Direct virucidal activity was assessed by compound treatment of immobilised virus, prior to washing and culture with permissive T cells as previously described 22. Both compounds demonstrated potent activity against the X4 isolate, with 50% inhibitory concentrations (IC₅₀) observed at 14.8 (± 1.9) and 9.3 (± 2.1) μg/ml of PRO 2000 and DxS respectively (Figure 1Ai &1Bi). In contrast, both compounds failed to exert any effect against R5 virus, even at concentrations of 1 mg/ml (Figure 1Ai &1Bi). Receptor mediated blockade was assessed by incubating target cells with compound prior to compound removal and culture with immobilised virus; this was poor or absent for both compounds (Figure 1Aii &1Bii). Inhibition of attachment/fusion was assessed by pre-treatment of virus with test compound for 1 hour prior to culture with permissive cells in the presence of compound. Both compounds exhibited potent activity against R5 and X4 infection, although greater activity was observed against X4 than R5 virus with IC₅₀ values of 1.9 (± 1.6) and 20.8 (± 1.5) μg/ml respectively for PRO 2000, and 0.38 (± 1.9) and 782.8 (± 2.4) μg/ml respectively for DxS (Figure 1Aiii &1Biii).

Before the activity of compounds against HIV-1 infection of female genital tissue was investigated, it was important to ensure that neither compound would elicit a toxic effect. This was evaluated using genital mucosal tissue explants obtained from seronegative women undergoing therapeutic hysterectomy as previously described 1221. Tissue explants were immersed in test compound for 2 or 24 hours and tissue viability determined using the principle of MTT dye reduction (see Methods). Compounds were tested to a maximal concentration of 1 mg/ml and toxicity was compared to the known toxic agent Nonoxynol-9 (N9) 23. Only mild toxic effects were observed with both PRO 2000 and DxS following 2 hour compound treatment, with 50% toxic doses (TD₅₀) of greater than 1 mg/ml for both compounds (Figure 2A and 2B). This was in contrast to N9, which caused significant toxicity with a TD₅₀ of 700 (± 2) μg/ml following a 2 hour treatment period (Figure 2C). In fact, N9 caused significant toxicity at 1 mg/ml, causing a 65% reduction in viability. Furthermore, 24 hour treatment of tissue with N9 caused significant damage (TD₅₀ = 34 ± 1 μg/ml), whilst only mild toxicity was observed following 24 h treatment with either PRO 2000 or DxS (TD₅₀ > 1 mg/ml).

The potential of PRO 2000 and DxS to block infection of the female genital mucosa was investigated using ectocervical explants, cultured in a non-polarised manner as previously described 2122. Explants were treated with test compound (PRO 2000 or DxS) for 1 hour prior to exposure to R5 HIV-1_BaL for 2 hours in the presence of compound as described in the Methods. Viral infection was evaluated by p24 released into culture supernatants. The activity of polyanions against HIV-1_BaL infection of cervical explants was dose-dependent (Figure 3). Both PRO 2000 and DxS were able to completely inhibit infection at 1 mg/ml (p < 0.001), but allowed breakthrough of infection to occur at 100 μg/ml, with DxS being 10 fold better than PRO 2000 with an IC₅₀ of 6.9 (± 1.6) versus 79.5 (± 3.7) μg/ml (Figure 3i).

We have previously shown spontaneous migration of CD4⁺ dendritic cells (DC) from cervical explant tissue during overnight culture, a population of cells able to bind virus via mannose C-type lectin receptors (MCLR) and/or CD4 21. Migratory cells were harvested from explant cultures (exposed to compound and virus as described) following overnight culture, washed to eliminate cell free virus, and co-cultured with permissive PM-1 T cells. The effect of both compounds in preventing dissemination of virus by these migratory cells was dose-dependent. PRO 2000 completely inhibited viral transfer at 1 mg/ml, and demonstrated significant inhibition (>90%) at 100 μg/ml, with an IC₅₀ of 29.1 (± 2.5) μg/ml (Figure 3Aii). DxS provided 95% protection at 1 mg/ml (Figure 3Bii) demonstrating an IC₅₀ of 62.4 (± 2.9) μg/ml.

Having observed that both compounds showed some efficacy against dissemination of HIV-1 by migratory cells, subsequent experiments were carried out to determine whether either compound blocked DC-SIGN binding and/or transfer. To this end, Raji-DC-SIGN⁺ CD4^- cells were incubated with candidate polyanions during exposure to virus (2 h). Excess virus and compound were removed by washing and cells either directly lysed to determine the amount of virus bound to cell surface receptors, or cultured with permissive T cells (PM-1) to assess trans infection. Mannan, the natural ligand for DC-SIGN and other MCLR, blocked most, but not all, binding of virus to Raji DC-SIGN⁺ cells. Viral binding to Raji DC-SIGN⁺ cells in the presence of mannan (100 μg/ml) mirrored values seen with Raji DC-SIGN^- cells (Figure 4), indicating a low level (20% of untreated controls) of DC-SIGN-independent binding of virus to Raji cells. PRO 2000 exhibited significant activity at 0.25 mg/ml against virus binding to DC-SIGN and trans infection of PM-1 cells (Figure 4A). DxS exhibited a lower level of inhibition, demonstrating a maximal 50% inhibition of both binding and trans infection at the highest concentration of 2.5 mg/ml (Figure 4B) while demonstrating no statistically significant effect at lower concentrations when compared to untreated controls (taken as 100%).

To investigate whether exposure of human cervical tissue to candidate polyanions would elicit an inflammatory response, tissue explants were exposed to compound (2 h) prior to compound removal by washing and overnight culture. Culture supernatant was assessed by Bioluminex assay for the presence of a panel of 9 cytokines (IL-1β, IL-6, IL-8, TNF-α, GM-CSF, MIP-1α, MIP-1β, RANTES, and MCP-1). Untreated tissue explants produced detectable levels of all cytokines except TNF-α and RANTES, which were towards the limits of detection. Treatment with either compound (1 mg/ml) had little or no effect on the production of most of the cytokines including IL-6, IL-8, GM-CSF, and MIP-1α (data not shown). However, treatment with 1 mg/ml of either PRO 2000 or DxS resulted in a 13 or 6 fold (respectively) increase in IL-1β release (Figure 5i), which was statistically significant (p = 0.006) for PRO 2000. Both compounds also induced increases in TNF-α and RANTES production (Figure 5ii and 5iii) although only the increase in RANTES induced by PRO 2000 reached statistical significance (p = 0.002). To aid the interpretation of this data, results were compared with explants treated with an equal dose of the toxic compound N9. Unfortunately, N9 caused significant (≥ 50%) toxicity to tissue at concentrations of ≥ 100 μg/ml. Although approximately 50% viability was still observed at 100 μg/ml, the effect such toxicity had on cytokine release could not be determined with complete confidence, therefore only concentrations causing no toxicity were used for comparison. In general, treatment of tissue with 10 μg/ml N9 caused little change in cytokine release. To determine whether there was any correlation between increasing compound dose and release of cytokines, data was analysed using Spearman rank correlation and significance determined using two-tailed significance testing of paired samples. However, none of the compounds demonstrated any significant correlation between increasing compound dose and modulation of cytokine release, suggesting the observed cytokine release was unlikely to reflect adverse response to compound treatment.

Due to the strong correlation reported between the presence of genital herpes and HIV-1 transmission 24, the effect of both PRO 2000 and DxS on the ability of HSV-2 to infect vaginal epithelial cells was investigated using the ME180 cell line. ME180 cells were exposed to HSV-2 (1 hour) in the presence of test compound and, following compound removal, cells were cultured for 48 hours in the absence of compound and virus, and viability determined by the principle of MTT dye reduction (see Methods). PRO 2000 and DxS demonstrated no significant toxicity towards ME180 cells, and both compounds demonstrated potent anti-HSV-2 activity, with IC₅₀ values of 11.5 (± 1.4) μg/ml (PRO 2000) and 5.2 (± 1.4) μg/ml (DxS) (Figure 6).

PM-1 (AIDS reagent project, National Institute for Biological Standards and Control, Potters Bar (NIBSC), UK), Raji, Raji/DC-SIGN (provided by V N Kewal-Ramani, HIV Drug Resistance Program, NCI, Frederick, MD) and Vero cells were grown in complete RPMI [RPMI 1640 medium supplemented with 10% fetal calf serum, 100 U/ml penicillin, 100 μg/ml streptomycin and 2 mM L-glutamine]). The adherent cell line ME180 was cultured in DMEM supplemented as complete RPMI (complete DMEM). All cells were grown in continual culture in a humidified environment of 5% CO₂ at 37°C and passaged every 3–4 days.

HIV-1 strains (HIV-1_BaL and HIV-1_RF, AIDS reagent project, NIBSC, UK) were grown in phytohaemagglutinin (PHA)-stimulated peripheral blood mononuclear cells as previously described 12. Cell-free viral stocks were passed through 0.2 μm pore-size filters. Infection was monitored by viral p24 antigen (HIV-1 p24 ELISA, AIDS Vaccine Program, National Cancer Institute (NCI) at Frederick, MD, USA), carried out according to manufacturers protocol) or reverse transcriptase (RT) 42 release into culture supernatants. The 50% tissue culture infectious dose (TCID₅₀) was determined in PM-1 cells for both viruses, and additionally in PHA-stimulated PBMC for HIV-1_BaL.

HSV-2 (G) (kindly donated by Dr. B. Herold (Mount Sinai School of Medicine, NY, USA)) was grown in Vero cells. Infectivity of viral stocks was assessed by plaque assay using ME180 cells as previously described 43.

Unformulated PRO 2000 was provided by Indevus Pharmaceuticals, USA, and DxS by ML Laboratories, UK. Both products were used at non-toxic concentrations as determined by MTT viability assays.

Solid phase immobilisation of HIV was carried out as previously described 22. In brief, HLA-DR Mab (L243, ATCC) was bound to 96 well, flat bottom, tissue culture plates (Nunc) for 1 hour at room temperature. Unbound antibody was washed off with 1 volume PBS prior to the addition of virus (RF or BaL, 10³ tissue culture infectious doses [TCID₅₀] as determined in PM-1 cells). Plates were centrifuged for a minimum of 1 hour (room temperature) at 3200 rpm. Unbound virus was washed away with 2 volumes of PBS. Direct virucidal activity was determined by compound pre-treatment of immobilised virus for 1 hour before culture with target cells (PM-1 cells, 4 × 10⁴ cells/well) in the absence of compound (compound was removed with 4 PBS washes). Receptor mediated blockade activity was determined by the pre-treatment of target cells (1 hour) prior to exposure to immobilised virus in the absence of compound (where compound was removed from treated cells by 4 PBS washes). Attachment/fusion inhibition was determined by the pre-treatment of immobilised virus with test compound prior to the addition of target cells in the presence of compound. Plates were cultured for 10 days, in the absence of media (or compound) replenishment, when viral replication was determined by measurement of RT in culture supernatants. The described assay allows topical administration of candidate compounds: previous studies have demonstrated no difference in compound activity against virus that is either in suspension of immobilised onto plastic (data not shown).

To determine whether compounds blocked either virus binding and/or transfer via DC-SIGN, CD4^--DC-SIGN⁺ or CD4^--DC-SIGN^- Raji cells (0.5 × 10⁴ cells/well) were treated with test compound for 1 hour at 37°C prior to exposure to virus (HIV-1_RF or HIV-1_BaL, 10⁴ TCID₅₀ determined in PM-1 cells) for 2 hours at 37°C in the presence of compound. Compound and unbound virus were removed by washing (4 volumes PBS) and cells either: i) lysed in 1% Triton X-100 to determine the level of virus bound to the cell surface (p24 ELISA); or ii) co-cultured with permissive T cells (PM-1 cells, 4 × 10⁴ cells/well) to evaluate trans infection. Co-cultures were assessed for viral replication by measurement of reverse transcriptase activity following 7 days in culture.

Cervical explant culture was performed as previously described 122122. Cervical tissue was obtained from women undergoing planned therapeutic hysterectomy (with written consent as per approval from the local Research Ethics Committee). Cervical tissue comprising both epithelium and stromal tissue was cut into 3 mm explants prior to culture submerged in RPMI 10%. Briefly, explants were pre-treated for 1 hour with test compound prior to exposure to HIV-1_BaL (10³ – 10⁵ TCID₅₀ determined in PHA-activated PBMC) for 2 hours at 37°C. After incubation with infectious virus and compound, explants were washed with 4 volumes of PBS. Explants were then cultured overnight prior to transfer to fresh plates and further culture for 12–14 days, with 50% media feeds every 2–3 days. Migratory cells present in the overnight culture plate were washed with 2 volumes of PBS and then co-cultured with 4 × 10⁴ PM-1 cells/well to assess blockade of virus transfer by migrating cells. At the end of the assay, HIV-1 infection was determined by the measurement of p24 in culture supernatants (ELISA): supernatants from explant cultures were assessed using Beckman Coulter p24 ELISA (lower detection limit of 15 pg/ml); supernatants from migratory cell co-cultures were analysed with the less sensitive p24 ELISA from NCI (lower detection limit 300 pg/ml).

Viability of cells and tissue was determined following compound treatment by the principle of MTT (3 [4,5-dimethylthiazol-2-yl]-2,5 dipbenyltetrazolium bromide or thiazolyl blue) dye reduction.

Cytokine production was determined by multiplex bead immunoassays (Biosource International Inc., UK) as per manufacturers instructions. Tissue explants were exposed to compound for 2 hours prior to compound removal by washing and overnight culture in the absence of compound. Culture supernatant (50 μl) was assessed for the presence of a panel of 10 cytokines (IL-1β, IL-6, IL-8, TNF-α, GM-CSF, MIP-1α, MIP-1β, RANTES, and MCP-1). Lower limits of detection for each cytokine were generally: IL-1β (7 pg/ml), IL-6 (8 pg/ml), IL-8 (8 pg/ml), TNF-α (6 pg/ml), GM-CSF (16 pg/ml), MIP-1α (15 pg/ml), MIP-1β (19 pg/ml), RANTES (23 pg/ml) and MCP-1 (30 pg/ml). Spiked control samples demonstrated that culture conditions and any residual compound did not interfere with assay sensitivity (data not shown). Plates were read using the Luminex 100 system (Luminex Corp., USA) and data analyzed using Bioplex Manager version 4.0 software (Biorad, UK). Cytokine concentrations present in culture supernatants were determined using non-linear regression analysis.

ME180 cells (1.5 × 10⁴ cells/well) were seeded in 96-well plates and cultured overnight. Cells were exposed to test compound alone (to determine compound toxicity), or virus (approximately 5 × 10⁴ pfu/well) in the presence of compound (to determine inhibitory effects of the compound) for 1 hour. Compound and unbound virus was removed by washing (3 × 200 μl PBS) and cells cultured in fresh media for 48 hours. Viability was then determined by MTT assay. Whilst a decrease in cell viability in wells exposed to virus reflects viral replication, a reduction in viability following exposure to compound alone indicates toxicity. Viability and infectivity values were calculated as percentage of viability from cells exposed to culture medium alone or percentage of infectivity from cells exposed to virus in the absence of compound.

50% inhibitory concentration analysis was determined using non-linear regression analysis, whilst correlation coefficients were calculated by non-parametric correlation (Spearman) and two-tailed p-value calculation (GraphPad PRISM, GraphPad Software, Inc.). Student's T-tests were performed in Excel (Microsoft Corporation).