Potent dual block to HIV-1 infection using lentiviral vectors expressing fusion inhibitor peptide mC46- and Vif-resistant APOBEC3G

Abstract

Gene therapy strategies that effectively inhibit HIV-1 replication are needed to reduce the requirement for lifelong antiviral therapy and potentially achieve a functional cure. We previously designed self-activating lentiviral vectors that efficiently delivered and expressed a Vif-resistant mutant of APOBEC3G (A3G-D128K) to T cells, which potently inhibited HIV-1 replication and spread with no detectable virus. Here, we developed vectors that express A3G-D128K, membrane-associated fusion inhibitor peptide mC46, and O⁶-methylguanine-DNA-methyltransferase (MGMT) selectable marker for in vivo selection of transduced CD34⁺ hematopoietic stem and progenitor cells. MGMT-selected T cell lines MT4, CEM, and PM1 expressing A3G-D128K (with or without mC46) potently inhibited NL4-3 infection up to 45 days post infection with no detectable viral replication. Expression of mC46 was sufficient to block infection >80% in a single-cycle assay. Importantly, expression of mC46 provided a selective advantage to the A3G-D128K-modified T cells in the presence of replication competent virus. This combinational approach to first block HIV-1 entry with mC46, and then block any breakthrough infection with A3G-D128K, could provide an effective gene therapy treatment and a potential functional cure for HIV-1 infection.

Keywords: APOBEC3G; C46; D128K; HIV; MGMT; MT: Delivery Strategies; Vif; gene therapy; hypermutation; lentivirus.

Graphical abstract

Figure 1

Design and titers of A3G-D128K direct repeat vector (A) Lentiviral vector pGFP expresses enhanced green fluorescent protein (eGFP), while pGFP.MGMT∗ expresses eGFP fused in-frame to O⁶-methylguanine-DNA-methyltransferase (MGMT) containing the P140K mutation (MGMT∗). (B) Lentiviral vector pA3x3G(DK).GFP.MGMT∗ contains an overlapping ∼900-bp homologous region of A3G-D128K (“3”, indicated by black arrows). A3 and 3G are the 5′ and 3′ fragments of A3G, respectively, containing the D128K mutation that confers Vif resistance. The direct repeats are separated by three stop codons (“stop”); the 3G fragment is fused in-frame with the self-cleaving porcine teschovirus-1 2A peptide (P2A), eGFP, and MGMT∗. All vectors contain intact long terminal repeats (LTRs), are expressed from the internal human elongation factor 1 alpha promoter (EF1a), and contain the woodchuck hepatitis virus post-transcriptional regulatory element (WPRE). Direct repeat deletion that occurs during reverse transcription to reconstitute A3G-D128K and the protein products A3G-D128K and GFP-MGMT∗ resulting after translation and P2A cleavage are shown. Ψ, HIV-1 RNA packaging signal. (C) Normalized vector virus (0.08–80 ng of p24 CA input) from GFP, GFP.MGMT∗, and A3x3G(DK).GFP.MGMT∗ produced in the presence of HIV-2 Vif was used to transduce 293T cells. The proportion of GFP⁺ cells was determined by flow cytometry analysis 48 h post-transduction. Average of ≥3 independent experiments. (D) Average titer represented in infectious units (IU)/mL (average of ≥3 independent experiments). IU/mL was determined by using the formula: (number of 293T cells at transduction) × ([%GFP⁺ cells]/[mL of virus stock used for transduction]). Error bars show mean ± standard deviation (SD).

Figure 2

MGMT selection and characterization of T cells expressing A3G-D128K (A) MGMT selection of T cell lines MT4, CEM, or PM1 transduced with GFP, GFP.MGMT∗, or A3x3G(DK).GFP.MGMT∗ vector virus. The proportion of GFP⁺ cells was measured every 2 days after initiating selection. As expected, cells transduced with GFP did not survive MGMT selection, while the proportion of cells transduced with GFP.MGMT∗ (GM) or A3x3G(DK).GFP.MGMT∗ (D128K.GM) reached 100% 6 days after initiating selection. Average %GFP⁺ cells from 3 to 5 independent experiments are shown. Error bars show mean ± standard deviation (SD). (B) After MGMT selection, the %GFP⁺ cells in the MT4/D128K.GM and CEM/D128K.GM cell pools remained stable for least 45 days. Average of 3 and 5 independent experiments for MT4 and CEM, respectively. Error bars show mean ± SD. (C) Western blot analysis of MGMT selected MT4 and CEM cell pools 1 to 3 expressing GFP.MGMT∗ (∼50 kDa) (MT4/GM and CEM/GM, respectively) probed with antibodies against GFP and housekeeping gene Hsp90 (90 kDa). GFP.MGMT size control shown from 293T cells transfected with pGFP.MGMT∗. (D) Western analysis post MGMT selection of MT4 and CEM cell pools transduced with A3x3G(DK).GFP.MGMT∗ (MT4/D128K.GM and CEM/D128K.GM, respectively). Probing with antibody against P2A (top blot) shows that each cell line expressed the correct deleted A3G(DK)P2A protein (∼49 kDa). Probing with antibody against A3G (middle blot) shows expression of both the A3G(DK)P2A (∼49 kDa) and the endogenous A3G (∼46 kDa) in CEM/D128K.GM cells. As expected, MT4 cells express little to no endogenous A3G, and therefore lack the endogenous A3G band. Probing with antibody against GFP (bottom blot) shows that all cell lines express the expected P2A cleaved GFP.MGMT∗ protein (∼50 kDa). Also shown is A3G(DK)P2A protein expressed in 293T cells transfected with A3G(DK)-P2A-eYFP. Housekeeping proteins α-tubulin or Hsp90 were detected and used as loading controls. (E) MGMT selected MT4 and CEM cell pools transduced with A3x3G(DK).GFP.MGMT∗ vector virus (MT4/D128K.GM and CEM/D128K.GM, respectively) were analyzed to determine the deletion frequency of the A3G-D128K direct repeat. Primers flanking the direct repeat were used to PCR amplify gDNA from the cell pools as well as the plasmid controls: A3G(DK) (1,159 bp) and undeleted A3x3G(DK) (2,077 bp). Band density analysis shows near 100% direct repeat deletion frequency for 3 independent pools for MT4 and 5 independent pools for CEM. Low levels of deletion were detected with the plasmid A3x3G(DK) control due to direct repeat deletion that can occur during PCR (avg ∼5%).

Figure 3

HIV-1 replication kinetics in MT4, CEM, or PM1 cell pools expressing A3G-D128K NL4-3 (∼70 pg p24 CA) was used to infect MT4, CEM, or PM1 T cell lines (1 × 10⁶ cells) expressing GFP.MGMT∗ (MT4/GM, CEM/GM, and PM1/GM, respectively) or A3x3G.GFP.MGMT∗ (MT4/D128K.GM, CEM/D128K.GM, and PM1/D128K.GM, respectively) vector virus. Supernatants were harvested every 2–3 days post infection up to 45 days, and p24 CA amounts were determined by ELISA. No detectable virus was observed 45 days post infection in MT4/D128K.GM and PM1/D128K.GM (3 independent pools) or CEM/D128K.GM (5 independent pools). Parental MT4, CEM, and PM1 cells, as well as MT4/GM, CEM/GM, and PM1/GM cell pools (3 independent pools each) supported replication and spread of NL4-3 with peak virus on days 5, 9, and 9, respectively. The average p24 amounts in culture supernatants from duplicate flasks for each cell pool are shown.

Figure 4

Transduction and MGMT selection of A3G-D128K expressing CD34⁺ cells (A) Mobilized peripheral blood CD34⁺ cells were transduced with vector virus from GFP.MGMT∗ and A3x3G.GFP.MGMT∗ (MOI = 15) and flow cytometry analysis was performed on day 7 to determine the %GFP marked cells. A3x3G.GFP.MGMT∗ vector virus was also transduced in the presence of transduction enhancers α-viniferin and PGE-2. (B) CD34⁺ transduced cells were placed on MGMT selection (at days 7 and 10 post transduction; black triangles) and flow cytometry analysis was performed 4 days later (day 14) to determine the %GFP. Error bars show mean ± SD.

Figure 5

MGMT selection and single-cycle HIV-1 infectivity of T cell lines expressing mC46 (A) Vectors pGFP.MGMT∗.mC46 and pA3x3G.GFP.MGMT∗.mC46 are shown; the vectors express entry inhibitor peptide mC46 using a T2A self-cleaving peptide downstream of GFP.MGMT∗. All other abbreviations are as described in Figure 1. (B) Generation of MT4, CEM, and PM1 cell pools expressing mC46 or A3G-D128K + mC46. MGMT selection of T cell lines MT4, CEM, or PM1 transduced with GFP, GFP.MGMT∗.mC46, or A3x3G(DK).GFP.MGMT∗.mC46 vector virus. The proportion of GFP⁺ cells was measured every 2 days post selection. Cells transduced with GFP did not survive MGMT selection, while the proportion of GFP⁺ cells for MT4/GM.mC46, CEM/GM.mC46, PM1/GM.mC46, and MT4/D128K.GM.mC46, CEM/D128K.GM.mC46, and PM1/D128K.GM.mC46 reached 100% 6–10 days post selection. Three independent cell pools were generated per T cell line (numbered pools 1–3). (C–F) HIV-1 infectivity in T cell pools expressing mC46 in a single-cycle assay. MT4, CEM, and PM1 cell pools expressing GFP.MGMT∗.mC46 alone (GM.mC46) or with A3G-D128K (D128K.GM.mC46) were transduced with an HIV-1 luciferase vector pseudotyped with either NL4-3 envelope (C–E) or Ad8 envelope (F). Shown is the average percentage of luciferase activity (relative light units) compared with the parental T cell lines (set to 100%). n = 3 per cell pool. Error bars show mean ± SD.

Figure 6

HIV-1 replication kinetics in mC46- and mC46 + A3G-D128K-expressing T cell pools (A) MT4, (B) CEM, and (C) PM1 cell pools expressing GFP.MGMT∗.mC46 with or without A3G-D128K were transduced with NL4-3 (∼70 pg p24 CA per 1 × 10⁶ cells). Supernatants were harvested every ∼3 days and assayed for p24 CA amounts by ELISA. Virus replication peaked in parental MT4, CEM, and PM1 cell lines on days 6, 9, and 9, respectively. T cell pools expressing GFP.MGMT∗ + mC46 (GM.mC46) either had no evidence of NL4-3 replication or delayed replication kinetics. For MT4/GM.mC46, 6/6 flasks had no or little (∼29-day delay) evidence of viral replication. For CEM/GM.mC46, 3 of 6 flasks and for PM1/GM.mC46 1 of 6 flasks showed virus replication which peaked on days 21–24 in CEM cells and on day 36 in PM1 cells. No detectable virus replication was observed for 45 days post infection in cell pools expressing A3G-D128K + mC46 (D128K.GM.mC46 cell pools). The average p24 amounts in culture supernatants from duplicate flasks for each cell pool (flasks a and b) are shown.

Figure 7

Preferential survival of CEM cells expressing mC46 after infection with replication competent HIV-1 (A) Mixtures of 95% CEM parental cells and 5% CEM cells expressing either GFP.MGMT∗ (CEM/GM) or GFP.MGMT∗.mC46 (CEM/GM.mC46) were infected with NL4-3 (1 ng of p24 CA per 1 × 10⁶ total cells). Viral replication was monitored by measuring p24 CA in culture supernatants (dashed lines) and the %GFP⁺ cells were monitored by flow cytometry analysis (solid lines). (B) Mixtures of 95% CEM parental cells and 5% CEM cells expressing either D128K.GFP.MGMT∗ (CEM/D128K.GM) or D128K.GFP-MGMT∗.mC46 (CEM/D128K.GM.mC46) were infected with NL4-3. Viral replication was monitored by measuring p24 CA in culture supernatants (dashed lines) and the %GFP cells were monitored by flow cytometry analysis (solid lines). Cell pools expressing mC46 (CEM/GM.mC46 and CEM/D128K.GM.mC46) exhibited a selective advantage over CEM/GM and CEM/D128K.GM cell pools lacking mC46 expression, reaching near 100% and ∼50%–80% GFP⁺ cells in the cell cultures after infection with NL4-3. (C) Western analysis of mC46 expression levels in CEM/GM.mC46 and CEM/D128K.GM.mC46 cell pools. Cytosolic (C) and membrane (M) fractions of cell lysates from CEM cell pools expressing mC46 (GM.mC46) or A3G-D128K + mC46 (D128K.GM.mC46) were probed for α-tubulin (cytosolic protein) and mC46 (membrane protein). On average, the CEM/D128K.GM.mC46 cell pools expressed ∼5-fold lower amounts of mC46 than the CEM/GM.mC46 cell pools.

Figure 8

Transduction and MGMT selection of CD34⁺ cells expressing A3G-D128K and mC46 (A) Mobilized peripheral blood CD34⁺ cells were transduced with vector virus from GFP.MGMT∗ and A3x3G.GFP.MGMT∗.mC46 (MOI 5) and flow cytometry analysis was performed on day 7 to determine the %GFP marked cells. A3x3G.GFP.MGMT∗mC46 vector virus was also transduced in the presence of transduction enhancer PGE-2. (B) CD34⁺-transduced cells (CD34⁺/GM- and PGE-2-treated CD34⁺/D128K.GM.mC46) were placed on MGMT selection (at days 7, 10, and 14 post transduction; black triangles) and flow cytometry analysis was performed on day 18 to determine the %GFP.