Antiretroviral activity of the aminothiol WR1065 against Human Immunodeficiency virus (HIV-1) in vitro and Simian Immunodeficiency virus (SIV) ex vivo

Fresh human peripheral blood mononuclear cells (PBMC, from the NIH Transfusion Center) were cultured in 250 ml flasks (2 × 10⁶ cells/ml) for 48 hr in RPMI-1640 media (ATCC, Manassas, VA) containing 10% fetal bovine serum (Hyclone, Logan, UT), 1% penicillin/streptomycin/glutamine (Invitrogen, Gaithersburg, MD), 10 U/ml interleukin 2 (IL2, BD Biosciences, San Jose, CA) and 20 μg/ml phytohemagglutinin (PHA, Sigma, St. Louis, MO). After 48 hr, the cells were washed to remove PHA and the resulting PHA-stimulated T-cell blasts (human TCBs) were transferred to 96 well microtiter plates (0.5 × 10⁶ cells/well), infected with HIV-1_BZ-167(gift from S. Sharpe, New York University, New York, NY) at 170-200 50% tissue culture infectious dose/10⁵ target cells for 2 hr, and subsequently incubated with 2.5-103.0 μM WR1065 (Chemical Carcinogen Reference Standard Repository, Kansas City, MO) and/or 0.002-0.117 μM AZT (Sigma-Aldrich Inc., St. Louis, MO) for 72 hr. Cells were then harvested and evaluated for HIV-1 replication by RETRO-TEK HIV-1 p24 Extended Range Elisa Kit (ZeptoMetrix, Buffalo, NY) or by HIV-1 p24 Antigen Capture Assay Kit (Biological Products Laboratory, FCRDC, Frederick, MD).

To compare the metabolite WR1065 with the parent compound amifostine, in one experiment 50.0 μM amifostine (Chemical Carcinogen Reference Standard Repository) was added. Due to the lack of alkaline phosphatase in cultured human cells, we pre-incubated the amifostine with alkaline phosphatase (Sigma-Aldrich Inc.), at 1 U per 100 μl of media containing 50 μM amifostine, to generate WR1065. In experiments designed to examine virus replication with the combination of AZT and WR1065, the standard curve for AZT included concentrations between 0 and 23.0 ηM and WR1065 was used at either 18.7 or 26.0 μM.

Drug-induced cell viability at 72 hr was determined by Trypan blue exclusion [20, 21] in human TCBs grown in a second 96-well microtiter plate, where cells were exposed to drugs in the absence of HIV-1 inoculation. Cells from triplicate wells were mixed with Trypan blue and counted twice by hemocytometer. Numbers of viable (unstained) cells were expressed as a percentage of total (stained plus unstained) cells.

To examine apoptosis as a measure of cell viability in human TCBs infected with HIV-1 and treated with drug, we assayed for Annexin V (as previously described [22]). Cells taken from the wells used for p24 protein analysis were subjected to flow cytometry for this analysis and sorted on the basis of Annexin V positivity (apoptotic) and negativity (non-apoptotic).

Flow cytometry was used to evaluate the integrity of cell cycle parameters in human TCBs exposed to 0, 9.5 and 18.7 μM WR1065 according to the protocol described above. Harvested cells were pelleted and washed with culture media without serum before they were fixed overnight in 1 ml of ice-cold 70% ethanol, pelleted by centrifugation and incubated with Ribonuclease A (Sigma-Aldrich Inc.) at room temperature for 20 min. Propidium iodide (20-50 μg/ml) (Molecular Probes, Eugene, OR) was added to each cell suspension and cells were kept in the dark at 4°C overnight. Cells were passed through a fluorescence activated flow cytometer (FACSCalibur, BD Biosciences, San Jose, CA) using the doublet discrimination module, and data were acquired using CellQuest (BD Biosciences) software. Cell cycle analysis was performed using ModFit software (Venty Software, Topsham, ME). Percentages of cells in G₀-G₁, S and G₂-M phases were calculated directly by the software.

Blood used to prepare macaque PBMC was collected from Macaca mulatta monkeys (macaques) numbered M612, M642 and M674. The macaques, housed at Advanced BioScience Laboratories (ABL), Inc. (Rockville, MD), had been infected with SIV_Mac251 for 14 months before these experiments were performed. The animals were maintained and treated under conditions approved by the Association for Assessment and Accreditation of Laboratory Animal Care, and all procedures were performed in accordance with humane principles for laboratory animal care. Protocols were reviewed and approved by the Institutional Animal Care and Use Committee of ABL, Inc.

Macaque PBMC (10⁶ cells/ml), prepared from blood using Ficoll gradient centrifugation, were depleted of CD8⁺ cells by magnetic bead separation using the CD8 Microbead Kit for non-human primates (Miltenyi, Auburn, CA). Briefly, whole PBMC were incubated with microbeads conjugated to an anti-CD8⁺ antibody and then washed. Cells were resuspended in Dulbecco's phosphate buffered saline (DPBS, Invitrogen, Carlsbad, CA) supplemented with 5% bovine serum albumin (BSA) and 2 mM EDTA, and run through a magnetic column. The flow-through material contained PBMC depleted (>99%) of CD8⁺ T-cells, which were then counted and cultured using the same media as for the human TCBs (above). Once in culture, PBMC were incubated for 48 hr in the presence of PHA to activate remaining T-cells, as described above for human TCBs. These cells, macaque CD8⁺ T cell-depleted, PHA-stimulated macaque T-cell blasts (TCBs) were transferred to 48-well plates (500 μL media/well, 0.5 × 10⁶ cells/well, 6 wells/macaque) and cultured for an additional 17 days in the presence of 0, 9.5 or 18.7 μM WR1065. The medium was changed twice weekly for a total of 4 times, and fresh WR1065 was added at each medium change. Cell survival was evaluated on days 10 and 17 using the Cell Titer 96^® Aqueous Non-Radioactive Cell Proliferation (MTS) Assay (Promega Corp., Madison, WI). SIV levels were assayed using the p27 Antigen Assay kit (Beckman Coulter, Fullerton, CA) on days 3, 7, 10, 14 and 17.

In HIV-1-infected human TCBs, the HIV-1 titers, determined in the absence of drug, ranged from 1,312 to 38,000 pg p24/ml (10,205 ± 2,367, mean ± SE, n = 19 experiments). The inter-experimental variability, likely a reflection of the variability of HIV-1 growth in cells from different individuals, was such that we chose to present "% Inhibition" in the graphs and tables to take advantage of the power of multiple experiments. We assayed for WR1065-induced inhibition of HIV-1 replication at three points on the dose-response curve in several replicate experiments. The HIV-1 inhibition data are shown in Table 1, where 26 and 52 μM WR1065 gave 65% and 89% inhibition of HIV-1, respectively. Parallel cell survival studies were performed using either Trypan blue exclusion in cells with drug but no virus, or Annexin V, an early marker of apoptosis, in the HIV-1-infected cells containing drug (Table 1). Because the Trypan blue assay showed extensive cell death at 52 and 103 μM WR1065, we chose to perform subsequent experiments at ≤ 26 μM WR1065. For the Annexin V assay, drug-exposed cultures ranged from 75% to 100% Annexin V-negative (non-apoptotic), with the majority of experiments showing 85-95% of the cells as Annexin-V negative (data not shown).

Table 1 also presents mean values for replicate experiments in which we exposed HIV-1-infected human TCBs to amifostine to compare the anti-HIV-1 activity of this compound with its active metabolite WR1065. Because cultured human cells lack alkaline phosphatase, we pre-incubated 50 μM amifostine with this enzyme for 30 minutes before adding the mixture to HIV-1-infected human TCBs to evaluate viral replication. The extent of HIV-1 inhibition and the fraction of cells surviving were similar to those observed in cells cultured with 52 μM WR1065 (Table 1), indicating that most of the amifostine had been converted to WR1065 and was available to inhibit virus replication.

Complete dose-response curves for % inhibition of HIV-1 replication with WR1065, and TCB % survival determined by Trypan blue are plotted in Figure 1A (mean ± SE, n = 4 experiments). The concentration of WR1065 giving 50% inhibition of virus replication was 13 μM, and at this dose the TCBs were 80% viable by the Trypan blue. Also by Trypan blue, 50% cell survival was observed at 52 μM WR1065, yielding a therapeutic index of 0.25 for the cell culture studies. However, this relatively-poor in vitro therapeutic index is not relevant for the in vivo potential because the parent drug amifostine can be administered at very high doses with virtually no toxicity (see Discussion).

As an additional test of WR1065-induced toxicity, flow-cytometric analysis of cell cycle parameters, was performed using human TCBs grown in the presence of 0, 9.5 and 18.7 μM WR1065. Table 2 shows values for percentage of cells in S-phase, G₂/M-phase and G₀/G₁ phase. We found that exposures of human TCBs to 9.5 and 18.7 μM WR1065 did not significantly alter the TCB cycling, as compared to unexposed cells, adding support to the notion that the TCBs did not sustain unacceptable toxicity at the doses chosen.

Figure 1B shows inhibition of HIV-1-replication, and cell survival determined by Trypan blue, for AZT dose-response experiments (mean ± SE, n = 4 experiments). The figure shows 50% inhibition of virus replication at 5.0 ηM AZT, a dose that was associated with 90% cell survival.

We performed three experiments to examine inhibition of HIV-1 replication with the combination of AZT and WR1065 (Table 3). In each experiment, we compared two AZT dose-response curves, one with increasing doses of AZT alone, and a second with identical concentrations of AZT plus a constant amount of WR1065 added to each well. A representative experiment is shown in Figure 2, in which the increase in % inhibition of virus replication with both AZT and WR1065 is evident by comparing the curves with AZT alone (solid diamond) and AZT plus WR1065 (solid square). In this experiment (Experiment 3 from Table 3) the only dose of AZT that gave less-than-saturating inhibition of HIV-1 replication was 2.2 ηM. This AZT dose was informative because it did not saturate virus inhibition, allowing for further inhibition when WR1065 was added (see Table 3, right column). Due in part to interindividual differences in growth, HIV-1 infection capacity, and specific drug dose used, variability was such that the experiments could not be combined. However, the consistent increase in the % inhibition of HIV-1 replication with the addition of WR1065 to non-saturating doses of AZT (see Table 3, right column) suggests that WR1065 did not inhibit the antiretroviral activity of AZT. On the contrary, combination of WR1065 with AZT did increase the antiretroviral efficacy of AZT.

TCBs from three macaques, which had been chronically infected with SIV for 14 months, were used to test the effect of WR1065 on SIV replication ex vivo. At the time of blood collection, the animals (612, 642 and 674), had plasma titers of 0.10, 0.03 and 6.40 × 10⁶ copies of SIV RNA/ml, and CD4 counts of 376, 635 and 547/ul, respectively. The PBMC were depleted of CD8⁺ T cells, PHA stimulated, and either cultured for 20 days in the absence of WR1065, or cultured for 3 days before the addition of 0, 9.5 or 18.7 μM WR1065 to the medium, and then for an additional 17 days. The medium was changed twice weekly in both culture groups and fresh WR1065 was added at each medium change. Using the MTS assay, cell survival was measured on days 10 (data not shown) and 17 of this experiment (Table 4).

Kinetic p27 data, generated in the cultures with and without WR1065, are illustrated in Figure 3. SIV replication by TCBs from macaque 612 (Figure 3A) cultured in the absence of WR1065 (solid triangle) peaked at day 10. In contrast, SIV replication was reduced approximately 5-fold, to background levels, in the two groups exposed to WR1065 (solid square, hollow diamond) (Figure 3A). The peak virus titer in the macaque 612 TCBs (5200 pg SIV/ml at 10 days) was the lowest of those examined. In TCBs from macaque 642 (Figure 3B), at 7 days, the SIV p27 levels in groups exposed to 0 (solid triangle) and 9.5 (solid square) μM WR1065 were measurable, and only in cells exposed to 18.7 μM WR1065 (hollow diamond) was the SIV titer lowered to background levels. At 7 days, TCBs cultured from macaque 642 had a virus titer of 30,000 pg SIV/ml (Figure 3B), which was much higher than the SIV titers for the other two macaques. Perhaps because of this, an antiviral effect was observed only at the 18.7 μM WR1065 dose in TCBs from macaque 642. In untreated TCBs from macaque 674 (Figure 3C), the virus titer (solid triangle) showed two viral peaks, one at day 7, followed by a decline at day 10, and a second increase for the remainder of the experiment. WR1065-exposed cells (solid square, hollow diamond) from this animal had baseline SIV levels throughout the 17 day culture period, indicating a persistent WR1065-induced inhibition of SIV replication. This inhibition was observed irrespective of fluctuations in the SIV replication pattern found in untreated cultures from different animals.

In summary, these ex vivo experiments performed using macaque TCBs obtained from three chronically-infected macaques demonstrate that WR1065 effectively inhibited production of SIV p27 throughout the 17-day culture period. Furthermore, our finding that 18.7 μM WR1065 was required to inhibit SIV replication in cultures with the highest SIV levels (Figure 3B) suggests that the inhibition of SIV replication is dose dependent.