ABSTRACT
In vitro models of drug efficacy have evolved over the past 10 years reflecting lessons learned from the first generation of candidate drugs, including pH buffering agents (ACIDFORM® and BufferGel®) and sulfated/sulfonated polymers (PRO 2000, CS and Carraguard®). The latter prevented HIV entry primarily by binding to the V3 loop of gp120, but were effective only in the microgram/milliliter concentration, were less potent against R5 viruses compared to CXCR4-tropic (X4) viruses, and the antiviral activity was diminished when virus was introduced in diluted seminal plasma [29-31]. These in vitro findings translated into lack of efficacy in clinical trials [9,10,14]. The acid buffering agents were developed based on observations that clade B HIV is susceptible to low pH, but subsequent studies found that several non-clade B HIV primary isolates were not inactivated at a low pH [32] and BufferGel® did
not provide protection in clinical trials [11]. These studies led to the modification of the algorithm for pre-clinical testing of candidate drugs, including the evaluation of R5 viruses and clinical isolates representing multiple clades, testing of drugs over a wide pH range and with rapid pH transition from 3.5-4.5 to 7-8 (as might occur following coitus), and evaluation of efficacy when cells are exposed to female genital tract secretions and the virus is introduced in semen or seminal plasma (see Section 2.5). These studies are typically conducted first using the API in cell lines (typically Jurkat, H9, PM1, TZM-bl, and GHOST cells) or peripheral blood mononuclear cells (PBMCs), and then advanced to studies using explant cultures (cervical, vaginal, penile, or colorectal tissue) and animal models.The optimal concentrations of drug that must be delivered orally or topically to protect against infection are not known. NHPs and explant culture studies suggest that protection with topical PrEP will require concentrations that are substantially greater than doses effective in cell culture systems, reflecting variable tissue penetration and distribution [33,34]. This differential may be greater for drugs such as TFV and related prodrugs because the drugs are internalized by epithelial cells, where they are converted to TFV-diphosphate (TDP) as evidenced by the anti-herpes simplex virus (HSV) activity [35]. This may reduce the effective dose of drug available for submucosal immune target cells as TDP has a prolonged intracellular half-life and could be trapped in the epithelial cells. We modified a dual-chamber culture model initially designed to evaluate microbicide safety to evaluate how well candidate drugs and formulations cross a multilayered epithelial barrier (Section 2.6) [36], which complements the classic Franz cell apparatus [37,38]. HEC-1-A cells, a human endometrial cell line that grows in a multilayered manner thus mimicking the multilayered vaginal epithelium, are cultured on Transwell® inserts and polarity is monitored by measuring transepithelial electrical resistance (TEER) and imperviousness to HIV. The model can be expanded to examine the impact of genital tract secretions on permeability by exposing the cells apically to drug in the presence of female genital tract secretions and/or semen and collecting basolateral supernatants over time for measurement of drug levels (PK) and antiviral activity (PD).
