Abasic Phosphorothioate Oligomers Inhibit HIV-1 Reverse Transcription and Block Virus Transmission across Polarized Ectocervical Organ Cultures

Abstract

In the absence of universally available antiretroviral (ARV) drugs or a vaccine against HIV-1, microbicides may offer the most immediate hope for controlling the AIDS pandemic. The most advanced and clinically effective microbicides are based on ARV agents that interfere with the earliest stages of HIV-1 replication. Our objective was to identify and characterize novel ARV-like inhibitors, as well as demonstrate their efficacy at blocking HIV-1 transmission. Abasic phosphorothioate 2′ deoxyribose backbone (PDB) oligomers were evaluated in a variety of mechanistic assays and for their ability to inhibit HIV-1 infection and virus transmission through primary human cervical mucosa. Cellular and biochemical assays were used to elucidate the antiviral mechanisms of action of PDB oligomers against both lab-adapted and primary CCR5- and CXCR4-utilizing HIV-1 strains, including a multidrug-resistant isolate. A polarized cervical organ culture was used to test the ability of PDB compounds to block HIV-1 transmission to primary immune cell populations across ectocervical tissue. The antiviral activity and mechanisms of action of PDB-based compounds were dependent on oligomer size, with smaller molecules preventing reverse transcription and larger oligomers blocking viral entry. Importantly, irrespective of molecular size, PDBs potently inhibited virus infection and transmission within genital tissue samples. Furthermore, the PDB inhibitors exhibited excellent toxicity and stability profiles and were found to be safe for vaginal application in vivo. These results, coupled with the previously reported intrinsic anti-inflammatory properties of PDBs, support further investigations in the development of PDB-based topical microbicides for preventing the global spread of HIV-1.

INTRODUCTION

Worldwide, there are >35 million human immunodeficiency virus (HIV)-positive individuals, with ∼6,300 new cases of HIV infection occurring per day and an estimated 36 million deaths having occurred since the beginning of the epidemic (1). Heterosexual transmission is responsible for the majority of HIV-1 infections (2), with women representing the fastest growing population of new cases (3). Previous efforts to reduce the risk of HIV-1 being transmitted through sexual intercourse have had only limited success (4). This is largely due to the fact that strategies focused on minimizing the probability of HIV-1 transmission, including abstinence education, promotion of safe sex practices, and socioeconomic equality between the genders, are not widely accepted. Although condoms lower the risk of infection (5, 6), their universal use is likely unachievable (7, 8). Such sociological and behavioral practices, coupled with increasing cases of newly infected women, have necessitated the development of novel strategies that can be implemented to prevent sexually transmitted diseases (9). The need for alternative HIV-1 prevention approaches has become even more critical when one considers the pessimistic prospects of developing an effective vaccine that elicits sterilizing immunity (10).

Currently, preexposure prophylaxis in the form of orally administered antiretroviral (ARV) drugs or topical microbicides appears to be the only effective method available for preventing HIV-1 transmission (11–13). The CAPRISA 004 trial, based on a topical gel containing 1% tenofovir, demonstrated the rather promising result that a vaginal microbicide could moderately reduce the risk of HIV-1 infection (12). The initial success of tenofovir has provided a much-needed impetus in the microbicide field, which has suffered from the failure of several microbicides in clinical trials (14–17; see also http://www.popcouncil.org/news/trial-shows-anti-hiv-microbicide-is-safe-but-does-not-prove-it-effective).

The clinical failures of nonoxynol-9 (N-9), Ushercell (cellulose sulfate), and Carraguard (carrageenan) (7, 14, 18, 19) have influenced the future development of microbicides in a number of ways. First, increased emphasis has been placed on agents that specifically inhibit or interfere with early (preintegration) events in the HIV-1 replication cycle. Second, greater importance has been given to preclinical investigations that provide deeper insights into compound safety, efficacy, and mechanisms of action. Even though the partial success of tenofovir gel has provided the first evidence that microbicides can have a significant impact on sexual transmission of the virus (12), more recent clinical trials involving ARV-based microbicides demonstrated that they lacked efficacy (20), suggesting a need for continuing to improve current strategies and identifying new and effective HIV-1 prevention agents.

We recently demonstrated that phosphorothioate 2′ deoxyribose backbone (PDB) oligomers are nontoxic, stable, and potent inhibitors of HIV-1 and simian immunodeficiency virus (SIV) infections (21). Importantly, PDBs also block Toll-like receptor 7/9 (TLR7/9) activation and are thus capable of suppressing innate immune responses triggered by virus-specific molecules (21). This feature would perhaps assist in preventing the early amplification of HIV-1 infection within mucosal membranes. Indeed, we investigated the anti-inflammatory capacities of these compounds in the context of HIV-1 infection (21), and our results are in agreement with the previously reported TLR7/9-specific antagonistic properties of deoxyribose phosphorothioates (22–24). However, the direct antiviral mechanisms of action of PDBs against HIV-1 are unknown.

In this study, we therefore sought to characterize the inhibitory effects of PDB oligomers against various HIV-1 strains and determine their antiviral mechanisms of action. Our second goal was to expand our previous findings on the safety and stability of PDB oligomers by using an in vivo model of murine cervicovaginal toxicity and by conducting efficacy testing in the presence of seminal plasma and genital secretions. Finally, we set out to assess the abilities of PDB-based compounds to inhibit HIV-1 transmission in a polarized human cervical explant model, as this may represent an important preclinical predictor of microbicide efficacy.

MATERIALS AND METHODS

Ethics statement.

Healthy donor blood and seminal plasma samples were collected following Drexel University Institutional Review Board (IRB) approval. Normal human ectocervical tissue samples were acquired from HIV-1-seronegative premenopausal women as part of their standard care. The genital tissue samples were deidentified and thus exempted from IRB review. Eight-week-old Swiss Webster and C57BL/6 (The Jackson Laboratory, Bar Harbor, ME) female mice were housed in animal facilities in accordance with the Drexel University College of Medicine Institutional Animal Care and Use Committee (IACUC) regulations on the care and protection of laboratory animals. Female rhesus macaques (Macaca mulatta) were housed at Bioqual, Inc. (Rockville, MD), in accordance with the standards of the American Association for Accreditation of Laboratory Animal Care. The protocol was approved by Bioqual's Institutional Animal Care and Use Committee under Office of Laboratory Animal Welfare (OLAW) assurance no. A-3086-01. Bioqual is International Association for the Assessment and Accreditation of Laboratory Animal Care (IAAALAC) accredited, and the procedures were carried out in accordance with the recommendations of the Weatherall report.

Compounds and reagents.

Abasic (1,2-dideoxyribose) phosphorothioate (PDB) compounds were synthesized by TIB-Molbiol (Adelphia, NJ) and IBA (Göttingen, Germany). All studies were performed with a PDB that carried a single 5′-(6-FAM) (FAM, 6-carboxyfluorescein)-tagged thymidine, except where indicated as unlabeled. The PDBs were reconstituted in phosphate-buffered saline (PBS) as 10-mM stocks and diluted appropriately in cell culture medium. Tenofovir, AMD-3100, TAK-779, and T-20 were obtained from the NIH AIDS Research and Reference Reagent Program (Germantown, MD). The synthesis of temacrazine was performed according to previously described procedures (25), and leptomycin B was purchased from Sigma-Aldrich (St. Louis, MO).

Cells, tissues, and viruses.

The following cell lines were obtained through the NIH AIDS Research and Reference Reagent Program: GHOST (3) X4/R5, H9, MOLT4-CCR5, P4-R5 multinuclear activation of a galactosidase indicator (MAGI), MAGI-CCR5, TZM-bl, Sup-T1, and HL2/3 cells; HeLa-R5-16 cells were a kind gift of Roche (Palo Alto, CA) (26). Peripheral blood mononuclear cells (PBMCs) were isolated using Ficoll-Hypaque (Amersham Pharmacia Biotech, Uppsala, Sweden) density centrifugation. Laboratory-adapted HIV-1 and the clinical subtype isolates used in this study were obtained from the NIH AIDS Research and Reference Reagent Program and include HIV-1_IIIB, HIV-1_Ba-L, HIV-1_JR-CSF, HIV-1 92UG031, HIV-1 92BR030, HIV-1 93IN101, HIV-1 CMU02, HIV-1 93BR029, and HIV-1 J_V1083. The origins of HIV-1_SK-1 (27, 28) and HIV-1 99UGA07412 (29) have been described. Multidrug-resistant HIV-1 MDR 769 was originally obtained from the laboratory of Thomas Merigan (Stanford University). This isolate is resistant to the reverse transcriptase (RT) inhibitors zidovudine, didanosine, lamivudine, stavudine, foscarnet, and nevirapine, as well as the protease inhibitors indinavir, saquinavir, and nelfinavir (30). Low-passage virus stocks were prepared using freshly isolated human PBMCs and stored in liquid nitrogen.

In vitro HIV-1 fusion and infection assays.

The methodology used for the in vitro antiviral activity assays as part of the standard algorithm for screening candidate microbicides was followed as previously described (31). Cell viability and efficacy were tested in parallel using a commercially available soluble tetrazolium-based procedure (CellTiter 96 cell proliferation assay; Promega, Madison, WI). Fusion assays assessed the abilities of the compounds to block cell-to-cell fusion mediated by the HIV-1 envelope (Env) present on one Tat-expressing cell line that interacts with CD4 and CXCR4 or CCR5 coreceptors expressed on a separate target cell line possessing a long terminal repeat (LTR)-β-galactosidase (β-gal) reporter. Compounds were used to pretreat target cells (P4-R5 MAGI or MAGI-CCR5) for 1 h at 37°C, prior to the addition of HL2/3 cells (expressing HIV-1 Tat and CXCR4-tropic Env from HXB2/3gpt) or HeLa-R5-16 cells (expressing HIV-1 Tat and CCR5-tropic Env). The incubation was continued for 48 h, after which fusion was monitored by measuring β-gal activity (Tropix Gal-Screen; Applied Biosystems, Grand Island, NY).

Cell-free HIV-1 infection assays were used to detect the capacities of PDB compounds to block the infection of MAGI and TZM-bl cells. The target cells were preincubated with compounds for 15 min at 37°C, following infection with 10 50% tissue culture infectious doses (TCID₅₀) of HIV-1_IIIB or HIV-1_Ba-L. The PDBs were also evaluated in a modified assay in which the compound and virus were preincubated for 1 h prior to the addition of the target cells. At 48 h postinfection, β-gal activity was determined by chemiluminescence.

CD4-dependent cell-to-cell transmission assays were performed to determine whether PDB 14-mer (PDB₁₄) prevented the transmission of HIV-1 from an infected cell to a target cell. CD4-positive GHOST (3) X4/R5 cells served as targets. The virus-transmitting cells were either H9 cells chronically infected with HIV-1_SK1 (CXCR4 tropic) or MOLT4-CCR5 cells chronically infected with HIV-1_JR-CSF (CCR5 tropic). Twenty-four hours prior to the assay, the target cells were seeded into 96-well flat-bottom plates. On the day of the assay, CD4⁺ GHOST (3) X4/R5 target cells were 100% confluent, and the HIV-1-transmitting cells were treated with mitomycin C (200 μg/ml) (Sigma-Aldrich) for 60 min at 37°C. PDB₁₄ was added to the target cells, followed by the addition of 1,000 (CXCR4-tropic) or 4,000 to 8,000 (CCR5-tropic) effector cells. The HIV-1-transmitting cells and targets were incubated with PDB₁₄ for 4 h, followed by three washes. Twenty hours later, the wells were washed again, and viral replication was assessed via an HIV-1 p24 ELISA (PerkinElmer, Shelton, CT).

The HIV-1 postentry assay procedure was the same as the cell-free infection protocol, with the modification of P4-R5 MAGI target cells first being infected with HIV-1_IIIB (10⁵ TCID₅₀/ml) for defined periods of time. After extensive washing to remove unbound virus, 1 μM PDB₁₄ was added to cultures, and β-gal activity was measured 48 h later.

To determine whether PDB₁₄ inhibited infection by primary HIV-1 isolates of various subtypes, phytohemagglutinin (PHA)- and recombinant human interleukin-2 (IL-2)-activated PBMC from HIV- and hepatitis B virus (HBV)-seronegative donors were diluted in fresh medium to a final concentration of 1 × 10⁶ cells/ml, and they were plated in the interior wells of a 96-well round-bottom microplate at 50 μl/well (5 × 10⁴ cells/well). Test dilutions of PDB₁₄ were prepared at a 2× concentration, and 100 μl of each concentration was placed in the appropriate wells, followed by 50 μl of a predetermined dilution of virus stock at a final multiplicity of infection (MOI) of ≈0.1. The PBMC cultures were maintained for 7 days following infection at 37°C in 5% CO₂. Subsequently, cell-free supernatant samples were collected for analysis of reverse transcriptase (RT) activity, as previously described (32). Following the removal of the supernatant samples, cell viability was assessed by adding 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) to the culture wells. The cells were also examined microscopically for any abnormalities during the course of the 7-day infection assay.

In vitro uptake of PDB.

For fluorescence microscopy analysis, P4-R5 MAGI cells (7 × 10⁴) were plated on sterile 12-mm glass covers in 24-well culture plates in Dulbecco's modified Eagle's medium (DMEM)–10% fetal bovine serum (FBS)–1 μg/ml puromycin and cultured for 18 h. The cells were incubated with medium containing 5 μM 6-FAM-labeled PDB₁₄ at 37°C for 30 min and washed 4 times with PBS. The cells were then incubated in the dark for 20 min with 2% paraformaldehyde, washed with PBS, and mounted with ProLong Gold anti-fade reagent-4′,6-diamidino-2-phenylindole (DAPI) (Invitrogen). The slides were analyzed using the Eclipse Ti microscope (Nikon, Melville, NY) with the NIS Element software (Nikon).

To test PDB₁₄ uptake by cells of the murine vaginal tract, vaginal tissue samples were harvested from C57BL/6 mice, cut into small pieces, and digested in RPMI–10% FBS–2 mM l-glutamine–100 U/ml penicillin–100 μg/ml streptomycin sulfate supplemented with 2 mg/ml collagenase D (Roche Diagnostics Corporation, Indianapolis, IN), 4 mg/ml collagenase type IV (Sigma, St. Louis, MO), and 150 μg/ml DNase (Worthington Biochemical Corporation, Lakewood, NJ) for 1.5 h at 37°C on a rotor. The cells were filtered through a 45-μm filter, washed once with RPMI–10% FBS, and counted. A total of 1 × 10⁶ cells were incubated with medium alone or 10 μM 6-FAM-labeled PDB₁₄ for 2 h, washed twice with medium, and stained with fluorochrome-conjugated antibodies directed against CD326 (epithelial cells), F4/80 (macrophages), Gr-1 (neutrophils), CD4 and CD8 (T cells), and CD19 (B cells). The cells were run on an LSRFortessa flow cytometry instrument (BD Biosciences) and analyzed with FlowJo software (Tree Star, Ashland, OR).

Optical biosensor binding assay.

Surface plasmon resonance (SPR) interaction experiments were performed on a ProteOn XPR360 protein interaction array system (Bio-Rad Laboratories, Hercules, CA). The assay was performed at 25°C using standard 1× PBS (pH 7.4), with 0.005% P-20 surfactant. A GLC sensor chip was derivatized by amine coupling using 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (0.2 M) and sulfo-N-hydroxysuccinimide (0.5 M) with HIV-1_YU-2 gp120. HIV-1_YU-2 gp120 was immobilized on the sensor surface (∼3,500 response units). PDB₁₄ and PDB 30-mer (PDB₃₀) were individually diluted in PBS buffer at concentrations ranging from 62.5 nM to 1 μM and subsequently passed over the surface at a flow rate of 100 μl min⁻¹, with a 2-min association phase and a 5-min dissociation phase. The previously characterized gp120-binding peptide, 12p1 (33), and soluble CD4 (sCD4) were used as positive controls in the assay. Regeneration of the surface was performed via a single 18-s pulse of 10 mM glycine (pH 1.5).

Real-time PCR analysis of HIV-1 postentry replication events.

A cellular assay was performed to determine whether PDB₁₄ could block the synthesis of a very early HIV-1 reverse transcription product. Sup-T1 cells (2 × 10⁶/sample) were pretreated with 1 μM PDB₁₄, 1 mg/ml tenofovir, or medium alone for 1 h at 37°C prior to the addition of HIV-1_IIIB at final titer of 10⁵ TCID₅₀/ml. Infections were allowed to continue for 4 h, and the total cellular DNA was subsequently extracted from each sample using the QIAamp DNA blood kit (Qiagen, Valencia, CA). The quantification of RU5 was performed with a forward primer (5′-GCCTCAATAAAGCTTGCCTTGA-3′), a reverse primer (5′-TGACTAAAAGGGTCTGAGGGATCT-3′), and a probe (5′-FAM-AGAGTCACACAACAGACGGGCACACACTA-TAMRA-3′) (TAMRA, 6-carboxytetramethylrhodamine). A standard CCR5 primer-probe set was used for each sample to normalize for DNA recovery (34). Real-time PCRs were performed using the TaqMan Gold PCR procedure (Applied Biosystems, Foster City, CA).

HIV-1 reverse transcriptase RNA-dependent DNA polymerase activity assay.

The ability of PDB₁₄ to inhibit the synthesis of DNA in an RNA-dependent DNA polymerase (RDDP) reverse transcriptase assay was assessed using the Colorimetric RT procedure, as described by the manufacturer (Roche Applied Science, Indianapolis, IN). Briefly, 10 μM PDB₁₄ was incorporated into a reaction mixture containing 0.5 ng of HIV-1 RT, a hybrid poly(rA) × oligo(dT)₁₅ as a template/primer, and digoxigenin- or biotin-labeled nucleotides (dUTP-dTTP mixture). The reaction mixture was incubated for 10 min at 37°C and terminated using 0.48 M EDTA. A sandwich enzyme-linked immunosorbent assay (ELISA) protocol was performed to quantify the reaction products.

HIV-1 reverse transcriptase DNA-dependent DNA polymerase activity assay.

The inhibition of HIV-1 RT DNA-dependent DNA synthesis by PDB₁₄ was determined using fluorescently labeled duplex DNA that was generated by annealing a 42-nucleotide template (5′-TACATACCCATACATAAATCCTAACCTTGAAGAACTCGTCAC-3′) to the 5′ Cy5-labeled primer (5′-ATGTATGGGTATGTATTTAGG-3′), as previously described (35). The reaction was initiated by adding 1 μl of 2 mM deoxynucleoside triphosphates (dNTPs) to 9 μl of a mixture containing 10 nM enzyme, 200 nM substrate, 0.0041 to 3 μM PDB₁₄ in 10 mM Tris-HCl (pH 7.8), 80 mM KCl, 1 mM dithiothreitol (DTT), 10 mM MgCl₂, and 10% dimethyl sulfoxide (DMSO) at 37°C, and the reaction was terminated after 10 min by adding an equivalent volume of a formamide-based gel-loading buffer. Reaction products were evaluated by denaturing polyacrylamide gel electrophoresis and fluorescence imaging (Typhoon Trio+; GE Healthcare, Mickleton, NJ).

Cell-based assays for assessment of HIV-1 Tat and Rev functions.

PDB oligomers were evaluated for their ability to inhibit the function of HIV-1 regulatory proteins using engineered cell lines with either HIV-1 Tat-dependent or Rev-dependent expression of renilla luciferase (36, 37). HeLa cells that stably express firefly luciferase and HIV-1_IIIB Tat or Rev from a bicistronic mRNA under the control of the tetracycline (Tet)-Off promoter were transfected with an HIV-1_NL4-3 LTR- or HIV-1_SF2 Rev-dependent renilla luciferase reporter construct, respectively. Therefore, in the absence of doxycycline, these cell lines express both luciferases and Tat or Rev, with renilla luciferase expression being dependent upon Tat or Rev function and firefly luciferase expression remaining independent of these regulatory proteins. Firefly luciferase was used to assess nonspecific or toxic effects of the test compounds and also as an indicator of Tat or Rev expression levels. HIV-1 Tat or Rev indicator cells were seeded in triplicate into wells of a 96-well plate (2 × 10⁴/well) in the presence of PDB₁₄ or PDB₃₀ (0 to 5 μM). Temacrazine (TMZ), an inhibitor of LTR transcription initiation, served as a positive control in the Tat reporter assay. An active control for the Rev reporter system consisted of leptomycin B, which blocks human chromosome region maintenance 1 (hCRM1)-mediated Rev nuclear export. Following 24 h of incubation with the above compounds, the luciferase expression levels were determined using dual-luciferase assay reagents (Promega, Madison, WI), according to the manufacturer's instructions.

Human vaginal epithelial cell and primary PBMC cytotoxicity assays.

One day prior to performing the assay, human vaginal Vk2/E6/E7 cells were seeded into 96-well flat-bottom plates at a concentration of 10⁵ cells/200 μl/well and maintained at 37°C in 5% CO₂. The PDB compounds were added to the cells at final concentrations ranging from 10 to 100 μM. Following 24 h of continuous exposure, the cells were washed three times with PBS and incubated with 7.5 mg/ml 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) (Sigma) for 2 h at 37°C in 5% CO₂. Following the removal of MTT, intracellular formazan crystals were solubilized for 15 min in 70% isopropyl alcohol and read at 570 nm using a SpectraMax Plus 384 spectrophotometer (Molecular Devices, Sunnyvale, CA), as previously described (38). Readings were also recorded at 690 nm in order to quantify nonspecific absorbance.

Freshly isolated primary human PBMCs were continuously exposed to 100 μM PDB₁₄ or PDB₃₀ for 24 h and subsequently washed and stained for apoptosis using annexin V Cy5.5 (BD Biosciences, San Diego, CA). The PBMCs were stained with annexin V in Hanks' balanced salt solution (HBSS) (Cellgro), 3% heat-inactivated horse serum (Invitrogen, Carlsbad, CA), 0.02% NaN₃, and 2.5 mM CaCl₂ for 30 min on ice, washed two times, and fixed with 1% paraformaldehyde. The samples were acquired on a FACSAria cell sorter and analyzed using the FlowJo software.

Lactobacillus sensitivity assay.

Lactobacillus crispatus and Lactobacillus jensenii were propagated in Lactobacilli MRS broth (Difco, Fisher Scientific, Pittsburgh, PA). This medium facilitates efficient growth of Lactobacillus spp. under facultative anaerobic conditions. To investigate the effect of PDB₁₄ on the growth of L. crispatus and L. jensenii, 10 ml of MRS broth was inoculated with a stab from the glycerol bacterial stock, and the culture was incubated for 24 h at 37°C. On the following day, the bacterial density was adjusted to an optical density at 600 nm (OD₆₀₀) of 0.06. PDB₁₄ was diluted and plated into 96-well round-bottom plates, followed by the addition of Lactobacillus organisms. A penicillin-streptomycin solution at high test concentrations of 1.25 U/ml and 1.25 μg/ml, respectively, was used as the positive control. The plates were incubated for 24 h at 37°C in a GasPak CO₂ bag, and bacterial growth was determined by measuring the OD₄₉₀ using a 96-well spectrophotometric plate reader.

In vivo toxicity testing.

A Swiss Webster mouse model of cervicovaginal toxicity was used to determine whether PDB oligomers induced inflammation and tissue damage within the cervicovaginal mucosa following microbicide application. Six- to 10-week-old female Swiss Webster mice were hormonally synchronized at 7 and 3 days prior to the start of each experiment with a subcutaneous injection of Depo-Provera (Pharmacia and Upjohn Company, Peapack, NJ) diluted in lactated Ringer's saline solution to a final dose of 3 mg/mouse. Following synchronization, the anesthetized mice received an intravaginal application (50 μl) of either 10 μM PDB or 1% nonoxynol-9 (N-9) formulated in a standard hydroxyethyl cellulose (HEC) gel topical delivery vehicle. Untreated mice and mice treated with HEC gel alone were used as negative controls to establish a baseline of the normal architecture and inflammation status of the cervicovaginal mucosa. Three independent experiments were performed, with a total of at least four mice in each treatment group. Each mouse was sacrificed at designated time points following compound application, and the entire reproductive tract was harvested. The excised cervicovaginal tracts and uterine horns were fixed in 4% paraformaldehyde and subsequently embedded in paraffin by standard procedures. Gross morphological analyses were performed on tissues stained with hematoxylin and eosin (H&E). Tissue sections from all animals within a treatment group were examined and scored for toxicity using an Eclipse 80i microscope (Nikon).

PDB stability testing in the presence of semen and vaginal lavage fluid specimens.

Cell-free HIV-1 infection inhibition assays were modified to include a pH transition from 4.0 to 7.0, seminal plasma, or vaginal lavage fluid samples in the medium. In the first set of tests, diluted virus was added (1:1) to medium containing 50% seminal plasma, and 50 μl of this was added to 50 μl of P4-R5 MAGI target cells that were preincubated for 15 min with 2× PDB, resulting in a final concentration of 12.5% seminal plasma (31). Quinobene was used as a control for the seminal plasma and pH transition assays. This compound is active in the standard assays, inactive in seminal plasma, and exhibits reduced activity in the pH transition assay. To assess their stability in the presence of vaginal fluid, lavage fluid samples from healthy female rhesus macaques were combined 1:1 with PDB compounds or tenofovir for 1 h at 37°C, followed by a standard HIV-1 infectivity assay using P4-R5 MAGI cells.

Sperm motility assay.

To determine whether the PDB compounds had contraceptive capacities, PDB₁₄ with and without a single 5′-(6-FAM) moiety was tested in a sperm motility assay. Aliquots of a sperm suspension were loaded into a 20-mm deep disposable chamber (MicroCell; Conception Technologies, USA), prewarmed to 37°C. The sperm samples were untreated or incubated with 10 μM of each PDB compound for 0, 1, 3, or 24 h. Computer-assisted sperm motion analysis was performed using a digital image analyzer (CASA, Hamilton Thorne IVOS V 10.8) at the rate of 60 frames/s.

Ex vivo cervical organ cultures.

Organ cultures were established by using a Transwell system (39, 40) in which cervical tissue was placed into the top chamber. The edges around the explant were then sealed with Matrigel (BD Biosciences). Human CD8-depleted PBMCs were present at the bottom of the chamber and served as indicator cells for HIV-1 transmission. Transwells consisting of Matrigel alone in the top chamber served as a negative control, whereas Transwells with membrane alone served as a positive control. HIV-1_Ba-L at a titer of 10⁵ TCID₅₀ was preincubated with PDB oligomers at a final concentration of 2.5 to 25 μM or medium alone for 1 h at 37°C. The mixtures were then added to the tissue, and incubation was allowed to continue at 37°C for 7 to 14 days. The HIV-1 p24 levels were measured in the culture supernatant present in the bottom chambers of the tissue wells.

Data analysis, quality control measures, and statistical tests.

For the in vitro HIV-1 assays, defined quality control limits involving the following parameters were used to validate the results: total virus infection or envelope fusion, triplicate reproducibility, and response to positive and negative controls. Both antiviral efficacy (measured by the concentration of compound that inhibits virus transmission in a cellular assay by 50% and 90% [EC₅₀ and EC₉₀, respectively]) and compound cytotoxicity (the concentration of compound that results in a 50% loss of cell viability [TC₅₀]) are determined by linear regression of the resulting data. From these parameters, a therapeutic index (TI₅₀, calculated by TC₅₀/EC₅₀) was derived and used to identify active compounds.

Fitted dose-response curves for the DNA-dependent DNA polymerase (DDDP) assay were obtained by setting the reaction yields in the absence of an inhibitor to 100%. The concentration of compound that inhibits enzymatic activity in a biochemical assay by 50% (IC₅₀) and the Hill slope parameter (analogous to the Hill coefficient n) were derived from a dose-response curve using GraphPad Prism software (La Jolla, CA) and the following modified Hill equation: % inhibition = inhibitor concentrationⁿ/(IC₅₀ⁿ + inhibitor concentrationⁿ), where n is the Hill slope (41, 42). All fitted curves exhibited an R² value of ≥0.98.

The Shapiro-Wilk W test for normality and Student's t test of the paired samples were used for statistical analysis with the JMP data analysis program (SAS, Cary, NC). P values of <0.05 were considered significant. The reported P values are 2 tailed.

RESULTS

Anti-HIV-1 activities of PDBs are size dependent.

PDB₁₄ is a potent inhibitor of HIV-1 infection (21). Based on these observations, we postulated that different lengths of PDB may exhibit various degrees of inhibitory activity. Therefore, we sought to determine the minimum length of PDB that is active, as this may increase the availability and effectiveness of the compound. PDB oligomers of different sizes were therefore examined for their inhibitory activity with respect to HIV-1 infectivity. With increasing length, the antiviral efficacy of PDB was enhanced (Fig. 1A). Oligomers composed of 9 deoxyribose monomers (PDB₉) showed little activity against HIV-1 infectivity compared to that of the 14-mer and 30-mer PDBs (Fig. 1A).

FIG 1.

The anti-HIV activities of PDB oligomers are size dependent and not affected by fluorophore labeling. (A) Inhibition of HIV-1_IIIB infectivity mediated by a PDB 30-mer (PDB₃₀), 14-mer (PDB₁₄), and 9-mer (PDB₉) in a single-cycle MAGI assay. HIV-1 was exposed to PDB oligomers (0.01 to 1 μM) for 1 h, and the HIV-PDB mixture was added to MAGI cells. β-gal activity was quantified 48 h later. Pooled data show the means from 4 independent experiments ± standard error of the mean (SEM). The EC₅₀ of each compound in this assay is as follows: PDB₃₀, ∼10 nM; PDB₁₄, 70 to 200 nM; PDB₉, >1,000 nM. (B) Chemical structure of PDB₁₄ labeled with 6-FAM on the 5′ position. A single thymidine is present on the 6-FAM-labeled oligomer for the purpose of quantification during the purification process. (C) Cellular penetration of 6-FAM-labeled PDB₁₄ analyzed by confocal microscopy. DAPI staining shows total nuclei. P4-R5 MAGI cells were incubated with PDB₁₄ for 30 min, fixed with 2% paraformaldehyde, mounted, and analyzed. DIC, differential interference contrast. (D) Ex vivo uptake of fluorescent PDB₁₄ compound by primary murine vaginal cells. Tissues were dissected, and single-cell suspensions were generated by enzymatic digestion. The cells were incubated for 2 h with PDB₁₄, washed, stained for subpopulations (epithelial cells, CD326; macrophages, F4/80; neutrophils, Gr-1) and analyzed by flow cytometry. (E) Dose response for inhibition of HIV-1_Ba-L by PDB₁₄ with and without the 5′,6-FAM substitution, as assessed using the TZM-bl assay. *, P < 0.05; **, P < 0.01; ***, P < 0.003; n.s., not statistically significant. P values were determined by Student's t test.

Cellular uptake and antiviral activity of PDB₁₄ are not diminished by fluorophore labeling.

Because a fluorescein-labeled conjugate of a PDB-based compound can be utilized for optical imaging studies and to establish an easy and inexpensive method for monitoring adherence during clinical trials in the field, it was important to determine whether such a compound retained cell-penetrating properties and activity against HIV-1. Therefore, PDB₁₄ was modified to carry a single 5′-6-FAM-tagged thymidine (Fig. 1B), and cellular uptake and antiviral efficacy were subsequently assessed. PDB₁₄ coupled with a commercially available fluorescein tag was rapidly internalized by P4-R5 MAGI cells (Fig. 1C) and ex vivo cervicovaginal cells, including epithelial cells, macrophages, and granulocytes (Fig. 1D). Notably, 5′-FAM modification did not significantly affect anti-HIV-1 activity, as a PDB derivative lacking this substitution exhibited equivalent antiviral efficacy (Fig. 1E).

PDB₁₄ does not inhibit HIV-1 attachment/fusion.

Because PDB oligomers are negatively charged, we reasoned that larger oligomers (30-mers) might interfere with viral attachment/fusion. Reporter cells were therefore incubated with PDB₃₀ or PDB₁₄, followed by the addition of effector cells expressing HIV-1 Env. This assay is not dependent on reverse transcription or integration but recapitulates the mechanism of virus-to-cell binding, fusion, and Tat-dependent transcription of a luciferase reporter. While PDB₃₀ blocked HIV-1 entry to the same extent as the T-20 fusion inhibitor (Fig. 2A), PDB₁₄ did not prevent binding/fusion in assays involving CXCR4- (Fig. 2A) and CCR5-utilizing (Fig. 2B) HIV-1 Env (Table 1). Cell viability was not affected by the presence of PDB₁₄ or T-20 (Fig. 2B). Furthermore, SPR analysis revealed that in comparison to sCD4 (Fig. 3A) and the 12p1 peptide (Fig. 3B), PDB₁₄ (Fig. 3C) did not directly interact with HIV-1 gp120. Interestingly, we also observed a lack of direct binding of PDB₃₀ to this HIV-1 surface glycoprotein that enables viral attachment and entry into host cells (Fig. 3D).

FIG 2.

PDB₁₄ does not inhibit attachment/fusion between the HIV-1 envelope and target cells. (A) Differential abilities of PDB₃₀ and PDB₁₄ to inhibit the interaction of HIV Env with CD4-expressing LTR β-gal reporter cells. P4-R5 MAGI cells were incubated with PDB oligomers or T-20 (active control), followed by the addition of HL2/3 effector cells that express CXCR4-tropic HIV-1 Env (HXB2/3gpt) on the cell surface and Tat intracellularly. (B) Inability of PDB₁₄ to prevent attachment/fusion of HeLa-R5-16 cells (expressing HIV-1 Tat and CCR5-tropic Env) with MAGI cell targets. Identical conditions were used for toxicity, but the endpoint was detected via an MTS reaction. The error bars depict the averages ± SEM from 3 independent experiments. *, P < 0.0003; **, P < 0.0001. P values were determined by Student's t test.

TABLE 1.

Activities of PDB₁₄ in selected cell culture-based HIV-1 infection assays compared to those of standard antiviral compounds

Compound (control status) by assay type	Antiviral activity (μM)a		TC50 (µM)b	TI50c
	EC50	EC90
CCR5-tropic HIV-1 fusion assay
PDB14	>10	>10	>10	NAd
T-20 (active)	0.02	0.08	>10	>500
TAK-779 (active)	0.003	0.01	>10	>3,333
AMD-3100 (inactive)	>10	>10	>10	NA
CXCR4-tropic HIV-1 fusion assay
PDB14	>10	>10	>10	NA
AMD-3100 (active)	0.001	0.01	>10	>10,000
TAK-779 (inactive)	>10	>10	>10	NA
Cell-free CCR5-tropic HIV-1 infection
PDB14	0.07	0.99	>10	>143
TAK-779 (active)	0.004	0.17	>10	>2,500
AMD-3100 (inactive)	>10	>10	>10	NA
Cell-free CCR5-tropic HIV-1 infection (virus pretreatment)
PDB14	0.07	1.96	>10	>143
TAK-779 (active)	0.004	0.2	>10	>2,500
AMD-3100 (inactive)	>10	>10	>10	NA
Cell-free CXCR4-tropic HIV-1 infection
PDB14	0.22	1.42	>10	>45
AMD-3100 (active)	0.002	0.02	>10	>5,000
TAK-779 (inactive)	>10	>10	>10	NA
CD4-dependent CCR5-tropic HIV-1 cell-to-cell infection
PDB14	>10	>10	>10	NA
TAK-779 (active)	0.008	0.09	>10	>1,250
AMD-3100 (inactive)	>10	>10	>10	NA
CD4-dependent CXCR4-tropic HIV-1 cell-to-cell infection
PDB14	>10	>10	>10	NA
AMD-3100 (active)	0.008	0.06	>10	>1,250
TAK-779 (inactive)	>10	>10	>10	NA

^aEC₅₀ and EC₉₀, concentration of compound that inhibits virus infection or HIV-1 Env attachment/fusion by 50% and 90%, respectively.

^bCompound cytotoxicity was measured by the TC₅₀, the concentration of compound that results in a 50% loss of cell viability.

^cThe 50% therapeutic index (TI₅₀) was calculated by TC₅₀/EC₅₀.

^dNA, not applicable.

FIG 3.

PDB compounds do not directly interact with HIV-1 gp120. Lack of direct binding of PDB₁₄ and PDB₃₀ to surface-immobilized HIV-1_YU-2 gp120. Sensorgrams for the direct binding of sCD4 (active control) (A), 12p1 peptide (active control) (B), PDB₁₄ (C), and PDB₃₀ to HIV-1 envelope (D) are shown. HIV-1_YU-2 gp120 was immobilized on the surface of a ProteON XPR 360 system. The response (denoted as response units [RUs]) curves are shown for increasing concentrations of compound binding to immobilized HIV-1_YU-2 gp120. The results are representative of 3 independent experiments. K_D (M), equilibrium dissociation constant.

PDB₁₄ inhibits early stage cell-free HIV-1 infection.

We next characterized PDB₁₄-mediated anti-HIV-1 activity in selected time-of-addition assays. Importantly, determining the stage of the viral replication cycle at which PDB₁₄ inhibits HIV-1 infection might reveal the mechanism of action. Therefore, PDB₁₄ was tested in a standard cell-free infection in which it was used to preincubate MAGI cells for 15 min before HIV-1 was added to the cultures and also in a modified assay in which the virus was pretreated with PDB₁₄ before its addition to the target cells. PDB₁₄ demonstrated the same EC₅₀ when tested against HIV-1_Ba-L in the standard or modified assay (Fig. 4A and Table 1) and was similarly active against HIV-1_IIIB in the standard assay (Fig. 4B and Table 1). PDB₁₄ did not inhibit CD4-dependent infection by CCR5- (Fig. 4C) or CXCR4-utilizing cell-associated virus (Fig. 4D and Table 1). Notably, however, PDB₁₄ was found to block infection when added to the target cells after HIV-1 adsorption (Fig. 3E), but this inhibition was time dependent, suggesting that PDB₁₄ inhibits events that occur early after HIV-1 entry.

FIG 4.

PDB₁₄ blocks early infection by CCR5- and CXCR4-tropic cell-free HIV-1. The activity of PDB₁₄ in selected time-of-addition HIV-1 infectivity assays is shown. (A) Inhibition of cell-free HIV-1_Ba-L following a standard 15-min pretreatment of P4-R5 MAGI cells with PDB₁₄ prior to infection or a 1-h preincubation of virus before addition to susceptible targets. (B) Ability of PDB₁₄ to block cell-free HIV-1_IIIB infection of pretreated (15 min) P4-R5 MAGI cells. HIV-1 infection was measured as a function of β-gal activity 48 h after the addition of the virus. Evaluation of PDB₁₄ in CCR5-tropic (C) and CXCR4-tropic (D) CD4-dependent cell-to-cell HIV-1 infection assays. HIV-1-infected effectors and CD4⁺ GHOST (3) X4/R5 target cells were incubated with PDB₁₄ for 4 h, followed by extensive washing. Cell-associated HIV-1 Gag p24 was quantified 24 h later by ELISA. TAK-779 (CCR5 inhibitor) and AMD-3100 (CXCR4 inhibitor) were included as active controls. (E) Time kinetics of PDB₁₄-mediated postentry HIV-1 suppression. P4-R5 MAGI cells were infected with HIV-1_IIIB for 30 min, 1 h, and 2 h, followed by washout of the virus inoculum and the addition of 1 μM PDB for 48 h. The cells were harvested, and the total enzymatic activity of β-gal was measured in the lysates. The results are representative of ≥3 independent experiments. (F) Representative results depicting PDB₁₄-mediated inhibition of the multidrug-resistant HIV-1 isolate MDR 769. Primary PBMC were treated with 5 μM PDB₁₄ prior to infection. Following 7 days of incubation, virus replication was assayed by determination of reverse transcriptase activity in cell-free supernatants. The means ± SEM of the results from triplicates are shown.

HIV-1 clinical isolates of various subtypes are inhibited by PDB₁₄, including a multidrug-resistant strain.

To be effective in the context of the HIV/AIDS pandemic, inhibitors of HIV-1 infection must possess antiviral efficacies against a number of different regional HIV-1 subtypes found throughout the world. PDB₁₄ was therefore evaluated for its activity against primary HIV-1 isolates of subtypes A, B, C, D, E, F, and G. In antiviral assays using activated primary human PBMCs, PDB₁₄ was an effective inhibitor of all the viral subtypes examined, with EC₅₀s ranging from 0.089 to 2.5 μM (Table 2). Most notably, PDB₁₄ reduced infections of activated primary human PBMCs (>60% inhibition) by the multidrug-resistant HIV-1 MDR 769 isolate (Fig. 4F).

TABLE 2.

Antiviral efficacy of PDB₁₄ against clinical HIV-1 isolates in primary human PBMCs^a

Primary HIV-1 isolate	HIV-1 subtype	Tropismb	EC50 (μM)c	TC50 (µM)d	TI50e
92UG031	A	CCR5	0.484	>5	>10
92BR030	B	CCR5	0.277	>5	>18
93IN101	C	CCR5	2.493	>5	>2
99UGA07412	D	CCR5	0.430	>5	>12
CMU02	E	CXCR4	0.089	>5	>56
93BR029	F	CCR5	1.213	>5	>4
JV1083	G	CCR5	0.357	>5	>14

^aShown is the PDB₁₄-mediated inhibition of clinical HIV-1 isolate replication in PBMC cultures as determined by reverse transcriptase activity in cell-free supernatants.

^bIndicates coreceptor tropism (CCR5 or CXCR4 utilizing) for each virus strain.

^cEC₅₀, concentration required to achieve 50% inhibition of virus replication.

^dCompound cytotoxicity was measured by the TC₅₀, the concentration of compound that results in a 50% loss of cell viability.

^eThe 50% therapeutic index (TI₅₀) was calculated by TC₅₀/EC₅₀.

PDB₁₄ blocks HIV-1 reverse transcription.

Because PDB₁₄ did not function as an HIV-1 entry inhibitor but did inhibit infection following the adsorption of virus to the target cells in a time-dependent manner (Fig. 4E), we proceeded to determine whether it inhibited the next stage of the viral replication cycle. Indeed, similar to tenofovir, PDB₁₄ inhibited the synthesis of HIV-1 minus-strand strong-stop DNA (RU5), the earliest product of HIV-1 reverse transcription (Fig. 5A). To explore whether PDB₁₄ can directly inhibit reverse transcription, we tested the oligomer in two different biochemical assays of HIV-1 RT activity. PDB₁₄ was found to suppress both RDDP (Fig. 5B) and DDDP HIV-1 RT activity at an IC₅₀ of 172 nM (Fig. 5C and D).

FIG 5.

PDB₁₄ inhibits HIV-1 reverse transcription. (A) Inhibition of HIV-1 reverse transcription by PDB₁₄ as depicted in Sup-T1 cells pretreated with 1 μM PDB₁₄, 1 mg/ml tenofovir (active control), or medium alone for 1 h, followed by the addition of HIV-1_IIIB. At 4 h postinfection, total cellular DNA was extracted for quantification of RU5 synthesis. A standard CCR5 primer-probe set was run for each sample to normalize for DNA recovery and loading. Pooled data showing the mean values from 4 independent experiments are shown. The error bars depict the SEM. (B) PDB₁₄-mediated inhibition of HIV-1 RT RDDP activity in vitro. PDB₁₄ was incorporated into a 10-min HIV-1 RT reaction using a hybrid poly(rA) × oligo(dT)₁₅ as a template/primer, enzyme, and digoxigenin- and biotin-labeled nucleotides. The quantification of the synthesized DNA as a parameter for RT activity was performed using a sandwich ELISA procedure. (C) Inhibition of DDDP activity of HIV-1 RT by PDB₁₄. PDB₁₄ was incorporated into a 10-min HIV-1 RT reaction using a hybrid oligo(dT)₁₅ as a template/primer, enzyme, and fluorescently labeled nucleotides. DDDP activity was determined at inhibitor concentrations of 4.1 nM to 3,000 nM. The migration positions of the unextended and fully extended primers (arrows) are indicated. (D) Representative dose-response curve for inhibition of DDDP activity by PDB₁₄. The IC₅₀ of PDB₁₄ in this assay is 172 nM. All error bars represent the SEM of 3 independent experiments. *, P < 0.05. A Student's t test was used to derive P values.

PDBs do not affect the functional activities of HIV-1 Tat or Rev regulatory proteins.

Because of the chemical nature of PDBs, we reasoned that these molecules may have the potential to nonspecifically interfere with Tat/transactivation response (TAR) element or Rev/Rev response element (RRE) RNA interactions. To explore this possibility, we assessed the effects of PDB₁₄ and PDB₃₀ in cell-based assays designed to directly measure the inhibition of HIV-1 Tat (Fig. 6A) and Rev (Fig. 6B) functions. Neither PDB₁₄ nor PDB₃₀ affected the ability of HIV-1_IIIB Tat to transactivate an HIV-1_NL4-3 LTR, as determined by expression of a downstream renilla luciferase reporter (Fig. 6A). This was in contrast to temacrazine, an inhibitor of LTR-driven transcription initiation (Fig. 6C).

FIG 6.

PDB-based antiviral compounds do not affect the functions of HIV-1 Tat and Rev regulatory proteins. Dose-response assays to determine whether PDB₁₄ and PDB₃₀ affect Tat-dependent (A) or Rev-dependent (B) renilla luciferase expression. HeLa cells containing stably integrated expression plasmids for renilla and firefly luciferases are shown. Renilla luciferase is expressed in a Tat- or Rev-dependent context, while the firefly luciferase is expressed in a Tat- or Rev-independent fashion. Engineered cells were treated with the indicated concentrations of each compound, and luciferase activity (in relative light units [RLUs]) was measured following 24 h of exposure. (C) TMZ, an inhibitor of LTR transcription initiation, served as an active control in the Tat reporter assay. (D) Leptomycin B, which blocks hCRM1-mediated Rev nuclear export, served as a positive control in the Rev reporter system. The assays were performed in triplicate, and the error bars represent the SEM. The results are representative of the values from ≥2 independent experiments.

Next, we assessed whether PDB compounds interfere with the regulatory function of HIV-1 Rev by employing a widely used assay based on a standard Rev expression vector and Rev-responsive reporter. The reporter construct contains the renilla luciferase gene and an RRE sequence, both of which are positioned between HIV-1 splice sites under the control of a Tet-Off promoter. Therefore, in the absence of doxycycline, unspliced renilla luciferase-encoding transcripts are shuttled out of the nucleus and translated in a Rev-dependent manner, resulting in elevated levels of renilla luciferase. We determined that PDBs do not affect HIV-1 Rev function (Fig. 6B), unlike leptomycin B, which blocks hCRM1-mediated Rev nuclear export (Fig. 6D).

PDB oligomers exhibit no toxicity on primary human PBMCs, vaginal epithelial cells, or normal vaginal flora.

In order to exclude any possible toxicity associated with PDB₁₄ or PDB₃₀, we incubated freshly isolated human PBMCs with PDBs for 24 h (continuous exposure) and subsequently stained them for cell death using annexin V. The cells were then analyzed by flow cytometry. Annexin V binds to phosphatidylserine that is exposed on the surface of early and late apoptotic and necrotic cells. In this very sensitive cellular death assay, neither PDB₁₄ nor PDB₃₀ at a concentration as high as 100 μM induced apoptosis or necrosis of PBMCs (Fig. 7A). In addition, the continuous exposure of human VK2 vaginal epithelial cells to PDB oligomers at concentrations of 10 to 100 μM did not affect viability after 24 h (Fig. 7B). Furthermore, PDB-based compounds were not toxic and did not inhibit the growth of two different Lactobacillus species (Fig. 7C) compared to that with standard antibiotics, which served as active controls (Fig. 7D).

FIG 7.

PDB compounds are nontoxic in vitro and in vivo. (A) A 24-h exposure to PDB compounds had no effect on the viability of primary human PBMCs. PBMCs from healthy donors were either untreated or treated with 100 μM PDB₁₄ or PDB₃₀ for 24 h. Cell death was measured by annexin V staining and flow cytometry. The percentages in the plots indicate the frequency of dead cells. The results are representative of the values from five different healthy donors. (B) A continuous 24-h exposure of VK2 cells to high concentrations of PDB compounds had no effect on their viability. Viability was measured by an MTT assay. Pooled data from 3 experiments are shown, with the error bars representing the SEM. PDB does not affect the growth of normal vaginal flora Lactobacillus species. (C) To assess the effect of PDB₁₄ on L. crispatus and L. jensenii growth, the compound was diluted and added to bacterial cultures. (D) A penicillin-streptomycin solution was used as the positive control for inhibition of bacterial growth. (E) Intravaginal application of a formulated PDB microbicide gel is nontoxic. The integrity of cervical and vaginal tissue following a single application of HEC gel alone, 5% N-9 HEC gel, or 10 μM PDB microbicide gel is shown. Each gel was applied to Swiss Webster mice, and the mice were sacrificed at 4 h after application. Representative H&E-stained sections from 3 independent experiments with ≥4 mice per treatment group are shown. The arrows denote epithelial damage and sloughing.

In vivo application of a formulated PDB microbicide gel is nontoxic.

To determine whether phosphorothioate oligomer exposure is associated with toxicity in vivo, 10 μM PDB formulated in HEC gel was instilled into the vaginal canals of Swiss Webster mice, and microscopic examinations of genital tissues were performed. Following intravaginal application of PDB microbicide gels, the gross morphological appearances of the vaginal and cervical epithelia were similar to that of the control (empty HEC gel) tissue following 4 h of exposure (Fig. 7E). In contrast, 5% N-9 gel induced a marked sloughing of epithelial tissue (Fig. 7E).

PDB compounds are active against HIV-1 in the presence of seminal plasma and vaginal lavage fluid.

Seminal plasma can abrogate the anti-HIV activities of microbicides due to positively charged polyamines, including spermine, spermidine, and putrescine. Additionally, hydrolytic enzymes produced in the vaginal tract may destroy the functions of topical antiviral inhibitors. It was therefore important to establish whether the anti-HIV activities of PDB oligomers are susceptible to seminal plasma or vaginal fluid inhibition. We found that PDB₁₄ retained activity in the nanomolar range following pH transition or exposure to seminal plasma (Fig. 8A). In contrast, the antiviral activity of quinobene was abrogated in the presence of 12.5% seminal plasma (Fig. 8B). Additionally, PDB compounds retained anti-HIV activity in the presence of rhesus macaque vaginal fluid, similar to that with tenofovir (Fig. 8C). To determine whether PDB oligomers possess contraceptive capacity, PDB₁₄ with and without FAM was tested using a sperm motility assay. A longitudinal evaluation of the sperm motion parameters demonstrated no effect of PDB on sperm motility during an incubation period of up to 24 h (Fig. 8D).

FIG 8.

PDB₁₄ is stable in the presence of seminal plasma and cervicovaginal fluid and does not affect sperm motility. (A) PDB₁₄ is stable in the presence of seminal plasma, compared to quinobene (B). Compounds were tested in the standard cell-free CCR5-tropic HIV-1 infection assay and protocols that were modified to include a pH transition step or 12.5% human seminal plasma. The increase above 100% virus control at the lower compound concentrations reflects assay variability and not the true infectivity enhancement of the compound. (C) PDB₁₄ (10 μM) stably inhibits HIV-1 infection in the presence of rhesus macaque cervicovaginal lavage fluid, and this is similar to the antiviral activity of tenofovir (1 mg/ml). PDB₁₄ was incubated with medium alone or medium containing 50% rhesus macaque vaginal lavage fluid and subsequently used to inhibit cell-free HIV-1_IIIB infection of P4-R5 MAGI cells. β-gal activity was assessed at 48 h postinfection. (D) PDB₁₄ does not affect sperm motility. Semen was combined with saline or 10 μM PDB₁₄ with or without FAM. Sperm motility was subsequently examined at 0, 1, 3, and 24 h. The results are representative of the values from ≥2 independent experiments. The error bars represent the SEM for triplicate samples.

PDB oligomers inhibit HIV-1 transmission across organ cultures.

We have demonstrated that PDB oligomers inhibit HIV-1_Ba-L infection in primary PBMC cultures (21). To further validate the potential use of PDB-based compounds as microbicides, it was important to determine whether they could block HIV-1_Ba-L transmission across organ cultures of human genital tissue. The antiviral activities of PDB oligomers were evaluated in an ex vivo ectocervical tissue explant model. This dual-chamber polarized system is ideal for studying the transmission of HIV-1 in the human female genital tract because of the broad range of cells present and the fact that the virus must cross mucosal and epithelial barriers, similar to what occurs during sexual HIV-1 transmission in vivo. As shown in Fig. 9, both PDB₁₄ and PDB₃₀ reduced HIV-1_Ba-L transmission to primary PBMCs across the cervical mucosa. At days 7 (Fig. 9A) and 14 (Fig. 9B) postinfection, a >60% inhibition of HIV-1 transmission was achieved with 25 μM PDB₁₄. Similar activity was observed with 2.5 µM PDB₃₀ (Fig. 9A and B).

FIG 9.

PDB oligomers block HIV-1 transmission in primary cervical histocultures. The inhibition of HIV-1 transmission across human cervical mucosa is shown. PDB oligomers were mixed with HIV-1_Ba-L and applied to polarized ectocervical explants. HIV-1 Gag p24 produced on day 7 (A) and day 14 (B) in the bottom well in the presence or absence of each PDB compound was measured by ELISA. The results shown represent the mean ± SEM of quadruplicate biopsy samples. Pooled data from 3 independent experiments performed with 3 different tissue donors are shown. *, P < 0.05; **, P < 0.005; n.s., not statistically significant. P values were determined by performing a Student's t test.

DISCUSSION

Currently, the most promising topical HIV-1 microbicide candidates are postentry ARV inhibitors, such as tenofovir and UC-781 (43). Despite the moderate success of 1% tenofovir gel at reducing the risk of sexual HIV-1 acquisition, a highly specific inhibitor based on a single mechanism of action raises safety concerns related to viral evolution and the emergence of escape mutants (44). Additionally, it has become increasingly clear that genital tract inflammation may influence the degree of protection that any microbicide needs to achieve in order to be efficacious (34, 45–48). Therefore, new ARV compounds are needed to complement current microbicide strategies, especially those that target multiple key drivers of HIV-1 transmission and those that are less likely to engender resistance based on their mechanisms of action.

PDB oligomers are a new promising class of microbicide candidates, as they exhibit very potent antiviral activity with no evidence of toxicity and possess intrinsic anti-inflammatory properties (21). The combination of such dual antiviral/anti-inflammatory action makes this family of microbicides unique. As we (21) and others (22–24) have previously reported, phosphorothioate deoxyribose molecules can inhibit the triggering of TLRs. The ability of PDB-based compounds to specifically inhibit TLR7/9 activation is a desirable property that may reduce HIV-1-elicited proinflammatory cytokine/chemokine production, which is believed to facilitate the early dissemination of virus from the cervicovaginal tract (46, 49). However, the direct antiviral mechanism of action of these inhibitors has not been elucidated.

In this study, we demonstrate that while large abasic PDB₃₀ oligomers block HIV-1 attachment or fusion, a smaller PDB₁₄ exerts an inhibitory effect following cellular entry and during the initial stages of infection. Accordingly, our results suggest that PDB₁₄ may directly inhibit an essential component of early viral replication. This is based on a number of results, including the observation that the addition of PDB₁₄ postinfection still has an inhibitory effect at 30 min and 1 h following virus adsorption but has no activity after 2 h. Furthermore, PDB₁₄ failed to block HIV envelope-mediated cell fusion or to interact with gp120 directly. Similar to PDB₁₄, we found that PDB₃₀ did not bind to the viral envelope, which suggests that this larger compound may be interacting with a cellular receptor required for HIV-1 or an entry step following the formation of the gp41 prehairpin intermediate (i.e., before mixing of the cell-viral membrane and pore formation). The inability of both compounds to interfere with Tat protein-TAR and Rev peptide-RRE interactions, as well as Tat- or Rev-mediated reporter gene expression, suggests that PDB-based inhibitors may not affect the transactivation of HIV-1 LTR-directed gene expression and splicing. Therefore, unlike PDB₃₀, PDB₁₄ does not inhibit attachment and/or fusion between HIV-1 and target cells, but it interferes with infection processes that occur early after entry.

Based on our previous report that PDB₁₄ inhibits HIV-1-triggered TLR7 activation (21), we speculate that this oligomer may directly interact with viral RNA. This is further supported by the observation that PDB oligomers can inhibit infection by a number of different RNA viruses, including HIV-1, SIV, and influenza type A virus (21). Kinetic studies demonstrate that the timing of PDB₁₄-mediated inhibition of HIV-1 infection corresponds to reverse transcriptional initiation, and this was confirmed by our finding that the compound prevents the synthesis of minus-strand strong-stop DNA. In addition, we showed that PDB₁₄ inhibits reverse transcription in cell-free RDDP and DDDP RT activity assays. Therefore, the exact inhibitory mechanism(s) exerted by PDB₁₄ on HIV-1 reverse transcription is the subject of ongoing studies. At this time, we cannot exclude the possibility that PDB₁₄ directly interacts with RT to inhibit its function. Similarly, we have not ruled out the potential for PDB₁₄ to act as a bona fide TLR-binding antagonist, as has been reported for similar compounds (22–24). Further investigations are required to identify and characterize the physical interactions between PDB₁₄ and elements of HIV-1 virions, as well as innate immune receptors.

Another important advantage of PDB₁₄ is that it is amenable to fluorophore labeling. As shown in this study, 5′-FAM substitution does not significantly alter its activity. The core fluorescent moiety in FAM is fluorescein, which is used extensively as a diagnostic tool (45, 50). FAM labeling may provide an additional benefit for PDB₁₄ in terms of establishing a fluorescence assay to detect and monitor the distribution/dilution of a PDB₁₄-based microbicide in vivo. This feature would be of great utility in view of the important issue of adherence, which was highlighted in the VOICE trial (20).

An effective microbicide needs to have broad inhibitory activities against multiple strains of HIV-1. The results presented in this study confirm our previous findings that PDB₁₄ is a potent inhibitor of HIV-1_Ba-L and HIV-1_IIIB, suggesting that inhibition is coreceptor independent. However, it was important to determine the breadth of PDB₁₄ antiviral activity by testing the compound against primary HIV-1 isolates of various subtypes. In addition to the most predominant subtypes, namely, A, B, C, and D, we expanded our analysis to include viruses belonging to clades E, F, and G. Notably, PDB₁₄ inhibited a panel of both CCR5- and CXCR4-using primary HIV-1 isolates, representing the currently circulating major subtypes. These results suggest that PDB₁₄ targets a conserved element that is essential to viral replication, and they further underscore its sequence-independent nature.

The failure of PDB₁₄ to prevent cell-associated HIV-1 infection may be attributed to limitations associated with currently available assays to explicitly monitor cell-to-cell transmission (51, 52). If more sensitive methods emerge and reveal that PDB₁₄ cannot efficiently block cell-associated HIV-1 transmission, these results should not preclude its development as a microbicide. This is supported by several factors: (i) cell-free virus in seminal fluid is considerably more infectious than is cell-associated virus (53), (ii) the inhibition of cell-free HIV-1 or its cellular targets in vivo is likely to be more efficacious than the targeting of cell-associated forms (44), and (iii) animal challenge studies involving strategies that do not inhibit cell-to-cell transmission demonstrate protection against infection (52, 54, 55).

The results presented in this study further support the selection of PDB₁₄ as a promising microbicide candidate. Unlike polyanionic compounds that lacked efficacy in clinical trials, previously described phosphorothioate oligodeoxynucleotides of similar size (56), and PDB₃₀, PDB₁₄ does not inhibit HIV-1 attachment to cellular targets. Instead, low-molecular-weight PDB₁₄ blocks early postentry HIV-1 infection. In addition, the ability of any putative microbicide to work intravaginally while suspended in gel will require transit from the delivery vehicle. When formulated in HEC gel, PDB₁₄ was previously shown to be effective at inhibiting HIV-1 infection (21). In contrast, PDB₃₀ is more prone to stick to a topical gel matrix and consequently to be released more slowly due to its overall greater negative charge. Furthermore, microbicides that function by disrupting the interaction between the positive charges on the viral envelope and the negative charges on the cell surface are typically susceptible to acidic or alkaline pH that can alter the secondary or tertiary structures and charges of these compounds, thereby affecting anti-HIV-1 activity (31). Semen possesses polyamines (57) that are positively charged at a neutral pH and can therefore nonspecifically interact with negatively charged polymers and abrogate their efficacy. A pH transition also occurs when the pH in the vaginal tract changes following the introduction of alkaline ejaculate during sexual intercourse. Notably, we found that PDB₁₄ is stable at a low pH, during transition to a neutral pH, and in the presence of both cervicovaginal secretions and seminal plasma.

We also examined the contraceptive potential of PDB₁₄. Microbicides that lack contraceptive activity are desirable because women may want protection from HIV-1 infection while retaining the option to conceive. Furthermore, men in many areas that are endemic for HIV-1 are more receptive to the idea of a noncontraceptive microbicide (58). To assess its contraceptive capacity, PDB₁₄ with and without a FAM moiety was tested. No differences in sperm motility were observed, demonstrating that PDB₁₄ does not appear to have contraceptive activity.

The lessons learned following the failures of early candidate microbicides, particularly N-9, have been invaluable in shaping preclinical strategies for demonstrating the safety of compounds under scrutiny as potential active pharmaceutical ingredients (APIs) in microbicides. Numerous in vitro experiments, studies performed in animal-based microbicide models, and clinical investigations involving potential microbicides and microbicide APIs have pointed to various potential determinants of topical vaginal microbicide safety. These include API concentration (9–11), application frequency (6), changes in epithelial integrity following topical application (10, 12, 13), subepithelial immune cell recruitment due to inflammation (10, 11, 14–16), and changes in the normal cervicovaginal microenvironment or barrier functions (see http://www.popcouncil.org/news/trial-shows-anti-hiv-microbicide-is-safe-but-does-not-prove-it-effective). In this study, we report that prolonged exposure to PDB compounds does not affect the viability of freshly isolated human PBMCs or vaginal epithelial cells, even at concentrations 1,000-fold higher than those required to inhibit infection by high-titer HIV-1. Additionally, PDB₁₄ had no effect on the growth of two vaginal Lactobacillus species, suggesting that these compounds will not adversely affect the normal flora of the female reproductive tract. Finally, by using a relevant in vivo model of cervicovaginal toxicity, we show that a formulated PDB microbicide gel does not cause any adverse changes in the integrity of genital tissue at the histological level. Therefore, PDB-based microbicides appear to be safe for intravaginal application. Ongoing and future studies of PDB safety will involve short- and long-term exposures to determine acute and chronic effects, the detection and release of inflammatory factors indicative of microbicide toxicity, the effect of topical formulation on epithelial tight junction integrity, and the incorporation of multiple-exposure protocols into our in vivo model for the purpose of investigating the contribution of application frequency to potential reductions in microbicide safety. Indeed, these studies may prove to be innovative for their roles in the development of a novel family of multifunctional antiviral molecules that are highly effective inhibitors of HIV-1 infection.

Given that by the very nature of its application, a successful microbicide will be applied repeatedly and for extensive periods, it is imperative to determine whether HIV-1 drug resistance can develop. Although drug resistance per se may not be a criterion for the rejection of a microbicide in preclinical development, as modification or combinations with other microbicides or antiretroviral agents may overcome this problem, viral evolution and the subsequent emergence of escape mutants are nevertheless concerning. A major potential advantage of PDB₁₄ over other ARV inhibitors is that its activity is sequence independent and not mediated by Watson-Crick base pairing, and it may therefore be less susceptible to mutation-generated drug resistance. It seems promising that PDB₁₄ possesses activity against a multidrug-resistant HIV-1 isolate at an EC₅₀ that is 80-fold lower than the highest reported nontoxic concentration of the compound. However, further investigations are required to determine whether long-term PDB₁₄ exposure will lead to resistance-engendering viral mutations.

Prior to screening microbicides in animals or launching clinical trials, it is important to perform testing in an adequate ex vivo organ culture system that not only preserves the native architecture of the tissue itself but also possesses the relevant cell types present within the cervicovaginal environment. We therefore evaluated PDB oligomers using a polarized cervical tissue matrix in a dual-chamber organ culture that recapitulates in vivo conditions (39). Regardless of oligomer size and the mechanism of action, the application of PDBs was found to potently and durably block HIV-1 transmission across cervical explants. These results further illustrate the potential success of using PDB-based compounds to prevent HIV-1 transmission.

In conclusion, abasic PDB molecules possess many properties of an ideal microbicide candidate, including strong antiviral efficacy and a mechanism of action that is similar to that of clinically successful topical prophylaxis agents (12). Additionally, by virtue of their sequence-independent anti-HIV-1 activities, the application of PDB microbicides repeatedly and for extensive periods may be less likely to generate resistance-engendering mutations in the virus. For these reasons, combined with excellent safety and stability profiles and their intrinsic anti-inflammatory properties, PDBs should be considered promising microbicide candidates for further preclinical development.