Microbicides: In Vitro and In Vivo
Thursday, 1:30 - 3:30 pm
Background: The first microbicide to be clinically evaluated contained the detergent Nonoxynol-9 (N9). It was found that N9 did not prevent HIV infection and that frequent use can result in increased susceptibility due to its toxicity to epithelial cells. Because of its wide availability, many laboratories have used N9 as a historical control for toxicity. However, no published comparisons of the results among laboratories or attempts to establish a standardized protocol for testing of microbicides exist. Our study is the first report to investigate N9 toxicity in a multi-center comparison.
Methods: The study included 81 toxicity assays from 5 laboratories between 1998 and 2004, employing 15 different cell lines or tissues. Additional variables included duration of exposure, N9 concentration, number of replicates, and laboratory site. Intra-assay reproducibility was measured at 2, 3, and 5-fold differences using standard deviations). Inter-assay reproducibility was assessed for each series of values for the same cell line, laboratory and N9 concentration using general linear models. A step-wise regression was used to study the interaction between the variables.
Results: The intra-assay reproducibility with the same N9 concentration, cell type, time, and laboratory was remarkably consistent down to the 2-fold level. For inter-assay comparison, cell line, duration of assay and N9 concentration were all significant sources of variability (p < 0.01 for all). Either long (24 of 48 hours) or short (< 2 hours) exposure of cells to N9 showed greater variability, while assays with N9 exposure for 4-to-8 hours were not significantly different. There was no effect of replicates (3, 6, or 7+) on reproducibility. In a step-wise regression analysis, the N9 concentration and laboratory strongly affected reproducibility (R2 = 0.268, p < 0.001 and R2 = 0.032, p = 0.005, respectively), while assay duration and number of replicates had no significant effect (R2 = 0.349, p = 0.085, and R2 = 0.347, p = 0.404).
Conclusion: This is the first analysis to compare toxicity levels of a microbicide obtained by different laboratories using different protocols. While the data within one laboratory were consistent, the data obtained between the laboratories were highly variable. This comparative work will be used to develop a standardized microbicide toxicity testing protocol and proficiency-testing program that will help advance potential microbicides to clinical trials.
Keywords: Microbicides; Nonoxynol 9; Multicenter evaluation