Abstract Body

Background: Clinical trials with daily dosing of tenofovir (TFV) } emtricitabine have shown disparate results for HIV protection. It is still unknown what target tissue concentrations must be achieved to confer protection. Here we describe the utility of a novel ex vivo human mucosal tissue culture model to determine the tenofovir diphosphate (TFVDP) target for protection. Methodology: Fresh vaginal tissue was collected from 8 women undergoing surgical repair for vaginal prolapse. 5 subjects were used to characterize TFVDP formation and elimination over 96h. 3 subjects were used for HIV challenge, whereby up to 35 3mm punch biopsies (Miltex) per donor were incubated in a range of TFV concentrations (0, 0.1, 1, 10, 100 ug/mL) for 24h. TFVDP was quantified using a validated LC-MS/MS method. Remaining explants were incubated for 3h with 10^7 TCID50 HIV-1JRCSF, rinsed, and transferred to gelfoam rafts. Explants were harvested every 24h from 0-72h post-infection, with total RNA extracted using RNeasy Kit (Qiagen) and first strand cDNA synthesis performed with Superscript Vilo cDNA Synthesis Kit (Invitrogen). A quantitative RT-PCR assay measured copies of a spliced RNA variant expressed in abundance early in the transcription life cycle, and normalized for tissue weight. Infection was quantified as AUC24-72h adjusted for baseline. EC50 estimates were generated using an Emax model (Sigma Plot) and compared to estimates obtained in the TZM-bl cell line using a luciferase reporter assay. Results: Vaginal explants phosphorylated TFV to TFVDP at a median molar conversion rate of 0.2% (range 0-2.0 %) and reached maximal conversion at 24h. TFVDP elimination half-life was 20h. Viral infection was confirmed by a peak in spliced RNA >1.5-fold baseline, and peaked at 2 days after infection. There was a poor relationship between TFV concentration and HIV protection. When normalized for TFVDP, standard error of the estimated EC50 was reduced by 4 logs. Estimated EC50 of TFVDP was 909 fmol/mg in explant tissue and 20 fmol/mg in TZM-bl cells. Conclusions: This method uses an R5, T-cell-tropic virus to mimic the virus implicated in mucosal HIV transmission. When coupled with a novel spliced RNA assay, an Emax model of TFVDP was generated using 72h incubations. Variable phosphorylation resulted in a more reliable EC50 estimate when TFVDP (the active intracellular moiety) was used, rather than TFV incubation concentrations. Likely representing a “worst case scenario”, this EC50 was ~45 times higher than a standard cellular model, but consistent with other recent clinical data. We are currently using this model to test additional drugs and drug combinations to identify target mucosal concentrations for protection against HIV infection, and determine its utility in PrEP development.