• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of aacPermissionsJournals.ASM.orgJournalAAC ArticleJournal InfoAuthorsReviewers
Antimicrob Agents Chemother. Apr 2005; 49(4): 1509–1520.
PMCID: PMC1068583

Comparative Safety Evaluation of the Candidate Vaginal Microbicide C31G


C31G is currently the focus of clinical trials designed to evaluate this agent as a microbicidal and spermicidal agent. In the following studies, the in vivo safety of C31G was assessed with a Swiss Webster mouse model of cervicovaginal toxicity and correlated with results from in vitro cytotoxicity experiments and published clinical observations. A single exposure of unformulated 1% C31G resulted in mild-to-moderate epithelial disruption and inflammation at 2 and 4 h postapplication. The columnar epithelium of the cervix was the primary site of damage, while no perturbation of the vaginal mucosa was observed. In contrast, application of unformulated 1.7% C31G resulted in greater levels of inflammation in the cervical epithelium at 2 h postapplication and severe epithelial disruption that persisted to 8 h postapplication. Application of a nonionic aqueous gel formulation containing 1% C31G resulted in no apparent cervicovaginal toxicity at any time point evaluated. However, formulation of 1.7% C31G did not substantially reduce the toxicity associated with unformulated C31G at that concentration. These observations correlate with findings gathered during a recent clinical trial, in which once-daily applications resulted in no adverse events in women receiving the formulation containing 1% C31G, compared to moderate-to-severe adverse events in 30% of women receiving the 1.7% C31G formulation. The Swiss Webster mouse model was able to effectively discriminate between concentrations and formulations of C31G that produced distinct clinical effects in human trials. The Swiss Webster animal model may be a highly valuable tool for preclinical evaluation of candidate vaginal microbicides.

Transmission of human immunodeficiency virus type 1 (HIV-1) continues to be a major global health concern, having resulted in an estimated 5 million new infections and 3 million deaths in 2003 alone (3). An estimated 80 to 90% of new HIV-1 infections occur via heterosexual intercourse. Furthermore, females may be more susceptible to HIV-1 infection than males are (25). These aspects of the HIV-1 epidemic, combined with the lack of preventative options, have created an urgent need for a female-controlled method to prevent HIV-1 infection. The development of microbicides has emerged as one of the leading strategies to reduce the risk of HIV-1 transmission, with more than 60 compounds now in preclinical and clinical development for use as topical microbicides.

Early approaches to topical microbicide development focused on nonoxynol 9 (N-9), a commercially available spermicide that has been widely used for more than 40 years. Clinical trials evaluating the efficacy of N-9 for use as a microbicide demonstrated that N-9 failed to reduce the incidence of HIV-1 infection (22, 24, 29), despite its potent anti-HIV-1 activity in vitro (13, 21). N-9 use was associated with genital inflammation and lesions (19, 23, 26) and an increased incidence of HIV-1 infection in frequent users of an N-9-containing vaginal gel (28). In addition, clinical studies evaluating the impact of sexually transmitted disease (STD) prevalence on the incidence of HIV-1 infection suggested that genital inflammation (14, 15) and the integrity of the cervicovaginal epithelium are critical risk factors associated with the incidence of HIV-1 infection (10). These studies clearly emphasized the need to evaluate candidate microbicides in the context of parameters that may increase the risk of HIV-1 transmission during sexual intercourse. To examine these factors, our efforts have focused on the development of a small-animal model for the preclinical evaluation of cervicovaginal toxicity and inflammation associated with microbicide application.

Recent clinical trials of potential N-9-containing microbicidal products have raised concerns that disruption of the cervicovaginal epithelium by spermicides or microbicides may increase the susceptibility to HIV-1 infection by providing a direct portal of entry for the virus to subcutaneous tissues and/or by recruiting HIV-1-susceptible immune cells to the genital tissues as part of an inflammatory response to microbicide application (26). These and other detrimental and undesirable consequences of microbicide exposure need to be better understood and documented in order to screen out compounds and formulations that may otherwise perform poorly in clinical trials of safety and efficacy. With at least 60 compounds under development, there is an urgent need for predictive model systems able to expeditiously and accurately evaluate the safety of topical microbicides. To address this need, we used a murine model of cervicovaginal toxicity and inflammation for preclinical screening of candidate vaginal microbicides (6). In these studies, N-9 was used as an example of a topical microbicide with an undesirable safety profile.

One of the compounds currently under investigation as a topical microbicide for the prevention of HIV-1 infection and STDs is C31G. C31G is an amphoteric surfactant containing myristamine oxide and cetyl betaine. C31G has demonstrated in vitro spermicidal activity equivalent to that of the most widely used topical contraceptive agent, N-9 (27). In addition, in vitro studies have shown that C31G has activity against numerous gram-positive and gram-negative bacteria, possesses antifungal properties, and is a potent antiviral agent with activity against herpes simplex virus and HIV-1 (5, 7, 13). Recent clinical trials of C31G-containing formulations have demonstrated that C31G has effective postcoital contraceptive activity (16) and favorable safety profiles with respect to exposure of both female (18) and male (17) genitalia.

Previous studies of the impact of C31G on human cell viability indicated that the in vitro cytotoxicity of C31G is dependent on cell type, exposure duration, and compound concentration (11, 12). The studies presented herein examined the in vivo safety of C31G by assessing toxicity and inflammation associated with introduction of unformulated and formulated C31G into the murine cervicovaginal environment. Results from these animal studies paralleled observations made during recent clinical trials of formulated C31G, suggesting the utility of the mouse model as a preclinical screen of compound safety.



C31G was obtained from Biosyn, Inc. (Huntington Valley, Pa.), as a 10% concentrate (unformulated) and as a nonionic aqueous gel (4) containing either 1 or 1.7% C31G. A placebo gel containing the vehicle alone was also supplied for use as a control. N-9 was obtained from Rhone-Poulenc Rorer (now Aventis), Strasbourg, France. Conceptrol (4% N-9) was obtained as an over-the-counter product (Ortho Pharmaceutical Laboratories, Raritan, N.J.).

Cell lines.

Cell lines were chosen to be representative of the epithelial cell types found in the female reproductive tract. Human ectocervical (Ect1), endocervical (End1), and vaginal keratinocyte (Vk2) cell lines (9), which were previously established by immortalization with human papillomavirus 16/E6E7, were used. HeLa cells (human cervical carcinoma; ATCC CCL-2) were maintained in Dulbecco's modified Eagle's medium (Invitrogen Life Technologies, Carlsbad, Calif.) supplemented with fetal bovine serum (10%), l-glutamine (0.3 mg/ml), and antibiotics (penicillin, streptomycin, and kanamycin at 0.04 mg/ml). End1 and Vk2 cells were maintained in keratinocyte serum-free medium (Invitrogen Life Technologies) supplemented with bovine pituitary extract (50 μg/ml), epidermal growth factor (0.1 ng/ml), penicillin (100 U/ml), and streptomycin (100 μg/ml). The medium was further supplemented with CaCl2 to a final concentration of 0.4 mM.


Female Swiss Webster mice were purchased from Charles River Laboratories (Wilmington, Mass.). Research with animals conformed to the guiding principles in the care and use of animals approved by the Council of the American Physiological Society and was approved by The Pennsylvania State University College of Medicine Institutional Animal Care and Use Committee.

Assessment of vaginal and cervical cell sensitivity to C31G in vitro.

Confluent epithelial cell monolayers of Ect1, End1, Vk2, or HeLa cells, seeded in a 96-well plate at a density of 4 × 104 to 5 × 104 cells per well, were incubated with a range of C31G concentrations (0.000125 to 0.25%) for 10 min, 2 h, 4 h, or 8 h. Following exposure to the test compound, cells were washed and assessed for viability with the CellTiter 96 AQueous nonradioactive cell proliferation assay (Promega, Madison, Wis.). In this assay, 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium is bioreduced by cells into a formazan product that is soluble in tissue culture medium. The absorbance of formazan was measured directly at 490 nm (corrected for background at 690 nm) from 96-well assay plates with an MRX II Revelation microplate photometer (Thermo Labsystems, Vantaa, Finland). The quantity of formazan product is directly proportional to the number of viable cells (data not shown). Results were obtained from two independent experiments that included triplicate wells at each concentration.

Histological evaluation of the cervicovaginal mucosa following exposure to C31G.

The previously described Swiss Webster mouse model (6) was used to evaluate the toxicity and inflammation associated with C31G application. Six- to 10-week old female Swiss Webster mice were hormonally synchronized 7 and 3 days prior to the start of each experiment with a subcutaneous injection of Depo-Provera (Pharmacia and Upjohn Company, Peapack, N.J.) diluted in lactated Ringer's saline solution to a final concentration of 3 mg per mouse. Following synchronization, anesthetized mice received an intravaginal inoculation (60 μl) of the test compound. Untreated mice and mice treated with the diluent (water) or placebo alone were used as controls to evaluate the normal tissue architecture and inflammation status in the cervicovaginal mucosa. Mice were sacrificed at 10 min, 2 h, 4 h, or 8 h following application, and the entire reproductive tract was surgically excised. Tissues were either formalin fixed and embedded in paraffin or frozen by standard procedures.

Gross morphological analyses were performed on tissues stained with hematoxylin and eosin (H&E). Tissue sections were visually examined with an Olympus IX81 microscope to assess the gross morphological condition of the cervicovaginal mucosa. Each micrograph is representative of approximately six to eight sections per mouse across two to four mice per data point in duplicate experiments. Sections on coded slides were scored blindly for epithelial disruption as follows. Mild disruption described localized loss of tissue integrity and epithelial sloughing over less than 5% of the epithelial surface, which was otherwise contiguous and intact. Moderate disruption described multiple areas of epithelial disturbance representing 5 to 25% of the total epithelial surface and small regions of sloughing that exposed the basal cell layer. Severe disruption described large sections of the epithelial surface (>25%) where sloughing, which exposed the basal cell layer, was generally apparent throughout the section.

Immunohistochemical staining to identify inflammatory cell infiltrate.

To visualize all cells of hematopoietic origin (with the exception of erythrocytes), tissues harvested from control and microbicide-treated mice were stained with a rat anti-mouse monoclonal antibody specific to CD45 as previously described (6). To visualize specific neutrophil and CD4-positive cell populations, immunohistochemical analyses were performed on frozen sections because of the lack of an available antibody for staining formalin-fixed, paraffin-embedded sections. Briefly, excised genital tracts were placed in a base mold containing optimal cutting temperature compound (Sakura Finetek, Torrance, Calif.) and plunged into 2-methylbutane prechilled with liquid nitrogen. When blocks were almost solidified, blocked tissues were placed on dry ice and then stored frozen in a −70°C freezer. Tissue sections were then cut to an 8-μm thickness and allowed to adhere to the slide by drying at room temperature. Prior to staining, tissue sections were fixed in cold acetone (−20°C) for 2 min and dried at room temperature for 1 h. The tissue was then rinsed three times in phosphate-buffered saline (PBS) for 5 min each time and incubated at room temperature in blocking buffer (antibody diluent; BD Biosciences) for 10 min. Tissue was again rinsed three times in PBS for 5 min each time. Anti-CD45, anti-Ly-6G, and anti-CD4 primary antibodies (all from BD Biosciences) were diluted (all to 1:20) with antibody diluent (BD Biosciences) and then applied to tissue sections for 1 h at room temperature in a humidified chamber. As a staining control, the corresponding isotype control antibody for each primary antibody was also applied to tissue sections at a concentration equivalent to that of the primary antibody. Following incubation with primary antibody, tissue sections were rinsed with three changes of PBS for 5 min each time. The appropriate biotinylated secondary antibody was then diluted with antibody diluent (BD Biosciences) and applied to tissue sections at room temperature for 30 min as follows: anti-rat immunoglobulin G2b (1:50; BD Biosciences) for anti-CD45 and anti-Ly-6G and polyclonal anti-rat immunoglobulin (1:100; BD Biosciences) for anti-CD4. Addition of Vectastain ABC reagent (Vector Laboratories) and diaminobenzidine (Pierce Biotechnology), mounting, and visualization were performed as previously described (6).

Statistical analyses.

Means and standard deviations of viability index values for each concentration, cell line, and time point were calculated. The concentrations that corresponded to average viability indexes just above and below 0.5 (CC50) were identified. A linear regression analysis (concentration-versus-viability index) was then performed and used to calculate the predicted CC50 and its 95% confidence interval for each concentration, cell line, and time point. Pairwise comparisons for each time point were performed to determine statistical significance. The P value for each pairwise comparison was calculated on the basis of the Wald statistic. All analyses were performed with the PROC REG procedure in SAS (2).


Vaginal and cervical cell sensitivity to C31G in vitro.

Cellular sensitivity to C31G after 10 min, 2 h, 4 h, or 8 h of exposure was measured in cell lines of genital tract origin (Fig. (Fig.1).1). Comparisons of cellular sensitivity were made with CC50s (concentrations of C31G that reduced cellular viability by 50% relative to that of mock-exposed cells) calculated from data presented in Fig. Fig.1.1. HeLa cells were significantly less sensitive (P < 0.0001) to C31G-mediated cytotoxicity than were Vk2, End1, or Ect1 cells at all of the time points examined. However, these differences could be attributed to the presence of serum in the HeLa cell medium and the absence of serum in the medium used to maintain the immortalized epithelial cell lines. Interestingly, End1 cells were particularly sensitive to C31G following short-term (2-h) exposure (Fig. (Fig.1).1). As expected, the sensitivity of all cell lines to C31G increased with concentration and exposure duration.

FIG. 1.
In vitro sensitivities of vaginal and cervical cell lines to C31G. Vaginal (Vk2) and cervical (Ect1 [ectocervical], End1 [endocervical], and HeLa) cell lines were exposed to C31G for 10 min (A), 2 h (B), 4 h (C), or 8 h (D). Cellular viability was determined ...

Assessment of cervicovaginal toxicity and inflammation following application of unformulated preparations of C31G.

To determine if microbicide exposure resulted in differential toxicity and inflammation of the vaginal and cervical mucosa, test compounds were applied intravaginally to Swiss Webster mice and pathological examinations of genital tissues were performed at 10 min, 2 h, 4 h, or 8 h postapplication by morphological and immunohistochemical assessment methodologies. Following intravaginal application of 1% C31G, the gross morphological appearance of the vaginal epithelium and distribution of immune cells throughout the vaginal mucosa were similar to those of control (water-treated) tissue at all of the time points examined. The vaginal epithelium remained protected by a covering of keratin, and the integrity of the squamous epithelium appeared intact. In addition, immune cells remained distributed throughout the submucosa and no signs of mucosal inflammation were observed. Representative tissue sections obtained at 2 h postapplication are presented in Fig. Fig.22.

FIG. 2.
Vaginal mucosa following a single 2-h application of C31G. Swiss Webster mice were inoculated with water or 1% C31G (60 μl) and sacrificed at 2 h postapplication, and genital tract tissue was collected. Representative sections stained with H&E ...

Mice treated with either 1 or 1.7% C31G exhibited no apparent epithelial damage or inflammation within the cervix at 10 min postapplication (data not shown). However, following a 2-h exposure to 1% C31G, very mild disruption of the cervical columnar epithelium was evident and small regions of focused inflammatory infiltrate were observed just below the basal cell layer (Fig. (Fig.3).3). Application of 1.7% C31G also resulted in mild epithelial disruption. However, compared to exposure to 1% C31G, exposure to 1.7% C31G resulted in relatively intense inflammation localized just below the basal cell layer in large regions of the columnar epithelium in the cervix. Mice treated for 2 h with either 1 or 1.7% C31G exhibited less severe epithelial disruption and inflammation than did mice treated with 1% N-9. Application of 1% N-9 resulted in severe disruption of the cervical epithelium and elevated levels of inflammatory infiltrate (compared to controls) localized just below the basal cell layer. Several areas of damage were observed where the columnar cells were stripped off the mucosal surface, exposing the basal cell layer.

Morphological and histological analyses of cervical mucosa following a 2-h exposure to C31G. Swiss Webster mice were inoculated intravaginally with water, 1% N-9, 1% C31G, or 1.7% C31G (60 μl of each) and sacrificed at 2 h postapplication, and ...

Following a 4-h exposure to 1% C31G (Fig. (Fig.4),4), disruption of the columnar epithelium was observed in multiple regions of the cervix. In addition, the number of inflammatory cells present within the cervical mucosa was substantially increased over the levels observed at 2 h postapplication of 1% C31G. Application of 1.7% C31G resulted in severe epithelial disruption at 4 h postapplication. Most of the columnar epithelium was sloughed, and breaks in the basal cell layer exposing the underlying lamina propria were evident. However, the number of CD45-positive cells present within the submucosa was substantially lower than that observed at 2 h and the distribution of these cells was more dispersed, suggesting a diminution of the initial inflammatory response.

FIG. 4.
Epithelial disruption and inflammation of the cervical mucosa following a 4-h exposure to C31G. Swiss Webster mice were inoculated with 60 μl of either 1 or 1.7% C31G and sacrificed at 4 h postapplication, and genital tract tissue was collected ...

Despite decreased levels of inflammatory infiltrate, severe disruption of the cervical epithelium down to and including the basal cell layer was observed at 8 h following application of 1.7% C31G (Fig. (Fig.5).5). In contrast, following an 8-h exposure to 1% C31G, CD45-positive cell numbers and distributions were similar to those observed in control-treated tissue. In addition, epithelial sloughing was no longer apparent and signs of tissue regeneration (columnar epithelial cells with small cytoplasmic content) were evident.

FIG. 5.
Cervical mucosa following an 8-h exposure to C31G. Swiss Webster mice were inoculated with 60 μl of 1 or 1.7% C31G and sacrificed at 8 h postapplication, and genital tract tissue was collected for staining with H&E or anti-CD45 antibody. ...

Cervicovaginal toxicity and inflammation associated with intravaginal inoculation of formulated C31G and N-9 (Conceptrol).

As in previous experiments with unformulated C31G, the vaginal mucosa was relatively insensitive to damage or inflammation mediated by formulated C31G or Conceptrol at all of the time points evaluated (data not shown). In postexposure analyses of cervical tissue, application of the placebo formulation, which contained the formulation vehicle without C31G, resulted in no apparent damage or inflammation throughout the time course examined. A representative tissue section following a 2-h exposure to the placebo formulation is presented in Fig. Fig.6.6. As demonstrated in previous studies (6), application of formulated N-9 (Conceptrol) for 2 h (Fig. (Fig.6)6) and 4 h (Fig. (Fig.7)7) caused epithelial disruption in the cervix and substantial recruitment of immune cells to tissues directly below the basal cell layer. However, despite the higher concentration of N-9 in Conceptrol (4%), the appearance of severe epithelial disruption was delayed relative to damage caused by unformulated N-9.

Cervical mucosa following a 2-h exposure to Conceptrol or formulated C31G. Swiss Webster mice were inoculated with 60 μl of placebo, Conceptrol (4% N-9), or C31G formulated at 1 or 1.7% and sacrificed at 2 h postapplication, and genital tract ...
FIG. 7.
Cervical mucosa following a 4-h exposure to Conceptrol or formulated C31G. Swiss Webster mice were inoculated with 60 μl of Conceptrol (4% N-9) or C31G formulated at 1 or 1.7% and sacrificed at 4 h postapplication, and genital tract tissue was ...

Exposure to 1% formulated C31G for 2 h (Fig. (Fig.6)6) or 4 h (Fig. (Fig.7)7) resulted in no epithelial disruption. However, there were a few small regions of localized inflammation below the basal cell layer at 4 h postapplication that did not coincide with epithelial disruption. In contrast, intravaginal inoculation of 1.7% formulated C31G resulted in a pattern of epithelial disruption and inflammation similar to that observed following application of unformulated 1.7% C31G. A 4-h exposure to 1.7% formulated C31G resulted in sloughing of large regions of columnar epithelium and persistent localization of inflammatory cells directly below the basal cell layer. At 8 h postapplication, severe epithelial disruption of the cervical mucosa was still apparent and immune cells remained elevated relative to those in mock-treated tissues but were dispersed throughout the submucosa (data not shown).

Characterization of inflammatory cell infiltrate in the cervical mucosa following microbicide application.

Immunohistochemical analyses of frozen tissue sections were performed to identify specific immune cell populations within the cervical mucosa following an inflammatory response to either 1% N-9 or 1.7% formulated C31G (Fig. (Fig.8).8). CD45-positive cells were distributed throughout the submucosa and epithelium following a 2-h exposure to water. Staining with cell type-specific antibodies demonstrated that neutrophils (Ly6) represented <5% of the total number of immune cells. In contrast, CD4-positive cells, which may represent a population of HIV-1-susceptible target cells within the female genital tract, were approximately 20% of the total immune cell population and were found distributed throughout the submucosa and epithelium. Application of 1% N-9 resulted in elevated levels of inflammatory cells (CD45), which were distributed throughout the submucosa and localized just below the cervical epithelium. The inflammatory infiltrate responding to N-9 application consisted primarily of neutrophils. CD4-positive cell numbers and distribution did not appear substantially different following N-9 application. In contrast, application of formulated 1.7% C31G resulted in an infiltration of inflammatory cells localized directly below the cervical epithelium. This population of immune cells consisted of a smaller fraction of neutrophils than that observed following N-9 application. Few neutrophils were found throughout the submucosa but were instead localized below the cervical epithelium. The large, subepithelial region of relatively intense staining in the formulated 1.7% C31G Ly6 panel corresponded to the presence of a lymphatic vessel. In contrast to the inflammatory response observed following N-9 application, a dramatic increase in CD4-positive cells was observed following exposure to formulated 1.7% C31G, primarily in regions of severe epithelial disruption. Control staining with isotype antibody and secondary antibody alone verified that the observed staining patterns were specific to the antibodies used in these studies (data not shown).

FIG. 8.
Characterization of inflammatory cell infiltrates in the cervical mucosa. Swiss Webster mice were inoculated with 60 μl of water, 1% N-9, or C31G formulated at 1.7% and sacrificed at 2 h postapplication, and genital tract tissue was collected ...


The continuous mucosal epithelium of the vagina and cervix plays an important role, both as a physical barrier and as an immunological barrier, in preventing entry of bacterial and viral pathogens into host tissues. The integrity of the epithelial barrier and the presence of genital tract inflammation are risk factors associated with an increased incidence of HIV-1 infection, as well as other sexually transmitted infections. Disruption of the mucosal epithelium, specifically, microabrasions or breaks in the basal cell layer, creates a direct portal of entry for HIV-1 and other STD pathogens. Inflammation within the mucosal tissues may result in recruitment of high levels of HIV-1-susceptible cells, concentrating them at sites of potential viral entry. Studies have indicated that genital inflammation resulted in an increased number of macrophages and CD4-positive T lymphocytes within the genital mucosa (15, 26). Therefore, it is critical not only for a vaginal microbicide to directly target an STD pathogen but also for microbicide application to cause little or no inflammation of the cervicovaginal mucosa. The focus of this study was to determine if exposure to the candidate vaginal microbicide C31G results in differential toxicity and inflammation within the vaginal and cervical mucosa relative to the conventional spermicide N-9.

Epithelial disruption and genital inflammation were evaluated following intravaginal inoculation of unformulated or formulated C31G with the Swiss Webster mouse model. These studies produced several important observations. First, these results demonstrated that microbicide-mediated toxicity was localized strictly to the cervical mucosa. This observation was true for all of the microbicidal formulations tested in this study, and these results correlated with our previous observations (6), which indicated that the primary site of N-9-mediated damage in the Swiss Webster mouse model is the columnar epithelium of the cervix. The relevance of this finding was supported by observations made in a clinical trial evaluating the ulceration and irritation caused by application of three N-9-based spermicides. In this study, N-9-mediated damage was localized to the cervical mucosa (20). In contrast, a recent study demonstrated acute vaginal toxicity following microbicide application in a CF-1 mouse model (1). However, a single application of Depo-Provera 7 days prior to microbicide treatment resulted in transformation of the stratified squamous epithelium of the vagina into columnar epithelium similar to that found in the cervix, suggesting that the observed toxicity was related to the change in epithelial morphology. Collectively, these results emphasize the importance of evaluating the cervical epithelium following exposure to vaginal microbicides, particularly after exposures of short duration.

The second important observation of these studies was that inflammatory responses following microbicide application varied with respect to intensity and timing between different microbicide formulations and with different concentrations of the same microbicide. Specifically, the duration and intensity of epithelial disruption and genital inflammation following 1% C31G application were less severe than those found after application of 1% N-9. Electrostatic interactions of the amphoteric surfactant C31G with the cervical mucus may have resulted in the observed reduction in toxicity relative to that of N-9, which is a nonionic surfactant. Furthermore, inflammation associated with C31G exposure was characterized by an infiltration of CD45-positive immune cells localized directly below the basal cell layer. In the 1% C31G test group, the observed inflammation at 2 h postapplication consistently preceded epithelial disruption, which was observed at 4 h postapplication. Although epithelial disruption and inflammation were present by 2 h postapplication of 1% N-9, a similar pattern of inflammation, followed by sloughing of the upper epithelium, was indicated by the localization of epithelial disruption only in regions of the epithelium with elevated levels of inflammatory infiltrate directly below the basal cell layer. These results suggested that the observed toxicity following exposure to both N-9 and C31G is likely due to the release of inflammatory mediators that subsequently cause sloughing of the upper epithelium.

Although the studies described above were confined to single applications of microbicidal agents and formulations, we recognize the necessity of assessing the impact of multiple exposures on epithelial integrity and inflammation, in light of the clinical consequences of multiple N-9 exposures (28). In experiments to examine the effect of multiple exposures to C31G, daily applications of 1% C31G for 5 consecutive days did not result in epithelial damage beyond the levels associated with a single application of the same concentration (data not shown). More extensive studies are planned to further address the effect of multiple microbicide exposures on cervicovaginal integrity and inflammation.

Additional studies are also needed to examine the long-term effects of microbicide application. We previously demonstrated that severe disruption of the cervical epithelium caused by a single application of N-9 (1%) was apparently repaired between 8 and 24 h postapplication (6). Initial experiments with C31G formulated at 1.7% demonstrated similar epithelial recovery at 24 h postapplication (data not shown), suggesting a general capacity for relatively rapid epithelial regeneration in response to microbicide application. The implication of this epithelial repair process is that clinical trials of microbicide safety may underestimate the damage associated with microbicide exposure if patient examinations are conducted after epithelial regeneration has begun. This reconstructive capacity will be examined in further detail in future studies with the mouse model of toxicity.

Nonionic aqueous gels containing either 1 or 1.7% C31G were also examined to determine if the formulation of C31G used has any impact on its toxicity profile. In the murine model, a single exposure to formulated 1% C31G resulted in no apparent genital toxicity over the observed time course (10 min to 8 h). In contrast, moderate-to-severe epithelial disruption and inflammation were observed with exposure to formulated 1.7% C31G. However, the inflammation resulting from exposure to formulated 1.7% C31G was delayed relative to the inflammation resulting from the application of unformulated 1.7% C31G. Comparisons with an N-9-containing formulation indicated that both C31G formulations exhibited less toxicity than the commonly used spermicide Conceptrol. These results demonstrated that the nonionic aqueous gel formulation effectively eliminated any associated toxicity of unformulated 1% C31G but was unable to substantially reduce the observed inflammation and toxicity associated with unformulated 1.7% C31G.

Comparisons between the in vitro and in vivo observations did not indicate substantial correlations between C31G cytotoxicity in cell lines and toxicity in the Swiss Webster mouse model, particularly with respect to the regional origins of the immortalized cervicovaginal cell lines. Vk2 cell sensitivity to C31G was generally similar to the sensitivities of the Ect1 and End1 cell lines at most time points, which is in contrast to the observation that mouse vaginal tissue was relatively impervious to C31G exposure. Furthermore, monolayer cultures of the human-derived vaginal keratinocyte cell line were much more sensitive to the effects of C31G (50% inhibitory concentration, approximately 0.0005 to 0.003%) than were mouse vaginal epithelial tissues, which were unaffected after exposure to a 3-log greater concentration of compound. Several factors likely provide explanations for these observed discrepancies. First, within the vaginal epithelium, vaginal keratinocytes are protected by a layer of keratin that may prohibit interaction of C31G with the epithelium. There is no corresponding protective layer over the cervical epithelium, which may contribute to the increased sensitivity of this tissue to microbicide exposure. Second, vaginal keratinocytes and ectocervical cells are found in a stratified squamous epithelium, as opposed to the single layer of columnar epithelial cells that lines the endocervix. As a result, the columnar epithelium may be inherently more sensitive to microbicide-mediated damage than the stratified squamous epithelium. Finally, vaginal environmental factors may impact the interaction of a compound with the genital epithelium. In addition, localization of these factors to specific regions of the genital tract may differentially impact this interaction. For example, the keratin over vaginal keratinocytes may provide increased protection from microbicide toxicity relative to the cervical mucus covering the columnar epithelium.

Most importantly, recent clinical trials evaluating the safety of multiple formulations of C31G provide support for the use of the Swiss Webster mouse model as a platform for preclinical evaluations of candidate vaginal microbicides. In a phase I randomized, multicenter clinical trial, formulated 1% C31G was the best tolerated and most acceptable formulation following a 3-day, once-daily treatment regimen (4). In addition, a recently completed once- or twice-daily application of either 1 or 1.7% formulated C31G for 14 consecutive days indicated that moderate-to-severe adverse events were not observed in women treated with formulated 1% C31G, while 30% of women treated with the formulation containing 1.7% C31G experienced moderate-to-severe adverse events (18). These clinical findings correlate very well with observations made in the Swiss Webster mouse model, strongly suggesting the accuracy of the model in predicting the outcomes of these safety trials of C31G. These results also suggest that this model may be predictive of the clinical safety of other candidate vaginal microbicides and may help preclude the expenditure of funds and resources on compounds with less than desirable safety profiles.

The development of a small-animal model for drug screening is advantageous for several reasons. A mouse model is inexpensive, allows rapid analyses, and provides representative histological and immunological responses in tissues relevant to those found in the human female reproductive tract. These results suggest that the Swiss Webster mouse model may accurately predict a compound's genital tract toxicity and could be used to complement studies of microbicide toxicity with other animal models, including the rabbit vaginal irritation model (8). Collectively, these results support the utility of the Swiss Webster mouse model for the preclinical evaluation of candidate vaginal microbicides.


These studies were supported by Public Health Service grant PO1 AI37829 (M.K.H., projects 1 and 3A; B.W., project 3C).

We thank Hung-Mo Lin for assistance with the statistical analysis of differences in cell line sensitivity.


1. Achilles, S. L., P. B. Shete, K. J. Whaley, T. R. Moench, and R. A. Cone. 2002. Microbicide efficacy and toxicity tests in a mouse model for vaginal transmission of Chlamydia trachomatis. Sex. Transm. Dis. 29:655-664. [PubMed]
2. Anonymous. 2001. The SAS/STAT user guide, version 8, p. 2875-3026. SAS Institute, Inc., Cary, N.C.
3. Anonymous. 2004. 2004 report on the global AIDS epidemic: 4th global report. Joint United Nations Programme on HIV/AIDS, Geneva, Switzerland.
4. Bax, R., K. Douville, D. McCormick, M. Rosenberg, J. Higgins, and M. Bowden. 2002. Microbicides—evaluating multiple formulations of C31G. Contraception 66:365-368. [PubMed]
5. Calis, S., N. Yulug, M. Sumnu, A. Ayhan, and A. A. Hincal. 1992. A non-antibiotic antimicrobial mixture (C31G): evaluation of the antimicrobial efficiency of C31G on vaginal cultures. Boll. Chim. Farm. 131:335-338. [PubMed]
6. Catalone, B. J., T. M. Kish-Catalone, L. R. Budgeon, E. B. Neely, M. Ferguson, F. C. Krebs, M. K. Howett, M. Labib, R. Rando, and B. Wigdahl. 2004. Mouse model of cervicovaginal toxicity and inflammation for preclinical evaluation of topical vaginal microbicides. Antimicrob. Agents Chemother. 48:1837-1847. [PMC free article] [PubMed]
7. Corner, A. M., M. M. Dolan, S. L. Yankell, and D. Malamud. 1988. C31G, a new agent for oral use with potent antimicrobial and antiadherence properties. Antimicrob. Agents Chemother. 32:350-353. [PMC free article] [PubMed]
8. Fichorova, R. N., M. Bajpai, N. Chandra, J. G. Hsiu, M. Spangler, V. Ratnam, and G. F. Doncel. 2004. Interleukin (IL)-1, IL-6, and IL-8 predict mucosal toxicity of vaginal microbicidal contraceptives. Biol. Reprod. 71:761-769. [PubMed]
9. Fichorova, R. N., J. G. Rheinwald, and D. J. Anderson. 1997. Generation of papillomavirus-immortalized cell lines from normal human ectocervical, endocervical, and vaginal epithelium that maintain expression of tissue-specific differentiation proteins. Biol. Reprod. 57:847-855. [PubMed]
10. Fleming, D. T., and J. N. Wasserheit. 1999. From epidemiological synergy to public health policy and practice: the contribution of other sexually transmitted diseases to sexual transmission of HIV infection. Sex. Transm. Infect. 75:3-17. [PMC free article] [PubMed]
11. Krebs, F. C., S. R. Miller, B. J. Catalone, R. Fichorova, D. Anderson, D. Malamud, M. K. Howett, and B. Wigdahl. 2002. Comparative in vitro sensitivities of human immune cell lines, vaginal and cervical epithelial cell lines, and primary cells to candidate microbicides nonoxynol 9, C31G, and sodium dodecyl sulfate. Antimicrob. Agents Chemother. 46:2292-2298. [PMC free article] [PubMed]
12. Krebs, F. C., S. R. Miller, B. J. Catalone, P. A. Welsh, D. Malamud, M. K. Howett, and B. Wigdahl. 2000. Sodium dodecyl sulfate and C31G as microbicidal alternatives to nonoxynol 9: comparative sensitivity of primary human vaginal keratinocytes. Antimicrob. Agents Chemother. 44:1954-1960. [PMC free article] [PubMed]
13. Krebs, F. C., S. R. Miller, D. Malamud, M. K. Howett, and B. Wigdahl. 1999. Inactivation of human immunodeficiency virus type 1 by nonoxynol-9, C31G, or an alkyl sulfate, sodium dodecyl sulfate. Antiviral Res. 43:157-173. [PubMed]
14. Laga, M., A. Manoka, M. Kivuvu, B. Malele, M. Tuliza, N. Nzila, J. Goeman, F. Behets, V. Batter, M. Alary, et al. 1993. Non-ulcerative sexually transmitted diseases as risk factors for HIV-1 transmission in women: results from a cohort study. AIDS 7:95-102. [PubMed]
15. Levine, W. C., V. Pope, A. Bhoomkar, P. Tambe, J. S. Lewis, A. A. Zaidi, C. E. Farshy, S. Mitchell, and D. F. Talkington. 1998. Increase in endocervical CD4 lymphocytes among women with nonulcerative sexually transmitted diseases. J. Infect. Dis. 177:167-174. [PubMed]
16. Mauck, C. K., M. D. Creinin, K. T. Barnhart, S. A. Ballagh, D. F. Archer, M. M. Callahan, S. W. Schmitz, and R. Bax. 2004. A phase I comparative postcoital testing study of three concentrations of C31G. Contraception 70:227-231. [PubMed]
17. Mauck, C. K., R. G. Frezieres, T. L. Walsh, S. W. Schmitz, M. M. Callahan, and R. Bax. 2004. Male tolerance study of 1% C31G. Contraception 70:221-225. [PubMed]
18. Mauck, C. K., D. H. Weiner, M. D. Creinin, K. T. Barnhart, M. M. Callahan, and R. Bax. 2004. A randomized phase I vaginal safety study of three concentrations of C31G vs. Extra Strength Gynol II. Contraception 70:233-240. [PubMed]
19. Niruthisard, S., R. E. Roddy, and S. Chutivongse. 1991. The effects of frequent nonoxynol-9 use on the vaginal and cervical mucosa. Sex. Transm. Dis. 18:176-179. [PubMed]
20. Poindexter, A. N., III, H. Levine, H. Sangi-Haghpeykar, M. L. Frank, A. Grear, and K. O. Reeves. 1996. Comparison of spermicides on vulvar, vaginal, and cervical mucosa. Contraception 53:147-153. [PubMed]
21. Polsky, B., P. A. Baron, J. W. Gold, J. L. Smith, R. H. Jensen, and D. Armstrong. 1988. In vitro inactivation of HIV-1 by contraceptive sponge containing nonoxynol-9. Lancet i:1456. [PubMed]
22. Richardson, B. A., L. Lavreys, H. L. Martin, Jr., C. E. Stevens, E. Ngugi, K. Mandaliya, J. Bwayo, J. Ndinya-Achola, and J. K. Kreiss. 2001. Evaluation of a low-dose nonoxynol-9 gel for the prevention of sexually transmitted diseases: a randomized clinical trial. Sex. Transm. Dis. 28:394-400. [PubMed]
23. Roddy, R. E., M. Cordero, C. Cordero, and J. A. Fortney. 1993. A dosing study of nonoxynol-9 and genital irritation. Int. J. STD AIDS 4:165-170. [PubMed]
24. Roddy, R. E., L. Zekeng, K. A. Ryan, U. Tamoufe, S. S. Weir, and E. L. Wong. 1998. A controlled trial of nonoxynol 9 film to reduce male-to-female transmission of sexually transmitted diseases. N. Engl. J. Med. 339:504-510. [PubMed]
25. Saracco, A., M. Musicco, A. Nicolosi, G. Angarano, C. Arici, G. Gavazzeni, P. Costigliola, S. Gafa, C. Gervasoni, R. Luzzati, et al. 1993. Man-to-woman sexual transmission of HIV: longitudinal study of 343 steady partners of infected men. J. Acquir. Immune Defic. Syndr. 6:497-502. [PubMed]
26. Stafford, M. K., H. Ward, A. Flanagan, I. J. Rosenstein, D. Taylor-Robinson, J. R. Smith, J. Weber, and V. S. Kitchen. 1998. Safety study of nonoxynol-9 as a vaginal microbicide: evidence of adverse effects. J. Acquir. Immune Defic. Syndr. Hum. Retrovirol. 17:327-331. [PubMed]
27. Thompson, K. A., D. Malamud, and B. T. Storey. 1996. Assessment of the anti-microbial agent C31G as a spermicide: comparison with nonoxynol-9. Contraception 53:313-318. [PubMed]
28. Van Damme, L., G. Ramjee, M. Alary, B. Vuylsteke, V. Chandeying, H. Rees, P. Sirivongrangson, L. Mukenge-Tshibaka, V. Ettiegne-Traore, C. Uaheowitchai, S. S. Karim, B. Masse, J. Perriens, and M. Laga. 2002. Effectiveness of COL-1492, a nonoxynol-9 vaginal gel, on HIV-1 transmission in female sex workers: a randomised controlled trial. Lancet 360:971-977. [PubMed]
29. Wilkinson, D., M. Tholandi, G. Ramjee, and G. W. Rutherford. 2002. Nonoxynol-9 spermicide for prevention of vaginally acquired HIV and other sexually transmitted infections: systematic review and meta-analysis of randomised controlled trials including more than 5000 women. Lancet Infect. Dis. 2:613-617. [PubMed]

Articles from Antimicrobial Agents and Chemotherapy are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...