HYPOTHETICAL MECHANISMS OF CERVICOVAGINAL TOXICITY BY MICROBICIDES
A microbicide is supposed to be in contact with the cervicovaginal epithelium for a variable amount of time before sexual intercourse. These applications are assumed to be repeated with a variable frequency and interval. As depicted in Figure 3, a microbicide compound may elicit three types of epithelial responses associated with the induction of inflammation. The first, a homeostatic cell response that induces transient changes in the mucosal tissues, does not lead to inflammation. The two other pathological epithelial responses would entail cell hyperactivation, with significant production and secretion of proinflammatory cytokines (e.g. IL-8 and IL-6), and cell injury and death via necrosis or apoptosis, with a massive release of prepackaged IL-1 and other proinflammatory factors. These changes would lead to the recruitment and activation of immune cells that constitute new and susceptible targets for HIV infection.
IN VITRO EVALUATION OF THE PROINFLAMMATORY POTENTIAL OF MICROBICIDE CANDIDATES
On the basis of this model we have studied numerous microbicidal compounds in vitro for their capacity to induce cell toxicity as well as the release of proinflammatory cytokines. We used a human vaginal cell line, VK-2/E6E7, as the target cell system.17 Serial dilutions of compounds were incubated with confluent monolayers of VK-2 cells in 96-well plates for 6 h at 37°C and 5% carbon dioxide. Medium supernatants were assessed for interleukin concentrations using commercial enzyme-linked immunosorbent assay kits (R&D Systems Inc., Minneapolis, MN, USA and Endogen, Rockford, IL, USA), and cells were checked for viability with a 3-(4,5-dimethyl thiazol-2-yl)-2,5-diphenyl tetrazolium bromide assay (Promega, Madison, WI, USA). Using in vitro cytotoxicity and cytokine data we were able to predict the rank order of test agents, from the most to least ‘irritating’, in a 3-day exposure RVI model. Figure 4 shows data on cytotoxicity, IL-1 release (both in vitro and in vivo) and RVI scores of three surface-active microbicidal compounds (nonoxynol-9, sodium dodecyl sulphate, benzalkonium chloride), which illustrate the correlation of these endpoints.15
In order to assess microbicide-induced cytotoxicity, other authors have used human primary vaginal keratinocytes and even cervical explants.18,19 In vitro systems are very useful for an early selection of compounds with low epithelial toxicity and proinflammatory potential. However, as they are inadequate to support a dynamic process of inflammation, they cannot replace in vivo testing during the more advanced stages of compound preclinical development.
A REFINED RABBIT VAGINAL IRRITATION MODEL
Since its description and standardization in the late 1960s, the RVI model has been used to determine the local toxicity of vaginal topical formulations. Its two main end-points are the anatomical and histopathological evaluation of the cervicovaginal mucosa after 10 days of daily compound application. The US Food and Drug Administration recommends the use of this model to assess the local toxic impact of vaginal microbicides. The results of an RVI study will describe the cervicovaginal irritation caused by a compound as minimal, mild, moderate and severe, also indicating whether it is acceptable or not for human use.10 The RVI model presents several advantages over other animal models. It allows for the administration of relatively large volumes of compound (1.0–1.5 ml) without leaking. Because of the simple columnar epithelium of the rabbit's vagina, it also offers higher testing sensitivity, conferring a reasonable margin of safety for the human use of the test compound. Additionally, its results have been validated against the outcome of Phase I clinical trials and extensive post-marketing surveillance.
The RVI model, however, was essentially designed to identify compounds that could cause major irritation and injury of the vaginal mucosa. The advent of the AIDS epidemic and the fact that HIV uses CD4 and other immune cells as portals of entry into the organism have rendered the classical endpoints of the RVI model insufficiently detailed for the proper assessment of a compound's cervicovaginal toxicity. In fact, vaginal formulations with up to 4% nonoxynol-9 have consistently received RVI scores that are ‘acceptable’ for human use.
Because the classic RVI model underestimates the impact of the immunoinflammatory component of the mucosal reaction, we decided to refine the model, expanding the characterization of such a component as well as the compound's effects on the epithelial lining. A secondary objective of this research was to identify soluble and cellular markers of tissue inflammation in CVLs, with the ultimate aim of providing validated endpoints for the non-invasive clinical evaluation of microbicide-induced cervicovaginal inflammation.
Description of the model
Figure 5 shows the endpoints of the refined RVI model. Test compounds are applied intravaginally to rabbits, once a day for 3–10 days, with the aid of a flexible catheter hooked to a 3 cc-disposable syringe. CVLs are collected after the instillation of 5 ml sterile saline (0.9% sodium chloride) with the use of a similar device, at baseline, 24, 48 and 96 h after the last compound application, and at one or two intermediate timepoints. Tissues are collected 24 h after the last dose or at the end of the study.
Intravaginally applied compounds come into contact with the epithelia of vaginal, ectocervical and uterine mucosae, causing a reaction that may lead to a varied degree of inflammation. We propose to characterize such immunoinflammatory reactions by studying the number, phenotype and activation status of the immune cells recruited to the mucosal and submucosal tissues. Activated immune cells typically release cytokines and other soluble factors that may be evaluated in cervicovaginal secretions through swabs or lavages. Appropriately validated soluble markers of inflammation enable a close monitoring of the tissue reaction and may be sufficiently predictive to replace its direct assessment by biopsy.
Soluble and cellular markers of inflammation in cervicovaginal lavages
Three doses, one per day, of a formulation containing 2% nonoxynol-9 and 4.5% carboxymethyl cellulose as vehicle were sufficient to identify the proinflammatory properties of nonoxynol-9.15 CVLs collected from the animals showed a clear and significant increase in IL-1? levels 24 h after the third dose of nonoxynol-9, an increment barely noticeable in the vehicle-treated group. Forty-eight hours after the last dose, the IL-1 concentration in CVLs of nonoxynol-9-treated animals was significantly higher than baseline levels (Figure 6).
CVLs are quickly centrifuged (at +4°C) after collection, and the cell pellet is separated from the supernatant. Whereas the latter is used to quantify soluble markers of inflammation such as cytokines, the cells are phenotyped using specific monoclonal antibodies and either an immunocytochemistry technique for microscopic assessment or an immuno-fluorescence technique for fluorescence-activated cell sorter analysis. Figure 7 shows some of the phenotypes being studied. Antibodies against CD25 and nuclear factor kappa B (NF?B), for example, provide important information about the activation status of the cells.
Although the number of CD3+ T cells increased in the CVLs of animals treated with carboxymethyl cellulose vehicle at 24 and 48 h after the third dose, the percentages of these cells in nonoxynol-9-treated animals were twice those of the controls (Figure 8).
Phenotype and activation status of mucosal and submucosal immune cells
Increased levels of cytokines and activated T cells in the CVLs of animals treated for 3 days with a 4% nonoxynol-9 formulation indicated a cervicovaginal immuno-inflammatory reaction, even though the histopathology of the corresponding tissues showed ‘minimal’ irritation and ‘acceptable’ scores. Using specific monoclonal antibodies and an immunohistochemistry technique, we characterized the phenotype and activation status of the leukocytes infiltrating mucosal and submucosal tissues (Figure 9).
Figure 10 shows an example of these findings. Tissues from nonoxynol-9-treated animals revealed 10 times more CD4 cells than those of its vehicle-treated controls. It is noteworthy that CD4 cells are the main targets of HIV infection, therefore their localization in tissues that may come in close contact with the virus represents a clear sign of an increased risk of HIV transmission, regardless of other signs of inflammation or irritation.
Nuclear factor kappa B activation at the epithelial level
One of the unexpected and very interesting findings derived from the immunohistochemical characterization of the cervicovaginal tissues was the differential intensity and localization of NFκB staining in the epithelial cells of nonoxynol-9 and control-treated animals (Figure 11). Nonoxynol-9-treated epithelium showed intense nuclear and cytoplasmic staining (especially in the apical region), which was clearly evident even when the epithelium appeared otherwise normal. Vehicle and saline controls, conversely, revealed pale and diffuse staining. These differences in staining pattern would indicate significantly higher levels of neosynthesized NF?B and its active (nuclear) form in epithelial cells that have come in contact with nonoxynol-9. Being a transcription factor capable of transactivating cytokine and chemokine genes, elevated levels of NF?B in the epithelial cells of the cervicovaginal mucosa may be an early sign of the ensuing immunoinflammatory reaction.
The efficacy of a microbicide depends on the balance between its specific activity and its safety. The experience with nonoxynol-9 provides a good example of a compound with high in vitro activity and poor clinical performance because of underestimated local safety issues. It is of paramount importance to assess the local safety profile of a microbicide candidate as early as possible in the preclinical development. Given that HIV penetrates the cervicovaginal mucosa and infects immune cells, it is crucial to identify compounds that may induce epithelial toxicity and inflammation.
The local toxic effects of test compounds should be progressively characterized using human cervicovaginal cell lines, human cervical explants and animal models. We have described and validated in vitro and in vivo models that fit well with these goals. Microbicide candidates may be assessed for their cytotoxicity and proinflammatory potential using the VK-2/E6E7 cell line early in the discovery and/or preclinical phases, perhaps in combination with the characterization of specific activity. The proposed refined RVI model, which expands the endpoints of the classic RVI model, especially in the areas of mucosal and submucosal immunoinflammatory response and epithelial integrity and reaction, represents a valuable, almost indispensable assay in the evaluation of the local safety profile of an anti-HIV vaginal microbicide. This model has also proved to be a useful tool to further the understanding of microbicide-mucosa interactions as well as to develop non-invasive markers of mucosal inflammation with the potential for clinical use.
The authors wish to thank Ms Christine J. Farrigan for her excellent editorial assistance in the preparation of this manuscript. This work was supported by CONRAD intramural funds from USAID and the Bill and Melinda Gates Foundation to Dr Gustavo F. Doncel and a CONRAD extramural grant to Dr Raina N. Fichorova. The views of the authors do not necessarily reflect those of the funding agencies.
1. The Joint United Nations Programme on HIV/AIDS (UNAIDS) and World Health Organization. AIDS epidemic update: December 2002.
2. Miller CJ, Shattock RJ. Target cells in vaginal HIV transmission. Microbes Infect
3 Mayaud P, McCormick D. Interventions against sexually transmitted infections (STI) to prevent HIV infection. Br Med Bull
, 2001;58: 129–153.
4 Moodley P, Connolly C, Sturm AW. Interrelationships among human immunodeficiency virus type 1 infection, bacterial vaginosis, trichomoniasis, and the presence of yeasts. J Infect Dis
, 2002;185: 69–73.
5. van De Wijgert JH, Mason PR, Gwanzura L, et al. Intravaginal practices, vaginal flora disturbances, and acquisition of sexually transmitted diseases in Zimbabwean women. J Infect Dis
, 2000;181: 587–594.
6. 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
7. Doncel GF, Mauck C, Colvard DS, et al. From conception to contraception. In: DeJonge C, Barratt C, editors. ART and beyond. Cambridge, UK: Cambridge University Press; 2002. pp. 329–354.
8. Turpin JA. Considerations and development of topical microbicides
to inhibit the sexual transmission of HIV. Expert Opin Investig Drugs
9. Mauck C, Doncel G. An update on vaginal microbicides
. Curr Infect Dis Rep
10. Eckstein P, Jackson MC, Millman N, et al. Comparison of vaginal tolerance tests of spermicidal preparations in rabbits and monkeys. J Reprod Fertil
11. 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
12. Van Damme L, Ramjee G, Alary M, et al. Effectiveness of COL-1492, a nonoxynol-9 vaginal gel, on HIV-1
transmission in female sex workers: a randomised controlled trial. Lancet
13. Benes S, McCormack WM. Inhibition of growth of Chlamydia trachomatis
by nonoxynol-9 in vitro. Antimicrob Agents Chemother
14. Jennings R, Clegg A. The inhibitory effect of spermicidal agents on replication of HSV-2 and HIV-1 in vitro. J Antimicrob Chemother
, 1993; 32:71–82.
15. Fichorova R, Bajpai M, Chandra N, et al. Interleukins IL-1, IL-6 and IL- 8 predict mucosal toxicity of vaginal microbicidal contraceptives. Biol Reprod
, 2004; in Press.
Also available at
16. 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.
17. Fichorova RN, Anderson DJ. Differential expression of immuno- biological mediators by immortalized human cervical and vaginal epithelial cells. Biol Reprod
18. Krebs FC, Miller 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
19. Greenhead P, Hayes P, Watts PS, et al. Parameters of human immunodeficiency virus infection of human cervical tissue and inhibition by vaginal virucides. J Virol
Keywords:© 2004 Lippincott Williams & Wilkins, Inc.
Cytokine; HIV-1; inflammation; irritation; microbicides; rabbit vaginal irritation; vaginal cell line