In this study, we used normal cycling female pigtailed macaques in a repeat-challenge model to evaluate the window of protection conferred by a vaginal 1% TFV gel. We predicted that both the high concentration of TFV in the gel and the long intracellular half-life of TFV-DP would result in durable protection. Our data show that 1% TFV gel maintained significant efficacy (74%) at 3 days after application, thus providing a wide window of protection to macaques against vaginal SHIV transmission. To better understand the tissue drug levels needed for protection, we examined TFV-DP concentrations in vaginal lymphocytes collected between 4 h and 3 days after administering a single gel dose and related these concentrations to the TFV-DP levels associated with in vitro
inhibition. Intracellular drug analysis revealed persistently high TFV-DP concentrations detected as early as 4 h and up to 2 days, reflecting rapid accumulation and long persistence in vaginal lymphocytes. Furthermore, these TFV-DP concentrations exceeded the intracellular TFV-DP IC95
, which likely explains the complete protection previously observed in macaques challenged 30 min after gel application (21
). The persistence of these TFV-DP levels in vaginal lymphocytes for up to 2 days further implies that this level of high efficacy would likely extend to 2 days. In contrast, our data show an unexpected 7-fold decrease in TFV-DP concentrations in vaginal lymphocytes at 3 days and, thus, a shorter half-life (25 h) than that in PBMCs after oral dosing (49 h). The reduced TFV-DP levels in the vaginal lymphocytes are in the range of the in vitro
and likely explain the reduction in efficacy from 100 to 74% when TFV gel was applied at 3 days as opposed to 30 min before virus exposure. While our findings could be strengthened by larger numbers of macaques and more than two efficacy measurements, the strong correlation we observed between TFV-DP concentrations, their associated in vitro
activity, and in vivo
protection suggests that TFV-DP in vaginal lymphocytes is a good predictor of efficacy. Thus, data from this model provide the first information on the protective TFV-DP levels and point to TFV-DP concentrations in vaginal lymphocytes that are above the in vitro
as predictors of high protection by TFV gels.
Although we have not measured the TFV-DP concentrations in vaginal lymphocytes after oral TDF dosing in macaques, recent data for humans are consistent with our macaque data and have demonstrated much higher TFV-DP exposures in vaginal tissues after TFV gel compared to oral TDF dosing, implying that oral TDF may not confer the same level of protection as topical TFV gel against vaginal infection (3
). An interim analysis of two human clinical trials has recently shown conflicting efficacy results. In the VOICE trial, oral TDF was found to be ineffective, while the Partners PREP study showed a reduction in HIV-1 infection by 68% among women who have HIV-1-infected partners (J. Baeten et al., presented at the 6th IAS Conference on HIV Pathogenesis Treatment and Prevention, 17 to 20 July 2011, Rome, Italy, and MTN-003). The reasons for the discrepancy between the two study results are not clear at this time, as further analyses of adherence and other parameters are needed. If oral TDF is found to be highly protective in women despite the lower vaginal TFV exposures, this would suggest a mechanism of dual protection whereby systemic and local antiviral activities both contribute to prevent infection.
We also found that vaginal permeability reduces intracellular TFV-DP concentrations in vaginal lymphocytes by about two orders of magnitude relative to in vitro
dosing, which underscores the need for highly concentrated (mM) TFV gels to achieve in vivo
protection. The scale-up in drug concentration needed for in vivo
protection with topical gels has been previously observed with other drugs and likely reflects the permeability of drug through the vaginal mucosa, as well as the need to protect a large vaginal surface area for a significant period of time (32
). The TFV-DP concentrations in vaginal lymphocytes associated with high efficacy that were identified in this study may provide reference values to guide the development of devices for vaginal TFV delivery by different gel formulations, vaginal rings, or films.
The persistently high levels of TFV-DP detected in vaginal lymphocytes for up to 2 days, which were previously shown to be associated with complete protection, suggest that the same level of protection could be maintained within 2 days after dosing. These data and the 74% efficacy observed at 3 days all point to a long window of protection that supports coitus-independent gel use, which may be more desirable to women and help improve adherence. The window of protection for TFV gel is longer than that recently reported for macaques with a gel containing the CCR5 entry inhibitor maraviroc (33
). Maraviroc gel (6 mM) fully protected macaques from vaginal SIV infection when administered 30 min before exposure, but the protection was short-lived, as complete loss of protection was observed by 12 h (33
). The reasons for the short window of protection by the maraviroc gel are not known but could be due to rapid CCR5 receptor recycling and lymphocyte turnover. Lymphocyte turnover in vaginal tissues could also explain the precipitous drop in TFV-DP levels at 3 days following TFV gel dosing. It will be important to compare the half-lives of TFV-DP in vaginal lymphocytes from macaques with those in women following TFV gel dosing.
Drug resistance emergence when antiretroviral prophylaxis fails is a concern, particularly if the intervention includes drugs that are widely used for treatment, as is the case with TFV. Consistent with human data from CAPRISA 004 showing only wild-type virus detected in women who became infected while using TFV gel, we demonstrated that two macaques who failed TFV gel treatment had no evidence of the TFV-associated K65R mutation in plasma by conventional sequencing and ultrasensitive testing, despite continued gel dosing postinfection (1
). We also found only wild-type SHIV populations in the vaginal secretions of both macaques, which was reassuring because it indicated that the higher drug selection in vaginal tissues, as a result of vaginal dosing, had no impact on K65R emergence.
We further showed that TFV was detected in 60% of plasma samples collected 30 min after gel dosing, confirming very rapid absorption through vaginal tissues. Plasma TFV was detected at higher levels in macaques than in women who applied vaginal TFV gel containing 40 mg (16
). This difference may be explained by the approximately 12-fold-smaller blood volume in macaques, which may concentrate TFV, thus making it more easily detected in macaques than in women, especially if vaginal absorption is similar in both species (26
). We also found no reduction in the acute viremia in breakthrough infections compared to controls, reflecting insignificant antiviral activity from any systemic TFV exposure by the once-weekly gel dosing modality.
Our study has several limitations. First, the virus inoculum was clonal, and all challenges were done in the absence of semen or semen-derived factors shown to enhance HIV infection in vitro
). Second, exposures were to an intact vaginal mucosa without trauma associated with coitus, concurrent genital ulcers, or bacterial vaginosis, all of which can increase the risk of HIV acquisition (2
). Third, the gel volume was decreased to 3 ml (30 mg TFV), compared to 4 ml (40 mg TFV) in women, to accommodate the smaller vaginal cavity in macaques, which may not fully mimic the drug exposure in women. Although the drug exposure may be greater based on the smaller vaginal surface area in the macaques than in humans, it is unclear whether the difference in scale is critical because the drug may not be limiting in either humans or macaques given the high local drug dose and the small amount (~0.02%) that is absorbed (21
). Importantly, data from humans dosed with 4 ml of 1% TFV gel show intracellular TFV-DP concentrations in cells collected from vaginal tissue at 24 h (~1,700 fmol/106
cells) that are similar to the levels in macaques at 24 h (~1,100 fmol/106
cells), suggesting little impact of the difference in scale (27
In conclusion, we demonstrate that topical application of a gel containing 1% TFV administered 3 days before virus exposure provided durable protection in macaques. Data from this model suggest that gel application immediately before sex may not be required to provide high protection. We also show that gel dosing results in very high TFV-DP levels in vaginal lymphocytes but demonstrate a faster decay of TFV-DP than previously predicted from oral dosing. Our data are the first to identify protective intracellular drug levels in vaginal lymphocytes, which should facilitate the development of improved methods for vaginal delivery of TFV and inform next-generation efficacy and PK studies in humans.