An experimental mouse model of Candida vaginitis has been established and historically used for the past few decades to study mucosal host response to Candida as well as for testing antifungal therapies3,4,11,13,16,17,19,21,24,25,37. The protocols presented here incorporate efficient and less labor-intensive methods, and appear to be one of the most optimized model systems of Candida vaginitis described to date. These techniques enable rapid quantification of fungal burden and collection of vaginal specimens. Furthermore, previous studies testing several haplotypic strains of mice (n=13) and various time points post-inoculation showed the similar levels of variability in fungal burden and host responses to vaginal colonization with Candida2,6,41. Hence, this model can be adapted to existing protocols without restriction on mouse strains or the duration of infection. However, one should recognize that the variability in vaginal fungal burden can be high between animals given the same inoculum (Figure 6B). There is evidence supporting that the variability occurs due to differing degrees of early adherence to vaginal epithelium41. Therefore, 7-10 mice/group is suggested for statistical purposes.
Variability in vaginal fungal burden has also been evidenced between different strains of C. albicans, suggesting that not all strains of C. albicans have the capacity to similarly colonize the mouse vagina. For instance, C. albicans 3153A (lab strain) used in this protocol has been reported to show higher vaginal colonization than SC5314 (clinical isolate), where a higher inoculum (> one log) would be required to obtain equivalent levels of vaginal fungal burden seen with 3153A27. In fact, a highly variable vaginal colonization with several clinical isolates has been documented36. Hence, care should be taken when determining an optimal inoculum for each C. albicans strain to ensure consistent colonization in mice. Typically, mean CFUs ranging from 5 x 104 to 5 x 105/100 μl lavage fluid should be detected for consistent colonization. For assessment of vaginal fungal burden, quantification of CFUs by lavages is appropriate for this model as Candida blastoconidia and hyphae normally do not penetrate beyond the superficial layer of vaginal epithelium. Our previous histological evaluation of post-lavage vaginal tissues rarely showed residual Candida25,41. However, we recognize the possibility that hyphae could be miscounted as they grow as a single colony on the agar plate and might not reflect accurate fungal burden. As an alternative method to CFU enumeration, a newly developed in vivo imaging technique has been reported to successfully assess vaginal fungal burden when using genetically engineered luminescent C. albicans9,29. Furthermore, the protocol can be modified to induce C. glabrata colonization in diabetic animals15,20. To date, mouse models for other non-C. albicans-induced vaginitis have not been reported.
Estrogen administration is critical for this model, initiating robust vaginal colonization with Candida6,14,23. In addition to subcutaneous injection of estradiol in sesame oil suspension, injection of water-soluble estradiol in PBS and intradermal implantation of a controlled-release estradiol pellet are alternative methods for estrogen administration and have been employed in other mouse models of female genital tract infections10,35. Requirement of high estrogen in this model may be explained by two physiological factors. One is that elevated estrogen promotes stratification of vaginal epithelium5,8,30. Thickened epithelium from increased epithelial cell proliferation may allow Candida to gain better adherence to the vaginal wall and subsequent colonization. Secondly, vaginal epithelial cells are known to have high glycogen content. An increased tissue estrogen level results in deposition of glycogen into the vaginal walls30. Elevated glycogen in the vaginal microenvironment may in turn allow Candida to flourish by providing additional nutrients. Previous immune analyses showed that exogenous estrogen did not affect cell adhesion molecule expression40. A drawback of treating the animals with exogenous estrogen is that elevated estrogen has also been known to promote overgrowth of vaginal commensal flora30. Since the composition of flora may vary significantly between animals, this may add a variable where commensal organisms could influence Candida growth or modulate host responses. In consideration of this issue, the model requires estrogen-treated control (uninoculated) animals to be included in all experiments and analyzed in parallel.
In this mouse model, the efficacy of anti-fungal agents can be accurately assessed by the protocols described herein, which may provide crucial in vivo testing leading to the development of potential therapies for clinical use. In fact, the robust nature of the model provides good indications of clinical efficacy. In addition to the use of estrogen, neutral vaginal pH in mice promotes growth of hyphae, a pathogenic form of Candida seen in all inoculated mice. Similar to clinical observations, a robust vaginal neutrophil migration occurs in a subset of mice without affecting fungal burden31,41. The presence of neutrophils is prominent in women during symptomatic vaginitis and appears to parallel mice following inoculation12,41. Since other clinical symptoms (i.e., itching, swelling) cannot be precisely measurable in mice, the assessment of vaginal neutrophils serves as a simple yet reliable indication of inflammation (symptoms) in this model.
The microscopy of the vaginal lavage fluid shown in Figure 5 is routinely performed in our laboratory to assess Candida colonization and vaginal inflammation. Moreover, hyphal scoring can also be used as a measure of infection. Thereafter, the cellular and soluble fractions of the lavage fluid can be preserved and cryo-archived for future analyses. Of note, an advantageous feature of the protocol is that vaginal lavages can be performed longitudinally in the same mice under anesthesia. Our previous studies confirmed that longitudinal evaluation does not influence assessment of vaginal fungal burden. This approach is particularly advantageous in cases where longitudinal analyses of infection are desired.
Finally, we have developed a less-invasive excision method for the vagina. This effective and quick technique requires no incision in the abdomen or any internal organs, leaving the rest of the genital tract intact. Another useful feature of this model is the lack of requirement for immunosuppressive agents to initiate Candida colonization. This is particularly important because maintaining natural immune status of the host is a critical aspect of immunological studies and host responses to a microbial challenge. Hence, tissues and cells isolated by these methods could be applied in various in vitro immune assays.
In conclusion, we provide several important features and representative results of the experimental model of vaginal candidiasis. In addition to C. albicans-induced vaginitis, infections by other pathogens of the female lower genital tract, including Neisseria gonorrhoeae, Trichomonas vaginalis and Herpes simplex virus have been studied using mouse models where the organisms are intravaginally introduced in hormone-treated mice1,7,10,22,26,35,38. Therefore, the techniques that are useful for studies involving Candida vaginitis can be applied to and potentially advance methodologies to study pathogenesis and host immune responses for other infectious diseases of the female lower genital tract.