Long lasting control of viral rebound with a new drug ABX464 targeting Rev - mediated viral RNA biogenesis

Abstract

Background: Current therapies have succeeded in controlling AIDS pandemic. However, there is a continuing need for new drugs, in particular those acting through new and as yet unexplored mechanisms of action to achieve HIV infection cure. We took advantage of the unique feature of proviral genome to require both activation and inhibition of splicing of viral transcripts to develop molecules capable of achieving long lasting effect on viral replication in humanized mouse models through inhibition of Rev-mediated viral RNA biogenesis.

Results: Current HIV therapies reduce viral load during treatment but titers rebound after treatment is discontinued. We devised a new drug that has a long lasting effect after viral load reduction. We demonstrate here that ABX464 compromises HIV replication of clinical isolates of different subtypes without selecting for drug resistance in PBMCs or macrophages. ABX464 alone, also efficiently compromised viral proliferation in two humanized mouse models infected with HIV that require a combination of 3TC, Raltegravir and Tenofovir (HAART) to achieve viral inhibition in current protocols. Crucially, while viral load increased dramatically just one week after stopping HAART treatment, only slight rebound was observed following treatment cessation with ABX464 and the magnitude of the rebound was maintained below to that of HAART for two months after stopping the treatment. Using a system to visualize single HIV RNA molecules in living cells, we show that ABX464 inhibits viral replication by preventing Rev-mediated export of unspliced HIV-1 transcripts to the cytoplasm and by interacting with the Cap Binding Complex (CBC). Deep sequencing of viral RNA from treated cells established that retained viral RNA is massively spliced but importantly, normal cellular splicing is unaffected by the drug. Consistently ABX464 is non-toxic in humans and therefore represents a promising complement to current HIV therapies.

Conclusions: ABX464 represents a novel class of anti-HIV molecules with unique properties. ABX464 has a long lasting effect in humanized mice and neutralizes the expression of HIV-1 proviral genome of infected immune cells including reservoirs and it is therefore a promising drug toward a functional cure of HIV.

Figure 1

ABX464 Inhibits HIV-1 production in PBMC- and macrophages-infected cells. a Drawing of 10-chloro-2,6-dimethyl-2H-pyrido[3′,4′:4,5]pyrrolo[2,3-g]isoquinoline (IDC16), 8-chloro-N-(4-(trifluoromethoxy)phenyl)quinolin-2-amine (ABX464) and 8-chloro-N-glucuronide-N-(4-(trifluoromethoxy)phenyl)quinolin-2-amine) (ABX464-N-glucuronide) compounds. b HIV-1 strain Ada-MR5 was used to infect triplicate of activated PBMCs from different donors (stimulated for two days with PHA and IL2) in the absence or presence of increasing concentrations of ABX464. Supernatant was harvested 6 days post-infection (pi) and viral capsid protein p24 antigen was quantitated using a standard ELISA protocol. Each point represents 5 donors. c Concurrently, cell viability was measured by MTS assay after 6 days of incubation and cytotoxicity was indicated as percentage as compared with untreated cells. d Histogram plots of CFSE fluorescence of CD4⁺ (upper panel) or CD8+ (lower panel) after 3 days culture without (blue and green curves, respectively) or with increasing concentrations of ABX464; 15 μM (red curves), 31 μM (yellow curves) or 61 μM (pink curves). CD4⁺ and CD8+ T cells purified from PBMCs were labeled with CFSE and cultured with the indicated concentration of ABX464 and then analysed by flow cytometry. The plots show the CFSE profiles of viable CFSE-labeled CD4⁺ and CD8+ T cells. e HIV-1 strain YU2 was used to infect triplicate of monocyte-derived macrophages from different donors in the absence or presence of increasing concentrations of ABX464. Supernatant was harvested 8 days pi and viral capsid protein p24 antigen was quantitated using standard ELISA protocol. Each point represents 9 donors.

Figure 2

ABX464 inhibits different HIV-1 clades, resistant viruses and did not select for resistance. a Different HIV-1 strains (clade B, clade C and recombinants clades) were used to infect PBMCs from three different donors in the absence or presence of 5 μM of ABX464. Supernatant was harvested 6 days pi and viral capsid protein p24 antigen was quantitated using standard ELISA protocol. b RT activity (cpm) measured in human PBMCs infected with different resistant mutants of NL4.3 strain (K103N, K65R and M184V) and treated with ABX464 or 3TC. c Resistance to ABX464 was tested on human PBMCs and compared to current therapies (see Methods). There were no resistance-inducing mutations detected after treatment with ABX464 for at least 24 weeks.

Figure 3

Effect of ABX464 on cellular and HIV-1 RNA splicing. a Heat map showing percent-spliced-in (psi) shifts for the 264 alternate splicing events in positive control fibroblasts (fibroblasts) and their iPSCs (stem) untreated, PBMCs untreated (Cells) or PBMCs treated by either DMSO without drug (DMSO) or with various compounds (ABX35, ABX36, ABX273, ABX388, ABX387 ABX402, ABX415, ABX449, ABX464 and Darunavir). b Scatter plots comparing splicing changes for the 264 exons in PBMCs treated with ABX464 versus DMSO (left panel) or stem cells differentiated into fibroblast (middle panel). The right panel shows a scatter plot comparing psi changes for the 264 exons induced by ABX464 to those induced during stem cell differentiation. Pearson correlations (R values) are shown. c Visualisation of HIV-1 splice junctions after capture and sequencing of HIV-1 RNAs extracted from PBMCs infected with the YU2 strain, either untreated (DMSO) or treated with ABX464 (ABX464) 3 days pi (d3) or 6 days pi (d6) using 454 pyrosequencing (according to GS junior method manual). Estimate of the distribution of read mappings, positions of known acceptor (A1, A2, A3, A4a, A4c and A7) and donor splice sites (D1a, D2G4, D3 and D4), the exon-exon junctions (the line of accolade which is dependent of the junction abundance) and the known coding regions of HIV-1, are shown.

Figure 4

ABX464, influences REV-mediated HIV RNA biogenesis and export, and interacts with the CBC complex. a Schematic representation of the HIV-1 reporter gene. b Visualization of GFP-MS2-Bound RNA in HeLa cells expressing Tat, MS2-GFP and HIV-reporter transcripts in non-transfected or transfected cells with Rev. Rev-transfected cells were untreated (DMSO) or treated with ABX464 (ABX464). Arrow indicates transcription site. c Quantification of images corresponding to untreated (DMSO) or treated (ABX464) in (b) of total GFP (upper panel), in the nucleoplasm (left panel) or at the start site (right panel). Box-plots show the average GFP intensity (foci numbers or starting transcription point numbers) of Rev-transfected HeLa cells for each condition. Whiskers correspond to the minimum and maximum, boxes, to the 25–75 percentiles and the band inside the box, to the median. Statistical analysis was performed on at least 15 nuclei of Rev positive cells using an unpaired /t/-test (**: p < 0.005 and ****: p < 0.0001). d Purified recombinant CBC20 and CBC80 proteins were incubated with increasing concentrations of ABX464 (left panel) or ABX464-N-glucuronide (right panel, Gluc) and treated for 30 minutes with UV light. The proteins were revealed by Western Blotting using CBC20 and CBC80 antibodies. e Unlike m7GpppG cap structure, neither ABX464 nor ABX464-N-glucuronide interferes with the binding of capped RNA to CBC complex. Recombinant human CBC was incubated with a capped RNA substrate and analysed by native gel electrophoresis in order to resolve the different RNA and RNA-protein complexes: free RNA (lane 1), and CBC-RNA complexes (lanes 2–10) in the presence of 12 mM of m7GppG (lanes 10) or 5 μM, 10 μM or 50 μM of ABX464 (lanes 2–4, respectively) or 5 μM, 10 μM, 50 μM or 100 μM of ABX464-N-glucuronide (lanes 5–9, respectively).

Figure 5

Efficacy of ABX464 to inhibit viral replication in humanized mice. a Reconstituted SCID mice were infected with JRCSF HIV-1 strain by intraperitoneal injection. The control group received labrafil and 5% DMSO by gavage (n = 15) and treated group 20 mg/kg b.i.d of ABX464 in labrafil and 5% DMSO (n = 14) for 15 days. Two independent experiments were performed with 5 and 10 reconstituted mice for each group. Viral load was assessed by measuring viral RNA using the Amplicor HIV-1 Monitor from Roche (limits of detection are delimited by interrupted lines). b FACS analysis was performed on cells recovered by peritoneal wash at day 15 post-treatment to assess the CD8/CD4 ratio. c Engrafted NSG humanized mice were treated by oral gavage with ABX464 at either 20 mg or 40 mg/kg once a day for 30 days and indicated lymphocyte populations were monitored by FACS analysis. d NSG humanized mice were infected with the YU2 HIV-1 virus and treated either by oral gavage with ABX464 at 40 mg/kg once a day for 30 days or by HAART (3TC-Tenofovir-Raltegravir) (grey boxes of left and middle panels). Right panel represents viral loads after treatment cessation. For HAART, food pellets were made by mixing 2.5 g of 3TC and TDF each, and 5 g of RTV with 5 kg of ground protein-rich, vitamin-fortified food (Nafag 3432, Provimi Kliba AG, Switzerland) which was subsequently formed to food pellets and sterilized by gamma-irradiation with 25 kGy. Viral load was assessed by measuring viral RNA using the Amplicor HIV-1 Monitor from Roche (limits of detection are delimited by interrupted lines).