|Home | About | Journals | Submit | Contact Us | Français|
The potent nonnucleoside reverse transcriptase inhibitor UC781 has been safety tested as a vaginal microbicide gel formulation for prevention of HIV-1 sexual transmission. To investigate whether UC781 retained anti-infective activity following exposure to the female genital tract, we conducted an ex vivo analysis of the UC781 levels and antiviral activity in cervicovaginal lavage (CVL) fluids from 25 Thai women enrolled in a 14-day safety trial of twice-daily vaginal application of two concentrations of the UC781 microbicide gel. CVL samples were collected from women in the 0.1% (n = 5), 0.25% (n = 15), and placebo (n = 5) gel arms following the first application of gel (T15 min) and 8 to 24 h after the final application (T8-24 h) and separated into cell-free (CVL-s) and pelletable (CVL-p) fractions. As UC781 is highly hydrophobic, there were significantly higher levels of UC781 in the CVL-p samples than in the CVL-s samples for the UC781 gel arms. In T8-24 h CVL-p samples, 2/5 and 13/15 samples collected from the 0.1% and 0.25% UC781 gel arms, respectively, efficiently blocked infection with ≥4 log10 50% tissue culture infective dose (TCID50) of a CCR5-tropic CRF01_AE HIV-1 virus stock. Independent of the arm, the 11 CVL-p samples with UC781 levels of ≥5 μg/CVL sample reduced infectious HIV by ≥4 log10 TCID50. Our results suggest that the levels and anti-infective activities of UC781 gel formulations are likely to be associated with a cellular or pelletable component in CVL samples. Therefore, cellular and pelletable fractions should be assayed for drug levels and anti-infective activity in preclinical studies of candidate microbicides.
The development and testing of vaginally applied microbicides to prevent HIV infection during sexual intercourse have produced both disappointing and promising results (10, 18, 21, 22, 28). Initial nonspecific microbicides, such as Carraguard, SAVVY, and PRO2000, showed no efficacy for preventing male-to-female sexual transmission of HIV in clinical trials, while the spermicide Nonoxynol-9 was reported to increase susceptibility to HIV infection (13, 17, 21, 26, 28). Promising results of the CAPRISA-004 trial demonstrated that a vaginal microbicide gel containing the reverse transcriptase inhibitor tenofovir was effective in preventing HIV infection in some populations (1). In contrast, lack of efficacy in the more recent VOICE trial led to a halt in the tenofovir gel arm of that study (24). The preclinical testing outcomes for all of these products suggested that they might significantly reduce sexual transmission of HIV when used appropriately in a clinical trial. However, a recent report using ex vivo testing of PRO2000 claimed that this product may lose some of its anti-infective activity after vaginal application and sexual intercourse (16). Except for PRO2000, it is not known whether exposure of microbicide gels to female genital secretions in vivo lowers their anti-HIV activity and may have contributed to the disappointing outcomes in some clinical efficacy trials.
UC781 is an extremely hydrophobic nonnucleoside reverse transcriptase inhibitor (NNRTI) developed as a potent antiretroviral drug with a high resistance threshold, requiring the accumulation of multiple drug-resistance mutations with fitness costs to become ineffective (9). UC781 inhibits in vitro infection by multiple HIV-1 subtypes at concentrations of <10 nM (3, 4, 8, 12, 19, 29) and retains its anti-HIV activity for 4 days in cell culture (6). However, UC781 has low oral bioavailability (10). UC781 formulated as a drug suspension in Carbopol gel was tested as a candidate vaginal microbicide. Phase I safety testing showed it was safe for vaginal application at concentrations of less than 1% (25). Further preclinical testing found that UC781 was detectable in the genital fluids of female pig-tailed macaques for up to 6 h after vaginal application (20). In comparison, the time from application to detection for UC781 in macaque vaginal fluids was much shorter than that reported for its anti-HIV activity in cell culture (6). Due to the hydrophobicity of UC781, it is possible that its residence time in the genital mucosa of female macaques may have been underestimated by sampling only their genital fluid supernatants.
We investigated drug levels in the genital fluid supernatant and pelleted CVL samples collected from women enrolled in a phase I study of the safety and toxicity of vaginally applied UC781 gel formulations conducted in Thailand. In addition, we assessed the anti-infective capacity of the UC781-containing fractions to block virus infection ex vivo using an HIV titer reduction assay. Our results indicate the need to examine both cell-free and insoluble fractions of female genital samples to better estimate retention times of mucosal drug and anti-infective levels.
Forty-five HIV-uninfected Thai women were enrolled in a randomized, placebo-controlled, double-blind phase I study of the safety and acceptability of UC781 topical vaginal microbicide in Chiang Rai, Thailand. HIV-1-uninfected women (ages 18 to 50) with a normal Pap smear, regular menses, no evidence of reproductive tract infection at screening, and in a monogamous sexually active relationship were eligible for enrollment. Participants were randomized in a 1:1:1 ratio to one of 3 study arms for twice-daily application of one of the following gel formulations: (i) a Carbopol gel formulation of 0.1% UC781 (0.1% UC781), (ii) a Carbopol gel formulation of 0.25% UC781 (0.25% UC781), or (iii) hydroxyethyl cellulose (HEC) placebo gel (placebo). UC781 gel formulations were provided by CONRAD (Arlington, VA). UC781 and placebo gels were packaged in overwrapped single-dose (3.5 g, approximately 3.5 ml) prefilled gel applicators. Informed consent was obtained from all study participants, and the study was approved and reviewed by the Centers for Disease Control and Prevention (CDC) Institutional Review Board, the Thai Ministry of Public Health (MOPH) Ethical Review Committee, and the Chiang Rai Hospital Internal Ethical Committee for Research in Human Subjects (IEC).
Participants visited the clinic on the day of enrollment, with follow-up visits on days 7 and 14. Cervicovaginal lavage (CVL) samples were obtained by introducing 5 ml of sterile phosphate-buffered saline (PBS, pH 7.2) into the vagina and collecting the pooled fluid in the posterior vaginal fornix on day 0 (T0) before gel application and 15 min following gel application (T15 min). Women were instructed to insert the gel product twice daily, once in the morning and once in the evening, for 14 days and not to insert gel on the morning of and to abstain from sexual intercourse for 24 h prior to the day 7 and day 14 study visits. At 8 to 24 h after the last gel application, on day 14, a CVL sample (T8-24 h) was collected. CVL samples were fractionated by centrifugation at 400 × g for 15 min at 4°C. The soluble supernatant fractions (CVL-s) and pelletable material (CVL-p) were collected, aliquotted, and stored at −70°C.
Stored CVL-s and CVL-p samples were gamma irradiated with 2 × 104 Gy while on dry ice to prevent microbial contamination (see below). Preliminary studies demonstrated that gamma irradiation did not change the ability of UC781 to be detected or measured and did not affect UC781 levels or antiviral activity (data not shown). Following gamma irradiation, CVL-p samples were resuspended in 200 μl PBS, and a 100-μl aliquot was used to determine the UC781 concentration by high-performance liquid chromatography (HPLC). Similarly, 100 μl of gamma-irradiated CVL-s was used to determine the UC781 concentration. Samples (100 μl) were mixed with 200 μl of dimethyl sulfoxide and briefly centrifuged to clarify them. A 100-μl aliquot of the supernatant was injected into a Luna C18(2) 250- by 3-mm column (Phenomenex, Torrance, CA) connected to a Shimadzu Prominence HPLC system (Columbia, MD). The HPLC mobile phase A was 0.1% trifluoroacetic acid (TFA) in water, and the mobile phase B was 0.1% TFA in acetonitrile. UC781 detection was monitored by UV absorbance at 297 nm, and it was eluted from the column first at 70% mobile phase B for 5 min, followed by a gradient from 70% to 95% B in 10 min. The concentration of UC781 was determined by linear regression analysis using a standard curve generated by measurement of known UC781 concentrations (0.007 to 10 μg/ml). The sensitivity of UC781 quantification was 0.01 μg/ml in CVL-s and 0.09 μg/ml in CVL-p. A sample with UV absorbance at the UC781 peak that was above background but less than that of the lowest standard (0.007 μg/ml) was reported to have a detectable but less than quantifiable level of UC781. UC781 levels were reported as the amount in the total CVL-s or CVL-p sample. Specifically, the measured concentration (μg/ml) of UC781 in an extracted aliquot of CVL-s or CVL-p was multiplied by the total supernatant volume or the total number of cell pellets obtained from the whole CVL sample. Thus, the lower limits of quantification per total CVL-s and CVL-p sample varied slightly depending on minor differences in supernatant volumes and the cell pellet numbers among samples (data not shown). We therefore set minimum values for quantification of 0.1 μg and 5 μg per total CVL-s and CVL-p sample, respectively. Statistical analyses were performed using the SigmaPlot 11.0 software package (Systat Software, Inc., Chicago, IL). Mann-Whitney rank sum tests were used to compare UC781 concentrations between time points, groups, and CVL fractions (a P value of <0.05 was considered statistically significant).
The anti-infective activities of CVL-s and CVL-p samples were determined using a titer reduction assay with TZM-bl cells in a transwell format to allow cells to grow without being in direct contact with the cellular and particulate debris in CVL-p samples. TZM-bl cells were plated at 5 × 103 cells per well in the upper chamber of a 96-well transwell plate (1.0-μm-pore-size polyester [polyethylene terephthalate] membrane; Corning Life Sciences, Inc., Lowell, MA) with the appropriate CVL-s or CVL-p sample in the receiver plate. CVL-s samples were thawed, and 100 μl was aliquoted into each of 9 receiver wells. A 100-μl aliquot of resuspended CVL-p sample was diluted 1:15 in complete Dulbecco's modified Eagle's medium (cDMEM), and 100 μl of the dilution was aliquoted into each of 12 receiver wells. The upper chamber with TZM-bl cells was placed in the receiver plate containing cDMEM, CVL-s, or CVL-p samples and allowed to equilibrate overnight at 37°C (7% CO2 atmosphere). HIV-1 CRF01_AE was serially diluted (1:5, 1:25, 1:125, 1:625, and 1:3,125) in cDMEM, and 25-μl amounts of appropriate virus dilution or cDMEM (negative control) were added in triplicate to TZM-bl cells in the upper chamber of the cDMEM-equilibrated transwells to provide a positive control and determine the 50% tissue culture infective dose (TCID50) for each virus assay. cDMEM was added to 3 wells of CVL-p sample-equilibrated TZM-bl cells to provide a negative control and determine the cutoff for positive virus infection in the presence of each woman's CVL-p sample (average luminescence plus 2 standard deviations). Only dilutions 1:5, 1:25, and 1:125 were used in triplicate with CVL-s or CVL-p samples due to limited specimen volume. CVL-s samples were not used in the absence of virus due to limited specimen volume. Using this transwell method, UC781 from the CVL-s or CVL-p samples would need to diffuse through the transwell membrane and be present in the solution phase of the upper chamber or be absorbed by TZM-bl cells contacting the transwell membrane in order to block virus infection of TZM-bl cells in the transwell. Virus and cells were allowed to incubate for 48 h, and HIV-1 infection was detected in the cells using the luciferase assay system (Promega Corp., Madison, WI). The manufacturer's protocol was followed except that the lysis buffer volume was increased from 20 μl to 100 μl/well and lysed contents were transferred from the upper chamber of the transwell to a 96-well plate to measure luminescence on a Fluoroskan Ascent FL luminometer (Thermo Fisher Scientific, Inc., Waltham, MA). Luminescence was reported as relative light units (RLUs), and the TCID50 of virus was calculated using the Reed and Muench method (23). Infectious virus titer reductions for CVL-s or CVL-p samples were reported as the difference in their TCID50 values compared to the TCID50 value of the corresponding CVL-s or CVL-p T0 sample collected from the same patient.
Between 22 May 2007 and 18 September 2007, 114 women were screened and 45 were enrolled in the trial. All women enrolled in the trial completed the study. All women were married or in a monogamous heterosexual relationship during the study. During the trial, 45 (100%) were sexually active and had sex a median of 5 times, with condom use a median of 5 times, during the 14 days using the product.
Self-reported adherence to gel use was high, with 42 (93%) women using at least 25 of 27 scheduled doses, and none of the women interrupted or completely stopped gel use for reasons related to adverse effects. The numbers and types of adverse events were similar between study arms (data not shown).
UC781 drug levels were assayed in CVL-s and CVL-p samples from all 15 women in the 0.25% UC781 arm and 5 women randomly selected from the 0.1% UC781 and placebo arms, and total UC781 quantities were calculated for the soluble and insoluble fractions of each CVL.
UC781 was detected in T15 min CVL-s samples from 20/20 women in the 0.1% and 0.25% UC781 groups (Fig. 1). In T8-24 h samples, UC781 levels were detectable in 1/5 and 7/15 CVL-s samples from the 0.1% UC781 and 0.25% UC781 groups, respectively. UC781 levels could only be measured in 0/5 and 4/15 CVL-s samples from the 0.1% UC781 and 0.25% UC781 groups, respectively. The UC781 levels were significantly higher in CVL-s samples from the T15 min time point than in T8-24 h samples for both the 0.1% UC781 (P = 0.008) and 0.25% UC781 groups (P < 0.001) (Fig. 1). When quantifiable, the levels of UC781 in CVL-s samples were greater than the range of 50% effective concentration (EC50) calculations reported by Balzarini et al. (5).
UC781 levels were higher in the T15 min samples and lower in the T8-24 h samples. UC781 was detected in 20/20 T15 min CVL-p samples in both the 0.1% and 0.25% UC781 groups. While UC781 was detected in all T8-24 h CVL-p samples in both the 0.1% and 0.25% UC781 groups, the UC781 levels were only high enough to be quantified in 1/5 and 10/15 CVL-p samples from the 0.1% UC781 and 0.25% UC781 groups, respectively. The T15 min CVL-p samples showed significantly higher levels of UC781 than the T8-24 h samples for both the 0.1% UC781 (P = 0.008) and 0.25% UC781 (P < 0.001) groups (Fig. 1). Additionally, the levels of UC781 in T8-24 h CVL-p samples were significantly higher in the 0.25% UC781 group than in the 0.1% UC781 group (P = 0.036) (Table 1). CVL-p samples showed significantly higher UC781 levels than CVL-s samples at both time points (Table 1). All CVL-p samples with measurable UC781 contained drug levels as much as 10-fold higher than that deemed necessary for in vitro sterilization by Watson et al. (30). These data show that hydrophobic compounds persist at levels inhibitory to HIV infection in vitro for at least 8 to 24 h following vaginal application.
The detection of UC781 in T8-24 h samples after extended exposure to the cervical and vaginal mucosa suggests that UC781 is not degraded following application, which is important for potential use as a coitally independent method of HIV prevention. The vast majority of UC781 was found in the insoluble pelletable fraction of all CVLs, suggesting that future examination of preclinical and clinical studies of UC781 must focus on both cellular and cell-free fractions of vaginal samples. A previous preclinical study of UC781 microbicide gel in macaques was unable to detect UC781 in clarified CVL samples, as most UC781 was likely pelleted and removed prior to analysis of clarified CVL samples (20). Focusing on insoluble and pelletable fractions of CVL samples may be especially important when examining other candidate hydrophobic compounds with low solubility, such as dapivirine (14).
The levels of UC781 detected in most CVL-p samples suggest the capacity to prevent HIV infection; however, it was important to determine whether UC781 retained its anti-infective activity following vaginal application. Most UC781 levels in CVL samples were detected in the CVL-p fraction (Table 1); therefore, T8-24 h CVL-p samples from all 20 women in the 0.1% and 0.25% UC781 groups and a random sample from the placebo group were analyzed for their ability to block HIV infection. CVL-p samples collected at T0, prior to the first gel application, were used as an internal control for anti-infective activity. None of the T0 CVL samples examined had anti-infective activity in the titer reduction assay (data not shown). T15 min and T8-24 h CVL-s samples from the 4 women with the highest levels of UC781 in T15 min CVL-s samples were also selected for evaluation of anti-infective activity.
UC781 anti-infective activity was detected in CVL-p samples from 2/5 women from the 0.1% UC781 group and 13/15 women from the 0.25% UC781 group (Table 2). In all instances, CVL-p samples completely blocked infection with 105 TCID50 of CRF01_AE HIV-1. Therefore, using the titer reduction calculations, all samples that showed anti-infective activity were able to block ≥4 log10 TCID50 CRF01_AE HIV-1, and one sample (from patient 11) was able to block ≥5 log10 TCID50 CRF01_AE HIV-1 (Table 3). Importantly, every CVL-p sample that had >5 μg UC781/CVL sample (11/20) was able to block HIV infection of TZM-bl cells. Additionally, 4/9 samples with UC781 levels below the quantifiable level (5 μg/CVL sample) still had anti-infective activity. Importantly, there was no measurable anti-HIV activity in the CVL-p samples from the placebo group or T0 samples collected prior to UC781 gel application (data not shown).
Examination of CVL-s T15 min samples revealed that 3/4 samples were able to inhibit ≥4 log10 TCID50 CRF01_AE HIV-1 (Table 3). The 3 samples able to inhibit infection of TZM-bl cells had the highest levels of UC781 in the CVL-s fraction. Moreover, while 15/20 CVL-p samples from T8-24 h were able to prevent infection of TZM-bl cells, no CVL-s samples (0/4) selected from T8-24 h were able to block infection of TZM-bl cells (Table 3). Of note, the UC781 levels in T8-24 h samples collected from patients 20 and 21 were concentrated in the CVL-p fractions such that the CVL-s fractions contained little to no UC781 drug and no detectable anti-HIV activity, while the CVL-p fractions contained quantifiable levels of UC781 as well as potent anti-HIV activity (Table 3).
A primary concern of clinical evaluations of microbicides is that the compound being tested should demonstrate anti-HIV activity following exposure to the female genital tract. In this study, we investigated whether UC781 was inactivated in vivo, possibly through contact with enzymes or microflora in the vaginal mucosa that might render intravaginal products ineffective over time, as was suggested previously (14). Testing T8-24 h CVL-p samples for anti-HIV activity indicated that UC781 blocked large inocula (≥4 TCID50) of a virus most likely to be sexually transmitted to HIV-negative Thai women. Recently, Anton et al. demonstrated that rectal biopsy specimens collected after 30 min but not after 7 days of exposure to UC781 rectal gel were protected from HIV challenge ex vivo (2). The differing results between the study by Anton et al. and these results are likely a reflection of the differences between rectal and vaginal tissues, as well as the use of alternative methods for ex vivo evaluation of anti-HIV activity. Our transwell method can be used to perform ex vivo analysis of any vaginally or rectally applied microbicide with poor solubility or a high level of cellular uptake, such as NNRTI-based microbicide candidates, in order to evaluate the impact of mucosal secretions on microbicide efficacy.
Despite the high level and antiviral activity of UC781 following exposure to the female genital tract, UC781 microbicide gel may not provide protection against HIV infection in clinical settings. It is possible the microbicide gel may not effectively deliver UC781 to HIV target cells within the female genital tract. It is also possible that semen may inhibit UC781 antiviral activity in vivo as has been suggested for other microbicide candidates, although previous in vitro studies indicate that UC781 is unaffected by seminal plasma (15, 18). Additionally, UC781 bioavailability may be reduced during coitus as recently reported by Keller et al. for PRO2000, potentially rendering UC781 microbicide gel less effective against HIV infection (16). Through the use of multiple tests incorporating seminal plasma, coitus, and ex vivo assays, future safety and preclinical trials will be able to generate important results regarding the potential efficacy of microbicide candidates prior to extensive clinical efficacy trials. Lead candidate microbicides need to pass rigorous tests of in vitro, safety, and ex vivo testing that will ensure only the safest and most active microbicide candidates are pursued in clinical efficacy trials.
This study questions whether UC781 in CVL-p fractions has been absorbed into epithelial cells and leukocytes that are collected in the CVL or formed an insoluble coating on the surface of the cellular material. It is possible that in T8-24 h samples, UC781 has been absorbed by cells within the female genital tract and persists to prevent HIV infection. Following collection, storage, and thawing, UC781 may have been released from these cells to be absorbed by TZM-bl cells in the anti-infective activity assay. It is also possible that UC781 has remained in an insoluble extracellular form that is also able to block HIV infection in this ex vivo assay. We observed crystalline occlusions in some CVL-p samples with the highest levels of UC781 (unpublished observations), suggesting that high levels of UC781 may have remained in an insoluble extracellular form that was released upon resuspension during the assay. We previously saw similar crystalline occlusions that were stuck to cells in laboratory-based assays for more than 2 days after a single application of UC781 gel in vitro that was followed by multiple wash steps to remove excess gel (unpublished observations). An important limitation of our study is that we could not determine whether the anti-infective drug levels we measured at T8-24 h came from a single application of gel or whether the multiple, twice-daily applications were required. However, this study was able to determine that UC781 retained its anti-infective activity, regardless of its subcellular localization within the female genital tract. It is also important to note that 4/9 CVL-p samples with UC781 levels below the level of quantification were still able to prevent infection with high levels of CRF01_AE HIV-1 (Table 2), suggesting that levels of UC781 within the EC50 range were still fully capable of preventing HIV infection in our titer reduction assay.
In conclusion, ex vivo analysis of vaginal lavage samples collected in a phase I clinical safety trial of UC781 vaginal microbicide revealed that the majority of UC781 is found in an insoluble fraction but is still able to inhibit HIV infection. While further development of UC781 has been halted due to difficulties with solubility and stability, the results presented here highlight the need to examine pelletable fractions from mucosal lavage fluids for the quantities and activities of antiviral agents.
We thank the study participants for their contribution of time, energy, and commitment to this effort. The authors also acknowledge colleagues who contributed to the implementation of the study: the clinical study team at the Chiang Rai Health Club, Sara Whitehead, Tepnaruemit Medtanavyn, Philip Mock, Chalinthorn Sinthuwattanawibool, Wanna Leelawiwat, Punneporn Wasinrapee, Lisa Grohskopf, Gustavo Doncel, and Christine Mauck.
Financial support for this study was provided by the United States Centers for Disease Control and Prevention.
The findings and conclusions in the manuscript are those of the authors and do not necessarily represent the views of the United States Centers for Disease Control and Prevention or the Department of Health and Human Services.
Published ahead of print 16 April 2012