Bryostatin-1 Decreases HIV-1 Infection and Viral Production in Human Primary Macrophages

ABSTRACT

While combination antiretroviral therapy maintains undetectable viremia in people living with HIV (PLWH), a lifelong treatment is necessary to prevent viremic rebound after therapy cessation. This rebound seemed mainly caused by long-lived HIV-1 latently infected cells reverting to a viral productive status. Reversing latency and elimination of these cells by the so-called shock-and-kill strategy is one of the main investigated leads to achieve an HIV-1 cure. Small molecules referred to as latency reversal agents (LRAs) proved to efficiently reactivate latent CD4⁺ T cells. However, the LRA impact on de novo infection or HIV-1 production in productively infected macrophages remains elusive. Nontoxic doses of bryostatin-1, JQ1, and romidepsin were investigated in human monocyte-derived macrophages (MDMs). Treatment with bryostatin-1 or romidepsin resulted in a downregulation of CD4 and CCR5 receptors, respectively, accompanied by a reduction of R5 tropic virus infection. HIV-1 replication was mainly regulated by receptor modulation for bryostatin-1, while romidepsin effects rely on upregulation of SAMHD1 activity. LRA stimulation of chronically infected cells did not enhance HIV-1 production or gene expression. Surprisingly, bryostatin-1 caused a major decrease in viral production. This effect was not viral strain specific but appears to occur only in myeloid cells. Bryostatin-1 treatment of infected MDMs led to decreased amounts of capsid and matrix mature proteins with little to no modulation of precursors. Our observations revealed that bryostatin-1-treated myeloid and CD4⁺ T cells respond differently upon HIV-1 infection. Therefore, additional studies are warranted to more fully assess the efficiency of HIV-1 eradicating strategies.

IMPORTANCE HIV-1 persists in a cellular latent form despite therapy that quickly propagates infection upon treatment interruption. Reversing latency would contribute to eradicate these cells, closing the gap to a cure. Macrophages are an acknowledged HIV-1 reservoir during therapy and are suspected to harbor latency establishment in vivo. However, the impact of latency reversal agents (LRAs) on HIV-1 infection and viral production in human macrophages is poorly known but nonetheless crucial to probe the safety of this strategy. In this in vitro study, we discovered encouraging antireplicative features of distinct LRAs in human macrophages. We also described a new viral production inhibition mechanism by protein kinase C agonists that is specific to myeloid cells. This study provides new insights into HIV-1 propagation restriction potentials by LRAs in human macrophages and underline the importance of assessing latency reversal strategy on all HIV-1-targeted cells.

INTRODUCTION

Although combination antiretroviral therapy (cART) leads to undetectable plasma levels of HIV-1 mRNA in people living with HIV (PLWH) (1), a functional cure is still out of reach. One of the main reasons for HIV-1 resilience is thought to be an early establishment of viral reservoirs throughout the body (2). These viral reservoirs have been shown to be composed of latently infected cells that carry silent replication-competent provirus (3). Early work could identify latently infected cells as part of resting memory CD4⁺ T-cell subsets (4). Since HIV-1 transcription is silenced (5), these long-lived cells are not affected by cART or the host immune system. However, upon cART termination, these cells are quickly reactivated, which leads to viremic rebounds (6, 7). With the latent reservoir estimated to survive a lifelong cART treatment, eradication of latently infected cells represents a key concept to achieve an HIV-1 cure (8).

It is thought that by reactivating or shocking latently infected cells, the ensuing viral production recovery would lead to their termination either by the immune system or by virus-associated toxicity, otherwise known as the kill phase (9). Hence, when combined with antiretroviral therapies to limit viral spread, this so-called shock-and-kill strategy represents a great asset to making a cure.

Although precise mechanisms of latency establishment are not yet fully understood, epigenetic chromatin silencing of the HIV-1 long terminal repeat (LTR) region (10), cellular deactivation, limitation of transcription or elongation factors (11), transcriptional interferences (12), and microRNA silencing (13) were shown to be involved. Hence, small pharmacological molecules targeting these major mechanisms have been under review for more than a decade and are termed latency reversal agents (LRAs). Among them, histone deacetylase inhibitors (HDACis) inducing chromatin remodeling (14), protein kinase C activators (PKCas) (15, 16) recruiting transcription factors to the HIV-1 LTR, and bromodomain and extra terminal domain inhibitors (BETis) (17, 18) favoring p-TEFb/Tat association represent the most studied families of LRAs.

The reactivation potency of these main LRA classes has been extensively studied in CD4⁺ T cells and cell line models with encouraging results (18–24), but none of the LRAs has been able to diminish significantly the size of the HIV-1 reservoir in vivo (22, 25, 26). Combining different classes of LRAs enhanced these agents’ potencies (23, 27, 28). However, in vivo data of combinatory LRAs are still lacking. Because LRAs are nondiscriminant and since transcriptional and functional features vary among cell types, these treatments could result in different outcomes on HIV-1 propagation in non-T-cell populations. For example, romidepsin was shown to decrease de novo HIV-1 infection in CD4⁺ T cells (29) but not in monocyte-derived macrophages (MDMs), where HDACis and JQ1 decrease HIV-1 viral production (30, 31). On the other hand, bryostatin-1 decreases permissiveness to HIV-1 infection and reactivates viral production in lymphoid and myeloid cell lines (24), but its effects on human primary macrophages remain unknown. Moreover, some LRAs were shown to either improve or degrade HIV-1-infected cell clearance functions in NK (32, 33) or in CD8⁺ T cells (34), potentially affecting the kill phase. Hence, with such distinct features among cell types and LRA molecules, these agents need to be better studied on other cell types.

Macrophages were long suspected to harbor latency, fueling many debates. High levels of differentiation, self-renewal capability (35), resistance to HIV-1-associated toxicity (36, 37) and cytotoxic T-lymphocyte responses (38), antiretroviral efflux drug transporter (39), and localization into sanctuaries featured macrophages as excellent candidate for reservoir establishment and latency promotion. However, evidence to support this hypothesis was hindered by the complexity of tissular macrophage sample collection. Nevertheless, Honeycutt et al. (40) have recently described a persistent viral reservoir in myeloid-only mice. Further work evidenced the presence of a lipopolysaccharide-inducible reservoir in urethral macrophages from combination antiretroviral therapy-treated patients (41). These elegant studies provided proof of concept that human tissue macrophages constitute a persistent and inducible HIV-1 reservoir that might harbor latently infected cells.

Nevertheless, few studies have focused on the reactivation potential of LRAs in human primary macrophages. Hence, in this study, we decided to investigate the impact of 3 classes of LRAs used alone or in dual combinations on human macrophage susceptibility to HIV-1 infection and viral production in vitro. Our results show that bryostatin-1 and romidepsin treatments decreased HIV-1 receptors and de novo infection mainly through CD4 downregulation or by increasing SAMHD1 activity, respectively. While none of the LRAs, when used alone or in combination, could enhance either viral production or viral mRNA in infected macrophages, surprisingly, bryostatin-1 induced a major decrease of mature gag protein production.

RESULTS

Treatment with LRAs decreases macrophage susceptibility to HIV-1 infection.

We first tested whether LRAs could alter permissiveness of human macrophages to virus infection. To do so, MDMs were treated with LRAs from the three major classes, bryostatin-1 (PKCa), romidepsin (HDACi), and JQ1 (BETi), used alone or in various combinations, and were then infected with a reporter virus for 16 h before addition of EFZ to ensure a single round of infection. Flow cytometry analysis, shown in Fig. 1A, revealed that all tested LRAs negatively impacted HIV-1 infection. This inhibitory effect was minor for JQ1, mild for romidepsin and its combination with JQ1 (2-fold), and potent with bryostatin-1 used alone or in combinations (10-fold). Interestingly, compared to a treatment with tumor necrosis factor plus interferon (TNF+IFN) (used as a control), which is known to lessen HIV-1 infection (42), romidepsin or bryostatin-1 treatment achieved a more robust impediment of macrophage permissiveness to HIV-1 infection. No additive effect was observed when LRAs were used in dual combinations. LRA treatment did not decrease cell viability, and bryostatin-1 even tends to increase MDM survival, which infers that an LRA-induced reduction in HIV-1 infection rate cannot be attributed to cell death (Fig. 1B). Since the percentage of HSA⁺ cells is relatively low and close to the detection threshold for bryostatin-1, we quantified the p24 content as an indicator of viral production (Fig. 1C). This p24 detection paralleled our flow cytometry observations. Indeed, the use of bryostatin-1 almost abolished viral production.

FIG 1.

Bryostatin-1 and romidepsin decrease HIV-1 infection in MDMs. (A to C) MDMs were treated with the listed LRAs before infection with Bal-HSA. Virus infection was analyzed at 3 days postinfection. (A and B) Flow cytometry analysis of HSA-positive cells relative to the untreated condition (n = 5) (A) or cell viability by dye exclusion (n = 8) (B) is shown. (C) Viral production was quantified based on extracellular p24 detection by ELISA (n = 3). Each symbol represents an independent donor. Horizontal lines indicate the mean value obtained from biological replicates for each condition. One-way analysis of variance (ANOVA) with Dunnett’s multiple-comparison test was used (*, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.005; ****, P ≤ 0.001).

Bryostatin-1 and romidepsin downregulate CD4 and CCR5 expression, respectively.

Since some HDACis (43, 44) and bryostatin-1 (24, 28) have been shown to modulate surface expression of CCR5 and CD4, respectively, in other cell types, we investigated the potential downregulation on human macrophages. To this end, MDMs were treated with the listed LRAs for 6 to 24 h, and CCR5 or CD4 gene and membrane expression were assessed by reverse transcription-quantitative PCR (RT-qPCR) and flow cytometry, respectively (Fig. 2).

FIG 2.

Bryostatin-1 and romidepsin decrease HIV-1 cellular receptors. (A) MDMs were treated with LRAs for 6 h (circles) or 24 h (red squares), and CCR5 (left) or CD4 (right) gene expression was quantified by RT-qPCR. Minimum, mean, and maximum values are represented by horizontal lines. Data were normalized on 18S gene expression for each donor and expressed as fold change over the untreated condition (n = 5). (B) MDMs were treated with LRAs for 24 h and were next analyzed for CCR5 (left) and CD4 (right) membrane expression by flow cytometry. Results are represented as the percentage of positive cells for CCR5 or as the mean fluorescence intensity (MFI) relative to untreated cells for CD4 (n = 4). One-way ANOVA with Dunnett’s multiple-comparison test was used (*, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.005; ****, P ≤ 0.001).

CCR5 gene expression was markedly reduced by more than 10-fold after a 24-h treatment with romidepsin used either alone or in combinations (Fig. 2A, left). This potent effect was not detected with the other LRAs. This transcriptional downregulation was paralleled by a 5-fold decrease in cell surface expression (Fig. 2B, left). Likewise, CD4 gene expression was modestly reduced by brostatin-1 at 6 h posttreatment with a maximum of downregulation at 24 h, alone or in combination with JQ1 (Fig. 2A, right). Even though all LRAs decreased CD4 membrane expression to some extent, 24 h of bryostatin-1 treatment resulted in the most robust diminution (i.e., 4-fold) (Fig. 2B, right). Moreover, a treatment with bryostatin-1 was also accompanied by a 2-fold reduction in the percentage of CD4-positive cells (data not shown).

In summary, our results indicate that bryostatin-1 and romidepsin downregulate HIV-1 receptors in human macrophages.

Reduced HIV-1 infection by bryostatin-1 but not romidepsin relies mainly on receptor downregulation.

To investigate the importance of receptor downregulation on MDM resistance to HIV-1 infection after LRA treatment, we first investigated the effect on viral entry. Macrophages were stimulated for 24 h with the listed LRAs and infected for 2 h with CD4/CCR5-dependent viruses (i.e., Bal-HSA). Trypsin was used to remove cell surface-bound viruses, and virus entry was assessed by the quantification of intracellular p24 content by enzyme-linked immunosorbent assay (ELISA). As depicted in Fig. 3A, Bal-HSA entry was dampened by every treatment with similar potency for bryostatin-1 and romidepsin when used as a single LRA. Romidepsin combinations achieved the most potent reduction, while the bryostatin-1 and JQ1 combination seemed to rescue HIV-1 entry compared to bryostatin-1 alone. These results suggest that antiviral effects of LRAs rely on receptor downregulation. We then quantified early steps of the HIV-1 intracellular cycle to investigate the contribution of postentry mechanisms to LRA antiviral effects. LRA-stimulated cells infected with Bal-HSA were analyzed for HIV-1 total and completed RT DNA content by qPCR. Cells treated with the reverse transcriptase inhibitor efavirenz (EFZ) were used as negative controls for RT product completion. As expected, total and completed transcripts decreased under all conditions except when using JQ1 alone (Fig. 3B and C). Treatment with bryostatin-1 achieved a similar decrease in HIV-1 total and completed transcripts up to 70%, consistent with entry downregulation, while romidepsin treatment more profoundly affected the detection of completed (70%) versus total (50%) HIV-1 transcripts. Since total transcripts are a marker of all HIV-1 DNA products, including early and late RT, a specific decrease of completed transcripts suggests a postentry impairment directly affecting reverse transcription completion rather than its initiation (Fig. 3C).

FIG 3.

Dependency of CD4 and CCR5 on de novo infection modulation by LRAs. (A) Intracellular p24 content was quantified in cells treated with LRAs for 24 h before incubation with Bal-HSA viruses (75 ng of p24) for 2 h. Results are represented as p24 concentration relative to untreated cells determined by ELISA (n = 4). (B to D) Cells were first treated for 8 h with LRAs and next infected for 16 h with Bal-HSA (B and C) or VSV-g-HSA (D) viruses. (B and C) Cells were cultured for another 24 h, after which total DNA was extracted to quantify total (B) and completed (C) HIV-1 RT products by qPCR (n = 3). Data are displayed as HIV-1 products copies relative to the untreated condition. EFZ (100 nM) treatment was used as a negative control of RT completion. (D) Cells were incubated for another 3 days in complete culture medium, and infection rate was assessed as the percentage of HSA-positive cells relative to untreated cells by flow cytometry (n = 6). One-way ANOVA with Dunnett’s multiple-comparison test was used (*, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.005; ****, P ≤ 0.001).

Our results imply the existence of an early postentry HIV-1 restriction mechanism triggered by romidepsin, probably during the reverse transcription step. To confirm this hypothesis, we used single-round HSA-expressing HIV-1 viruses pseudotyped with the vesicular stomatitis virus glycoprotein (VSV-g HSA) to bypass CD4 and CCR5 binding for cellular entry. Results depicted in Fig. 3D show that bryostatin-1 alone and in combination with JQ1 induced only a minor decrease of VSV-g HSA-infected cells (25%) with high donor-to-donor variation. This reduction was far weaker than what could be observed using Bal-HSA viruses (Fig. 1A), underlining the major contribution of CD4 receptors in the bryostatin-1-mediated diminution of HIV-1 infection. However, romidepsin tends to achieve a similar reduction of HIV-1 infection regardless of the entry route (Fig. 1A and 3D), suggesting a receptor-independent antiviral feature.

Overall, these results imply that (i) HIV-1 restriction by romidepsin is independent of its effect on receptor expression and relies on a postentry mechanism and (ii) a CD4 decrease represents the main but not the only mechanism of bryostatin-1 antiviral features.

Romidepsin antiviral activity is linked to its effect on SAMHD1 activity.

SAMHD1 is a highly potent restriction factor in macrophages, preventing viral DNA synthesis and more profoundly affecting completed versus early transcripts (45, 46), reminiscent of the romidepsin-induced pattern. Moreover, previous work uncovered potent SAMHD1 activity upregulation as a consequence of HDACi stimulation in MDMs (47). Thus, we investigated whether romidepsin or bryostatin-1 treatment could modulate SAMHD1 activity. Total SAMHD1 protein or its phosphorylated inactive form was quantified by Western blotting on uninfected and LRA-treated macrophages (Fig. 4A and B). Romidepsin treatment did not affect total SAMHD1 protein level but induced a nearly complete depletion of the phosphorylated SAMHD1 form. Conversely, bryostatin-1 resulted in a downregulation of both total and inactive forms of SAMHD1, which, despite high donor-to-donor variations, ultimately did not affect these protein ratios (Fig. 4B).

FIG 4.

Romidepsin- but not bryostatin-1-mediated effect is linked with an upregulation of SAMHD1 activity. (A and B) Immunoblot experiments of 5 donors stimulated for 24 h with the listed LRAs were probed on different membranes with antibodies directed against total or the phosphorylated inactive form of SAMHD1. (A) One representative donor is shown along with β-tubulin band loading control. (B) Protein band intensities normalized on their corresponding tubulin or actin band and compared to those of untreated cells ± standard deviations (SD) (n = 5). (C to E) MDMs were pretreated with Vpx-containing VLPs for 2 h, stimulated for 8 h with LRAs, infected for 16 h with VSV-g (C and D) or Bal-HSA (E), and finally treated with EFZ for 3 additional days. Quantifications of HSA-positive cells (C) (n = 5) or p24 supernatant detection (D and E) (n = 5) were assessed by flow cytometry and ELISA, respectively. (B) One-way ANOVA with Dunnett’s multiple-comparison test was performed on log₂ ratios of normalized band intensities over untreated conditions. (C to E) One-way ANOVA with Dunnett’s multiple-comparison test was used (*, P ≤ 0.05; ***, P ≤ 0.005; ****, P ≤ 0.001).

To assess whether SAMHD1 is involved in LRA-induced antiviral activity, macrophages were pretreated with Vpx-containing virus-like particles (VLPs) to induce its degradation. Cells were then treated with LRAs and infected with VSV-g-pseudotyped HSA viruses to bypass receptor restrictions. Results depicted in Fig. 4C indicate that SAMHD1 degradation by Vpx can restore HIV-1 infection to the untreated level in romidepsin-treated cells. In contrast, SAMHD1 degradation did not modulate further HIV-1 infection when cells were stimulated with bryostatin-1 (Fig. 3D and 4C). These results were confirmed by assessing extracellular p24 detection with Bal or VSV-g HSA viruses in Vpx-pretreated cells (Fig. 4D and E). Regardless of the virus used, SAMHD1 depletion abolished romidepsin-induced resistance to HIV-1 infection but had no effect on bryostatin-1 inhibition. Again, bryostatin-1 affected more strongly R5-using HIV-1 viruses (approximately 50-fold reduction) than its VSV-g-pseudotyped counterpart (approximately 2-fold reduction), confirming the critical contribution of CD4 receptor to the antiviral effect (Fig. 4D and E). Nevertheless, HIV-1 production impairment remained severe in VSV-g HSA-infected cells despite an infection rate that was almost unchanged. This may confer on bryostatin-1 an additional antiviral feature by restricting HIV-1 production.

To summarize, our results suggest that (i) despite HIV-1 entry modulation, infection downregulation by romidepsin relies mainly on a SAMHD1-dependent mechanism, (ii) bryostatin-1’s effect is mainly CD4 dependent but also relies on the modulation of a SAMHD1-independent restriction, and (iii) bryostatin-1 might regulate late events in the viral life cycle.

Bryostatin-1 rapidly and reversibly decreases HIV-1 viral production in chronically infected cells.

Based on our previous observations and because HIV-1 latency of in vivo macrophages is highly suspected (41), we investigated whether LRA treatment after HIV-1 infection could impact viral production. MDMs infected with HIV-1 were cultured for 11 days under EFZ pressure to promote possible latency events. Cells were then treated with LRAs for 24 h, cell-free supernatants were collected, and their p24 content was quantified by ELISA. Results presented in Fig. 5A revealed that none of the listed LRAs could significantly increase viral production in MDMs. More surprisingly, bryostatin-1 and its combinations caused a drastic reduction (more than 80%) in extracellular p24 content (Fig. 5A), a trend readily detected as soon as 4 h posttreatment (Fig. 5B). The bryostatin-1-mediated reduction in viral production could be attributed to neither a reduction of cell viability (Fig. 5C) nor a reduced rate of HIV-1-positive cells (Fig. 5D). Similar results were obtained using the R5-tropic ADA viral strain, after a 72-h treatment period with LRAs or with human fetal microglia (Fig. 5E, F, and G, respectively). As shown in Fig. 5H, extensive washes of bryostatin-1-treated cells resulted in an increase of virus production as soon as 48 h posttreatment and reached a level similar to that of untreated cells at 72 h. To evaluate cell viability, a 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium assay was performed at 6 days posttreatment, which showed no metabolic changes in bryostatin-1-treated cells and a modest decrease in romidepsin-treated cells (Fig. 5I). Compared to untreated conditions, neither JQ1 nor romidepsin seemed to modulate HIV-1 viral production at any time point under all conditions tested (Fig. 5A and E to H). Together, these results suggest that bryostatin-1 can rapidly but transiently decrease HIV-1 viral production in infected human macrophages.

FIG 5.

Treatment of productively infected cells with bryostatin-1 induces a marked decrease in viral production. (A, C, and D) Bal-HSA-infected cells were kept for 11 days in complete culture medium before treatment with LRAs for 24 h. Culture medium was supplemented continuously with EFZ to prevent new rounds of virus infection. Supernatants and cells (n = 4) were collected to quantify p24 expression by ELISA (A) or cellular viability (C) and infection rate (D) by flow cytometry. (B) MDMs were infected with Bal-HSA viruses for 24 h and cultured for 6 additional days in complete culture medium. Cells were then treated for 4 h with bryostatin-1 supplemented with EFZ. Supernatants were collected either immediately (4 h) or after an additional 20 h following thorough washes (4 to 24 h). Data are depicted as the p24 concentration relative to untreated conditions for 6 different donors determined by ELISA. (E) Cells were infected with R5-tropic ADA viruses (n = 6) and cultured for 6 additional days in complete culture medium. Cells were then treated for 24 h with listed LRAs supplemented with EFZ. Supernatants were collected to quantify p24 expression by ELISA. (F) MDMs were infected for 24 h with Bal-HSA and cultured for 32 additional days in complete culture medium continuously supplemented with EFZ before their treatment for 24 h. Cells were finally washed and cultured for another 2 days with EFZ. p24 quantification by ELISA relative to untreated cells is depicted (n = 4). (G) Fetal microglia were infected with a dual reporter virus called NL4.3 eGFP-IRES-Crimson for 21 or 27 days in the presence of EFZ. Cells were then washed and incubated for 48 h with listed LRAs supplemented with EFZ. Data are displayed as the extracellular p24 mean detection for 2 donors by ELISA. (H and I) Cells were infected with Bal-HSA (n = 4) and cultured for 7 additional days in complete culture medium. (H) Time course of p24 expression relative to untreated cells at the indicated days poststimulation (dps). Cells were extensively washed and EFZ was added after every harvest. Symbols represent mean values from 4 donors. (I) Metabolic activity was evaluated by an MTS-phenazine methosulfate assay at 6 dps. (A, C to F, H, and I) One-way ANOVA with Dunnett’s multiple-comparison test was used. ns, not significant. (B) Student ratio paired t test was used (*, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.005; ****, P ≤ 0.001).

Bryostatin-1 decreases production of HIV-1 mature Gag proteins.

To investigate the cellular origin of virion production impairment by bryostatin-1, the p24 intracellular content in HIV-1-infected and then LRA-treated MDMs was also quantified. As shown in Fig. 6A, JQ1 and romidepsin alone caused a marginal, but not statistically significant, increase in intracellular p24. Meanwhile, bryostatin-1 alone or in combinations decreased p24 detection by approximately 2-fold, suggesting that downregulation of extracellular viral production originates from an intracellular impediment.

FIG 6.

Bryostatin-1 decreases accumulation of mature Gag proteins. (A to F) Bal-HSA-infected cells were kept in complete culture medium for the period indicated in Materials and Methods before LRA treatment for 24 h. Cells (A to E) or extracellular particles (F) were finally lysed to quantify intracellular p24 content by ELISA (A) (n = 5), Gag (B), and spliced Tat (C) mRNA by RT-qPCR (n = 5) or HIV-1 proteins (D to F) by Western blotting (n = 3 to 4). (D and F) Membranes of representative donors are shown. (E) Band intensities for Gag, gp160, and Nef were normalized on the corresponding tubulin band and compared to those of untreated cells. Vertical lines symbolize separately processed membranes. (A to C) One-way ANOVA with Dunnett’s multiple-comparison test (***, P ≤ 0.005). (E) Student paired t tests were performed on log₂ ratios of normalized tubulin band intensities over untreated conditions. Calculated P values for p17 (0.0692), p24 (0.0682), and pr160 (0.0511) are close to statistical significance.

Since the lack of viral production is a hallmark of cells latently infected with HIV-1, we explored the possibility of a transient latency state induced by bryostatin-1 in macrophages. We quantified HIV-1 Gag (Fig. 6B) and spliced Tat mRNA (Fig. 6C) in LRA-treated cells by RT-qPCR to investigate a potential modulation at the transcriptional level. Our results show that none of the listed LRAs could significantly alter the expression of targeted genes, suggesting that bryostatin-1’s effect on viral production could not be attributed to alteration in Gag RNA expression. Next, the abundance of viral precursor and mature Gag, Nef, or Env proteins was analyzed by Western blotting using specific antibodies. As revealed in Fig. 6D and E, bryostatin-1 induced a 2-fold decrease in mature Gag proteins CAp24 and MAp17 proteins, concordant with ELISA-based intracellular p24 determination, while their precursors, pr55Gag and p41, were not affected. A modest accumulation of pr160Gag-Pol was observed in bryostatin-1-treated cells. However, the impact of this variation on mature protein abundance is questionable given the low abundance of this precursor. Conversely, other viral proteins encoded by spliced transcripts, such as Nef or gp160, were not modulated (Fig. 6D). It is worth noting that we could not detect mature proteins resulting from gp160 processing, i.e., gp41 or gp120, probably because of their low abundance on mature virions combined with the low rate of HIV-1 infection.

Impairment of mature Gag proteins may be the consequence of improper pr55Gag proteolysis by the HIV-1 protease. Even though we could not detect any variations of this precursor in whole-cell lysates, we investigated its potential accumulation in extracellular virions. In contrast to whole-cell lysates, pr55Gag was hardly detected, compared to Cap24 protein, in extracellular HIV-1 particles (Fig. 6F), consistent with mature virion content. Nevertheless, as observed in cell lysates, bryostatin-1 treatment induced a major decrease of Cap24 with little to no variation in pr55Gag amounts. Taken together, these results suggest that bryostatin-1 does not modify posttranscriptional mechanisms such as splicing, RNA transport, translation, and proteolytic processing but specifically impairs Gag mature protein accumulation and/or transport by an unknown mechanism.

Ingenol-3-angelate mediates a minor effect on HIV-1 infection and virus production despite showing a better latency reversal in J-Lat cells.

A recent study demonstrated more potent HIV-1 latency reversal features in various subsets of CD4⁺ T cells with ingenol-3-angelate, another PKCa, over bryostatin-1 (48). Thus, we evaluated whether ingenol-3-angelate could alter the susceptibility of MDMs to virus infection, production, and reactivation compared to bryostatin-1. Using the protocols described in the legends to Fig. 1 and 5, we found that pretreatment with ingenol-3-angelate caused only a minor decrease in HIV-1 infection (Fig. 7A). Similarly, inhibition of virus production by chronically infected MDMs was less pronounced with ingenol-3-angelate, with a 2-fold reduction compared to a 4-fold reduction achieved with bryostatin-1 (Fig. 7B). On the other hand, treatment of J-Lat clone 10.6, a widely used model of latency (49), revealed that after TNF+IFN combination, ingenol-3-angelate was the most potent LRA tested, either alone or in combination (Fig. 7C). Similar observations were made when using the J-Lat A2 clone (data not shown). Altogether, these experiments suggest that although ingenol-3-angelate displays a better reactivation potency than bryostatin-1 in a latency T-cell line model, this compound possesses a limited HIV-1 restriction capacity in primary human macrophages. This implies that the antiviral capacities of PKCa family members are not linked to their latency reversal potential in human macrophages.

FIG 7.

Ingenol-3-angelate provides better reactivation potency but weaker HIV-1 restriction than bryostatin-1. (A) LRA-treated MDMs were infected with Bal-HSA and then cultured for 3 days. Flow cytometry analysis of HSA-positive cells relative to the untreated condition is depicted for 4 distinct donors. (B) Bal-HSA-infected cells were kept for 7 days in complete culture medium before their LRA treatment for 24 h under an EFZ pressure. Data are depicted as the extracellular p24 quantification relative to untreated cells by ELISA (n = 5). (C) J-Lat clone 10.6 cells were treated for 24 h with the indicated single or combinatory LRA before their processing by flow cytometry. Latency reversal is depicted as the percentage of GFP-positive cells. One-way ANOVA with Dunnett’s multiple-comparison test was used (*, P ≤ 0.05; ****, P ≤ 0.001).

DISCUSSION

Despite their broad spectrum, few studies were conducted to investigate the overall effect of LRAs on nonlymphoid CD4⁺ T cells during HIV-1 infection. In this regard, macrophages are often an overlooked cell population in HIV-1 studies. However, these cells are one of HIV-1’s natural targets and are well known to be more resistant to virus-induced cytopathic effects and to hold reduced intracellular antiretroviral concentrations. Recent proof of inducible reservoir establishment in vivo and reactivation potency of disulfiram on latent myeloid but not lymphoid cell lines (50) underline the importance of monitoring the putative impact of LRAs on macrophages to ensure the success of the shock-and-kill strategy. The current work was aimed at unravelling this issue by assessing the impact of LRAs on HIV-1 infection and viral production in human primary macrophages.

We first found that LRAs did not impact viability of macrophages when used alone or in combination. Our work points toward an antiviral feature on de novo infection by romidepsin and bryostatin-1, the latter rapidly crippling HIV-1 viral production as well.

As previously shown for panobinostat and SAHA in macrophages, we could associate the romidepsin-mediated antiviral effect on SAMHD1 activity with a complete disappearance of phosphorylated SAMHD1 (47). Compared to its resting counterpart, SAMHD1 is highly phosphorylated in activated CD4⁺ T cells, resulting in a more permissive state for HIV-1 infection (51, 52). Thus, potentiating SAMHD1 activity by HDACis could restrict HIV-1 propagation in this cell subset. However, despite an envelope-independent restriction of HIV-1 infection in activated CD4⁺ T cells by romidepsin, involvement of SAMHD1 in these cells is questionable. Indeed, while panobinostat failed to restrict HIV-1 infection in CD4⁺ T cells (29), its SAMHD1-related antiviral activity in MDMs suggests a cell type specificity (47).

In contrast to romidepsin, the antiviral potential of bryostatin-1 was not related to that of SAMHD1, although a decrease in both total and phosphorylated SAMHD1 protein levels was observed upon bryostatin-1 treatment. Nonetheless, these modulations could influence latent cell reactivation, with recent work proposing that both SAMHD1 phosphorylation and dNTPase activity downregulate HIV-1 LTR gene expression (53). Thus, latent reversing potential of SAMHD1-mediated regulation by HDACis and bryostatin-1 would need further characterization.

The antiviral activity mediated by bryostatin-1 was shown here to be mainly dependent on CD4 downregulation and partially on an unidentified factor limiting the HIV-1 life cycle. Bryostatin-1 is known to downregulate multiple PKC classes, including PKC delta, during sustained stimulations (24, 54). Inhibition of this PKC isoform was shown to interfere with the HIV-1 reverse transcription step during the process of infection (55). Moreover, novel PKC family members such as the delta isoform are involved in the downregulation of CD4 receptors by prostratin, being partially responsible for limiting HIV-1 infection (15, 56, 57). Hence, we suspect HIV-1 de novo restriction by bryostatin-1 is the consequence of PKC delta activation.

Previous work on human MDMs does not concur with our conclusions. Indeed, romidepsin was previously not reported to mediate CCR5 downregulation or inhibition of HIV-1 infection (29, 30). Rather, authors showed a decrease in HIV-1 production by numerous HDACis and JQ1, which was linked to an upregulation of the autophagy process (30, 31). Although we could not definitely exclude this pathway’s involvement, in our hands, no modulations were seen with these LRAs in MDMs or in microglia, and common autophagy inhibitors did not restore HIV-1 production in bryostatin-1-stimulated cells (data not shown). Discrepancies among the aforementioned studies may be attributed to the differences in the MDM differentiation protocol, with serum source (animal versus human) being known to modulate HIV-1 infection linked to transcriptomic alterations (47). Moreover, studies from Campbell and colleagues were performed under constant LRA pressure, and p24 levels were assessed after long-term treatment (48 to 72 h) (30, 31). Considering that romidepsin half-life in plasma has been shown to be 3 h (58), an extended treatment period with LRAs may not depict physiologic behavior. Finally, the absence of EFZ supplementation and a higher concentration of romidepsin used for most of their experiments (50 nM) could also explain these results. Nevertheless, we believe that by using human serum to generate human MDMs, our results might more closely parallel in vivo situations.

Physiological relevancy has always been a major hurdle when investigating human macrophages. Indeed, most studies are conducted on in vitro-generated MDMs rather than tissue-resident macrophages relative to their obvious paucity and technical issues surrounding isolation. Thus, data obtained from MDMs may not fully replicate the complex physiology of tissue-resident macrophages. Nonetheless, we did observe a comparable bryostatin-1-mediated inhibition of HIV-1 replication in primary human fetal microglial cells (Fig. 5G). These results deserve to be validated in other tissue-resident macrophage populations.

In contrast to CD4⁺ T cells, we did not observe any sign of HIV-1 reactivation in MDMs induced by LRAs. Even long-term culture up to 32 days postinfection of MDMs under EFZ pressure to favor potential latency events did not yield an increase in viral production after a treatment for 72 h with LRAs (Fig. 5F). These observations could lead to several conclusions that are not mutually exclusive: (i) HIV-1 latency in in vitro macrophages may not be achievable or may be more uncommon than that in CD4⁺ T cells, (ii) our culture model is not optimal to promote latency events, (iii) latency reversal may not be detected because p24 quantification is not sensitive enough, and/or (iv) selected LRAs and concentrations used in our work might not reverse latency efficiently in MDMs. Thus, additional studies are warranted to assess whether HIV-1 latency and its reversal can be established in human macrophages.

While we did not elucidate the exact mechanism of the underlying bryostatin-1-mediated effect on viral production, our work gives rise to several leads. First, the PKCa ingenol-3-angelate also decreases, albeit to a lesser extent, HIV-1 viral production. Thus, at least 2 different PKCas can dampen HIV-1 viral production within macrophages, suggesting that the mechanism of action is related to PKC activity. The divergence in potency between the two compounds may be associated with different activation potentials mediated by distinct PKC isoforms. Second, bryostatin-1 does not silence the HIV-1 genome or prevent Gag polyprotein, Nef, or gp160 translation. A first hypothesis could involve Gag polyprotein processing. Indeed, pr55Gag precursor is sequentially cleaved by the HIV-1 protease, resulting in subsequent release of p24, p17, p7, and p6 mature proteins (59). Since HIV-1 protease synthesis and activity were not quantified in this work, one could suspect either one of these steps to be modulated. However, we revealed a peculiar pattern of Gag protein modulation whereby precursor pr55Gag and p41 protein did not accumulate despite the decrease of p24 and p17 detection, suggesting that HIV-1 protease activity is not impaired by bryostatin-1.

Another hypothesis could rely on proteasomal rerouting of Gag mature proteins toward degradation. However, incubation of the proteasome inhibitor MG132 during or after bryostatin-1 treatment did not rescue extracellular p24 levels (data not shown), suggesting that proteasomal degradation is not involved in the reduced amounts of p24 and p17. Incidentally, while bryostatin-1 treatment diminished HIV-1 viral production in fetal microglial cells (Fig. 5G), it had no effect on the HIV-1-infected reporter cell line TZM-bl (data not shown) and fetal astrocytes (60). With no reports indicating HIV-1 production inhibition in CD4⁺ T cells, this may point to a myeloid-specific mechanism induced by bryostatin-1 rather than direct regulation of HIV-1 proteins. In this regard, some unique features in HIV-1 production were acknowledged within MDMs and more recently in tissular macrophages (41), whereby instead of the plasma membrane, these cells harbor unusual viral production and retention within intracellular plasma membrane-connected compartments (IPMCs) (61). Although their composition and formation are not yet fully understood, studies linked their origin to specialized plasma membrane domains characterized by specific markers and limited access to the extracellular milieu (61, 62). Since bryostatin-1 is involved in rapid and potent perturbations in MDM morphology (unpublished observations), we hypothesize that such disturbances impact IPMC formation or fate, which could lead to a degradative pathway of mature viruses. In this regard, bryostatin-1 might perturb membrane association of PI(4,5)P2 necessary for Gag binding or tamper with recruitment of ESCRT components, such as TSG101 or ALIX, involved in the normal virus budding process (62).

Finally, owing to their synergic reactivation potencies, LRA combinations were assessed for their potential to modulate HIV-1 infection and viral productions. While some modest additive or antagonistic effects could be observed, we did not witness any major modulation in this work. This could be explained by the intense magnitude of impairments by single LRA agents on HIV-1 infection and viral production combined with an absence of HIV-1 promoting features. Thus, we can expect that LRA combinations would lead to a higher latency reversal without interfering with their respective HIV-1 antiviral effect.

Overall, our study unveiled the impact of some specific LRAs on HIV-1 infection and viral production in human macrophages in vitro. While some features are shared with CD4⁺ T cells, PKCa-induced decrease of viral production revealed cell-specific features. By limiting de novo infection and viral production in productively infected myeloid cells and by their potency to reactivate HIV-1 in latent CD4⁺ T cells, PKCa appears to be a promising LRA class to limit HIV-1 propagation during the shock-and-kill strategy. However, owing to their different side effects on CD8 and NK cells, careful assessment of LRA impact in other cell types involved in HIV-1 pathogenesis would be required to select the most potent and safest agents for an effective HIV-1 eradication strategy.

MATERIALS AND METHODS

Ethics statement and cell culture.

The current study was approved by the Bioethics Committee from the Centre Hospitalier Universitaire de Québec-Université Laval. Peripheral blood samples were collected from healthy donors, respecting guidelines of the Institutional Bioethics Committee with written consent provided by all participants.

Peripheral mononuclear cells were collected after Ficoll-Hypaque (Corning Life Science, Tewksbury, MA) gradient centrifugation and seeded for 2 h at 37°C to allow adherence of monocytes. Cells were then washed extensively with Dulbecco's phosphate-buffered saline (DPBS) (Corning Life Science) to remove nonadherent cells. Monocytes were maintained for 3 days in RPMI 1640 culture medium (Corning Life Science) supplemented with 10% (vol/vol) AB human serum (Valley Biomedical, Winchester, VA), penicillin-streptomycin (Gibco, Thermofisher Scientific, Waltham, MA) (here referred to as complete culture medium), and 25 ng/ml macrophage colony-stimulating factor (GenScript, Piscataway, NJ). Cells were washed extensively in DPBS and maintained for 3 additional days in complete culture medium to obtain nonpolarized monocyte-derived macrophages (MDMs). MDM purity was assessed in a previous work based on CD68 expression and was greater than 97% (63). Next, cells were washed extensively to remove nonadherent cells, incubated for 1 h at 37°C with accutase (Invitrogen, Thermofisher Scientific), and detached with gentle scraping using a cell scraper (Sarstedt, Nümbrecht, Germany). MDMs were seeded at various cell concentrations, usually 1.5 × 10⁵ to 2 × 10⁵ for a minimum of 24 h before further processing either in ultra-low attachment plates (Corning) for experiments requiring flow cytometry studies or with regular tissue culture-treated plates (Corning) when cell detachment was not needed.

Human embryonic kidney (HEK293T) cells were kindly provided by Warner C. Greene (The J. Gladstone Institutes, San Francisco, CA). These cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) (Invitrogen) supplemented with 10% (vol/vol) heat-inactivated fetal bovine serum (FBS) (Corning Life Science) and antibiotics. J-Lat cell clone 10.6 was obtained from the NIH AIDS reagent program (Germantown, MD) and maintained in RPMI 1640 culture medium supplemented with 10% (vol/vol) heat-inactivated FBS and antibiotics. Fetal microglia were isolated and cultured as previously described (64).

Antibodies and reagents.

Bryostatin-1 (used at 20 nM), ingenol-3-angelate (20 nM) (Sigma-Aldrich, Oakville, ON), JQ1 (500 nM), and romidepsin (5 nM) (Cayman Chemicals, Ann Arbor, MI) diluted in dimethyl sulfoxide (DMSO), with a final DMSO concentration ranging from 0.005% to 0.02%, were used either alone or in combinations. Untreated cells (0.02% DMSO) and cells treated with a combination of TNF (20 ng/ml) and IFN-γ (10 ng/ml) from BioLegend (San Diego, CA) were used as controls.

Hybridomas producing 183-H12-5C and 31–90-25 monoclonal antibodies (MAbs), which recognize different epitopes of the HIV-1 major viral core protein p24Gag, were supplied by the NIH AIDS Reagent Program and ATCC (Manassas, VA), respectively.

The following reagents were obtained through the NIH HIV Reagent Program, Division of AIDS, NIAID, NIH: efavirenz (EFZ), ARP-4624 (contributed by DAIDS/NIAID), polyclonal anti-human immunodeficiency virus type 1 gp120 (rabbit) and polyclonal anti-human immunodeficiency virus type 1 p17 protein (antiserum, rabbit), ARP-4811 (contributed by Paul Spearman and Lingmei Ding). Mouse monoclonal SAMHD1 (I19-18) and rabbit monoclonal phospho-SAMHD1 (Thr592) (D7O2M) were purchased from Millipore (Etobicoke, ON, Canada) and Cell Signaling Technology (Danvers, MA), respectively. Anti-actin (C-2) from Santa Cruz Biotechnology (Dallas, TX) and anti-β-tubulin (ab6046) from Abcam (Cambridge, UK) were used to normalize protein contents.

Monoclonal mouse anti-Nef clone AG11 (MAB899), phycoerythrin-labeled anti-mouse CD24 antibody clone M1/69 (12-0242-83), fixable viability dyes eFluor 450 (65-0863-18) and 780 (65-0865-18), and 2.5% trypsin were all purchased from ThermoFisher Scientific. Allophycocyanin (APC) mouse anti-human CD4 clone RPA-T4 (555349) and APC mouse anti-human CD195 clone 2D7 (561748) were purchased from BD Bioscience (San Jose, CA).

Plasmid and virus stock production.

pNL4.3Bal-IRES-HSA is an infectious molecular clone of HIV-1 composed of the X4 tropic virus NL4.3 backbone, within which the env gene was replaced by the R5 tropic Bal strain and contained the murine reporter gene heat-stable antigen (HSA or CD24) (65). Cell surface expression of the HSA protein allows quantification and sorting of cells productively infected with HIV-1. Plasmid coding for wild-type R5-tropic ADA was used to generate fully infectious ADA viruses (66). Pseudotyped HSA viruses expressing the vesicular stomatitis virus VSV glycoprotein (VSV-g), here called VSV-g-HSA, were produced by cotransfection of an env-deficient vector (pNL4.3Δenv-IRES-HSA) with the VSV-g coding vector (pHCMV-G). NL4.3 eGFP-IRES-Crimson was produced as previously described (60). Vpx-containing virus-like particles (VLPs) were obtained by cotransfecting HEK293T cells with pHCMV-G and a Vpx plasmid (pSIV3+) (67). The pSIV3+ vector was a kind gift from F. Margottin-Goguet (Université Paris Descartes, Paris, France). Viruses and VLPs were produced by calcium chloride transfection of HEK293T cells. After 48 h of transfection, cell supernatant containing particles was collected, filtered, and ultracentrifuged at 28,000 × g for 1 h at 4°C in an Optima L-90K ultracentrifuge apparatus (Beckman Coulter, Brea, CA). Pelleted viruses and VLPs were resuspended in endotoxin-free DPBS, aliquoted, and stored at −80°C for further use. To normalize the quantity of inoculated viruses in each experiment, p24 viral protein content was quantified by our in-house sandwich enzyme-linked immunosorbent assay (ELISA) using 183-H12-5C and 31–90-25 monoclonal antibodies specific for p24 (68). For infection studies, unless stated otherwise, viruses (20 ng of p24) were inoculated per 10⁵ cells.

HIV-1 infection and treatment with LRAs.

MDMs were plated in complete culture medium and treated for 8 h with LRAs before being infected with NL4.3 Bal-HSA or VSVg-HSA virus for 16 h. Alternatively, cells were not preincubated or were preincubated with Vpx-containing VLPs for 2 h before treatment with LRAs and their subsequent virus infection. SAMHD1 degradation by Vpx-containing VLPs was assessed by Western blotting on two donors (data not shown). After gentle washes with PBS, complete culture medium was replaced and supplemented with the anti-HIV-1 agent EFZ (100 nM) to prevent additional rounds of infection. After 3 additional days, cells were finally detached and processed for HSA cell surface quantification by flow cytometry, or cell-free supernatants were collected to quantify extracellular p24 content by ELISA.

In some experiments, MDMs were infected for 16 h with NL4.3 Bal-HSA or ADA viruses, washed, and cultured in complete culture medium, supplemented- when specified- with EFZ, for the indicated days postinfection. Half of the medium was changed every 3 days. After several days, cells were washed with PBS and stimulated with LRAs for 24 h in complete culture medium with EFZ. Supernatant containing extracellular p24 was collected and quantified with our in-house p24 ELISA. To evaluate the long-term effect of LRAs, supernatants were collected at different time points poststimulation, and cells were then washed extensively with PBS and were finally supplemented with complete culture medium containing EFZ. Cells were finally assessed for metabolic activity after 6 days of LRA stimulation using CellTiter 96 AQueous nonradioactive cell proliferation assay (Promega, Madison, WI) by following the manufacturer’s instructions. In some other experiments, cells were detached and processed for viability and HSA cell surface quantification by flow cytometry.

J-Lat reactivation assay.

J-Lat clone 10.6 cells were seeded at 10⁵ cells per well in a 96-well plate for 24 h and stimulated with listed LRAs for another 24 h of cell culture. J-Lat cells were then collected, stained with a viability dye, and fixed in 2% formaldehyde, and reactivated cells were quantified based on their green fluorescent protein (GFP) expression by flow cytometry.

Gene expression.

MDMs were plated in complete culture medium and treated with LRAs for 6 or 24 h, after which total mRNA was extracted by following instructions from the manufacturer (Macherey-Nagel’s NucleoSpin RNA kit; Duren, Germany). Purified mRNA was reverse transcribed into cDNA using Maloney murine leukemia virus reverse transcriptase (Promega), random primers (Roche, Basel, Switzerland), and deoxynucleoside triphosphate mix (Thermofisher Scientific). Gene expression of CD4, CCR5, Tat splice, and Gag were quantified by qPCR using specific listed oligonucleotides (Table 1) and PowerUp SYBR green master mix (Applied Biosystems) on a Quanstudio 6 Flex system apparatus (Applied Biosystems). Amplified target genes were normalized based on 18S RNA expression using the 2^−ΔΔCT method (69). No variations in 18S cycle threshold were observed between LRA-treated or untreated cells.

TABLE 1.

List of primers used in this study

Target	Forward sequence (5′–3′)	Reverse sequence (5′–3′)
CCR5	TCTCTTCTGGGCTCCCTACA	CTGAACTTCTCCCCGACAAA
CD4	TCAGTATGCTGGCTCTGGAAACCT	AGACCTTTGCCTCCTTGTTCTCCA
Tat	GAAGCATCCAGGAAGTCAGC	GGAGGTGGGTTGCTTTGATA
Gag	AGTAAGAATGTATAGCCCTACCAGCAT	CTTAGAGTTTTATAGAACCGGTCTACATAGTC
18S	TAGAGGGACAAGTGGCGTTC	CGCTGAGCCAGTCAGTGT
β-Globin	TGGTCTATTTTCCCACCCT	TGGCAAAGGTGCCCTTGA
AA55	CTGCTAGAGATTTTCCACACTGAC
M661	CCTGCGTCGAGAGATCTCCTCTG
M667	GGCTAACTAGGGAACCCACTGC
HIV-1 probe	5′-6-carboxyfluorescein-TAGTGTGTGCCCGTCTGTTGTGTGAC-black hole quencher-3′
β-Globin probe	5′-VIC-TCTGTCCACTCCTGATGCTG-nonfluorescent quencher-minor groove binder-3′

For spliced Tat and Gag gene expression, MDMs were inoculated with Bal-HSA viruses for 24 h in complete culture medium. Cells were then washed, cultured in complete culture medium supplemented with EFZ for 6 additional days, and stimulated for 24 h with LRAs and EFZ. MDM mRNAs were finally extracted and processed as described above.

Intracellular p24 quantification and entry assay.

For virus entry assay, MDMs (1.5 × 10⁵) were treated with LRAs for 24 h and next inoculated with 75 ng of NL4.3 Bal-HSA for 2 h at 37°C. For intracellular assays, MDMs were infected for 6 to 8 days with NL4.3 Bal-HSA viruses with EFZ added after 24 h of infection. Finally, MDMs were stimulated with LRAs for 24 h. Cells were then extensively washed with PBS and incubated with trypsin for 5 min at 37°C to remove uninternalized virions and treated with lysis buffer (0.05% [vol/vol] Tween 20, 2.5% [vol/vol] Triton X-100, and 2% [vol/vol] trypan blue in PBS) before being frozen at −20°C for further use. After one freeze/thaw cycle to ensure proper cell lysis, p24 was quantified using our in-house p24-specific ELISA.

Western blotting.

Uninfected MDMs or MDMs infected with HIV-1 for 6 to 7 days (8 × 10⁵) were treated with LRAs and EFZ for 24 h, washed extensively, and lysed in 1× radioimmunoprecipitation assay buffer (NaCl at 150 mM, EDTA at 5 mM, Tris at 50 mM, NP-40 at 1%, sodium deoxycholate at 0.5%, SDS at 0.1%) supplemented with phosSTOP phosphatase inhibitor (Roche) and protease inhibitor cocktails (Sigma-Aldrich). Protein amount was then quantified using the Pierce bicinchoninic acid protein assay from Thermofisher Scientific.

Alternatively, cell supernatants were collected, filtered, and ultracentrifuged at 28,000 × g for 1 h at 4°C in an Optima L-90K ultracentrifuge apparatus. Extracellular particles were then processed similarly to whole-cell lysate.

Proteins extracts (5 to 20 μg) were separated by SDS-PAGE on 8% to 16% polyacrylamide gel, transferred on Immobilon-P polyvinylidene difluoride membrane, 0.45-μm pore size (Thermofisher Scientific), and blocked for 1 h at room temperature in Tris-buffered saline with 0.1% Tween 20 (TBST) supplemented with 5% milk. After several washes with TBST, membranes were incubated overnight at 4°C with specific primary antibodies. Probed membranes were washed thoroughly with TBST and incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies (Jackson Immunoresearch, West Grove, PA). Proteins were finally revealed using Amersham ECL select (GE Healthcare Life Sciences, Mississauga, ON), and signal was captured on a Fusion FX7 Spectra chemical documentation apparatus. Signal quantification was assessed with Fusion software (Vilber-Lourmat, Collégien, France) and normalized on actin or β-tubulin signal for each condition. For extracellular particles, identical volumes of protein extracts were loaded onto the gel.

HIV-1 DNA quantification.

MDMs (4 × 10⁵) were treated for 8 h with LRAs or EFZ before inoculation with 80 ng of NL4.3 Bal-HSA for 16 h. Stock viruses were previously treated for 45 min with DNase I (Roche) at room temperature to degrade remaining HIV-1 plasmid DNA input from virus production in 293T cells. MDMs were then washed and cultured for an additional 24 h in EFZ supplemented complete medium to prevent multiple rounds of infection. Cells were finally washed and lysed using a NucleoSpin tissue extraction kit by following the manufacturer’s instructions (Macherey-Nagel). HIV-1 DNA products were all quantified by qPCR using TaqMan fast advanced master mix (Applied Biosystems) and HIV-1 TaqMan probes on a QuantStudio 6 apparatus (Applied Biosystems). For HIV-1 RT products, 25 ng of extracted DNA was amplified using AA55/M667 or M661/M667 primers for total and completed transcripts, respectively (70). The number of DNA copies obtained in every assay was normalized using the β-globin gene copy number for each diluted sample. Standard curves for HIV-1 replication intermediates were obtained with a serial dilution of the NL/Bal-HSA plasmid, starting at 6 × 10⁶ copies. All probes and primers were purchased from Integrated DNA Technologies, except for the β-globin probe (Applied Biosystems), corresponding to sequences listed in Table 1.

Flow cytometry.

Cells were detached with 5 mM PBS-EDTA by 30 min of incubation at 37°C and washed twice with the same buffer. Cells were then incubated with a fixable viability dye for 30 min at 4°C. For cell surface labeling, cells were blocked for 30 min at 4°C with blocking buffer (PBS containing 5 mM EDTA, 1% BSA, 20% normal goat serum, and 10% AB-human serum) and stained in this buffer with specific antibodies for 15 min at 4°C. Finally, cells were fixed in a 2% formaldehyde solution for 30 min at 4°C. Flow cytometry experiments were performed on a BD FACSCelesta (BD Biosciences) apparatus and analyzed on FlowJo software version 10 for Windows.

Statistical analysis and graphical display.

All statistical analyses were performed using GraphPad Prism, version 9.03, on raw data. Test description and number of independent donors (n), portrayed as symbols, are indicated in the figure legends. Despite matching symbols, donors from different experiments may be unrelated. Statistical tests on raw data expressed as percentages and p24 concentrations were calculated on their logit and log-transformed values, respectively. Data compared to the untreated condition were considered statistically significant for P values of ≤0.05.