Abstract
The mucosal safety of the combination antiretroviral spermicide, WHI-07 [5-bromo-6-methoxy-5,6-dihydro-3′-azidothymidine-5′-(p-bromophenyl)-methoxy alaninyl phosphate] and vanadocene dithiocarbamate (VDDTC), was evaluated in 3 different animal models. Twenty-seven NZW rabbits in four subgroups were exposed intravaginally to a gel-microemulsion (GM) with and without three dose levels of WHI-07 plus VD-DTC (0.5 + 0.06%, 1.0 + 0.12% and 2.0 + 0.25%) or 4% nonoxynol-9 (N-9; Conceptrol®) for 14 consecutive days. Ten nonestrus gilts (Duroc) in three subgroups received either a single or daily intravaginal application of GM with and without 2.0% WHI-07 plus 0.25% VDDTC or 2.0% benzalkonium chloride (BZK)-containing gel for 6 and 4 consecutive days, respectively. Five cats received a single intravaginal application of GM incorporating 2.0% WHI-07 plus 0.25% VDDTC. Genital tract histopathology was performed in the pig and rabbit at the end of dosing period but after 18 weeks post-dosing in the cat. Porcine cervicovaginal lavage (CVL) fluid was obtained for up to 72 hours after a single exposure and changes in the levels of inflammatory cytokines (IL-1β, IL-8, IFN-γ, and TNF-α) were quantitated by a multiplexed chemiluminescence-based immunoassay. Rabbit vaginal tissues were evaluated for localized cellular inflammation and in situ apoptosis by immunohistochemical staining for CD45, nuclear factor (NF)-κB, and terminal deoxynucleotidyl transferase-mediated FITC-deoxyuridine triphosphate nick-end labeling (TUNEL) using confocal laser scanning microscopy (CLSM), respectively. Vanadium content in selected organs and body fluids from rabbits and pigs was determined by atomic absorption spectroscopy. When compared with 4% N-9 (total irritation score 13–14 out of a possible 16), none of the rabbits given WHI-07 plus VDDTC intravaginally, developed histological alterations such as epithelial erosion, edema, leukocyte influx or vascular congestion characteristic of inflammation (total irritation score 4–6). CD45 and NF-κB immunoreactivity was limited to cells within the vascular lumen of both control and WHI-07 plus VDDTC-treated vaginal tissues. TUNEL assay revealed lack of increased apoptotic cells in vaginal mucosa exposed to increasing concentrations of WHI-07 plus VDDTC. Basal levels of proinflammatory cytokines (IL-1β, IL-8, IFN-κ and TNF-α) in porcine CVL were unaffected by intravaginal exposure to WHI-07 plus VDDTC when compared with BZK used as a positive control. Endpoint histology of the reproductive tract from cats and pigs after a single or repeated intravaginal exposure to WHI-07 plus VDDTC, respectively, revealed lack of irritation/inflammation in the epithelium, subepithelium/lamina propria, vessels/perivascular tissues, and underlying/surrounding muscles. Vanadium was not preferentially incorporated into rabbit or porcine tissues and body fluids at levels above 1 μg/g. Based on comparative histologic data and surrogate markers for inflammation, repeated intravaginal administration of WHI-07 plus VDDTC via a gel-microemulsion did not result in vaginal irritation, mucosal toxicity, or systemic absorption of vanadium. Therefore, the combined use of WHI-07 and VDDTC via gel-microemulsion appears safe for topical use as a prophylactic anti-HIV microbicide.
Keywords
Introduction
Heterosexual transmission is the main mode of HIV-1 transmission worldwide and accounts for nearly 90% of all HIV infections in women (UNAIDS Report, 2006). Currently, about half of the 42 million people living with HIV/AIDS are women (NIAID Fact Sheet, 2006). Because most women at risk for HIV infection are of reproductive age, the development of safe and efficacious spermicidal microbicides is important to prevent HIV-1 transmission and unintended pregnancies [Uckun and D’Cruz, 1999]. Anti-HIV microbicides can provide protection by inactivating viruses or preventing viruses from replicating either in semen or in the infected host cells that line the vaginal wall.
Several candidate microbicides are moving forward clinically but many unanswered questions remain relative to benchmarks for evaluating efficacy and safety (D’Cruz and Uckun, 2004a, 2007). An essential prerequisite for the clinical development of microbicides and spermicides is the careful monitoring of vaginal irritation potential in a physiologically relevant and sensitive animal model. An ideal topical microbicide would not only be effective at preventing HIV-1 but would also be safe for repeated use thereby preserving the inherent defenses of the genital tract without causing genital inflammation. Since the entire reproductive tract is a potential target for HIV-1 uptake, it is imperative that potential microbicides and spermicides do not disrupt or irritate the epithelium to attract and activate HIV-1 infected host cells.
Animal models for assessment of in vivo microbicide/spermicide efficacy or toxicity have included monkeys, dogs, guinea pigs, rabbits, rats, and mice (Eckstein et al., 1969; Chvapil et al., 1980; Gray et al., 1984; Kaminsky et al., 1985; Milligan et al., 2002; Patton et al., 1999, 2004a,b; 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 differences in genital tract physiology and histology. The estrus cycle of animals and their genital tract infectivity are important variables. Due to variable sensitivity of these models to topically applied test agents, the toxicological endpoints also differ.
At present, animal models best suited for vaginal microbicides seem to be the chimeric simian/HIV (SHIV)-infected or SIV/HIV-2-infected rhesus monkeys (Manson et al., 2000; Tsai et al., 2004; Veazey et al., 2005) and the feline immunodeficiency virus (FIV)-infected domestic cats [Bishop et al., 1996; Jordan et al., 1996; Obert and Hoover, 2000]. The macaque species closely parallels humans with respect to their vaginal anatomy, physiology, pH, and micro-flora (Ratterree et al., 2005). However, this model may not accurately reflect the dynamics of human sexual transmission since infection is achieved using a cell-free virus challenge and requires progestin pre-treatment to enhance susceptibility to virus. Feline AIDS is strikingly similar to AIDS in humans (Willett et al., 1997). FIV is similar to HIV-1 in morphology, protein composition, properties of the Mg2+ -dependent reverse transcriptase (RT), cellular tropism as well as the clinical symptoms produced, and is therefore considered an ideal preclinical model for testing new antiretroviral agents (North et al., 1989; Bendinelli et al., 1995).
FIV can be efficiently transmitted to cats by vaginal inoculation of infected feline T cells (Bishop et al., 1996; Jordan et al., 1996; D’Cruz et al., 2004). Consequently, FIV/cat model has been particularly useful for studies of viral resistance to nucleoside analog RT inhibitors (NRTIs) because its RT is similar to the HIV-1 RT in its physical properties, catalytic activity, and sensitivity to the active forms of NRTIs (Amacker et al., 1998; Auwerx et al., 2002).
We previously reported the discovery of WHI-07 (5-bromo-6-methoxy-5,6-dihydro-3′-azidothymidine-5′-(p-bromophenyl)-methoxy alaninyl phosphate), a novel aryl phosphate derivative of zidovudine (ZDV) with potent anti-HIV-1 and spermicidal activities (D’Cruz et al., 1998a, 1999, 2000c). WHI-07 was rationally designed to enhance the antiretroviral activity as well as bypass the kinase activation step especially in thymidine kinase (TK) deficient or lacking monocytes/macrophages, the main carriers of HIV-1 in semen and female genital tract secretions (D’Cruz et al., 2001a). The primary metabolite of WHI-07 (3′-azidothymidine-5′-[p-bromophenyl methoxyalaninyl phosphate) exhibits broad-spectrum anti-HIV-1 activity against drug-resistant strains. WHI-07 formulated via a lipophilic gel-microemulsion for mucosal delivery provided significant protection from systemic FIV infection via the vaginal and rectal routes in the FIV/cat model of natural AIDS (D’Cruz et al., 2004). However, it is becoming apparent that no single agent will be able to prevent sexual transmission of HIV-1 long term. In addition, prophylaxis against sexually transmitted HIV-1 also requires blocking the transfer of infected leukocytes among sexual partners
In an attempt to develop an effective antiretroviral spermicide composed of two classes of drugs, we have evaluated the utility of WHI-07 in combination with vanadocene dithiocarbamate (VDDTC), a chelated vanadocene as two distinct classes of dual-function contraceptives with antileukocytic activity (D’Cruz and Uckun, 1998, 2003 , 2005; D’Cruz et al., 1998a,b,c; Uckun et al., 2005; Ghosh et al., 2000). These two classes of compounds fundamentally differ from the currently used membrane-active surfactant spermicides. WHI-07 and VDDTC exerted a synergistic effect on the antiretro-viral effect in the FIV/domestic cat, a natural model for AIDS (D’Cruz et al., 2004). The synergistic in vivo antiretroviral and contraceptive efficacy of the combination of WHI-07 and VDDTC is most likely due to the potent antiviral activity of WHI-07 as well as the combined effects of WHI-07 and the organovanadium complex on key cellular pathways essential for nucleic acid synthesis, viral replication, and sperm motility. Consequently, gel-microemulsion composed of WHI-07 and VDDTC is expected to effectively block vaginal transmission of cell-free and cell-associated HIV-1 via semen as well as provide a reliable method of contraception.
Previous studies demonstrated that repeated intravaginal exposure to WHI-07 (0.5–2.0(0.06–0.25(mice) and/or non-rodent species (rabbits, pigs, and/or cats (D’Cruz et al., 1999, 2001b, 2002, 2003, 2004, 2005; D’Cruz and Uckun, 2001a,b,c, 2002, 2004b). This study investigated the mucosal irritation potential of WHI-07 plus VDDTC via gel-microemulsion as a combination anti-HIV spermicide evaluated in three different animal models for local tolerance. The tissue inflammatory response, if any, resulting from single or repeated intravaginal exposure of rabbits, pigs, and cats to WHI-07 plus VDDTC-containing gel-microemulsion was evaluated at the extracellular, cellular, and/or histological level. Our recent studies established that the porcine model is superior to the lagomorph as an animal model for human vaginal pathology (D’Cruz et al., 2005). Using species-specific reagents, the levels of porcine cervicovaginal lavage (CVL) fluid proinflammatory cytokines were shown to predict the mucosal irritation potential than can be used to rank vaginal products. Because spermicidal activity of vanadocenes is invariably associated with induction of apoptosis (D’Cruz et al., 1998b,c, 2000a; D’Cruz and Uckun, 2000; Ghosh et al., 2000), we also investigated the ability of intravaginally administered WHI-07 plus VDDTC to inducein situ apoptosis in the female genital tract. Our findings revealed that intravaginal administration of increasing doses of WHI-07 plus VDDTC via a gel-microemulsion did not cause vaginal irritation, mucosal toxicity, or preferential absorption of vanadium. Therefore, the combined use of WHI-07 and VDDTC via gel-microemulsion appears safe for topical use as a prophylactic anti-HIV microbicide.
Materials and Methods
Synthesis of WHI-07 and VDDTC
WHI-07 and VDDTC, the structures of which are shown in Figure 1, were synthesized according to our published procedures (D’Cruz et al., 1998a,c; Ghosh et al., 1998). The purity of WHI-07 was >98%, as assessed by the proton (1H), carbon (13C), phosphorous (31P) nuclear magnetic resonance (NMR) spectra and Fourier-transform infrared (FTIR) spectroscopy (FT-Nicolet model Protege 460; Nicolet Instruments Corp., Madison, WI), and UV-visible spectroscopy (DU 7400 spectrophotometer, Beckman Instruments, Fullerton, CA). The purity of VDDTC (bis(cyclopentadienyl)N,N -diethyl dithiocarbamato triflate salt) as determined by 1H NMR, FTIR spectra, UV-visible spectroscopy, and elemental analysis exceeded 99%.
Gel-Microemulsion Formulation of WHI-07 plus VDDTC
WHI-07 and VDDTC were solubilized in a lipophilic sub-micron (30–80 nm) particle size microemulsion-based vehicle using selected pharmaceutical excipients identified through systemic mapping of ternary phase diagrams and drug solubilization studies (D’Cruz and Uckun, 2001d). Microemulsions can deliver larger amounts of topically applied WHI-07 and VDDTC into the vaginal mucosa than traditional vehicles because of their capacity for enhanced solubilization and greater bioavailability (Tenjarla, 1999). The microemulsion-based system composed (w/w) of Phospholipon 90G (5.1%) and Captex 300 (10.8%) as the oil phase with Cremophor EL (7.6%) as surfactant, propylene glycol (4.2%) and polyethylene glycol 200 (4.2%) as cosurfactants, and water (66.1%) as a carrier and a preservative (0.2% sodium benzoate) (D’Cruz and Uckun, 2006). Polymer suspensions of SeaSpen PF (0.9%) and Viscarin GP-209 (0.9%) carrageenans were selected as additives to the microemulsion to obtain a gel with desirable viscosity containing WHI-07 and VDDTC with thickening capability and compatibility with vaginal mucosa. The lipophilic microemulsion-based vehicle developed for WHI-07 and VDDTC offers several benefits for vaginal delivery, including increased absorption, potent microbicide as well as contraceptive activity, and decreased toxicity (D’Cruz and Uckun, 2001d).
Dose Level Selection
The doses of WHI-07 and VDDTC chosen for the mucosal safety studies were based on the lack of local, systemic, and reproductive toxicity observed in two 13-week subchronic studies and a two-year carcinogenicity study conducted in mice (D’Cruz and Uckun 2001a,b; D’Cruz et al., 2000d, 2001b). Since no adverse effects were observed with the highest concentration of WHI-07 (2.0%) or VDDTC (0.25%) tested in mice and rabbits, we selected the highest doses for the intravaginal safety study. Stable microemulsions were prepared by combining separate mixtures of Captex 300/Phospholipon 90G and Cremophor EL/propylene glycol/PEG 200. WHI-07 and VDDTC were solubilized in the microemulsion at three dose levels (0.5 + 0.06%, 1.0 + 0.12% and 2.0 + 0.25% (by weight), respectively) followed by the addition of aqueous polymer suspensions to increase the viscosity. The gel-microemulsion with and without WHI-07 plus VDDTC were prepared daily. The neat gel-microemulsion has been found to be stable when stored at room temperature for 5 years. Particle size determination was made using Nicomp Model 380 laser diode source (Particle Sizing Systems, Santa Barbara, CA). Viscosity measurements were made using Brookfield Model DV-II+ digital viscometer (Brookfield Engineering Laboratories, Spoughton, MA).
Animals
Rabbits
Twenty-seven female, sexually mature (>6 months old) New Zealand White (NZW) rabbits were obtained from Bakkam Rabbitry (Madison, WI). All rabbits were identified with specific metal ear tags and housed individually in stainless steel cages under standard conditions (temperature, 20 ± 2°C, 50 ± 10% relative humidity, and 12-hour light:dark cycle). Food (2031 Global High Fiber Rabbit Diet, Harlan Teklad, Madison, WI) and water were available ad libitum. The rabbits were isolated for 4 weeks before the intravaginal study.
Pigs
Ten adult, female domestic Duroc strain pigs (>7 months old) free of porcine reproductive and respiratory syndrome virus and pseudorabies virus were obtained from Zierke Company (Morris, MN). 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 identified with specific ear tags and housed in groups of two in large interconnected pens that provided aural, visual, and olfactory contacts. The pigs were given pelleted food (LabDiet 5084, PMI Nutrition International, LLC, Brent-wood, MO); 4 kg/day) once daily in amounts calculated to maintain their adult weights and access to automatic waterers. The animals were maintained in rooms that were at 22 ± 2°C with relative humidity of 60 ± 20% and a 12-h light/dark cycle. The animal rooms were cleaned daily.
Cats
Eight adult, female domestic cats (>6 months old) obtained from Liberty Research, Inc. (Waverley, NY) were used in this study. The cats were housed in environmentally enriched large pens and maintained under standard conditions: 18 to 29°C, 30 to 70% relative humidity, and 12-hour light:dark cycle. Cats were provided with dry and wet Purina cat chow and tap water ad libitum. The animal rooms were cleaned daily, and cat litters were changed daily. Cats were acclimated to the study room conditions for at least one week prior to initiation of the experiment.
The animals were housed in accordance with the American Association for Accreditation Laboratory Animal Care standards. 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).
Rabbit Vaginal Irritation (RVI) Test
For the vaginal mucosal toxicity assessment, 27 female rabbits were randomly assigned 5 five treatment groups as follows: gel-microemulsion control (n = 6), 0.5 + 0.06% (n = 6), 1.0 + 0.12% (n = 6), 2.0 + 0.25% WHI-07 plus VDDTC (n = 6). The doe was held in a supine position and one mL of gel-microemulsion with and without test agents was administered intravaginally via sterile tuberculin syringe to a depth of 6–8 cm for 14 consecutive days. The formulation of WHI-07 plus VDDTC in gel-microemulsion was prepared daily.
Rabbits (n = 3) administered with one ml of 4% Nonoxynol-9 (N-9) gel (Conceptrol, Ortho-McNeil Pharmaceutical Inc., Raritan, NJ) in parallel were used as positive control. Body weights were obtained before and after completion of the 14-day intravaginal application. All animals were individually observed daily for signs of toxic effects (inappetency, genital swelling, redness as well as bleeding). On day 15, rabbits were sedated with ketamine plus xylazine and euthanized by an intravenous injection with euthasol (Delmarva Laboratories, Midlothian, VA, USA) and their genital tracts were examined grossly and microscopically after completion of the study. The vaginal tissues were rapidly removed and parts of the upper (cervico-vagina), middle (mid-vagina), and lower (uro-vagina) regions of each vagina were fixed in 10% neutral-buffered formalin. The remainder of the vaginal tissue was frozen at −80°C for vanadium analysis. The bone, heart, kidney, liver, lung, muscle, spleen, ovary, blood and urine/bladder were also collected and kept frozen for vanadium analysis.
Fixed vaginal tissues were embedded in paraffin, sectioned at a thickness of 4–5-μm, stained with hematoxylin and eosin (H & E), and examined by light microscopy by a board certified veterinary pathologist. Tissue sections were viewed under ×200 and ×400 magnification using an Olympus BX40 light microscope (Olympus Optical Co., Ltd, Japan) attached to an Olympus PM-C35DX camera. Each of the three regions of vagina was examined for epithelial cell damage, leukocyte infiltration, stromal edema, and vascular congestion. The irritation scores were assigned based on a semiquantitative scoring system (Eckstein et al., 1969): Individual score: 0 = none, 1 = minimal, 2 = mild, 3 = moderate, 4 = intense irritation. This scoring system has been shown to correlate to human irritation potential as follows: Total scores of 0 to 8 are acceptable, scores of 9 to 10 indicate borderline irritation potential, and scores of 11 and above are potentially irritating (unacceptable). Results were expressed as the mean ± SD irritation scores.
Porcine Mucosal Safety Test
Test gel application
Intravaginal gel application and CVL fluid collection was performed during the nonestrus stage of the 21-day porcine reproductive cycle to avoid the effect of estradiol on genital IL-1-mediated proinflammatory responses during estrus. The study was conducted in two separate phases. In experiment I, gilts in subgroups of three were assessed for mucosal safety of gel microemulsion alone and a gel formulation incorporating 2% benzalkonium chloride [BZK] as a positive control (D’Cruz et al., 2005). In experiment 2, 4 gilts were assessed for mucosal safety of 2% WHI-07 plus 0.25%-loaded gel-microemulsion. Estrus was synchronized in four sexually mature gilts by administering (15 mg/gilt) orally active synthetic progestin (MATRIX [Altrenogest); Intervet Inc, DeSoto, KS) for 14 consecutive days and the gilts were allowed to come to standing heat for up to 8 days. On post-treatment day 11, gilts were administered intravaginally a gel-microemulsion incorporating 2.0% WHI-07 plus 0.25% VDDTC.
The test agents were administered by means of a plastic insemination catheter with a soft foam tip (‘Golden Pig’; Fox A. I., Cedar, IA, USA), which was gently inserted upwards into the vagina ensuring that the tip did not penetrate the urethra. The catheters were rotated anti-clockwise and steadily pushed through the vagina until resistance was felt at the opening of the cervix. A plastic syringe containing gel was then connected to the inserted catheter and approximately 50 ml of test gel was expelled using gentle pressure and drawn into the pig by wave-like muscular contractions and to prevent any back-flow.
CVL fluid collection
CVL lavage fluid was collected similarly before and after test gel application. Following insertion of the catheter it was connected to a 60 mL syringe containing sterile phosphate buffered saline (PBS) containing 50 μg/ml gentamicin sulfate and approximately 50 mL was instilled into the cervico-vaginal lumen. As much lavage fluid as possible was then collected during gentle manipulation and the suction created by the foam-tip during the withdrawal of the catheter. To examine the lymphocytic cellular profiles and cytokine levels following intravaginal exposure to WHI-07 plus VDDTC, CVL fluid was collected at 24, 48, and 72 hours post application. CVL fluid collected from three pigs that received gel-microemulsion alone and three pigs that received gel 2% BZK gel were used as negative and positive controls for surrogate markers for inflammation and mucosal toxicity (D’Cruz et al., 2005). Within 30 min 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 CVL fluid were viewed microscopically to verify the absence of bacterial, fecal or red cell contamination. The remainder of the CVL fluid was centrifuged at 300g for 10 min at 4°C to remove mucus and cells. The cell-free supernatants were aliquoted, stored at −20°C and was used to measure the cytokine levels.
Flow Cytometry
Normal control blood was obtained from the ear vein for gating the flow cytometric scatter profile of specific blood cell types. Flow cytometric 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). Peripheral blood mononuclear cells were isolated from whole heparinized blood by density centrifugation using Histopaque-1077 (Sigma Chemical Co, St. Louis, MO) and suspended in PBS. One ml aliquots of freshly collected CVL fluid fixed in 2% paraformaldehyde 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 gated cells.
Cytokine/chemokine assay
Levels of proinflammatory cytokines in porcine CVL fluid were quantitated using multiplex SearchLight proteome array (Pierce SearchLight Proteome Arrays, Boston, MA) (Moody et al., 2001). Each sample was tested in duplicate for porcine-specific cyokines: interleukin (IL)-1β, IL-8; interferon (IFN)-γ, and tumor necrosis factor (TNF)-α. The values for the test samples were compared to those on standard curves, and exact values were determined.
The SearchLight multiplex assay uses multiplexed sandwich ELISA (Moody et al., 2001). The different capture antibodies were pre-spotted using robotic arrayers on a 96-well plate. Supernatant samples and/or standards were added to the wells and proteins within the sample bind the appropriate capture antibodies. Following a wash step, a cocktail of biotinylated antibodies was added. These antibodies bind to the target proteins that have been captured. Following a second wash step, streptavidin-horseradish peroxidase (HRP) conjugate was added, followed by the addition of SuperSignal ELISA femto chemiluminescent substrate. The luminescent signal produced from the HRP-catalyzed oxidation of the substrate was measured by imaging the plate using a Black Ice cooled charge-coupled device (CCD) camera system. The data was then analyzed using array analyst software. Density values for test samples in duplicate were analyzed using the standard curve for each analyte to calculate actual pg/ml values. Values for each protein were identified based on the spotting position of each capture antibody within the well. The detection limits for the tests were 0.4, 0.8, 1.6, and 0.4 pg/ml for IL-1β, IL-8, IFN-γ, and TNF-α, respectively.
Histologic scoring
In experiment 2, four gilts approximately 11 days following chemical-induced estrus received intravaginally 50 ml of gel-microemulsion incorporating 2.0% WHI-07 + 0.25% VDDTC via a catheter for 6 consecutive days. Pigs were individually observed daily for overt clinical signs (inappetency, genital swelling, redness, vaginal discharge including bleeding). On day 7, pigs were sedated with telazol (Fort Dodge Animal Health, Fort Dodge, IA, USA), and euthanized by an intravenous injection of euthasol via a suitable ear vein. Animals were placed in dorsal recumbency and following a midline abdominal incision, the genital tract was retrieved and parts of the vagina, cervix, uterus, Fallopian tube, and ovary were excised and fixed in 10% neutral-buffered formalin for microscopic evaluation. The remainder tissues were frozen at −80°C for vanadium analysis. In addition, the bone, kidney, liver, lung, muscle, spleen, cervix, uterus, ovary, blood, urine and bladder were collected from WHI-07 plus VDDTC-treated pigs and kept frozen for vanadium analysis.
Fixed vaginal, cervical, uterine, and Fallopian tube specimen were trimmed, embedded in paraffin, stained with H & E and examined by light microscopy by a board-certified veterinary pathologist. 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. The muscle scoring was for the smooth muscle overlying the epithelium and lamina propria of the vagina and cervix, the myometrium of the uterus and including scant muscle fibers of the Fallopian tubes. Scoring for muscle integrity was based on separation of muscle fibers with leukocyte infiltration and fracture, scarring or mineralization of muscle fibers.
Tissues previously collected from 3 pigs that received gel-microemulsion alone served as negative control and 3 pigs that received 2% BZK gel for 4 consecutive days were used as positive control for mucosal toxicity. Based on the morphological changes observed with the irritant spermicide BZK, tissue sections from each WHI-07 plus VDDTC-treated specimen were scored blindly for the following ten histological features: epithelial ulceration/erosion; 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 (D’Cruz et al., 2005).
Mucosal safety studies in cats
For the evaluation of long-term inflammatory associated obstructions in the reproductive tract, 5 cats received a single application of 0.4 mL of gel-microemulsion incorporating 2% WHI-07 + 0.25% VDDTC and three cats were used as untreated control. The formulation was administered intravaginally using a blunt-tipped pipette in cats sedated with a combination of ketamine, atropine, and acepromazine. Cats were monitored for signs of toxic effects (inappetency, lethargy, vomiting, and vaginal discharge). To determine long-term histopathological changes on the gel application site and upper reproductive tract, cats were electively sacrificed 18 weeks after vaginal application of the test gel-microemulsion. Following euthanization with euthasol, the genital tract was retrieved and parts of the vagina, cervix, uterus, and Fallopian tube were excised and fixed in 10% neutral-buffered formalin for microscopic evaluation. H & E-stained sections were examined by light microscopy by a board-certified veterinary pathologist for the presence of cervicovaginal abnormalities including cervicitis, vaginitis, neutrophil influx, scarring or fibrosis.
Evaluation of In Situ Vaginal Inflammation by CD45 and Activated NFκB Antigen Staining
CD45 and activated NFκB/p65 (Rel A) immunostaining was used to evaluate the vaginal tissues for the presence of inflammatory cells. Briefly, tissue sections were deparaffinized with xylene and rehydrated through graded series of alcohols. Tissue sections were rinsed in PBS, pre-treated with citrate buffer at 93°C, blocked with PBS containing 2% BSA, and then incubated with an optimal dilution of FITC-conjugated mouse anti-rabbit CD45 (VMRD Inc, Pullman, WA or Antigenix America, Huntington Station, NY) or antibody reactive against rabbit activated p65 subunit of NFκB (RelA; clone 12H11, Chemicon International Inc., Tamecula, CA or SC-372, Santa Cruz Biotechnology Inc, Santa Cruz, CA). The sections were counter-stained with propidium iodide (PI). Bound primary antibodies were detected using appropriate FITC-conjugated secondary antibodies (Biosource International, Carlsbad, CA). FITC-labeled antibody complexes were visualized by confocal scanning laser microscopy (CSLM). The nature of staining and the distribution of CD45 and NF-κB immunoreactivities were determined by scoring a minimum of 500 cells in the vaginal epithelium and the stromal region in several random fields and the percentage of CD45 and NF-κB-positive cells for each region of the tissue section was calculated.
Confocal Laser Scanning Microscopy
Confocal microscopy was performed using a BioRad MRC-1024 Laser Scanning Confocal Microscope (Bio-Rad, Hercules, CA) equipped with a krypton/argon mixed gas laser (excitation lines at 488, 568, and 647 nm) and mounted on a Nikon Eclipse E800 series upright microscope with high numerical objectives. Using fluorescence imaging, the fluorescence emission of FITC and PI localized in cells was simultaneously recorded using 598/40 nm, and 680 DF32 emission filter, respectively. Confocal images were obtained using a Nikon 60 × (NA 1.4) objective and Kalman collection filter. Digitized data was processed using Lasersharp (Bio-Rad) and the digitized images were processed with the Adobe Photoshop version 8.0 software (Adobe Systems Inc, San Jose, CA).
Evaluation of In Situ Apoptosis in Vaginal Tissue by DNA Strand Break Labeling
The ability of WHI-07 plus VDDTC to induce in situ apoptosis in the genital tissue was evaluated by the terminal deoxynucleotidyl transferase (TdT)-mediated digoxigeninuridine triphosphate (dUTP) nick-end labeling (TUNEL) assay which labels exposed 3′-hydroxyl (3′-OH) ends of fragmented nuclear DNA (Gavrieli et al., 1992; D’Cruz and Uckun, 2001b). Briefly, following deparaffinization and rehydration, control and WHI-07 plus VDDTC-treated vaginal tissue sections were treated with proteinase K (20 μg/mL) for 15 minutes at room temperature to digest nuclear matrix proteins and expose the chromatin. As a positive control, control sections were treated with DNAse (2 μg/mL) in the reaction buffer (40 mM Tris-HCl, pH 7.9, 10 mM NaCl, 6 mM MgCl2, 10 mM CaCl2) for ten minutes prior to TUNEL staining. Sections were then incubated for 1 hour with TdT and FITC-dUTP (Promega Corporation, Madison, WI) in a humidified chamber at 37°C. After terminating the reaction, sections were washed in PBS. Nuclear staining was performed by counterstaining the sections with PI (1 μg/mL). For negative controls, the TdT enzyme was omitted from the reaction mixture. The numbers of apoptotic cells in the vaginal epithelium and stroma were counted in several random fields per tissue (×400), and the percentage of apoptosis was taken as (number of cells with fragmented nuclei)/(total cells counted) ×100. Three independent determinations were made for each tissue section. Vaginal tissue sections from all animals were evaluated by CLSM for the presence of apoptotic nuclei.
Vanadium Analysis
Because the absorption of vanadium into the body depends on its physicochemical state, chemical composition of the compound, the species exposed as well as the route of exposure [Ramanadham et al., 1991], the vanadium content of tissues harvested from rabbits and pigs was determined by atomic absorption spectroscopy. Frozen tissue samples were thawed and 1 g of these tissues/fluids were transferred to Folin-Wu tubes and wet ashed with nitric/perchloric acid. To wet ash the samples, 10 mL of concentrated nitric acid was added to the tissue/fluid and allowed to stand overnight for a pre-digestion. The samples were then heated to 150°C for one hour and cooled. Two ml of concentrated perchloric acid was added and the mixture was ramped slowly to 220°C until white fumes appeared. The digestion was continued for two additional hours after the appearance of white fumes. The samples were removed from the heating block and allowed to cool slightly. The digest was then diluted to 12.5 mL with 10% hydrochloric acid and mixed thoroughly. Vanadium analysis was performed on the digest using a Perkin Elmer 3000 DV ICP atomic emission spectrometer (Norwalk, CT). Digest controls consisted of positive (spiked with known amount of vanadium) and negative controls. Under the digesting conditions used, the detection limit for vanadium was 0.075 μg/g. The recovery was >90%. The instrument was calibrated with multi-element solutions (Spex CertiPrep, Inc., NJ). The calibration solution was analyzed as an unknown at the beginning and after every tenth sample.
Statistical Analysis
Data are presented as the mean ± SD or SEM. 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). Power analysis and sample size determinations were calculated from the mean and SD values using a two-sided two-sample t-test. Values of P < 0.05 were considered statistically significant.
Results
Light Microscopic Changes in the Rabbit, Porcine, and Cat Reproductive Tract
The potential of WHI-07 plus VDDTC via a gel-microemulsion to cause toxicity to the female reproductive tract was evaluated in three different animal models. In the rabbit and porcine model, tissue toxicity was evaluated following a 14-day and 6-day intravaginal exposure, respectively. Cats were evaluated histologically for long-term effects on the female reproductive tract after a single intravaginal exposure. No clinical signs of toxicity were observed in rabbits, pigs, and cats given WHI-07 plus VDDTC via gel-microemulsion intravaginally.
Figure 2 shows the representative light micrographs of H & E-stained vaginal sections of rabbits given gel-microemulsion with and without increasing concentrations of WHI-07 plus VDDTC for 14 days. Repeated intravaginal exposure to three doses of WHI-07 plus VDDTC via a gel-microemulsion did not result in significant microscopic abnormalities. Light microscopic examination revealed intact vaginal epithelium and lack of leukocyte influx in the representative mid vaginal sections of rabbits following daily intravaginal administration of gel-microemulsion alone [Fig. 2A] or gel-microemulsion containing 0.5 + 0.06% (Fig. 2B), 1.0 + 0.12% (Fig. 2C), or 2.0 + 0.25% (Fig. 2D) WHI-07 plus VDDTC for 14 consecutive days. In contrast, vaginal tissue sections from rabbits treated with 4% N-9 gel (Conceptrol®) used as a positive control, revealed extensive ulceration and denudation of the epithelial cell layer, submucosal edema, leukocyte infiltration, and vascular congestion (Fig. 2E and Fig. 2F).
Figure 3 shows the total irritation scores for histological changes in three different regions of the rabbit vagina after 14 days of exposure to increasing concentrations of WHI-07 plus VDDTC and 4% N-9 (Conceptrol®). No substantive differences in the incidence or severity of histopatho-logic changes were evident among any of the WHI-07 plus VDDTC-exposed groups at the end of the 14-day exposure period. The histopathological changes observed for WHI-07 plus VDDTC were considerably less than that for 4% N-9 gel used as a positive control (total score 13–14 out of 16). At the highest dose tested, WHI-07 plus VDDTC (2.0% + 0.25%) induced only minimal to mild irritation (total score 4–6). Morphological changes observed in the treatment groups included mild vascular congestion and leukocyte infiltration, which were within the acceptable range for a clinical trial (≤ 8). No significant differences in body weight were observed between control and WHI-07 plus VDDTC-treated rabbits (4777 ± 280 g compared with 4368 ± 367, 4576 ± 272, 4399 ± 263 g, respectively) at the end of 14-day treatment period. Power analysis based on the SD of the total irritation score of 0.7 units observed for vehicle control, a sample size of 6 per treatment group was calculated to detect a difference of 2.63 units at 95% power (α = 0.05, two sample, 2-tailed test).
Figure 4 shows the representative light micrographs of H & E-stained vaginal (
In all 4 WHI-07 plus VDDTC-treated gilts, the cervical epithelium infrequently displayed minimal inflammatory changes and was essentially within normal limits. Within the subepithelial lamina propria, all gilts displayed minimal-to-mild infiltration of neutrophils, while 2 had mild-to-moderate infiltration of mononuclear leukocytes. Three gilts had occasional fibrinous necrosis within small vessels with a slight degree of accompanying congestion and edema. The cervical smooth muscle was essentially within normal limits (Figure 4B). The uterine epithelium was largely composed of polygonal cells (Figure 4C). In 2 of these gilts, there was mild to moderate infiltration of both granulocytes and monocytes into the epithelial cell layers. Mild to moderate infiltration of inflammatory cells within the lamina propria of the subepithelium was present in three animals. Endometrial glands had few monocyte infiltrate in both the epithelium and subepithelium. Vascular congestion and edema was present in all pigs and a few capillaries in one animal displayed fibrinous necrosis related to the phase of the physiological cycle. The Fallopian tubes were examined in two pigs. In the Fallopian tubes, except for moderate congestion in one pig, all histopathology parameters were within normal limits. In the ampulla, the elaborately branched folia (the plicae), were intact with epithelial lining composed of ciliated and nonciliated simple columnar cells (Figure 4D). The lamina propria contained fibroblasts, reticular fibers, and a few lymphocytes and macrophages.
The results of histological changes observed in the vaginal, cervical, uterine, and Fallopian tube are summarized in Table 1. There were no significant differences between mean individual scores for pigs exposed to WHI-07 plus VDDTC versus gel-microemulsion control when compared with 2% BZK group. The total histopathological score was 5.6 in the control, 6.8 in the 2% WHI-07 plus 0.25% VDDTC group and 29.3 in the 2% BZK group (p < 0.001); the cumulative range being 0–40 (Figure 5). Power analysis based on the total irritation score of 5.6 with a observed SD of 1.2 obtained for gel-microemulsion control, a sample size of three, four, and six pigs/treatment group at 95% power was calculated to yield cutoff values for mucosal toxicity at 10.42, 9.31, and 8.38, respectively.
The long-term potential for histopathology changes in the female reproductive tract was evaluated in cats. Histopathological examination of H & E -stained vaginal, cervical, uterine, and Fallopian tube sections from tissues obtained at 18 weeks after a single application 2% WHI-07 plus 0.25% VDDTC via gel-microemulsion did not reveal scarring or fibrosis in all five cats examined. Only mild vaginitis, cervicitis, and cystic uterine glands were noted in 1 of 5 cats treated with WHI-07 plus VDDTC gel-microemulsion (Table 2). Notably, the histology of Fallopian tubes was within normal limits.
Cellular and Proinflammatory Cytokine Profile in the Porcine CVL Fluid
The cellular profile of PBS-flushed CVL fluid collected before and at 24, 48 and 72 hours after gel application was assessed by flow cytometry (Figure 6A–D). The pre CVL fluid collected from sexually mature nonestrus pigs (14-day post-treatment with altrenogest) lacked significant leukocyte population. The scatter profiles showed no significant increase in leukocyte population in CVL fluid obtained at 24, 48, and 72 hours following intravaginal exposure to WHI-07 plus VDDTC gel-microemulsion (Figure 6). Although, CVL fluid cells showed variation in the proportion of non-leukocyte cells, their presence had little effect on the light scatter characteristics of granulocyte populations. Time-course analysis revealed no significant increase in the vaginal granulocyte population in CVL fluid obtained for four pigs exposed to WHI-07 plus VDDTC (10.7 ± 1.7% versus 7.8 ± 2.9%, 9.1 ± 1.1%, and 13.1 ± 3.1, respectively) (Figure 7). Similarly, there was no significant increase in baseline levels of monocytes and lymphocytes (5.6 ± 3.7% versus 1.7 ± 0.2%, 1.7 ± 0.5%, and 2.0 ± 1.5%, respectively, for monocytes and 3.8 ± 0.7% versus 2.8 ± 0.8%, 3.2 ± 0.8%, and 4.6 ± 1.4%, respectively, for lymphocytes).
Although in vivo results using rabbit and porcine model indicated that at the highest concentration tested, WHI-07 plus VDDTC did not induce vaginal irritation, it was necessary to quantitate the levels of key proinflammatory cytokines in the porcine CVL fluid using porcine-specific mAbs. The cytokine levels, which are hallmarks for vaginal inflammation, were quantitated using a sensitive chemiluminscence-based multiplex approach. Time kinetics of simultaneous evaluation of four secreted porcine cytokines in CVL fluid by the sensitive chemiluminescence-based immunoassay showed no significant increase from baseline levels of IL-1β, IL-8, TNF-α and IFN-γ in pigs exposed to 2% WHI-07 plus 0.25% VDDTC via gel-microemulsion (Figure 8A–D). Under identical experimental conditions, CVL from 2% BZK-treated pigs showed significant increase (p < 0.05) in IL-1β, IL-8, and TNF-α levels at 24 hours when compared to the baseline controls and WHI-07 plus VDDTC-treated groups (Figure 8).
Evaluation of Immunoreactive Markers for Vaginal Inflammation
Because the number and distribution of CD45 positive cells within the vaginal mucosa is indicative of vaginal inflammation to test agents (Catalone et al., 2004), the rabbit vaginal mucosa was further evaluated for localized inflammation for the expression of CD45 and activated NFκB. By indirect immunofluorescence and CLSM, aggregates of CD45 immunoreactive cells were localized only to blood vessels in the lamina propria of all control and WHI-07 plus VDDTC-treated tissues examined (Figure 9A–D). Extravascular cells within the lamina propria were nonreactive in all treatment groups. No increased reactivity was apparent in the epithelial and stromal cells of gel-microemulsion-exposed and vaginal tissues from rabbits exposed to gel-microemulsion incorporating 0.5 + 0.06%, 1.0 + 0.12%, or 2.0 + 0.25% WHI-07 plus VDDTC for 14 consecutive days (Figure 9A–D).
Similarly, with mAb against the activated p65 subunit of NF-κB, no increased reactivity was apparent in the vaginal mucosa of tissues from rabbits repeatedly exposed to three increasing concentrations of WHI-07 plus VDDTC via gel-microemulsion. NF-κB-positive nuclei were localized only to blood vessels in the lamina propria of all vehicle control and WHI-07 plus VDDTC-treated tissues examined (Figure 10A–D).
Evaluation of Apoptosis in the Vaginal Mucosa
Although vanadocenes are potent apoptosis-inducing agents, their differential concentration-dependent spermicidal and apoptosis-inducing properties can be exploited for their development as vaginal spermicides (D’Cruz et al., 1998b, 2000a; D’Cruz and Uckun, 2000, 2001b). We used the in situ TdT-mediated labeling of 3′-OH termini with FITC-digoxigenin-conjugated UTP assay method to demonstrate whether repeated intravaginal administration of WHI-07 plus VDDTC results in increased apoptosis in the vaginal mucosa. By the TUNEL assay and CLSM, no increased reactivity was noted in the vaginal epithelial and lamina propria regions of vehicle control and WHI-07 plus VDDTC-exposed rabbit tissues (Figure 11). Tissue sections from rabbits repeatedly exposed to increasing concentrations of WHI-07 plus VDDTC contained only a low incidence (range 0.9% to 3.9%) of TUNEL-positive apoptotic cells in the epithelium and lamina propria (Table 3). Confocal laser scanning micrographs revealed lack of increased apoptotic cells in the vaginal tissues repeatedly exposed to gel-microemulsion (Figure 11A) or to three increasing concentrations of WHI-07 plus VDDTC (Figure 11B–D). Under identical experimental conditions, vaginal tissue section treated with DNAse prior to TUNEL assay revealed intense staining of the majority of cell nuclei (Figure 11E).
Evaluation of Vanadium Retention in Tissues and Body Fluids
Table 4 shows the results from the atomic absorption spectroscopy analysis of the bone, heart, kidney, liver, lung, muscle, spleen, vagina, cervix, ovary as well as blood and urine or bladder of control and 2.0% WHI-07 plus 0.25% VDDTC-treated rabbits and/or pigs. Despite daily intravaginal administration of 2.0% WHI-07 plus 0.25% VDDTC for 14 and 6 consecutive days, respectively, the vanadium content of all tissues and body fluids examined was <1 μg/g. The mean vanadium content in control rabbit tissues and body fluids ranged from 0.065 to 0.213 μg/g. The mean vanadium content in tissues and body fluids of dosed rabbits given 2.0% WHI-07 plus 0.25% VDDTC ranged from 0.065 to 0.456 μg/g. Similarly, the mean vanadium content in tissues and blood of pigs given intravaginal 2.0% WHI-07 plus 0.25% VDDTC gel-microemulsion was essentially undetectable.
Discussion
The present study based on the comparative histologic data and surrogate markers for inflammation demonstrated that repeated intravaginal administration of the microbicidal nucleoside prodrug WHI-07 in combination with a vanadocene VDDTC via a gel-microemulsion did not result in vaginal irritation, mucosal toxicity, or systemic absorption of vanadium in the rabbit and porcine model. Furthermore, intravaginal application of a high dose of WHI-07 plus VDDTC was not associated with long-term inflammatory-associated obstructions in the reproductive tract of cats. Consequently, the combined use of WHI-07 and VDDTC via gel-microemulsion appears safe for topical use as a prophylactic anti-HIV-1 microbicide.
WHI-07, a novel spermicidal prodrug of ZDV was designed to bypass the rate-limiting TK dependency of ZDV activation as well as protect the parent drug from hydrolysis before releasing the drug intracellularly (D’Cruz et al., 1998a, 1999, 2001a). Unlike, ZDV, WHI-07 exhibits enhanced lipophilicity, superior pharmacokinetics, and retains full anti-HIV activity in TK-deficient cells. VDDTC, a chelated vanadocene complex, is an effective spermicide as well as anticancer agent (D’Cruz et al., 1999; Ghosh et al., 2000). Vanadocenes rapidly immobilize sperm without affecting the sperm membrane integrity or viability of normal human vaginal or cervical epithelial cells (D’Cruz et al., 1998a,b, 1999, 2000a,b). Consequently, the spermicidal and antileukocytic effects of WHI-07 and VDDTC are particularly useful for eliminating cell-associated HIV-1 in semen and genital tract secretions (Uckun et al., 2005).
Notably, the spermicidal properties of WHI-07 and VD-DTC fundamentally differ from those of currently used membrane-active surfactant spermicides such as N-9 and BZK that are cytotoxic to human genital epithelial cells at spermicidal concentration (Chantler, 1992; D’Cruz et al., 1999, 2000a,b). Clinical failure of currently used surfactant spermicides is attributed to subclinical inflammatory lesions and attraction of target cells to mucosal tissues (Roddy et al., 1993; Check, 2007). Activation of leukocytes that accumulate at the site of inflammation results in the production of cytokines (Cohen et al., 2004). Proinflammatory cytokines, IL1-β, IL-6, TNF-α, and chemokines such as IL-8, macrophage inflammatory protein (MIP)-α, MIP-β and RANTES (regulated on activation, normal T cell expressed and secreted), have been detected in genital secretions of women (Belec et al., 1995; Sha et al., 1997). Cervicovaginal epithelial cells also secrete cytokines and chemokines (Fichorova and Anderson, 2000).
These cytokines have different effects on HIV-1 replication. Proinflammatory cytokines stimulate HIV-1 replication through activation of NF-κB (Pantaleo et al., 1993; Rasmussen et al., 1997). Conversely, chemokines, such as MIP-α, MIP-β and RANTES, block HIV-1 infection by interfering with attachment to the viral co-receptor (Deng et al., 1996; Dragic et al., 1996). The relative concentration of these cytokines and chemokines could influence HIV-1 expression and transmission in a positive or negative fashion. Women exposed to N-9 develop high levels of inflammation in the vagina and increased levels of proinflammatory cytokines in CVL fluid (Fichorova et al., 2001). N-9 and BZK induce a cumulative inflammatory reaction by IL-1 mediated NF-κB activation in cervicovaginal epithelial cells. Repeated N-9 use is associated with inflammatory sequelae, increased levels of IL-1, IL-8, MIP-1 and macrophage/neutrophil influx. Clinical studies have confirmed that detergent-type spermicides lead to increased risks of contracting sexually transmitted diseases (Hooten et al., 1991; Kreiss et al., 1992; Mati et al., 1995; Rosenstein et al., 1998; Stafford et al., 1998; Moszynski, 2007). Consequently, an animal model, which can accurately imitate vaginal irritation potential of microbicides and spermicides with consistent and predictable outcome prior to expensive human clinical trials, is highly desirable.
The simple columnar epithelium of rabbit vagina is highly sensitive to mucosal irritants when compared to the stratified squamous epithelia of human vagina (Patton et al., 2000). Accordingly, this model has been used extensively in the pharmaceutical industry because of earlier reports on the correlation between rabbits and humans with respect to the irritation potential of vaginal products and remains the FDA recommended model for safety evaluation of vaginal products (Eckstein et al., 1969). A predetermined semi quantitative scoring system is used to assess changes in the continuity of the epithelial lining, edema of the submucosal layer, and the inflammatory cell infiltrate of this layer after exposure to increasing doses of the test agent for 5 to 14 days. One limitation of the RVI model is that two thirds of the rabbit vaginal mucosa is lined by columnar epithelium that is structurally distinct from the stratified, squamous epithelium of the human vagina. In addition, the classical histologic endpoints for inflammation established in 1969 using this model for the development of vaginal spermicides is not rigorous enough for the development of safe vaginal microbicides.
We previously established a physiologically relevant and sensitive porcine model to investigate the mucosal toxicity potential of candidate microbicides and spermicides (D’Cruz et al., 2005). The porcine and human vaginal epithelia from various anatomical sites consist of a stratified, squamous epithelium that is supported by connective tissue lamina propria (Bal and Getty, 1972). Porcine and human mucosal epithelia share similarities in lipid composition, histologic condition, and ultrastructural organization that are determinants of mucosal barrier function (Swartzendruber et al., 1989). The porcine vaginal epithelium is nonkeratinized and shares a similar structural organization, which makes it a good surrogate for the human tissue (Kremer et al., 2001; Thompson et al., 2001; Davis et al., 2003). Additionally, this model allows the examination of immune cell activation and cytokine release using species-specific reagents as surrogate markers for genital inflammation. Our recent studies established that porcine CVL cytokines can predict mucosal irritation and can be used to rank vaginal microbicides and spermicides (D’Cruz et al., 2005).
When grading epithelial ulceration, inclusion of epithelial cell layers, cell type and changes within the epithelia (i.e., vacuolization/crypt formation), can be more specific grading of subclinical changes. Because, vaginal epithelium, when ulcerated or sloughed, can rapidly regenerate, cell proliferation is a valid criterion for epithelial cell damage that can occur from chronic, persistent irritation. The 10 morphologic endpoints chosen for mucosal toxicity based on the observed effects of a 2% BZK gel separates leukocytes into acute and subacute/chronic categories by grading both neutrophils and mononuclear leukocytes. Neutrophils and suppuration signify bacterial infection.
Addition of fibrinous necrosis and hemorrhage within the vaginal tissues gives more pathological meaning to vascular/perivascular changes when congestion and edema are separated. Extension of irritation/inflammation changes from the subepithelial connective tissue to the surrounding and underlying skeletal muscle signifies more widespread pathology and thus is significant to positive findings. This further enhances the 10 criteria for vaginal pathology. These preclinical endpoints are statistically more significant when pathological changes are found. Vaginal irritation or infection can affect all 4 regions (epithelium, subepithelium/lamina propria, vessels/perivascular tissues, and underlying/surrounding muscles). The endometrial lining of the porcine uterus showed variable sloughing, erosion and/or ulceration. This is unlikely related to the test agent and most likely due to physiologic changes that occur during the estrus cycle since the vaginal and cervical histology from all treated gilts, appeared normal.
Additionally, a number of genital tract infections also involve the upper genital tract endocervical canal, endometrial cavity and the Fallopian tubes). The pig model provides a large surface area and longer retentive capacity of the gel to determine the potential toxic effect of intravaginally applied spermicides and microbicides on the upper tract. Recent studies suggest that detergent-based spermicides induce apoptosis in the upper female reproductive tract (Jain et al., 2005). These findings imply that vaginally administered topical agents may cause upper reproductive tract toxicity and possibly facilitate HIV-1 infection. The entire reproductive tract is a potential target for HIV-1 uptake and systemic spread. The free virus and/or cell-associated virus present in semen can infect at any site along the tract independently of lesions or sites of trauma. Since small numbers of poorly motile sperm are found in the Fallopian tubes within minutes of vaginal deposition (Settlage et al., 1973), both the virus and virus-infected cells could establish an initial infection in cells within the Fallopian tubes. Indeed, purified cultures of both epithelial and stromal cells from the Fallopian tube and uterus can be infected directly with HIV-1 (Howell et al., 1997).
In the present study, histopathological evaluation of vagina, cervix, uterus, and Fallopian tube from pigs as well as the TUNEL assay in the rabbit tissues revealed lack of histo-logic alterations as well as apoptotic cells in vaginal tissues repeatedly exposed to combination of WHI-07 and VDDTC. Because the contraceptive effect can also be induced by local inflammation and chemical scarring of endometrium and Fallopian tubes, we tested the potential long-term effect of intravaginally administered WHI-07 plus VDDTC in the cat reproductive tract. Trancervical insertion of the sclerotizing agent quinacrine chloride results in chemical scarring of Fallopian tubes as a non-surgical permanent contraceptive method (Zipper et al., 1970). Following a single application of a high dose of the WHI-07 plus VDDTC, histologic evaluation of the entire reproductive tract 18 weeks after test agent exposure revealed lack of inflammation-related changes in the cat reproductive tract.
Furthermore, in molecular genotoxicity studies, WHI-07 and vanadocene complexes did not increase the DEL recombination frequency in yeast nor did they activate any of the DNA damage-associated promoters in HepG2 cells using the CAT-Tox(L) assay (Aubrecht et al., 1999). The inflamma-tory cell marker CD45, accumulates preferentially at sites of inflammation with greater ability of CD45-positive cells to adhere to endothelium and form clusters (Pitzalis et al., 1988).
In the rabbit vagina, N-9 mediated leukocyte infiltration has been shown to correlate with CD45-positive cells in the submucosa and NFκB activation by the vaginal epithelium (Doncel et al., 2004). In the present study, no increased immunoreactive CD45-positive cells were detected in the lamina propria of WHI-07 plus VDDTC-treated rabbit vaginal mucosa. The absence of extravascular CD45-positive cells in the lamina propria also suggested the lack of vaginal inflammation. We also investigated whether NF-κB active complex occurred in the vaginal mucosa following repeated topical exposure to WHI-07 plus VDDTC. The NFκB is involved in the inducible expression of a variety of cellular genes that regulate inflammatory response (Umansky et al., 1998). IL-1β and TNF-α activate signaling pathways that stimulate transcription factors AP-1 and NF-κB and thereby induce the expression of genes coding for inflammatory mediators (Stanocovski and Baltimore, 1997; DiDonato et al., 1997). The lack of immunoreactive cells against CD45 and activated NF-κB antigens observed in the vaginal tissues is consistent with the lack of induction of proinflammatory mediators by WHI-07 plus VDDTC in the porcine CVL fluid. Based on the preclinical results reported here, it is highly unlikely that repetitive intravaginal application of WHI-07 plus VDDTC will have any adverse systemic side effects in the clinical setting.
In summary, our comparative in vivo mucosal irritation studies using rabbit, porcine, and cat models suggest that combined use of WHI-07 and VDDTC via gel-microemulsion is unlikely to induce mucosal injury or apop-tosis in the female reproductive tract. Vanadium from the vanadocene complex is not preferentially absorbed systemically or incorporated into organs following repeated intrav-aginal administration. Consequently, the combined use of WHI-07 and VDDTC via gel-microemulsion appears safe for topical use as a prophylactic anti-HIV-1 microbicide.
Footnotes
Acknowledgments
O. J. D. is supported by research grants HD042884 and HD042889 from the National Institutes of Child Health and Human Development, Bethesda, MD. The authors thank Barbara Waurzyniak, D.V.M. and Douglas Erbeck, D.V.M, Ph.D., for the histological grading of rabbit, porcine, and cat reproductive tract tissues.