AN ESTIMATED 15 MILLION new episodes of STDs occur annually in the United States alone. 1 Infections due to Chlamydia trachomatis are among the most common, resulting in approximately 4 to 5 million new cases each year in the United States. 2 Nearly 70% of genital chlamydial infections in women are asymptomatic, 3 and many of these infections go untreated. Untreated infections can result in serious reproductive tract complications, including pelvic inflammatory disease, ectopic pregnancy, and infertility. For this reason, C trachomatis is a leading cause of preventable infertility worldwide. 4
Potent, nonirritating, broad-spectrum vaginal microbicides are needed to combat the pandemics of unintended pregnancies and STDs. Although many candidate microbicides are active against Chlamydia in vitro, very few have been tested in vivo. Moreover, those that have been tested in vivo often yield results that do not correlate with their in vitro activities. For instance, tests in vitro demonstrate that N9 does not inactivate C trachomatis, 5–7 but in well-controlled clinical trials N9 products provide women with partial protection against chlamydial infections. 8,9 Similarly, in an earlier study on genital herpes (HSV-2), we found that the in vitro activities of several candidate vaginal microbicides against HSV-2 differed from their abilities to prevent vaginal transmission of HSV-2 infections in the mouse. 10 Here we present results of testing several candidate microbicides for preventing transmission of a human oculogenital serovar of C trachomatis in a mouse vaginal challenge model. We selected agents with known in vitro activity against multiple STD pathogens or against Chlamydia specifically. Although N9 has been shown not to inactivate Chlamydia, it was included in this study since previous reports indicate that it provides partial protection against vaginal chlamydial infections in both women and mice.
The recent results of large N9 clinical studies 11–13 revealed that N9 does not protect commercial sex workers from infection with HIV. These studies suggest that N9 may cause disruptions in the vaginal epithelium, which may increase the risk of acquiring HIV. Furthermore, a recent study by Fichorova et al. showed that N9 stimulates secretion of proinflammatory cytokines that might promote HIV infections by recruiting target cells to the site of viral entry. 14 These findings emphasize how critical it is that vaginal microbicides be nontoxic and that screening of candidate microbicides in animal models should be done both for toxicity as well as efficacy against HIV/STD pathogens.
The traditional animal toxicity model is the rabbit vaginal irritation model. However, this model, as currently practiced and scored, failed to predict the vaginal toxicity caused by N9 in humans. Furthermore, this model is typically used relatively late in the development of a candidate microbicide. One goal of this research was to create an in vivo animal toxicity model appropriate for early screening that would be highly sensitive, rapid, economical, and more predictive of human toxicity than in vitro cytotoxicity tests or the rabbit vaginal irritation model as commonly practiced. Here we report results of testing candidate microbicides in the progestin-sensitized mouse model for acute toxicity to genital tract columnar epithelium, tissue similar to human cervical columnar epithelium, which is the site of infection for many STD pathogens, including Chlamydia trachomatis.
Materials and Methods
Six- to 8-week-old specific pathogen-free outbred CF-1 female mice were purchased (Charles River Breeding Laboratories, Wilmington, MA) and housed for at least 1 week without manipulation before being used experimentally. All experiments reported here were performed with protocols approved by the Animal Care and Use Committee at The Johns Hopkins University. The mice were anesthetized only for acute toxicity tests since the Chlamydia challenge procedures caused no signs of discomfort and CF-1 mice posed no difficulty in handling. Mice were pretreated with 2.5 mg of medroxyprogesterone acetate (Depo-Provera; UpJohn, Kalamazoo, MI), administered subcutaneously in the hind quarters 7 days before inoculation with C trachomatis. Treatment with this long-acting progestin has been shown to increase the susceptibility of mice to vaginal infection, 15,16 most likely because this treatment renders the entire vaginal squamous epithelium into a monolayer of columnar epithelium. Similarly, the cervix of women is a columnar epithelium.
C trachomatis serovar D (Ct-D) obtained from the American Type Culture Collection (ATCC VR-885; ATCC, Rockville, MD) was propagated in McCoy cells (ATCC). A stock was stored in aliquots at −86 °C, and material from this stock was used in all experiments described. The 50% vaginal infectious dose (ID50) was determined experimentally to be 104 ifu (Figure 1).
The following compounds are suitably formulated for vaginal use in women and thus were used neat in mice: BufferGel and neutralized BufferGel (supplied by ReProtect, LLC, Baltimore, MD), composed of a high-molecular-weight cross-linked polyacrylic acid gel formulated to pH 3.9 and 6.0, respectively; Gynol II and Extra Strength Gynol II (Ortho Pharmaceutical, Raritan, NJ), over-the-counter spermicides containing 2% and 3% nonoxynol-9, respectively; Liqua-Gel (Paddock Laboratories, Minneapolis, MN), a methylcellulose/propylene-glycol lubricating gel; and 0.5% chlorhexidine digluconate (chlorhexidine) gel and an otherwise identical placebo gel containing no chlorhexidine (both supplied by Advanced Care Products, Raritan, NJ).
Concentrated aqueous solutions of the following compounds were diluted in Liqua-Gel, an infection-neutral gel vehicle: 500,000 molecular weight (MW) polystyrene sulfonate (Scientific Polymer Products, Ontario, NY); κ-carrageenan (FMC Corporation, Rockland, ME); chlorhexidine digluconate (chlorhexidine; Sigma, St. Louis, MO); 5,000-MW dextran sulfate (Sigma); 500,000-MW dextran sulfate (Sigma); nonoxynol-9 (Rhône-Poulenc, Cranbury, NJ); and sodium dodecyl sulfate (J. T. Baker, Phillipsburg, NJ).
The following lectins were used at a concentration of 0.1% (w/v) in saline (0.9% NaCl): Concanavalin A (Con A; Calbiochem, La Jolla, CA), wheat germ agglutinin (WGA; Sigma), and Phaseolus vulgaris agglutinin (PHA-L; Sigma).
Neem pessaries (PR-075) were provided as solid, 2.5-g vaginal suppositories by G. P. Talwar (International Center for Genetic Engineering and Biotechnology, New Delhi, India). A single pessary was dissolved in 1.0 ml of saline at 37 °C before use.
Inoculation of Mice
Mice were inoculated 7 days after progestin treatment. A 50-μl dose of the test microbicide was delivered to the vagina with a 50-μl fire-polished glass Wiretrol pipette (Drummond Scientific, Broomall, PA). Before withdrawal of the pipette from the vagina, the pipette was stirred with both circular and in-and-out motions to distribute the test compound over the vaginal surface and to mimic the stirring action of human coitus. Immediately following delivery of the test microbicide, 105 ifu of Ct-D suspended in 10 μl of sucrose-phosphate transport medium (2SP) were delivered into the vagina with a 10-μl fire-polished glass Wiretrol pipette, which was moved in and out five times to mimic human coitus. Control animals were either untreated or treated with Liqua-Gel, saline, or a placebo gel (Advanced Care Products) as appropriate. Note that this procedure differs from that described in our prior publication of this mouse model 17 in that the mice were not anesthetized.
In two experiments, the 50-μl dose of test microbicide was premixed with the 105 ifu Chlamydia inoculum and incubated for 10 minutes before delivery of the mixture vaginally.
Assay for Ct-D Infection
The presence of Ct-D in the lower genital tract was determined by culturing the material obtained by swabbing of the vagina (Type 1 Dacroswab; Spectrum Laboratories, Dallas, TX). Swabbing was performed on each mouse on days 4 and 8 after inoculation. All swabs were placed in 2SP and frozen at −86 °C until cultured.
After thawing, one half of the available specimen volume was placed over a confluent monolayer of McCoy cells in each of two duplicate 96-well plates. The plates were centrifuged at 1990 g for 1 hour at 37 °C, and following centrifugation, the inoculum was replaced by fresh McCoy cell culture media supplemented with cycloheximide (1 μg/ml) and D-glucose (4.2 mg/ml).
After 72 hours of incubation, one (primary) plate was iodine-stained, and the duplicate plate was frozen at −86 °C. Specimens with iodine-stained inclusions on the primary plate were scored as positive for chlamydial infection. Specimens found to be negative on the primary plate were selected from the thawed duplicate plate and transferred onto fresh McCoy cell monolayers (secondary culture). The use of this culture amplification step yielded a sensitivity comparable to that of ligase chain reaction assays for detecting chlamydial infection. 18 Any positive result of either primary or secondary culture on one or both days was considered to represent a productive infection.
Susceptibility Tests in Mice With Microbicide-Induced Vaginal Friability
Forty-eight mice were sensitized with progestin treatment 10 and 7 days before inoculation with C trachomatis. Three days before Chlamydia inoculation, 24 of these mice were additionally pretreated with 50 μl of 0.5% chlorhexidine to induce vaginal friability, a state in which the epithelium easily bleeds to mild mechanical stimulation (e.g., when gently swabbed). On the day of inoculation, 12 mice that had been pretreated with chlorhexidine and 12 that had not were inoculated with 102 ifu of Ct-D suspended in 10 μl of 2SP delivered with a 10-μl fire-polished glass Wiretrol pipette that was moved in and out five times to mimic coitus. Likewise, the remaining 24 mice were inoculated with 103 ifu of Ct-D suspended in 10 μl of 2SP.
Sensitivity of Elementary Bodies (EBs) to SDS and N9 In Vitro
Stock EBs (6 × 107 ifu/ml) were exposed to SDS (0.0001%–1%) or N9 (2%) for 10 minutes at 37 °C with agitation, followed by either extensive dilution (1:1000) or centrifugation (3×) to remove the detergent. The detergents were removed by dilution so the viability of EBs could be assessed by culture without interference of the detergents with target cells. Control experiments demonstrated that the concentrations of detergents remaining after dilution (≤0.002%) did not reduce the sensitivity of the target cells in detecting EBs.
Test of Acute Toxicity to the Vaginal Epithelium
To evaluate acute toxicity, we used a fluorescent “dead-cell” dye, ethidium homodimer-1, which is impermeant to the plasma membranes of live cells. The nuclei of cells whose membranes have been compromised (by exposure to microbicide, for example) take up the dye and fluoresce brightly. Mice that had been progestin-treated 7 days earlier were anesthetized by inhalation of methoxyflurane (Mallinckrodt Veterinary, Mundelein, IL) and positioned on a ∼20% incline with their heads lower than their tails. Fifty microliters of the candidate microbicide or the control agent (Liqua-Gel or saline), containing 20 μmol/l of ethidium homodimer-1 (Molecular Probes, Eugene, OR), was vaginally delivered without stirring (to minimize potential cell damage) by means of a 50-μl fire-polished glass Wiretrol pipette. The mice were killed 15 minutes later and their vaginas were dissected out, cut longitudinally to open the luminal surface, and rinsed with a stream of saline from a squirt bottle. Each vagina was then compressed between two glass slides to open and flatten folds, and the ends of the slides were taped together to maintain compression. The luminal surfaces of the vaginas were immediately examined (unblinded) with epifluorescence and an Eclipse 800 microscope (Nikon) and FITC/TRITC filter cube. The entire luminal surface was scanned with a low-power (10×) objective. Representative images were captured with a CCD camera (MicroMax; Princeton Instruments, Trenton, NJ) and IP Lab Spectrum software (Signal Analytics Corp., Vienna, VA).
Each treatment group was assigned a staining score, ranging from + to ++++, as compared with the low intrinsic fluorescence of the control group, which was defined as 0 staining. Intensities were assigned on the basis of visual estimates of the fraction of cell nuclei that were stained in the vaginal epithelium: a score of + indicates that only marginal staining above that seen in controls was observed (<5% of the nuclei); ++ indicates scattered regions of staining (5–25%); +++ indicates substantial areas of staining (25–80%); and ++++ indicates that nearly every cell nucleus in large regions of the vaginal epithelium was stained (>80%).
Two of the candidate microbicides, SDS and BufferGel, were found to be incompatible with the dye and therefore could not be premixed with ethidium homodimer-1. Instead, after a 15-minute exposure to microbicide, vaginas were dissected and thoroughly rinsed; then 50 μl of saline containing 20 μmol/l of ethidium homodimer-1 was pipetted over the luminal surface of each vagina and allowed to incubate for 15 minutes before compression between glass slides and imaging.
Statistical significance was determined with use of the Fisher exact test (two-tailed) to compare the number of animals infected between test and control groups during the same experiment (Table 1).
Determining Infectious Dose
A dose–response study was performed to determine an appropriate dose of Chlamydia for vaginal challenge studies (Figure 1). We chose a dose of 105 ifu, approximately 200 times larger than the dose from the urethra of infected human males (∼500 ifu). 19 The 105 ifu dose delivered corresponded to ∼10 intravaginal ID50 in this model and resulted in infection in nearly all animals. In humans, 68% of partners were found to be infected after multiple coital contacts with an infected index case, 20 indicating that the rate of transmission per intercourse must be <68%. This suggests that the usual dose in a human inoculum is ≤1 ID50 Therefore, we reasoned that a 10× ID50 dose would provide a stringent test for these microbicide studies but would not be so high as to overwhelm the protective efficacy of test agents.
The chlamydial inoculum was highly infectious: 54 of 59 untreated mice (92%) became infected. The infection rates in the Liqua-Gel and untreated control groups for individual experiments were a highly reproducible 98% (SD, 4%) and 92% (SD, 12%), respectively.
Microbicide Efficacy Tests
Table 1 shows the results of testing each candidate microbicide for preventing vaginal transmission of Chlamydia in progestin-sensitized mice. The protective efficacy of each candidate is indicated by the relative risk of infection. Relative risk and statistical significance were calculated by comparing the results of each experimental group with its paired control group.
N9 is a nonionic detergent that is the active ingredient in most vaginal spermicides in the United States. Since N9 has been shown previously to protect mice 21 and monkeys 22 from infection by Chlamydia, it was included in this study as a measure of protective efficacy. However, since there are conflicting reports in the literature regarding the sensitivity of EBs to N9 in vitro, 5–7,23–25 we also tested 2% N9 for in vitro activity against our stock of EBs and carefully ensured that the results were not confounded by the action of N9 on the target cells. We found that the viability of EBs was not affected by exposure to N9 (Figure 2).
We tested two over-the-counter formulations of N9 for protective efficacy in mice: Gynol II, containing 2% N9, and Extra Strength (XS) Gynol II, containing 3% N9. N9 was also tested at 3% and 5% by diluting concentrated N9 with Liqua-Gel. All four formulations of N9 provided significant though incomplete protection (Table 1).
Recent in vitro studies indicate that C trachomatis is highly susceptible to chlorhexidine at concentrations that were nontoxic to target cells; they showed complete inactivation immediately upon exposure to 0.0625% chlorhexidine digluconate gel. 26 Furthermore, several studies in mice and monkeys have indicated that chlorhexidine provides protective efficacy against Chlamydia in vivo. 21,27 We tested a high dose of chlorhexidine (10%) in Liqua-Gel as well as a low dose of chlorhexidine (0.5%) supplied in an investigational gel formulation by the manufacturer. We observed no protection at the low dose and significant but incomplete protection at the high dose tested (Table 1). However, at both the low and high concentrations of chlorhexidine tested, we found that the vagina subsequently became friable. Vaginal swabs emerged covered with blood in 75% and 100% of the mice, respectively, when inserted 4 days after delivery of the dose of chlorhexidine. Vaginal friability was never observed with any other test or control agent.
To determine if vaginal friability after a single exposure to chlorhexidine enhances the susceptibility of mice to infection by Chlamydia, we used low-dose inocula, 102 ifu (∼0.01 ID50) and 103 ifu (∼0.1 ID50), in mice treated with 0.5% chlorhexidine 3 days earlier. We found that the 7-day interval between progestin treatment and chlorhexidine treatment was needed to observe a high frequency of chemically induced friability (presumably because columnar epithelium induced by the progestin is more sensitive to chlorhexidine damage). Therefore, we gave 2 doses of progestin (10 days and 7 days before inoculation) to both control and experimental animals so that we could maintain the 7-day interval between progestin treatment and both chlorhexidine pretreatment (experimentals) and C trachomatis inoculation. When vaginal swab specimens were obtained on days 4 and 8 after inoculation (corresponding to 7 and 11 days after exposure to chlorhexidine), 21/24 (88%) and 19/24 (79%), respectively, of the mice pretreated with chlorhexidine exhibited vaginal friability. In contrast, none of the control mice exhibited vaginal friability. As expected, none of the 12 mice in the 102 ifu control group and one of the 12 mice in the 103 ifu control group became infected. However, exposure to a single dose of 0.5% chlorhexidine 3 days before vaginal inoculation with either 102 ifu or 103 ifu C trachomatis resulted in 6/12 mice and 9/12 mice, respectively, becoming infected, indicating a significant (P < 0.007) increase in susceptibility to infection. As can be seen in Figure 1, a single vaginal exposure to 0.5% chlorhexidine elevated the postexposure susceptibility ∼100-fold.
In addition, we tested 0.5% chlorhexidine for protective efficacy using the premix method, in which the dose of microbicide gel was premixed with the Chlamydia inoculum before delivery of the mixture vaginally (Figure 3). In this experiment, we used two additional groups of mice for comparison: one group received the 0.5% chlorhexidine gel vaginally followed by the Chlamydia inoculum (two-step) and the other group received a premixed formulation of Liqua-Gel and Chlamydia. None of the 12 animals in the chlorhexidine premix group became infected, whereas all 12 of animals in the Liqua-Gel premix group became infected (P < 0.00001). As before, 7 of the 12 animals in the chlorhexidine two-step group became infected.
SDS has been shown to be a potent inactivator of both enveloped (HSV-2 and HIV-1) and nonenveloped (bovine papillomavirus) sexually transmitted viruses in vitro. 28 We find that SDS has potent in vitro activity against chlamydial EBs, with >105-fold inactivation when exposed to 0.1% SDS for 10 minutes (Figure 2). We therefore tested several concentrations of SDS for protective efficacy against Chlamydia in our mouse vaginal challenge model and found that 1%, 3%, and 4.5% SDS all gave significant but incomplete protection (Table 1).
SDS (1%) was also tested with use of the premix method. None of the 12 animals in the SDS premix group became infected, whereas 7 of 12 of animals in the Liqua-Gel premix group became infected (P < 0.005).
Sulfated and sulfonated polymers.
Sulfated and sulfonated polymers have been shown to block infection of cultured cells by sexually transmitted pathogens, including Chlamydia. 29,30 We tested the following polymers: κ-carrageenan, low-MW dextran sulfate (5,000 MW), high-MW dextran sulfate (500,000 MW), and polystyrene sulfonate. We observed no protection with any of these sulfated or sulfonated polymers (Table 1). These results are consistent with those of Burillo et al. 31 in a study of other sulfated polysaccharides, all of which failed to protect in a mouse vaginal challenge model.
Lectins have been shown to bind to the major outer membrane protein (MOMP) on the surface of C trachomatis32–34 and to inactivate Chlamydia in vitro. 35,36 Therefore, we hypothesized that lectins could interfere with the binding and attachment steps necessary for cell entry and infection. We tested concanavalin A (ConA, specific for α-D-mannose and α-D-glucose), wheat germ agglutinin (WGA, specific for N-acetyl-D-glucosamine), and red kidney bean leucoagglutinin (PHA-L) for protective efficacy. We observed no protection with any of the lectins tested (Table 1).
BufferGel contains a high-MW polymer of cross-linked polyacrylic acid with substantial pH buffering capacity. It has been formulated to maintain the mild acidity of the human vagina by having sufficient buffering capacity to acidify an ejaculate that otherwise abolishes the acidity of the vagina for several hours. BufferGel has been shown to be protective in a variety of STD animal models. 17 Here we have expanded on earlier published studies and included neutralized BufferGel. BufferGel (pH, 3.9) was tested in five separate experiments, and neutralized BufferGel (pH, 6.0) was tested in two separate experiments (Table 1). BufferGel and neutralized BufferGel both provided significant but incomplete protection.
Tests of Acute Toxicity to the Vaginal Epithelium Following a Brief Exposure to Microbicide
We tested the acute toxicity of all the candidate microbicides that gave significant protection against Chlamydia (N9, SDS, chlorhexidine, BufferGel, and neutral BufferGel). We have previously reported that neem, a candidate microbicide currently being investigated in India, showed strong protective efficacy in this Chlamydia vaginal challenge model. 37 We therefore tested one of the neem formulations, pessary PR-075, for acute toxicity to the vaginal epithelium.
After a single 15-minute exposure to 2% N9 in Liqua-Gel, the lowest dose tested in the vaginal challenge experiments, the nuclei of a major fraction of the columnar epithelial cells became fluorescent (++++ staining), indicating that their membranes had been compromised. Fluorescent staining of epithelial cell nuclei increased in a dose-dependent fashion, beginning at an N9 concentration of ∼0.05% (+ staining). An interesting finding was that in all samples with maximal (++++) staining, narrow stripes of unstained tissue became apparent as the tissue was stretched for microscopic evaluation. The unstained stripes appeared to reveal folded regions of the vaginal epithelium that had not been contacted by microbicide gel. Representative fluorescence images of cell staining for mice exposed to Liqua-Gel (gel-vehicle control) and Liqua-Gel containing 2% N9 are shown in Figures 4A and 4C, respectively.
Gel containing 0.5% chlorhexidine was tested as supplied by the manufacturer. A 15-minute exposure to this low-dose gel also resulted in maximal staining intensity (++++), as seen in Figure 4D.
SDS could not be tested by mixing the ethidium homodimer-1 dye with the SDS gel because the anionic SDS complexed with the cationic dye such that even the rare stained cells normally seen in controls were completely absent. Therefore, an alternative method was used in which the ethidium homodimer-1 was delivered to the vaginal epithelium after the epithelium had already been exposed to SDS for 15 minutes and then thoroughly rinsed. SDS treatment resulted in a dose-dependent epithelial cell staining that became detectable (+) at ∼0.05% SDS and increased to +++ with 1% SDS and ++++ with 4.5% SDS. Representative fluorescence images of cell staining for mice exposed to saline and saline containing 1% SDS are shown in Figures 4B and 4E, respectively.
The neem formulation (PR-075) was tested for toxicity at the same concentration that was previously found protective in this mouse Chlamydia model. 37 We found a major fraction of epithelial cells stained after exposure to this neem formulation (intensity score of ++++) (Figure 4F).
Both BufferGel (pH of 4) and neutralized BufferGel (pH of 6) were also tested for epithelial toxicity with this acute toxicity assay. Like SDS, BufferGel required testing via an alternative method since the polyanionic BufferGel complexed the cationic dye, making it less available for staining. In vaginal epithelia treated with BufferGel or neutralized BufferGel, only a minor fraction of nuclei were stained (++ and +, respectively). Figure 4 (A, G, and H) shows representative images of vaginal epithelia exposed to Liqua-Gel, BufferGel, and neutralized BufferGel.
Several recent studies have led to growing concern that use of the majority of spermicides/microbicides on the market (those that contain N9) may be linked to an increased susceptibility to transmission STD pathogens: HSV 38 and HIV. 14,39 Frequent use of products containing N9 can damage the vaginal epithelium, 40,41 one of the natural protective barriers to disease. The areas of damage might then become portals of entry for HIV/STD pathogens, especially if the product containing N9 is used intensively but inconsistently. Our aim here was to identify nontoxic microbicides that may prevent chlamydial infection in vivo.
The following candidate agents showed protective efficacy that was similar to or greater than that of commercially available products containing N9: 10% chlorhexidine, 4.5% SDS, BufferGel (pH of 4), and neutralized BufferGel (pH of 6). Although chlorhexidine significantly prevented Chlamydia transmission at the highest dose tested (10%), vaginal friability subsequently occurred in 100% of these mice. Mice given 0.5% chlorhexidine also experienced vaginal friability of similar severity and nearly equal frequency. We examined the vaginal epithelium for acute toxicity to a brief exposure to the 0.5% chlorhexidine gel and found a major portion of the vaginal epithelium had been damaged as a result of the exposure. Most important, 3 days after exposure to a single vaginal dose of 0.5% chlorhexidine, long after most of the gel would have been shed from the vagina, the susceptibility of these mice to chlamydial infection was greatly increased. Thus, even though freshly applied chlorhexidine was protective, the damage it causes might subsequently increase susceptibility to STD transmission. These findings raise concern about the use of surface-active agents as vaginal microbicides.
In addition to the acute toxicity caused by chlorhexidine, we found all three classes of surface-active agents (nonionic, cationic, and anionic) to be highly toxic in this sensitive assay of acute columnar epithelium toxicity. In contrast, the controls (saline and Liqua-Gel) were nontoxic, and the nondetergent candidate vaginal microbicide, BufferGel, had only a mild effect on the columnar epithelium. These results are consistent with clinical observations of genital tract toxicity in women after extensive exposure to N9 (showing vaginal disruptions) 41–43 and BufferGel (showing low epithelial toxicity). 44,45 Therefore, this mouse model for acute toxicity may serve as a highly sensitive, rapid, inexpensive initial screen for candidate microbicides.
The data presented here may also offer an explanation to the paradoxical finding that even though N9 does not inactivate Chlamydia in vitro, N9 does provide partial protection from vaginal chlamydial infection to both mice and women. As shown here, brief exposure of the vaginal epithelium to N9 at concentrations used by women damaged >80% of the surface layer of columnar epithelial cells. Therefore, N9 may protect women by killing the target cells for Chlamydia.
Damaging the target cells may provide effective protection when the microbicide is present. However, if a subsequent contact with an infectious pathogen occurred when the detergent was no longer present, but before the epithelium had fully recovered, the epithelium could be more susceptible to infectious entry. Indeed, we observed markedly increased susceptibility to C trachomatis 3 days after a single exposure to a low concentration of chlorhexidine. Similarly, a marked increase in susceptibility to HSV-2 has been observed in mice following a single rectal exposure to N9. 38
Since neutral BufferGel was comparable to acidic BufferGel in preventing infections, this candidate microbicide may protect against Chlamydia via a mechanical barrier mechanism. BufferGel is bioadhesive (it adheres to the epithelium) and has a high yield strength (it resists convective movements). These gel characteristics help prevent pathogens from contacting target cells since the gel creates an unstirred layer that the pathogen must diffuse through to contact target cells. Moreover, polyacrylic acids, such as the Carbopol used in BufferGel, adhere to a wide variety of proteins 46 and may thus trap pathogens within the unstirred layer of gel.
One of the most important findings of this investigation is that none of the candidate microbicides tested here gave complete protection against chlamydial infections, even though two in particular, SDS and chlorhexidine, potently inactivated Chlamydia in vitro and also potently injured major fractions of potential target cells in the vagina. Our results suggest that the microbicide gels must have failed to cover all susceptible surfaces of the vagina, so that chlamydial EBs could still contact vaginal surfaces that were not covered by microbicide. The stripes of unstained epithelia that we observed were located on the interior surfaces of vaginal folds, indicating that the microbicides failed to enter all the infolded surfaces despite the stirring actions performed during vaginal application. Like sunscreens for the skin, microbicides can protect only those areas to which they have been applied, and the infectious inoculum may have flowed into some of the unprotected vaginal infoldings, especially since the viscosity of the watery inoculum was markedly less than the viscosity of the test compounds. This hypothesis is supported by our finding that both SDS and chlorhexidine gave complete protection when they were premixed with the infectious inoculum shortly before being delivered to the vagina.
In contrast with the partial protection against Chlamydia transmission that we observed here, in similar tests of microbicides in our progestin-sensitized mouse model for vaginal transmission of HSV-2, 10 several candidates completely blocked HSV-2 transmission, even though the vaginal distribution of the microbicides was likely similar in both studies. Complete protection against HSV-2 but not Chlamydia might be due to differences in potency (rate of inactivation) of the microbicides on these two pathogens. If so, complete protection against Chlamydia may be achievable with microbicides that act with greater speed (more potent) and/or that persist in the vagina for sufficiently long times to achieve more effective coverage of all susceptible epithelial surfaces. Indeed, long persistence of microbicides in the vagina, perhaps enabling vaginal application hours in advance of coitus, may also be a desirable feature for the women who will ultimately use these products.
In women, the most susceptible site for chlamydial infections is the cervix, not the entire vagina as in the progestin-sensitized mouse. Thus, a strategy to enhance protective efficacy may be to protect the highly susceptible cervix with a vaginal device (diaphragm, cervical cap, etc.) used in combination with a microbicide. In addition to studies of microbicide efficacy and toxicity to epithelia, our results also emphasize the need to investigate dose volumes, formulations, and methods of microbicide delivery that will provide maximal coverage of the highly folded surfaces of the vagina.
1. Cates JR, Cates W Jr. STD prevention in the United States. Lessons from history for the new millennium. Am J Prev Med 1999; 16: 75–77.
2. Centers for Disease Control and Prevention. US Department of Health and Human Services, Public Health Services. Division of STD/HIV Prevention Annual Report. Atlanta: Centers for Disease Control and Prevention, 1996.
3. Stamm WE. Chlamydia trachomatis
infections of the adult. In: Holmes KK, Sparling PF, Mardh PA, et al, eds. Sexually Transmitted Diseases. 3rd ed. New York: McGraw-Hill, 1999: 407–422.
4. Cates W, Brunham RC. Sexually transmitted diseases and infertility. In: Holmes KK, Sparling PF, Mardh PA, et al, eds. Sexually Transmitted Diseases. 3rd ed. New York: McGraw-Hill, 1999: 1079–1087.
5. Kappus EW, Quinn TC. The spermicide nonoxynol-9 does not inhibit Chlamydia trachomatis
in vitro. Sex Transm Dis 1986; 13: 134–137.
6. Herold BC, Kirkpatrick R, Marcellino D, et al. Bile salts: natural detergents for the prevention of sexually transmitted diseases. Antimicrob Agents Chemother 1999; 43: 745–751.
7. Knight ST, Lee SH, Davis CH, Moorman DR, Hodinka RL, Wyrick PB. In vitro activity of nonoxynol-9 on McCoy cells infected with Chlamydia trachomatis
. Sex Transm Dis 1987; 14: 165–173.
8. Rosenberg MJ, Rojanapithayakorn W, Feldblum PJ, Higgins JE. Effect of the contraceptive sponge on chlamydial infection, gonorrhea, and candidiasis: a comparative clinical trial. JAMA 1987; 257: 2308–2312.
9. Louv WC, Austin H, Alexander WJ, et al. A clinical trial of nonoxynol-9 for preventing gonococcal and chlamydial infections. J Infect Dis 1988; 158: 518–523.
10. Zeitlin L, Whaley KJ, Hegarty TA, Moench TR, Cone RA. Tests of vaginal microbicides in the mouse genital herpes model. Contraception 1997; 56: 329–335.
11. Van Damme L. Advances in topical microbicides. In: UNAIDS, ed. XIII International AIDS Conference, July 2000, Durban, South Africa. 2000.
12. Rustomjee R, Abdool Karim Q, Abdool Karim SS, Laga M, Stein Z. Phase 1 trial of nonoxynol-9 film among sex workers in South Africa. AIDS 1999; 13: 1511–1515.
13. Roddy RE, Zekeng L, Ryan KA, Tamoufe U, Weir SS, Wong EL. A controlled trial of nonoxynol 9 film to reduce male-to-female transmission of sexually transmitted diseases. N Engl J Med 1998; 339: 504–510.
14. Fichorova RN, Tucker LD, Anderson DJ. The molecular basis of nonoxynol-9-induced vaginal inflammation and its possible relevance to human immunodeficiency virus type 1 transmission. J Infect Dis 2001; 184: 418–428.
15. Tuffrey M, Taylor-Robinson D. Progesterone as a key factor in the development of a mouse model for genital-tract infection with Chlamydia trachomatis
. FEMS Microbiol Lett 1981; 12: 111–115.
16. Whaley KJ, Barratt RA, Zeitlin L, Hoen TE, Cone RA. Nonoxynol-9 protects mice against vaginal transmission of genital herpes infections. J Infect Dis 1993; 168: 1009–1011.
17. Zeitlin L, Hoen TE, Achilles SL, et al. Tests of BufferGel for contraception and prevention of sexually transmitted diseases in animal models. Sex Transm Dis 2001; 28: 417–423.
18. Achilles S. Microbicides for blocking vaginal transmission of Chlamydia: efficacy and mechanisms of action [dissertation]. Baltimore: The Johns Hopkins University, 2000.
19. Lin JS, Jones WE, Yan L, et al. Underdiagnosis of Chlamydia trachomatis
infection: diagnostic limitations in patients with low-level infection. Sex Transm Dis 1992; 19: 259–265.
20. Quinn TC, Gaydos C, Shepherd M, et al. Epidemiologic and microbiologic correlates of Chlamydia trachomatis
infection in sexual partnerships. JAMA 1996; 276: 1737–1742.
21. Lyons JM, Ito JI Jr. Reducing the risk of Chlamydia trachomatis
genital tract infection by evaluating the prophylactic potential of vaginally applied chemicals. Clin Infect Dis 1995; 21 (suppl 2): S174–S177.
22. Patton DL, Kidder GG, Sweeney YC, Rabe LK, Clark AM, Hillier SL. Effects of nonoxynol-9 on vaginal microflora and chlamydial infection in a monkey model. Sex Transm Dis 1996; 23: 461–464.
23. Benes S, McCormack WM. Inhibition of growth of Chlamydia trachomatis
by nonoxynol-9 in vitro. Antimicrob Agents Chemother 1985; 27: 724–726.
24. Kelly JP, Reynolds RB, Stagno S, Louv WC, Alexander WJ. In vitro activity of the spermicide nonoxynol-9 against Chlamydia trachomatis
. Antimicrob Agents Chemother 1985; 27: 760–762.
25. Patton DL, Wang SK, Kuo CC. In vitro activity of nonoxynol 9 on HeLa 229 cells and primary monkey cervical epithelial cells infected with Chlamydia trachomatis
. Antimicrob Agents Chemother 1992; 36: 1478–1482.
26. Lampe MF, Ballweber LM, Stamm WE. Susceptibility of Chlamydia trachomatis
to chlorhexidine gluconate gel. Antimicrob Agents Chemother 1998; 42: 1726–1730.
27. Patton DL, Sweeney YT, McKay TL, et al. 0.25% chlorhexidine gluconate gel: a protective topical microbicide. Sex Transm Dis 1998; 25: 421–424.
28. Howett MK, Neely EB, Christensen ND, et al. A broad-spectrum microbicide with virucidal activity against sexually transmitted viruses. Antimicrob Agents Chemother 1999; 43: 314–321.
29. Zaretzky FR, Pearce-Pratt R, Phillips DM. Sulfated polyanions block Chlamydia trachomatis
infection of cervix-derived human epithelia. Infect Immun 1995; 63: 3520–3526.
30. Bose SK, Paul RG. Purification of Chlamydia trachomatis
lymphogranuloma venereum elementary bodies and their interaction with HeLa cells. J Gen Microbiol 1982; 128 (Pt 6):1371–1379.
31. Burillo CA, Fontenot JD, Phillips DM. Sulfated polysaccharides block chlamydia infection in vitro, but do not protect mice from vaginal inoculation. Microb Pathog 1998; 24: 253–256.
32. Swanson AF, Kuo CC. Evidence that the major outer membrane protein of Chlamydia trachomatis
is glycosylated. Infect Immun 1991; 59: 2120–2125.
33. Swanson AF, Kuo CC. Identification of lectin-binding proteins in Chlamydia species. Infect Immun 1990; 58: 502–507.
34. Siridewa K, Froman G, Hammar L, Mardh PA. Characterization of glycoproteins from Chlamydia trachomatis
using lectins. APMIS 1993; 101: 851–857.
35. Levy NJ. Wheat germ agglutinin blockage of chlamydial attachment sites: antagonism by N-acetyl-D-glucosamine. Infect Immun 1979; 25: 946–953.
36. Swanson AF, Kuo CC. Binding of the glycan of the major outer membrane protein of Chlamydia trachomatis
to HeLa cells. Infect Immun 1994; 62: 24–28.
37. Talwar GP, Raghuvanshi P, Mishra R, et al. Polyherbal formulations with wide spectrum antimicrobial activity against reproductive tract infections and sexually transmitted pathogens. Am J Reprod Immunol 2000; 43: 144–151.
38. Phillips DM, Zacharopoulos VR. Nonoxynol-9 enhances rectal infection by herpes simplex virus in mice. Contraception 1998; 57: 341–348.
39. Uckun FM, D'Cruz OJ. Prophylactic contraceptives for HIV/AIDS. Hum Reprod Update 1999; 5: 506–514.
40. Patton DL, Kidder GG, Sweeney YC, Rabe LK, Hillier SL. Effects of multiple applications of benzalkonium chloride and nonoxynol 9 on the vaginal epithelium in the pigtailed macaque (Macaca nemestrina). Am J Obstet Gynecol 1999; 180: 1080–1087.
41. Stafford MK, Ward H, Flanagan A, et al. Safety study of nonoxynol-9 as a vaginal microbicide: evidence of adverse effects. J Acquir Immun Defic Syndr Hum Retrovirol 1998; 17: 327–331.
42. Kreiss J, Ngugi E, Holmes K, et al. Efficacy of nonoxynol 9 contraceptive sponge use in preventing heterosexual acquisition of HIV in Nairobi prostitutes. JAMA 1992; 268: 477–482.
43. Roddy RE, Cordero M, Cordero C, Fortney JA. A dosing study of nonoxynol-9 and genital irritation. Int J STD AIDS 1993; 4: 165–170.
44. Mayer K, Peipert J, Fleming T, et al. Safety and tolerability of BufferGel, a novel vaginal microbicide, in women in the United States. Clin Infect Dis 2001; 32: 476–482.
45. van De Wijgert J, Fullem A, Kelly C, et al. Phase 1 trial of the topical microbicide BufferGel: safety results from four international sites. J AIDS 2001; 26: 21–27.
46. Sternberg M, Hershberger D. Separation of proteins with polyacrylic acids. Biochim Biophys Acta 1974; 342: 195–206.