Purcell, Damian F.J.a,d; Cameron, Paul U.a,b,c; Churchill, Melissa J.g; Lewin, Sharon R.a,b; Lu, Hao K.a

Author Information
doi: 10.1097/QAD.0000000000002075



Despite the successes of antiretroviral therapy (ART), HIV cannot be cured because of the persistence of latently infected resting and proliferating memory T cells [1–5]. One strategy being developed to eliminate latently infected T cells is to stimulate HIV transcription using latency-reversing agents (LRAs) [6]. The response to LRA stimulation has recently been shown to depend on the viral integration site [7], however other factors such as the proliferative history and cellular activation state may also play a role.

HIV latency in primary CD4+ T cells can be established either through direct infection of resting cells (preactivation latency) [8–11], or through the infection of activated CD4+ T cells that then revert to a ‘resting’ state (postactivation latency) [12–15]. In postactivation models, a ‘resting’ phenotype is often defined by the lack of expression of activation markers CD69, CD25 and human leukocyte antigen (HLA)-DR [14–17]. In addition, most in vitro models use laboratory-derived virus strains and this may not accurately represent HIV sequences in vivo, which are diverse, especially in individuals who initiate ART during chronic infection [18]. We recently demonstrated that the activity of LRAs was different dependent upon how latency was established [19]. Here we hypothesized that the proliferation history of a latently infected cell will also affect the response to an LRA.

To test this hypothesis, we developed a model of postactivation latency using a luciferase reporter construct that was able to integrate and persist in activated and resting T cells. We confirmed latent infection by the demonstration of inducible luciferase expression in two distinct T-cell subsets that did not express activation markers: nonproliferating resting cells (nonblasts) and proliferating cells (CD69CD25HLA-DR transitional blasts). Finally we used this model to assess the response to LRAs following infection with either wild type virus or virus that contained patient-derived long terminal repeat (LTR) sequences.


Infection of CD4+ T cells and cell sorting

Resting CD4+ T cells were isolated from healthy donor peripheral blood mononuclear cells (Australian Red Cross Blood Service, Southbank, Australia) as previously described [20]. Resting CD4+ T cells were prestained with the proliferation dye, eFluor450 (eBioscience, San Diego, California, USA) and either left in culture, stimulated with CCL19 (100 nmol/l; R&D Systems, Minneapolis, Minnesota, USA), or activated with phytohemagglutinin (PHA; 10 μg/ml; Sigma-Aldrich, St. Louis, Missouri) and IL-2 (10 IU/ml; Roche Applied Sciences, Penzberg, Germany) at a density of 2 × 106 cells/ml in RF10 media [RPMI1640 (Life Technologies, Carlsbad, California, USA), 10% fetal bovine serum (Bovogen Biologicals, Melbourne, Victoria, Australia) and 1× penicillin-streptomycin-glutamine (Life Technologies)]. After 48 h, cells were infected with NL4.3-enhanced green fluorescent protein (EGFP) at 50% tissue culture infectious dose (TCID50) of 0.5, or 20× concentrated vesicular stomatitis virus-G (VSV-G) pseudotyped viral supernatant at a ratio of ∼10 ng p24 to 1 million cells. Uninfected ‘mock’ and raltegravir controls were conducted concurrently; raltegravir (1 μmol/l; Selleck Chemicals, Houston, Texas, USA) was added 5 min prior to infection. Cells treated with the chemokine CCL19 were cultured in 100 nmol/l CCL19 for 24 h prior to infection, as previously described [20].

After 12 days, infected cells were stained with anti-CD69, anti-CD25 and anti-HLA-DR (all BD Biosciences, Franklin Lakes, New Jersey, USA). Cells were then sorted by flow cytometry (MoFlo Astrios; Beckman Coulter, Brea, California, USA) based on forward-scatter and side-scatter, eFluor and expression of CD69, CD25 and HLA-DR ± EGFP into three populations: eFluorhigh CD69CD25HLA-DR/EGFP nonblasts, eFluorlow CD69CD25HLA-DR/EGFP blasts (transitional blasts), and CD69+CD25+HLA-DR+ blasts (for VSV-G pseudotyped virus only). Cells were harvested for baseline luciferase expression or reactivated with various LRAs.

Cloning of patient long terminal repeats into a vector able to integrate and express luciferase

We amplified and cloned the HIV LTR from integrated virus in memory CD4+ T cells from four HIV-infected individuals prior to and following ART, as previously described [21]. To clone patient-derived LTRs into a retroviral vector, we first added the polypurine tract (PPT) to each LTR isolate using an overlap extension PCR (Table S1, A lentiviral vector with a luciferase reporter gene (pGBFM) was then used to evaluate LTR function as previously described [22]. Purified PPT-LTR fragments and the pGBFM vector were digested using KpnI (Promega, Madison, Wisconsin, USA) and XbaI restriction enzymes (New England Biolabs, Ipswich, Massachusetts, USA), and ligated using T4 DNA ligase (New England Biolabs). The ligation product was transformed into Stbl3 Escherichia Coli competent cells (Life Technologies). Successful clones were identified using colony PCR and confirmed by sequencing (AGRF, Melbourne, Victoria, Australia).

Production and quantification of vesicular stomatitis virus-G pseudotyped and NL4.3-enhanced green fluorescent protein viruses

For VSV-G pseudotyped virus production, 293 T cells (5 × 105) were transfected as previously described [22]. Viral production was quantified by measuring p24 levels by ELISA with antip24 sheep antibody D7320 (Aalto Bio Reagents, Dublin, Ireland). NL4.3-EGFP was generated and the TCID50 was determined as described in [23].

Reactivation of cells with latency-reversing agents

Infected CD4+ T cells were harvested for baseline luciferase expression at day 4 (unsorted population) and day 12 postinfection (unsorted and sorted populations). The unsorted and sorted cell populations (nonblast and blasts) harvested at day 12 were reactivated with the positive control PHA (10 μg/ml) and phorbol myristate acetate (PMA, 10 nmol/l; both Sigma-Aldrich), the vehicle control dimethyl sulfoxide (DMSO; Sigma Aldrich), or a panel of LRAs: romidepsin (40 nmol/l), panobinostat (50 nmol/l), vorinostat (0.5 μmol/l; all Selleck Chemicals), JQ-1 (1 μmol/l) and chaetocin (10 nmol/l; both Sigma-Aldrich). Five hundred thousand cells were plated and treated with raltegravir (1 μmol/l, Selleck Chemicals) for 5 min prior to the addition of LRAs. All cells were reactivated in the presence of 10 IU/ml IL-2 at a cell density of 1 × 106 cells/ml. After 24 h, 200 000 viable cells were replated and luciferase quantified using the Luciferase Assay System (Promega). For infection with the EGFP expressing wild-type virus, raltegravir was added to the cultures prior to addition of either romedepsin (40 nmol/l) or PMA/PHA (10 nmol/l/10 μg/ml) and after 24 h, EGFP was quantified by flow cytometry. Data are shown as fold-change over DMSO control or percentage of PMA/PHA stimulation.

Cytotoxicity assay

Resting CD4+ T cells were treated with varying concentrations (range dose 10–100 000 nmol/l for all drugs) of the histone deacetylase inhibitors (HDACi) romidepsin, panobinostat and vorinostat; the bromodomain inhibitor JQ-1, and the histone methyltransferase inhibitor chaetocin. All drugs were diluted in DMSO to establish a final DMSO concentration of less than 1%. Two hundred thousand resting CD4+ T cells were plated and incubated for 48 h with the appropriate concentration of LRAs. A negative control with DMSO was included where no drug was added. A positive control was used where NP-40 (Thermo Fisher Scientific, Waltham, Massachusetts, USA) was added to induce cell death. After 48 h, the proliferation dye (CellTiter 96 Aqueous One Solution Cell Proliferation Assay, Promega) was added to each well and incubated for 6–8 h. Colorimetric analysis was performed using a plate reader (Thermo Fisher Scientific) measured at OD 492 nm. Data were analyzed using GraphPad Prism software (Version 6, Graphpad, La Jolla, California, USA) using nonlinear regression analysis.

Phenotyping CD69-CD25-human leukocyte antigen-DR blasts and nonblasts

The negatively sorted CD69CD25HLA-DR blast and nonblasts were stained with anti-CD45RO-FITC, anti-CD45RA-PE-Cy7 (BD Biosciences), anti-CCR7-APC, eFluor-780, anti-CD27 PE-Cy5 (all eBioscience) and anti-CD4-PE-Texas Red (Invitrogen, Carlsbad, California, USA), and analyzed using a flow cytometer (LSRFortessa; BD Biosciences).

Statistical analysis

A one-way analysis of variance with Tukey posttest was used to measure the difference in basal luciferase expression between unstimulated cells, CCL19 and PHA/IL-2 stimulated cells; or between different sorted and unsorted cell populations; or between different LRAs. The Student t tests were used to measure differences between DMSO and PMA/PHA-treated populations; or in the PMA/PHA and romidepsin-treated nonblast populations.


Establishing latency in vitro using a vesicular stomatitis virus-G pseudotyped long terminal repeat reporter virus

We first developed a VSV-G pseudotyped luciferase reporter virus containing the wild-type NL4-3 LTR that was able to integrate into primary T cells. Using this virus, we were able to infect PHA/IL-2 activated CD4+ T cells but not resting or CCL19 treated CD4+ T cells (Fig. 1a).

Fig. 1
Fig. 1:
Infection of phytohemagglutinin/IL-2-activated CD4+ T cells with NL4-3 long terminal repeat vesicular stomatitis virus-G pseudotyped pGBFM virus.(a) Primary resting CD4+ T cells isolated from healthy blood donors were left unstimulated, treated with CCL19, or were activated with phytohemagglutinin in the presence of IL-2 (with or without the integrase inhibitor raltegravir) for 48 h prior to infection with NL4-3-long terminal repeat vesicular stomatitis virus-G pseudotyped pGBFM virus. Infection was quantified by measuring the level of luciferase produced on day 4. Mean ± SD values are shown from four independent experiments. * P less than 0.05, ** P less than 0.01. (b) Phytohemagglutinin-activated CD4+ T cells were infected as in (a) and the level of luciferase was measured every 1–3 days postinfection. On days 12 and 19 postinfection, cells in culture were further stimulated with phorbol myristate acetate/phytohemagglutinin and luciferase measured 2 days after reactivation. The mean ± SD luciferase expression is shown from three independent donors.

Given this pseudotyped virus was able to infect activated, but not quiescent T cells, we investigated whether latency could be established in primary T cells by allowing activated cells to return to a resting state through long-term culture with 10 IU/ml IL-2. Cells were harvested every 2–3 days to monitor luciferase activity. Luciferase expression increased rapidly during the first 4 days after infection and plateaued up to day 21 (Fig. 1b), consistent with constitutive expression of luciferase. Because this was a single round noninfectious virus, constitutive expression of luciferase did not result in new rounds of viral replication. The addition of PMA/PHA at days 12 and 19 postinfection, in the presence of raltegravir, led to large increases in luciferase production 48 h later (mean ± SEM increase above mock of 8.6 ± 0.4 and 11.6 ± 0.9 fold respectively; Fig. 1b). These findings demonstrate that in this culture system, even in the setting of constitutive luciferase expression, there were cells infected with inducible virus, consistent with latency.

Changes in cell size over time in a model of postactivation latency

To better understand the kinetics of T-cell activation and proliferation in this model, we assessed expression of the activation markers CD69, CD25 and HLA-DR as well as history of proliferation using the cytoplasmic dye eFluor. The early activation marker CD69 peaked 2 days after T-cell activation, followed by CD25 at day 4 and then both declined. The late activation marker HLA-DR only began to increase at 3–4 days after and remained elevated to day 14 (Fig. 2a). Cell viability declined shortly after PHA/IL-2 stimulation and continued to decline over the next 30 days (Fig. 2b). By staining the T cells with eFluor prior to stimulation, we demonstrated the existence of two distinct populations: eFluorlow large cells that had proliferated (blasts); and eFluorhigh small cells (nonblasts) that had never proliferated (Fig. 2c). The proportion of blasts progressively increased over time and plateaued at about 80% (Fig. 2d), while the percentage of nonblasts decreased proportionally.

Fig. 2
Fig. 2:
The phenotype and viability of CD4+ T cells during long-term culture.(a) Kinetics of expression of T-cell activation markers following stimulation with phytohemagglutinin/IL-2. Primary resting CD4+ T cells were activated with phytohemagglutinin in the presence of IL-2 and the percentage of CD69, CD25 and human leukocyte antigen (HLA)-DR in the viable cell population was measured by flow cytometry. Data represent the mean from two individual donors. (b) Resting CD4+ T cells were activated with phytohemagglutinin/IL-2 with or without infection with NL4.3-long terminal repeat vesicular stomatitis virus-G pseudotyped pGBFM virus and cell viability was measured by flow cytometry. Linear regression of the uninfected (grey line) and infected (black line) donors is also shown. (c) Gating strategy to determine the proportion of nonblast, blast and nonviable cells by flow cytometry based on forward-scatter and side-scatter, respectively. A representative flow plot of cells collected 12 days postactivation, prior to sorting is shown. Cells stained with the proliferation dye, eFluor, were also used to confirm the gating strategy. (d). Percentage of blast and nonblasts over time is shown. Percentage was calculated based on the total number of viable cells. Each line represents an individual donor, uninfected (dashed lines) and infected (solid line) are shown.

Inducible virus in CD69-CD25-HLA-DR-blasts and nonblasts

To determine which populations of cells were infected with the VSV-G pseudotyped LTR reporter virus, we sorted the cells on day 14 postactivation based on their morphology and expression of CD69, CD25 and HLA-DR (Fig. 3a). Luciferase was measured prior to and following reactivation with PMA/PHA. The nonblasts had minimal constitutive luciferase expression (mean 1.9 ± 0.4 fold increase above mock, Fig. 3b). As expected, constitutive expression of luciferase was highest in the CD69+CD25+HLA-DR+ blasts (60.5 ± 14.3 fold above mock), followed by the unsorted cells (31.9 ± 7.5 fold) and the CD69CD25HLA-DR blasts, that we here call transitional blasts (24.0 ± 4.4 fold) (Fig. 3b).

Fig. 3
Fig. 3:
Inducible virus is detected in both CD69CD25HLA-DR transitional blasts and nonblasts.(a) Schematic of the method used to establish HIV latency in vitro using a vesicular stomatitis virus-G pseudotyped reporter virus. Phytohemagglutinin/IL-2-activated CD4+ T cells were infected and maintained in culture for 12 days. CD69CD25HLA-DR transitional blast (eFluorlow) and nonblast (eFluorhigh) and CD69+/CD25+/HLA-DR+ cells were sorted and luciferase activity was quantified. Only the activation marker negative cells were then incubated with various latency-reversing agents and the change in luciferase activity was quantified again. (b) Activated CD4+ T cells were infected with a vesicular stomatitis virus-G pseudotyped reporter virus (using both patient-derived and NL4-3 long terminal repeats) and luciferase production was measured in the unsorted, CD69CD25HLA-DR transitional blast and nonblast, or the CD69+/CD25+/HLA-DR+ blast populations immediately after sorting. Only statistically significant differences are shown. Each symbol represents a different donor and the black lines represent the mean ± SEM. (c) The sorted cells from (b) were either reactivated with phorbol myristate acetate and phytohemagglutinin or treated with the negative vehicle control dimethyl sulfoxide for 24 h and luciferase production measured. The fold increase in luciferase following phorbol myristate acetate/phytohemagglutinin treatment, relative to dimethyl sulfoxide, is shown as the ratio of the arithmetic mean ± SEM. * P less than 0.05; ** P less than 0.01; *** P less than 0.001. Each symbol represents a different virus derived from long terminal repeats isolated from CD4+ T cells from HIV-infected individuals, preantiretroviral therapy (open) or postantiretroviral therapy (closed) using CD4+ T cells from the same healthy donor.

To identify which subsets of cells contained inducible virus, PMA/PHA was then used to restimulate luciferase expression. Upon reactivation with PMA/PHA, a significantly higher fold-increase in luciferase expression was observed in the transitional blasts (4.1 fold, P = 0.012), compared with unsorted cells (2.5 fold, P = 0.003) or the nonblast population (2.6 fold, P = 0.02) (Fig. 3c). The transitional blasts contained a mixed population of central memory (TCM, CD45RACCR7+CD27+, 16.9 ± 2.1%), effector memory (TEM, CD45RACCR7CD27, 4.7 ± 1%), transitional memory (TTM, CD45RACCR7CD27+, 16.7 ± 2.2%), naive (CD45RA+CCR7+CD27+, 41.5 ± 9.2%) and negligible number of terminally differentiated cells (TTD, CD45RA+CCR7CD27, 0.1 ± 0.0%) (Fig. S1A, Significantly, the nonblasts were also comprised of similar cell subpopulations but with a higher proportion of naive T cells (51.9 ± 5.5%) cells (Fig. S1B, These data demonstrate that latent infection can be established in vitro in two subsets of cells that don’t express activation markers – nonblasts and transitional blasts.

Reversal of latency in nonblasts infected with the replication competent enhanced green fluorescent protein reporter virus

Given that the pseudotyped virus was a single round virus that also bypassed chemoreceptor signaling because of the VSV-G protein, we repeated the same experiments with a replication-competent full-length wild-type NL4.3 virus that expressed EGFP [24]. Following infection, the cells were cultured for 12 days and EGFP cells sorted into eFluorhigh and eFluorlow cells (Fig. 4a). The sorted cells were incubated with the integrase inhibitor raltegravir to block further rounds of infection and then stimulated with either romidepsin or PMA/PHA. We observed a significant increase in EGFP expression in the nonblasts (Fig. 4b, Fig. S2, but no overall increase in EGFP expression in the transitional blasts.

Fig. 4
Fig. 4:
Postactivation latency following infection with NL4.3-enhanced green fluorescent protein.(a) Activated CD4+ T cells were infected with NL4.3-enhanced green fluorescent protein virus and cultured for 12 days to allow cells to return to rest. Cells were sorted at day 12 postinfection for enhanced green fluorescent protein-negative CD69CD25HLA-DR transitional blasts (eFluorlow) and nonblasts (eFluorhigh) as shown. (b) Sorted blasts and nonblasts were reactivated with the histone deacetylase inhibitor romidepsin, phorbol myristate acetate/phytohemagglutinin or with control dimethyl sulfoxide. The number of enhanced green fluorescent protein+ cells per 10 000 events was quantified by flow cytometry. In each experiment between 10 000 and 100 000 events were collected. Data represent mean ± SEM of six individual donors. Each symbol represents a different donor. * P less than 0.05.

Latency reversing agents induced luciferase production from transitional blasts and nonblasts

We next assessed the responsiveness of infected transitional blasts to a panel of LRAs at noncytotoxic concentrations (Fig. S3, using VSV-G pseudotyped LTR luciferase reporter viruses carrying wild-type NL4-3 or patient-derived LTRs. Using the NL4-3 LTR pseudotyped virus, chaetocin displayed the highest potency (3.1 ± 0.3 fold above DMSO) in transitional blasts (Fig. 5a). Of the HDACi tested, romidepsin induced the highest fold-increase in luciferase activity (1.9 ± 0.7 fold), followed by panobinostat (1.8 ± 0.6 fold), and vorinostat (1.1 ± 0.1 fold). Cells treated with the positive control, PMA/PHA, induced the highest level of luciferase (3.7 ± 1.1 fold) (Fig. 5a). The effect of LRAs was also expressed as a percentage of maximum stimulation with PMA/PHA (Fig. 5b). We then compared the ex-vivo response of three patient-derived LTRs (Fig. 5c). Consistent with our previous in-vitro model of latency [21], patient-derived LTRs were responsive to LRA stimulation with no significant difference observed across patient-derived LTRs and wild-type NL4-3 (Fig. 5c and d, Fig. S4,

Fig. 5
Fig. 5:
Latency reversing agents induce luciferase expression in both transitional blasts and nonblasts.(a) Sorted CD69CD25HLA-DR transitional blasts infected with pseudotyped-virus containing NL4-3 long terminal repeat were treated with various latency reversing agents including the histone deacetylase inhibitors romidepsin, panobinostat and vorinostat; the histone-methyltransferase inhibitor chaetocin; and the bromodomain inhibitor JQ-1. The level of luciferase was measured 24 h after treatment with latency reversing agents. The level of reactivation is expressed as fold change over dimethyl sulfoxide (n = 3). (b). The level of reactivation from (a) is expressed as a percentage of maximum stimulation using phorbol myristate acetate/phytohemagglutinin. (c) CD69CD25HLA-DR transitional blasts were infected with pseudotyped-viruses carrying long terminal repeats derived from HIV-infected individuals obtained prior to antiretroviral therapy initiation (P1B11 and P6B7) or after full viral suppression on antiretroviral therapy (P1T10) and treated with the same panel of latency reversing agents. Data represent mean ± SEM from three independent experiments. (d). The level of reactivation from the three patient-derived long terminal repeats is expressed as a percentage of maximum stimulation using phorbol myristate acetate/phytohemagglutinin. (e). Sorted nonblasts infected with NL4-3 pseudotyped-virus were stimulated with romidepsin and the level of luciferase produced at 24 h is expressed as fold change over dimethyl sulfoxide (n = 3). (f) Reactivation is expressed as a percentage of maximum stimulation with phorbol myristate acetate/phytohemagglutinin. * P less than 0.05; ** P less than 0.01; *** P less than 0.001; **** P less than 0.0001.

Since HIV latency was detected in the resting nonblast population (Figs. 3c and 4b), we also tested the responsiveness of these cells to the most potent HDACi, romidepsin, using NL4-3 LTR pseudotyped-virus. Following romidepsin stimulation, there was a 2.1 ± 0.4 fold increase in luciferase activity (Fig. 5e), while PMA/PHA induced a 2.9 ± 0.5 fold increase in luciferase over DMSO control. When expressed as percentage of maximal stimulation with PMA/PHA, romidepsin induced 57.1 ± 8.6% increase in luciferase expression (Fig. 5f).


We developed a novel model of postactivation latency allowing the assessment of patient-derived LTRs in proliferating and nonproliferating T-cell subsets. Using a single round pseudotyped virus, inducible virus was established in CD69CD25HLA-DR T cells that had never proliferated (nonblasts) and in CD69CD25HLA-DR transitional blasts: cells that had proliferated but did not express classical activation markers and retained a blast morphology. Inducible virus was demonstrated in both T-cell subsets using single round virus and a panel of LRAs and pseudotyped viruses with patient-derived LTRs. Although an artificial model, these data suggest that virus expression can be induced from both proliferating and nonproliferating cells that don’t express classical activation markers.

Transitional blasts are not found in the peripheral blood but they are readily detected in the lymph node, especially after antigen challenge such as vaccination [25,26]. In a recent study by Shan et al.[27], a similar transitioning cell was described, where activated CD4+ T cells undergoing effector-to-memory (ETM) transition had downregulation of gene transcription and increased expression of CCR5 with increased permissiveness to latent HIV infection. Similar to the transitional blasts described in our article, these ETM cells had low levels of activation marker expression, had mixed TCM and TEM phenotypes and were transitioning to a smaller cell size. Using a different model, we confirm that latency can be established in cells that have proliferated, have a blast morphology but don’t express activation markers consistent with cells undergoing ETM transition.

Our model is the first to evaluate latency reversal in two distinct cell populations distinguished by activation state, cell size and morphology in primary CD4+ T cells. Other investigators have assessed cell size in models of postactivation latency [13,28] but have not identified cells that are negative for all classical activation markers. In a model using anti-CD3/CD28 activation [28], cell size was observed to decrease, although didn’t fully revert to baseline resting size over an 11-day culture period and there was continued expression of the activation marker CD25. In the model developed by Tyagi et al.[13], cells reverted to a resting size after a culture period of up to 49 days, however CD25 expression continued to persist in these ‘resting’ cells. In our model, we were unable to culture cells for that duration given the high frequency of cell death. Here we found that a subset of both blasts and nonblasts didn’t express any of the activation markers, CD69, CD25 and HLA-DR.

Using the novel pGBFM luciferase reporter construct, we were able to assess the effect of LRAs on patient-derived HIV LTRs in primary CD4+ T cells. Use of patient-derived LTRs may provide a more accurate representation of LRA activity, given the heterogeneity of HIV LTR sequences in vivo. This contrasts to other models of latency that use only laboratory-derived strains. The use of primary cells in our model also allows for random integration following infection, as compared with cell line models which use a monoclonal cell population with a single HIV integration site [29,30]. Our model also provides the potential to evaluate patient-derived HIV LTRs that may be of particular interest for example LTRs isolated from patients in clinical trials shown not to respond to LRAs in vivo.

The lack of expression of the activation markers CD69, CD25 and HLA-DR is often used to define a ‘resting’ cell population in models of postactivation latency and when quantifying the frequency of latent infection on ART ex vivo[14–17]. The latently infected transitional blasts identified in our new model expressed persistent low levels of luciferase, suggesting a downregulated but not fully quiescent transcriptional state. In contrast, the nonblasts had virtually no constitutive luciferase expression. Latently infected cells are often described as being in a state of complete transcriptional dormancy. However, constitutive expression of cell-associated HIV RNA in HIV-infected individuals on ART has been demonstrated by us and other groups [31–33], and expression of viral protein Gag has also been shown to occur in latently infected cells in vitro[34,35]. This model, using single round LTR luciferase constructs that integrate, may allow for further insights into understanding the spectrum of transcriptional activity in HIV latency.

The establishment of latency in the nonblasts was surprising, given prior evidence has suggested resting T cells to be relatively nonpermissive to infection with VSV-G pseudotyped [36] or wild-type virus [20,37]. Infection of nonblasts could potentially have occurred through several mechanisms. First, it is possible that latent infection was established in a subset of nonblasts that were activated but did not proliferate [38]. This is supported by the finding of a comparable proportion of TCM and TTM phenotypes in the nonblasts compared with transitional blasts. Second, coculture of these cells with activated proliferating T cells could have facilitated latent infection either through exposure to secreted soluble factors, or through cell–cell transmission of virus [39,40]. Finally, it is also possible that a proportion of the nonblasts could be T memory stem cells, which possess a CD45RA+CCR7+ phenotype and have been shown to be permissive to infection with VSV-G pseudotyped viruses [41].

We acknowledge that there are several limitations of this model. First we used a single round pseudotyped envelope deleted virus that was unable to induce cell death. We sought to address this by repeating the experiments using a full length EGFP reporter virus. Using this virus we were able to demonstrate inducible EGFP in nonblasts but not in transitional blasts, most likely due to death of the transitional blasts during the prolonged phase of productive infection, when there were no antiretrovirals present. Following stimulation of cells infected with wild-type virus, there would be expression of envelope protein, which would also enhance cell death with syncytia formation even in the presence of raltegravir. In addition, we used different methods to quantify the input virus for the wild-type and pseudotyped viruses and this could potentially explain the difference in findings for the two viruses. Second, we only measured EGFP expression as a marker of productive infection and didn’t directly measure infectious virus, but based on previous reports of this virus, we assume that EGFP expression correlates with infectious virus [24]. Third, we defined latency here as the demonstration of inducible luciferase. In the setting of high constitutive luciferase expression with a virus that is unable to induce cell death, as seen in the transitional blasts, we cannot rule out productive infection. In addition, inducible luciferase expression does not reflect the frequency of infected cells, but rather the frequency of infected cells with inducible virus. We may have potentially underestimated the frequency of latently infected cells if the virus in some cells did not activate following T-cell stimulation, described in patient-derived cells as ‘noninduced intact proviruses’ [42]. This is clearly a limitation of single reporter viruses that express either luciferase or EGFP. Finally, we acknowledge that in blood, there are no circulating cells with features of transitional blasts (as described here) but these cells could potentially be found in tissue such as lymph node.


We show that following T-cell activation, latency is established in nonproliferating and proliferating cells that do not express classical activation markers. Both latently infected T-cell subsets responded to LRAs with similar efficiency. The contribution of each of these T-cell subsets to HIV persistence in vivo in blood and in tissue warrants further evaluation.


M.A.M., J.L.A., L.R.G., S.S., P.U.C., M.J.C., S.R.L., H.K.L. conceived and designed the experiments; M.A.M., J.L.A., S.A., G.K., J.J.C., A.M.E., W.J.C., H.K.L. performed experiments; M.A.M., J.L.A., G.K., J.J.C., J.J.Z. analyzed the data; J.C.J., J.J.Z., D.F.J.P., P.U.C., S.R.L. contributed reagents, materials and analysis tools; M.A.M., S.L.R., H.K.L. wrote the article and all authors reviewed and approved the article.

We thank the staff of the flow cytometry unit at the Peter Doherty Institute for Infection and Immunity for assistance with sorting and analysis by flow cytometry.

Conflicts of interest

The authors declare that they have no competing interests. S.R.L. is an Australian National Health and Medical Research Council (NHMRC) Practitioner Fellow. This work was supported by a program grant from the Australian NHMRC; the National Institute of Allergy and Infectious Disease (NIAID), US National Institutes of Health (NIH) (Delaney AIDS Research Enterprise, DARE; U19AI096109 and RFA-AI-15-029) and the American Foundation for AIDS Research.


1. Chun TW, Stuyver L, Mizell SB, Ehler LA, Mican JA, Baseler M, et al. Presence of an inducible HIV-1 latent reservoir during highly active antiretroviral therapy. Proc Natl Acad Sci U S A 1997; 94:13193–13197.
2. Finzi D, Blankson J, Siliciano JD, Margolick JB, Chadwick K, Pierson T, et al. Latent infection of CD4+ T cells provides a mechanism for lifelong persistence of HIV-1, even in patients on effective combination therapy. Nat Med 1999; 5:512–517.
3. Bui JK, Sobolewski MD, Keele BF, Spindler J, Musick A, Wiegand A, et al. Proviruses with identical sequences comprise a large fraction of the replication-competent HIV reservoir. PLoS Pathog 2017; 13:e1006283.
4. Hosmane NN, Kwon KJ, Bruner KM, Capoferri AA, Beg S, Rosenbloom DI, et al. Proliferation of latently infected CD4+ T cells carrying replication-competent HIV-1: potential role in latent reservoir dynamics. J Exp Med 2017; 214:959–972.
5. Lorenzi JC, Cohen YZ, Cohn LB, Kreider EF, Barton JP. Paired quantitative and qualitative assessment of the replication-competent HIV-1 reservoir and comparison with integrated proviral DNA. Proc Natl Acad Sci U S A 2016; 113:E7908–E7916.
6. Pitman MC, Lau JSY, McMahon JH, Lewin SR. Barriers and strategies to achieve a cure for HIV. Lancet HIV 2018; 5:e317–e328.
7. Chen HC, Martinez JP, Zorita E. Position effects influence HIV latency reversal. Nat Struct Mol Biol 2017; 24:47–54.
8. Lassen KG, Hebbeler AM, Bhattacharyya D, Lobritz MA, Greene WC. A flexible model of HIV-1 latency permitting evaluation of many primary CD4 T-cell reservoirs. PLoS One 2012; 7:e30176.
9. Saleh S, Wightman F, Ramanayake S, Alexander M, Kumar N, Khoury G, et al. Expression and reactivation of HIV in a chemokine induced model of HIV latency in primary resting CD4+ T cells. Retrovirology 2011; 8:80.
10. Swiggard WJ, Baytop C, Yu JJ, Dai J, Li C, Schretzenmair R, et al. Human immunodeficiency virus type 1 can establish latent infection in resting CD4+ T cells in the absence of activating stimuli. J Virol 2005; 79:14179–14188.
11. Evans VA, Kumar N, Filali A, Procopio FA, Yegorov O, Goulet JP, et al. Myeloid dendritic cells induce HIV-1 latency in nonproliferating CD4+ T cells. PLoS Pathog 2013; 9:e1003799.
12. Bosque A, Planelles V. Studies of HIV-1 latency in an ex vivo model that uses primary central memory T cells. Methods 2011; 53:54–61.
13. Tyagi M, Pearson RJ, Karn J. Establishment of HIV latency in primary CD4+ cells is due to epigenetic transcriptional silencing and P-TEFb restriction. J Virol 2010; 84:6425–6437.
14. Marini A, Harper JM, Romerio F. An in vitro system to model the establishment and reactivation of HIV-1 latency. J Immunol 2008; 181:7713–7720.
15. Sahu GK, Lee K, Ji J, Braciale V, Baron S, Cloyd MW. A novel in vitro system to generate and study latently HIV-infected long-lived normal CD4+ T-lymphocytes. Virology 2006; 355:127–137.
16. Contreras X, Barboric M, Lenasi T, Peterlin BM. HMBA releases P-TEFb from HEXIM1 and 7SK snRNA via PI3K/Akt and activates HIV transcription. PLoS Pathog 2007; 3:1459–1469.
17. Yang HC, Xing S, Shan L, O’Connell K, Dinoso J, Shen A, et al. Small-molecule screening using a human primary cell model of HIV latency identifies compounds that reverse latency without cellular activation. J Clin Invest 2009; 119:3473–3486.
18. Cuevas JM, Geller R, Garijo R, Lûpez-Aldeguer J, Sanjun R. Extremely high mutation rate of HIV-1 in vivo. PLoS Biol 2015; 13:e1002251.
19. Rezaei SD, Lu HK, Chang JJ, Rhodes A, Lewin SR, Cameron PU. The pathway to establishing HIV latency is critical to how latency is maintained and reversed. J Virol 2018; 92:e02225–e02317.
20. Saleh S, Solomon A, Wightman F, Xhilaga M, Cameron PU, Lewin SR. CCR7 ligands CCL19 and CCL21 increase permissiveness of resting memory CD4+ T cells to HIV-1 infection: a novel model of HIV-1 latency. Blood 2007; 110:4161–4164.
21. Lu HK, Gray LR, Wightman F, Ellenberg P, Khoury G, Cheng WJ, et al. Ex vivo response to histone deacetylase (HDAC) inhibitors of the HIV long terminal repeat (LTR) derived from HIV-infected patients on antiretroviral therapy. PLoS One 2014; 9:e113341.
22. Gray LR, On H, Roberts E, Lu HK, Moso MA, Raison JA, et al. Toxicity and in vitro activity of HIV-1 latency-reversing agents in primary CNS cells. J Neurovirol 2016; 22:455–463.
23. Anderson JL, Mota TM, Evans VA, Kumar N, Rezaei SD, Cheong K, et al. Understanding factors that modulate the establishment of HIV latency in resting CD4+ T-cells in vitro. PLoS One 2016; 11:e0158778.
24. Yamamoto T, Tsunetsugu-Yokota Y, Mitsuki YY, Mizukoshi F, Tsuchiya T, Terahara K, et al. Selective transmission of R5 HIV-1 over X4 HIV-1 at the dendritic cell-T cell infectious synapse is determined by the T cell activation state. PLoS Pathog 2009; 5:e1000279.
25. Johansen P, Haffner AC, Koch F, Zepter K, Erdmann I, Maloy K, et al. Direct intralymphatic injection of peptide vaccines enhances immunogenicity. Eur J Immunol 2005; 35:568–574.
26. Rush C, Mitchell T, Garside P. Efficient priming of CD4+ and CD8+ T cells by DNA vaccination depends on appropriate targeting of sufficient levels of immunologically relevant antigen to appropriate processing pathways. J Immunol 2002; 169:4951–4960.
27. Shan L, Deng K, Gao H, Xing S, Capoferri AA, Durand CM, et al. Transcriptional reprogramming during effector-to-memory transition renders CD4+ T cells permissive for latent HIV-1 infection. Immunity 2017; 47:766–775.e3.
28. Chavez L, Calvanese V, Verdin E. HIV latency is established directly and early in both resting and activated primary CD4 T cells. PLoS Pathog 2015; 11:e1004955.
29. Jordan A, Bisgrove D, Verdin E. HIV reproducibly establishes a latent infection after acute infection of T cells in vitro. EMBO J 2003; 22:1868–1877.
30. Jordan A, Defechereux P, Verdin E. The site of HIV-1 integration in the human genome determines basal transcriptional activity and response to Tat transactivation. EMBO J 2001; 20:1726–1738.
31. Lewin SR, Vesanen M, Kostrikis L, Hurley A, Duran M, Zhang L, et al. Use of real-time PCR and molecular beacons to detect virus replication in human immunodeficiency virus type 1-infected individuals on prolonged effective antiretroviral therapy. J Virol 1999; 73:6099–6103.
32. Elliott JH, Wightman F, Solomon A, Ghneim K, Ahlers J, Cameron MJ, et al. Activation of HIV transcription with short-course vorinostat in HIV-infected patients on suppressive antiretroviral therapy. PLoS Pathog 2014; 10:e1004473.
33. Pasternak AO, DeMaster LK, Kootstra NA, Reiss P, O’Doherty U, Berkhout B. Minor contribution of chimeric host-HIV readthrough transcripts to the level of HIV cell-associated gag RNA. J Virol 2016; 90:1148–1151.
34. Pace MJ, Graf EH, Agosto LM, Mexas AM, Male F, Brady T, et al. Directly infected resting CD4+T cells can produce HIV Gag without spreading infection in a model of HIV latency. PLoS Pathog 2012; 8:e1002818.
35. Passaes CP, Bruel T, Decalf J, David A, Angin M. Ultrasensitive HIV-1 p24 assay detects single infected cells and differences in reservoir induction by latency reversal agents. J Virol 2017; 91:e02296–e02316.
36. Yu D, Wang W, Yoder A, Spear M, Wu Y. The HIV envelope but not VSV glycoprotein is capable of mediating HIV latent infection of resting CD4 T cells. PLoS Pathog 2009; 5:e1000633.
37. Dai J, Agosto LM, Baytop C, Yu JJ, Pace MJ, Liszewski MK, et al. Human immunodeficiency virus integrates directly into naive resting CD4+ T cells but enters naive cells less efficiently than memory cells. J Virol 2009; 83:4528–4537.
38. Katzen D, Chu E, Terhost C, Leung DY, Gesner M, Miller RA, et al. Mechanisms of human T cell response to mitogens: IL 2 induces IL 2 receptor expression and proliferation but not IL 2 synthesis in PHA-stimulated T cells. J Immunol 1985; 135:1840–1845.
39. Skinner AM, Chakkaramakkil Verghese S, Kurre P. Cell-cell transmission of VSV-G pseudotyped lentivector particles. PLoS One 2013; 8:e74925.
40. Pan YW, Scarlett JM, Luoh TT, Kurre P. Prolonged adherence of human immunodeficiency virus-derived vector particles to hematopoietic target cells leads to secondary transduction in vitro and in vivo. J Virol 2007; 81:639–649.
41. Buzon MJ, Sun H, Li C, Shaw A, Seiss K, Ouyang Z, et al. HIV-1 persistence in CD4+ T cells with stem cell-like properties. Nat Med 2014; 20:139–142.
42. Ho YC, Shan L, Hosmane NN, Wang J, Laskey SB, Rosenbloom DI, et al. Replication-competent noninduced proviruses in the latent reservoir increase barrier to HIV-1 cure. Cell 2013; 155:540–551.

HIV; latency; latency reversing agents; nonproliferating; proliferating; transitional blasts

Supplemental Digital Content

Copyright © 2019 Wolters Kluwer Health, Inc.