Reactivation of latent HIV-1 by the glucocorticoid receptor modulator AZD9567

Abstract

One key determinant of HIV-1 latency reversal is the activation of the viral long terminal repeat (LTR) by cellular transcription factors such as NF-κB and AP-1. Interestingly, the activity of these two transcription factors can be modulated by glucocorticoid receptors (GRs). Furthermore, the HIV-1 genome contains multiple binding sites for GRs. We therefore hypothesized that glucocorticoids and other GR modulators may influence HIV-1 latency and reactivation. To investigate how GR signaling affects latent HIV-1 reservoirs, we assembled a representative panel of GR modulators including natural steroidal agonists, selective and non-selective GR modulators, and clinically approved GR-modulating drugs. The effects of these compounds on HIV-1 reactivation were assessed using latently HIV-1-infected cell lines and primary cells, as well as reporter assays that monitored GR and LTR activities. We found that AZD9567 (Mizacorat), a non-steroidal partial GR agonist, reactivates latent HIV-1 in both lymphoid and myeloid cell lines and primary CD4+ T cells. Conversely, the GR antagonist mifepristone suppresses HIV-1 LTR-driven gene expression. Mechanistic analyses revealed that AZD9567-mediated reactivation partially depends on both GR and AP-1 binding sites in the LTR. In summary, we, here, identify the GR modulator AZD9567 as novel latency-reversing agent that activates LTR-driven gene expression, which may aid in advancing current shock-and-kill approaches in the treatment of HIV-1 infection.IMPORTANCELatently infected cells of people living with HIV are constantly exposed to fluctuating levels of glucocorticoid hormones such as cortisol. In addition, many HIV-infected individuals regularly take corticosteroids as anti-inflammatory drugs. Although corticosteroids are known to affect the activity of the viral long terminal repeat (LTR) promoter and influence ongoing HIV-1 replication, relatively little is known about the effect of corticosteroid hormones and other glucocorticoid receptor (GR) modulators on latent HIV-1. By systematically comparing natural and synthetic GR modulators, we, here, identify a first first-in-class, oral, partial GR agonist that reactivates latent HIV-1 from different cell types. This drug, AZD9567, was previously tested in clinical trials for rheumatoid arthritis. Mutational analyses shed light on the underlying mode of action and revealed transcription factor binding sites in the HIV-1 LTR that determine responsiveness to AZD9567.

Keywords: AP-1; HIV; LTR; glucocorticoid receptor; latency; reactivation.

Fig 1

Glucocorticoid response elements in the HIV-1 genome and glucocorticoid receptor expression in HIV-1 target cells. (A) GRs (gray) may activate (top) or repress (bottom) transcription by directly binding to GRE in the DNA and/or by interfering with other transcription factors (TF, beige). (B) 15 bp core binding sequence of GR homodimers. The canonical consensus binding site is shown on top. Nucleotides that are enriched at experimentally determined GR binding sites are highlighted in green at the bottom. (C) Positions of GREs in the HIV-1 genome are indicated in green. (D) Sequence logo plots illustrating the conservation of GRE I–III and GRE_vif in the genomes of the HIV-1 subtype reference sequences (four representative sequences for each subtype). Green circles indicate the presence of nucleotides enriched in GREs. Beige circles indicate nucleotides that do not match the canonical GR binding site (see panel B). (E) Endogenous GR protein levels in U1, ACH-2, J-Lat 10.6, HEK293T, and primary CD4+ cells were analyzed by Western blotting. CD4+ cells were activated with phytohemagglutinin. HEK293T cells transfected with a GR expression plasmid served as positive control.

Fig 2

AZD9567 reactivates latent HIV-1 in lymphoid and myeloid cell lines. (A) J-Lat 10.6, ACH-2, and U1 cells were stimulated with increasing amounts of the indicated GR modulators (0.01 µM–10 µM) or their respective solvents (methanol, water, ethanol, or DMSO). Forty-eight hours or 96 h later, cells were fixed and the percentage of reactivated cells was determined by flow cytometry. Mean values of three to five biological replicates are shown. (B) J-Lat 10.6, ACH-2, and U1 cells were stimulated with increasing amounts of AZD9567 (0.01 µM–10 µM) or DMSO (0.0001%–0.1%) for 24 h, 48 h, 72 h, and 96 h and analyzed as described for (A). Data are shown as mean ± SEM of at least three biological replicates (n = 3–5). Mixed-effect analysis followed by Bonferroni’s multiple comparison test was used to compare cells treated with AZD9567 and cells treated with DMSO (*P < 0.01, **P < 0.01, ***P < 0.001, ****P < 0.0001).

Fig 3

AZD9567 reactivates latent HIV-1 in primary CD4+ T cells. (A) HIV-1 reporter virus expressing firefly luciferase via an internal ribosome entry site (IRES) and lacking an intact env gene. (B) Experimental setup: CD4+ T cells from healthy donors were infected with the reporter virus shown in (A), in the presence of interleukin 2 (IL-2), but without prior activation. Five days later, cells were treated with increasing amounts of PMA, AZD9567, or DMSO. Luciferase activity was determined 24 h and 48 later. (C) Cells were infected, stimulated, and analyzed as described in (B). Mean values ± SD of five independent donors, analyzed in technical triplicates, are shown.

Fig 4

Mechanisms of AZD9567-mediated HIV-1 reactivation. (A) Jurkat cells (clone E6.1) were stimulated with AZD9567 (10 µM) or DMSO (0.1%). Six hours later, the culture medium was removed and transferred to J-Lat 10.6 cells. In parallel, some J-Lat 10.6 cells were directly stimulated with 10 µM AZD9567. After 48 h, cells were fixed and percentage of GFP-positive cells was determined using flow cytometry to monitor reactivation. Exemplary primary data of one out of three biological replicates are shown in the center. Mean values ± SEM of three independent biological replicates are shown on the right. (B) The effect of AZD9567 on HIV-1 LTR activity was analyzed using a luciferase reporter construct (top). HEK293T cells were co‐transfected with the reporter plasmid expressing firefly luciferase under the control of the HIV-1 LTR promoter and a construct expressing Gaussia luciferase under the control of a minimal promoter. Cells were additionally co-transfected with an empty vector control (left) or a GR expression plasmid (right). Six hours post‐transfection, cells were stimulated with the indicated amounts of AZD9567 (0.01 µM–10 µM) or DMSO (0.0001%–0.1%). Firefly luciferase activity was determined and normalized to Gaussia luciferase activity. Mean values ± SEM of biological replicates (n = 3–16) are shown and were analyzed by two‐way analysis of variance with Bonferroni’s multiple comparison test (**P < 0.01, ****P < 0.0001).

Fig 5

Effects of GR modulators on the activity of the HIV-1 LTR promoter. HEK293T cells were transfected with an HIV-1 LTR firefly reporter plasmid (A) alone or (B) together with a GR expression plasmid. Cells were additionally co-transfected with a Gaussia luciferase reporter for normalization. Six hours post‐transfection, cell culture medium was removed and fresh medium containing GR modulators (0.01 µM–10 µM) or their respective solvents (0.0001%–0.1%) were added. One day post‐transfection, firefly luciferase activity was determined and normalized to Gaussia luciferase activity. Mean values ± SEM of three biological replicates (n = 3) are shown and were analyzed by two‐way analysis of variance with Bonferroni’s multiple comparison test (*P < 0.05).

Fig 6

Disruption of GRE I–III partially abrogates AZD9567-mediated HIV-1 LTR activation. (A) Cartoon illustrating the localization of GRE I–III in the HIV-1 LTR. Mutations disrupting the GREs are shown in red. (B) HEK293T cells were co‐transfected with a reporter plasmid expressing firefly luciferase gene under the control of the wild-type LTR or a mutated LTR with disrupted GRE I, II, or III. Cells were co-transfected with a construct expressing Gaussia luciferase under the control of a minimal promoter for normalization. Six hours post‐transfection, cells were stimulated with AZD9567 (0.0.1 µM–10 µM) or DMSO (0.0001%–0.1%). Thirty hours post-transfection, firefly luciferase activity was determined and normalized to Gaussia luciferase activity. Mean values ± SEM of biological replicates (n = 3–16) are shown. (C) HEK293T cells were co-transfected with a firefly luciferase reporter plasmid harboring either the wild-type LTR or a mutant LTR, in which GRE I–III were disrupted simultaneously. Cells were co-transfected with a construct expressing Gaussia luciferase under the control of a minimal promoter for normalization and analyzed essentially as described in (B). Mean values ± SEM of at least four biological replicates (n = 4–16) are shown. (D) HEK293T cells were co‐transfected, stimulated, and analyzed as in (C). Half of the samples were additionally co-transfected with an expression plasmid for GR. Mean values ± SEM of four independent biological replicates are shown. (E) HEK293T cells were co-transfected with a reporter plasmid containing the GRE I–III sequences of the HIV-1 LTR upstream of a minimal promoter, the Gaussia luciferase reporter for normalization, and a GR expression plasmid or the respective vector control. Cells were stimulated with increasing doses of AZD9567 and analyzed as described in (B–D). Data are shown as mean ± SEM (n = 3–4). Data were analyzed by two‐way analysis of variance with Bonferroni’s multiple comparison test (*P < 0.05, ***P < 0.001).

Fig 7

Efficient AZD9567-mediated HIV-1 reactivation depends on an AP-1 binding site in the LTR. (A) Schematic representation of the HIV-1 LTR reporter construct where the two NF-κB binding sites or an adjacent AP-1 binding site was disrupted. Mutated residues are highlighted in red. (B, C) HEK293T cells were co‐transfected with a reporter plasmid harboring either the wild-type LTR or a mutant with (B) disrupted NF-ᴋB binding sites or (C) a disrupted AP-1 binding site, and a construct expressing Gaussia luciferase under the control of a minimal promoter. Six hours post‐transfection, cells were treated with AZD9567 (0.0.1 µM–10 µM) or DMSO (0.0001%–0.1%). Thirty hours post-transfection, firefly luciferase activity was determined and normalized to Gaussia luciferase activity. Data are shown as mean ± SEM of three biological replicates (n = 3) and were analyzed by two‐way analysis of variance with Bonferroni’s multiple comparison test (*P < 0.05). (D) HEK293T cells were co-transfected with firefly luciferase LTR reporter constructs harboring mutation in the indicated transcription factor binding sites. Mean values ± SEM from three independent experiments each performed in triplicate are shown.