The current impact of the global AIDS epidemic is staggering, with even greater potential to come. Of the more than 14,000 estimated new HIV infections every day, half occur in young persons between the ages of 15 and 24 years, and teenage girls are disproportionately affected compared with boys.1 Attitudes, beliefs, and taboos surrounding sex and the status of women and adolescents complicate efforts to control HIV transmission and to provide appropriate preventive interventions.
According to a recent study that surveyed the opinions of eminent international experts in genome-related technologies and global health issues, female-controlled protection against sexually transmitted infections (STIs) ranked sixth of 10 biotechnologies as holding the greatest potential for improving global health within a decade, especially in developing countries.2 Microbicides, which can be applied topically for the prevention of HIV and other STIs, are one of the most promising preventive interventions, because they could be inexpensive, readily available, and widely accepted.
The first microbicidal product to be clinically evaluated contained Nonoxynol-9 (nonylphenoxypolyethoxyethanol [N-9]), a nonionic surfactant as the active agent.3 N-9 has been used at concentrations between 2% and 12% as the active component of spermicides for more than a quarter of a century. The spermicidal action of N-9 is attributed to solubilization and disruption of the spermatozoa plasma membranes, and it is used in the laboratory for extraction of membrane-bound proteins without denaturation.4,5 Early studies indicated that N-9 was active against a range of microbes and/or pathogens in vitro, including herpes simplex virus (HSV), cytomegalovirus, Neisseria gonorrhoeae, Trichomonas vaginalis, Treponema pallidum, Gardnerella vaginalis, Bacteroides, and Chlamydia trachomatis.6-12 Consequently, its potential for reducing the transmission of STIs had been suggested long before the onset of the HIV epidemic.
;)
A vaginal gel formulation containing a low concentration of N-9 (52.5 mg) was evaluated in a randomized, placebo-controlled, triple-blind phase 2/3 trial at 4 international sites for its effectiveness in preventing HIV infection. Results of the trial indicated that N-9 did not prevent HIV infection when used at low frequency. In addition, N-9 increased the risk of HIV infection in women who used the study gel more than 3 or 4 times a day compared with women in the placebo group.13 Increased susceptibility to HIV infection was associated with a higher incidence of lesions with epithelial disruption, suggesting that N-9, when used frequently, has an adverse effect on the integrity of the vaginal mucosa. This study clearly indicated that N-9 does not deserve further testing as a microbicide. The World Health Organization (WHO)/Contraceptive Research and Development (CONRAD) Technical Consultation on N-9 Report, a panel convened in October 2001, concluded that N-9 should not be used for the purpose of STI or HIV prevention. The wealth of preclinical and clinical data available for N-9 represents an important opportunity to assess the predictive value of this data in relation to the clinical outcome, however.
In this article, we review the in vitro, ex vivo, and animal model data on the safety of N-9 and present a critical analysis of their predictive power based on the results of multiple safety and efficacy trials. Although, ideally, a stepwise progression of microbicide development would begin with in vitro testing and proceed through animal model testing to clinical studies, this is rarely done. Likewise, in the case of N-9, much of the information we know on a preclinical basis was obtained while clinical trials were ongoing.
IN VITRO AND EX VIVO MODELS
The first indication that N-9 might have prophylactic potential against the heterosexual spread of HIV was raised by in 1985 Hicks et al,14 who demonstrated that a 1-minute exposure to 0.05% N-9 or greater blocked infection of HIV-pulsed cells. This observation, together with the known in vitro activity against other STIs, led to the suggestion that “…when HIV prevention is being considered, a protective effect of N-9-containing spermicides should be assessed….”14 Subsequent studies demonstrated that as little as a 30-second exposure to 0.05% N-9 directly inactivated HIV virions.15,16 This provided added optimism that N-9 might be efficacious against HIV transmission: “…the incidence of sexual transmission of infectious agents has been greatly reduced in those who have used spermicides…it is therefore likely that in vitro results can be extrapolated into the clinical setting and that preparations containing N-9 can rapidly inactivate HIV in vivo.”15 Although clinical evidence of genital irritation caused by N-9 products had been reported as early as 1964,17 the initial optimism over the use of N-9 was only dampened when in vivo studies suggested that N-9 might facilitate HIV transmission.18 What in vitro observations might have been predictive of these adverse events, and how could they be used to evaluate new products?
Early in vitro studies demonstrated that N-9 was markedly less active against uropathogenic bacteria than against hydrogen peroxide-producing strains of Lactobacillus, which provide protection against STIs by competitive colonization and maintenance of vaginal acidity (pH 4.5).19 Further studies demonstrated that N-9 products increased adhesion of Candida species20 and Escherichia coli20,21 to mucosal epithelial cells. Taken together, these studies suggested that N-9 products might provide a selective advantage in colonization for uropathogens such as E. coli and Candida, a possible mechanism for the reported enhanced incidence of STI infection in vivo.18 As a consequence of these early studies, all current microbicides in preclinical development are screened for their effect on vaginal microbial flora. In most cases, negative effects are now used as exclusion criteria for further development of a specific product. The screening of compounds to determine if they increase adherence of uropathogens to cervicovaginal epithelium is not included in current testing algorithms but could provide “go/no go” decision points when negative effects on the epithelia are predicted.
N-9 Cytotoxicity
The issue of selectivity for N-9 products was raised in 1994, when a study demonstrated that although N-9 was active against HIV at 0.01%, it was also cytotoxic for lymphocytes at the same concentration.22 Thus, N-9 had a selective or therapeutic index equal to 1, implying that N-9 displayed antiviral activity at doses that were cytotoxic for lymphocytes. The cytotoxic concentration (CC50) for lymphocytes (0.02%) is similar to that required to immobilize spermatozoa in vitro.23 It is interesting to note that with each study on the anti-HIV activity of N-9, the reported selectivity index for N-9 decreased (Table 1). Furthermore, HeLa epithelial cells and cervical explants seem especially susceptible to the cytotoxic activity of N-9, being at least 8 times more sensitive than the virus itself24; this confirms that N-9 displays antiviral activity at doses that are cytotoxic.25 These findings were obtained 6 to 10 years after the initial clinical studies of N-9 for anti-HIV activity.18 In consequence of these findings and because preclinical evaluation of microbicides has become more standardized, microbicides cannot enter clinical trials without extensive data on epithelial toxicity using cervicovaginal cell lines as a minimum criterion.
TABLE 1: In Vivo and Ex Vivo Toxicity of N-9
Further studies of N-9 demonstrated that the level of toxicity increased with duration of exposure; 0.00025% N-9 kills 89% of primary vaginal epithelial cells within 48 hours. In contrast, cells were observed to be relatively resistant during the first 24 hours of exposure. These data suggest that membrane damage may be cumulative and that once a certain threshold is reached, cell death occurs more rapidly. Repeat exposure (2 hours every day for 3 days) also resulted in increased cytotoxicity, in agreement with a cumulative damage hypothesis. Taken together, these results demonstrate that N-9 cytotoxicity is influenced by concentration, duration, and multiple exposure.11,26 The potential for cumulative damage should now be considered an important assessment factor, particularly for membrane-active compounds. Thus, it is no longer sufficient to provide short-term exposure data (hours) to justify the safety of a product.
More recent studies also have demonstrated that N-9 can cause release of proinflammatory cytokines interleukin (IL)-1a, IL-1b, and IL-8. Subtoxic doses of N-9 (10%-30% cell death within 24 hours) resulted in significant release of IL-1a and IL-1b (6 hours after exposure) and secondary release of IL-8 (24 hours after exposure). IL-8 production was dependent on IL-1a activation of the transcription factor nuclear factor (NF)-κB. In contrast, the anti-inflammatory factor secretory leukocyte protease inhibitor (SLPI) was decreased at subtoxic and nontoxic concentrations of N-9.27 Such data demonstrate the potential ability of N-9 to exert an influence on inflammatory processes. Serious consideration is now being given to the potential inflammatory effects of microbicide compounds. Screening of compounds for effects on cytokine induction has now become an important criterion for most preclinical evaluation programs, where compounds are selected on the basis of not having an inflammatory profile.
Damage to Epithelial Surfaces
Most studies have demonstrated the toxicity of N-9 for monolayers of primary epithelial cells, suggesting that direct contact of N-9 with single-layered columnar epithelium of the endocervix or uterus could result in potential damage of these epithelial surfaces. This may not directly reflect the effects of N-9 on stratified epithelium of the vagina and ectocervix, however. This multilayered structure can be 20 to 45 cells thick and is made up of 4 zones (basal, squamous, granular, and cornified).28 As epithelial cells progress in their upward migration from the dividing cells of the germinal basal layer, they become flattened and keratinized with small pyknotic nuclei. Intercellular desmosomes and amorphous lipoidal material within the cornified and granular layers restrict passive diffusion of molecules into the deeper layers of the epithelium.29,30 The pyknotic superficial keratinized cells provide a protective barrier for the underlying viable squamous epithelial cells. Indeed, limited in vivo exposure to N-9 does not lead to extensive lysis of superficial epithelial cells. Ex vivo studies using rabbit vaginal tissue have demonstrated that as little as 0.42% N-9 causes major changes in vaginal permeability, however.31 More recently, we have demonstrated that N-9 increases the permeability of reconstituted stratified vaginal epithelium (data not shown). Thus, N-9 may circumvent protective effects of keratinized cells by modifying epithelial permeability, eventually gaining access to underlying susceptible cells. Any preexisting physical microtrauma or ulceration would further increase the potential access of N-9 to the susceptible squamous epithelial cells, exacerbating preexisting damage. These data suggest that it may be important to screen potential microbicide compounds not only for direct epithelial toxicity but for effects on epithelial permeability (Fig. 1).
FIGURE 1: 1, Superficial cervicovaginal epithelial cells may provide a degree of protection against cytotoxic effects of Nonoxynol-9 (N-9). 2, Repeat or prolonged use may lead to changes in epithelial permeability. 3, Exposure of susceptible squamous epithelial cells to increasing levels of N-9 would cause cell damage and/or lysis, triggering interleukin-1 release. 4, Subsequent release of proinflammatory cytokines and chemokines and upregulation of adhesion molecules in a variety of cell types. 5, Recruitment of polymorphs and activated lymphocytes, with the latter increasing the number of localized target cells for HIV infection. 6, N-9 tissue damage provides incoming HIV (free virus and infected cells) direct access to activated target cells.
N-9 TOXICITY IN ANIMAL MODELS
The failure of N-9 to provide protection against HIV despite the in vitro virucidal activity stimulated interest in the effect of this detergent on cervicovaginal epithelia in animal toxicity models so that we could learn from this “negatively validated” microbicide candidate. N-9 was established as a vaginal spermicide decades before rigorous testing of drugs in animals became routine; thus, little or no animal testing preceded its widespread use by women. Moreover, because it was plausibly less toxic than many other candidate spermicides (eg, mercurial formulations), it was considered safe for vaginal use in women. Thus, when animal testing regimens for vaginal products were developed, N-9 soon became the “negative control,” that is, a comparison product “known” to be well tolerated by women; thus, almost by definition, the degree of toxicity subsequently observed with N-9 in animal models was judged to be acceptable.32
The rabbit vaginal irritation (RVI) test, the most widely used model, consists of 10 daily doses of test agent administered to the rabbit genital tract; histologic sections are then scored for epithelial ulceration, leukocyte infiltration, edema, and vascular injection, with scores for N-9 products ranging between 8 and 12 (16 being the most intense toxicity). A score in this range represents a substantial degree of epithelial ulceration and inflammatory change.33 In light of data from other animal models described elsewhere in this article and increasing data correlating these models with findings in women, the level of toxicity observed after N-9 exposure in the RVI test is now more appropriately becoming a “positive toxicity control,” that is, a marker for a level of toxicity that is unacceptable. The effect of N-9 in the RVI test should no longer be used as a marker for an acceptable level of toxicity, although it is still sometimes presented as such.34,35
The rat has also been used to assess the toxicity of N-9 products, and the histologic findings with N-9 are similar and confirmatory to those seen in the rabbit, with toxicity manifested as acute cervicovaginitis.36 An assessment of the effect of N-9 on an epithelial function, the maintenance of bioelectric potential (a feature of all intact epithelial surfaces), has been carried out in the rat model. The bioelectrical potential of the rat vagina is rapidly abolished after application of 2% N-9.37 Sensitivity to N-9 toxicity was shown to be maximal at diestrus or after ovariectomy. In both conditions, the rodent vaginal epithelium becomes thinned and resembles cervical columnar epithelium. This parallels experience in the RVI test, where detergent-induced toxicity is generally higher in the proximal two thirds of the genital tract, where the epithelium is columnar, and toxicity is less in the lower third, where it is squamous.38 This is relevant to human use of N-9, because columnar epithelium is exposed to N-9 on the face of the ectocervix in women with cervical ectopy. Moreover, N-9 spermicides have recently been shown to ascend into the human endocervix, which is invariably lined with columnar epithelium.39 Thus, it is important to test microbicides on columnar epithelium during preclinical studies, because columnar epithelium is exposed after vaginal application in women.
N-9 and other microbicide candidates have been assessed in primate models, of which the pigtailed macaque is the best developed, having important anatomic and histologic similarities to the human genital tract as well as a similar vaginal flora. Along with its significant strengths, the primate model has limitations. The cost and limited availability of primates limit the size of experimental groups, often limiting adequate statistical analysis. Also, because cervical ectopy is present in only approximately 20% of examinations (Y. Cosgrove-Sweeney, research scientist, personal communication, 2003), the model cannot reliably assess the effect of microbicides on columnar epithelium. Nevertheless, important observations have been reported with this model. A single exposure to N-9 causes colposcopically detectable cervicovaginal irritation,40 and multiple dosing causes progressively more severe epithelial disruption and inflammatory infiltrates on biopsy.41 As in rabbits and rats, cervical epithelium was more susceptible to N-9 toxicity than vaginal epithelium.42
N-9 is cytotoxic at low concentrations in a variety of in vitro assays. in vivo circumstances differ from in vitro systems in important ways; therefore, the results of direct assessments of the in vivo cytotoxic effect of N-9 on sensitive columnar epithelium are of interest. In progesterone-treated mice, the entire cervicovaginal epithelium becomes columnar. In this system, after a single vaginal exposure to 2% N-9, the vaginal epithelium was assessed for cellular cytotoxicity with a membrane-impermeable fluorescent “dead cell” stain.43 Extensive staining of the vaginal epithelium was observed, indicating damage to epithelial cell membranes, a toxicity compatible with the known surfactant mechanism of action of N-9 (Fig. 2).
FIGURE 2: Depo-Provera-treated mice received 20 μL of 2% Nonoxynol-9 (left) or phosphate-buffered saline (PBS) control (right). Ten minutes later, they were lavaged with PBS and then membrane-impermeable dye used to disclose cell death caused by membrane damage. After additional washing, the entire vagina was subsequently photographed using a ×1 objective with an epifluorescent microscope.
Microbicides may compromise epithelial barrier function by direct cell damage, but barrier function may also be compromised by induction of an inflammatory response. Damaged epithelia secrete cytokines and chemokines that can cause an influx of HIV target cells as well as increase expression of HIV coreceptors and replication factors.27 Milligan et al41 reported that mice treated with a single vaginal dose of N-9 had a robust infiltration of macrophages into cervicovaginal fluids that remained detectably elevated for more than 24 hours. The increased number of macrophages is of concern, because they are susceptible targets for HIV infection and they may serve as motile cellular vectors to carry infection across the epithelium.
Recently, data have been reported on the effect of microbicides on increasing the susceptibility to infection by STI pathogens in animal models. Thus, models that can directly measure whether prior exposure to a candidate microbicide can unintentionally increase susceptibility to STI pathogens are of great interest.
In one of those models, Phillips and Zacharova44 tested the effect of N-9 administered rectally to mice 5 minutes before rectal challenge with HSV-2. N-9 failed to protect mice from rectal challenge and substantially increased transmission of infections compared with saline pretreatment. In another report, Abusuwwa et al43 reported that a single vaginal exposure to 2% N-9 12 hours before vaginal challenge with HSV-2 exposure increased susceptibility 17-fold. Finally, Achilles et al45 further generalized the concept that microbicide toxicity may increase susceptibility to STIs in an experiment using a different surfactant microbicide and a bacterial STI: mice treated once with 0.5% chlorhexidine were found to be 100-fold more susceptible to infection when challenged 3 days later with C. trachomatis.
These models allow for the direct detection and quantitation of increased susceptibility to STI pathogens resulting from microbicide toxicity. They should be useful in supplementing, correlating, and validating the surrogate markers for susceptibility (colposcopy, biopsy, and inflammatory markers) that are available in early clinical studies.
As summarized in Table 2, N-9 shows substantial toxicity in a wide variety of genital tract animal models: mice, rats, rabbits, and non-human primates, with single or multiple doses. The toxicity is compatible with surfactant-mediated membrane damage and cytotoxicity and is more severe on columnar epithelium characteristic of the human cervix than on squamous epithelium. There are significant inflammatory changes secondary to this epithelial damage, and the primary damage and secondary inflammation might be expected to increase susceptibility to HIV and other STIs.
TABLE 2: Animal Models to Access the Toxicity of N-9
CLINICAL AND BEHAVIORAL STUDIES
A number of clinical studies have been conducted over the past decade to evaluate the safety of different formulations of N-9 in women. These studies have reported discrepant safety findings depending on the frequency, dose, and duration of N-9 use and timing of sampling. These trials have been invaluable in giving an indication of the predictive power of in vitro and animal testing as indicators of human safety, however. Inflammation has emerged as a key issue of microbicide safety; thus, these clinical studies have also provided critical information on the validity of methods to detect inflammation and the utility of less invasive indicators of inflammatory responses in the vaginal and cervical epithelium.
A randomized placebo-controlled trial of 100 mg of N-9 in a gel formulation was conducted by Stafford et al46 in 40 healthy female volunteers who were sexually abstinent during the study period. The volunteers had a baseline screening evaluation 21 days before product application. Participants applied the N-9 gel for 7 consecutive nights and recorded in a diary any symptoms of vulvar or vaginal soreness or pruritus, vaginal bleeding, or dysuria. The participants were evaluated by histopathologic (biopsy of the vaginal epithelium) and microbiologic methods at baseline after N-9 exposure at 7 and 14 days. Use of N-9 was associated with increased symptoms of irritation and colposcopic and histologic evidence of inflammation in the genital tract compared with placebo-treated women. A few adverse events were also observed in the placebo group. These could be a result of the potential activity of some excipients in the placebo gel. The correlation of symptoms and colposcopic erythema with evidence of inflammation was also assessed in the study, using vaginal biopsy for defining inflammation. Symptoms and colposcopy were found to be poor predictors of vaginal inflammation, as assessed by histologic evaluation of vaginal biopsy specimens. Although N-9 use was associated with increased numbers of symptoms, colposcopic abnormalities, and histologically detected inflammation, none of these findings correlated with each other. Thus, no symptom could be defined as a “gold standard” for assessing safety in human studies.
A randomized study to assess the symptoms and signs of genital irritation among women using different regimens of suppositories containing 150 mg of N-9 was conducted by Roddy et al.47 The study population included 35 healthy female patients at a family planning clinic in the Dominican Republic who were sexually abstinent during the study period. Participants were randomly allocated to 5 study groups: 1 N-9 suppository every other day for 2 weeks, 1 N-9 suppository daily for 2 weeks, 2 N-9 suppositories a day for 2 weeks, 4 N-9 suppositories daily for 2 weeks, and 4 placebo suppositories daily for 2 weeks. Women were examined by colposcopy at baseline and at 1 and 2 weeks after product use. Participants were asked to report any adverse events. There were no significant differences in the subjects' reports of symptoms of irritation between the placebo group and the 4 N-9 groups. Colposcopic evaluation identified that the use of N-9 once or twice a day was shown to double the frequency of epithelial disruption compared that in with placebo group, however. Use of N-9 4 times daily increased the frequency of epithelial disruption approximately 5 times compared with that of placebo use. The authors noted that, “in this dosing study and in the previous pilot study, having symptoms of irritation did not predict well the actual signs of irritation nor did women with obvious clinical signs of irritation always have symptoms to report…”.47 This dose-ranging study was in agreement with Stafford et al's study46 in demonstrating that although there was poor correlation of symptoms with objective signs of epithelial disruption, N-9 use was linked with epithelial disruption.
A randomized controlled trial was subsequently conducted to compare N-9 gel (3.5 g of gel containing 100 mg of N-9) plus condom use versus condom use alone for the prevention of male-to-female transmission of urogenital gonococcal and chlamydial infection. This trial suggested that N-9 gel use was safe and well tolerated during typical use in high-risk women. The study was not masked, however, and no objective measures of genital inflammation were obtained, which may limit the interpretability of the findings relating to safety.
In addition to its direct vaginal application in the form of suppositories or gels, N-9 is used as a condom lubricant. Ward et al48 compared the effect of different levels of N-9 and the relation between N-9 exposure and genital inflammation in a randomized trial of condoms lubricated with polyethylenglycol, 2% (8 mg) N-9 or 4% (16 mg) N-9. The study included 70 commercial sex workers in London who were evaluated by clinical examination, symptoms review, and assessment of vaginal microflora by Gram stain. No association was found between dose of N-9 and symptoms or erythema by examination. Higher cumulative exposure to N-9 resulted in more polymorphonuclear leukocytes (PMNs) on vaginal wall smears, however, indicating that N-9 caused an influx of inflammatory cells into the vaginal fluid. Again, symptoms and direct observation were poor surrogates for detection of inflammatory changes occurring with N-9 use.
Proinflammatory cytokines are generally beneficial and important for clearance of infections. Cervical and vaginal inflammation may enhance HIV transmission and acquisition by compromising the integrity of the physical barrier, recruiting potential viral targets, and enhancing viral replication (Fig. 1), however, as discussed in the section on in vitro and ex vivo models. Damage from inflammation is decreased by anti-inflammatory factors produced by healthy vaginal epithelium, including the SLPI, IL-1 receptor agonist, and soluble tumor necrosis factor (TNF) receptor. The balance between anti-inflammatory and proinflammatory mediators determines whether vaginal inflammation occurs.
Fichorova et al27 conducted a pilot clinical study of proinflammatory cytokines in 10 healthy volunteers at low risk for HIV infection. Cervicovaginal fluid was sampled from the 10 women before the study. One group of 6 women received a single application of a gel containing 150 mg of N-9 (Gynol II), and the second group of 4 women received 3 applications of the same formulation. Both groups were then sampled at 12, 36, and 60 hours during the study. A single application of N-9 did not alter the profile of cytokines, chemokines, and other inflammation regulators in cervicovaginal fluid. Multiple exposures to N-9 resulted in significantly increased levels of proinflammatory cytokines such as IL-1 IL-1α, IL-1β, IL-8, and TNF receptor II, however, and caused diverse effects on anti-inflammatory factors. SLPI concentration levels were consistently decreased when compared with the baseline levels observed in these women. These findings suggested that doses of N-9 not grossly disruptive of the genital epithelium may cause generalized inflammatory responses at the cellular level while decreasing some of the anti-inflammatory factors that protect the epithelium. In addition, injured or irritated epithelial cells release intracellular stores of IL-1α and/or IL-1β IL-1α and IL-1β are efficient inducers of proinflammatory signal transduction pathways, which promote and coordinate the expression of cytokines, chemokines, and adhesion molecules, potentially enhancing HIV transmission and dissemination through recruitment and/or activation of HIV permissive cells.
As discussed in the section on in vitro and ex vivo models, the normal vaginal ecosystem is increasingly recognized as an important host defense mechanism against acquisition of STIs. The interaction of hormonal contraceptives and the vaginal ecosystem is unknown; however, it is known from animal model studies that hormonal contraceptives result in thinning of the vaginal epithelium. To assess the effects of contraceptives on the vaginal microflora, Gupta et al49 conducted a prospective evaluation of women who were initiating use of birth control methods, including the diaphragm with application of N-9 and oral contraceptives. The study showed that oral contraceptives had little effect on the vaginal microbial flora, whereas diaphragms with added N-9 were associated with increased vaginal colonization by E. coli, Enterococcus species, and anaerobic gram-negative rods. These data indicated that N-9 use may also alter the microbial constituents of the vaginal ecosystem. These changes in flora could, in turn, affect the pro- and anti-inflammatory immune milieu in the reproductive tract.
There are several insights in regard to safety that we can glean from clinical trials of N-9 in women. Absence of inflammatory events is a key issue for safety of microbicides. Symptoms and colposcopic evaluation are unreliable predictors of local inflammation. Inflammation resulting from N-9 exposure has been detected by different methods, including number of neutrophils, detection of inflammation by histology, and cytokine-chemokine release. The best methods for detection of inflammatory processes in large trials evaluating safety of microbicides in hundreds or thousands of women are unknown, however. Future designs of trials should pay special consideration to the potentially adverse safety profile of multiple exposures to microbicides.
CONCLUSION
Recent clinical studies have clearly indicated that N-9 should no longer be considered a candidate microbicide. Nevertheless, the wealth of nonclinical and clinical data available from these studies represents an important opportunity to assess the predictive value of these data in relation to clinical outcome. What lessons can we learn from the in vitro, in vivo, and human studies conducted with N-9 that can inform the clinical development of safe microbicides? In vitro studies demonstrate that N-9 displays antiviral activity at doses that are cytotoxic for lymphocytes and primary epithelial cells. This cytotoxicity is influenced by the concentration, duration, and number of exposures. N-9 also shows substantial toxicity in vivo in the genital tract of mice, rats, rabbits, and non-human primates. Data obtained from N-9 studies in animal models correlate well with data that have been accumulating in women exposed to N-9.46,47,50 The epithelial disruption, inflammatory infiltrate in tissue and cervicovaginal secretions, increase in expression of cellular factors that promote HIV replication, and apparent increase in susceptibility to HIV that are observed in women after N-9 use are similar to the physical characteristics observed in those animal models; they argue against the pursuit of N-9 as a microbicide and likewise against the further study of other candidate microbicides that have similar toxicities. Thus, an important lesson learned is that although N-9 has been historically used as a negative control, the level of toxicity observed after N-9 exposure in the RVI test should become a positive toxicity control or marker for an unacceptable degree of toxicity. Additionally, new products that have toxicity profiles similar to N-9 should probably not move forward into clinical trials.
N-9 was never subjected to the extensive in vitro preclinical evaluation that is now required of all candidates before clinical testing. In the current scientific environment, it is highly unlikely that N-9 would have progressed further than early in vitro evaluation.
In vitro, in vivo, and clinical studies have suggested the potential ability of N-9 to exert an influence on inflammatory processes. These findings have started a debate on what is the best method to detect vaginal inflammation in clinical trials of microbicidal candidates. Evaluating cervical lesions has been a routine part of clinical safety assessments for microbicide trials; however, little is known about more subtle changes in the cervical and vaginal epithelium, including induction of subepithelial inflammation and interference with normal host defense mechanisms. The lack of correlation in clinical studies of inflammation, cervical lesions, and symptomology suggests that we need to address this issue further in all future trials.
N-9 has provided us with an invaluable learning experience in microbicide preclinical development. The lessons learned from N-9 studies have and should continue to be in force in the design of protocols to test other candidate microbicides. The most critical aspect of this experience has been to recognize the paramount importance of safety in microbicide development.
REFERENCES
1. Joint United Nations Programme on HIV/AIDS and World Health Organization.
AIDS Epidemic Update: December 2003. Geneva: Joint United Nations Programme on HIV/AIDS and World Health Organization; 2003.
2. Top 10 biotechnologies for improving health in developing countries. In:
Report of the University of Toronto Joint Centre for Bioethics. Toronto: University of Toronto; 2003.
3. Roddy RE, Zekeng L, Ryan KA, et al. A controlled trial of nonoxynol 9 film to reduce male-to-female transmission of sexually transmitted diseases.
N Engl J Med. 1998;339:504-511.
4. Buttar HS, Swierenga SHH, Matula TI. Evaluation of the cytotoxicity and genotoxicity of the spermicides Nonoxynol-9 and Octoxynol-9.
Toxicol Lett. 1986;31:65-73.
5. Helenius A, Simons K. Solubilization of membranes by detergents.
Biochem Biophys Acta. 1975;41:29-79.
6. Singh B, Posti B, Cutler JC. Virucidal effects of certain chemical contraceptives on type 2 herpesvirus.
Am J Obstet Gynecol. 1976;126:422-425.
7. Hermonat PL, Daniel RW, Shah KV. The spermicide nonoxynol-9 does not inactivate papillomavirus.
Sex Transm Dis. 1992;19:203-205.
8. Singh B, Cutler JC, Utidjian HMD. Studies on development of a vaginal preparation providing both prophylaxis against venereal disease, other genital infections and contraception. II. In vitro effects of contraceptive and noncontraceptive preparations on Treponema pallidum and Neisseria gonorrhoeae.
Br J Vener Dis. 1972;48:57-64.
9. Singh B, Cutler JC, Utidjian HMD. Studies on development of a vaginal preparation providing both prophylaxis against venereal disease, other genital infections and contraception. III. In vitro effects of vaginal contraceptive and selected vaginal preparation of Candida albicans and Trichomonas vaginalis.
Contraception. 1972;5:401-411.
10. Jones BM, Willcox LM. The susceptibility of organisms associated with bacterial vaginosis to spermicidal compounds, in vitro.
Genitourin Med. 1991;67:475-477.
11. 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.
12. Bird KD. The use of spermicide containing nonoxynol-9 in the prevention of HIV infection.
AIDS. 1991;5:791-796.
13. Van Damme L, Ramjee G, Alary M, et al. COL-1492 Study Group. Effectiveness of COL-1492, a nonoxynol-9 vaginal gel, on HIV-1 transmission in female sex workers: a randomised controlled trial.
Lancet. 2002;360:971-977.
14. Hicks DR, Martin LS, Getchell JP, et al. Inactivation of HTLV-III/LAV-infected cultures of normal human lymphocytes by nonoxynol-9 in vitro.
Lancet. 1985;2:1422-1423.
15. Malkovsky M, Newell A, Dalgleish AG. Inactivation of HIV by nonoxynol-9.
Lancet. 1988;1:645.
16. Jennings R, Clegg A. The inhibitory effects of spermicidal agents on replication of HSV-2 and HIV-1 in-vitro.
J Antimicrob Chemother. 1993;32:71-82.
17. Rovinsky JJ. Clinical effectiveness of a contraceptive cream.
Obstet Gynecol. 1964;23:125-131.
18. 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.
19. Hooton TM, Fennell CL, Clark AM, et al. Nonoxynol-9: differential antibacterial activity and enhancement of bacterial adherence to vaginal epithelial cells.
J Infect Dis. 1991;164:1216-1219.
20. McGroaty JA, Soboh F, Bruce AW, et al. The spermicidal compound nonoxynol-9 increases adhesion of
Candida species to human epithelial cells in vitro.
Infect Immun. 1990;58:2005-2007.
21. McGroarty JA, Chong S, Reid G, et al. Influence of the spermicidal compound nonoxynol-9 on growth and adhesion of urogenital bacteria
in vitro. Curr Microbiol. 1990;21:219-223.
22. Bourinbaiar AS, Lee-Huang SL. Comparative in vitro study of contraceptive agents with anti-HIV activity: gramicidin, nonoxynol-9, and gossypol.
Contraception. 1994;49:131-137.
23. Harrison C, Chantler E. The effect of nonoxynol-9 and chlorhexidine on HIV and sperm in vitro.
Int J STD AIDS. 1998;9:92-97.
24. Krebs FC, Miller SR, Malamud D, et al. Inactivation of human immunodeficiency virus type 1 by nonoxynol-9, C31G, or an alkyl sulfate, sodium dodecyl sulfate.
Antiviral Res. 1999;43:157-173.
25. Bourinbaiar AS, Fruhstorfer EC. The efficacy of nonoxynol-9 from an
in vitro point of view.
AIDS. 1996;10:558-559.
26. Krebs FC, Millar SR, Catalone BJ, et al. Sodium dodecyl sulfate and C31G as microbicidal alternatives to nonoxynol 9: comparative sensitivity of primary human vaginal keratinocytes.
Antimicrob Agents Chemother. 2000;44:1954-1960.
27. 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.
28. Robboy SJ, Prade M, Cunha G. In: Sternberg SS, ed.
Histology for Pathologists. North Holland, NY: Raven Press; 1992:881-892.
29. Shattock RJ, Griffin GE, Gorodeski GI.
in vitro models of mucosal HIV transmission.
Nat Med. 2000;6:607-608.
30. King BF. The permeability of nonhuman primate vaginal epithelium: a freeze- fracture and tracer-perfusion study.
J Ultrastruct Res. 1983;83:99-110.
31. Acaturk F, Robinson JR. Effect of the spermicide, nonoxynol 9, on vaginal permeability in normal and ovariectomized rabbits.
Pharm Res. 1996;13:950-951.
32. Eckstein P, Jackson MCN, Millman N, et al. Comparison of vaginal tolerance tests of spermicidal preparations in rabbits and monkeys.
J Reprod Fertil. 1969;20:85-93.
33. D'Cruz OJ, Zhu Z, Yiv SH, et al. WHI-05, a novel bromo-methoxy substituted phenyl phosphate derivative of zidovudine, is a dual-action spermicide with potent anti-HIV activity.
Contraception. 1999;59:319-331.
34. Zaneveld LJD, Waller DP, Ahmad N, et al. Properties of a new, long-lasting vaginal delivery systems (LASRS) for contraceptive and antimicrobial agents.
J Androl. 2001;22:481-490.
35. Ballagh SA, Baker JM, Henry DM, et al. Safety of single daily use for one week of C31G HEC gel in women.
Contraception. 2002;66:369-375.
36. Tryphonas L, Buttar HS. Morphologic evidence for vaginal toxicity of Delfen contraceptive cream in the rat.
Toxicol Lett. 1984;20:289-295.
37. Levin RJ. Bioelectric activity as a quantifiable index of acute spermicide (nonoxynol-9) actions on rat vaginal epithelial function during the oestrous cycle.
Pharmacol Toxicol. 1987;60:175-178.
38. Barberini F, De Santis F, Correr S, et al. The mucosa of the rabbit vagina: a proposed experimental model for correlated morphofunctional studies in humans.
Eur J Obstet Gynecol Reprod Biol. 1992;44:221-227.
39. Barnhart KT, Stolpen A, Pretorius ES, et al. Distribution of a spermicide containing nonoxynol-9 in the vaginal canal and the upper female reproductive tract.
Hum Reprod. 2001;16:1151-1154.
40. Patton DL, Kidder GG, Sweeney YC, et al. Effects of nonoxynol-9 on vaginal microflora and chlamydia infection in a monkey model.
Sex Transm Dis. 1996;23:461-464.
41. Milligan GN, Dudley KL, Bourne N, et al. Entry of inflammatory cells into the mouse vagina following application of candidate microbicides.
Sex Trans Dis. 2002;29:597-605.
42. Patton DL, Kidder GG, Sweeney YC, et al. 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.
43. Abusuwwa RF, Hoen TE, Cone RA, et al. Mouse model for quantifying acute cervicovaginal toxicity caused by microbicides upon contact with columnar epithelium [abstract A-176]. Presented at: Microbicides 2002; 2002; Antwerp.
44. Phillips DM, Zacharopoulous VR. Nonoxynol-9 enhances rectal infection by herpes simplex virus in mice.
Contraception. 1998;57:341-348.
45. Achilles SL, Shete PB, Whaley KJ, et al. Microbicide toxicity and efficacy tests in a mouse model of vaginal transmission for Chlamydia trachomatis.
Sex Trans Dis. 2002;29:655-664.
46. Stafford MK, Ward H, Flanagan A, et al. Safety study of nonoxynol-9 as a vaginal microbicide: evidence of adverse effects.
J Acquir Immune Defic Syndr Hum Retrovirol. 1998;17:327-331.
47. Roddy RE, Cordero M, Cordero C, et al. A dosing study of nonoxynol-9 and genital irritation.
Int J STD AIDS. 1993;4:165-170.
48. Ward H, De La Court A, Kitchen V. Noxoynol-9 in lubricated condoms. Results of a study in female prostitutes.
Sex Transm Dis. 1996;23:413-414.
49. Gupta K, Hillier SL, Hooton TM, et al. Effects of contraceptive method on the vaginal microbial flora: a prospective evaluation.
J Infect Dis. 2000;181:595-601.
50. 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.
