Background To effectively block the invasion of human immunodeficiency virus (HIV)-1 on mucosal surface, vaginal anti-HIV-1 microbicides should avoid inflammatory responses and disruption of mucosa integrity because these will facilitate transepithelial viral penetration and replication. However, existing models fail to predict and evaluate vaginal mucosal toxicity induced by microbicides, and most importantly, they are unable to identify subtle or subclinical inflammatory reactions. This study was designed to develop a cost-effective in vivo model to evaluate microbicide safety in a preclinical study which can recapitulate the mucosal topical reaction.Methods A murine model was employed with nonoxynol-9 (N-9) as the topical stimulant within the vagina. Different concentrations of N-9 (1%, 3%, and 4%) were topically applied to the vagina for five consecutive days. A panel of inflammatory cytokines including interleukine-2 (IL-2), IL-4, IL-6, IL-17A, interferon-Y (IFN-Y), tumor necrosis factor-α (TNF-α), and immuno-regulatory IL-10 were assayed in vaginal lavage. Cytokines were quantified by using cytometric bead array (CBA) and reverse transcript (RT) real-time PCR. Histopathological evaluation of vaginal tissues was conducted on hematoxylin-eosin stained slides and scored with a semi-quantitative system according to the severity of epithelial disruption, leucocyte infiltration, edema, and vascular injection. The association between the cytokines and histopathological scores was assessed by linear regression analysis.Results All three concentrations of N-9 induced inflammatory cytokine production. The 4% N-9 application resulted in a consistent production of cytokines in a time-dependent manner. The cytokines reached peak expression on day three with the exception of IL-4 which reached its peak on day one. Histopathological examination of 4% N-9 treated cervicovaginal tissues on day three showed intensive damage in four mice (sores: 10-13) and moderate damage in one mouse (score: 8), which were significantly associated with both inflammatory cytokines IL-17A and IL-6 and anti-inflammatory cytokines IL-4 and IL-10. Interestingly, IL-17A showed significant positive association with inflammatory cytokine TNF-α (r=0.739; P ＜0.05), anti-inflammatory cytokines IL-10 (r=0.804; P ＜0.01) and IL-4 (r=0.668; P ＜0.05).Conclusions Our data demonstrate that a panel of cytokines (IL-17A, IL-6, IL-4 and IL-10) could be used as surrogate biomarkers to predict the histopathological damage. Th17 may play a central role in orchestrating inflammatory cytokine responses. This Th17 based mouse model is cost-effective and suitable to assess the toxicity of candidate microbicides in preclinical studies.