A Study of the Potential of the Pig as a Model for the Vaginal Irritancy of Benzalkonium Chloride in Comparison to the Nonirritant Microbicide PHI-443 and the Spermicide Vanadocene Dithiocarbamate

Abstract

A porcine model was established to test the mucosal toxicity potential of a thiophene thiourea (PHI-443)-based anti-HIV microbicide and a vanadocene-based spermicide, vanadocene dithiocarbamate (VDDTC) in comparison to benzalkonium chloride (BZK). Nine domestic pigs (Duroc) in nonestrus stage received a single intravaginal application of 2% BZK, 2% PHI-443, or 0.1% VDDTC-containing gel. At various times after gel application, cell differentials and levels of inflammatory cytokines (IL-1β, IL-4, IL-6, IL-8, IL-10, IL-18, IFN-γ, and TNF-α) in cervicovaginal lavage (CVL) fluid were monitored by flow cytometry and ELISA, respectively. Eight pigs were exposed intravaginally to a gel with and without BZK or VDDTC for 4 consecutive days and vaginal tissues were scored histologically for inflammation using a new scoring system. Only CVL fluid from pigs exposed to BZK showed a significant increase of IL-1β, IL-8, and also IL-18 production when compared to the controls, PHI-443 or VDDTC-treated groups. Maximum levels of BZK-induced IL-1β (100-fold), IL-8 (2,500-fold), IL-18 (80-fold), and IFN-γ(10-fold) were found at 24 hours. In the in vivo porcine vaginal irritation model, increased levels of vaginal IL-1β, IL-8, and IL-18 were associated with histological changes consistent with vaginal inflammation. These results demonstrate that key cervicovaginal inflammatory cytokines are useful in vivo biomarkers for predicting the mucosal toxicity potential of vaginal products in the physiologically relevant and sensitive porcine model.

Introduction

An animal model, which can accurately imitate the vaginal irritation potential of spermicides and microbicides with consistent and predictable outcome prior to expensive human clinical trials, is highly desirable. Many animal species have been used as models to predict the in vivo mucosal toxicity. These include monkeys, dogs, guinea pigs, rabbits, rats, and mice (Eckstein et al., 1969; Chvapil et al., 1980; Gray et al., 1984; Kaminsky et al., 1985; Patton et al., 1999, 2004a, 2004b; Milligan et al., 2002; Abdel-Rahman et al., 2004; Catalone et al., 2004). Except for nonhuman primates, these animal models do not mimic the vaginal inflammation that is seen clinically in humans due to their differences in genital tract physiology, anatomy, and histology. The estrus cycle of experimental animals and their genital tract infectivity are important variables affecting the data collected. In addition, due to differential sensitivity of various animal species to topically applied test agents, the toxicological endpoints seem to differ (Kaminsky and Willigan, 1982; Gray et al., 1984). The lack of a specific, sensitive, reliable, reproducible, and physiologically relevant animal model has severely hindered the efforts to conduct standardized, controlled research on spermicide/microbicide preclinical development.

The rabbit has been the standard animal model of vaginal irritancy testing. The simple cuboidal or columnar epithelium is highly sensitive to mucosal irritants when compared to the stratified squamous epithelium of human vagina (D’Cruz et al., 2004b). A predetermined 4-point semiquantitative scoring system is used to assess changes in the continuity of the epithelial lining, edema of the submucosal layer, vascular congestion, and the inflammatory cell infiltrate of this layer after exposure to increasing doses of the test agent for 5 to 20 days (Eckstein et al., 1969; Chvapil et al., 1980; Kaminsky et al., 1985). Because of earlier reports on the correlation between rabbits and humans with respect to the irritation potential of vaginal formulations, the rabbit remains the only FDA recommended model for safety evaluation of vaginal products. However, the rabbit is markedly different from the human in terms of anatomy and cervicovaginal tissues, lacks cyclic reproductive stages, vaginal lactobacilli, vaginal acidity, cervical mucus production, species-specific markers for inflammatory processes, and is unresponsive to most human genital pathogens (Baker, 1998; Noguchi et al., 2003). Furthermore, it is now apparent that the semiquantitative scoring system established by Eckstein et al. (1969) using the rabbit vaginal irritation model for the clinical development of vaginal spermicides might not be rigorous enough for the current development of safe vaginal or rectal microbicides. Frequent use of currently used detergent-type spermicides such as Nonoxynol-9 (N-9) and benzalkonium chloride (BZK) has been shown to be associated with increased risks of vaginal irritation or ulceration (Niruthisard et al., 1991; Klebanoff, 1992; Stafford et al., 1998). Clinical studies have confirmed that such tissue irritation can actually promote the acquisition and spread of certain sexually transmitted pathogens (Kreiss et al., 1992; Weir et al., 1995; Roddy et al., 1998; Van Damme et al., 2002). Since detergent-based spermicides induce nonclinical irritation and inflammatory responses with potential to facilitate genital transmission of human immunodeficiency virus type-1 (HIV-1) infection, the new criteria for mucosal safety that will lead to successful clinical advancement of potential spermicides and microbicides needs steadfast guidelines.

As a subclinical response to physical or chemical stress, cervicovaginal tissue produce and release inflammatory cytokines including IL-1α, IL-1β, IL-18, TNF-α, IFN-γ, chemotactic cytokines (IL-8 and RANTES), growth promoting factors (IL-6, IL-7, GM-CSF, and TGF), and other signaling factors that rapidly generate vaginal inflammation (Belec et al., 1995; Donders et al., 2003; Simhan et al., 2003; Andrews, 2004; Genc et al., 2004; Kayem et al., 2004). Increased levels of IL-1β, IL-6, TNF-α, RANTES, and MIP-1, in cervicovaginal secretions have been associated with HIV-1 infection and bacterial vaginosis (Anderson et al., 1998; Crowley-Nowick et al., 2000; Cauci et al., 2003). Since cervicovaginal tissues synthesize and release a number of these inflammatory mediators, both basally and in response to inflammatory stimuli, measurement of such vaginal responses may allow the evaluation of toxicological properties of a potential spermicide or a microbicide. Topical application of vaginal products might also induce cytokine-mediated events in the vaginal tissue, via either the circulation or secondary induction within the vagina (Hedges et al., 1995).

In recent years, the pig has emerged as a well-characterized and appropriate “large” animal model for studies relevant to inflammatory processes (Biggar et al., 1985; Sullivan et al., 2001). The availability of species-specific, sensitive molecular and immunological probes to study the role played by inflammatory mediators as well as the production of transgenic and knockout models makes the pig one of the most preferred animal models for inflammation research (Wang et al., 2000; Peuster et al., 2004). Moreover, the pig is well suited for colposcopic observations and obtaining multiple biopsies for evaluating the cervicovaginal gene expression profile of inflammatory mediators to specific topical agents. Additionally, physiological factors influencing ovulation, artificial insemination, fertilization, early embryonic development, and establishment of pregnancy in pigs have been well documented (Hunter, 1975, 1977; Pond and Houpt, 1988; Cole and Foxcroft, 1999). Compared to humans, the pig shows similarities in acidic vaginal pH (during estrus), adequate vaginal secretion and cervical mucus production, sperm transport patterns in genital tract, which is of significant importance in contraception studies.

A porcine model was established to test the vaginal inflammation potential of candidate spermicide and microbicide formulations due to many similarities in genital tract physiology and histology compared to human vagina (Pond and Houpt, 1988; Cole and Foxcroft, 1999). Spermicidal organometallic complexes of vanadium(IV) with bis(cyclopentadienyl) rings or vanadocenes are a new class of experimental contraceptive agents (D’Cruz and Uckun, 1998; D’Cruz et al., 1998a, 1998b). Among the 45 vanadocene complexes that were synthesized and tested for human spermicidal activity, vanadocene dithiocarbamate (VDDTC) was identified as the lead candidate for preclinical studies (D’Cruz et al., 2001, 2004c; D’Cruz and Uckun, 2001b). Among the 30 thiourea nonnucleoside reverse transcriptase inhibitors (NNRTI) of HIV-1 that were rationally designed, synthesized, and tested for anti-HIV activity, PHI-443 (N′-[2-(2-thiophene)ethyl]-N′-[2-(5-bromopyridyl)]thiourea), was identified as a broad-spectrum noncontraceptive anti-HIV microbicide (D’Cruz et al., 2000, 2003, 2004a). Both compounds lacked mucosal toxicity in the rabbit vaginal irritation model (D’Cruz and Uckun, 2001b; D’Cruz et al., 2004a).

Using a physiologically relevant porcine model and species-specific reagents, this study further assessed the vaginal inflammatory potential of VDDTC and PHI-443 in comparison to BZK. Two over-the-counter contraceptive sponges use BZK as the active spermicide (Protectaid 6.25 mg as combination of 3 agents and Pharmatex, 60 mg). BZK is a well-documented vaginal irritant in the rabbit and macaque model (Patton et al., 1999; Fichorova et al., 2004). In the porcine model, BZK identified key cervicovaginal inflammatory markers for mucosal inflammation the levels of which correlated with the histologic inflammation scores observed for 10 identifiable biological endpoints that reflect pathologic changes in the epithelial/subepithelial as well as vascular/perivascular compartments.

Materials and Methods

Test Compounds

VDDTC [bis(cyclopentadienyl)N,N-diethyl dithiocarbamato triflate salt] and PHI-443 (N′-[2-(2-thiophenyl-ethyl)]-N′-[2-(5-bromopyridyl)]thiourea), the structures of which are shown in Figure 1, were synthesized according to published procedures (D’Cruz et al., 1998b, 2000, 2004a). Purity was determined by proton and carbon nuclear magnetic resonance spectra, infrared and ultraviolet-visible spectra, mass spectroscopy, and elemental analysis. The final product with a purity of >99% was used for the preclinical studies. BZK (a mixture of quaternary benzyldimethylalkylammonium chlorides) was obtained from Sigma Chemical Co (St. Louis, MO, USA).

Animals

Nine sexually mature (~8 months old; ~130 kg) female Duroc strain pigs (Zierke Co, Morris, MN, USA), free of porcine reproductive and respiratory syndrome virus and pseudorabies virus were used for this study. Duroc breed has been shown to have a haplotype associated with increased serum complement activity and appears to be a more sensitive indicator for inflammation-related research (Mekchay et al., 2003). Pigs were housed in large pens that provided aural, visual, and olfactory contacts. The pigs were given pelleted food (LabDiet 5084, PMI Nutrition International, LLC, Brentwood, MO, USA; 4 kg/day) once daily with access to automatic waterers and were maintained in rooms that were at 22 ± 2°C with relative humidity of 60 ± 20% and a 12-hour fluorescent light cycle. The care, husbandry, and use of all experimental animals used in this study was approved by the Parker Hughes Institute Animal Use and Care Committee, and all animal care procedures conformed to the National Institutes of Health Guide For the Care and Use of Laboratory Animals (NIH Publication No. 85-23, Revised 1996).

Gel Formulation of VDDTC, PHI-443, and BZK

An effective drug solubilization method was developed for the vaginal bioavailability of VDDTC, PHI-443, and BZK in a clinically applicable gel. A microemulsion-based gel formulation for VDDTC (0.1%) composed of Phospholipon 90G and Captex 300 as the oil phase with Pluronic F68 and Cremophor EL as surfactants, propylene glycol and polyethylene glycol 200 as cosurfactants, and xanthan gum as thickener, and water as a carrier (D’Cruz and Uckun, 2001a, 2001b). A nonemulsifying gel formulation for PHI-443 (2%) composed of microcrystalline cellulose, xanthan gum (Xantral 75), and sorbital as suspending agents with macrogol and polysorbate 80 as the surfactants in an aqueous system (D’Cruz et al., 2004a, 2005). VDDTC and PHI-443 were stable in respective gel formulations as determined by analytical HPLC and electron paramagnetic resonance spectra, respectively. A corresponding control gel formulations using the ingredients described for the formulation of VDDTC and PHI-443 was used as the placebo formulation. These gel formulations have been previously shown to be nonirritating to vaginal mucosa following multiple applications in the rabbit and mouse models (D’Cruz and Uckun, 2001a, 2001b; D’Cruz et al., 2001, 20004a). An optically clear and stable formulation of BZK (2%) was developed for intravaginal use. The rationale for selecting a 2% BZK-containing gel was based on a recent comparative rabbit vaginal irritation study of 4 surfactant spermicides that ranked 2% BZK as the most potent vaginal irritant (Fichorova et al., 2004).

Intravaginal Gel Application and CVL Fluid Collection and Processing

Intravaginal gel application and CVL fluid collection was performed during the nonestrus stage of the 21-day porcine reproductive cycle. Nine pigs in subgroups of 3 were administered 80 ml of a gel formulation containing 0.1% VDDTC, 2% PHI-443, or 2% BZK as positive control. These concentrations represented approximately 2,000-times higher than the in vitro spermicidal EC₅₀ values for BZK and VD-DTC or 30,000-fold higher than the in vitro anti-HIV IC₅₀ value for PHI-443. CVL was performed using a rubber spiral tip (Melrose; Zierke Co, USA) or a soft foam tip (Golden Pig; Fox A. I., Cedar, IA, USA) insemination catheter. The catheter tip was lubricated with K–Y jelly (McNeil PPC, Inc, Skillman, NJ, USA) and gently inserted upwards into the vagina until resistance was felt at the opening of the cervix. Using a plastic syringe approximately 80 ml gel was expelled using gentle pressure and drawn into the pig by wavelike muscular contractions and to prevent any back-flow.

CVL fluid was collected similarly before and after gel application using 80 ml of sterile phosphate buffered saline (PBS). CVL fluid was collected during gentle manipulation and gravitational flow back into a sterile ice-cold glass container. To examine the lymphocytic cellular profiles and cytokine levels following intravaginal exposure to BZK, VD-DTC, and PHI-443, CVL fluid was collected at 24 and 48 hours postapplication. The amount of CVL fluid recovered varied between 30 and 60% of the initial lavage fluid instilled. Within 30 minutes of collection, the specimens were transported on ice to the laboratory for processing. Aliquots of recovered CVL fluid were used for flow cytometry and microscopic slide preparation. Samples of each CVL fluid were viewed microscopically to verify the absence of bacterial or red cell contamination. The remainder of the CVL fluid was centrifuged at 400g for 10 minutes at 4°C to remove mucus and cells. The cell-free supernatants were aliquoted, stored at −20°C, and used to measure the cytokine levels.

Flow Cytometric Cell Evaluation

Normal control blood was obtained from the ear vein for gating the flow cytometric scatter profile of specific blood cell types. Fluorescence-activated cell sorter (FACS) analysis of whole blood, density fractionated leukocytes, and control and test CVL fluids was performed by using a FACScan flow cytometer and analysis was done with the Cell-Quest Pro program (both from Becton Dickinson, San Jose, CA, USA). Peripheral blood mononuclear cells were isolated from heparinized blood by density centrifugation using Histopaque-1077 (Sigma Chemical Co, St. Louis, MO, USA) and suspended in PBS. One ml aliquots of freshly collected CVL fluid were evaluated by flow cytometry for cell types based on forward and side scatter profiles. Vaginal epithelial cells, macrophages, granulocytes, lymphocytes, and other leukocytes were identified based on their characteristic scatter profile obtained with whole blood and density gradient purified leukocytes. Cells were gated using forward vs. side scatter to select for granulocytes, monocytes, lymphocytes, and epithelial cells and excluding dead cells and debris. The number of each cell type identified in a field was expressed as a percentage of the total number of cells.

Cytokine Immunoassays

Porcine-specific IL-lβ, IL-4, IL-6, IL-8, IL-10, IL-18, IFN-γ, and TNF-α levels in CVL fluids collected before and after 24 and 48 hours post gel application were measured by a solid phase sandwich ELISA. Samples were assayed for immunoreactive porcine IL-lβ, IL-4, IL-8, IL-10, IFN-γ, and TNF-α using ELISA kits obtained from BioSource International Inc (Camarillo, CA, USA). Porcine-specific IL-6 and IL-18 ELISA kits were obtained from R&D Systems Inc (Minneapolis, MN, USA) and Bender Med Systems, (Burlingame, CA, USA), respectively. In all experiments, CVL fluid from gel-only control was used in parallel. All ELISA kits used porcine-specific monoclonal antibodies as capture antibody and prebiotinylated porcine-specific secondary antibody. Bound biotinylated antibody was detected by the addition of Streptavidin-horse radish peroxidase-conjugate. Tetramethylbenzidine (TMB) plus hydrogen peroxide was used as substrate and the signal was amplified with 2 N sulfuric acid to convert the product from a blue to a yellow color. Absorbance of the oxidized TMB was read at 450 nm versus a substrate blank using a microplate reader (Labsystems Multiskan MS, Labsystems, Helsinki, Finland). The standard curve for the assays ranged from 0–2500 pg/ml for IL-1β, 0–1,000 pg/ml for IL-4, 0–1,250 pg/ml for IL-6, 0–2,000 pg/ml for IL-8, 0–500 pg/ml for IL-10 and IFN-γ, and 0–1500 pg/ml for TNF-α. Cytokine levels in neat and serially diluted CVL fluid were calculated based on respective standard curves. The reported detection limits for the tests were 15 (IL-lβ), 2 (IL-4), 10 (IL-6), 10 (IL-8), 3 (IL-10), 23 (IL-18), 2 (IFN-γ), and 4 (TNF-α) pg/ml, respectively [Bender Med Systems, Biosource and R&D Systems protocol booklets]. Cytokine quantitation was expressed as picograms per milliliter (pg/ml) in CVL fluid. The reported intraassay coefficients of variation were 4, 3.3, 2.8, 7.1, 5.1, 6.8, 4.3, and 2.8%, respectively. The reported interassay coefficients of variation were 8.9, 4.4, 5.2, 8.5, 7.1, 8.1, 7.8, and 3.7%, respectively. Baseline levels of CVL fluid cytokines were measured in 10 CVL samples from untreated and placebo-treated controls.

High Performance Liquid Chromatography (HPLC) Analysis of PHI-443 in CVL Fluid

Chromatographic analysis of PHI-443 in CVL fluid was carried out using a previously established and validated HPLC method (D’Cruz et al., 2004a), which was modified for CVL fluid. Briefly, a 1 ml aliquot of CVL fluid was extracted with acetone and the clarified supernatant was dried under nitrogen gas and re-extracted with 90% aqueous methanol and injected into the HPLC. The HPLC used for these studies was a Hewlett Packard series 1100 instrument equipped with a Chemstation software for data analysis (Agilent Technologies, Palo Alto, CA, USA). The analytical column used was a reverse phase Lichrospher 100, RP-18 (5 μm; 4 × 250 mm) column equipped with a 4 × 4 mm Lichrospher 100, RP-18 (5 μm) guard column and the detection wavelength was set at 275 nm with a slit width of 4 nm. The method employed an isocratic mobile phase of acetonitrile and 0.1% acetic acid (70:30, v/v). The column was equilibrated with the mobile phase prior to data collection. The flow rate was set at 1 ml/minute and the column temperature was set at 20°C. The retention time for PHI-443 was 7.5 minutes.

Porcine Vaginal Irritation Test

For the vaginal irritation study, 8 female pigs in 3 subgroups were randomly assigned to receive 80 ml of gel with and without (placebo control; n = 2) 0.1% VDDTC (n = 3), or 2% BZK (n = 3), which was administered intravaginally via a catheter for 4 consecutive days. Pigs were observed daily for overt clinical signs (genital swelling, redness, and vaginal discharge including bleeding). On day 5, pigs were sedated with telazol (Fort Dodge Animal Health, Fort Dodge, IA, USA; 4 mg/kg im), and euthanized with euthasol (Delmarva Laboratories, Midlothian, VA, USA; 0.2 ml/kg iv). Following euthanization, the genital area was examined macroscopically for swelling, redness, or bleeding. Animals were placed in dorsal recumbency and following a midline abdominal incision, the genital tract was retrieved and the vagina was excised and fixed in 10% neutral-buffered formalin for microscopic evaluation.

Fixed vaginal tissues were trimmed, embedded in paraffin, sectioned at a thickness of 4–6 μm and stained with Harris’ hematoxylin and eosin (H&E) method. Stained sections were examined by light microscopy by a board-certified veterinary pathologist (DE). Evaluations included the thickness of epithelial cell layer; epithelial cell arrangement and type (polyhedral, flattened, proliferative, desquamated, vacuolated, and degree of mitosis); degree of inflammation, necrosis, fibrosis, neutrophil and macrophage infiltration in the epithelial layer; degree of inflammation, necrosis, fibrosis, neutrophil and macrophage infiltration in the lamina propria; hemorrhage, integrity of musculature; squamous and mucinous metaplasia; necrotic mucosal tissue; edema; and vascular congestion.

Based on the pathological changes observed with the positive control (BZK), several (5–8) tissue sections from each specimen were scored blindly for the following 10 histological features: epithelial ulceration; epithelial leukocyte influx; subepithelial leukocyte influx; subepithelial hemorrhage; vascular/perivascular hemorrhage; subepithelial edema; vascular/perivascular edema; vascular/perivascular congestion; muscle integrity; and cell/tissue necrosis. The irritation scores for these features were assigned as follows: 0 = no abnormal finding, 1 = minimal, 2 = mild, 3 = moderate; 4 = marked degree of severity: the cumulative range being 0–40.

Statistical Analysis

Results are expressed as the mean ± SD or SEM values. Statistical significance of the treated group mean with that of control group was analyzed by a 1-way analysis of variance, followed by Dunnett’s multiple comparison test using Graph-Pad Prism version 4.0a software (San Diego, CA, USA). Differences with probability values of <0.05 were considered significant. Correlations between 2 variables were examined using Pearson’s correlation coefficient and linear regression.

Results

Effect of Single Intravaginal Application of BZK, VDDTC, and PHI-443 on Cervicovaginal Cells and Levels of Inflammatory Cytokines CVL leukocyte profile

The cellular profile of PBS-flushed CVL fluid collected before and at 24 and 48 hours after gel application was assessed by flow cytometry. The scatter plots of CVL fluid cells gated based on blood cell population showed marked vaginal granulocyte influx only in BZK-exposed pigs [Figure 2, upper panels]. The pre CVL fluid collected from sexually mature nonestrus pigs lacked significant leukocyte population. However, following exposure to BZK-containing gel, a marked increase in vaginal granulocyte population was apparent at 24 hours as well as 48 hours after gel application. The scatter profiles showed no significant increase in leukocyte population in CVL fluid obtained from all 3 pigs at 24 and 48 hours following intravaginal exposure to either VDDTC or PHI-443-containing gel [Figure 2, middle and lower panels]. Although, CVL fluid cells showed considerable differences in the proportion of nonleukocyte cells, their presence had little effect on the light scatter characteristics of granulocyte populations.

The CVL fluid obtained for 10 baseline controls had a mean percentage of 20.7 ± 11.9% of granulocytes [Figure 3]. No significant increase in vaginal granulocyte population was apparent in CVL fluid obtained for 9 VDDTC-exposed and 6 PHI-443-exposed specimen (23.2 ± 11.5% and 20.8 ± 6.7%, respectively). Similarly, there was no significant increase in baseline levels of monocytes and lymphocytes in these 2 treatment groups. In contrast, significantly enhanced percentage of granulocytes was apparent in all 6 CVL fluids recovered from all pigs exposed to 2% BZK-containing gel (60.7 ± 9.7, p < 0.01). The increased percentage of neutrophils, monocytes, or lymphocytes observed in CVL fluid from BZK exposed pigs was not due to blood contamination since these samples lacked detectable red cell contamination.

CVL fluid cytokine profile

Baseline CVL samples (n = 10) lacked detectable levels of the inflammatory cytokines assayed by the ELISA method (detection range 2–15 pg/ml). In contrast, CVL fluid collected from pigs exposed to 2% BZK that was used as a positive control, showed a significant increase ( p < 0.05) in immunoreactive IL-1β and IL-8 when compared to the baseline controls, VDDTC, and PHI-443 treated groups (Figure 4). Maximum levels of BZK-induced vaginal IL-1β (100-fold) and IL-8 (2,500-fold) were found at 24 hours, as compared to the controls. While CVL from controls and VDDTC-treated groups were negative, IL-18 (80–100-fold) and IFN-γ(10-fold) were also detected at 24 or 48 hours postapplication of BZK-containing gel (data not shown). In comparison, PHI-443 induced only a 4-fold increase in IL-18 in only 1 of the 3 pigs tested. Immunore-active IL-4, IL-6, IL-10, and TNF-α were not detectable in CVL fluid in any of the groups (the detection limits were 2–10 pg/ml).

Because IL-1β and IL-18 induce the production of IL-8 and IFN-γ, respectively, we looked for possible correlation between these cytokine concentrations in CVL fluid. Comparison of cytokine levels in 32 CVL fluids obtained from control and all treatment groups revealed that the high levels of IL-1β induced by the irritant BZK was significantly associated with CVL levels of IL-8 (r = 0.793; p < 0.0001; Figure 5A). Similarly, higher levels of IL-18 were also associated with detectable levels of IFN-γ(r = 0.463; p = 0.007) in the BZK group [Figure 5B]. IL-1β and IL-8 were present at a very high concentration only in BZK-exposed CVL fluids.

Evaluation of Stability of PHI-443 in the Acidic Vaginal Environment

Microbicides must have adequate chemical stability in the acidic vaginal environment. In vivo stability of PHI-443 was tested by obtaining post-24 hour CVL fluid from each of the 3 pigs given 2% PHI-443 containing gel. CVL extracts from PHI-443 exposed pigs obtained after 24 hours were assayed by a reverse-phase HPLC. Figure 6 shows the HPLC chromatograms of standard PHI-443 versus the CVL fluid extracts from each of the 3 pigs. Intact PHI-443 was recoverable from all 3 pigs 24 hours after intravaginal administration of 2% PHI-443 containing gel.

Effect of Multiple Intravaginal Application of BZK and VDDTC-Containing Gel on Vaginal Histology

Clinical findings

Because of the sensitivity of the porcine vagina to BZK, repeated intravaginal exposure of test gels was limited to 4 days. Pigs exposed to 2% BZK-containing gel showed visible redness and vaginal swelling 24 hours postapplication with no direct cumulative effects following repeated dosing. Pigs exposed to control gel, 0.1% VDDTC, or 2% PHI-443 containing gel showed no signs of distress or vaginal disturbance. The rectal temperature remained within the normal range (39.4°C–40°C) throughout the 4-day dosing period.

The histopathologic changes in the pig vagina following multiple applications of VDDTC were compared to BZK-exposed vaginal tissues. Because the pathological changes observed with BZK differed markedly from the scoring criteria for the rabbit vaginal irritation model, we used a new scoring system using 10 identifiable endpoints that reflect pathological changes in the epithelial/subepithelial as well as vascular/perivascular compartments. Tables 1A–C summarize the individual scores for detailed histological changes in the in vivo porcine vaginal irritation model. Intravaginal administration of 0.1% VDDTC, which is approximately 2,000-times its in vitro spermicidal EC₅₀ value, did not cause vaginal irritation in any of the 3 pigs evaluated (mean individual scores 0–1). In contrast, intravaginal exposure to BZK resulted in marked vaginal irritation (mean individual scores 3–4). Figure 7 shows representative vaginal sections of pigs repeatedly exposed to gel formulation alone and gel formulation containing 0.1% VDDTC or 2% BZK.

Vaginal epithelia

The intact proliferative vaginal epithelia of placebo control and VDDTC group consisted of 5–10 cell layer (diestrus) with polyhedral cell type without crypts, vacuolation or ulceration and with basal levels of neutrophils and mononuclear cells (Figures 7A and 7D). In contrast, the vaginal epithelia of BZK-treated group were severely ulcerated with desquamation, vacuolation, loss of mitotic cells, and with marked infiltration of neutrophils, and mononuclear cells (Figure 7G). The desquamated and flattened epithelium was overlayed by a thick ribbon of necrosis with mixed inflammatory cells and cellular debris. The crypts were filled with necrotic as well as suppurative debris. Microscopically, these sites were devoid of the vaginal epithelia. The underlying amorphous subcutis was necrotic and devoid of recognizable cell outlines.

Vaginal subepithelium and lamina propria

Microscopic examination of the vaginal subepithelium and lamina propria of control and VDDTC-exposed pigs revealed normal tissue with basal levels of leukocytes and without vascular congestion (mean individual scores <1.5; Figures 7B and 7E). In the BZK group, microscopic examination of the exposed submucosa revealed marked infiltration of neutrophils and mononuclear cells, inflammation, marked submucosal edema, and fibrinous necrosis (mean individual scores 3–4) (Figure 7H).

Vaginal vascular and perivascular tissue

In the placebo and VDDTC group, the vaginal vascular and perivascular tissue was normal (mean individual scores <1.5; Figures 7C and 7F). In the BZK group, microscopic examination revealed marked vascular congestion and edema (mean individual scores 3). The musculature showed lack of muscle integrity (separated, fragmented, or missing muscle fibers; Figure 7I). The muscle fibers were separated by edema.

Based on the data generated during this model development, we selected 10 identifiable criteria for the 4-point scoring of vaginal irritation. Table 2 summarizes the combined scores for histological changes observed for porcine vaginal irritation following 4 consecutive days of intravaginal application of a gel formulation with or without VDDTC or BZK. The total histopathological score was calculated by adding the quantitative assessment of histopathological findings in the investigated regions of the vagina for each of the animals. The mean of total histopathological score was 5.6 ± 0.6 in the control group, 3.7 ± 0.4 in the VDDTC group and 28.6 ± 2.8 in the BZK group (p < 0.001); the cumulative range being 0–40 [Figure 8].

Discussion

In this study, a single application of 2% BZK to the porcine vaginal mucosa revealed intense inflammatory changes, which were manifested by increased levels of species-specific IL-1β, IL-8, IL-18, and IFN-γ. Increased levels of these cytokines in CVL fluid were also associated with marked granulocyte influx and histological changes consistent with vaginal inflammation. These finding imply that key cervicovaginal inflammatory cytokines IL-1β and IL-8 and to a less extent IL-18 and IFN-γ are useful in vivo biomarkers for predicting the mucosal toxicity potential of vaginal products in the physiologically relevant and sensitive porcine model.

The production of IL-1β and IL-8 by the vaginal mucosa following exposure to an irritant spermicide or a microbicide can be expected to play a major role in the recruitment and activation of professional phagocytes at the gel application site. IL-1β induces the expression of adhesion molecules and the secretion of chemokines by epithelial and endothelial cells (Sha et al., 1997). Both mechanisms are involved in the recruitment and the subsequent activation of leukocytes in the tissue (Eckmann et al., 1993). IL-1β is able to induce the release of additional inflammatory mediators contributing to the production of reactive oxygen species or cytotoxicity toward endothelium. Notably, IL-1β has been shown to stimulate replication and mucosal transmission of HIV-1 (Belec et al., 1995; Sha et al., 1997; Sturm-Ramirez et al., 2000).

The chemokine IL-8 may be a key mediator of vaginal irritation due to its ability to recruit degranulating neutrophils to the vaginal mucosa, regulate endothelial cell adhesion molecules expression, promote binding of leukocytes to the adhesion molecules of endothelial cells as well as their migration into the vascular wall (Baggiolini et al., 1989; Bochner et al., 1991; Eckmann et al., 1993; Harada et al., 1994; Okamura et al., 1995). IL-1β also induces the synthesis of IL-8. Since an association between IL-1β and IL-8 concentration was observed in porcine CVL fluid, both IL-1β and IL-8 may contribute to the alteration of endothelial permeability in vaginal submucosa. Therefore, increased production of IL-8, by vaginal epithelial cells due to chemical stimuli can be expected to play a major role in the recruitment and activation of professional phagocytes at the gel application site.

Increased levels of IL-18 may be associated with promotion of inflammatory disease. IL-18 acts as IFN-γ inducing factor and is a potent inducer of IFN-γ production by activated T cells (Morel et al., 2001). A positive association between the levels of IL-18 and IFN-γ was noted in the porcine CVL fluid. Since IL-18 is known to induce the expression of CXC chemokines, stimulate angiogenesis, and is involved in leukocyte recruitment by up-regulation of vascular adhesion molecule-1 through nuclear factor- B-dependent mechanisms, its presence in CVL fluid may be related to the degree of vaginal inflammation (Leung et al., 2001; Morel et al., 2001; Park et al., 2001).

Microscopic examination of vaginal sections from pigs exposed to 2% BZK during the nonestrus stage demonstrated substantial changes in the vaginal architecture beyond those previously described in the rabbit (2% BZK) and pigtailed macaque (1.2% BZK) models of vaginal irritation (Patton et al., 1999; Fichorova et al., 2004). Therefore, a new semiquantitative scoring system was developed based on 10 identifiable biological endpoints in the porcine vaginal mucosa. Despite the stratified vaginal epithelia, the pig genital tract was highly sensitive to the toxic effects of BZK. Since a number of genital tract infections involve the upper genital tract (endocervical canal, endometrial cavity, and the Fallopian tubes), the pig model provides a large surface area as well as longer retentive capacity of the gel to determine the potential toxic effect of intravaginally applied spermicides and microbicides on the upper genital tract.

Because the immuno-inflammatory response has been shown to play an important role in the development of vascular lesions to vaginal irritants, particular attention was focused on the vascular and perivascular regions of vaginal tissue. The inflammatory and cellular changes observed with BZK were not only confined to the injured stratified epithelia and the immediate submucosa but were also found in the perivascular tissue extending several millimeters away from the stratum superficiale. Neutrophils and mononuclear cells were found deep within the vascular/perivascular compartments surrounding the injured vessels. The involvement of the distal perivascular tissues has implications for the potential recruitment of inflammatory mediators from these regions, which may be involved in the perivascular inflammation.

Extensive preclinical studies conducted in mice and rabbits after repeated intravaginal administration of VDDTC and PHI-443-containing gels did not induce marked vaginal irritation, mucosal toxicity, or systemic absorption of the test agents (D’Cruz et al., 2001, 2004a; D’Cruz and Uckun, 2001b). In agreement with these previous studies, histological examination of vaginal tissue exposed to 0.1% VDDTC (2,000 times its in vitro spermicidal EC₅₀ value) via a gel-microemulsion or to 2% PHI-443 (30,000 times its in vitro anti-HIV IC₅₀ value) via a nonemulsifying gel showed no evidence of local toxicity in any of the dosed animals. Furthermore, the present studies using the porcine model showed no significant increase in the levels of immunoreactive IL-1β and IL-8 in CVL fluids when compared to baseline levels. The anti-HIV microbicide, PHI-443 was found to be highly stable in the acidic vaginal environment. The demonstrated lack of detergent-type membrane toxicity supported by the lack of immuno-inflammatory response in the CVL fluid as well as the lack of histomorphological changes in the pig vagina indicates that VDDTC and PHI-443 in this animal model displays the properties required of an ideal spermicide and nonspermicidal microbicide, respectively, and warrants further preclinical research and development.