Research ArticleNanomedicine

Mucus-Penetrating Nanoparticles for Vaginal Drug Delivery Protect Against Herpes Simplex Virus

See allHide authors and affiliations

Science Translational Medicine  13 Jun 2012:
Vol. 4, Issue 138, pp. 138ra79
DOI: 10.1126/scitranslmed.3003453

Abstract

Incomplete coverage and short duration of action limit the effectiveness of vaginally administered drugs, including microbicides, for preventing sexually transmitted infections. We investigated vaginal distribution, retention, and safety of nanoparticles with surfaces modified to enhance transport through mucus. We show that mucus-penetrating particles (MPPs) provide uniform distribution over the vaginal epithelium, whereas conventional nanoparticles (CPs) that are mucoadhesive are aggregated by mouse vaginal mucus, leading to poor distribution. Moreover, when delivered hypotonically, MPPs were transported advectively (versus diffusively) through mucus deep into vaginal folds (rugae) within minutes. By penetrating into the deepest mucus layers, more MPPs were retained in the vaginal tract after 6 hours compared to CPs. After 24 hours, when delivered in a conventional vaginal gel, patches of a model drug remained on the vaginal epithelium, whereas the epithelium was coated with drug delivered by MPPs. We then developed MPPs composed of acyclovir monophosphate (ACVp). When administered before vaginal herpes simplex virus 2 challenge, ACVp-MPPs protected 53% of mice compared to only 16% protected by soluble drug. Overall, MPPs improved vaginal drug distribution and retention, provided more effective protection against vaginal viral challenge than soluble drug, and were nontoxic when administered daily for 1 week.

Introduction

Improved methods for sustained and more uniform drug delivery to the vagina may provide more effective prevention and treatment of conditions that affect women’s health, such as cervical cancer, bacterial vaginosis, and sexually transmitted infections. For example, women are disproportionately infected with HIV, partly owing to a lack of female-controlled prevention methods (1, 2). An easily administered, discreet, and effective method for protecting women against vaginal HIV transmission could prevent millions of infections worldwide. However, vaginal folds, or “rugae,” which accommodate expansion during intercourse and childbirth, are typically collapsed by intra-abdominal pressure, making the surfaces of these folds less accessible to drugs and drug carriers (3). Poor distribution into the vaginal folds, even after simulated intercourse, has been cited as a critical factor for failure to protect susceptible vaginal surfaces from infection (4). Distribution over the entire susceptible target surface has been proven important for preventing and treating infections (59). Additionally, to increase user acceptability, drug delivered to the vagina should be retained in the vaginal tract at effective concentrations over extended periods of time. Achieving sustained local drug concentrations is challenging because the vaginal epithelium is highly permeable to small molecules (10) and also because soluble drug dosage forms (gels, creams) can be expelled by intra-abdominal pressure and ambulation (7, 9). Last, drug delivery methods must be safe and nontoxic to the vaginal epithelium. Improvements in the distribution, retention, and safety profile of vaginal dosage forms may lead to a substantial increase in efficacy and decrease in the side effects caused by largely ineffective systemic treatments for cervicovaginal infections and diseases (11, 12).

Nanoparticles have received considerable attention owing to their ability to provide sustained local drug delivery to the vagina (1, 13). However, the mucus layer coating the vaginal epithelium presents a barrier to achieving uniform distribution and prolonged retention in the vaginal tract. Mucus efficiently traps most particulates, including conventional polymeric nanoparticles (CPs), through both adhesive and steric interactions (14). The efficiency with which mucus traps foreign pathogens and particulates implies that CPs would become trapped immediately upon contact with the lumenal mucus layer, preventing penetration into and, thus protection of, the rugae. Particles and pathogens trapped in the superficial lumenal mucus layer would be expected to be rapidly cleared from the tissue, limiting the retention time of mucoadhesive materials, such as CPs (14).

By mimicking viruses that have evolved to penetrate the mucus barrier to establish infection, we recently engineered mucus-penetrating particles (MPPs) for mucosal drug delivery by coating CPs with an exceptionally high density of low–molecular weight poly(ethylene glycol) (PEG) (1517). MPPs diffuse through human cervicovaginal mucus (CVM) at speeds comparable to their theoretical diffusion through water (16, 17). Here, we sought to test the hypothesis that MPPs would provide enhanced distribution and increased retention in vivo in the vagina by penetrating into the deepest mucus layers, including the more slowly cleared mucus in the rugae, thereby releasing drug in the optimal location for efficient tissue uptake (Fig. 1A). In addition to the common progestin-induced diestrus phase (DP) mouse model, we introduce an estradiol-induced estrus phase (IE) mouse model in which the mouse CVM (mCVM) more closely mimics human CVM (hCVM) and therefore provides a more human-like model for developing and translating MPPs for human use.

Fig. 1

Characterization of CPs and MPPs. (A) Graphical depiction of vaginal drug delivery from gel, CP, and MPP formulations. (B) Representative trajectories for particles exhibiting effective diffusivities within one SEM of the ensemble average at a time scale of 1 s. (C) Ensemble-averaged geometric mean square displacement (〈MSD〉) as a function of time scale. Data for particles on ex vivo estrus phase mouse vaginal tissue (mCVM) compared to the same particles in ex vivo hCVM (20) and the theoretical diffusion rate of 110-nm particles in water. (D) Percentage of particles capable of penetrating a 100-μm-thick layer of mCVM over time, on the basis of Fick’s second law of diffusion simulation of particles undergoing random diffusion. (E) Distributions of the logarithms of individual particle effective diffusivities (Deff) at a time scale of 1 s. Data represent the average of three independent experiments, with n ≥ 150 particles. Diffusivity values to the left of the dotted line indicate particles with MSD values less than the particle diameter. (F and G) 〈MSD〉 as a function of time scale, comparing MPPs to CPs (F) as well as MPPs to BD-MPPs (G) in estrus and IE mice. The theoretical diffusion rate of 110-nm particles in water is included. Data in (C), (D), (F), and (G) are means ± SEM (three independent experiments, n ≥ 130 particles).

Results

Transport of nanoparticles on mouse vaginal tissue ex vivo

Carboxylic acid–coated, fluorescent polystyrene nanoparticles (PS-COOH) were made into MPPs by covalently attaching a dense coating of low–molecular weight PEG, as previously reported (15, 16). Additionally, biodegradable MPPs (BD-MPPs) were formulated with a poly(lactic-co-glycolic acid) (PLGA) core and a physically adsorbed PEG coating, as previously reported (18), because biodegradable particles can be loaded with drugs and are suitable for dosing to humans. PS-COOH and PLGA nanoparticles have a highly negative surface charge, which is nearly neutralized when densely coated with PEG. Nanoparticles were determined to be well coated by measuring the ζ potential (Table 1). A ζ potential more neutral than −10 mV was previously found to be necessary for mucus-penetrating properties in hCVM (15).

Table 1

Particle characterization. Data are means ± SEM (n = 3).

View this table:

To ensure that MPPs were mucus-penetrating in native estrus phase mCVM, we administered the particles intravaginally to mice in the estrus phase. The entire vagina was then excised and opened to visualize the motions of hundreds of individual particles with a multiple particle–tracking method (19) developed in our lab (Supplementary Materials and Methods). Particle trajectories for MPPs were indicative of rapid diffusion through watery pores in the mCVM, whereas motions of uncoated PS-COOH nanoparticles (CPs) were smaller than the particle diameter (~100 nm) (Fig. 1B). The ensemble-averaged mean square displacement (〈MSD〉) of MPPs in mCVM was found to be comparable to that reported for MPPs in hCVM (20) (Fig. 1C), corresponding to ensemble-averaged effective diffusivity (〈Deff〉) only about eightfold slower than the theoretical diffusion of 110-nm particles in water (~4 μm2/s). On the basis of the measured Deff for individual particles, we estimated with Fick’s second law of diffusion that about half of the MPPs would diffuse through a 100-μm-thick layer of mCVM in about 4 hours, whereas even after 24 hours, there would be no appreciable penetration by CPs (Fig. 1D). Deff values for CPs at a time scale of 1 s corresponded to MSD values less than the particle diameter (Fig. 1E, dotted line), likely revealing thermal fluctuation of particles stuck to mucin fibers and not particle diffusion. Overall, the transport behavior of both MPPs and CPs in estrus phase mCVM was similar to their transport behavior in hCVM.

Synchronizing a large number of mice in the estrus phase for retention studies required hormonal treatment. Particle transport behavior was tested in IE mice to confirm that estradiol treatment, which has been used routinely for inducing estrus-like behavior in many animal models (21, 22), did not alter MPP and CP transport behavior before distribution and retention studies (Fig. 1F). Additionally, BD-MPP transport behavior was indistinguishable from that of MPPs in IE mucus (Fig. 1G).

Distribution of nanoparticles in the vagina

We next investigated in the estrus phase mouse and IE mouse whether the ability to rapidly penetrate mucus would lead to more rapid and uniform vaginal distribution of MPPs compared to CPs. We applied MPPs and CPs in hypotonic medium to mimic the way osmotically driven water flux (advective transport) rapidly transports nutrients from the intestinal lumen to the brush border epithelial surface. Ten minutes after particle administration, the entire vagina was dissected out and stained for cell nuclei. CPs aggregated in the lumenal mucus and did not penetrate into the vaginal rugae (Fig. 2A). In contrast, MPPs—both nonbiodegradable and biodegradable—formed a continuous particle layer that coated the entire vaginal epithelium, including all the surfaces of the rugae. MPPs penetrated more than ~100 μm of mucus via advection within 10 min compared to the ~4 hours it would take them to diffuse that distance through mucus (Fig. 1E). This behavior was also consistent for BD-CPs and BD-MPPs, and CPs and MPPs administered to IE mice (Fig. 2A). Videos illustrating the movement of MPPs through hCVM past mucoadhesive CPs can be found in video S1 (no flow, diffusion) and video S2 (with flow, advection). To further characterize the effects of the mucus barrier, we found that removing vaginal mucus by lavage (23, 24) before particle administration (Supplementary Materials and Methods) markedly improved CP distribution, indicating that their mucoadhesive character prevents uniform distribution in the vagina (Fig. 2B).

Fig. 2

Particle distribution in the mouse vagina. (A) Distribution of red fluorescent nonbiodegradable and biodegradable CPs and MPPs in transverse cryosections of estrus phase and IE mouse vaginal tissue. (B) Distribution of red fluorescent nonbiodegradable and biodegradable CPs in transverse cryosections of mouse vaginal tissue with either an intact mucus layer (no treatment) or mucus removed by lavage and swabbing (mucus removed). Images are representative of n ≥ 3 mice.

To quantify the difference in distribution of MPPs and CPs, we obtained fluorescence images of freshly excised, opened, and flattened mouse vaginal tissue. The adhesion of CPs to lumenal vaginal mucus layers created “stripes” of mucus with particles alternating with dark stripes of mucus without particles, the latter corresponding to the rugae that were opened when the vaginal tissue was flattened (Fig. 3A). In contrast, transport of MPPs toward the epithelium and into the rugae created a continuous particle coating on the flattened vaginal surface (Fig. 3A). Quantification of the fluorescence on the vaginal and ectocervical tissue indicated that 88% of the flattened vaginal surface and 87% of the ectocervical surface were densely coated with MPPs, whereas only 30% of the vaginal surface and 36% of the ectocervical surface were coated with CPs. Upon further inspection at higher magnification of darker areas of the vaginal and ectocervical surfaces, a continuous, less-concentrated coating of MPPs was seen (Fig. 3A, insets), implying that there was nearly complete coverage of the vaginal and ectocervical epithelium. For CPs, a less-concentrated coating was not found at higher magnification (Fig. 3A, insets). Similar trends were found with BD-MPPs, with 85% vaginal coverage and 86% ectocervical coverage, as well as BD-CPs, with 31% vaginal coverage and 27% ectocervical coverage (Fig. 3A).

Fig. 3

Quantification of nanoparticle and drug coverage in vagina and ectocervical tissue. (A) Distribution of red fluorescent nonbiodegradable and biodegradable CPs and MPPs on flattened estrus phase mouse vaginal and ectocervical tissue. Insets are images of dark areas at higher magnification. (B) Distribution and retention of a model drug, FITC, in the estrus mouse vagina delivered in gel form or encapsulated in BD-MPPs. Fluorescence images of flattened mouse vaginal tissue after 24 hours. All images are representative of the averages calculated for n ≥ 3 mice and were quantified as percent coverage ± SEM. *P < 0.05 compared to CPs or FITC/gel, Student’s t test.

We then sought to determine whether the improved distribution of BD-MPPs could improve the delivery of small molecules compared to a gel dosage form. Lipophilic molecules are likely to enter the first epithelial surface they contact, failing to contact cells in the rugae. Conversely, hydrophilic molecules can diffuse rapidly through the vaginal epithelium and be carried away by blood and lymph circulation, leading to brief periods of coverage. We loaded BD-MPPs with a fluorescent, water-soluble small molecule, fluorescein isothiocyanate (FITC), as a model drug (FITC/MPPs). To mimic conventional vaginal delivery, we administered soluble FITC (FITC/gel) in the universal vaginal placebo gel hydroxyethylcellulose (HEC). Twenty-four hours after administration to estrus phase mice, the vaginal tissues were excised and flattened to expose the vaginal folds. Patches of FITC coated 42% of the vaginal surface when administered as FITC/gel, whereas FITC/MPPs provided a well-retained FITC coating of 87% of the vaginal surface (Fig. 3B), even 24 hours after particle administration.

Retention of nanoparticles in the vagina

Using our IE model, we next sought to determine vaginal retention of MPPs compared to mucoadhesive CPs. Fluorescent MPPs and CPs were administered intravaginally to IE mice. At specified time points, the entire reproductive tract (vagina and uterine horns) was excised and analyzed quantitatively with fluorescence imaging (Fig. 4A). After an initial decrease in particle fluorescence that was similar for MPPs and CPs (likely owing to initial “squeeze out” preceding mucus penetration), the remaining amount of MPPs stayed constant at about 60% (Fig. 4B). In contrast, the amount of CPs steadily decreased with time to 10% (6 hours). Although CPs were distributed along the length of the vagina, this longitudinal coverage did not indicate that CPs penetrated mucus to reach the epithelium, nor surfaces inside the vaginal folds, as shown in Fig. 2A.

Fig. 4

Retention of nonbiodegradable MPPs and CPs in the IE mouse cervicovaginal tract. (A) Overlay of particle fluorescence intensity and photographic images for CPs and MPPs in whole cervicovaginal tract tissue. Images chosen are representative of the average at each time point. (B) Fraction of particles remaining over time on the basis of quantification of particle fluorescence in (A). Data are means ± SEM (n ≥ 7). *P < 0.05 compared to CPs, Student’s t test.

Nanoparticle toxicity in the progestin-treated diestrus mouse vagina

The immune system is highly active at mucosal surfaces (25), especially those with surfaces covered with living cells, such as columnar epithelia in the endocervix in humans. Inflammatory effects of nanoparticles were investigated using mice pretreated with Depo-Provera (medroxyprogesterone acetate), a long-acting progestin treatment that synchronizes mice in a diestrus-like state, during which the vaginal epithelium thins and becomes covered with living cells (this is important for experiments lasting 24 hours or longer). In contrast, in estrus, the mouse vagina thickens from 4 to 7 cell layers to about 12 cell layers, and the epithelial surface is protected with many layers of dead and dying cells (26). Additionally, the DP mouse vaginal epithelium has an increased immune cell population, leading to enhanced acute inflammatory responses, whereas the estrus phase is characterized by an absence of immune cells (27).

Standard hematoxylin and eosin (H&E) staining was used to investigate potential toxic effects of intravaginally administered nanoparticles (Supplementary Materials and Methods). Nonoxynol-9 (N9), a nonionic detergent known to cause vaginal toxicity (28), was used as a positive control, and phosphate-buffered saline (PBS) (saline) was used as a negative control. The same (BD-)MPPs and (BD-)CPs that were used for distribution and retention studies were tested for toxicity. As expected, N9 caused acute inflammation at 24 hours that was not seen after PBS treatment (Fig. 5A). CPs, like N9, caused pronounced neutrophil infiltration into the lumen, but MPPs did not cause this inflammatory effect (Fig. 5A).

Fig. 5

Acute toxicity and cytokine concentrations with daily administration. (A) H&E-stained cross-sections of mouse DP vaginal tissue excised 24 hours after intravaginal administration of 5% N9, PBS, CPs, MPPs, BD-CPs, and BD-MPPs. Scale bar applies to all images. Arrowheads point to clusters of neutrophils (stained purple). Images are representative of n ≥ 5 mice. (B) Cytokine concentrations in DP mouse cervicovaginal lavage after daily vaginal treatments for 7 days. Data are means ± SEM. *P < 0.05 compared to ACVp-MPPs, Student’s t test.

Cytokine release with repeated vaginal application

Recent studies indicate that in response to certain vaginal products, the vaginal epithelium can secrete immune mediators that may enhance susceptibility to sexually transmitted infections (29, 30). Thus, it is important that a vaginal product not induce such an immune response, particularly after repeated dosing. Because our ultimate goal was to test MPPs for protection against herpes simplex virus 2 (HSV-2), we compared an MPP formulation containing acyclovir monophosphate (ACVp) to N9, HEC placebo gel, PBS, and a gel vehicle (TFV vehicle) used in recent tenofovir clinical trials. Nanoparticle and control formulations were administered vaginally to Depo-Provera–treated mice daily for 7 days. Vaginal lavages were collected on day 8 from each mouse and assessed for cytokines that have been found to be elevated in response to epithelial irritation: interleukin-1β (IL-1β), IL-1α, tumor necrosis factor–α (TNF-α), and IL-6. It was found that both IL-1α and IL-1β levels were elevated in response to both the TFV vehicle and the N9 solution (Fig. 5B). This was not surprising in the case of N9 treatment, considering that IL-1α and IL-1β are secreted by the vaginal epithelium in response to injury (29). In contrast, the cytokine levels associated with ACVp-MPPs were equivalent to the levels associated with HEC placebo gel (Fig. 5B), which has been used in clinical trials without any associated increase in susceptibility to infection (31, 32). There was no detectable elevation of either IL-6 or TNF-α associated with any vaginal treatment compared to untreated controls.

Vaginal protection against HSV-2 in the progestin-treated diestrus mouse

We finally investigated whether the improved distribution, retention, and toxicity profile of MPPs would lead to improved protection against vaginal HSV-2 challenge in mice. Depo-Provera treatment markedly increases the vaginal susceptibility of mice to infections, and candidate microbicides have provided only partial protection in the mouse model used here, even when administered immediately before the infectious inoculum (4, 33). Moreover, several vaginal product excipients actually increase susceptibility to infection in this model (34, 35). We chose to test ACVp for blocking vaginal transmission of HSV-2 infections, because acyclovir provides viral suppression in animals with repeated dosing multiple times per day (36). However, a single vaginal pretreatment with ACVp at 50 mg/ml (5%) in guinea pigs resulted in 70% of animals infected compared to controls (37). Therefore, ACVp provided a test case to determine whether MPPs could significantly improve protection by a water-soluble and quickly metabolized drug by prolonging therapeutically relevant drug concentrations after a single application. Additionally, the mechanism of action of nucleotide analogs, such as ACVp, is prevention of intracellular viral replication, such that successful protection implies efficient uptake and retention in susceptible target cell populations in the vaginal and cervical mucosa.

We formulated ACVp nanoparticles with the same muco-inert coating used for all other studies. The size and ζ potential of ACVp-MPPs were similar to those of PS-based MPPs (Table 1). Mice were administered soluble ACVp or ACVp-MPPs intravaginally 30 min before HSV-2 challenge. Soluble drug administered at the same concentration as the ACVp-MPPs (1 mg/ml) was ineffective at protecting mice from viral infection (84.0% infected compared to 88.0% of controls), whereas only 46.7% of mice in the ACVp-MPPs group were infected (Table 2). Groups of mice given soluble drug at 10 times the concentration in ACVp-MPPs were still infected at a rate of 62.0% (drug in PBS) or 69.3% (drug in water). Comparing soluble drug to ACVp-MPPs in the same vehicle (pure water), soluble drug was significantly less protective, even at 10-fold higher concentration than ACVp-MPPs (Table 2).

Table 2

HSV-2 vaginal challenge results. Infections in DP mice resulting from vaginal challenge with HSV-2 30 min after drug administration. Statistical significance was determined using Fisher’s exact test, two-tailed distribution. N/A, not applicable.

View this table:

Discussion

The female reproductive tract is susceptible to a wide range of sexually transmitted infections (1). Biological vulnerability, a lack of female-controlled prevention methods, and inability to negotiate condom use all contribute to male-to-female transmission worldwide (1, 2). An easily administered, discreet, and effective method for protecting women against vaginal HIV, HSV-2, and other virus transmission could prevent millions of infections worldwide. After 11 unsuccessful microbicide trials, CAPRISA 004 was the first to demonstrate partial protection against HIV with a vaginally administered microbicide (tenofovir) in a gel formulation (31). An important difference between previous-generation microbicides, such as N9, and the current generation of microbicides is the site of action. Many current-generation microbicides, such as nucleotide analogs tenofovir and ACVp, work intracellularly to inhibit viral replication, whereas previous generations directly inactivated pathogens in the vaginal lumen. However, some previous-generation microbicides caused toxicity to the vaginal epithelium that increased susceptibility to infection (38).

For vaginal drug delivery to be maximally effective, topically delivered drugs must be distributed uniformly, maintained at sufficiently high concentration, and remained in close proximity to the folded vaginal epithelium (rugae) and cervical mucosa. Several techniques have been used to observe distribution of gels and drugs after vaginal administration, such as magnetic resonance imaging (6), γ-scintigraphy (6, 9), colposcopy (8), and fiber optics (6). These techniques are adequate to observe gross distribution along the vaginal tract, but do not reveal entry into vaginal folds. Our work demonstrates that, although a topical treatment may be well distributed longitudinally along the vaginal tract, much of the folded epithelium can be left untreated and unprotected. Such untreated surfaces could have contributed to recent failures of several candidate microbicides against HIV in clinical trials (39). Additionally, when a fluid or gel is administered to the vagina, it directly contacts the rapidly shed outer lumenal mucus layer. Mucoadhesive particles, such as CPs, are trapped in this superficial mucus layer and thereby excluded from the rugae. In contrast, we demonstrated that MPPs can penetrate deep into the mouse rugae and, when delivered hypotonically, provided complete coverage of the epithelium within only 10 min.

Diffusion of particles is not rapid enough to result in such a uniform epithelial coating within minutes. Diffusion over ~100 μm would take on the order of hours. However, the vaginal epithelium has a great capacity for fluid absorption induced by osmotic gradients. Absorption of water through the mucus barrier assists MPPs in rapidly reaching the entire epithelial surface by advection, where the drug payload can then be released for optimal tissue uptake. In contrast, water absorption was not beneficial for CPs, because they became adhesively trapped and immobilized in the lumenal mucus (video S2).

Inadequate retention of therapeutically active compounds in the vaginal tract is another limiting factor for vaginal protection. For example, many vaginal spermicides provide protection for no more than 1 hour (40). Other vaginal products are not well retained even after 6 hours (79), necessitating repeated administration for adequate protection. Similarly, more than 90% of CPs were shed from the vagina within 6 hours because they did not penetrate deep into the mucus layers. In contrast, MPPs provided enhanced delivery of an encapsulated model drug (FITC) for at least 24 hours compared to soluble drug in a gel formulation. Thus, MPPs may provide a means for achieving potent, once-daily, topical vaginal administration for treatments such as microbicides against sexually transmitted diseases.

In previous attempts to develop mucosal drug delivery systems for the vaginal tract, a variety of “pretreatments” have been used that diminish the mucus barrier. Administering fluids (23, 24, 41), swabs (24, 42), or degradative enzymes (43) before administration of mucoadhesive delivery vehicles was likely essential to the drug or gene delivery achieved in these studies. Here, we found that a lavage plus swab pretreatment markedly improved distribution of CPs in the vagina, allowing the particles to coat the epithelium similarly to MPPs (Fig. 2B). Barrier-removing pretreatments may be impractical for human use and especially inappropriate for microbicides intended to prevent sexually transmitted diseases. Healthy CVM itself is a somewhat effective barrier to viral infections (44). We show that effective epithelial coverage can be achieved by use of MPPs, without the need to degrade or remove the mucus barrier.

PEG coatings have been widely used in developing polymeric drug carriers that are not easily recognized by the immune system (17). We demonstrated that dense PEG coatings produce MPPs that rapidly penetrate mucus without causing inflammation in the mouse vaginal tract. In contrast, administration of uncoated CPs resulted in an acute inflammatory response similar to administration of N9. Additionally, cytokine levels associated with daily administration of MPPs were indistinguishable from HEC placebo gel. Elevated levels of IL-1α and IL-1β, which are associated with epithelial injury, occurred after daily dosing with both N9 and TFV vehicle gel. The tenofovir-containing version of this gel was shown to have complete protection against HIV in a tissue explant model, and complete protection occurred despite visible epithelial shedding (45). Previous work suggests that glycerol in the TFV gel may be responsible for the observed toxicity in mice (35).

Mice are useful animal models for developing vaginal products, but there are key differences in vaginal physiology between mice and humans. First, the estrous cycle occurs over a 4- to 5-day period in contrast to the 28-day human menstrual cycle. Throughout the four stages of the mouse estrous cycle, substantial growth is followed by sloughing of the epithelium, whereas there is relatively little change in the human vaginal epithelium throughout the menstrual cycle (46, 47). The late proestrus and early estrus phases of the mouse estrous cycle are the most similar to that of the human vaginal epithelium (47, 48). In these stages, there is significant bacterial colonization, including a peak in the presence of lactobacilli (49). Additionally, the estradiol influence causes active secretion of mucus (49, 50), which we found in mice is both penetrable by MPPs and cleared in a matter of hours, similar to humans. Thus, we believe the IE mouse model is a valuable model in addition to the commonly used DP model for investigating vaginal delivery methods. Estradiol can be used to synchronize mice in the estrus phase, but does not “arrest” them in estrus; they continue to cycle, whereas DP treatment can arrest mice in a diestrus-like phase for days to weeks (51).

We have not yet investigated certain conditions that might affect MPP-mediated protection; for example, the potential effects of semen on MPP movement through mucus, or the impact of the MPP formulation on vaginal pH and bacterial flora, were not tested. Mice and nonhuman primates have essentially neutral vaginal pH because their vaginal flora is not lactobacilli-dominated. Therefore, it would be appropriate to investigate potential alterations in vaginal pH and bacterial flora in humans before any large-scale clinical trials. After MPPs were deemed suitable for repeated vaginal application in humans, partner studies could be done that would investigate the effect of coitus and the presence of semen on vaginal tissue drug concentrations. Nevertheless, we have shown that vaginally administered MPPs loaded with ACVp were more effective at protecting mice against vaginal HSV-2 infection than soluble drug, even at 10-fold higher soluble drug concentration. These results motivate further development of MPPs for safe and effective vaginal drug delivery, prevention and treatment of sexually transmitted infections, contraception, and treatment of other cervicovaginal disorders.

Materials and Methods

Mouse vaginal epithelium model

To study the distribution and retention of nanoparticles at the vaginal mucosal surface and the effects of repeated dosing, we used 6- to 8-week-old CF-1 mice (Harlan). Mice were housed in a reversed light cycle facility (12-hour light/12-hour dark). For naturally cycling estrus, mice were selected for external estrus appearance and confirmed upon dissection (52). For hormonally induced estrus (IE), mice were acclimated for 3 weeks and injected subcutaneously with 100 μg of 17β-estradiol benzoate (Sigma) 2 days before the experiments. It has been demonstrated in numerous studies that treatment with estradiol induces an “estrus-like” state with analogous epithelial characteristics and vaginal cell populations (22, 50). For vaginal toxicity and cytokine release, mice were injected subcutaneously with 2.5 mg of Depo-Provera (150 mg/ml) (Pharmacia & Upjohn Company) 7 days before the experiments.

Water was used as the hypotonic medium for all particle solutions. For ex vivo tracking, 5 μl of particles was administered intravaginally. After about 10 min, the vagina was removed and carefully sliced open to lay flat. The whole tissue was placed in a custom-made well constructed such that a coverslip could be placed on top to contact the mucus without deforming the tissue. The well was a rectangle about 1 mm by 0.5 mm cut out of three layers of electrical tape adhered to a standard glass slide. Coverslips were sealed around the edges with superglue and imaged immediately to prevent drying.

Mice were anesthetized before experimental procedures, including sacrifice by cervical dislocation. For all studies, mice were prevented from self-grooming by a collar of mildly adhesive tape around the abdomen, and from intergrooming by housing in individual cages. All experimental protocols were approved by the Johns Hopkins Animal Care and Use Committee.

Nanoparticle preparation and characterization

The preparation and characterization of CPs, BD-MPPs, and ACVp-MPPs are provided in the Supplementary Materials.

Distribution of nanoparticles in the mouse vagina

Five microliters of either CPs or MPPs was administered intravaginally. The entire vagina was then removed and frozen in Tissue-Tek OCT Compound (Sakura Finetek U.S.A. Inc.). Transverse sections were obtained at various points along the length of the tissue (between the introitus and the cervix) with a Microm HM 500 M Cryostat (Microm International). The thickness of the sections was set to 6 μm to achieve single-cell layer thickness. The sections were then stained with ProLong Gold (Invitrogen) antifade reagent with 4′,6-diamidino-2-phenylindole (DAPI) to visualize cell nuclei and retain particle fluorescence. Fluorescence images of the sections were obtained with an inverted fluorescence microscope. To quantify nanoparticle distribution, we administered 5 μl of either CPs or MPPs intravaginally. Within 10 min, vaginal tissues, including a “blank” tissue with no particles administered, were sliced open longitudinally and clamped between two glass slides sealed shut with superglue. This procedure completely flattens the tissue, exposing the folds. The blank tissue was used to assess background tissue fluorescence levels to ensure that all images taken were well above background levels. Six fluorescence images at low magnification and at least one image at high magnification were taken for each tissue. The images were thresholded to draw boundaries around the fluorescence signal, and then the area covered was quantified with ImageJ software. An average coverage was determined for each mouse, and then these values were averaged over a group of n ≥ 3 mice.

The cervix from each mouse was cut from the uterine horns and mounted with the same custom-made wells used for ex vivo particle tracking. The wells were sealed with a coverslip, and the background fluorescence levels were determined with the blank tissue. One fluorescence image, constituting nearly the entire ectocervical surface, was taken at low magnification above tissue background levels. These images were thresholded in the same manner to determine the area covered with particles. At least one higher-magnification image was taken for each tissue to show individual particles.

Distribution and retention of drug molecules in the mouse vagina

FITC dye (Sigma-Aldrich) was mixed at 1 mg/ml in HEC gel provided by T. Moench (ReProtect). BD-MPPs were prepared as described, loaded with FITC dye, and suspended in 1% Lutrol F127. To evaluate distribution, we administered 10 μl of either gel or particle solution intravaginally. After 24 hours, the vaginal tissue was removed and cut open to lie flat. The tissue was then mounted between two microscope slides and squeezed to flatten the rugae. A blank tissue was included to determine background autofluorescence from the vaginal tissue and to ensure that the exposure setting used was indicative of FITC presence. Fluorescent images of the dye distribution on the flattened tissue surface were obtained with a Nikon E600 inverted microscope equipped with a 2× objective. These images were thresholded in the same manner with ImageJ to determine the coverage area.

Retention of nanoparticles in the mouse vagina

To evaluate nanoparticle retention, we administered 5 μl of red fluorescent CPs or MPPs intravaginally. Whole cervicovaginal tracts were obtained at 0, 2, 4, and 6 hours and placed in a standard tissue culture dish. For each condition and time point, n > 7 mice were used. Fluorescence images of the tissues were obtained with the Xenogen IVIS Spectrum imaging device (Caliper Life Sciences). Quantification of fluorescence counts per unit area was calculated with the Xenogen Living Image 2.5 software.

HSV-2 challenge in the mouse vagina

Female 6- to 8-week-old CF-1 mice were subcutaneously injected with Depo-Provera and, 1 week later, received 20 μl of test agent or PBS intravaginally with a fire-polished positive displacement capillary pipette (Wiretrol, Drummond Scientific). Thirty minutes later, mice were challenged with 10 μl of inoculum containing HSV-2 strain G [American Type Culture Collection, 2.8 × 107 median tissue culture infectious dose (TCID50) per milliliter]. HSV-2 was diluted 10-fold with Bartel’s medium to deliver 10 ID50, a dose that typically infects ~85% of control mice. Mice were assessed for infection 3 days later after inoculation by culturing a PBS vaginal lavage on human foreskin fibroblasts (Diagnostic Hybrids), as described previously (34). In this model, input (challenge) virus is no longer detectable in lavage fluid if it is collected more than 12 hours after the challenge.

Vaginal cytokine release with repeated administration

Twenty microliters of each test agent was administered intravaginally to the DP mouse model once a day for 7 days. HEC gel and N9 were provided by T. Moench (ReProtect), and TFV vehicle gel was provided by C. Dezzutti (University of Pittsburgh). On the eighth day, each mouse was lavaged twice with 50 μl of PBS. Each lavage sample was diluted with an additional 200 μl of PBS and centrifuged to remove the mucus plug. Supernatant (200 μl) was removed and split into 50-μl aliquots for each of the four (IL-1β, IL-1α, TNF-α, and IL-6) Quantikine ELISA (enzyme-linked immunosorbent assay) kits (R&D Systems). ELISAs were conducted per the manufacturer’s instructions.

Statistical analysis

All data are presented as means with SEM indicated. Statistical significance was determined by a two-tailed Student’s t test (α = 0.05) assuming unequal variance. In the case of HSV-2 challenge, statistical significance was determined with Fisher’s exact test, two-tailed distribution.

Supplementary Materials

www.sciencetranslationalmedicine.org/cgi/content/full/4/138/138ra79/DC1

Materials and Methods

Video S1. Nanoparticle motions without flow.

Video S2. Nanoparticle motions with induced flow.

References and Notes

  1. Acknowledgments: We thank H. Patel (ELISA analysis), K. Maisel (particle tracking), the animal husbandry staff at Johns Hopkins, and the Wilmer Microscopy and Imaging Core Facility. Funding: This work was supported by the NIH (grants R33AI079740, R01CA140746, R21AI094519, and R21EB008515) (J.H. and R.C.), NSF (L.M.E. and Y.-Y.W.), and Howard Hughes Medical Institute (L.M.E.) graduate research fellowships. Author contributions: L.M.E and Y.-Y.W. conducted imaging experiments; L.M.E. and B.C.T. conducted animal experiments; L.M.E. and T.A.T. conducted particle formulation and characterization experiments; and T.H. conducted HSV-2 infection experiments. L.M.E., R.C., and J.H. directed all studies. L.M.E. wrote the manuscript with input and editing contributions from all authors. Competing interests: The MPP technology is being developed by Kala Pharmaceuticals, of which J.H. is a cofounder, consultant, and director. J.H. and R.C. own company stock, which is subject to certain restrictions under University policy. The terms of this arrangement are being managed by the Johns Hopkins University in accordance with its conflict of interest policies.
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article