Optimization of a lentivirus-mediated gene therapy targeting HIV-1 RNA to eliminate HIV-1-infected cells

Abstract

Persistence of HIV-1 in cellular reservoirs results in lifelong infection, with cure achieved only in rare cases through ablation of marrow-derived cells. We report on optimization of an approach that could potentially be aimed at eliminating these reservoirs, hijacking the HIV-1 alternative splicing process to functionalize the herpes simplex virus thymidine kinase (HSVtk)/ganciclovir (GCV) cell suicide system through targeted RNA trans-splicing at the HIV-1 D4 donor site. AUG1-deficient HSVtk therapeutic pre-mRNA was designed to gain an in-frame start codon from HIV-1 tat1. D4-targeting lentiviral vectors were produced and used to transduce HIV-1-expressing cells, where trans-spliced HIV-1 tat/HSVtk mRNA was successfully detected. However, translation of catalytically active HSVtk polypeptides from internal AUGs in HSVtk _ΔAUG1 caused GCV-mediated cytotoxicity in uninfected cells. Modifying these sites in the D4 opt 2 lentiviral vector effectively mitigated this major off-target effect. Promoter choice was optimized for increased transgene expression. Affinity for HIV-1 RNA predicted in silico correlated with the propensity of opt 2 payloads to induce HIV-1 RNA trans-splicing and killing of HIV-1-expressing cells with no significant effect on uninfected cells. Following latency reversing agent (LRA) optimization and treatment, 45% of lymphocytes in an HIV-1-infected latency model could be eliminated with D4 opt 2/GCV. Further development would be warranted to exploit this approach.

Keywords: HIV-1; HSVtk/GCV; MT: Oligonucleotides: Therapies and Applications; RNA trans-splicing; gene therapy; herpes simplex virus; latency reversing agents; latent reservoir; lentiviral vector; shock and kill.

Graphical abstract

Figure 1

A lentivirus-mediated gene therapy targeting HIV-1 RNA to eliminate HIV-1-expressing cells (A) Schematic for delivery and activation of HIV-1-dependent HSVtk/GCV CSS. Following LVV transduction, the HIV-1-targeting HSVtk_ΔAUG1 payload is expressed at the RNA level and subsequently functionalized in HIV-1-expressing cells through an RNA trans-splicing reaction with the HIV-1 tat 1 exon. Near-full-length HSVtk is translated from the chimeric HIV-1 tat/HSVtk mRNA and phosphorylates the prodrug GCV, initiating a phosphorylation cascade by cellular kinases that culminates in GCV-TP, a cytotoxic metabolite that induces cell death through DNA damage. (B) Schematic by which HIV-1 pre-mRNA is targeted for therapeutic trans-splicing. The therapeutic pre-mRNA localizes to HIV-1 pre-mRNA by means of a complementary binding domain. This positions the therapeutic splice acceptor (3′ss) for RNA trans-splicing with the HIV-1 D4 splice donor, which is proximal to the binding domain target sequence in HIV-1 vpu. The approximate locations of HIV-1 open reading frames and splice sites are annotated above and within the pre-mRNA, respectively. The HIV-1_NL4-3:binding domain RNA duplex is adapted from Figure 3A, with the spacer region requisite for RNA secondary structure modeling digitally removed. ss, splice site; D, donor (5′ss); A, acceptor (3′ss). (A and B) Diagrams are not drawn to scale.

Figure 2

Optimization of HSVtk translational initiation from HIV-1 RNA-targeting LVVs is necessary to restrict HSVtk/GCV-mediated killing to HIV-1-expressing cells (A) Maps of HIV-1 RNA-targeting trans-splicing cassettes D4, D4 opt 1, and D4 opt 2 illustrating modifications made to HSVtk and BbvCI to alter downstream translational initiation sites and an additional predicted splice acceptor site, respectively. The full-length HSVtk-positive control cassette is shown for comparison. BD, binding domain. 3’ss, splice acceptor site. (B–E) Sanger sequencing confirmation of trans-spliced HIV-1/HSVtk RNA in HIV-1-expressing Jurkat T cells following delivery of (B) D4, (C) D4 opt 1, or (D) D4 opt 2 therapeutic LVV. Chromatogram snapshots depict the HIV-1/HSVtk splice junction. Refer to Figure S4 for experimental details. In brief, trans-spliced products were amplified from Jurkat RNA by RT-PCR and cloned into TOPO plasmids (where they could be inserted in either orientation) for sequencing. M13 primers were used to read from the TOPO backbone into the insert. (E) Representative BLAST alignment between the predicted splice junction and the amplified trans-spliced product from HIV-1_NL4-3ΔE-expressing Jurkat cells transduced with D4 opt 2. (F) Anti-HSVtk western blot of polypeptide expression (P1, 43 kDa; P2, 40.4 kDa; P3, 39.8 kDa; P4, 37.0 kDa²³) from therapeutic constructs D4, D4 opt 1, and D4 opt 2 in the absence of HIV-1, with anti-vinculin (124 kDa) loading control. Full-length HSVtk construct (HSVtk) and transfer plasmid backbone (pSico) used as positive and negative controls for HSVtk expression, respectively. MW, molecular weight. (G and H) Viability screens in HIV-1-expressing and uninfected cells. (G) 5 × 10³ Jurkat T cells/well or (H) 2 × 10⁴ HEK293T cells/well were seeded on day 1, (G) transduced with HIV-1_NL4-3ΔE (MOI = 6) or (H) transfected with 100 ng HIV-1_NL4-3ΔE plasmid (pNL4-3ΔE) on day 2, transduced with (G) CkRhsp (CR)-driven LVV panel (MOI = 14) or (H) EF1α-driven LVV panel (MOI = 2) on day 3, treated with 50 μM GCV doses on days 4 and 5, and subjected to MTT cell viability assay on day 8. Mock treatments performed with media only. Data presented as mean with SD (N = 3 wells/condition). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; one-way ANOVA with Tukey’s multiple comparisons test.

Figure 3

In silico RNA secondary structure modeling of the HIV-1 RNA-targeting potential of RNA binding domains (A–C) Centroid secondary structure predictions. Minimum free energy (MFE): (A) −84.50 kcal/mol, (B) −11.20 kcal/mol, (C) 0.00 kcal/mol (D–F) MFE secondary structure predictions. MFE: (D) −84.50 kcal/mol, (E) −16.00 kcal/mol, (F) −6.28 kcal/mol. Duplexes formed between HIV-1_NL4-3 and either the (A and D) HIV-1 D4-targeting or (B and E) Scramble Version 1 (V1) binding domains. In comparison, HIV-1_NL4-3 and (C and F) the Scramble Version 2 (V2) binding domain were predicted to make no or reduced contact.

Figure 4

In silico HIV-1 RNA-targeting potential correlates with the levels of HIV-1 D4 trans-splicing and GCV-mediated killing achieved by opt 2 LVV in HIV-1-expressing cells (A–G) HEK293T cells were seeded in duplicate wells at 2 × 10⁵/well on day 1, transfected with 200 ng full-length HIV-1_NL4-3 plasmid (pNL4-3) on day 2, transduced with EF1α-driven LVV panel (MOI = 2) on day 3, subjected to a media change on day 4, and lysed for DNA or RNA extraction on day 5. Mock treatments performed with media. (A) Viral DNA copies in extracted cellular DNA following delivery of opt 2/positive control LVV panel and HIV-1 pNL4-3 to HEK293T, assessed by WPRE and tat qPCRs, respectively, normalized to ALB qPCR. Values at the dashed line were considered background amplification but did not meet the criteria to be excluded from the analysis. (B) Levels of opt 2/positive control RNA payload and HIV-1_NL4-3 RNA target (per microgram of total cellular RNA) in HEK293T cells, assessed by RT-qPCR for HSVtk and tat, respectively, normalized to β-actin. (A and B) Data presented as mean with SD (N = 2 qPCR replicates). (C) RT-PCR detection of putative chimeric HIV-1/HSVtk RNA sequences (291 bp) in HIV-1-expressing HEK293T cells following delivery of D4 opt 2 or Scramble V1 (SV1) opt 2 LVV, with β-actin RT-PCR (202 bp) for normalization. PCR products were gel extracted for sequencing. (D) PCR primer design for amplification of the splice junction of chimeric HIV-1/HSVtk transcripts (long amplicon), with the forward primer positioned in HIV-1 tat exon 1 and the reverse positioned in HSVtk. Diagram not to scale. (E) Densitometric quantification of (C). (F and G) Sanger sequencing confirmation of trans-spliced HIV-1/HSVtk RNA in HIV-1-expressing HEK293T cells following delivery of (F) Scramble V1 opt 2 or (G) D4 opt 2 LVV. Sequencing performed with trans-splice PCR primers. Chromatogram snapshots depict the HIV-1/HSVtk splice junction. (H) Viability screen in HIV-1-expressing and uninfected cells. 2 × 10⁴ HEK293T cells/well were seeded on day 1, transfected with 100 ng HIV-1 pNL4-3 on day 2, transduced with EF1α-driven LVV panel (MOI = 3) on day 3, treated with 50 μM GCV doses on days 4 and 5, and subjected to MTT cell viability assay on day 8. Mock treatments performed with media only. Data presented as mean with SD (N = 3 independent experiments, each performed in triplicate). ∗p < 0.05, ∗∗∗p < 0.001; one-way ANOVA with Tukey’s multiple comparisons test.

Figure 5

Dose-response of chronically HIV-1-infected J-Lat 10.6 cells to LRA treatment Dose-response curves for HIV-1 reactivation (green line), based on the percentage of live J-Lat 10.6 cells determined by flow cytometry to express EGFP from the HIV-R7/E^–/GFP 5′ long terminal repeat (LTR). Dose-response curves for cell viability (red line) were based on the MTT assay. J-Lat 10.6 cells were treated with (A) the tumor necrosis factor receptor (TNFR) agonist TNFα, (B) the protein kinase C (PKC) agonist phorbol 12-myristate 13-acetate (PMA), (C) the T cell receptor (TCR) agonist phytohemagglutinin (PHA), (D) the 26S proteasome inhibitor bortezomib (BTZ), (E) the DNA methyltransferase (DNMT) inhibitor decitabine (DAC), (F) the histone deacetylase (HDAC) inhibitor romidepsin (RMD), (G) the HDAC inhibitor valproic acid (VPA), or (H) the bromodomain and extraterminal domain protein (BET) inhibitor JQ1 at the specified doses and were assayed for HIV-1 reactivation and cell viability levels subsequent to (A, C, E, G, and H) five or (B, D, and F) six additional days of incubation. Where duplicate wells were assayed, data presented as mean with SD. Across the LRA panel, relative EC₅₀ values for HIV-1 reactivation were estimated to be (A) 0.40 ng/mL, (B) 607 pM, (C) 8.30 μg/mL, (D) 11.7 nM, (E) 60.2 nM, (G) 2.15 mM, and (H) 163 nM. The relative EC₅₀ value for (F) could not be reliably determined due to low HIV-1 reactivation coupled with sharp increase in toxicity over RMD dosing range. The baseline level of GFP+ across the LRA panel ranged from 0.36% to 2.21% of unstimulated J-Lat 10.6 cells.

Figure 6

Chronically HIV-1-infected J-Lat 10.6 cells transduced with Trans-splicing Opt 2 LVVs can produce chimeric HIV-1/HSVtk mRNA when LRA treatment is used to stimulate HIV-1 expression 1 × 10⁵ J-Lat 10.6 cells were seeded in duplicate wells on day 1, stimulated with the LRA TNFα (1.56 ng/mL) on day 2, transduced with EF1α-directed LVVs at an MOI of 14 on day 3, and lysed for DNA or RNA extraction on day 5. Mock treatments performed with media. (A) Evaluation of opt 2/positive control LVV delivery and payload expression in J-Lat 10.6 cells ± TNFα stimulation. (Left axis) LVV delivery (based on VCN) assessed by WPRE qPCR on cellular DNA, normalized to ALB. (Right axis) LVV payload expression (HSVtk RNA copies per microgram of total cellular RNA) assessed by RT-qPCR on cellular RNA, normalized to β-actin. ND, below limit of detection. (B) Evaluation of HIV-1 transcription in J-Lat 10.6 cells ± TNFα stimulation. HIV-1 target expression (tat RNA copies per microgram of total cellular RNA) assessed by RT-qPCR on cellular RNA, normalized to β-actin. (A and B) Data presented as mean with SD (N = 2 or 3 qPCR replicates/condition). (C-F) Evaluation of HIV-1 trans-splicing in J-Lat 10.6 cells transduced with LVV panel ± TNFα stimulation. (C) (Top) Chimeric HIV-1/HSVtk splice junctions amplified by RT-PCR on cellular RNA; 291-bp amplicon expected (see Figure 4D for primer design). PCR products were gel extracted for sequencing; refer to Figure S11 for analysis of product from TNFα-stimulated cells transduced with HSVtk LVV. (Bottom) β-actin RT-PCR for normalization; 202-bp amplicon expected. (D and E) Chromatograms of HIV-1/HSVtk splice junctions from TNFα-stimulated J-Lat 10.6 cells transduced with (D) Scramble V2 opt 2 and (E) D4 opt 2 trans-splicing LVVs. Sequencing performed with trans-splice PCR primers. (F) Densitometric quantification of select lanes in (C). SV2, Scramble V2.

Figure 7

LRA treatment can influence the susceptibility of chronically HIV-1-infected J-Lat 10.6 cells to killing by an HIV-1 RNA-targeted CSS, D4 opt 2 LVV and GCV (A–C) Effect of TNFα on J-Lat 10.6 cells in isolation and in combination with HIV-1 RNA-targeted CSS for shock and kill. (A and B) Viability screen based on population-level metabolism of MTT. 5 × 10³ J-Lat 10.6 cells/well were seeded on day 1, treated with 1.56 ng/mL TNFα (or mock; media) on (A) day 2 or (B) day 3, transduced with EF1α-driven LVV panel (MOI = 14 or mock; media) on (A) day 3 or (B) day 2, treated with 50 μM GCV doses (or mock; media) on days 4 and 5, and assayed on day 8. (A) N = 3 wells/condition. SV2, Scramble V2. (B) N = 7 independent experiments; two or three wells/condition in each experiment. (C) Percentage of J-Lat 10.6 cell population determined to be live (DRAQ7−) by flow cytometry. N = 2 or 3 independent experiments, including data from experiment described in Figure 5A. Refer to Figure 5A for experimental details. (D–F) Effect of DAC on J-Lat 10.6 cells in isolation and in combination with HIV-1 RNA-targeted CSS for shock and kill. (D and E) Viability screen based on population-level metabolism of MTT. 5 × 10³ J-Lat 10.6 cells/well were seeded on day 1, treated with (D) 12.5 nM or (E) 25 nM DAC (or mock; media) on day 2, transduced with EF1α-driven LVV panel (MOI = 14 or mock; media) on day 3, treated with 50 μM GCV doses (or mock; media) on days 4 and 5, and assayed on day 8. (D) N = 3 wells/condition with exception of HSVtk TDNs (one well/condition). (E) N = 4–6 wells/condition across two independent experiments. (F) Flow plots depicting percentage of J-Lat 10.6 cell population determined to be live (DRAQ7−) by flow cytometry. Data from experiment described in Figure 5E; refer to Figure 5E for experimental details. (A–E) Data presented as mean with SD. (A, C, D, and E) One-way ANOVA with Tukey’s multiple comparisons test. (B) Mann-Whitney test. ns, not significant; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.