Cellulose acetate 1,2-benzenedicarboxylate protects against challenge with pathogenic X4 and R5 simian/human immunodeficiency virus

Objectives: To evaluate the protective efficacy of cellulose acetate 1,2-benzenedicarboxylate (CAP) formulated in a glycerol-based gel against infection with CXCR4 (X4) and CCR5 (R5) viruses in the simian/human immunodeficiency virus (SHIV)/rhesus macaque model of HIV-1 transmission.

Design: Mucosal infection of non-human primates is a reasonable model for use in the investigation of HIV-1 intervention strategies.

Methods: Rhesus macaques treated with Depo-Provera 5 weeks prior to challenge were inoculated intravaginally twice, over a period of 6 h with mixed inocula of pathogenic X4- and R5-SHIV in the presence or absence of CAP. Plasma viral load, peripheral and mucosal CD4 T cell counts as well as the genotype of the circulating virus were determined.

Results: CAP protected seven of ten macaques against transmission of both X4- and R5-SHIV, reaching statistically significant values (P = 0.0256). Delayed and/or reduced virus replication, as well as blunting of peripheral and mucosal CD4 T cell loss was noted in the three macaques that were infected in the CAP treated group compared to those in the placebo group. Further, protection conferred by CAP appeared to be more effective against X4- than R5-SHIV infection.

Conclusions: CAP is protective against highly permissive challenges with X4 and R5 viruses in vivo. Research on further development of this promising compound as a candidate microbicide for the prevention of sexual HIV-1 transmission is therefore warranted.

Introduction

Sexual transmission is the major mode of HIV-1 infection worldwide ^[1,2]. The lack of success in the development of an effective vaccine against HIV-1 coupled with the stifling costs and logistics for administering effective antiviral prophylactics have focused attention on research into new preventive approaches such as the use of chemical barriers ^[3,4]. Cellulose acetate 1,2-benzenedicarboxylate (CAP) is a promising candidate topical microbicide that prevents infection by a variety of sexually transmitted disease pathogens, including HIV-1 ^[5,6]. CAP inactivates and blocks the coreceptor binding site on HIV-1, and induces six helix bundle formation within the gp41 transmembrane envelope glycoprotein, exerting potent antiviral activity against cell-free and cell-associated infection of a variety of cell types and tissue explants by diverse viral strains/subtypes ^[7-11].

In this study, we assessed the activity of CAP in preventing HIV-1 transmission in vivo. The close genetic and biologic homologies of HIV-1 and SIV have firmly established infection of rhesus macaques with SIV as a valuable model for studying heterosexual transmission of HIV-1 ^[12]. In this regard, the effect of a CAP topical gel on vaginal transmission of SIV in rhesus macaques (RM) has been evaluated, and shown to confer 67% protection ^[13]. Nevertheless, there are differences in the structure and function of the SIV and HIV-1 envelope glycoproteins that could prove important in the dynamics of virus-host interactions that occur during sexual transmission. Importantly, whereas most SIV viruses utilize the CCR5 coreceptor for entry, HIV-1 strains can use CXCR4 in addition to or instead of CCR5 ^[14,15]. Although the majority of HIV-1 strains transmitted between humans use CCR5 (R5 viruses), anecdotal transmission cases have been reported where CXCR4 (X4)-tropic strains were successfully transmitted but eventually becoming a minor species ^[16-18]. Rare infections of individuals defective for the CCR5 gene with X4 HIV-1 strains have also been reported ^[19]. Thus, there is a need to test prevention of X4, as well as R5 virus transmission by candidate microbicides in relevant animal models to evaluate their antiviral activities.

We have used infection of macaques with pathogenic CXCR4- or CCR5-specific simian/human immunodeficiency virus (SHIV) to study HIV-1 transmission in vivo. Infection of macaques with X4-SHIV_SF33A is accompanied by a rapid and severe depletion of peripheral blood and lymph node CD4 T lymphocytes, whereas infection with R5-SHIV_SF162P3 causes a more protracted peripheral CD4 T cell loss but dramatic depletion in the gut ^[20-22]. In more recent studies, we found that both the X4- and R5-SHIV are transmitted simultaneously when introduced atraumatically onto the vaginal vault of rhesus macaques as mixed inocula ^[23], indicating that there is no intrinsic block to X4 virus transmission in the presence of the R5 strain. Infection with a mixture of X4-SHIV_SF33A and R5-SHIV_SF162P3 was therefore used to assess the protective effect of CAP in rhesus macaques.

Methods

Cells and virus production

The cell line, CEMX174 5.25M7 (kindly provided by N. Landau, Salk Institute, San Diego, California, USA) expresses CD4 and CXCR4, and was stably transduced with an HIV-1 long term repeat (LTR)-green fluorescent protein (GFP) and HIV-1 LTR-luciferase reporter construct, and with CCR5. The cell line is maintained in RPMI medium supplemented with 10% fetal calf serum (FCS), 200 μg/ml gentamycin (G418), 200 μg/ml hygromycin, 0.5 μg/ml puromycin, and 100 U/ml each of penicillin and streptomycin. Cell free virus stocks of X4-SHIV_SF33A and R5-SHIV_SF162P3 were propagated and titrated in the 5.25M7 cell line. The titer of both viruses used in this study is 5.9 × 10³ 50% tissue culture infectious doses (TCID₅₀)/ml.

CAP formulation

Micronized CAP (Aquateric CD910 containing approximately 67% CAP, FMC Corporation, Philadelphia, Pennsylvania, USA) in a formulation with hydroxypropylmethyl cellulose E4M, Avicel PH105, Innovatol PD60, dimethicone (DC360) and glycerin (99%) was prepared, produced and analyzed by Dow Pharmaceutical Sciences (Petaluma, California, USA). The concentration of aquateric in the gel was 18% (w/w), corresponding to 13% CAP. Placebo gel lacks the active ingredient.

In vitro inhibition assays

For inhibition studies, serial dilutions of a soluble form of CAP (Eastman Chemical Company, Kingsport, Tennessee, USA), supplied as a 30 mg/ml stock solution in 300 mM sodium acetate buffer (pH 5.8) by the New York Blood Center, were incubated with equal volumes (50 μl) of each virus (100 TCID₅₀) for 1 h at 37°C. CEMX174 5.25M7 cells (2 × 10⁴ in 100 μl media) were then added to the virus-compound mixtures and cultured for 3 days at 37°C. Control cultures received virus incubated in the absence of compound. At the end of the incubation period, cells were harvested, lysed, and processed according to the manufacturer's instructions (Promega, Madison, Wisconsin, USA). Luciferase activity associated with the cell lysate was detected with a Dynex MLX microtiter plate luminometer (Dynex Technologies, Inc., Chantilly, Virginia, USA).

Animal infections and specimens

All infections were carried out in adult female rhesus macaques (Macaca mulatta) individually housed at the Tulane National Primate Research Center in compliance with the Guide for the Care and Use of Laboratory Animals. Animals were confirmed to be serologically negative for simian type D retrovirus, SIV and simian T cell lymphotropic virus prior to infection. For all procedures, macaques were sedated with ketamine-HCL (10 mg/kg) and intravaginal (IVAG) inoculations were performed by placing the virus inoculum atraumatically on the cervicovaginal mucosa. For protection studies, CAP or placebo gel was applied using a 5 ml syringe. The animals were kept with their pelvis elevated for 20 min after virus challenge. Whole blood was collected at designated time intervals postchallenge. Plasma virus was quantified by branch DNA analysis (Bayer Diagnostics, Emeryville, California, USA) and T cell subsets (CD3, CD4, and CD8 lymphocytes) were determined by Trucount according to the manufacturer's instructions (Becton Dickinson, San Jose, California, USA). Bronchial aspirations were performed to obtain lung lavages. Cells were washed with phosphate buffered saline (PBS) and stained with monoclonal antibodies for CD3, CD4 and CD8 followed by flow cytometry using a FACSCalibur (Becton Dickenson Immunocytometry Systems). Analysis was performed using CELLQuest software. The monoclonal antibodies used in this study included antimonkey CD3-FITC (clone FN-18, Biosource), antihuman CD4-PE (clone L200, BD Biosciences) and antihuman CD8-PerCP (clone SK1, BD Biosciences).

Virus genotype

The relative proportion of replicating X4- and R5-SHIVs in the plasma of co-infected macaques was determined by a real-time PCR that distinguishes the two viruses. Briefly, viral RNA was extracted from 500 μl clarified plasma using commercial kits (Qiagen, Valencia, California, USA), reversed transcribed using random hexamer primers and Moloney murine leukemia virus reverse transcriptase (Invitrogen, Carlsbad, California, USA), and cDNAs amplified in a 50 μl reaction mixture with the conserved Env-specific primers ED5 and ED12 ^[24]. Three 10-fold dilutions of the Env PCR products were then subjected to separate X4- and R5-specific real-time reactions in 25 μl volumes containing the following: 1 × Invitrogen PCR buffer, 5 mM MgCl₂, 0.1 × SYBR Green (Molecular Probes, Eugene, Oregon, USA), 10 mM dNTPs, 120 nM of each primer (from either the X4- or R5-specific sets) and 1.25 U of Platinum Tag (Invitrogen). Standard curves were generated in parallel with serial 10-fold dilutions of a known quantity of the X4 and R5 reference strain plasmids. Oligonucleotide primers for the separate quantitation of the two viruses were chosen within the C3 (5′) and V5 (3′) regions of the envelope glycoprotein gp120 and were shown to be specific and sensitive, detecting as few as one copy of a plasmid containing the envelope glycoprotein of the target viral strain in a background of 1 × 10⁵ copies of the opposite strain in spiking experiments (data not shown). The sequences for the X4-SHIV_SF33A and for the R5-SHIV_SF162P3 primers are: 5′-GAACAAGTAGCTACAAAATTAAG, 3′-TCTCGGTGGTACTGTTCCCGTTA; 5′-AAACAGATAGTTACAAAACTACA, 3′-TCTCGGTGGTGTTACTGACCTCTC, respectively. PCR was performed on the ABI 7700 PRISM spectroflurometric thermal cycler using the following conditions: one cycle of denaturation (95°C for 10 min) and 50 cycles of amplification (95°C for 10 s, 55°C for 30 s, 72°C for 40 s). Copy numbers of X4-SHV_SF33A and R5-SHIV_SF162P3 sequences in plasma samples were calculated from the standard curves and expressed as relative percentages.

Detection of antiviral humoral responses

SHIV-specific antibody responses were analyzed by strip immunoblot assay according to the manufacturer's instructions (Chiron, Emeryville, California, USA). This assay uses recombinant viral antigens derived from HIV-1_SF2 p24 Gag, gp120 and gp41, p31 from the endonuclease portion of Pol, and p27 Gag from HIV-2_UC1. Cross-reactivity of anti-SIV and anti-HIV-1 antibodies in plasma with viral antigens allowed detection of the antiviral humoral response. Each independent assay was carried out with internal positive and negative controls.

Results

CAP is active against X4-SHIV_SF33A and R5-SHIV_SF162P3 in vitro

The inhibitory effect of soluble CAP on infection with SHIV_SF33A and SHIV_SF162P3 was first assessed in CEMX174 5.25M7 cells. We found that CAP inhibited infection by both viruses, with a 50% inhibition dose (ID₅₀) of 25 and 180 μg/ml against R5-SHIV_SF162P3 and X4-SHIV_SF33A, respectively (data not shown).

Progesterone treatment enhances vaginal transmission of both X4-SHIV_SF33A and R5-SHIV_SF162P3 in rhesus macaques

We previously reported a 50-70% intravaginal transmission efficiency of X4-SHIV_SF33A and R5-SHIV_SF162P3 when inoculated singly in rhesus macaques ^[23]. The challenge dose used (approximately 5000 TCID₅₀), however, differs from the dose of HIV-1 in most sexual transmissions which is believed to be extremely small (based on the low probability of transmission per act) ^[25-27], although this notion has recently been challenged ^[28-30]. Modulation of the thickness of the vaginal epithelium by administration of medroxyprogesterone had been suggested to enhance mucosal SIV transmission ^[31,32], but a role of Depo-Provera in compromising the antiviral immune responses cannot be excluded ^[33]. In an attempt to increase the transmission efficiency and at the same time, lower the challenge dose of SHIV_SF33A and SHIV_SF162P3 for preclinical evaluation of CAP, we treated animals with a single intramuscular dose of 30 mg medroxyprogesterone acetate (Depo-Provera; Pharmacia & Upjohn, Kalamazoo, Michigan, USA) and challenged 5 weeks later with varying amounts of the two viruses in mixtures to determine the dose required to achieve 100% vaginal transmission.

Three macaques each were inoculated intravaginally once with a mixed inoculum of the two SHIVs containing 1250, 250 and 50 TCID₅₀ each of X4-SHIV_SF33A and R5-SHIV_SF162P3. Results showed that all animals were infected (Fig. 1), with peak viremia of 1 × 10⁶-1 × 10⁸ RNA copies/ml plasma at 2-3 weeks postinoculation (wpi) that were comparable regardless of the inoculum dose. Acute CD4 T cell loss, indicative of the presence of the X4 virus, was noted in six of the nine infected animals, with the extent of this loss correlating with the level of virus replication. Viral load and CD4 T cell loss were sustained in all three animals that received the highest challenge dose (2500 TCID₅₀ total) (Fig. 2a), with two of the animals (P460 and R902) succumbing to simian AIDS at 24 and 32 wpi. Virus replication was controlled in two of the three animals in the 500 and 100 TCID₅₀ challenge groups (Fig. 1b and c), with animals in both groups still alive at 42 wpi. An inverse correlation between viral load and the degree of sustained peripheral CD4 T cell loss was seen in these animals that were challenged with lower virus dose. Genotypic analyses of the circulating virus at first samplings of blood (2 wpi) revealed the presence of R5-SHIV_SF162P3 in all macaques, with X4-SHIV_SF33A present in only six of the nine infected animals (data not shown). At later samplings, however, both viruses can be detected in most animals. Based on the results of this titration study, a 300 TCID₅₀ mixed inoculum of the two viruses (150 TCID₅₀ each) was chosen as the challenge dose in Depo-Provera treated macaques to assess the efficacy of CAP.

Fig. 1

Fig. 2

CAP protects against challenge with X4-SHI_VSF33A and R5-SHIV_SF162P3

Fifteen animals, all treated with Depo-Provera (30 mg, intrasmuscular) 5 weeks before challenge, were used. The animals were challenged twice, once in the morning and then 6 h later. Ten animals received 3 ml of a CAP gel formulation and the remaining five received 3 ml of a placebo gel 20 min before each challenge. Results are shown in Fig. 2.

All five animals that received the placebo gel were infected, with peak viremia of 1 × 10⁶- × 10⁸ RNA copies/ml plasma, and showing varying degrees of peripheral CD4 T cell loss. Three of the 10 animals (BD72, BI44, BI83) that received CAP gel were also infected. Nevertheless, average peak viremia was lower for infected animals in the CAP group compared to the placebo group. Furthermore, two of the CAP infected animals (BD72, BI44) showed delayed virus replication, and one (BI44) replicated to 2-3 log lower titers. Virus replication (1 × 10³-1 × 10⁶ RNA copies/ml plasma) was sustained in three of the five infected animals in the placebo group (BE97, BH33, T185), but in only one of the three macaques in the CAP treated group (BD72).

The lung is a repeatedly accessible mucosal effector site that provides evidence of CD4 T cell depletion in SIV infected macaques when total CD4 T cell counts in blood show little change ^[34]. Indeed, at 3 wpi, we found significant loss of CD4 T cells in the lung lavages of four of the five infected animals in the placebo group despite the fact that only one of them, T185, showed dramatic CD4 T cell loss in peripheral blood. The 5.2-14.8% CD4 T cells seen in the lung of infected animals in the placebo group were lower than the average 30% for control uninfected animals, and in the two uninfected macaques in the CAP treated group (M770, BD52) that were analyzed in parallel (Fig. 3). At 16 wpi, CD4 T cell loss in the lung was sustained for macaques BE97, N966 and T185, with loss in the periphery now also detected in BE97. In contrast, none of the three infected animals in the CAP treated group showed CD4 T cell depletion in the periphery at 16 wpi, but BD72, the macaque with the highest viral load, did suffer loss in the lung (8.9%).

Fig. 3

Genotyping analyses of replicating virus at first sampling of plasma (2 wpi) again demonstrated the presence of R5-SHIV_SF162P3 in all infected animals, with the X4 virus detectable only in two of the five infected animals in the placebo group (T185 and N966) (Fig. 4). The greater proportion of circulating X4 virus in T185 at the peak of virus replication (2 wpi) explained for the precipitous drop in CD4 T cell counts seen in this macaque (Fig. 2). Analyses of later plasma samples (up to 16 wpi), however, revealed the presence of X4-SHIV_SF33A at varying proportions in all but one (BF52) of the infected animals in the placebo group, whereas none of the CAP treated animals that were infected harbored detectable X4-SHIV_SF33A during the same sampling period.

Fig. 4

Virus-specific antibodies developed by 1-2 months after challenge in the three CAP treated macaques that became infected as well as in four of the five macaques that received placebo gel (data not shown). The exception was BE97 in the placebo group. This animal had the highest acute viremia and set-point, features that are characteristics of rapid progressors which do not usually seroconvert, and was sacrificed with clinical symptoms of simian AIDS (SAIDS) at 33 wpi. Our findings are summarized in Table 1, showing that CAP protected seven of ten macaques from challenge with both pathogenic X4- and R5-SHIVs, with a P = 0.0256 (Fisher's exact test).

Table 1

Discussion

In this study we showed that CAP protected seven of 10 macaques from challenge with a mixed inoculum of pathogenic X4- and R5-SHIVs, reaching statistically significant values (P = 0.0256). All five placebo gel treated control animals were infected, with high levels of virus replication detected within 1-2 weeks after intravaginal exposure. In contrast, delayed and/or blunting of virus replication were noted in the three macaques that were infected in the CAP treated group, suggestive of partial protection. Further, protection conferred by CAP appeared to be more effective against X4-SHIV_SF33A than R5-SHIV_SF162P3 infection.

The efficacy of most candidate microbicides scheduled for testing in human clinical trials had been evaluated in macaque models using X4-tropic SHIV ^[35-37]. Given that most sexual transmission in humans initiates with R5-tropic strains, but that X4 virus can be transmitted as well ^[15], we use co-infection of macaques with pathogenic X4- and R5-SHIV to assess the relative protection conferred by CAP against X4- and R5- HIV-1 infection in vivo. Our data showed that X4-SHIV_SF33A could be detected in the plasma of four of five infected animals that were treated with placebo gel, but in none of the three infected animals in the CAP treated group up to 16 wpi (Fig. 4). The apparent greater suppressive effect of CAP against X4-SHIV_SF33A than R5-SHIV_SF162P3 infection in vivo contrasts with the greater resistance of the X4 virus infection to CAP inhibition in vitro. The basis for this difference is unclear, but could be due to the presence of selective blocks to X4-SHIV_SF33A transmission or replication and dissemination in vivo. Further studies are required to examine these various possibilities.

In the SIV macaque model, a 1-ml glycerol-based CAP gel applied 5 min prior to each of two high dose (1 × 10⁵ TCID₅₀) administrations of SIVmac251 protected four of six rhesus monkeys from vaginal infection ^[13]. Additionally, 2 ml CAP administered 15 min prior to virus challenge was effective around 92% of the time against low-dose vaginal exposure of 10 TCID₅₀ R5-SHIV_SF162P3 in pigtail macaques ^[38]. Since Depo-Provera enhances SIV/SHIV infection, our challenge design in which the animals were treated with the injectable contraceptive and challenged twice, even with low dose inoculum, represents a highly permissive transmission model. Yet the degree of protection reported in all these studies that are performed under very different challenge conditions and with different viruses is comparable, illustrating that CAP is responsible for the inhibition of infection seen. Furthermore, the results suggest that protection may not be related to the amount of product used, at least within the ranges (1-3 ml CAP) tested. Collectively, these findings further demonstrate the potency of CAP as a candidate topical microbicide for HIV-1 and indicate that the efficacy a particular vaginal product can be tested using different macaque challenge models with comparable results.

Microbicidal compounds that specifically inactivate infectious virus have only a limited time frame within which to act before the virus is bound to susceptible cells, while those that block HIV-1 attachment/fusion pathways may become diminished during the inevitable dilution over time. In this regard, CAP is an ideal microbicide in that it functions as a dual acting microbicide, blocking HIV-1 attachment/fusion at low concentrations and inactivating both X4 and R5 isolates of distinct clades at high concentrations in vitro ^[7,8]. We now show that CAP protected against vaginal transmission of pathogenic X4 and R5 viruses under highly permissive challenge conditions. Further testing and development of this promising compound as a topical microbicide for the prevention of HIV-1 transmission are therefore warranted.

Acknowledgements

This work was supported by a grant from the NIH (PO1 HD41761). We thank Nurjehan Jivani and Gordon J. Dow at Dow Pharmaceutical Sciences for the preparation and provision of CAP and placebo gel formulations used in this study, Elena Serbinova (Dow Pharmaceutical Sciences) for regulatory advice, and Wendy Chen for help with the graphics.

Sponsorship: Funded by NIH PO1 HD41761.