ORIGINAL RESEARCH published: 22 January 2019 doi: 10.3389/fimmu.2018.03162 Antigen Production After Latency Reversal and Expression of Inhibitory Receptors in CD8+ T Cells Limit the Killing of HIV-1 Reactivated Cells Alba Ruiz 1,2 , Oscar Blanch-Lombarte 1,2 , Esther Jimenez-Moyano 1 , Dan Ouchi 1 , Beatriz Mothe 1,3 , Ruth Peña 1 , Cristina Galvez 1,2 , Meritxell Genescà 4 , Javier Martinez-Picado 1,3,5 , Philip Goulder 6 , Richard Barnard 7 , Bonnie Howell 7 , Bonaventura Clotet 1,3 and Julia G. Prado 1,2* 1 IrsiCaixa AIDS Research Institute, Badalona, Spain, 2 Germans Trias i Pujol Research Institute (IGTP), Universitat Autonoma de Barcelona, Badalona, Spain, 3 Faculty of Medicine, University of Vic - Central University of Catalonia (UVic-UCC), Vic, Spain, 4 Department of Infectious Diseases, Hospital Universitari Vall d’Hebrón, Institut de Recerca (VHIR), Universitat Autònoma de Barcelona, Barcelona, Spain, 5 Catalan Institution for Research and Advanced Studies (ICREA), Barcelona, Spain, 6 Department of Paediatrics, University of Oxford, Oxford, United Kingdom, 7 Department of Infectious Disease, Merck & Co. Inc., Kenilworth, NJ, United States Edited by: Francesca Chiodi, Karolinska Institute (KI), Sweden The so-called shock and kill therapies aim to combine HIV-1 reactivation by Reviewed by: latency-reversing agents (LRA) with immune clearance to purge the HIV-1 reservoir. Elisa Vicenzi, The clinical use of LRA has demonstrated detectable perturbations in the HIV-1 San Raffaele Hospital (IRCCS), Italy reservoir without measurable reductions to date. Consequently, fundamental questions Sarah Rowland-Jones, University of Oxford, United Kingdom concerning the limitations of the recognition and killing of LRA-reactivated cells by *Correspondence: effector cells such as CD8+ T cells remain to be answered. Here, we developed a Julia G. Prado novel experimental framework where we combine the use of cytotoxic CD8+ T-cell jgarciaprado@irsicaixa.es lines and ex vivo CD8+ T cells from HIV-1-infected individuals with functional assays of Specialty section: LRA-inducible reactivation to delineate immune barriers to clear the reservoir. Our results This article was submitted to demonstrate the potential for early recognition and killing of reactivated cells by CD8+ T Viral Immunology, cells. However, the potency of LRAs when crossing the barrier for antigen presentation in a section of the journal Frontiers in Immunology target cells, together with the lack of expression of inhibitory receptors in CD8+ T cells, Received: 14 September 2018 are critical events to maximize the speed of recognition and the magnitude of the killing Accepted: 21 December 2018 of LRA-inducible provirus. Taken together, our findings highlight direct limitations in LRA Published: 22 January 2019 potency and CD8+ T cell functional status to succeed in the cure of HIV-1 infection. Citation: Ruiz A, Blanch-Lombarte O, Keywords: human immunodeficiency virus, HIV-1 reservoir, HIV-1 immunogen, shock and kill, CTL (Cytotoxic T Jimenez-Moyano E, Ouchi D, lymphocyte), inhibitory receptors Mothe B, Peña R, Galvez C, Genescà M, Martinez-Picado J, Goulder P, Barnard R, Howell B, INTRODUCTION Clotet B and Prado JG (2019) Antigen Production After Latency Reversal and The introduction of antiretroviral treatment (ART) has increased life expectancy and improved the Expression of Inhibitory Receptors in CD8+ T Cells Limit the Killing of HIV-1 health of people living with HIV-1 infection. However, ART does not cure HIV-1 infection, and Reactivated Cells. treatment is needed for decades, thus raising the alarm about long-term health and sustainability of Front. Immunol. 9:3162. treatment and care of HIV-1-infected individuals (1). The current limitations of ART highlight the doi: 10.3389/fimmu.2018.03162 need for therapeutic strategies to eradicate HIV-1 from the body. A major obstacle to eradication is Frontiers in Immunology | www.frontiersin.org 1 January 2019 | Volume 9 | Article 3162 Ruiz et al. Barriers to HIV-1 Shock and Kill the establishment of the HIV-1 reservoir (2). Besides, between LRA potency and the speed and magnitude of the the reservoir is perpetuated through cellular homeostatic killing of reactivated cells by CD8+ T cells. Although we found proliferation and clonal expansion, even after years of effective increased killing of reactivated cells by ex vivo CD8+ T cells ART (3, 4). in response to LRA, the magnitude of the response was highly One therapeutic approach, the so-called shock and kill variable across HIV-1-infected individuals and was associated strategies, has been tested in several clinical trials and proposes with a lack of expression of inhibitory receptors in CD8+ T cells. the use of latency-reversing agents (LRA) for the transcriptional Our data highlight several limitations in the efficacy of shock and activation of HIV-1 (shock) and the clearance of reactivated kill strategies and point to the need for a trade-off between LRA cells by immune responses (kill). Although several studies have potency and CD8+ T-cell functional status in HIV-1-infected demonstrated reactivation of HIV-1 by LRA both in vitro individuals if the reservoir is to be cleared. and in vivo (5–7), no measurable reduction in the HIV-1 reservoir has been found to date (8). Consequently, ensuring the immune recognition of LRA-reactivated cells by effector RESULTS responses will be essential for eradication of the HIV-1 reservoir LRA Allow HIV-1 Protein Expression and (9, 10). Several studies have proposed CD8+ T cells as effector cells for recognition and clearance of LRA-reactivated cells HLA-Class I Antigen Presentation for (11) based on their ability to control the reservoir size in CD8+ T-Cell Recognition to Increase natural controllers (12–14), their potent in vitro antiviral activity Killing of Latently Infected Cells (15, 16), and their role in controlling viral replication despite First, we developed the “resting-like” or RELI-model to evaluate ART (17). Although the frequency of HIV-1–specific CD8+ HIV-1 reactivation by LRA (shock) simultaneously with the T cells decays with ART (18, 19), the cells retain effector elimination of reactivated cells by HIV-1-specific cytotoxic and cytotoxic properties that enable them to recognize and CD8+ T-cell lines (CTL) (kill), as schematized in Figure 1A and kill HIV-1-infected cells (11, 17, 20, 21). However, functional detailed in the Materials and Methods section. Briefly, we infected barriers to CD8+ T-cell antiviral activity upon treatment with U937-HLA-B∗ 27:05 cells with the HIV-1NL43GFP reporter virus as LRA can affect the success of shock and kill strategies. These previously reported (31, 32). After 72 h of infection, we sorted the barriers may be associated with CD8+ T-cell dysfunction, GFP negative cells and cultured them for 4 days in the presence of which is a consequence of LRA treatment itself, and with the protease inhibitor ritonavir (RTV). At this point, we obtained the pro-inflammatory environment driven by HIV-1 infection. a RELI population enriched in uninfected and non-productively Several studies suggest an immunosuppressive effect of LRA, HIV-1-infected cells. RELI cells were then washed, treated with particularly histone deacetylase inhibitors (HDACi), on CD8+ the HIV-1 integrase inhibitor raltegravir (RAL) to avoid multiple T-cell antiviral activity (22, 23). Data remain controversial, and rounds of infection, and exposed to LRA shock for 48 h. We while some studies suggest a time-dependent or direct effect also treated RELI cells with azidothymidine (AZT) to avoid the of HDACi on CD8+ T-cell function (24), others do not find formation of unintegrated episomal forms that may contribute a measurable impact on ex vivo CD8+ T-cell function after to the expression of viral proteins (data not shown) (33). LRA- in vivo administration of HDACi (7). Moreover, the chronic treated cells were then extensively washed to uncouple the LRA pro-inflammatory environment and the persistence of antigen effect from CD8+ T-cell function against reactivated cells (24), exposure affect the functional profile of HIV-1–specific CD8+ T- and co-cultured with HLA-class I–matched HIV-1-specific CTLs cell responses (25, 26). This pro-inflammatory environment leads for 20 h (kill) (Figure 1A). to the reduction of cytotoxic potential and the upregulation of We measured HIV-1 reactivation in RELI cells in response inhibitory receptors, such as PD-1, LAG-3, and TIM-3 in CD8+ to LRA for 48 h including the HDACi panobinostat (PNBN), T cells associated with dysfunction and immune exhaustion in romidepsin (RMD), trichostatin A (Tricho), and suberoylanilide HIV-1-infected individuals (27–30). hydroxamic acid (SAHA). We also included combinations of In this context, fundamental questions regarding the HDACi and bryostatin A (Bryo) to evaluate potential additive limitations of LRA activity on target cells and CD8+ T-cell effects between HDACi and protein kinase C agonists (PKCag) sensing in response to HIV-1 reactivation remain unanswered. as previously reported (34). Drug concentrations were selected In this study, we design a novel experimental framework where based on previous clinical or in vitro studies (7, 35–37) and we combine cytotoxic HIV-1 CD8+ T-cell lines (CTL) and in the absence of direct cellular toxicity (Table S1). The use ex vivo CD8+ T cells from HIV-1-infected individuals with an of HIV-1NL43GFP enabled us to monitor viral reactivation by in vitro model of LRA-dependent HIV-1 reactivation. In this flow cytometry using two markers the expression of p24 protein framework, we evaluate the so-called window of opportunity (Figure 1B) and the expression of GFP encoded in frame with the between latency reversal and killing of reactivated cells by HIV-1 Nef protein, as previously reported (38, 39) (Figure 1C). CD8+ T cells. We characterize HIV-1 protein expression upon We observed a direct correlation between HIV-1 fold induction treatment with LRA and its association with antigen presentation measured according to p24 or GFP expression (p < 0.0001; and delineate the kinetics of recognition and killing of HIV-1 ρ = 0.94, Figure 1D). In this context, the expression of p24 was reactivated cells by CD8+ T cells. We also analyze the functional consistently higher than that of GFP (p24 vs. GFP p < 0.005, limitations of CD8+ T cells from HIV-1-infected individuals in Wilcoxon matched-pairs signed-rank test) (Figure 1E), thus the elimination of reactivated cells. We observed a correlation indicating its greater sensitivity as a marker for viral antigen Frontiers in Immunology | www.frontiersin.org 2 January 2019 | Volume 9 | Article 3162 Ruiz et al. Barriers to HIV-1 Shock and Kill FIGURE 1 | The HIV-1 RELI “shock and kill”model. (A) Schematic representation of the HIV-1 RELI model for “shock and kill.” RELI cells were reactivated with LRA in the presence of RAL before being extensively washed and cultured in the presence or absence of HLA-class I–matched CTL. After 20 h of co-culture, we evaluatedHIV-1 reactivation (shock) and the killing of HIV-reactivated by CTL (kill). (B) Representative dot plots of intracellular p24 expression in live RELI cells under untreated conditions (UT) compared with PMA/Ionomycin and SAHA and PNBN alone or combined with Bryo. (C) Representative dot plots of intracellular expression of p24 and GFP from live RELI cells reactivated with LRA. The green square indicates GPF- positive cells, and the red square indicates p24-positive cells. (D) Correlation between HIV-1 fold induction measured according to the expression of p24 or GFP in RELI cells after LRA treatment. The HIV-1 fold induction was calculated as the ratio between % of p24- or GFP-expressing cells under LRA conditions and the % of p24- or GFP- expressing cells left untreated. The line indicates the fit of the data to a linear regression. The Spearman correlation coefficient (ρ) and the two-tailed p-value are shown. The graph represents the mean ± SEM of three independent experiments performed in replicates. (E) HIV-1 reactivation measured by the percentage of p24 positive cells or GFP-positive cells. The graph represents the mean ± SEM of three independent experiments performed in replicates. (F) HIV-1 fold induction of RELI cells. The HIV-1-fold induction was calculated as the ratio between the % of p24 expressing cells under LRA conditions and the % of p24-expressing cells in the left untreated. The data represents the mean ± SEM of five independent experiments performed in replicates. The p–values were calculated using the one-sample t-test. Only significant values are shown in the figure (*p < 0.05, **p < 0.005). Frontiers in Immunology | www.frontiersin.org 3 January 2019 | Volume 9 | Article 3162 Ruiz et al. Barriers to HIV-1 Shock and Kill FIGURE 2 | LRA treatment increased the magnitude of killing of HIV-1 reactivated cells by CD8+ T cells. (A) Representative dot plots showing the frequency of p24 cells from live HIV-1 reactivated cells in the absence (top) or in the presence of CTL1 for 20 h. (B) Frequency of p24 cells from HIV-1 reactivated cells inthe absence or presence of CTLs (CTL1 and CTL2). The graph represents the mean ± SEM of three independent experiments performed in duplicate. (C) Fold-p24 expression relative to SAHA-reactivated cells in the presence of HLA-matched CTLs treated with isotype control antibody IgG. SAHA-reactivated cells in the absence or presence of HLA-matched CTLs (CTL1, CTL2), treated isotype control antibody IgG (IgG), and HLA-blocking antibody (W6/32), and in the presence of HLA-mismatched CTL3 (MM). The graph represents the mean ± SD of three experiments performed in duplicate. (D) Relative killing of RELI cells by untreated CTL1 and CTL2 and presence of HDACi (including PNBN and SAHA) alone and combined with Bryo. Box and whisker plots included data from four independent experiments performed in duplicate. The p-values were calculated using the Mann-Whitney test. Only significant values are shown in the figure (*p < 0.05). Frontiers in Immunology | www.frontiersin.org 4 January 2019 | Volume 9 | Article 3162 Ruiz et al. Barriers to HIV-1 Shock and Kill expression upon reactivation and supporting its application in monitored the kinetics of recognition and elimination of HIV- subsequent experiments. However, the use of GFP expression as 1 reactivated cells by CTLs at 3, 6, and 20 h post co-culture. a marker of HIV-1 induction is particularly relevant given the We measured changes in the frequency of p24-positive cells correlation between p24 and GFP expression for the screening from time 0 by flow cytometry at 3, 6, and 20 h after co- of large compound libraries and, thus, the increased cost- culture of untreated and LRA-treated RELI cells with CTLs effectiveness of the method. (Figure 3A). We observed a reduction in the frequency of p24- As shown in Figure 1F, we compared HIV-1 reactivation expressing cells over time in culture with CTLs (Figure 3B). HIV- after LRA treatment with Tricho, PNBN, and SAHA alone or 1 reactivated cells were eliminated rapidly after 3 h of co-culture combined with Bryo with untreated or DMSO controls. The in SAHA, PNBN, SAHA/Bryo, and PNBN/Bryo conditions levels of HIV-1 reactivation ranged between a 2- and 4-fold viral compared with untreated cells for both CTLs (untreated vs. induction, with a significant increase for PNBN and SAHA alone HDAC ± Bryo; p < 0.0001) (Figure 3C). These differences were (p < 0.05) or combined with Bryo (p < 0.005). The additive sustained at 6 h after co-culture (untreated vs. HDAC ± Bryo; effects of combining HDACi/Bryo were modest or absent. p < 0.0005) and 20 h after co-culture (untreated vs. HDAC ± Next, we assessed whether LRA treatment enabled HIV-1 Bryo; p < 0.005) (Figures 3D,E). These data indicate a general antigen presentation in the context of HLA-class I molecules for increase in the speed of recognition and killing of reactivated cells their recognition and killing by HIV-1 cytotoxic CD8+ T-cell by CTLs starting at 3 h in the presence of PNBN, PNBN/Bryo, lines (CTL). For this purpose, we used two CTLs, CTL1, and SAHA, and SAHA/Bryo, although our results do not reveal a CTL2, which recognized HLA-B∗ 27:05 restricted epitopes (Gag hierarchy for LRAs. The data suggest a prolonged effect over KK10 and Pol KY9, respectively) as previously described (31). the first 6 h by PNBN/Bryo, SAHA, and SAHA/Bryo compared Briefly, we exposed RELI cells to PNBN and SAHA alone or with untreated cells (Figures 3D,E) (untreated vs. PNBN/Bryo; combined with Bryo and co-cultured them with CTLs restricted p < 0.05, untreated vs. SAHA, p < 0.005; and untreated vs. by HLA-B∗ 27:05. After 20 h of co-culture, we assessed killing by SAHA/Bryo; p < 0.005). By contrast, the use of Bryo alone measuring the frequency of p24-expressing cells in the absence did not significantly increase the speed or magnitude of killing or presence of CTL1 (Figure 2A and Figure S1). As shown compared with untreated condition at 3, 6, or 20 h of co-culture in Figure 2B, we observed significantly increased killing by (Figures 3C–E). Although we observed a marginal augment in CTL1 and CTL2 after LRA treatment (p < 0.0005, one-way the killing at early time points, suggesting an indirect effect of ANOVA). In fact, CTL1 and CTL2 were highly functional and Bryo in CTL activation despite extensive washout of the drug as had similar 50% effective concentrations (EC50 , −4.8, and −5.5, previously reported (24). respectively), despite differences in epitope specificity and a Interestingly, as shown in Figure 3F, we found a direct higher response for CTL1 at low HIV-1 antigen concentrations correlation between the magnitude of CTL killing and the levels (Figure S2). of HIV-1 inducible reactivation by LRA at early time points (3 h In addition, we added an anti-HLA-A/B/C (W6/32) antibody p < 0.0005, ρ = 0.74; 6 h p < 0.005, ρ = 0.68), thus indicating or an HLA-mismatched (MM) CTL3 to demonstrate that, upon that the potency of the LRA for HIV-1 antigen expression dictates viral reactivation, CTL killing was mediated through contacts the speed of CTL recognition. Based on these data, we propose with class I HLA HIV-1 antigen complexes. The elimination of a threshold of >2-fold HIV-1 induction upon LRA treatment reactivated cells was significantly abolished both by means of the to allow antigen presentation and engage rapid recognition and W6/32 antibody and by impeding the recognition of the HLA- killing of HIV-1 reactivated cells by CTL. class I peptide complexes with an HLA-MM co-culture (p < 0.05, Besides, we investigated whether HIV-1 killing by CTL can be one-way ANOVA) (Figure 2C). Thus, the compiled analysis of monitored by changes in cytokine expression and degranulation all LRA tested showed a significant increase in the magnitude of upon antigen recognition using CTL as biosensors (40), we killing of cells latently infected by HIV-1 after LRA treatment in measured CD107a/MIP1β and IFNγ expression (Figure 3G, dot both CTL (p < 0.05, Figure 2D). However, we did not observe plots). Despite the increase of HIV-1 expression by LRAs, the use any additional benefit in the frequency of reactivated cell killing of SAHA or PNBN did not alter the kinetics of CD107a/MIP1β or through the combination of HDACi with Bryo (Figure 2D). IFNγ expression in the tested CTLs, as compared with untreated These data confirm that inducible reactivation of HIV-1 by conditions. By contrast, Bryo indirect effects increased both LRA enables antigen presentation in the context of HLA-class I CD107a/MIP1β and IFNγ secretion by CTLs at early time points molecules to increase killing of latently infected cells by CTL. (Figure 3H). We also found different profiles for CD107a/MIP1β and IFNγ expression over time, with increasing frequencies for LRA Potency Modulates the Speed of CD107a/MIP1β and decreasing frequencies for IFNγ secretion. Thus, the kinetics of cytokine expression by CTLs did not mirror HIV-1 Antigen Recognition and the changes associated with HIV-1 inducible reactivation or killing. Magnitude of the Killing of Reactivated Overall, these data indicate a direct relationship between the Cells by CD8+ T Cells potency of inducible reactivation of HIV-1 by LRA and the speed It is essential to investigate the kinetics of recognition and of recognition and killing of reactivated cells by CD8+ T cells. killing of LRA-reactivated cells and to establish a window Therefore, the potency of LRA to cross the threshold of HIV- of opportunity between shock and kill by CD8+ T cells 1 antigen presentation in reactivated cells is a crucial event to for therapeutic optimization. By using the RELI model, we shortening the window of opportunity for antigen recognition, Frontiers in Immunology | www.frontiersin.org 5 January 2019 | Volume 9 | Article 3162 Ruiz et al. Barriers to HIV-1 Shock and Kill FIGURE 3 | Kinetics of killing in response to HIV-1 reactivation by CTL. (A) Representative dot plots showing the percentages of intracellular p24 from HIV-1 RELI cells. The top line indicate the level of HIV-1 reactivation of RELI cells at baseline (0 h) in untreated and LRA-treated conditions and 3, 6, and 20 h after co-culture with CTL1 (B) Kinetics of elimination of HIV-1 reactivated by CTL. The graph shows the mean ± SEM of two independent experiments for CTL1. (C) Kinetics of killing of HIV-1 reactivated cells by CTLs at 3 h; (D) at 6 h; (E), and at 20 h after co-culture. Values in graphs C to E correspond to the mean ± SEM of three independent experiments for CTL1 and three independent experiments for CTL2 in replicates. The p-values were calculated using the unpaired t-test for comparisons between untreated cells and cells treated with HDACi (top line) and one-way ANOVA for multiple comparisons (asterisks above the bars). Only significant p-values are represented in the figure (*p < 0.05, **p < 0.005, ***p < 0.0005, ****p < 0.00005). (F) Correlation between HIV-1 fold induction and frequency of killing mediated by CTLs. The line indicates the fit of the data to a linear regression. The Spearman correlation coefficient (ρ) and the two-tailed p-value are indicated. The graph represents the frequency of killing from three independent experiments for CTL1 in replicates and three independent experiments for CTL2 in replicates. (G) Representative dot plots showing percentages of CD107a/MIP1β from CTL1 after co-culture with HIV-1 reactivated RELI cells. (H) Kinetics of CD107a/MIP1β and IFNγ secretion from CTLs after co-culture with HIV-1 reactivated RELI cells. The graphs show the mean ± SEM of two independent experiments for CTL1 and two independent experiments for CTL2 in replicates. Frontiers in Immunology | www.frontiersin.org 6 January 2019 | Volume 9 | Article 3162 Ruiz et al. Barriers to HIV-1 Shock and Kill thus maximizing CD8+ T-cell sensing for elimination of HIV-1 functional and sensitive to HIV-1 antigen presentation and may reactivated cells. not fully recapitulate the diversity of HIV-1–specific CD8+ T cells in infected individuals. To delineate the kinetics of recognition and killing of LRA- Inter-individual Divergences in ex vivo reactivated cells by CD8+ T cells, we initially co-cultured CD8+ T-Cell Recognition and Killing of SAHA-treated cells with ex vivo CD8+ T cells obtained from LRA-Reactivated Cells three of the eight HIV-1-infected individuals tested (participants Although the use of CTL supports rapid recognition of HIV-1 characteristics in Table 1). As shown in Figure 4A, we observed reactivated cells based on LRA potency, these CTL are highly variation in the kinetics of recognition and killing of reactivated FIGURE 4 | Killing of HIV-1 reactivated cells by ex vivo CD8+ T cells. (A) Percentage of killing of untreated or SAHA-treated HIV-1 RELI cells at 3 and 20 h by ex vivo CD8+ T cells from HIV-1-infected individuals (n = 3). (B) Percentage of killing of untreated or SAHA-treated HIV-1 RELI cells by ex vivo CD8+ T cells at 20 h (n = 7). (C) Paired comparison of the percentage of killing of untreated or SAHA-treated HIV-1 cells by ex vivo CD8+ T cells shown in (B). The graph represents the percentage of the killing of RELI cells. The p-values were calculated using the paired t-test. (D) Percentage of killing of untreated or SAHA-treated RELI cells by ex vivo CD8+ T cells from PT3. Conditions with a 1X SAHA and 2X SAHA and co-cultures at an E:T ratio of 3:1 in 1X SAHA conditions are shown. (E) Percentage of CD8+ T cells in contact with untreated or SAHA-treated HIV-1 RELI cells in the absence or presence of a W6/32 antibody after 20 h of co-culture. The p-values were calculated using the paired t-test. (F) Percentage of CD107a/MIP1β and IFNγ secretion by CD8+ T cells after 20 h in contact or not (CT/non-CT) with SAHA-reactivated cells. The p–values were calculated using the paired t-test. Only significant values (*p < 0.05, **p < 0.005 and ***p < 0.0005) are shown. Frontiers in Immunology | www.frontiersin.org 7 January 2019 | Volume 9 | Article 3162 Ruiz et al. Barriers to HIV-1 Shock and Kill TABLE 1 | Characteristics of study participants. Participant ID Age at sample Gender, Ethnicity ART regimen HLA-B serotype VL (copies/ml) PT1 34 M, CAU TDF/FTC/RAL B*27:05 <40 PT2 43 F, CAU Naïve B*27:02 <40 PT3 36 M, CAU ABC/3TC/DTG B*27:05 <40 PT4 38 M, CAU TDF/FTC/RAL B*27:05 <40 PT5 36 M, CAU Naïve B*27:05 44000 PT6 40 M, CAU TDF/FTC/ELV/COB B*27:05 <40 PT7 37 F, CAU Naïve B*27:05 982 MM 47 M, CAU AZT/DDC Non-B*27 <40 VL, viral load; MM, mismatch; TDF, tenofovir; FTC, emtricitabine; RAL, raltegravir; ABC, abacavir; 3TC, lamivudine; DTG; dolutegravir; ELV, elvitegravir; COB, cobicistat; AZT, zidovudine; DDC, zalcitabine. cells by ex vivo CD8+ T cells. While we observed killing in Expression Profiles of Inhibitory Receptors response to SAHA at 3 h for PT1, we did not observe elimination in HIV-1–Specific CD8+ T Cells Modulates of reactivated cells for PT2; in addition, elimination of reactivated the Killing of HIV-1 Reactivated Cells cells was only observed after 20 h of co-culture for PT3. These Inter-individual differences in functional ability to recognize divergences suggest functional differences between CD8+ T cells and eliminate HIV-1-infected cells ex vivo by CD8+ T cells from HIV-1-infected individual associated a particular profile of could be the result of differences in time to initiation of recognition and killing of LRA-reactivated cells (41). ART (42), in the persistence of antigen exposure and in the Next, we extended our initial analyses to an additional group chronic pro-inflammatory environment despite ART (30). The of HIV-1-infected individuals (Table 1). As in the previous pro-inflammatory environment and the long-term exposure experiments, we found high inter-individual variability in the to antigen lead to dysfunctional HIV-1–specific CD8+ T-cell killing capacity of HIV-1-reactivated cells by CD8+ T cells responses and eventually to immune exhaustion. ranging from 2% (PT2 and PT7) to 70% (PT4 and PT5) (Figure 4B). The killing observed was specific for CD8+ T- To delineate the determinants of CD8+ T-cell dysfunction cell recognition in the context of HLA-class I presentation, as in response to LRA-reactivated cells, we cultured CTLs with supported by the lack of killing observed by mismatch CD8+ HIV-1 cognate peptide for 9 weeks with weekly rounds of T cells (MM). Despite this biological divergence, we observed a peptide re-stimulation to simulate in vitro long-term antigen significant increase in the killing of SAHA-treated cells by CD8+ exposure (Figure 5A). At weeks 0, 5, 7, and 9 we performed T cells (p < 0.05, Figure 4C). Moreover, we were able to enhance an immune phenotype of inhibitory receptors (PD-1, LAG3, the magnitude of CD8+ T-cell killing ex vivo, either by increasing and TIM-3) and the immune-metabolic marker CD39. All these the concentration of SAHA 2-fold or the effector-to-target ratio markers have been previously associated with the exhaustion 3 times (Figure 4D). These results demonstrate the possibility of of HIV-1–specific CD8+ T-cell responses (28, 43). As shown modulating and improving the efficacy of CD8+ T-cell killing in in Figure 5B, we found a significant reduction in the killing response to HIV-1 inducible expression by higher LRA dosage or of LRA-reactivated cells by CTL after 9 weeks of continuous higher frequency of effector cells. We also carried out a detailed antigen exposure (p < 0.005). Despite the overall reduction in study of cell-to-cell interactions in our co-culture system. We the killing by CTL after 9 weeks, we did observe an increase observed an increase in the number of CD8+ T-cell contacts in killing between untreated and SAHA conditions (35–58%) with target cells upon reactivation of SAHA (p < 0.005). These (Figure 5B). contacts were specifically blocked by the addition of an anti- We also monitored phenotypic changes in CTL and HLA-A/B/C antibody (p < 0.05) (Figure 4E), thus demonstrating found a contraction in the frequency of subpopulations the dependency of cell-to-cell contact on the interactions between without expression or expressing a single inhibitory receptor TCR and HLA-class I/peptide complexes after LRA treatment. concomitant with the expansion of cells co-expressing three In addition, CD8+ T cells that established contact were highly or four inhibitory receptors over time (Figure 5C, pie charts). functional, as measured by the levels of CD107a/MIP1β and IFNγ This phenotype of expansion/contraction was coincident when compared with CD8+ T cells lacking contacts with target between CTL1 and CTL2, which vary in HIV-1 antigen cells (p < 0.0001) (Figure 4F). specificity. However, the specific combinations of inhibitory These findings support an overall increase in the killing of receptors diverge between CTL lines and time, thus highlighting HIV-1-infected cells by CD8+ T cells upon LRA induction of the diversity of T-cell inhibitory pathways. Also, CD39+ viral reactivation despite inter-individual divergences. Moreover, subpopulations appeared after 5 weeks in cells already expressing our data point to the opportunity to optimize the immune two or more inhibitory receptors (Figure 5C, arcs). These clearance of HIV-1 reactivated target cells by CD8+ T cells by data suggest the presence of terminally exhausted CD8+ T increasing the LRA dosage and the frequency of effector CD8+ T cells identified by CD39 expression, as previously described cells. (43). Frontiers in Immunology | www.frontiersin.org 8 January 2019 | Volume 9 | Article 3162 Ruiz et al. Barriers to HIV-1 Shock and Kill FIGURE 5 | CD8+ T-cell dysfunction impairs clearance of HIV-1 reactivated cells associated with an increase in the co-expression of inhibitory receptors. (A) Schematic representation of the in vitro exhaustion process, where CTL were exposed to HIV-1 cognate peptide over a period of 9 weeks. (B) Percentage of killing of HIV-1 reactivated cells by CTL after antigen stimulation at week 0 and at 9 weeks. The graph includes one representative experiment. The p-value was calculated using the paired t-test. Statistical significance is shown in the figure (* p < 0.05). (C) Pie charts representing the fraction of cells expressing different combinations of PD-1, LAG3, TIM-3, and CD39 at weeks 0, 5, 7, and 9 after antigen stimulation. Based on these data and the biological diversity found in we did not find a correlation between the frequency of the recognition and killing of LRA-reactivated cells by ex vivo polyfunctional cells among the HIV-1–specific CD8+ T cells CD8+ T cells, we measured HIV-1–specific CD8+ T-cell and the killing in response to LRA-reactivated cells, as it has function in response to a pool of Gag overlapping peptides. We previously been associated with spontaneous control of HIV-1 monitored expression of CD107a, IFNγ, and MIP1β cytokines replication (44, 45). Importantly, individuals harboring HIV-1– and of PD-1, LAG3, TIM-3, and CD39 by multiparametric flow specific CD8+ T cells lacking expression of inhibitory receptors cytometry. As shown in Figure 6A, we observed variation in the retained the highest suppressive capacity against HIV-1 LRA- profile of cytokine production among individuals in response to reactivated cells (Figure 6B). Thus, PT5 responded to LRA HIV-1. Meanwhile, the functional response to SEB control treatment with a 68% in vitro suppressive capacity and 63% stimuli was very homogeneous, as expected. Furthermore, of HIV-specific CD8+ T cells lacking inhibitory receptors. Frontiers in Immunology | www.frontiersin.org 9 January 2019 | Volume 9 | Article 3162 Ruiz et al. Barriers to HIV-1 Shock and Kill FIGURE 6 | Expression profiles of cytokines and inhibitory receptors in HIV-1–specific CD8+ T cells. (A) Functional profiles of HIV-1–specific CD8+ T cells. Pie charts represent functional profiles associated with the secretion of IFNγ, CD107a, and IL-2 in relation with IFNγ+ single, double (IFNγ+CD107a+IL-2-,IFNγ+CD107a- IL-2+), and triple positive cells (IFNγ+CD107a+IL−2+). Data from four HIV-1-infected individuals following 6 h of stimulation with SEB or a pool of HIV-1 Gag peptides is shown. (B) Profiles of inhibitory receptor (iR) expression in HIV-1–specific CD8+ T cells from infected individuals taken from IFNγ positive cells. For (A,B) the pie slices represent the frequency of cells with expression of one or more markers, and the pie arcs represent the specific markers expressed. The gray gradient below indicates the magnitude of killing of HIV-1 reactivated cells by four HIV-1-infected individuals. (C) Correlation of the frequency of HIV-1–specific CD8+T cells lacking iR expression and the magnitude of SAHA-reactivated cell killing. The line indicates the fit of the data to a linear regression. The Spearman correlation coefficient (r) and the two-tailed p-value are indicated. Participants PT3, PT1, and PT7 with 36, 35, and 24% of HIV-1–specific CD8+ T cells lacking inhibitory receptors HIV-1–specific CD8+ T cells lacking inhibitory receptors had and their immune responsiveness to killing LRA treatment suppressive capacities achieving 19, 10, and 2% of killing (p < 0.005, Figure 6C). of SAHA reactivated cells. Although limited by sample size, According to these data, changes in the profile of inhibitory our data delineate a correlation between the percentages of receptors in HIV-1-specific CTL and primary CD8+ T cells Frontiers in Immunology | www.frontiersin.org 10 January 2019 | Volume 9 | Article 3162 Ruiz et al. Barriers to HIV-1 Shock and Kill modulate their antiviral potency. Therefore, the levels of in HIV-1-infected individuals in the context of LRA in order inhibitory receptor co-expression in CD8+ T cells can limit the to ensure accurate delineation of the early reactivation events magnitude of the killing of HIV-1-reactivated cells. and broad immune recognition of LRA-reactivated cells. The consistency of our data across the CTLs used may indicate that clearance of reactivated cells is independent of CTL DISCUSSION antigen specificity when an adequate threshold of reactivation is reached. Clinical use of LRA revealed detectable perturbations in the Furthermore, we did not find differences in the levels of HIV-1 reservoir with no measurable impact on reservoir size cytokines secreted by CTL between untreated and HDACi- (5–8). Therefore, the potential of shock and kill strategies treated conditions, which may constrain their use as biosensors to purge HIV-1 reservoirs should be carefully evaluated to of HIV-1 reactivation events, as previously suggested (40). As delineate current therapeutic limitations. On the one hand, mentioned, the use of highly sensitive CTL in our experiments LRA should be able to induce HIV-1 latency reversal for may limit their capacity to detect small differences in the recognition by immune effector cells. On the other, the functional number of antigens presented between the conditions tested. characteristics of CD8+ T cells should ensure the effective Thus, we propose to explore additional markers of CD8+ T- clearance of reactivated cells. Here, we report the barriers cell activation to validate their use as biosensors of HIV-1 that account for the immune recognition of HIV-1 latently reactivation. We also detected a significant increase in cytokine infected by CD8+ T cells in the context of shock and kill secretion and killing by CTL under Bryo conditions in the cure strategies. Our data demonstrate the potency of LRA for absence of HIV-1 reactivation (22, 24). Such a pre-activation speeding up the recognition of reactivated cells and CD8+ status driven by residual Bryo despite extensive washout, or by T-cell dysfunction associated with the expression of inhibitory the direct effect of Bryo on target cells did not show a measurable receptors as limitations to ensure the killing of HIV-1 reactivated benefit concerning faster or better killing when compared with cells. Thus, we characterize the level of HIV-1 protein expression HDACi alone. These data alert to the potential effect of using upon treatment with LRA as a mechanism that is essential for Bryo, which leads to undesirable and non-specific CD8+ T-cell rapid and potent killing of reactivated cells by CD8+ T cells. activation and cytokine release, as suggested in other studies Moreover, our findings suggest high inter-individual variation (23, 24). in CD8+ T-cell responsiveness to HIV-1 reactivation, where the Together with LRA potency, CD8+ T-cell functional status expression of inhibitory receptors modulates the magnitude of is essential to guaranty the efficacy of shock and kill strategies killing. to clear the HIV-1 reservoir. Our data indicate biological As shown by previous reports, we found how reactivation of diversity across ex vivo CD8+ T cells from HIV-1-infected latently infected cells by LRA enables HIV-1 antigen presentation individuals in response to LRA-reactivated cells. In our for recognition and killing by CD8+ T cells (7, 46). Our data experimental framework, we can exclude some of the barriers are consistent with those from previous studies and suggest for CD8+ T-cell function, including viral escape (20), anatomical that once LRA induced HIV-1 antigen production, there was compartmentalization (48), and lack of viral protein translation no interference in antigen presentation in the context of HLA- (21). Therefore, the differences observed here are mainly due class I molecules for recognition by CD8+ T cells (22). In to functional differences between CD8+ T cells from HIV-1- addition, we were able to characterize in detail the characteristics infected individuals. Although we demonstrate an increase in the of the so-called window of opportunity between reactivation killing of reactivated cells by ex vivo CD8+ T cells after treatment and killing for CD8+ T cells. In our experimental framework, with SAHA, we found differences in the kinetics and the total the killing of reactivated cells can happen as early as 3 h after magnitude of killing between individuals. These differences in co-culture, under conditions where LRA treatment achieves response to LRA treatment reflect the heterogeneity of the HIV- the highest reactivation level. The correlation between LRA 1–specific CD8+ T cell responses, even in the context of the potency measure as HIV-1-fold induction, and the speed of HLA-B∗ 27 allele expressed in our target cells, which has been recognition by CD8+ T cells uncover the presence of a threshold associated with strong antiviral responses and elite control of of antigen production after administration of LRA for killing. HIV-1 (14). Although our experimental data propose a threshold above a Multiple factors account for the functional diversity of HIV- 2-fold induction of HIV-1 reactivation for CD8+ T-cell killing, 1–specific-CD8+ T-cell responses irrespective of the clinical we need to remember that the data were obtained in the stage of HIV-1, including the low frequency of HIV-1–specific context of two particular CTL. Our results constitute a proof CD8+ T cells upon antigen clearance by long-term ART of principle on the need for potent LRAs to ensure the rapid and continuous immune CD8+ T-cell activation despite ART recognition of HIV-1 reactivated cells for effective shock and (18, 19). Moreover, HIV-1–specific CD8+ T-cell dysfunction is kill strategies, even in the context of fully functional CD8+ associated with a high level of expression of inhibitory receptors T-cell responses. However, we cannot extrapolate the current including PD-1, TIM-3, LAG-3, and TIGIT (26, 27, 29). These findings as this theoretical threshold may vary across HIV-1- markers have been associated with CD8+ T-cell dysfunction infected individuals with differences in TCR affinities (31, 47). and terminal exhaustion in chronic infections (25, 30). Our in It is necessary to perform further experiments that evaluate vitro data demonstrate that upon continuous exposure to HIV- diversity in TCR affinities between CD8+ T-cell clonotypes 1 antigens, there is a contraction of subpopulations lacking Frontiers in Immunology | www.frontiersin.org 11 January 2019 | Volume 9 | Article 3162 Ruiz et al. Barriers to HIV-1 Shock and Kill inhibitory receptors. This is concomitant with the expansion MATERIALS AND METHODS of CTL subpopulations co-expressing at least three inhibitory receptors, including PD-1, LAG3, TIM-3, and the metabolic Virus and Cell Lines marker CD39 (43). In our experimental setting, the expansion The virus NL43GFP was generated by co-transfection of of CTL co-expressing inhibitory receptors over time diminished p83–2 and p83–10eGFP plasmids in MT4 cells (NIH AIDS the magnitude of the killing of LRA-reactivated cells. However, Reagent program, USA) based on electroporation, as previously our in vitro model does not account for the pro-inflammatory described (38, 39). The cell-free supernatant was concentrated environment of HIV-1 infection. Further studies including the by spinoculation at 24,000 rpm at 4◦ C for 90 min, and stored use of pro-inflammatory cytokines and characterization of the at −80◦ C. The 50% tissue culture infective dose (TCID50 ) was expression of inhibitory receptors in tetramer-sorted HIV-1– determined on TZM-bl cells (NIH AIDS Reagent program, USA) specific cells expose to HIV-1 peptides will help to clarify whether as described in Kloverpris et al. (31). The HIV-1-permissive our findings can be extended to HIV-1–specific circulating CD8+ U937 cell line transfected with the HLA-B∗ 27:05 gene (kindly T cells. provided by Paul Bowness, UK) (53) was cultured in RPMI media Despite the limitations mentioned, our findings were supplemented with 10% fetal calf serum (FCS) (HyClone), 2 mM consistent with the direct use of ex vivo CD8+ T cells, where L-glutamine, 100 U/ml penicillin, 100 µg/ml streptomycin, and the frequency of HIV-1–specific CD8+ T cells lacking inhibitory 0.5 mg/ml geneticin (Invitrogen, Life Technologies). The HIV-1 receptors directly correlated with the magnitude of killing of cytotoxic CD8+ T-cell lines (CTL) used were restricted by HLA- LRA-reactivated cells. This is the first study to our knowledge B∗ 27:05. The CTL1 was specific for the KK10 epitope in Gag that suggest an inter-dependency between CD8+ T-cell function and CTL2 for the KY9 epitope in Pol. In addition, CTL3 was and the expression patterns of inhibitory receptors in response to restricted by HLA-B∗ 57:01 and specific for the KF11 epitope in HIV-1 reactivation (28). Gag (31, 53). The identification of novel inhibitory receptors as markers of HIV-1–specific CD8+ T-cell dysfunction has increased in The HIV-1 Shock and Kill RELI Model recent years. The interest has risen due the unprecedented The U937 HLA-B∗ 27:05 permissive cell line was infected with success of clinical trials targeting inhibitory receptors in cancer NL43GFP by magnetofection at a dose equivalent to a nominal regression (49, 50). Consequently, additional efforts are needed multiplicity of infection (MOI) of 0.1, as previously reported to clarify the impact of inhibitory receptor blockade in HIV- (31, 32). After magnetofection, cells were washed twice with 1-infected individuals. Immunotherapy based on blockage of 1X PBS and cultured in R10 + geneticin (0.5 mg/ml). At inhibitory receptors has proven effective in various types of day three post-infection, the GFP-negative population, which cancer. The use of PD-1 inhibitors have demonstrated to be included the non-productively infected cells enriched in HIV-1 safe and effective in HIV-infected individuals with lung cancer integrated provirus, was sorted with a FACS Aria II Cell Sorter (51). Moreover, the use of an anti-PD-L1 in HIV-1-infected (BD Biosciences) and cultured in the presence of 1 µM of RTV individuals has already been evaluated in a Phase I clinical (Sigma, Spain). After 4 days, we obtained a culture enriched in trial (52), in which anti-PD-L1 had a limited effect on the “resting-like” cells (RELI). The mean levels of cell-associated total recovery of HIV-1–specific CD8+ T-cell responses (2 out of HIV-1 DNA in RELI-lysates was 5.1–log ± 0.19 log10 copies/106 6 patients). Consequently, the delineation of the functional RELI as measured by droplet digital PCR previously described in characteristics of HIV-1–specific CD8+ T cells behind successful Martínez-Bonet et al. (54). RELI cells were treated with LRA in shock and kill strategies will play a key role in the assessment the presence of RAL at 100 nM (Sigma, Spain) to monitor HIV- of treatment efficacy. The efforts to define these traits should 1 reactivation or “shock.” The LRA panel included the HDACi: accompany the development of ultrasensitive technologies for panobinostat (PNBN) at 30 nM (SelleckChem, USA), trichostatin the monitoring of small perturbations in the size of the A (Tricho) at 250 nM (Sigma, Spain), SAHA (suberoyl aniline HIV-1 reservoir in response to low fractions of reactivated hydroxamic acid) at 1 µM (SelleckChem, Spain), and romidepsin cells killed by CD8+ T cells in blood or tissues. These (RMD) at 40 nM (SelleckChem, Spain), and the PKC agonist technologies will be essential to guide the outcome of curative bryostatin A (Bryo) at 10 nM (Sigma, Spain). RELI cells were strategies. treated for 48 h with LRA except for RMD-treated cultures, In summary, our study suggests the need for a strong flush where the drug was washed-off after 4 h to avoid cellular of latent virus to speed up recognition of HIV-1 reactivated toxicity. cells and an adequate functional capacity of HIV-1–specific To monitor the “shock” or HIV-1-fold induction at 48 h after CD8+ T cells to maximize elimination of LRA-reactivated cells. treatment with LRA, the RELI cells were harvested and stained An adequate trade-off between shock and kill is essential to with a Live/Dead probe (APC-Cy7, Invitrogen). After washing, ensure the success of potential curative strategies. The immune cells were fixed and permeabilized with the Fix&Perm kit (A characterization of HIV-1–specific CD8+ T-cell responses, and B solutions, Thermo Fisher Scientific) for intracellular HIV- including patterns of inhibitory receptors associated with the 1 p24 (PE, clone KC57-RD1, Beckmann Coulter). Cells were killing of reactivated cells, will allow the design of novel immune- acquired on an LSR II flow cytometer (Beckmann Coulter), monitoring and immuno based therapeutics to ensure effective and data were analyzed using FlowJoV (Tree Star Inc.). The antiviral responses before latency reversal in HIV-1-infected HIV-1-fold induction was calculated based on the following individuals. equation: % of p24+ or GFP+ in CD4+ under LRA conditions/% Frontiers in Immunology | www.frontiersin.org 12 January 2019 | Volume 9 | Article 3162 Ruiz et al. Barriers to HIV-1 Shock and Kill of p24+ or GFP+ under untreated conditions. To proceed (10 µg/ml) and irradiated PBMCs from three healthy HIV- with the co-cultures to monitor the “killing” of reactivated seronegative donors weekly. We used flow cytometry to evaluate cells by CD8+ T cells the RELI cells were extensively washed variations in the expression of the inhibitory receptors PD-1, with 1X PBS and left in R10 media plus RAL. Thus, we TIM-3, LAG-3, and the immune-metabolic marker CD39. cultured LRA-reactivated RELI cells in the absence or presence Briefly, cells were taken from the culture at weeks 0, 5, 7, and of HIV-1-specific CTLs or B∗ 27-HLA-matched ex vivo CD8+ T 9 and stained with a Live/Dead probe (APC-Cy7, Invitrogen) cells obtained from HIV-1-infected individuals. Cells were co- and surface markers CD3 (Alexa700, BD Biosciences), CD4 cultured to a 1:1 effector: target ratio for 20 h. After 20 h of (APC-Cy7, BD Biosciences), CD8 (V500, BD Biosciences), PD-1 co-culture, we simultaneously measured by flow cytometry the (BV421, BD Biosciences), TIM-3 (Alexa 647, BD Biosciences), “shock” by intracellular p24, and the “kill” by the reduction of LAG-3 (PE, BD Biosciences), and CD39 (FITC, BD Biosciences) p24-positive RELI cells in the absence or presence of effector and incubated at room temperature for 25 min. Samples were cells. We also assessed LRA-related toxicity both in uninfected washed twice with 1X PBS, fixed in 1% formaldehyde, and U937 cells and in the RELI model by measuring cellular viability acquired on an LSR Fortessa. Data were analyzed with FlowJoV using flow cytometry and staining with a Live/Dead probe (Tree Star Inc.). Patterns of co-expression of PD-1, LAG-3, (APC-Cy7, Invitrogen) at the end of the co-culture with RELI TIM-3, and CD39 were analyzed using Pestle and SPICE v5 and CTL. software (55). CD8+ T-Cell Killing and Activation Immunophenotype of HIV-1–Specific CD8+ We performed CD8+ T-cell killing experiments of LRA- reactivated cells with CTL or ex vivo CD8+ T cells at 20 h after T-Cell Responses culture. We also monitored the kinetics of CD8+ T-cell killing The PBMCs from HIV-1-infected individuals were stimulated of LRA-reactivated cells with CTL or ex vivo CD8+ T cells with HIV-1 Gag peptide pool (2 µg/peptide/ml, EzBiolab), or after 3, 6, and 20 h after co-culture. We monitored CTL and the positive control SEB (enterotoxin B from Staphylococcus CD8+ T-cell activation by intracellular expression of CD107a, aureus, Sigma, Ref. S4881) or no stimuli in the presence of MIP-1β, and IFNγ. For analysis of HLA-class I restriction, the CD28/49d co-stimulatory molecules (1 µl/ml, BD Biosciences), W6/32 antibody (BioLegend, Spain) or the isotype control LEAF Golgistop solution (Monensin A, BD Biosciences, Spain), and purified IgG2aÎ (BioLegend) was added to the co-culture at anti-human CD107a (PE-Cy5, BD Biosciences) for 6 h at 37◦ C 10 µg/ml. For analysis of CD8+ T-cells in contact with RELI in a 5% CO2 incubator and overnight at 4◦ C. Stimulation cells cells, we analyzed the percentage of doublets gated cells that were then washed with 1X PBS and stained for 25 min with the expressed the CD8+ surface marker in the target cells gate. Live/Dead probe (APC-Cy7, Invitrogen). Next, cells were washed For analysis of early CD107a secretion, the antibody CD107a with 1X PBS and surface stained for 25 min with CD3 (A700, (PerCP-Cy5.5, BD Biosciences) was added to the co-culture. BD Biosciences), CD4 (APC-Cy7, BD Biosciences), CD8 (V500, For the immune-phenotype, samples were collected and washed BD Biosciences), PD-1 (BV421, BD Biosciences), LAG3 (PE, BD twice with 1X PBS and incubated for 3 h in the presence of Biosciences), TIM-3 (A647, BD Biosciences), and CD39 (FITC, the protein transport inhibitor Golgiplug (Brefeldin A solution, BD Biosciences). Subsequently, cells were washed twice in 1X BD Biosciences, Spain) and Golgistop (Monensin solution, BD PBS and fixed and permeabilized with the Fix&Perm kit (A and Biosciences, Spain). Cells were then harvested, stained with a B solutions, Thermo Fisher Scientific) for intracellular cytokine Live/Dead probe (APC-Cy7, Invitrogen). Next, cells were washed staining of IFNγ (BV711, BD Biosciences) and IL-2 (BV650, and stained to identify the surface marker CD8 (Pacific blue, BD Biosciences). After 20 min at room temperature in the dark, clone RPA-T8, BD Biosciences). After washing, cells were fixed stained samples were washed twice with 1X PBS, resuspended in and permeabilized with the Fix&Perm kit (A and B solutions, 1% formaldehyde and acquired on an LSR Fortessa. Data were Thermo Fisher Scientific) for intracellular cytokine staining of analyzed with FlowJoV (Tree Star Inc.). Patterns of simultaneous MIP1β (FITC, clone 24,006, R&D Systems), IFNγ (PE-Cy7, expression of IFNγ, CD107a, and IL-2, or combinations of PD-1, clone 4S.B3, BD Biosciences), and HIV-1 p24 (PE, clone KC57- LAG-3, TIM-3, and CD39 were analyzed using Pestle and SPICE RD1, Beckmann Coulter). Cells were acquired on an LSR II v5 software (55). flow cytometer (Beckmann Coulter). Data were analyzed with FlowJoV (Tree Star Inc.). The killing of reactivated HIV-1 cells Statistics by CTL/CD8+ was calculated based on the following equation: Statistical analyses were performed using Prism v4 (GraphPad 100—[(% p24+ in CD4+ co-cultured with CD8+/% p24+ in Software, Inc.). Reported p-values were calculated using the CD4+ in the absence of CD8+) × 100]. Wilcoxon single-rank test in paired comparisons and the t-test and Mann-Whitney test for independent median comparisons. In vitro CTL Exhaustion Experiments Also we used one-way ordinary analysis of variants (ANOVA), To characterize potential markers associated with CTL and one-way ANOVA for multiple group comparisons adjusting dysfunction in response to LRA-reactivated cells, we exposed the pairwise analyses by Tukey. The relationship between HIV- CTL1 and CTL2 to cognate HIV-1 antigen over nine weeks. 1 induction measured according to the expression of p24 and To do so, we stimulated CTL with a 1:1 mixture of irradiated GFP, and the relationship between the lack of inhibitory receptors autologous B-cell lines (BCLs) pulsed with cognate HIV peptide in CD8+ and percentage of killing were assessed using the Frontiers in Immunology | www.frontiersin.org 13 January 2019 | Volume 9 | Article 3162 Ruiz et al. Barriers to HIV-1 Shock and Kill Spearman correlation coefficient. All statistical tests were under Nacional de Investigación Científica, Desarrollo e Innovación a significance level of 0.05. Tecnológica 2008-2011; Instituto de Salud Carlos III; Fondos FEDER. JGP holds a Miguel Servet II contract (CPII15/00014) ETHICS STATEMENT funded by ISCIII. OB-L holds an AGAUR-FI Ph.D. fellowship from the Catalan Government. CG holds a FPU15/03698 All methods and experimental protocols of the study were Ph.D. fellowship from the Spanish Ministry of Education, approved by the Ethics Committee of Hospital Germans Trias Culture and Sport. MG was supported by the National Health i Pujol. For the study, both infected and healthy subjects Institute Carlos III (ISCIII; Grants PI14/01235 PI17/01470). provided their written informed consent to participate. The study MG holds the Pla estratègic de recerca i innovació en was conducted according to the principles expressed in the salut (PERIS) contract, from the Catalan government. The Declaration of Helsinki. funders had no role in study design, data collection and interpretation, or the decision to submit the work for AUTHOR CONTRIBUTIONS publication. AR and JGP: design of the study. AR, OB-L, EJ-M, RP, and CG: ACKNOWLEDGMENTS performance of the experiments. AR, OB-L, EJ-M, JGP, MG, BM, DO, CG, and RB: analysis of the data and results. AR, OB-L, BM, We thank the IGTP Flow Cytometry Core Facility and MF and JM-P, PG, MG, RB, BH, BC, and JGP: manuscript drafting and GR from the staff for their contribution to this publication. We discussion. thank all HIV+ individuals participating in the study and the National Institutes of Health (NIH) AIDS Reagent Program for reagents. FUNDING This study was supported by the National Health Institute Carlos SUPPLEMENTARY MATERIAL III (ISCIII; Grants PI14/01058, PI17/00168) and the Gilead Fellowship Program GLD 15/00298. EJ-M was supported by the The Supplementary Material for this article can be found following: Redes Temáticas de Investigación en SIDA (ISCIII online at: https://www.frontiersin.org/articles/10.3389/fimmu. RETIC RD16/0025/0041); Acción Estratégica en Salud; Plan 2018.03162/full#supplementary-material REFERENCES replication-competent HIV-1: potential role in latent reservoir dynamics. J Exp Med. (2017) 214:959–72. doi: 10.1084/jem.20170193 1. Deeks SG, Lewin SR, Havlir DV. The end of AIDS: HIV infection as a chronic 11. Shan L, Deng K, Shroff NS, Durand CM, Rabi SA, Yang H-C, et al. disease. Lancet (2013) 382:1525–33. doi: 10.1016/S0140-6736(13)61809-7 Stimulation of HIV-1-specific cytolytic T lymphocytes facilitates elimination 2. Chun TW, Finzi D, Margolick J, Chadwick K, Schwartz D, Siliciano RF. In vivo of latent viral reservoir after virus reactivation. Immunity (2012) 36:491–501. fate of HIV-1-infected T cells: quantitative analysis of the transition to stable doi: 10.1016/j.immuni.2012.01.014 latency. Nat Med. (1995) 1:1284–90. 12. Descours B, Avettand-Fenoel V, Blanc C, Samri A, Mélard A, Supervie V, 3. Maldarelli F, Wu X, Su L, Simonetti FR, Shao W, Hill S, et al. Specific HIV et al. Immune responses driven by protective human leukocyte antigen integration sites are linked to clonal expansion and persistence of infected alleles from long-term nonprogressors are associated with low HIV reservoir cells. Science (2014) 345:179–83. doi: 10.1126/science.1254194 in central memory CD4 T cells. Clin Infect Dis. (2012) 54:1495–503. 4. Chomont N, El-Far M, Ancuta P, Trautmann L, Procopio FA, Yassine-Diab doi: 10.1093/cid/cis188 B, et al. HIV reservoir size and persistence are driven by T cell survival and 13. Graf EH, Mexas AM, Yu JJ, Shaheen F, Liszewski MK, Di Mascio M, et al. homeostatic proliferation. Nat Med. (2009) 15:893–900. doi: 10.1038/nm.1972 Elite suppressors harbor low levels of integrated HIV DNA and high levels of 5. Søgaard OS, Graversen ME, Leth S, Olesen R, Brinkmann CR, Nissen SK, et 2-LTR circular HIV DNA compared to HIV+ patients On and Off HAART. al. The Depsipeptide romidepsin reverses HIV-1 latency in vivo. PLoS Pathog. PLoS Pathog. (2011) 7:e1001300. doi: 10.1371/journal.ppat.1001300 (2015) 11:e1005142. doi: 10.1371/journal.ppat.1005142 14. Noel N, Peña R, David A, Avettand-Fenoel V, Erkizia I, Jimenez E, et al. 6. Leth S, Schleimann MH, Nissen SK, Højen JF, Olesen R, Graversen ME, et Long-term spontaneous control of HIV-1 is related to low frequency of al. Combined effect of Vacc-4x, recombinant human granulocyte macrophage infected cells and inefficient viral reactivation. J Virol. (2016) 90:6148–58. colony-stimulating factor vaccination, and romidepsin on the HIV-1 reservoir doi: 10.1128/JVI.00419-16 (REDUC): a single-arm, phase 1B/2A trial. Lancet HIV (2016) 3:e463–72. 15. Kiepiela P, Ngumbela K, Thobakgale C, Ramduth D, Honeyborne I, Moodley doi: 10.1016/S2352-3018(16)30055-8 E, et al. CD8+ T-cell responses to different HIV proteins have discordant 7. Archin NM, Liberty AL, Kashuba AD, Choudhary SK, Kuruc JD, Crooks associations with viral load. Nat Med. (2007) 13:46–53. doi: 10.1038/nm1520 AM, et al. Administration of vorinostat disrupts HIV-1 latency in patients on 16. Sáez-Cirión A, Lacabaratz C, Lambotte O, Versmisse P, Urrutia A, Boufassa antiretroviral therapy. Nature (2012) 487:482–5. doi: 10.1038/nature11286 F, et al. HIV controllers exhibit potent CD8 T cell capacity to suppress HIV 8. Sluis-Cremer N. Therapeutic approaches to eradicate latent HIV-1 in resting infection ex vivo and peculiar cytotoxic T lymphocyte activation phenotype. CD4+ T Cells. Curr Top Med Chem. (2016) 16:1191–7. Proc Natl Acad Sci USA. (2007) 104:6776–81. doi: 10.1073/pnas.0611244104 9. Ho Y-C, Shan L, Hosmane NN, Wang J, Laskey SB, Rosenbloom DIS, et al. 17. Cartwright EK, Spicer L, Smith SA, Lee D, Fast R, Paganini S, et al. CD8 + Replication-competent noninduced proviruses in the latent reservoir increase Lymphocytes are required for maintaining viral suppression in SIV-infected barrier to HIV-1 cure. Cell (2013) 155:540–51. doi: 10.1016/j.cell.2013. macaques treated with short-term antiretroviral therapy. Immunity (2016) 09.020 45:656–68. doi: 10.1016/j.immuni.2016.08.018 10. Hosmane NN, Kwon KJ, Bruner KM, Capoferri AA, Beg S, Rosenbloom 18. Gray CM, Lawrence J, Schapiro JM, Altman JD, Winters MA, Crompton M, et DIS, et al. Proliferation of latently infected CD4 + T cells carrying al. Frequency of class I HLA-restricted anti-HIV CD8+ T cells in individuals Frontiers in Immunology | www.frontiersin.org 14 January 2019 | Volume 9 | Article 3162 Ruiz et al. Barriers to HIV-1 Shock and Kill receiving highly active antiretroviral therapy (HAART). J Immunol. (1999) 36. Sharma S, Witteveen PO, Lolkema MP, Hess D, Gelderblom H, Hussain SA, et 162:1780–8. al. A phase I, open-label, multicenter study to evaluate the pharmacokinetics 19. Casazza JP, Betts MR, Picker LJ, Koup RA. Decay kinetics of human and safety of oral panobinostat in patients with advanced solid tumors immunodeficiency virus-specific CD8+ T cells in peripheral blood after and varying degrees of renal function. Cancer Chemother Pharmacol. (2015) initiation of highly active antiretroviral therapy. J Virol. (2001) 75:6508–16. 75:87–95. doi: 10.1007/s00280-014-2612-8 doi: 10.1128/JVI.75.14.6508-6516.2001 37. Wei DG, Chiang V, Fyne E, Balakrishnan M, Barnes T, Graupe M, 20. Deng K, Pertea M, Rongvaux A, Wang L, Durand CM, Ghiaur G, et al. Broad et al. Histone deacetylase inhibitor romidepsin induces HIV expression CTL response is required to clear latent HIV-1 due to dominance of escape in CD4 T cells from patients on suppressive antiretroviral therapy at mutations. Nature (2015) 517:381–5. doi: 10.1038/nature14053 concentrations achieved by clinical dosing. PLoS Pathog. (2014) 10:e1004071. 21. Pollack RA, Jones RB, Pertea M, Bruner KM, Martin AR, Thomas AS, et al. doi: 10.1371/journal.ppat.1004071 Defective HIV-1 proviruses are expressed and can be recognized by cytotoxic 38. Weber J, Weberova J, Carobene M, Mirza M, Martinez-Picado J, Kazanjian P, T lymphocytes, which shape the proviral landscape. Cell Host Microbe. (2017) et al. Use of a novel assay based on intact recombinant viruses expressing green 21:494–506.e4. doi: 10.1016/j.chom.2017.03.008 (EGFP) or red (DsRed2) fluorescent proteins to examine the contribution of 22. Jones RB, O’Connor R, Mueller S, Foley M, Szeto GL, Karel D, pol and env genes to overall HIV-1 replicative fitness. J Virol Methods (2006) et al. Histone deacetylase inhibitors impair the elimination of HIV- 136:102–17. doi: 10.1016/j.jviromet.2006.04.004 infected cells by cytotoxic T-lymphocytes. PLoS Pathog. (2014) 10:e1004287. 39. Prado JG, Honeyborne I, Brierley I, Puertas MC, Martinez-Picado J, Goulder doi: 10.1371/journal.ppat.1004287 PJR. Functional consequences of human immunodeficiency virus escape from 23. Walker-Sperling VE, Pohlmeyer CW, Tarwater PM, Blankson JN. The effect of an HLA-B∗ 13-restricted CD8+ T-cell epitope in p1 Gag protein. J Virol. latency reversal agents on primary CD8+ T cells: implications for shock and (2009) 83:1018–25. doi: 10.1128/JVI.01882-08 kill strategies for human immunodeficiency virus eradication. EBioMedicine 40. Jones RB, Mueller S, O’Connor R, Rimpel K, Sloan DD, Karel D, et al. A (2016) 8:217–29. doi: 10.1016/j.ebiom.2016.04.019 subset of latency-reversing agents expose HIV-infected resting CD4+ T-cells 24. Clutton G, Xu Y, Baldoni PL, Mollan KR, Kirchherr J, Newhard W, et al. The to recognition by cytotoxic T-lymphocytes. PLoS Pathog. (2016) 12:e1005545. differential short- and long-term effects of HIV-1 latency-reversing agents on doi: 10.1371/journal.ppat.1005545 T cell function. Sci Rep. (2016) 6:30749. doi: 10.1038/srep30749 41. Yang H, Buisson S, Bossi G, Wallace Z, Hancock G, So C, et al. Elimination of 25. Kuchroo VK, Anderson AC, Petrovas C. Coinhibitory receptors and CD8 T latently HIV-infected cells from antiretroviral therapy-suppressed subjects by cell exhaustion in chronic infections. Curr Opin HIV AIDS (2014) 9:439–45. engineered immune-mobilizing T-cell receptors. Mol Ther. (2016) 24:1913– doi: 10.1097/COH.0000000000000088 25. doi: 10.1038/mt.2016.114 26. Okoye IS, Houghton M, Tyrrell L, Barakat K, Elahi S. Coinhibitory receptor 42. Macatangay BJC, Rinaldo CR. Preserving HIV-specific T cell responses: does expression and immune checkpoint blockade: maintaining a balance in CD8+ timing of antiretroviral therapy help? Curr Opin HIV AIDS (2015) 10:55–60. T cell responses to chronic viral infections and cancer. Front Immunol. (2017) doi: 10.1097/COH.0000000000000124 8:1215. doi: 10.3389/fimmu.2017.01215 43. Gupta PK, Godec J, Wolski D, Adland E, Yates K, Pauken KE, 27. Trautmann L, Janbazian L, Chomont N, Said EA, Gimmig S, Bessette et al. CD39 Expression identifies terminally exhausted CD8+ T B, et al. Upregulation of PD-1 expression on HIV-specific CD8+ T cells cells. PLoS Pathog. (2015) 11:e1005177. doi: 10.1371/journal.ppat.10 leads to reversible immune dysfunction. Nat Med. (2006) 12:1198–202. 05177 doi: 10.1038/nm1482 44. Ndhlovu ZM, Proudfoot J, Cesa K, Alvino DM, McMullen A, Vine S, et al. 28. Yamamoto T, Price DA, Casazza JP, Ferrari G, Nason M, Chattopadhyay PK, Elite Controllers with low to absent effector CD8+ T cell responses maintain et al. Surface expression patterns of negative regulatory molecules identify highly functional, broadly directed central memory responses. J Virol. (2012) determinants of virus-specific CD8+ T-cell exhaustion in HIV infection. 86:6959–69. doi: 10.1128/JVI.00531-12 Blood (2011) 117:4805–15. doi: 10.1182/blood-2010-11-317297 45. Graw F, Regoes RR. Predicting the impact of CD8+ T cell 29. Day CL, Kaufmann DE, Kiepiela P, Brown JA, Moodley ES, Reddy S, et al. polyfunctionality on HIV disease progression. J Virol. (2014) 88:10134–45. PD-1 expression on HIV-specific T cells is associated with T-cell exhaustion doi: 10.1128/JVI.00647-14 and disease progression. Nature (2006) 443:350–4. doi: 10.1038/nature 46. Sung JA, Sholtis K, Kirchherr J, Kuruc JD, Gay CL, Nordstrom JL, Bollard 05115 CM, et al. Vorinostat renders the replication-competent latent reservoir of 30. Blackburn SD, Shin H, Haining WN, Zou T, Workman CJ, Polley A, human immunodeficiency virus (HIV) vulnerable to clearance by CD8 T cells. et al. Coregulation of CD8+ T cell exhaustion by multiple inhibitory EBioMedicine (2017) 23:52–8. doi: 10.1016/j.ebiom.2017.07.019 receptors during chronic viral infection. Nat Immunol. (2009) 10:29–37. 47. Bennett MS, Ng HL, Dagarag M, Ali A, Yang OO. Epitope-dependent avidity doi: 10.1038/ni.1679 thresholds for cytotoxic T-lymphocyte clearance of virus-infected cells. J Virol. 31. Kloverpris HN, Payne RP, Sacha JB, Rasaiyaah JT, Chen F, Takiguchi M, (2007) 81:4973–80. doi: 10.1128/JVI.02362-06 et al. Early antigen presentation of protective HIV-1 KF11Gag and KK10 48. Petrovas C, Ferrando-Martinez S, Gerner MY, Casazza JP, Pegu A, Deleage C, gag epitopes from incoming viral particles facilitates rapid recognition et al. Follicular CD8 T cells accumulate in HIV infection and can kill infected of infected cells by specific CD8+ T cells. J Virol. (2013) 87:2628–38. cells in vitro via bispecific antibodies. Sci Transl Med. (2017) 9:eaag2285. doi: 10.1128/JVI.02131-12 doi: 10.1126/scitranslmed.aag2285 32. Sacha JB, Watkins DI. Synchronous infection of SIV and HIV in vitro for 49. Gong J, Chehrazi-Raffle A, Reddi S, Salgia R. Development of PD-1 and PD- virology, immunology and vaccine-related studies. Nat Protoc. (2010) 5:239– L1 inhibitors as a form of cancer immunotherapy: a comprehensive review of 46. doi: 10.1038/nprot.2009.227 registration trials and future considerations. J Immunother Cancer (2018) 6:8. 33. Bonczkowski P, De Scheerder M-A, Malatinkova E, Borch A, Melkova Z, doi: 10.1186/s40425-018-0316-z Koenig R, et al. Protein expression from unintegrated HIV-1 DNA introduces 50. Wykes MN, Lewin SR. Immune checkpoint blockade in infectious diseases. bias in primary in vitro post-integration latency models. Sci Rep. (2016) Nat Rev Immunol. (2017) 18:91–104. doi: 10.1038/nri.2017.112 6:38329. doi: 10.1038/srep38329 51. Ostios-Garcia L, Faig J, Leonardi GC, Adeni AE, Subegdjo SJ, Lydon 34. Laird GM, Bullen CK, Rosenbloom DIS, Martin AR, Hill AL, Durand CA, et al. Safety and efficacy of PD-1 inhibitors among HIV-positive CM, et al. ex vivo analysis identifies effective HIV-1 latency-reversing patients with non-small cell lung cancer. J Thorac Oncol. (2018) 13:1037–42. drug combinations. J Clin Invest. (2015) 125:1901–12. doi: 10.1172/JCI doi: 10.1016/j.jtho.2018.03.031 80142 52. Gay CL, Bosch RJ, Ritz J, Hataye JM, Aga E, Tressler RL, et al. Clinical 35. Rasmussen TA, Tolstrup M, Brinkmann CR, Olesen R, Erikstrup C, Solomon trial of the anti-PD-L1 antibody BMS-936559 in HIV-1 infected participants A, et al. Panobinostat, a histone deacetylase inhibitor, for latent-virus on suppressive antiretroviral therapy. J Infect Dis. (2017) 215:1725–33. reactivation in HIV-infected patients on suppressive antiretroviral therapy: doi: 10.1093/infdis/jix191 a phase 1/2, single group, clinical trial. Lancet HIV (2014) 1:e13–21. 53. Payne RP, Kløverpris H, Sacha JB, Brumme Z, Brumme C, Buus S, doi: 10.1016/S2352-3018(14)70014-1 et al. Efficacious early antiviral activity of hiv gag- and pol-specific Frontiers in Immunology | www.frontiersin.org 15 January 2019 | Volume 9 | Article 3162 Ruiz et al. Barriers to HIV-1 Shock and Kill HLA-B∗ 2705-restricted CD8+ T cells. J Virol. (2010) 84:10543–57. The remaining authors declare that the research was conducted in the absence of doi: 10.1128/JVI.00793-10 any commercial or financial relationships that could be construed as a potential 54. Martínez-Bonet M, Puertas MC, Fortuny C, Ouchi D, Mellado MJ, Rojo P, et conflict of interest. al. Establishment and replenishment of the viral reservoir in perinatally HIV- 1-infected children initiating very early antiretroviral therapy. Clin Infect Dis. Copyright © 2019 Ruiz, Blanch-Lombarte, Jimenez-Moyano, Ouchi, Mothe, (2015) 61:1169–78. doi: 10.1093/cid/civ456 Peña, Galvez, Genescà, Martinez-Picado, Goulder, Barnard, Howell, Clotet 55. Roederer M, Nozzi JL, Nason MC. SPICE: exploration and analysis of post- and Prado. This is an open-access article distributed under the terms of cytometric complex multivariate datasets. Cytometry A (2011) 79:167–74. the Creative Commons Attribution License (CC BY). The use, distribution doi: 10.1002/cyto.a.21015 or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in Conflict of Interest Statement: RB and BH were employed by the company Merck. this journal is cited, in accordance with accepted academic practice. No use, JGP and JM-P report research funding from Merck. JGP and AR have a patent on distribution or reproduction is permitted which does not comply with these the methods for screening of HIV-1 latency reversing agents issue (62590081). terms. Frontiers in Immunology | www.frontiersin.org 16 January 2019 | Volume 9 | Article 3162