Hypotonic, gel-forming delivery system for vaginal drug administration

Abstract

Vaginal drug delivery is often preferred over systemic delivery to reduce side effects and increase efficacy in treating diseases and conditions of the female reproductive tract (FRT). Current vaginal products have drawbacks, including spontaneous ejection of drug-eluting rings and unpleasant discharge from vaginal creams. Here, we describe the development and characterization of a hypotonic, gel-forming, Pluronic-based delivery system for vaginal drug administration. The rheological properties were characterized with and without common hydrogel polymers to demonstrate the versatility. Both qualitative and quantitative approaches were used to determine the Pluronic F127 concentration below the critical gel concentration (CGC) that was sufficient to achieve gelation when formulated to be hypotonic to the mouse vagina. The hypotonic, gel-forming formulation was found to form a thin, uniform gel layer along the vaginal epithelium in mice, in contrast to the rapidly forming conventional gelling formulation containing polymer above the CGC. When the hypotonic, gel-forming vehicle was formulated in combination with a progesterone nanosuspension (ProGel), equivalent efficacy was observed in the prevention of chemically-induced preterm birth (PTB) compared to commercial Crinone^® vaginal cream. Further, ProGel showed marked benefits in reducing unpleasant discharge, reducing product-related toxicity, and improving compatibility with vaginal bacteria in vitro. A hypotonic, gel-forming delivery system may be a viable option for therapeutic delivery to the FRT.

Graphical Abstract

1. Introduction

The development of drug delivery systems designed to treat diseases and conditions of the female reproductive tract (FRT) is an under-served area of research [1-3]. For many applications, such as hormone replacement therapy or treating infections, local vaginal drug administration is employed via dosage forms such as gels and creams [4, 5]. Such products can be messy, difficult to administer, and cause unpleasant discharge [4, 6]. The vaginal tissue surface is highly folded to accommodate expansion and collapsed due to intra-abdominal pressure, so thick gels and creams do not distribute well or uniformly coat the vaginal surface [7-10]. Further, the most common formulation approaches and excipients used are based off of skin products, rather than tailored specifically to the vaginal mucosa. Such formulation approaches render vaginal products hypertonic, resulting in fluid secretion from tissue, facilitating product leakage and clearance [11, 12]. Further, hypertonicity causes local epithelial toxicity, which has been demonstrated to increase susceptibility to infection [12-14]. There is a need to develop vaginal products that are tailored to the unique physiology of the FRT.

We previously described the application of mucoinert nanosuspensions to provide enhanced drug absorption locally in the FRT [15, 16]. When dosed in combination with a hypotonic delivery vehicle, the water absorption resulted in uniform distribution and retention over the mucosae without causing irritation or toxicity, leading to improved drug absorption and therapeutic effect [11, 12]. However, while liquid products are used for vaginal cleansing, it is not common for liquid formulations to be employed for drug administration. With the drawbacks of thick gels and creams, a viable option could be to take advantage of the distribution achieved by a hypotonic liquid and the retention time of a gel. Thermoreversible or thermosensitive polymers such as Pluronics have the characteristic of being a liquid that undergoes a gel transition when sufficiently concentrated above the critical gel concentration (CGC) and as temperature increases past the critical gelation temperature for that polymer concentration. We previously described a unique application of thermosensitive polymers as an in situ gelling system for topical application to the eye [17]. The combination of employing a hypotonic formulation with polymer below the CGC resulted in the formation of a thin, uniform gel layer that provided sustained drug absorption across the ocular mucosa. Formulations above the polymer CGC gel immediately upon warming to body temperature, leading to an uneven, clumpy, and easily cleared gel layer. Polymer concentration below the CGC ensured that the formulation remained a liquid in vitro, even at elevated temperature, and only became concentrated sufficiently to form a gel once there was water absorption by the epithelial cells to restore the osmotic disequilibrium in vivo [17]. Furthermore, the hypotonic formulation was safe and non-irritating with repeated dosing for at least several weeks in rabbits [17]. While such an approach may also have potential for improving vaginal drug absorption compared to conventional gels and creams, as well as thermosensitive formulations that gel in the lumen immediately upon application, the polymer concentration and osmolality of the formulation must be tailored specifically to the vaginal mucosa.

Here, we describe the further development of hypotonic, Pluronic F127-based thermosensitive hydrogels for vaginal drug administration. We first performed in vitro characterizations to determine compatibility with other common hydrogel polymers and small molecule acids and excipients used in vaginal products. We demonstrate that the thermally triggered Pluronic gel formation was not impaired by the other components, providing a versatile design space for formulation. We used qualitative (formulation leakage) and quantitative (multiple particle tracking) methods to identify the range of F127 concentration that was sufficient for gel formation only in vivo, and identified 10% (compared to the CGC of 15.5-16% (w/w)) as the composition to proceed with in efficacy testing. Using fluorescently-labeled F127, we observed that the hypotonic 10% formulation provided a more uniform gel coating on the vaginal tissue surface compared to the 18% (above CGC) formulation that gelled immediately in the lumen. To determine the impact of the improved distribution on vaginal drug delivery and retention, progesterone was selected as a model drug compared to the commercial progesterone cream product, Crinone^®. Crinone is approved for use in in progesterone supplementation in Assisted Reproductive Technology, and has been extensively tested clinically for prevention of preterm birth (PTB) [18, 19]. We loaded the hypotonic gel-forming vehicle with a progesterone nanosuspension [16] for comparative dosing to Crinone in a mouse model of progesterone withdrawal-induced PTB. Overall, the hypotonic gelling vehicle provided similar effectiveness in preventing PTB, but in contrast to Crinone, did not have a negative impact on the mucosal epithelia, did not cause unpleasant discharge, and did not negatively impact Lactobacillus crispatus survival in vitro. The hypotonic gel-forming delivery system described herein has the potential to provide improved treatment of a range of conditions affecting the FRT.

2. Materials and Methods

2.1. Materials

Pluronic F127 (P2443, source BCCH3308), DL-Lactic acid (~90%(T)) (69785), Sodium lactate (60% w/w) (L4263), Xanthan gum (G1253, 800-1200 cps 1% solution), Hydroxypropylmethyl cellulose (HPMC) (H9262, 80-120 cP, 2% in H₂O (20°C), progesterone (P8783), β-estradiol (E8875), Evans Blue (E2129), sodium dodecyl sulfate (SDS) (71729), phosphoric acid (438081), Rose Bengal (330000) was purchased from Sigma-Aldrich. HPLC grade acetonitrile (A998-4) and HPLC grade water (W5-4) were purchased from Fisher Chemical. Dulbecco’s Phosphate Buffered Saline (DPBS) (21-031-CV) was purchased from Corning Cellgro. Normal saline (114-055-101) was purchased from quality biological. Carboxymethylcellulose (CMC, 173 kDa) (CA193), hydroxyethyl cellulose (H1148, 3,400 cps), polyethylene glycol 400 (PEG400) (P0110), was purchased from Spectrum. Hyaluronic acid sodium (HA) (FH145201, MW 1.8-2.5 MDa) was purchased from Biosynth. Polyethylene glycol 6000 (PEG6000) (1546580) and Mifepristone (RU486) (M8046) was purchased from Millipore Sigma. Carbopol974P (CBP1053H, lot 106362, Carbomer Homopolymer Type B USP NF) was gifted from Lubrizol. Crinone 8% (Allergan) was purchased from the JHMI clinical pharmacy. 200 nm red carboxylate modified Fluospheres (F8786) and AlexaFluor 568 (A20003) was purchased from Invitrogen. 35mm glass bottom dishes (P35GC-0-14-C) were purchased from MatTek. For gel erosion studies, 30 mL syringes (53548-024) were purchased from VWR, 27Gx1-1/4 PrecisionGlide needles (306136) were purchased from BD, and NE-300 syringe pumps were acquired from New Era Pump Systems, Inc.

For bacteria experiments, GasPak jars (260629) and anaerobe satchets (160001) were purchased from BD. MRS agar (110660) and MRS broth (110661) were purchased from Millipore Sigma. For NYCIII media and agar, enzyme grade HEPES (BP310) was purchased from Fisher, Bacto Proteose Peptone No. 3 (211693) was purchased from Gibco, sodium chloride (J21618.A1), fetal bovine serum (26140079), and glucose (49139) was purchased from Thermo Fisher, Bacto Yeast Extract (212750) and Bacto Agar (0140-01) was purchased from BD. Gardnerella vaginalis (JCP8481B) and Lactobacillus crispatus (EX533959VC06) were obtained from BEI Resources.

2.2. Vehicle formulation

Formulations below the polymer CGC, particularly at the concentrations employed herein with Pluronic F127, will not undergo a gel transition when heated to body temperature in vitro. Thus, in order to probe the formulation design space and potential compatibility of Pluronic F127 gels with additional polymers in vitro, the F127 concentration must be above the CGC (~15.5-16% (w/w)). F127 forms micelles in solution, and above the CGC and corresponding critical gel temperature, the micelles pack together to form a gel structure [17, 20, 21]. As this is not a typical physical entanglement or crosslinking of high molecular weight polymers, it is possible that the addition of excipients could alter or entirely disrupt the gelation kinetics. To investigate this, 20% (w/w) F127, with or without various excipients (1.6% HPMC (w/w), 1.6% PEG400 (w/w), and 0.5% HEC (w/w)), was dissolved in water overnight at 4C. Sample osmolality was then measured using an EliTechGroup Vapro Osmometer in triplicate. The linear measurement range of the instrument is 100-1,000 mOsm/kg, so samples were mixed 1:1 with 0.9% normal saline prior to measurement. The formulation osmolality was then calculated from the measured value and the measured osmolality of the normal saline by linear extrapolation. Each measurement was repeated 3 times.

2.3. Rheological characterization

Gel tackiness is an important consideration for comfort and vaginal retention. Viscosity as a function of temperature and gel tackiness and adhesion force was measured using an Anton Paar cone and plate rheometer (model MCR302) with a PP25 probe and temperature control plates. For viscosity, the loading gap was set to 0.3 mm, and a temperature ramp from 10-40°C with 40 points (0.25°C minimum temperature change) were selected with gamma=1 (1/s) (n=3 per sample). For max tack measurements (n=3-4 per sample group), 200 μL of sample was transferred to the sample holder of the rheometer using a 1000 μL Drummond Wiretrol^® plunger pipette. The sample was equilibrated between the probe and sample holder with a 1 mm gap for 2 min at 37°C. The probe was placed in contact with the sample, then moved vertically perpendicular to the sample holder at a controlled rate to measure the normal force. A total of 100 points were measured over 60 seconds with a ramp linear model as the probe was pulled vertically upward at −0.5 mm/s. The peak force on the resulting curve was considered the max tack force. Adhesion was calculated by taking the area under the curve of the tack force. Data was exported from the Anton Parr RheoCompass software and analyzed in Microsoft Excel before being graphed in GraphPad Prism 10.

2.4. Erosion time

To evaluate the effects of excipients on resistance to erosion, an apparatus was constructed to allow for heating the formulations to 37C while flowing fluid at a controlled rate tangentially over the surface. Rose Bengal was added at 1 mg/mL for visualization and quantification of gel erosion. Holes were drilled into opposing sides of a 35mm glass bottom dish to allow for inserting syringes with needles attached, one for fluid entry, and one for fluid exit. Each syringe was connected to a syringe pump set to 100 mL/min flowrate. The glass bottom dish was placed on a 37C hot plate, and 50 μL of sample containing Rose Bengal was carefully pipetted in the center. Dulbecco’s phosphate buffered saline (PBS) was flowed over the samples, and the eluted fluid was collected every 4 min to measure the amount of Rose Bengal. For the comparison with an eluent more representative of the vaginal environment, 380 mg sodium lactate, 350 mg lactic acid, 37.5 mL of DI water, and 12.5 mL of saline were combined to make a vaginal fluid simulant (1% lactic acid, pH 3.7, ~260 mOsm/kg). Osmolarity was measured on a Vapro osmometer and pH was measured on a Fisher Scientific accumet XL200 pH meter. Samples were read in triplicate on a plate reader at 560 nm absorbance and analyzed against a dye standard curve to calculate the amount of gel eroded. 3 technical replicates were run per formulation.

2.5. Drug loaded formulation

P4 (160 mg/mL) was wet-milled in 6% (w/w) F127 for 10 hours to form a nanosuspension [15]. The nanosuspension was characterized using a Malvern ZetaSizer Nano Series (173° scattering angle). Formulations were diluted 1:100 in ultrapure water for particle size and polydispersity index measurements, and diluted 1:40 in 10 mM NaCl (pH 7) for ζ-potential measurements. The mean particle size was measured to be 314.3 ± 10.1 nm, ζ-potential of −1.12 ± 0.38 mV, and a polydispersity index of 0.16 ± 0.09. The P4 nanosuspension in 6% F127 was then combined 1:1 with 14% (w/w) F127. The final composition comprised 8% P4 in 10% F127 (ProGel). To create P4-loaded 18% or 20% F127, an additional 12% or 14% (w/w), respectively, of F127 was added to the P4 nanosuspension in 6% F127. The mixtures were then continually mixed in a rotating tube holder at 4°C for 72 hours to ensure the F127 was fully dissolved.

2.6. Acute accelerated stability

Viscosity over time after storage under accelerated degradation conditions was characterized using a Viscometer Brookfield DV2T equipped with a CPA-52Z cone spindle. Both 10% F127 and ProGel formulations were prepared and dispensed into sealed 15 mL Falcon tubes for storage. The samples were placed in an environmental chamber maintained at 40°C with 25% relative humidity, adhering to the FDA guidelines accelerated stability testing in semi-permeable containers (2.2.7.3). Sampling occurred on days 0, 7, and 14, with three samples retrieved from the environmental chamber each time. These samples were allowed to equilibrate to room temperature over a period of 6 h before viscosity measurements were conducted. To prepare for viscosity measurements, the Fisher A40 bath circulator was utilized to stabilize the viscometer cup and CPA-52Z cone spindle at 37°C for 60 min. Subsequently, 500 μL of each formulation was transferred to the viscometer cup for measurement. The Viscometer Brookfield DV2T was set to a shear rate of 13 rpm and a temperature of 37°C. Data collection followed a single-point averaging method over a 10-second measuring period, with an end condition set to 1 min. The reported data point for each sample represented the average viscosity measured between 50 and 60 seconds. This allowed for stabilization of the formulation viscosity, ensuring consistency and accuracy in the assessment under standardized conditions.

2.7. In vitro drug release study

Since ProGel contains F127 at 10%, which is too low to form a gel with heating in vitro, a group was included that contained 8% P4 nanosuspension in 20% F127. Ten μL of ProGel, Crinone, or 20% F127+8% P4 was pipetted into a 1.5 mL Eppendorf tube, each in triplicate. Tubes were placed in an Accublock digital dry bath (Labnet) at 37C for 5 min. Preheated PBS with 3% sodium dodecyl sulfate (SDS) (w/v) (1 mL) was carefully placed on top of each sample to ensure a boundary layer between the release media and the gel layer (if present) was maintained. At 15 min, 30 min, 45 min, 60 min, 90 min, 2 h, 3 h, 4 h, 5 h, 6 h, 7 h, 8 h, and 24 h, 0.9 mL of the 3% SDS solution was removed and replenished with fresh release media. P4 in the collected release media was quantified via HPLC for each timepoint. The HPLC mobile phase comprised 70:30 (v/v) acetonitrile:water containing 0.1% phosphoric acid with a flow rate of 1 mL/min with a run time of 10 min [22]. Detection was performed at an Ex/Em of 290/500 and λmax = 242 nm. Chromatographic separation was achieved using a Luna reversed phase C18 column (150 x 4.6 mm, 5 μm particle size, OOF-4252-EO).

2.8. Animal welfare statement

All experimental procedures were approved by the Johns Hopkins University Animal Care and Use Committee. Johns Hopkins program of Animal care and Use is accredited by AAALAC international. Animal care, and procedures follow the Guide for the Care and Use of Laboratory Animals 8th ed [23]. Unless otherwise specified, CD-1 mice were purchased from Charles River Laboratories at 6-8 weeks and were housed in a reverse light cycle room in cages of 5 if non-pregnant, or individually housed when pregnant. Pregnant animals were ordered timed pregnant to arrive at gestational day E13 and allowed to acclimate in reverse light cycling housing until dosing and experimentation beginning on day E15. Animals were anesthetized with isoflurane prior to euthanasia via cervical dislocation.

2.9. Assessment of formulation leakage

As the hypotonic gelling formulations do not form a gel in vitro, the gel-forming potential must be evaluated in vivo. Pluronic F127 solutions at 2%, 8%, 10%, and 18% (w/w) were prepared by mixing overnight at 4C, and 100 μg/mL of Evans blue was added for visualization. To synchronize the estrous cycle stage, naturally cycling mice were staged by visual observation of the external appearance of the introitus and surrounding tissue as previously described [24]. Mice in the estrus stage (n = 3) were dosed vaginally with 20 μL of F127 solution containing Evans blue. The mouse was allowed to rest for 1 minute inverted, then placed in a contained area lined with paper towels. The appearance of blue dye spots on the paper towels was counted as leakage, whereas a lack of dye spots within 3 minutes of ambulation was considered to be an absence of leakage and an indication of gel formation.

2.10. Multiple Particle Tracking (MPT)

To quantitatively characterize gel formation, visualization of fluorescent nanoparticles via multiple particle tracking (MPT) was employed. Polystyrene nanoparticles (200 nm red fluorescent) were chemically modified with 2 kDa polyethylene glycol (PEG) as previously described [25]. Particles were diluted 1:100 (0.02% solids) in 2%, 8%, 10%, and 18% (w/w) F127 solutions. Mice (n = 5-6) were staged in the estrus phase as described above [24, 26]. Mice were anesthetized with 3% isoflurane during vaginal dosing and subsequent inversion and sacrifice. Particle-containing solutions (20 μL) were dosed vaginally using a Wiretrol^®. Mice were held inverted for 1 minute and then sacrificed. The vagina was immediately dissected, sliced open vertically, and laid lumen-side-up on a slide with a custom-sized rectangular well cut from multiple layers of electrical tape. The top of the well was sealed with a coverslip and superglue, such that the coverslip contacted the mucosal surface without extensive compression of the tissue. The slide was inverted on a Zeiss Axiovert Observer-D1 and a minimum of 7 videos each at least 30 seconds long were taken per tissue. Videos were then exported and run through a proprietary MATLAB code that tracked the centroids of each particle and produced a mean square displacement (MSD) value for each particle [27]. The ensemble-averaged MSD (〈MSD〉) was calculated as the geometric mean of individual-particle MSDs at a time scale of 1 second.

2.11. Visualization of gel distribution in vivo

F127 was fluorescently labeled by chemical conjugation to AlexaFluor 568 as previously described [17]. The content of fluorescently labeled F127 was kept at 2% (w/w) to ensure that the labeled polymer did not interfere with the gel formation as previously described [17, 28]. For higher concentration F127 solutions, unlabeled polymer was added to reach a total concentration of 10% and 18% (w/w) F127. 20 μL of fluorescently labeled polymer solution was administered vaginally to estrus staged mice, and vaginal tissue was obtained 20 min after administration. Tissue was flash frozen in OCT, then sectioned on a Leica Cryostat at 10 μm thickness. To prevent smearing of the polymer on the slide, the vaginal tissue sections were not fixed or further stained. Imaging was performed immediately following sectioning using both a Texas Red filter and phase contrast Dark Low (DL) with bright field transmitted light settings on a Zeiss Axiovert. The brightfield and fluorescent images were overlaid using ImageJ.

2.12. Progesterone delivery to vaginal tissue

Pregnant mice were delivered from Charles River at gestational day 13 (E13) and were allowed to rest until E15. The morning of E15, mice received a vaginal dose (20 μL) of either Crinone cream, 10% F127 with 8% P4 (ProGel), 18% F127 with 8% P4, or saline. At each given timepoint, mice were sacrificed, and tissues were collected and flash frozen. Samples were sent to the JHMI Analytical Pharmacology Core for measurement of P4 concentrations using LC-MS/MS as previously described [16, 29]. Briefly, P4 was quantified in mouse vaginal tissue. Tissue samples were homogenized in 200 μL or 400 μL of blank charcoal stripped mouse plasma using Ultra Turrax T25 homogenizer (IKA) before extraction. P4 was extracted from 50 μL of plasma or tissue homogenates with 0.5 mL of acetonitrile/n-butyl chloride (1:4, v/v) containing 40 ng/mL of the internal standard, progesterone-d9 (Toronto Research Chemicals). After centrifugation, the top layer was then transferred to a clean glass tube and dried in a 40 °C water bath under a stream of nitrogen gas. The samples were reconstituted with 100 μL of water/acetonitrile (50:50, v/v) and then transferred into autosampler vials for LC-MS/MS analysis. Separation was achieved with an Agilent Zorbax XDB, C18 (4.6 × 50 mm, 5 μm) column at room temperature with water/acetonitrile/formic acid mobile phase (30:70:0.1, v/v) using isocratic flow at 1 mL/min for a total of 4 min. The column effluent was monitored using a Sciex triple quadrapole^™ 5500 mass-spectrometric detector (Sciex) using electrospray ionization operating in positive mode. The spectrometer was programmed to monitor the following MRM transition 315.3 → 109.1 for P4 and 324.3 → 100.0 for the internal standard, progesterone-d9. Calibration curves for P4 were computed using the area ratio peak of the analysis to the internal standard by using a quadratic equation with a 1/x2 weighting function over the range of 2 to 2000 ng/mL with dilutions up to 1:1000 (v:v). Tissue samples were then quantitated in ng/g as: nominal concentration (ng/mL) x dilution factor. Area under the curve from t = 0 to t = 6 h (AUC0–6) or the last measured concentration (AUClast) were calculated using sparse sampling noncompartmental analysis in Phoenix 64 WinNonlin^® software [16, 29]. Outliers were identified using Grubb’s test (1/5 data points was removed from the saline control group and 1/7 was removed from the P4-loaded 18% F127 group) and results were graphed using Prism GraphPad 10.

2.13. Efficacy in RU486 preterm birth model

Timed pregnant mice were delivered at E13 and housed in a reverse light cycle room to preserve their native circadian rhythm. Mifepristone (RU486), a progesterone receptor antagonist, was dosed to dams at a minimum dose that would achieve >50% PTB as previously described [16]. RU486 doses ranging from 3-50 μg in 100 μL DMSO were tested until a dose of 6.125 μg/100 μL was chosen (not shown). On E15, each mouse received a subcutaneous dose of 6.125 μg RU486/100 μL, followed by a vaginal dose of either 20 μL of Crinone, ProGel, 8% P4 in 18% F127, or saline via Wiretrol^® [16]. Animals were dosed daily E15-E18 and were monitored for PTB on or before E18. There were two technical replicates of this experiment, with half of the animals in each experiment iteration with an additional 3 RU486 control animals included from a separate experiment.

2.14. Histology with repeated dosing

Mice were housed together in the reverse light cycle room for 1 week for acclimation prior to dosing. Mice were dosed vaginally once per day for two weeks with 20 μL of either Crinone cream, 10% F127 vehicle, saline, or ProGel. Animals were housed with others who received the same treatment. One day after the last dose, mice were sacrificed and both vagina and cervix were collected and fixed in formalin for 24 hours. Tissues were then sent to the JHMI Reference Histology core for paraffin embedding, sectioning (6 μm), and staining with either hematoxylin and eosin (H&E) or mucicarmine. Sections were imaged and analyzed using a Nikon light microscope. Histology assessments were made in partnership with Johns Hopkins Phenotyping and Experimental Pathology Core.

2.15. Characterization of gel discharge

8-week-old CD-1 mice were visually estrus staged and vaginally dosed with 20 μL of Crinone, ProGel, or saline and allowed to ambulate for 30 minutes. Mice were then either lavaged with 25 uL of sterile saline (n=3 per treatment) and lavage fluid was placed on a slide, or the vagina was dissected, sliced opened, and flattened on a slide (n=3 per treatment). Lavage fluid and tissue were imaged at 10X on a Nikon Eclipse Ni-U light microscope to visualize particulates.

2.16. Incubation with vaginal microbiota strains

Laboratory strains of Lactobacillus crispatus and Gardnerella vaginalis were grown in triplicate and sub-cultured during their exponential growth phase. Bacteria in media was mixed 4:1 with drug solutions, including Crinone diluted with saline 1:5, saline, 10% F127, and ProGel. These mixtures were allowed to grow hypoxically with oxygen scavenging sachets for 3 hours, then were diluted and plated. Plates were incubated for 48 hours, then counted. CFUs were counted if Prism 10. Sample sizes were similar in all experiments using ANOVA/Tukey. The statistically significant threshold for these comparisons was p<0.05. Experiments were performed they were within the countable range of 30-300 CFUs. CFUs were graphed using Prism Graphpad 10.

2.17. Statistical Analysis

One-way analysis of variance (ANOVA) was used for comparing two groups in Graphpad Prism 10, and was followed by Tukey’s multiple comparison test when comparing three or more groups in Graphpad multiple times to ensure reproducibility. Log-rank test (Mantel-Cox) was used for comparing survival statistics during the PTB study in Graphpad Prism 10 testing the null hypothesis that all samples come from populations with the same survival and differences are due to chance and the statistically significant threshold for these comparisons was p<0.05.

3. Results

3.1. Common excipients used in vaginal products were compatible with Pluronic F127-based gels in vitro

Pluronic solutions formulated below the CGC do not form a gel in vitro, so for in vitro characterization of rheological properties, concentrations above the CGC (i.e. greater than 15.5-16% (w/w)) must be used. Thus, 20% (w/w) Pluronic F127 solutions were formulated with and without select “generally regarded as safe” (GRAS) excipients used in gel and cream products to assess the impact on gelation. Excipient concentration was chosen such that the resulting solutions remained liquid and did not gel at room temperature. Formulations tested included 20% F127, 20% F127 + 1.5% (w/w) HPMC, 20% F127 + 1.6% (w/w) PEG400, and 20% F127 + 0.5% (w/w) HEC. As shown in Figure 1A, excipient addition did not significantly increase the viscosity in the liquid phase at 17°C (range 64.3 – 375.0 mPa*s). At 37°C, all formulations formed a gel phase associated with a 1,000-2,000-fold increase in viscosity (range 182,560 – 224,700 mPa*s) (Figure 1B). When assessing the tackiness and adhesion of the formed gels at 37°C, only HPMC significantly increased the maximum tack force (Figure 1C) and adhesion forces (Figure 1D) of the formulation ~1.5 fold (p<0.0001). To assess how rheological properties would influence the gel erosion, a constant fluid flow system was employed in vitro. There were no significant differences in time to full erosion between formulations, with the average times observed to be in the range of 28 - 40 min (Figure 1E). Thus, it appeared that increased tackiness and adhesion with the HPMC formulation did not lead to a significant increase in erosion time. To test in an environment more similar to the vaginal environment, the 20% F127 formulation was tested in a 1% lactic acid eluent and had a similar erosion time to formulations tested in PBS eluent (Figure 1E). When characterizing the osmolality, it was observed that the addition of the polymer excipients increased the overall osmolality, though all formulations remained below 290 mOsm/kg (range 68 - 162 mOsm/kg) (Figure 1F). Overall, these studies confirmed that the added excipients did not interfere with the F127 gel formation, nor was there any obvious apparent improvement in the gel properties that would be expected to improve the in vivo performance. Thus, we moved forward with Pluronic F127 alone without additional additives to simplify the process of in vivo validation.

Figure 1. Physicochemical characterization.

The viscosity of F127-containing formulations at (A) 17C show the effect of adding polymer excipients on the liquid phase, and at (B) 37C show the effect of polymer excipients on the gel phase. (C) Maximum tack force and (D) adhesion force per unit area for each formulation was measured at 37C. (E) Time to full gel erosion in an in vitro constant flow system. (F) Osmolality of 20% F127 with polymer excipients. Dotted line represents isoosmolal, ~ 290 mOsm/kg. n=3-4 for each formulation with each measurement repeated in triplicate and averaged. *p < 0.05, one way ANOVA. Data shown as mean ± SD.

3.2. Hypotonic formulation below the CGC provides uniform gelation in vivo

As the gelation phenomena of formulations below the CGC cannot be assessed in vitro, we went on to in vivo characterizations to identify Pluronic F127 compositions that would form a gel when dosed hypotonically. We first performed a qualitative experiment to visualize leakage, which would be minimized if there was a liquid to gel transition after vaginal dosing. Using dye-loaded solutions, we consistently observed leakage after dosing 2% (w/w) (n=3/3) and 8% F127 (n=3/3), while no leakage was observed after dosing 10% (n=0/3) and 18% F127 (n=0/3) (Figure 2A).

Figure 2. Determination of polymer concentration to achieve gelation in vivo.

(A) Representative images of spots of dye-laden fluid that leaked out after vaginal administration of hypotonic solutions containing 2, 8, 10, or 18% F127. 3/3 mice experienced leakage in both the 2% F127 and 8% F127 groups, while 0/3 mice experienced leakage in both the 10% F127 and 18% F127 groups during 3 minutes of ambulation. (B) Ensemble averaged mean square displacement (<MSD>) at a time scale of 1 second for 200 nm particles administered in various concentrations of F127. Vaginal tissue was excised and flattened to view particle mobility on the tissue surface. Each data point represents the ensemble average of 50–100 nanoparticles tracked in an individual sample. n=5-6 mice per formulation. *p<0.05 compared to 2% F127 group, one way ANOVA. Data shown as mean ± SEM.

To quantitatively characterize the gelation phenomenon, we then vaginally dosed F127 solutions containing polyethylene glycol (PEG) coated nanoparticles. As previously described, nanoparticles become physically entrapped within the gel, which is reflected in low mobility observed via multiple particle tracking (MPT) [17, 30]. Thus, when excising the vaginal tissue and directly observing the nanoparticle mobility on the vaginal tissue surface, high particle mobility is indicative of a viscous liquid environment, whereas low mobility (trapped) nanoparticles reflect the presence of a gel. We found that while the nanoparticles dosed vaginally in the 2% and 8% F127 showed high mean squared displacement (MSD) indicative of thermal motion in a liquid, nanoparticles dosed vaginally in the 10% F127 showed a ~308-fold reduction in MSD compared to dosing in 2% F127 (p = 0.028), similar to the ~195-fold reduction in MSD for nanoparticles dosed in 18% F127 (p = 0.029) (Figure 2B). This indicates that despite being below the CGC, the hypotonic 10% F127 formulation formed a gel in the vagina in vivo.

We then sought to assess the vaginal distribution of the hypotonic 10% F127 gel compared to the standard 18% F127 formulated above the CGC. We utilized fluorescently labelled F127 to visualize the gel inside the vaginal lumen and the uniformity of contact with the highly folded vaginal epithelium 20 min after administration. As shown in Figure 3A, the 10% F127 formed a distinct uniform coating of the vaginal epithelium, whereas the 18% F127 was excluded from vaginal folds (denoted by arrows in Figure 3B). Similarly, the 10% F127 appeared to accumulate on the tissue surface due to the hypotonically-driven absorption and concentration of the polymer (Figure 3A), whereas the 18% F127 immediately formed a gel in the lumen (denoted by asterisks in Figure 3B).

Figure 3. Comparison of gel distribution after vaginal administration in mice.

Fluorescently-labeled F127 formulated to be (A) hypotonic, below the CGC (10%) administered vaginally results in a defined gel layer up against the epithelium, compared to (B) F127 formulated above the CGC (18%) which only partially coats the epithelium with large amounts of gel in the lumen. Three representative tissue sections from different mice (n = 4 per group) collected 20 min after vaginal dosing are shown for each formulation. Arrows indicate areas where gel did (A, black arrows) or did not (B, white arrows) reach the folded surfaces of the rugae, and asterisks in (B) indicate areas where gel remained in the lumen after dosing 18% F127. Scale bar applies to all images.

3.3. The hypotonic gel-forming formulation provides more consistent vaginal drug delivery

To demonstrate comparative efficacy in vaginal drug administration, we chose P4 as drug that is dosed vaginally for preterm birth (PTB) prevention [16]. P4 at 8% was nanomilled [15] with a size of 312.8 ± 10.4 nm, PDI of 0.18 ± 0.11, and zeta potential of −0.43 ± 0.10 mV, and suspended in 10% Pluronic F127 (ProGel) to dose match to Crinone 8% P4 vaginal cream. We then compared the release of P4 from the ProGel compared to Crinone, though since ProGel does not gel at 37°C in vitro, we included a group containing the 8% P4 nanosuspension in 20% F127 to characterize the effect of gelation on drug release. It was evident that when the F127 was in the gel state in vitro, the release of P4 from the nanosuspension was slowed compared to micronized P4 in Crinone (60% release after 24 h vs. 100% release by 7 h) (Supplementary Figure 1). Additionally, when stored under accelerated degradation conditions, the viscosity of the 10% F127 vehicle and ProGel was consistent for at least 2 weeks, supporting the potential for long-term stability (Supplementary Figure 2).

We previously reported limited ability to remove excess Crinone gel from the vaginal tissue surface to measure progesterone concentration in the vaginal tissue [16], so we characterized tissue concentrations at 24 h after dosing when excess material should no longer be retained prior to the next daily dose. Notably, mice dosed with Crinone had vaginal P4 concentrations ranging over four orders of magnitude from 93.3 ng/g to 1.64 x 10⁶ ng/g, with nearly equal groupings around lower and higher values, suggesting that some mice had excess cream remaining in the vaginal lumen (Figure 4A). In contrast, the vaginal tissue concentrations were more consistent after ProGel dosing (median 137.3 ng/g), and approximately 19-fold higher than the median endogenous P4 concentration in saline dosed mice and 3.6-fold higher than the median P4 concentration in mice dosed with progesterone in 18% F127 (Figure 4A). We next tested the efficacy of vaginal P4 product administration in a mouse model of systemic progesterone withdrawal. While 66% (n=6/9) of mice in the control group gave birth preterm, only 15% (n=2/14) of mice dosed vaginally with either Crinone or ProGel daily delivered preterm, compared to 25% (n = 3/12, p = 0.006) of mice dosed vaginally with 8% P4-loaded 18% F127 (p = 0.028) (Figure 4B).

Figure 4. Vaginal progesterone (P4) delivery for prevention of preterm birth (PTB).

(A) Pregnant mice at E15 were vaginally dosed either Crinone, ProGel, or 8% P4-loaded 18% F127 (n=4-6 mice per formulation) containing 8% (w/w) P4. Vaginal tissues were collected 24 h later to measure P4 concentrations. *p<0.05, one way ANOVA. Data shown as median ± IQR. (B) Mifepristone (RU486) was dosed on E15 (black arrow), followed by daily vaginal administration with ProGel (n = 14), Crinone (n = 14), 8% P4-loaded 18% F127 (n=12), or saline (RU486, n = 9) from E15-E18 (gray arrows). Full-term delivery was considered to occur on E19-20. Statistical significance was determined by log-rank test (Mantel-Cox), *p<0.05.

3.4. The hypotonic gel-forming formulation is more compatible with the healthy vaginal environment

While ProGel was non-inferior to Crinone in preventing PTB in the mouse model, Crinone was previously shown to have adverse effects related to toxicity [16]. Indeed, we observed that repeated vaginal dosing with Crinone resulted in increased polymorphonuclear cells (PMNs) and mononuclear cells in the vaginal wall and sometimes serosa (Figure 5). However, histology findings were milder and similar among ProGel and vehicle or saline groups (Figure 5). Further, we observed that dosing with Crinone gel resulted in a reduced cervical mucus layer and mucus producing cells, whereas mice dosed with ProGel had a cervical mucus layer that was indistinguishable from mice treated with vehicle or saline (Figure 5).

Figure 5. Histological analysis of vaginal and cervical tissue after repeat dosing.

Vagina (H&E, original magnification 20x) and cervix (mucicarmine, original magnification 10x) after daily vaginal dosing for 14 days (n=4 each). Mice were sacrificed and tissues were collected 24 hours after the last vaginal dose. Black arrows indicate PMNs, and asterisks indicate regions with reduced mucus and mucus producing cells (lack of pink staining) in mucicarmine stained sections after Crinone dosing.

In addition to being hypertonic, Crinone is a thick, opaque white cream that is known to cause unpleasant discharge in patients [31]. Similarly, visible clumps of cream were evident on the vaginal tissue surface of mice dosed with Crinone, whereas mice dosed with ProGel did not have any apparent residue or visually observable material in the mouse vagina (Figure 6). Similarly, a milky white haze and clumps of cream were observed in lavage fluid from mice dosed with Crinone, which also showed large particulates in solution under higher magnification (Figure 6). In contrast the lavage fluid from mice dosed with ProGel looked more similar to saline dosed mice at low magnification, showing smaller particulates at high magnification (Figure 6).

Figure 6. ProGel may reduce the incidence of unpleasant product discharge.

After dosing Crinone vaginally in mice, clumps of opaque gel could be observed on both dissected vaginal tissue (top row) and in vaginal lavage fluid (middle row), as indicated by red circles. In contrast, vaginal dosing with ProGel resulted in no visualizable residue on on excised vaginal tissue, and lavage fluid appeared similar to that of mice receiving saline treatment. Further, microscopy of lavage fluid showed large particulates after Crinone dosing, with smaller nanoparticulates in the lavage fluid after ProGel dosing (bottom row). Lavage and dissections were performed 30 min after vaginal dosing, n=3 for each group.

The vaginal microbiome is crucial for overall female reproductive health, and disruption and dysbiosis in the vaginal microbiota is associated with an increased risk for PTB, pelvic inflammatory disease, etc. [32, 33]. Thus, it is important that vaginal products do not negatively affect the endogenous Lactobacillus bacteria associated with reproductive tract health. L. crispatus is often found as the dominant bacteria in a healthy bacterial community, so we assessed the impact of our formulations on L. crispatus survival in vitro. We observed that exposure to Crinone resulted in a ~2 log reduction in the number of colony forming units (CFUs) after 3 hours of exposure, whereas ProGel had no effect on L. crispatus survival (Figure 7).

Figure 7. Formulation impact on L. crispatus survival in vitro.

Exposure to Crinone significantly reduced L. crispatus viability as measured by colony forming units (CFU), while ProGel or the 10% F127 vehicle did not impact L. crispatus survival compared to saline. Bacteria stock was grown with three biological replicates and plated in triplicate per treatment. *p < 0.05, one way ANOVA. Data shown as mean ± SD.

5. Discussion

While the majority of pharmaceutical products are oral and intravenously administered, many diseases and conditions of the FRT lend themselves to the opportunity for local delivery. Direct administration to the vagina can allow for the use of lower drug doses with fewer systemic side effects. Vaginal drug delivery can also be beneficial for the upper tract, as the “uterine first pass” effect provides access to the local vasculature for preferential distribution throughout the FRT before reaching the rest of the body [34, 35]. Indeed, some drugs, such as misoprostol, are known to be more efficacious when administered vaginally [36]. However, many vaginal products have significant drawbacks, and any related side effects, difficulty administering, high frequency of dosing, and discharge often prevent patients from complying with the full and consistent treatment regimen [37]. Patient adherence is a concern, as medications are only effective when used as prescribed.

There are a variety of vaginal drug delivery platforms both on the market and in development, including ovules, tablets, creams, foams, films, rings, washes, suppositories, and implantable devices [38, 39]. Most creams and gels contain stabilizing and solubilizing excipients, including glycerin, oil emulsions, and other lipids [40]. These formulations are often hypertonic, which can lead to discharge and potential toxicity [11]. Indeed, our prior work demonstrated that Crinone P4 cream causes cytotoxicity in the cervix and inflammatory gene expression in the myometrium [16]. In contrast, hypotonic delivery of mucoinert nanosuspensions provided enhanced vaginal drug delivery while preserving tissue integrity, and increasing epithelial coverage and drug delivery to the upper tract [11, 15, 16]. Here, we combined the hypotonic gelling formulation with a P4 nanosuspension to provide delivery of a high dose of P4 (ProGel) in the context of PTB prevention, though the vehicle is also compatible with water soluble drugs and can potentially solubilize hydrophobic drugs at lower concentrations [17, 41-43]. The ProGel formulation showed a marked reduction in the inflammatory response in the FRT compared to Crinone, which is consistent with other studies showing that hypotonic products typically do not have the same toxicity concerns as hypertonic products [11, 12, 15, 16].

Pluronic F127, or Poloxamer 407, a thermosensitive GRAS polymer, has been used for drug stabilization and delivery through a variety of mucosal membranes, including the eye, colon, and vagina [15-17, 44]. F127 is a tri-block co-polymer comprised of a polypropylene oxide (PPO) center chain flanked by two polyethylene glycol (PEG) chains [45]. Our lab has also demonstrated that F127 can enhance nanoparticle penetration of CVM without impacting mucus barrier properties to viral infections [46]. Additionally, we have used a much lower concentration of F127 as a nanosuspension stabilizer for P4 delivery in the vagina, though the incorporation of the gel-forming aspect described here may represent a more pharmaceutically appropriate dosage form [16]. The conventional approach for utilizing the thermosensitive gel-forming properties of F127, with or without other polymer blends, for vaginal drug administration utilize the polymer at or above its CGC (15-16%) [17, 47, 48]. Indeed, Pluronic F127 (also known as Poloxamer 407) is used at concentrations above the CGC in an FDA-approved vaginal gel formulation containing 2% clindamycin phosphate (Xaciato^™) [49, 50]. Here, we formulated F127 at 10% w/w, well below the CGC, to take advantage of the distribution and formation of a thin coating provided by fluid absorption by the mucosal epithelia. Thus, only once the polymer has been spread uniformly and concentrated does it reach or surpass the CGC to form a thin, uniform gel layer coating the epithelial surface. In contrast, when F127 is dosed to the vagina above the CGC, it immediately gels in the lumen as the gelation temperature is surpassed. This leaves a bulk of gel material in the lumen, which reduces drug absorption and facilitates clearance.

While our hypotonic gel-forming P4 formulation, ProGel, was equally as effective as Crinone in the PTB model, we demonstrated benefits in the vaginal distribution and uniformity, reduction in the presence of unpleasant discharge, reduced local toxicity, and minimized impact on Lactobacillus survival in vitro. It is possible that superiority in effectiveness could be demonstrated at different doses or with drugs that are dosed at lower concentrations than vaginal P4. It is worth noting that the studies performed here were in mice, and there are many similarities and differences in the FRT in mice compared to humans [51-53]. Further, it would be necessary to determine whether age or disease state would impact the behavior and drug delivery efficacy of the vehicle in the FRT, and whether formulations would need specific optimization for use in a particular disease or condition.

6. Conclusion

Herein, we described the development and characterization of a hypotonic, gel-forming delivery system for vaginal drug administration. Thermosensitive Pluronic F127 polymer was formulated to be hypotonic to the vaginal epithelium and below the polymer CGC to take advantage of the distribution achieved by a hypotonic liquid and the retention time of a gel. Rheological characterization confirmed that Pluronic F127 could be formulated with additional polymer excipients without interfering with the gel-forming properties in vitro, including viscosity at different temperatures and erosion time. To demonstrate the proof of principle of a thermosensitive gel formulated below the CGC, such that it only undergoes a gel transition after sufficient water absorption by the cervicovaginal epithelial surface in vivo, Pluronic F127 at different concentrations was dosed and characterized for leakage and gel-formation. It was found that 10% (w/w), well below the CGC of 15.5-16%, formed a gel when formulated to be hypotonic to the mouse vagina. Fluorescently labeled polymer was used to confirm that there was uniform distribution of the hypotonic 10% F127 compared to the conventional 18% F127. When 10% F127 was loaded with a P4 nanosuspension (ProGel), similar efficacy to Crinone 8% vaginal cream was observed in a mouse model of progesterone withdrawal-induced PTB. Additionally, dosing ProGel was associated with reduced discharge, reduced epithelial toxicity, and increased compatibility with vaginal microbiota in vitro. Such a versatile gel-forming platform technology has potential to be used with a variety of drugs for different indications affecting the FRT, and may increase patient compliance and satisfaction.

Data Availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.