WO2010061284A2 - A controlled release intravaginal polymeric pharmaceutical dosage form - Google Patents

Abstract

This invention relates to a controlled release, intravaginal, pharmaceutical dosage form. The dosage form consists of at least one desired pharmaceutically active ingredient which is admixed with a combination of biocompatible and biodegradable polymers and shaped for insertion into the vagina of a patient. The polymers are selected so as to biodegrade at a predetermined rate and, on biodegrading, the pharmaceutically active ingredient is released.

Description

A CONTROLLED RELEASE INTRAVAGINAL POLYMERIC PHARMACEUTICAL DOSAGE FORM

FIELD OF THE INVENTION

This invention relates to a controlled release intravaginal polymeric pharmaceutical dosage form and, more particularly, to a pharmaceutical dosage form suitable for the delivery of a pharmaceutical composition(s) in a rate-modulated site-specific manner via the intravaginal route or as an implantable embodiment in a human or animal body. The invention also relates to a method of manufacturing the said polymeric pharmaceutical dosage form.

BACKGROUND TO THE INVENTION

Heterosexual transmission of the Human Immunodeficiency Virus and Sexually Transmitted Infections (HIV/STIs) account for 90% of infections and transmission from male-to-female occurs more than female-to-male due to the larger mucosal exposure to seminal fluid in the vagina. Once infected, most women are asymptomatic and may remain untreated, thus increasing infection rates.

Strategies have been proposed or adopted to prevent the spread of HIV/STIs but these possess limitations. Currently, delivery of microbicides to the vagina has been limited almost exclusively to semi-solid aqueous gels and vaginal rings (Mitchel, 2006) which have been found to lack numerous essential requirements of an ideal intravaginal drug delivery system. It has been suggested that a vaginal drug delivery system that combines several mechanisms of prevention would be more effective or have fewer side-effects than a single mechanism of prevention as is the case of preparations currently being researched, partly due to the minimal attention to formulation, design and delivery technology development ((D'Cruz and Uckun, 2004; Panttinen, 2005; Shattock, 2006; Ndesendo et al., 2008).

COMFIRWiOW G^QPϊ It should be noted that the term "caplet", when used in this specification, means a smooth, coated, oval-shaped medicine tablet which is intended to be tamper resistant.

OBJECT OF THE INVENTION

It is an object of this invention to provide a controlled release intravaginal polymeric pharmaceutical dosage form and a method of manufacturing the said polymeric pharmaceutical dosage form.

SUMMARY OF THE INVENTION

In accordance with this invention there is provided an intravaginal polymeric pharmaceutical dosage form for releasing, in the vagina in a controlled and rate modulated manner, at least one active pharmaceutical composition, the dosage form comprising a suitably shaped combination of biocompatible and biodegradable polymers admixed with the active pharmaceutical composition or compositions such that, in use and over a predetermined time period, the or each active pharmaceutical composition is released from the dosage form as the polymers degrade.

There is further provided for the dosage form to be in the form of a caplet, alternatively a tablet.

There is also provided for the biocompatible and biodegradable polymers to be selected so as to biodegrade at a predetermined rate and, thus, to release the active pharmaceutical composition or compositions at a predetermined rate and achieve a selected release profile.

There is further provided for the polymers to have bioadhesive properties and adhere, in use, to the vaginal wall, preferably in the region of the posterior fornix, alternatively, the surface of the cervix, until the polymers have degraded and the active pharmaceutical composition or compositions has or have been released. Alternatively there is provided for the polymers to have gelling properties and further alternatively for the polymers to have bioadhesive and gelling properties.

There is further provided for the polymers or at least one polymer of a blend of polymers to also function to inhibit disease causing microorganism infection and preferably to inhibit infection by microorganisms causing sexually transmitted diseases, preferably the Human lmmunodefficiency Virus (HIV).

There is further provided for the polymer or polymers to be hydrophilic, alternatively hydrophobic, further alternatively a blend of hydrophilic and hydrophobic polymers, for the polymer or polymers to be selected from amongst the group comprising poly(acrylic acids) (PAA), poly(lactic acids) (PLA), carageenans, polystyrene sulfonate, polyamides, polyethylene oxides, cellulose derivatives, poly(vinylpyrrolidone) (PVP), polyvinyl alcohol) (PVA), Chitosan, poly(ethyl acrylate, methyl-methacrylate, and chlorotrimethyl- ammoniumethylmethacrilate (PMMA), hydroxyapatite, gum-based polymers such as xanthan gum and their variants or various permutations and derivatives of the said polymer-types.

There is further provided for at least one polymer to be a poly(acrylic acid) (PAA) polymer which, in use, the said polymer/s to function as a bioadhesive substance..

There is further provided for the polymer or polymers to be crosslinked and for the crosslinking reagents to be selected from a class of biocompatible inorganic or organic salts, used in the crosslinking reactions of the polymer or polymer and pharmaceutical composition(s), and are ionic of either mono-, di-, or trivalent nature, examples of which are sodium chloride, aluminium chloride or calcium chloride.

There is also provided for the dosage form to achieve, in use, a rate-modulatable release of an active pharmaceutical composition admixed with the polymer or polymers, said rate modulation being, at least partly, achieved by the the architectural structure of the polymeric network which is, in turn, a function of the different permutations of the hydrophilic and hydrophobic polymers, pharmaceutical composition or compositions, and inorganic and/or organic salts which are selected to achieve a desired rate modulated release of the or each active pharmaceutical composition.

There is further provided for a pre-determined rate-modulated release profile to be controlled, in use, by the rate of polymeric hydration within the system which depends on the pKa, concentration and valence of the release rate-modulating chemical substances used.

There is also provided for the pharmaceutical dosage form to be capable of inducing an acidic pH environment in the vagina upon degradation and thus maintain a wide spectrum of activity against pathogens.

There is also provided for the polymers to be multifunctional polymers, preferably barium sulphate, alternatively hydroxyapatite.

The invention extends to a method of manufacturing a pharmaceutical dosage form as described above.

DESCRIPTION OF PREFERRED EMBODIMENTS

The above and additional features of the invention will become apparent from the embodiments details of which are presented below. The embodiments result from a study aimed at designing and developing a novel controlled release Vaginal Polymeric Device (VPD) possessing certain perceived desirable criteria. The study makes reference to the following Figures:

Figure 1 : Is a schematic of the upper vagina region depicting the posterior fornix as the site of the VPD application. Figure 2: Is a schematic depicting the various mechanisms of preventing the transmission of Sexually Transmitted Infections and HIV by employing microbicide delivery systems. (Adapted: Stone, 2002).

Figure 3: a) 3'-azido-3'-deoxythymidine (AZT), molecular weight 267.24 and a solubility 20.1 mg/mL, b) Polystyrene Sulfonate (PSS).

Figure 4: Illustrates the synthesis of modified polyamide 6,10 by interfacial polymerization.

Figure 5: Illustrates the dissection process to remove the vaginal tissue from the pig

Figure 6: Shows a textural analysis method employed to generate Force-Distance profiles for assessing the caplet bioadhesivity to freshly excised pig vaginal tissue.

Figure 7: Illustrates a constructed multilayer perceptron network.

Figure 8: Shows the network topology depicting the hidden input and output layers.

Figure 9: Digital images depicting a) insertion of the VPD into the vagina of the pig and b) tracking the location of the VPD in the vagina using a speculum.

Figure 10: a) Calibration curves for: a) AZT in simulated plasma fluid 267nm (N=3, SD<0.001 in all cases); b) PSS in simulated plasma fluid 244nm (N=3, SD<0.003 in all cases).

Figure 11 : Permeation studies of AZT and PSS across pig vaginal tissue using a Franz Diffusion Cell apparatus. Figure 12: Summary of the in vivo study.

Figure 13: a) X-Ray imaging process and b) blood sampling procedure from the jugular vein of the pig.

Figure 14: Calibration curves for: a) AZT in blank plasma 267nm (N=5, SD<0.001 in all cases); b) PSS in blank plasma 244nm (N=5, SD<0.009 in all cases).

Figure 15: The vaginal polymeric devices (VPDs) a) Planar view, b) Side view, c) Oblique view, d) Longitudinal view.

Figure 16: A comparison of work of adhesion (AUCF_D) (J) between the AS-PAA and APE-PAA caplets (N=3; SD<0.03).

Figure 17: A typical response optimization plot for the AS-PAA and APE-PAA caplets.

Figure 18: Typical textural profiles elucidating the peak adhesive force (PAF) (N) and work of adhesion (AUCFD) (J) for the optimized composed of: a) AS-PAA, and b) APE-PAA.

Figure 19: Typical textural profile elucidating the force (N) and work of adhesion (AUCFD) (J) for the optimized APE-PAA caplets on freshly excised vaginal tissue.

Figure 20: A chemometric structural model developed in our laboratories depicting caplet bioadhesion to freshly excised pig vaginal tissue with muco- epithelial cell secretions and surface bio-molecule interactions. Figure 21 : Rheological behavior of 2%^w/_v AS-PAA and APE-PAA solutions at a shear rate between 0-50Os^"1.

Figure 22: A typical bar chart graph showing the sensitivity coefficients of each polymer type against the matrix erosion following the secondary training.

Figure 23: Micro-environmental pH variation in the simulated vaginal fluid containing the VPD (N=3).

Figure 24: TMDSC thermograms for a) the hydrated physical polymer blend, b) the unhydrated physical polymer blend, c) the hydrated VPD and d) the unhyd rated VPD.

Figure 25: A typical 3D UPLC profile showing a complete separation between AZT and MP (internal standard

Figure 26: UPLC chromatograms depicting the separation of a) AZT and MP (internal standard) and b) PSS and MP (internal standard) in simulated vaginal fluid (pH 4.5; 37°C).

Figure 27: A typical profile showing the effect of coating on the model hydrophilic drug AZT from a) uncoated VPD and b) shellac/APE-PAA-coated VPD in simulated vaginal fluid (pH 4.5; 37°C) (N=3; SD<0.18 in all cases).

Figure 28: Drug release profiles of a) AZT (AZT-loaded VPD), b) AZT (AZT/PSS- loaded VPD), c) PSS (AZT/PSS-loaded VPD) and d) PSS (PSS-loaded VPD), in simulated vagina fluid (pH 4.5; 37⁰C) (N=3; SD<0.38 in all cases).

Figure 29: Profiles showing the flux of AZT and PSS across pig vaginal tissue over a period of 24 hours (N=3, SD<0.23 in all cases). Figure 30: Typical Force-Distance textural profiles used for computing the Peak Adhesion Force (PAF) and Work of Adhesion (AUCFD) for a) uncoated devices and b) PAA-coated devices on freshly excised pig vaginal tissue.

Figure 31 : X-ray images depicting the presence of the intravaginal bioadhesive polymeric device at a) day 1 , b) day 14, and c) day 30 after insertion into the posterior fornix of the pig vagina.

Figure 32: Molecular model mechanistically depicting the VPD dissolution process with lesser H-bond formation due to the excessive of simulated vaginal fluid providing more freedom to polymeric strands to disentangle.

Figure 33: Chemometric model depicting the development of a diffusion channel with a) a single polymer strand P, situated perpendicularly to a forming pore C, b) a group of strands also denoted collectively as P giving rise to the channel C which is formed perpendicular to the polymer strands backbone with F=the direction of flow from the

Figure 34: UPLC chromatograms depicting the retention times for the standard solutions of a) AZT, b) PSS and MP as an internal standard in blank plasma.

Figure 35: a) AZT and PSS plasma concentration profiles (N=5, SDO.01), b) AZT and PSS mean vaginal tissue and plasma concentration values on day 28 (N=5, SD<0.014 in all cases).

Figure 36: Correlation between a) AZT, b) PSS concentrations in the blood plasma and vaginal tissue. Figure 37: Histological images of haematoxylin and eosin stained pig vaginal tissue samples depicting a) Epithelial hyperplasia of the vagina x 40, b) Epithelial hyperplasia, exocytosis and superficial exudate x 40, c) Lamina propria with no inflammatory cells and normal vagial epithelium x 20, d) Hyperplastic epithelium with exocytosis x 40, e) Mononuclear cell infiltrates in the lamina propria and f) Perivascular inflammation in the submucosal wall x 40.

This study was aimed at designing and developing a novel controlled release Vaginal Polymeric Device (VPD) which possessed the following desirable criteria: i) simple to manufacture, cost effective and easy to apply thus facilitating patient compliance; ii) non-irritative and not causing any physical discomfort; iii) provide immediate protection by releasing bioactives in a controlled manner over a prolonged period of time; iv) have suitable vaginal retention and distribution; v) be versatile against various sexually transmitted pathogens; and vi) simultaneously release an ARV and microbicide. Furthermore this study aimed at developing and evaluating a microbicide and ARV- loaded bioadhesive VPD that may be used in preventing the transmission of HIV/STIs by intravaginal release of bioactives in a controlled manner. The VPD would be easily inserted into the posterior fornix of the vagina (Figure 1) as opposed to gels or ring systems currently being used or tested to deliver microbicides. The VPD would be vaginal retentive and control bioactive release for longer periods of time and will utilize more than one preventative mechanism when inserted into the posterior fornix of the vagina. Bioadhesive polymer-based devices were used in the development of the VPD for intravaginal drug delivery. The devices comprised formulations containing various grades of poly(acylic acid) (PAA) such as allyl sucrose-crosslinked PAA (AS-PAA Matrix) and allyl penta erythritol-crosslinked PAA (APE-PAA Matrix) which were compressed and used as gold standards for assessing the bioadhesivity of the selected polymer blends in conjunction with other polymers and the bioactives. The model microbicidal agent employed in the formulation was polystyrene sulfonate (PSS) together with the anti-retroviral drug (ARV), 3'-azido-3'-deoxythymidine (AZT). The aim of designing a controlled release drug delivery system is to reduce the frequency of dosing, to increase the effectiveness of the drug by localization at the site of infection, to reduce the drug-load required and to provide uniform drug delivery. Controlled drug delivery occurs when a natural or synthetic polymer, is combined with a drug or other bioactive agent in such a manner that the bioactive agent is released from the polymeric material in a pre-determined manner. Various FDA approved biocompatible and biodegradable polymers were employed to deliver a microbicide and an ARV with the purpose of attaining bioadhesivity and controlled release from the posterior fornix of the vagina. The polymers used in this study include modified polyamide 6,10, (_mPA 6,10), poly(acrylic acid) (PAA), poly(lactic-co-glycolic acid) (PLGA), poly(ethylene oxide) (PEO), carrageen (CG), ethylcellulose (EC), polyvinyl alcohol) (PVA), polyvinylpovidone (PVP) xanthan gum (XG), gelatin (GL), tragacannth (TG), methylcellulose (MC), hydroxyethylcellulose (HEC), hydroxypropylcellulose (HPC), hydroxypropylmethylcellulose (HPMC), beeswax (BWX), and the poly (ethyl acrylate, methyl-methacrylate, and chlorotrimethyl-ammoniumethylmethacrilates) (PMMA) (ED-S100 and ED-RS100). These polymers have their own inherent properties when used independently. However, blends of these polymers were employed in this study in order to provide superior physicochemical and physicomechanical properties, such as abrasion resistance, chemical inertness, a high modulus, thermoplasticity as well bioadhesivity and therefore ideal for developing the VPD. In order for a drug delivery system to have superior retention within the vagina it must be highly bioadhesive. PAA, PEO and PVA are known to have substantially superior bioadhesive properties due to the presence of hydrogen bonding groups, strong anionic/cationic charges, high molecular mass, chain flexibility and surface energy interactions favoring spreading onto mucus. Their adhesiveness is also partially due to the intensity of the interfacial interactions developed between the hydrogel and the adhesion substrate. Thus, they confer bioadhesiveness to the VPD, an effect that will favor the extension of the drug residence time within the vagina as well as reducing the dosing frequency.

For an intravaginal drug delivery system to be effective as an anti-HIV microbicidal agent, the drug employed should be able to inactivate HIV replication in lymphocytes, epithelial cells and sperm cells (D'Cruz and Uckun, 2002). Given the fact that the passage of HIV-infected mononuclear cells in semen contributes to the sexual transmission of HIV (Zhang et al., 1998; D'Cruz and Uckun, 2002), the anti-HIV microbicide should be metabolized with equal efficiency by both the seminal cells and the epithelial cells of the cervico-vaginal region (D'Cruz and Uckun, 2002). AZT (Figure 3a) is one such a compound and therefore the reason for being employed. It undergoes intracellular hydrolysis yielding monophosphate derivatives which further become phosphorylated by thymidylate kinase to produce a bioactive triphosphate derivative.

Apart from the microbicidal activity of PSS (Figure 3b) it is a viscous polymer which is known to form a bioadhesive dispersion facilitating retention in the vagina for prolonged periods of time (Garg et al., 2004). Furthermore, PSS is FDA approved and has been identified and as an alternative topical vaginal microbicide following the drawbacks found with nonoxynolol-9 when used as a spermicidal and contraceptive (nonoxynolol-9 potentiated bacterial vaginosis as well as HIV transmission) (Rosenstein et al., 1998; Richardson et al., 1998; Anderson et al., 2000; Simoes et al., 2002). PSS has been shown to possess some ARV activity (Simoes et al., 2002). Likewise, it has gelling properties which may aid in preventing HIV infection by acting as a physical barrier.

Apart from playing a role of microencapsulation, PLGA will be used for the purpose of producing an acidic pH environment within the vagina since it will degrade into its respective lactic and glycolic acid units. This helps in maintaining the normal vaginal ecology by favouring the growth of Lactobacilli-containing microflora which is important entities for preventing bacterial vaginosis that commonly occur due to the presence of Gardenerela vaginalis or Mycoplasma hominus.

The main advantages of the proposed VPD is that: i) it is easier and more convenient to apply a tablet-like system than a gel or ring; ii) a bioadhesive tablet/caplet will have the desirable vaginal retention as opposed to the poor retention conferred by gel formulations (Gavin et al., 2002); iii) controlling the release of bioactives will mean less frequent applications therefore more enhanced patient compliance; iv) Gels currently under phase three clinical trials are reported to be messy and not able to effectively control bioactive release (Broumas et al., 2000; Justin-Temu et al., 2004; Bentley et al., 2006;) and need to be administered immediately prior to coitus (Kathambi, 2006).

The proposed VPD is a non-messy system and will reside in the posterior fornix of the vagina long before coitus, releasing the bioactives gradually. The VPD will utilize more than one mechanism of preventing HIV/STIs unlike existing vaginal delivery systems that utilize a single mechanism. The VPD has a diameter of not more than 10mm and is deeply inserted by an applicator into the posterior fornix of the vagina. The formulation is designed in such a manner that a steady-state concentration of bioactive agents is maintained locally (i.e. within the vaginal membrane) and in the vaginal inner mucosal muscularated stratum.

Embodiments of the invention are illustrated by the following non-limiting examples beginning with the apparent limitations of previous studies performed in an attempt to address the delivery of a pharmaceutical composition(s) for site-specific delivery and more particularly of polymers and dosage forms according to the invention.

HIV/AIDS and other sexually transmitted infections (STIs) are a burgeoning problem globally. AIDS has caused the mortality of more than 25 million people since it was first recognized in 1981 and has become one of the most destructive pandemics in history. With reference to the recent global estimates, more than 40 million people have contracted HIV/AIDS, 5 million became infected in 2005 and more than 3 million deaths occurred from HIV-related diseases in 2005 (Malcolm et al., 2006). Furthermore, it has been estimated that heterosexual transmission of HIV accounts for 90% of adult infections worldwide and that male-to-female transmission is eight times more likely to occur than female-to-male transmission. The reason is that, biologically women are more vulnerable to STIs and HIV/AIDS than men due to larger mucosal exposure to seminal fluids as a result of the anatomical structure of the vagina. Thus, once infected with an STI, most women are asymptomatic and may remain untreated, thereby increasing their susceptibility to HIV infection. The effects of untreated STIs in women tend to be more serious than in men. For instance, it may result into the development of a pelvic inflammatory disease (PID), which can subsequently lead to infertility or life- threatening ectopic pregnancy. Also, from a social-economic perspective, women are more vulnerable to STIs and HIV/AIDS due to the low incomes and meager or no authoritative status which renders them inability to refuse unsafe sexual practices. In lieu of the above, coupled with the lack of effective protective measures, it is apparent that many women are at risk of acquiring STIs or HIV/AIDS.

Thus far, several strategies have been proposed and adopted over the years to prevent sexually-acquired HIV infection. These include among others, promotion of abstinence, monogamy, condom use, reduction in the number of sexual partners and treatment of sexually related infections. Unfortunately, social-economic factors particularly in the developing world have made these strategies not as successful as expected. The most compelling solution to HIV/AIDS is an effective vaccine. However, with 25 years into the HIV pandemic, development of an effective vaccine has remained to be as elusive as ever. This has been due to various obstacles including inadequate resources, clinical trials, regulatory capacity concerns, intellectual property issues and to a large extent, the scientific challenges (Malcolm et al., 2006). It is thus evident that despite major efforts and new approaches, it is unlikely that a HIV vaccine will be available for human use within the next decade (Weber et al., 2005). It is envisaged that the biggest potential for the prevention of STIs and subsequently HIV/AIDS currently lies in a method that can entirely be controlled by the woman. The most promising are HIV preventive approaches which emphasize the need of developing effective delivery systems for microbicides. These are bioactive agents that when administered to the vagina prior to coitus, have the potential to either prevent or reduce HIV/AIDS and STI transmission through numerous well established mechanisms as shown in Figure 2.

The ultimate success of an intravaginal drug delivery system depends on the formulation and the bioactive agent. This requires consideration of several variables including the bioactive agent, vaginal physiology and the design of the delivery system. Furthermore, bioavailability is affected by numerous physiological factors and a formulation's ability to effectively deliver the drug may vary over the menstrual cycle, pH variations and the presence of co-pathogens. Understanding the mucosal immunity, could be a critical aspect in preventing HIV transmission since 80% percent of HIV transmission is through mucosal surfaces during sexual intercourse. Mucosal integrity and/or inflammation are key factors in HIV acquisition because the vaginal mucosa is the site of expression for HIV co-receptors (Van Damme, 2002). The presence of STIs is another pre-disposing factor for HIV/AIDS.

For intravaginal drug delivery, the challenge is to design a delivery system that provides a high drug concentration in the vagina over a prolonged period of time (Benkorp- Schnurch et al., 2003). For controlled, zero-order release, sustained over prolonged periods (days extending to months), solid polymeric systems may be most suitable provided they are compatible with the physicochemical nature of the drug to be delivered and the body. To date, drug release durations for intravaginal delivery systems are as follows: i) Vaginal gels (6 hours) (Wang and Lee, 2004; Bilensoy et al., 2006); ii) vaginal tablets (8 hours) (El-Kamel et al., 2002; Gavin et al., 2004); and iii) vaginal rings (71 days) (Van Laarhoven et al., 2002; Malcolm et al., 2006). For vaginal gels and tablets, the release is too rapid which ultimately requires the drug to be applied several times daily. For vaginal rings, the release period is adequate but they have been formulated for preventing HIV infection only.

This study proposes the development of a drug delivery system that will be active against both HIV and STIs with a controlled release effect from one month up to six months. The caplet-like device is referred to as a "Vaginal Polymeric Device (VPD)" throughout this study. The delivery system will have desirable bioadhesive properties and will be capable of inducing an acidic pH environment in the vagina upon degradation thus maintaining a wide spectrum of activity against pathogens. Bioadhesivity is essentially defined as the interfacial force between synthetic or natural polymeric materials and the mucus layer that covers a mucosal tissue. It occurs in three stages namely wetting, interpenetration and finally mechanical interlocking between mucin and the polymer (Jast et al., 2003). It has been reported that multifunctional polymers exhibiting bioadhesive and/or gelling properties such as thiolated polyacrylates and poloxomers represent a useful approach for the design of vaginal drug delivery systems (Bemkop-Schnurch et al., 2003). Dosage forms comprising such auxiliary agents include vaginal tablets, rings and gels.

EXPERIMENTAL SECTION

Materials

Modified polyamide 6,10 (_mPA 6,10) was synthesized using hexamethylenediamine, sebacoyl chloride, anhydrous n-hexane and cyclohexane, all purchased from Sigma- Aldrich Chemie (Sigma-Aldrich Chemie, Steinheim, Germany). The remainder of the polymers employed were commercially available. These were poly(acrylic acid) (Carbopol^® 734 and 974) (PAA 734 and 974) (Noveon lnc Cleveland, OH, USA), carageenan (CG) (Type one-kappa and alpha), ethylcellulose (Ethocel-10) (EC), xanthan gum (XG), tragacanth (TG), bovine serum albumin (BSA) (Sigma-Aldrich Chemie, Steinheim, Germany); poly(ethylene oxide) (PEO) (Union Carbide Corporation, Danbury, CT, USA); poly(lactic-co-glycolic) acid (PLGA) (Resomer^® RG504; Boehringer Ingelheim, Ingelheim, Germany); polyvinyl alcohol) (PVA), polyvinyl povidone (PVP) (Merck-Schuchardt, Hohenbrunn, Germany); gelatine (GL), beeswax (BWX) (Saarchem (Pty) Ltd., Krugersdorp, South Africa); methylcellulose (MC), hydroxyethylcellulose (HEC), hydroxypropylcellulose (HPC), hydroxypropylymethylcellulose (HPMC) (Merck- Schuchardt^®, Hohenbrunn, Germany); poly(ethyl acrylate, methyl-methacrylate, and chlorotrimethyl-ammoniumethylmethacrilate) (PMMA 100) (Eudragit^® S-100) (ED S100) and poly(ethyl acrylate, methyl-methacrylate, and chlorotrimethyl- ammoniumethylmetha-crilate) (PMMA RS 100) (Eudragit^® RS-100) (ED RS100) (Rohm & Co., Pharma Polymers, Darmstadt, Germany); calcium hydroxide, glycerol, acetic acid (Associated Chemical Enterprises (Pty) Ltd., Southdale, South Africa), polystyrene sulfonate (PSS) (sulfonated sodium salt; Scientific polymer products, Inc., Ontario, Canada); 3'-azido-3'-deoxythymidine (AZT) (Evershine Ind., Naejar Malad, Mumbai, India), methylparaben (Merck (Pty) Ltd., Darmstadt, Germany), shellac (Roeper GmbH, Hamburg, Germany), castor oil (Jayant Oils and Derivatives Ltd., Mumbai, India), barium sulphate (BaSO₄) (Merck-Schuchardt, Hohenbrunn, Germany) and simulated vaginal membranes (dialysis flat sheet membrane, M_w 12,000-14,000) (Spectrum Laboratories Inc., Rancho Dominguez, CA, Canada). The mobile phase solvents comprised of acetonitrile and methanol that were purchased from Romil-SpS™ (Cambridge, UK) including UPLC grade water (Milli-Q^® A10 System, Millipore^®, Molsheim, France). All other reagents used were of analytical grade and employed as received.

Synthesis of the polyamide 6,10 for incorporation into the VPD

The modified polyamide 6,10 was synthesized using a previous method developed by Kolawole and co-workers (2007) in which a Placket-Burman experimental design template was employed using combinations of hexamethylenediamine (HMD), sebacoyl chloride (SC), hexane (HXN), cyclohexane (C-HXN), sodium hydroxide (NaOH) and deionized water (DW). The overall chemical reaction is illustrated in Figure 4.

The modification focused on exploring the effect of volume ratios, stoichiometric variations and the addition of solvent phase modifiers such as NaOH and C-HXN on the physicochemical and physicomechanical properties of the PA 6,10 (Kolawole et al., 2007). For each variant, two solutions were prepared. The first solution comprised SC dissolved in a mixture of HXN and C-HXN while the second solution comprised specific quantities of HMD and NaOH dissolved in DW (Table 1). The concentrations of each solution were based on the combinations derived from the Plackett-Burman design after optimization in which scaling-up was performed by doubling the quantities of the solutes while keeping the quantities of the solvents constant.

The first solution was gradually added to the second to form two immiscible phases which resulted in a polymeric film being formed at the interface (i.e. by interfacial polymerization process). The polymeric film was collected as a mass by slowly rotating a glass rod at the interface. Upon collection of the polymeric mass, it was thoroughly washed, first with HXN to remove any un-reacted SC and then DW (3*300mL) to remove any un-reacted NaOH. The polymeric mass was then lightly rolled on filter paper (diameter 110mm, pore size 20μm) to remove any excess solvent and dried to constant mass at 4O⁰C over 48 hours.

Table 1 : The mass and volume relationships of the reactants and solvents employed for the synthesis of _mPA 6, 10 variants

Mass and Volume Relationships of Reactants and Solvents

HMD^{7 "^}SC² HEX⁴ C-HXN^S NaOH⁶ (9) (g) (mL) (mL) (mL) (9)

1.75 0.63 10.00 40.00 40.00 0.10

⁷ Hexamethylenediamine, ²sebacoyi chloride, "deionized water, *hexane, "cyclohexane, ^esodium hydroxide

Preparation AS-PAA and APE-PAA caplet devices for bioadhesivity testing

A Mixture Design (Extreme Vertices) template (Minitab^® software, V15, Minitab Inc., PA, USA) was statistically generated to produce various formulations comprising 11 polymer combinations as shown in Table 2. Each formulation had a equal mass of 800mg. Response optimization was then performed in which the D-optimal process was used to optimize the formulations by combining the mixture components and process factors and thereafter selecting the optimal settings for the process factors which enabled the determination of the appropriate proportions of polymers resulting in optimized AS-PAA and APE-PAA caplet devices. Table 2: Extreme Vertices Mixture Formulation Template for caplet preparation

F# _mPA 6,10 PLGA EC (mg) PVA (mg) PAA^a PAA"

(mg) (mg) (mg) (mg)

1 150 50 100 100 200 200

2 100 50 100 150 250 200

3 100 50 100 100 205 250

4 130 55 105 105 205 205

5 105 80 105 105 200 205

6 100 100 100 100 205 200

7 105 55 130 105 230 205

8 105 55 105 105 210 230

9 110 60 110 110 200 210

10 100 50 150 100 205 200

11 105 55 105 130 200 205

³AS-PAA: ally/ sucrose-crosslinked PAA b APE-PAA: ally! penta erythritol-crosslinked PAA

The polymers were weighed in triplicate and blended. Magnesium stearate (0.5%^w/_w) was added to each mixture and blended using an Erweka cube blender (Erweka Apparatebau, Heusenstamm, Germany). The blends were then granulated with 96.5 % ethanol and dried at room temperature (21⁰C) over 24 hours and thereafter compressed into polymeric caplet devices using a Carver press (Carver Inc. Hydraulic Laboratory Press, Wabash, IN, USA) at a force of 5 tons.

In-process validation tests on the polymeric devices

In-process validation tests that involved diametric hardness, mass uniformity and friability analysis were conducted to ensure the manufacturing reproducibility of the devices. A sample of 10 devices from each formulation was used for each test. Diametric hardness test was conducted on a Hardness Tester (Pharma Test, Hainburg, Germany) while friability was conducted on a Friabilator (Erweka D-63150, Heusenstamm, Germany) at 25rpm for 4 minutes with 1% set as the upper limit of acceptability. The weight of each device was determined using an electronic balance (Mettler, Model AE 240, Griefensee, Switzerland) with readings recorded to 2 decimal places.

Assessment of the bioadhesivity for the polymeric devices

Bioadhesivity testing was performed on both sets of devices (i.e. AS-PAA and APE-PAA devices). A Texture Analyser (TA.XTplus Stable Microsystems, UK) was used to conduct the bioadhesivity testing by adapting a method developed by Tambwekar et al., 2002. Principally, measuring adhesiveness ('stickiness') is conventionally performed with a cylindrical probe which is pushed (application of force) onto the surface of a sample, after which the force to pull the probe apart from the sample surface is measured. The maximum force required to detach the simulated vaginal membrane (dialysis flat sheet membrane, M_w 12,000-14,000) secured on the upper probe from the polymeric matrix, known as Peak Adhesive Force (PAF) and the Area Under the Force- Distance Curve (AUCFD-). representing the work of adhesion, were determined by the textural profiles generated profiles. In each case, the membrane and the caplet devices were hydrated for 30 minutes in simulated vaginal fluid (pH 4.5; 37⁰C) (Owen and Katz, 1999) before conducting the test. The optimized caplet devices were then subjected to the same test. Each measurement was performed in triplicate. The instrument parameter settings are shown in Table 3. Table 3: Textural settings employed for the determination the bioadhesivity of the devices.

Parameters Settings

Pre-test speed 2.0mm/sec

Test speed 2.0mm/sec

Post-test speed 1 mm/sec

Post test speed 10.0mm/sec

Trigger type Auto

Trigger force 0.4903N

Load cell 5kg

Contact time 5sec

Ex vivo bioadhesivity testing on the VPD employing a vaginal tissue from White Large pig model

The optimized APE-PAA matrix was selected for determining the bioadhesivity of the polymer-based system, on freshly excised vaginal tissue obtained from the pig model. The APE-PAA matrix was selected for ex-vivo studies due to its superior in vitro bioadhesivity potential when compared to the optimized AS-PAA matrix.

Removal of vaginal tissue from the pig model for bioadhesivity testing

A Large White female pig (84kg) was euthanized with 4OmL of sodium pentobarbitone (200mg/mL) administered intravenously. The pelvic canal of the pig was opened by dissecting through the symphysis pubis and then exposing the intra-abdominal vaginal tract. The external vaginal tract was carefully dissected from the surrounding tissues before removing the vaginal tissue (Figure 5). An incision was made through the vaginal canal to expose the inner lining of the tissue, which was then placed in an airtight specimen jar and immediately subjected to bioadhesivity testing.

Determination of the bioadhesivity of the optimized APE-PAA devices

The freshly excised vaginal tissue was secured to the textural probe. An optimized APE- PAA device was fixed on the textural platen after exposure to simulated fluid vaginal (pH 4.5; 37⁰C) for 30 minutes. Bioadhesive testing was then conducted by measuring the maximum force (N) required to detach the vaginal tissue on the upper probe from the secured optimized APE-PAA device on the textural platen as well as the work of adhesion (AUC_FD.) (Figure 6). The work of adhesion per unit area (wA αβδ), is the work performed on the system when two condensed phases α and β, forming an interface of unit area are separated reversibly to form unit areas of each of the αδ- and βδ- interfaces and is represented by Equation 1.

wAάβδ =γ α δ + γ β δ - γ αβ . Equation 1

Where γ αβ, γ αδ and γ βδ are the surface tensions between two bulk phases (i.e. vaginal tissue and optimized APE-PAA device phases) α, β; α, δ and β, δ respectively. The work of adhesion as defined in Equation 1 , and traditionally used, may be called the work of separation and a parameter for determining bioadhesivity.

Rheological analysis of AS-PAA and APE-PAA polymer solutions

Rheological behavior of 2%^w/_v AS-PAA and APE-PAA solutions were determined using a Rheometer (HAAKE MARS, Modular Advanced Rheometer system, Thermo Fischer Scientific, Karlsruhe, Germany) set at 25°C±1.00, analysis time 200s, controlled rate starting at 5.001/s dropping to 0.001/s and a shear rate ranging from 0 to 500 1/s. Preparation of an optimal VPD

From the Extreme Vertices Mixture Formulation template, the following optimized formulation was generated (Table 4).

Table 4: Statistically derived optimized matrix formulation Optimal D PA 6,10 PLGA PEO PAA CG

(mg) (mg) (mg) (mg) (mg)

1 50 300.0 125 125 200.0

The optimal formulation (Table 4) displayed superior bioadhesiveness on both the simulated vaginal membrane and freshly excised pig vaginal tissue. The maximum force and work of adhesion that indicated the extent of bioadhesivity for analysis employing the vaginal simulated membrane were 1.673±0.195N and 0.0006±8.9552x10^"4 while results for freshly excised pig vaginal tissue were 0.883±0.065N and 0.0003+0.4379x10^{" 5} respectively. However, the integrity and stability of the matrix was not desirable. The approach undertaken was to search for lead formulations which could provide the optimum matrix stability and integrity at both pH values i.e. 4.5 (human vaginal pH) and 7.0 (human seminal pH) from which an ideal formulation would finally be obtained. A One Variable at a Time (OVAT) approach was employed in searching for the lead formulations.

Preparation of preliminary lead VPDs

A total of 18 different polymers namely _mPA 6,10, PLGA, PEO, PAA, CG, EC, PVA, PVP, XG, GL, TG, MC, HEC, HPC, HPMC, BWX, ED-S100 and ED-RS100, were employed to result into 62 polymer combinations (Table 5) which were subjected to extensive screening to determine their matrix stability using the OVAT approach. Equilibrium swelling ratio was used as a determining parameter. Table 5: The sixty two formulations obtained from different polymer combinations

Polymeric Composition (mg)

F P6 PL PE PA CG EC PV XG GL ML HE HP BW ES ER MA MB MC MO ME MF MG MH Ml WT

1 150 150 100 200 200 800

2 100 150 150 200 100 800

3 100 100 200 250 150 800

4 130 155 170 205 140 800

5 120 180 160 210 130 800

6 200 150 150 200 100 800

7 105 155 105 205 196 800

8 110 165 120 230 175 800

9 140 160 110 240 150 800

10 230 120 100 110 240 800

11 210 105 130 205 250 800

12 205 150 145 200 100 800

13 150 300 150 200 800

14 250 200 150 200 800

15 50 300 125 125 - 200 800

16 50 200 125 125 - 300 800

17 50 150 125 125 - 350 800

18 50 100 125 125 - 400 800

19 150 250 - 125 - 275 800

20 150 225 - 175 - 250 800

21 150 200 - 225 - 225 800

22 160 250 - 190 - 200 800

23 170 250 - 205 - 175 800

24 180 250 - 220 - 150 800

25 190 250 - 235 - 125 800

26 195 250 - 225 - 100 - - - - - - 800

27 200 250 - 25 - 300 25 800

28 175 250 - 25 - 300 50 800

29 150 250 - 25 - 300 75 800

30 125 250 - 25 - 300 100 800

31 100 250 - 25 - 300 125 800

32 220 275 - 25 - 250 25 5 800

33 215 275 - 25 - 250 25 10 800

34 210 275 - 25 - 250 25 15 800

35 205 275 - 25 - 250 25 20 800

36 200 275 - 25 - 250 25 25 800

37 195 275 - 25 - 250 25 30 800

38 190 275 - 25 - 250 25 35 800

39 185 275 - 25 - 250 25 40 800

40 180 275 - 25 - 250 25 45 - 800

41 175 275 - 25 - 250 25 50 800

42 195 300 - 25 - 250 25 - 5 800

43 185 300 - 25 - 250 25 - 15 800

44 175 300 - 25 - 250 25 - 25 800

45 165 300 - 25 - 250 25 - 35 800

46 155 300 - 25 - 250 25 - 45 800

47 175 300 - 25 - 250 25 - - 25 800

48 175 300 - 25 - 250 25 - - - 25 800

49 175 300 - 25 - 250 25 - - - - 25 800

50 175 300 - 25 - 250 25 25 800

51 175 300 - 25 - 250 25 25 800

52 175 300 - 25 - 250 25 25 800

53 150 400 - 25 - 150 25 50 800

54 150 400 - 25 - 150 25 50 800

55 150 400 - 25 - 150 25 50 800

56 150 400 - 25 - 150 25 50 800

57 150 400 - 25 - 150 25 50 - - - - 800

58 150 400 - 25 - 150 25 50 - - - 800

59 75 400 - 25 - 200 25 75 - - 800

60 75 400 - 25 - 200 25 75 - 800

61 75 400 - 25 - 200 25 75 800

62 150 400 - 25 - 200 25 - 800

F: Formulatio number; P6: Modified polyamide 6,10; PL: Poly(lactic-co-glycolic acid); PE: Polyethylene oxide PA: Polyacrylic acid; CG: Carrageenan; EC: Ethycellulose; PV: Polyvinylakohol; XG: Xanthan gum; GL: gelatin; ML: Methylcellulose; HE:

Hydroxyethylcellutose; HP: Hydroxypropylcellulose; BW: beeswax; ES: Eudragit S100; ER: Eudragit RS 100; MA, MB, MC,

MD, ME, MFMixture of two polymers among gelatin, beeswax, xanthan gum and Eudragit S 100 25mg each; MG, MH, Ml: Mixture of two polymers among gelatin, beeswax, xanthan gum and Eudragit S 10025mg; WT: Weight Equilibrium swelling studies conducted on the preliminary formulations as a critical indicator of the formulation's matrix stability

The polymer combination derived from the Extreme Vertices Mixture Formulation template i.e. _mPA 6,10 (50mg), PLGA (300mg), PEO (125mg), PAA (125mg) and CG (200mg) was employed as the working formulation onto which the OVAT approach was applied in an attempt to find lead VPDs. A series of swelling tests were conducted on various sets of the formulations shown in Table 5 to investigate which device would have the optimal swelling at both human and pig vaginal pH (i.e. -4.5) and human/pig seminal pH (i.e. ~7.4) by employing the OVAT approach. Due to the physiological similarity between the human and pig (Buddhikot et al. 1999; Quintanar-Guerrero et al., 2001 ; Sandri et al., 2004; D'cruz et al., 2005b), simulated human vaginal and seminal fluids (Owen and Katz, 1999, 2005) prepared as indicated in Table 6, were employed for the tests. Each of the tested devices was weighed, immersed into both simulated fluids and then placed in an orbital shaking incubator (MRC Laboratory Instruments Ltd., Hahistadrut, Holon, Israel), maintained at 20rpm and a temperature of 37^°C for 24 hours. After 24 hours each device was removed from the orbital shaking incubator, gently blotted on a filter paper and then re-weighed. The swelling behavior was determined in terms of the equilibrium swelling ratio (ESR) which was calculated using Equation 2. The ESR (after 24 hours) which was the critical indicator of the formulation's matrix stability (i.e. the degree of matrix robustness), was used as a screening parameter for each formulation.

W_o Equation 2

and out of these 15 lead formulations were selected as ideal in terms of matrix stability, which were later subjected to matrix erosion studies a criterion upon which the optimization was based on. Table 6: Constituents used to prepare the simulated human vaginal and seminal fluids

SHVF¹ SHSF²

Component Quantity(g/L) Component Quantity(g/L)

NaCI 3.510 NaH₂PO4 H2O 16.974

KOH 1.400 Na₂HPO₄ 17.466 Ca(OH)₂ 0.222 Na₃C₃H5θ(Cθ2)3 8.130

Bovine serum albumin 0.018 KCI 0.908

Lactic acid 2.000 KOH 0.881 Acetic acid 1.000 CaCI₂ 1.010 Glycerol 0.160 MgCI₂ 0.920 Urea 0.400 ZnCI₂ 0.344 Glucose 5.000 Glucose 1.020

Fructose 2.720

Urea 0.450

Lactic acid 0.620

Bovine serum albumin 50.400

^''Simulated human vaginal fluid according to Owen and Katz, (1999) 2Simulated human seminal fluid according to Owen and Katz, (2005)

In vitro matrix erosion of the VPD

The 15 lead formulations were subjected to in vitro matrix erosion studies. Formulations were immersed in 10OmL of simulated vaginal fluid (pH 4.5; 37⁰C) using a sealable glass vessel (15OmL) and then placed in an orbital shaking incubator (LM-530-2, MRC Laboratory Instruments Ltd., Hahistadrut, Holon, Israel) maintained at 20rpm. After 24 hours, each formulation was removed from the medium, blotted on filter paper (diameter 110mm and pore size 20μm) and dried to constant weight at 40±0.5°C in an oven. All determinations were conducted in triplicate. The mathematical expression stated in Equation 3 was employed to determine the percentage matrix erosion (ME) (i.e. %^w/_w) of each formulation.

_ME (%) ₌ ^{0M ~ RM} _{x 100} Equation 3

OM where ME is the matrix erosion, OM original mass, and RM the residual mass.

Optimization by ANN for the best polymer combination selection

Optimization was conducted by employing the feedback Multilayer Perceptron (MLP) neural network to train the empirical input matrix erosion data with static back propagation. Figure 7 illustrates the typical construction of the MLP network while Table 7 shows the input matrix erosion data (obtained from the matrix erosion studies) that was trained. The main advantage of these networks is that they can approximate any input/output map.

A genetic algorithm with a Sigmoid Axon transfer function and Conjugated Gradient learning rule was employed for the hidden input and output layers. Figure 8 shows the network topology for the hidden input and output layers.

Table 7: The input matrix erosion data that was trained by the Multilayer Perception Network

Polymeric Composition ~F# P6 PG PA EC PV XG GL BW ES ER ME%

62 150 400 25 200 25 : : : : : 1.21

27 200 250 25 300 25 1.43

51 175 300 25 250 25 25 - 2.89

52 175 300 25 250 25 - 25 3.26

41 175 300 25 250 25 25 3.58

39 185 275 25 250 25 40 3.81

28 175 250 25 300 50 3.83 _ _ _ _ ___

40 180 275 25 250 25 45

44 175 275 25 250 25 50 - 4.45

36 200 275 25 250 25 25 - 5.57

29 150 250 25 300 100 - 6.31

50 175 300 25 250 25 - 25 - - 6.48

37 195 275 25 250 25 30 - 7.20

38 190 275 25 250 25 35 7.53

31 100 250 25 300 125 _ - 7.68

P6: Modified polyamide 6, 10; PL: Poly(lactic-co-glycolic acid); PE: Polyethylene oxide PA: Polyacrylic acid; CG: Carrageenan; EC: Ethycellulose; PV: Polyvinylalcohol; XG: Xanthan gum; GL: gelatin; BW: beeswax; ES: Eudragit SWO; ER: Eudragit RS 100; ME%: Matrix erosion percentage

Preparation of the optimized, coated, drug-loaded VPD

Biodegradable and biocompatible polymers namely _mPA 6,10 (150mg), PLGA (400mg), APE-PAA (25mg), PVA (25mg) and EC (200mg) were blended with model drugs AZT and PSS (separately and in combination) as well as radio-opaque barium sulfate (BaSO₄), using a cube blender (Erweka^® GmbH, Heusenstamm, Germany), and then compressed into robust devices on a Manesty D3B 16 station tableting press equipped with D3B oblong tooling of 22*9 mm in dimension (Manesty D3B L249LQ, Liverpool, England).. One set of devices was coated with 2%^w/_v APE-PAA while another set remained uncoated. In process validation tests were performed (employing the method described before for the devices developed for bioadhesivity testing) to ensure that the VPD device had desirable quality attributes in terms of diametric hardness, uniformity in mass and friability

Pan coating of the VPD

A dual coating process using the Thai Coater^® (Pharmaceutical and Medical Supply Limited Partinership, Yannawa, Bankok, Thailand) was employed with a protective undercoat comprising shellac and thereafter a mixture of XG and APE-PAA as an overcoat in order to prevent any irritation to the vaginal tissue during device insertion. The addition of APE-PAA was to facilitate bioadhesion of the VPD to the posterior fornix of the vagina. The process involved firstly undercoating the VPD with a combination of shellac (4mg/device), cold pressed castor oil (3mg/device) and ethanol (96%). This was followed by an overcoat of XG (2%^w/_v) and APE-PAA (2%7_V). XG was used for its viscoelastic non-collapsible swellability in order to facilitate bioadhesion of the VPD in conjunction with APE-PAA that was employed as a biodhesive polymer (Iseki et al., 2001 ; Gimeno et al., 20003; Verhoeven et al., 2006). The processing conditions utilized for effective coating of the VPDs are listed in Table 8. A non-coating period of 30 minutes was allowed after each coating phase to effect a reduction in pan temperature and avoid sticking or fracture of the undercoat or overcoat seal. The increase in weight after coating the VPD was determined using an electronic balance (Mettler, Model AE 240, Griefensee, Switzerland) while the increase in thickness was determined using a digital vernier caliper (Taizhou hangyu tools gauge and blades Co., Ltd, Wenqiao, Zhejiang, China) with a sensitivity of 0.01.

Table 8: Parameters and settings employed for coating the intravaginal bioadhesive polymeric device

Parameter Settings

Temperature 50-56⁰C

Relative humidity 23-28%

Warming-up period 10min

Pan rotation 2-3rpm

Undercoating duration 30min

Over-coating duration 60min Influence of the VPD on the micro-environmental pH of the vagina

The changes in micro-environmental pH within simulated vaginal fluid due to the presence of the VPD were assessed by incubation of 3ml_ simulated vaginal fluid (containing the VPD) in a Multi-Purpose Titrator (MPT-2) equipped with a rapid response, liquid filled glass pH micro-electrode supported on a vertical puller (Malvern Instruments Ltd., Worcestershire, UK). The changes in pH were evaluated from a pH- time profile over 30 days. The electrode calibration standards were adjusted to cover the buffer range from pH 3.5-5.5 with a linear Nemstian response maintained.

Thermal analysis of the polymeric composition and the optimized, coated, drug- loaded VPD

TMDSC was performed on the unhydrated and hydrated physical mixtures of the polymers and the VPD (Mettler Toledo, DSC1, STAR⁶ System, Schwerzenback, Switzerland). The thermal events were explicated in terms of the glass transition (T₉) measured as the reversible heat flow (ΔH) due to changes in the magnitude of the C_p- complex values (ΔC_P), melting (T_m) and crystallization (T_c) temperature peaks which are consequences of irreversible and reversible ΔH values corresponding to the total heat flow. The temperature calibration was accomplished with the melting transition of indium. The transitions of the individual polymers and their physical mixtures were compared with the transition of the composite VPD matrix. Samples were weighed (5mg) on perforated 40μL aluminum pans, crimped and then ramped from -35-230⁰C under a nitrogen (Afrox, Germiston, Gauteng, South Africa) atmosphere in order to diminish oxidation at a rate of 1°C/min. The instrument parameters and settings employed are listed in Table 9. Table 9: Temperature modulated differential scanning calorimetry settings employed for thermal analysis of the intravaginal bioadhesive polymeric device and its formulation components

Segment Type Parameter Setting "SINE PHASE³

Start

-35°C

Heating rate TC/min

Amplitude 0.8⁰C

Period 0.8⁰C

LOOP PHASE*

To segment 1

Increment 0.8°C

End 230°C

Count 436

^aSinusoidal oscillations bOscillation period

Ex vivo bioadhesivity testing of the optimized, coated, drug-loaded VPD

The excision of vaginal tissue from the pig model for bioadhesivity testing was undertaken following the method previously described for the rabbit model.

Textural profiling analysis to determine the bioadhesivity of the optimized coated VPD

Bioadhesivity of the VPD was determined using a method developed earlier employing the pig model and also described in one of our recent publications (Ndesendo et al., 2009). Briefly, the freshly excised pig vaginal tissue was secured on the textural probe and the VPD was fixed onto the heated textural platen after exposure to simulated vaginal fluid (pH 4.5, 37°C) for 30 minutes. Testing was then conducted by measuring the maximum force (N) required to detach the vaginal tissue from the fixed device. This was determined by measuring the Peak Adhesive Force (PAF) or the Work of Adhesion that was computed as the area under the curve of a Force-Distance textural profile (AUCFD).

Insertion of the VPD into the vagina of the pig

Three Large White pigs each weighing 35kg were anaesthetized with midazolam (0.3mg/kg I. M.) and ketamine (11mg/kg I. M.). 2% isoflurane in 100% oxygen was administered via a face mask to maintain anesthesia. The VPD was then deeply Inserted into the posterior fornix of the vagina of each pig with the aid of an applicator and a speculum as shown in Figure 9a and b.

X-ray imaging of the pig for detection of the VPD

To detect the presence and position of the VPD in the pig vagina after insertion, animals were X-rayed (Siemens AG, Medical Engineering Group, Erlangen, Germany) directly after device insertion and thereafter three times weekly for 2 weeks, then twice weekly for a further 2 weeks to confirm the retention of the VPD in the vagina and to qualitatively analyze its swellability and bioerosion dynamics.

In vitro drug release analysis on the optimized VPD

Analysis of the effect of device coating on the drug release

To assess the effect of coating on drug release, analysis was conducted on VPDs (coated and uncoated) containing AZT as a representative drug model due to its hydrophilicity. An VPD was immersed in a 10OmL (Umamaheshwari et al., 2004; Charde et al., 2008) simulated vaginal fluid (pH 4.5; 37°C) using a sealable glass vessel (15OmL) and placed in an orbital shaking incubator (LM-530-2, MRC Laboratory Instruments Ltd., Hahistadrut, Holon, Israel) maintained at 20rpm and a temperature of 37°C. For the determination of AZTconcentration, 3mL samples were withdrawn at predetermined time intervals over a period of 30 days and subjected to Ultra Performance Liquid Chromatography (UPLC) analysis. An equivalent volume of drug- free simulated vaginal fluid was replaced into the release medium to maintain sink conditions. The analysis was conducted in triplicate. A correction factor was appropriately applied in all cases where dilution of samples was required.

Analysis of the drug release from the coated devices containing AZT and PSS separately and in combination

For the analysis of the drug release from the coated devices containing AZT and PSS separately and in combination, the same procedure as above was employed the only difference being that in this case samples for analysis were withdrawn over a period of 72 days.

Chromatographic conditions for the analysis of AZT and PSS concentration

Quantitative analysis was performed using a Waters^® Acquity Ultra Performance Liquid Chromatographic (UPLC) system (Waters Corp., Milford, MA, USA), equipped with a photodiode array (PDA) detector and interchangeable columns, namely, a UPLC^® BEH phenyl column (1.7μm; 2.1^χ50mm) for AZT separation, and a UPLC^® BEH Ci₈ column (1.7μm; 2.1^χ 100mm) for PSS separation. The binary mobile phases were composed of water/acetonitrile (60:40^v/_v) and methanol/water (50:50^v/_v) for AZT and PSS respectively. All solutions were filtered using a 0.22μm membrane filter (Millipore Corp., Bedford, Massachusetts, USA) prior to injection onto the UPLC column. A gradient assay method was used for AZT separation with a column temperature set at 25°C, injection volume of 2μL and a UV detection wavelength of 267nm. The gradient settings for the assay method are shown in Table 10. An isocratic assay method was used for PSS separation employing methanol/water (50:50^v/_v) as the mobile phase, a flow rate of 0.2ml_/min, a column temperature of 25°C, an injection volume of 1.7μL and a UV detection wavelength of 244nm.

Table 10: Chromatographic mobile phase gradient settings used for the separation of AZT

Separation Time Flow Rate % A % B events (min) (ml_/min) (Water) (Acetonitrile)

1 0.00 0.500 60.0 40.0

2 1.00 0.500 5.0 95.0

3 2.60 0.500 5.0 95.0

4 3.50 0.500 60.0 40.0

5 3.60 0.500 60.0 40.0

Preparation of standard solutions and calibration curves

The internal standard employed for both model drugs was methylparaben (MP). Standard solutions of AZT, PSS and MP (internal standard) were separately prepared by mixing specific quantities in water/acetonitrile (60:40^v/_v) for AZT and methanol/water (50:50^v/_v) for PSS to yield a concentration of 0.1mg/mL in each case. The standard solutions employed in preparing the calibration curve of the test drug and internal standard were obtained by further serial dilutions with a final concentration range of 25- 10,000ng/mL The internal standard solution was prepared at a concentration of 5000ng/mL and was added to all samples prepared for UPLC analysis. Calibration curves were developed using blank simulated vaginal fluid (pH 4.5) and computed as a ratio of the Area Under the Curve (AUC) of AZT and PSS chromatographic peaks to that of the internal standard MP against the corresponding standard concentrations of AZT and PSS (Figures 10a and b). The solid phase extraction procedure employed in the extraction of the drugs from simulated vaginal fluid samples for UPLC analysis

This was carried out by using single use Water Oasis^® HLB 3cc (60mg) extraction cartridge (Waters Corporation, Milford, Massachusetts, USA) adapting a method developed by Notari and co-workers (2006). The Solid Phase Extraction (SPE) cartridge was conditioned with 1.OmL methanol followed by 1.OmL water MiIIi-Q. For AZT sample preparation, 1mL of sample was mixed with 1mL of acetonitrile vortexed for 1 minute and centrifuged (Nison Instrument (Shangai) Limited, Shangai, China) at 13,000rpm for 6 minutes at 24⁰C. 650μL of the supernatant was diluted by adding water MiIIi-Q (1mL) and loaded in the catridge. Thereafter, the cartridges were washed with 1.OmL of 5%^v/_v methanol in water MiIIi-Q. Analytes were eluted by washing cartridges with 550μL 0.01 M KH₂PO₄ followed by 2mL methanol. The eluate was evaporated to dryness in a slow stream of high purity nitrogen gas (Afrox, Germiston, Gauteng, South Africa). The extracted sample was re-constituted with 100μL absolute methanol, mixed with 400μL of MP, and then filtered into the injection vials using 0.22μm syringe-driven filter units (Millipore Corporation, Bedford, Massachusetts, USA) for UPLC analysis. The same procedure was followed for PSS samples. However, methanol was used as the mobile phase.

Ex vivo permeation studies through pig vaginal tissue from the VPD using the Franz Diffusion Cell Apparatus

To assess the extent of permeation of drugs across pig vaginal tissue, diffusion studies were performed using a Franz Diffusion Cell (FDC) apparatus (PermeGear Inc. Bethlehem, PA, USA) equipped with a 12mL receptor compartment, clamp and stir-bar. Freshly excised pig vaginal tissue obtained from a Large White pig was placed between the donor and receptor compartments of the FDC apparatus (Figure 11). Simulated plasma (1OmL) (pH 7.4; 37°C) (Table 11) was used in the receptor compartment and continuously agitated. VPDs containing AZT and PSS were dissolved in 5OmL simulated vaginal fluid (pH 4.5; 37°C) and assessed for their diffusivity across the porcine vaginal tissue (thickness=1.5±0.19mm; diffusion area=2.0±0.01cm²) and into the systemic circulation that was represented by simulated plasma in the receptor compartment of the FDC apparatus. Samples were withdrawn from the receptor compartment, filtered using a 0.45μm pore size Cameo Acetate membrane filter (Milipore Co., MA, USA) and analyzed by UV spectroscopy (Cecil Instruments, Cambridge, UK) at 254nm for AZT and 244nm for PSS at predetermined intervals over a period of 24 hours. An equivalent volume of drug-free simulated plasma was replaced into the receptor compartment to maintain sink conditions throughout the diffusivity study. The analyses were conducted in triplicate. A correction factor was appropriately applied in all cases where dilution of samples was required. The diffusivity of AZT and PSS across the porcine vaginal tissue was determined in terms of drug mass flux. The mass flux (mg.cirf ².h^{" 1}) of drug across the vaginal tissue was calculated at steady-state per unit area by linear regression analysis of permeation data using Equation 4.

J_ς = -^- (Equation 4) s A x t i ^f /

Where, Q_r (mg) is the quantity of AZT or PSS that diffused through the porcine vaginal tissue into the receptor compartment, A (cm^"2) is the effective cross-sectional area available for drug diffusion and t (h) is the time of drug exposure to the vaginal tissue.

Table 11: Constituents used to prepare the simulated plasma vaginal fluid.

Simulated Plasma Component Quantity(g/L)

KH₂PO₄ 0.144

Na₂HPO₄ 0.795

NaCI 9.000 Postulated mechanism of the drug permeation and dissolution dynamics from the VPD employing chemometric and molecular modelling

Chemometric and molecular structural modeling was used to deduce the transient mechanisms of diffusion and dissolution, chemical interactions and inter-polymeric interfacing during the dissolution of the VPD device and the permeation of AZT and PSS across the vaginal tissue. This approach allowed us to make predictive findings based on the chemical and physical interactions underlying the dissolution of the VPD and the diffusion of AZT/PSS from the VPD (contained in the simulated vaginal fluid) and finally the permeation of these drugs to simulated plasma fluid across the pig vaginal tissue. In addition, semi-empirical quantum mechanics were employed to generate molecular interactions and computational energy paradigms of the VPD components based on inherent interfacial phenomena underlying the mechanisms of dissolution and diffusion as provided by the inter-polymeric blended VPD. Models and graphics supported on the step-wise molecular VPD-simulated fluids and VPD-tissue interactions, polymeric interconversion, dissolution and diffusion as envisioned by the molecular behavior and stability of the gelled VPD network were generated on ACD/I- Lab, V5.11 (Add-on) software (Advanced Chemistry Development Inc., Toronto, Canada, 2000).

In vivo studies in the pig model using the optimized, coated, drug-loaded VPD

The study involved 20 female, healthy Large White pigs (35kg) divided into 4 groups of 5 each. Group 1 was a control group in which no VPDs were inserted. Groups 2 and 3 were used for testing the model drugs, and therefore VPDs with drugs were inserted. Group 4 was used as a placebo group in which VPDs with no drug (i.e. formulated with native polymers) were inserted. The study period for each group was 1 month spanning across a total of 4 months in staggered manner. In Group 2 and Group 3, VPDs containing AZT and PSS respectively were inserted after anaesthetizing the pigs. X-ray imaging was done three times a week (1^st, 3^rd and 5^th day) for two weeks and then twice a week (1^st and 5^th day) for another two weeks, each time under anesthesia. Furthermore, blood samples (1OmL) were withdrawn from the jugular vein of each pig on days 1 , 3, 5, 7, 14 and 28 while the pigs were still under anesthesia. In Group 4, placebo VPDs were inserted and then subjected to the same procedure as above. Group 1 was subjected to blood sampling procedures on the scheduled days. On the 28^th day, each pig was euthanized, followed by dissection of the vaginal tissue for drug content analysis using UPLC, and toxicity studies. The toxicity studies involved histological examination for inflammation, exocytosis, hyperplasia, hypoplasia, epithelial exudate, ulceration, polymorphonuclear infiltration and any evidence of infection. A summary of the in vivo study is shown in Figure 12.

Anesthesia, X-ray imaging and blood sampling

Each group of pigs was anesthetized with ketamine (11mg/kg I.M.) and midazolam, (0.3mg/kg I.M.)- The pigs were then intubated and anesthesia was maintained with 2% isoflurane in 100% oxygen. X-ray imaging of the pelvic region was performed (Figure 13a), and blood samples were taken following the procedure stipulated above as shown in Figure 13b. All blood samples were immediately transferred into heparinised vacutainers (BD Vacutainer^®, Plymouth, BD Beliver Industrial State, UK) and then, stored in a refrigerator at 4⁰C for one hour. The blood was then centrifuged (Nison Instrument (Shangai) Limited, Shangai, China) at 15,000rpm for 10 minutes and the supernatant was pipetted as blank plasma using an adjustable volume micropipette (Boeco Gmbh, Hamburg, Germany). The blank plasma was stored in a refrigerator at - 7O⁰C for the quantitative drug analysis using UPLC.

Vaginal tissue removal for drug analysis and histopathological studies

On 28^th day, each pig (weighing between 80-85kg) was euthanized with 4OmL of sodium pentobarbitone (200mg/mL) administered intravenously. The pelvic canal of the pig was opened by dissecting through the symphysis pubis and then exposing the intraabdominal vaginal tract. The external vaginal tract was carefully dissected from the surrounding tissues before removing the vaginal tissue (as previously done during bioadhesivity testing). An incision was made through the vaginal canal to expose the inner lining of the tissue. Transverse cuts of tissues (at the anterior, middle and posterior part of the vagina) with a cross-sectional size 1.5x3.0cm were made through the vaginal wall. Half of the tissues were kept in 10OmL of 10% formalin and then subjected to histopathological studies (including toxicity evaluation). The remaining tissue was immediately frozen using liquid nitrogen, and then stored in a refrigerator at - 70⁰C for quantitative drug analysis.

Quantitative analysis of AZT and PSS in the blood and the vaginal tissue of the Pig-

Preparation of standard solutions and calibration curves

The same procedure as previously described in the method for developing calibration curves using simulated vaginal fluid was used except that in this case calibration curves were developed using blank pig plasma. The constructed calibration curves are as shown in Figures 15a and b.

The drug extraction method from the plasma samples for UPLC analysis

This was conducted employing a single use Water Oasis^® HLB 3cc (60mg) extraction cartridge (Waters Corporation, Milford, Massachusetts, USA) adapting a method developed by Notari and co-workers (2006). The Solid Phase Extraction (SPE) cartridge was conditioned with 1.OmL methanol followed by 1.OmL ultra pure water MiIIi-Q. 1mL of acetonitrile (for AZT) was added to 600μL of pig plasma. The solution was vortexed for 1 minute and then centrifuged (Nison Instrument (Shangai) Limited, Shangai, China) at 15,000 rpm for 6 minutes at 24⁰C. 650μL of the supernatant was diluted by adding ultra pure water MiIIi-Q (1 mL) and loaded in the cartridge. Thereafter, the cartridges were washed with 1.OmL of 5%^v/_v methanol in ultra pure water MiIIi-Q. Analytes were eluted by washing cartridges with 550μL 0.01 M KH₂PO₄ followed by 2mL methanol. The eluate was evaporated to dryness in a slow stream of high purity nitrogen gas (Afrox, Germiston, Gauteng, South Africa). The extracted sample was re-constituted with 100μl_ absolute methanol, mixed with 400μL of MP, and then filtered into the injection vials using 0.22μm syringe-driven filter units (Millipore Corporation, Bedford, Massachusetts, USA) for UPLC analysis. The same procedure was followed for PSS but substituting acetonitrile with methanol as a mobile phase.

Drug extraction from the vaginal tissue for UPLC analysis

To extract the drug from the tissues, a method developed by Wang and co-workers (2009) was adapted but with some modifications. Briefly, each tissue sample (8g each) was homogenized in 16mL of simulated plasma (pH 7.4). The homogenized sample was digested by incubating it with 10 mg of subtilisin, vortexed for 1 minute, and then placed in a thermostatic bath for 1 hour at 56°C while mixing after every 10 minutes to ensure complete degradation of the tissue. After enzymatic digestion, the sample was centrifuged at 10,000rpm for 15 minutes. For the AZT sample preparation, 1mL of the supernatant was withdrawn, mixed with 1mL of acetonitrile and then pipetted into the centrifuge tubes into which 550μL of 0.01 M KH₂PO₄ was added as an extracting medium. The tubes were vortexed for 1 minute and then centrifuged at 13,000 rpm for 6 minutes at 24⁰C. 650μL of the supernatant was diluted by adding ultra pure water MiIIi- Q (1mL) followed by the addition of 2mL absolute methanol. The supernatant was placed into sample vials and then evaporated to dryness in a slow stream of high purity nitrogen gas (Afrox, Germiston, Gauteng, South Africa). The samples were reconstituted with 100μL absolute methanol mixed with 400μL of MP, then filtered into the injection vials using 0.22μm syringe-driven filter units (Millipore Corporation, Bedford, Massachusetts, USA). The same procedure was followed for the PSS sample preparation but substituting acetonitrile with methanol as the mobile phase. Instrumentation and chromatographic conditions for the quantitative analysis of AZT and PSS from plasma and vaginal tissue samples

Quantitative analysis was performed using the Waters Acquity Ultra Performance Liquid Chromatographic (UPLC) system (Waters Corporation, Milford, Massachusetts, USA), equipped with the Acquity Photodiode Array (PDA) and Evaporative Light Scattering (ELS) detectors. The columns employed were Acquity UPLC^® BEH Phenyl 1.7μm, 2.1^χ50mm column and Acquity UPLC^® BEH Ci₈, 1.7μm, 2.1^χ100mm for AZT and PSS analysis respectively. The mobile phases were composed of water/acetonitrile (60/40^v/_v) and methanol/water (50/50^v/_v) for AZT and PSS respectively. The wash solutions used, namely strong and weak washes were composed of 90/10 acetonitrile/water, 10/90 water/acetonitrile, and 100% ultra pure water MiIIi-Q, respectively. All prepared solutions were filtered using 0.22μm membrane filters (Millipore Corporation, Bedford, Massachusetts, USA) under vacuum and degassed before their use. For AZT analysis, a gradient method was used at a column temperature of 25°C, injection volume was 2μL and UV detection wavelength of 267nm. The employed gradient settings were as shown in Table 12. For PSS analysis, an isocratic method was employed using methanol/water (50/50^v/_v) as a mobile phase, flow rate 0.2mL/minute, column temperature 25⁰C, total run time of analysis was 3 minutes, injection volume was 1.7μL and UV detection wavelength of 244nm.

Table 12: Parameter settings for the AZT gradient method.

Separation Time Flow % A (Water) % B events (min) (mL/min) (Acetonitrile)

1 0.00 0.500 60.0 40.0

2 1.00 0.500 5.0 95.0

3 2.60 0.500 5.0 95.0

4 3.50 0.500 60.0 40.0

5 3.60 0.500 60.0 40.0 Histopathological evaluation of the vaginal tissue

The vaginal tissue specimens obtained were cut into three tissue blocks containing the anterior, middle and posterior sections. These blocks were processed with routine histological methodology in an automated tissue processor. They were then sectioned at 5μm, placed on slides and stained with haematoxylin and eosin in an automated stainer (Rankin Biomedical Corporation, Michigan, USA) and finally subjected to thorough histological evaluation. The evaluation was divided into three parts. The first was an evaluation on epithelial histological lesions; the second one was an assessment of the lamina propria and the third was an evaluation on the subepithelial tissues and vaginal wall. The evaluation of epithelial histological lesions encompassed: i) hyperplasia which is an increase in epithelial cell layers which finally results in acanthosis and thickening of the epithelium; ii) exocytosis which refers to transepithelial leukocyte migration of inflammatory cells; iii) exudate on the epithelial surface; and iv) ulceration. In the lamina propria, mononuclear inflammation, polymorphonuclear infiltration and foreign body inflammation were assessed. The evaluation of subepethelial tissues and vaginal wall was mainly concerned with a search for perivascular inflammation.

RESULTS AND DISCUSSION

PA 6,10 synthesis and in-process validation tests

The PA 6,10 product obtained presented as a strong white crystalline compact, sphere- like solids with irregular edges. The polymeric devices produced were sufficiently strong and robust with an average hardness of 286±0.01 N. They presented with uniformity in mass (800±0.48mg) and the friability was at an average of 0.029% which was within the set limit. Assessment of the bioadhesivity of the VPD

Values for work of adhesion for APE-PAA devices were higher than those of AS-PAA devices. The percentage difference in work of adhesion was 94.1% (Figure 16). This clearly indicated that APE-PAA devices were more adhesive than APE-PAA devices.

Response optimization of the AS-PAA and APE-PAA devices

Response optimization performed on the AS-PAA and APE-PAA devices using Minitab^® software (Minitab^® Inc., V15, PA, USA), resulted in optimum levels for each polymer i.e. PA 6,10, PEO, PLGA, CG and PAA as shown in Figure 17. The bioadhesivity values for the in vitro test were significantly higher compared to the in vivo test (P<0.01) (Table 13 and 14). The reason is mainly based on the nature of the surfaces with which the caplets made contact with to establish the bioadhesive force. The in vitro test employed artificial simulated vaginal membrane whereas the in vivo tests employed freshly excised vaginal membrane. The mucus present on the surface of the freshly excised vaginal membrane underwent a constant turn-over with the old layer being replaced by a new one. This process interrupted the establishment of hydrogen bonds and van der Waal forces between the caplet and the vaginal membrane thereby resulting in low bioadhesion. With articial membrane, there was no turn-over effect and therefore once a bond formed, it remained intact (i.e. the interaction holds) which may thus be the reason for the higher bioadhesivity observed. Furthermore, mucus degrades while artificial membranes do not and therefore the biodhesive forces established in the later, remained intact and undisturbed. The difference in bioadhesivity was 46.7% (Table 13 and Figure 18). The desirability (D) values were 0.98 and 0.92 respectively. The difference in bioadhesivity may be due to the variation in the crosslinking approach that was employed commercially for synthesizing AS-PAA and APE-PAA. AS-PAA is crosslinked with allyl sucrose while APE-PAA is crosslinked with allyl penta erythritol which has superior crosslinking ability. Table 13: Optimization of the textural parameters for AS-PAA and APE-PAA devices.

Textural Fitted Experimental SD⁰

Parameters

Force of 1.300 1.16810 0.13270

Adhesion^a/b (N) 2.000 1.67160 0.18995

Work of 0.0005 0.00032 8.6480 X 10^"6

Adhesion^a/b (J) 0.0009 0.00060 8.9552 X 10^~4

³AS-PAA matrix, ^b APE-PAA matrix, ^{c '}standard deviation

Bioadhesivity of optimized APE-PAA caplet devices on freshly exercised vaginal tissue

The peak adhesive force was 0.883±0.065N while the work of adhesion (AUCFD) was 0.0003±00.4379 X10^"5J. Results portrayed a direct correlation between the force and the work of adhesion (Table 14 and Figure 19). The optimized matrix which had a desirability (D) value of 1 , disclosed that there was a close correlation between the fitted and experimental results (Table 4). Figure 20 shows a chemometric structural model developed in our laboratories depicting caplet bioadhesion to freshly excised pig vaginal tissue with muco-epithelial cell secretions and surface bio-molecule interactions .

Table 14: Optimal bioadhesivity results for the APE-PAA matrix on freshly exercised vaginal tissue.

Textural Parameters Fitted Experimental SD^a

Force of Adhesion (N) 1.000 0.883 0.065

Work of Adhesion (J) 0.00035 0.0003 0.4379e^"5

^a Standard deviation ~~~ Rheological analysis

AS-PAA and APE-PAA are polymers that produce mucilage with short flow rheology, a property that can be associated with the high degree of crosslinking in both polymers. In terms of rheological properties, APE-PAA appeared to be different from AS-PAA as it was found to be highly viscous in comparison to AS-PAA with a subsequent higher shear stress (Figure 14). Justifiably, this may be one of the reasons contributing to the superior bioadhesion of the APE-PAA devices. In both cases, viscosity decreased as the shear rate increased (Figure 21).

Matrix swelling

The ESRs for the selected 15 lead formulations are summarized in Table 15. Depending on the polymer combination and hydrophilic/hydrophobic proportions, all devices demonstrated a different swelling equilibrium as depicted by their ESR values (Table 15). Swelling ratio describes the amount of water that was contained within the device at equilibrium and is a function of the proportion between hydrophilicity and hydrophobicity in the device network structure. Ionization of the polymer functional groups, crosslinking density, charge density, and simulated vaginal fluid ionic strength, may have played a role as well in this regard. The higher the hydrophobicity the lower the ESR and the higher the hydrophilicity the higher the ESR. The opposite holds true in both cases. Low ESR is an indication of low swelling rate and therefore high marix stability and vice versa (Baumgartner et al., 2000; Girish et al, 2008; Wen et al., 2008). Among the 15 lead formulations, F62 presented with the lowest ESR (0.011). These findings may be associated with the presence of a high quantity of PLGA (400mg) in the formulation which prevented the influx of water into the VPD matrix due to its high hydrophobicity. The presence of EC in the formulation at a relatively high quantity (200mg) may have as well attributed to the obtained results since EC is also a polymer with high degree of hydrophobicity. Most certainly, minimal quantities of the hydrophilic polymers PAA and PVA coupled the superior matrix resilience of _mPA 6,10 contributed as well to the small ESR value of F62. ED-S100 and ED-RS100 are both pH-dependent polymer materials that are only soluble at pH above 6.0 (Ohmura et al., 1991). Thus, at the simulated vaginal fluid pH (4.5), these polymers were insoluble indicating that they may have played a role in preventing the entrance of water molecules to the VPD matrix, therefore leading to the low ESR values obtained in F51 (ESR=0.023) and F52 (ESR=0.033) respectively (Table 15).

The high tendency GL, BWX and XG to form a non-collapsible networked-structure may have improved the veracity of the VPD matrix and therefore the relatively low values of ESR values obtained in the formulations containg these polymers (Table 3.19).The ESR for the best 15 formulations was in the following order F62>F27>F51>F52>F41>F39>F28>F40>F44> F36>F29>F50> F37> F38>F31 (i.e. lowest to the highest) (Table 15). Overall, the high degree of hydrophobicity in the polymers constituting these formulations, prevented rapid penetration of water molecules into the VPD matrix resulting in a low swelling tendency (Table 15) and therefore lower ESR values when compared to the rest of the tested formulations (i.e. in the group of 62 formulations).

Table 15: The selected fifteen lead formulations screened through the OVAT approach

Polymeric Composition (mg)

F# P6 PG PA EC PV XG GL BW ES ER ESR

62 150 400 25 200 25 - 0.011

27 200 250 25 300 25 - 0.020

51 175 300 25 250 25 - 25 0.023

52 175 300 25 250 25 - 25 0.033

41 175 300 25 250 25 - 25 0.041

39 185 275 25 250 25 40 0.044

28 175 250 25 300 50 - 0.050

40 180 275 25 250 25 45 0.053

44 175 275 25 250 25 50 0.060

36 200 275 25 250 25 25 0.064 . _ _ _ ₀₀₇₁

29 150 250 25 300 100 -

50 175 300 25 250 25 - 25 - - 0.073

37 195 275 25 250 25 30 0.080

38 190 275 25 250 25 35 0.082

31 100 250 25 300 125 _ 0.084

P6: Modified poly amide 6, 10; PL: Poly(lactic-co-glycolic acid); PE: Polyethylene oxide PA: Poly acrylic acid; CG: Carrageenan; EC: Ethycellulose; PV: Polyvinylalcohol; XG: Xanthan gum; GL: gelatin; BW: beeswax; ES: Eudragit S100; ER: Eudragit RS 100; ESR: Equilibrium swelling ratio

in vitro matrix erosion

The results of the matrix integrity as determined in terms of the matrix erosion percentage for the fifteen lead formulations are as shown in Table 16. Depending on the polymer combination and proportions of hydrophilic/hydrophobic polymer, all devices demonstrated different degrees of erosion as depicted in Table 16. F1 presented with the least matrix erosion (1.21%) while Formulation F31 presented with the highest matrix erosion (7.68%) (Table 16). Generally, the rate of matrix erosion from the VPDs appeared to depend mainly on the networking level of their hydrophilic and hydrophobic polymer-based microstructure. Matrix erosion in a polymer matrix is usually determined by the rate at which the polymer undergoes hydration and swelling (Roy et al., 2002; Khamanga and Walker, 2006; Sriamornsak et al., 2007; Choonara et al., 2008). Therefore, matrix erosion from the VPDs depended on the relative magnitude of polymer hydration at the moving rubbery/glassy front within the VPDs. F1 contained the highest content of PLGA (400mg) and a relativey high quantity of EC (200mg) (Table 16). These polymers are highly hydrophobic with high compressibility properties. The presence of relatively low quantities of the hydrophilic polymers in this formulation (_mPA 6,10: 150mg; PAA: 25mg; PVA 25mg) coupled with the hydrophobic nature of PLGA and EC are certainly the reasons for the minimal matrix erosion observed. Thus, the high degree of hydrophobicity prevented rapid penetration of water molecules into the polymer matrices, leading to less hydration, swelling and erosion and therefore the lowest degree of matrix erosion. Table 16: The matrix erosion percentage for the fifteen lead formulations

Polymeric Composition

F# P6 PG PA EC PV XG GL BW ES ER ME %

62 150 400 25 200 25 - 1.21

27 200 250 25 300 25 - 1.43

51 175 300 25 250 25 - 25 2.89

52 175 300 25 250 25 - 25 3.26

41 175 300 25 250 25 - 25 3.58

39 185 275 25 250 25 40 3.81

28 175 250 25 300 50 - 3.83

40 180 275 25 250 25 45 4.37

44 175 275 25 250 25 50 4.45

36 200 275 25 250 25 25 5.57

29 150 250 25 300 100 - 6.31

50 175 300 25 250 25 - 25 6.48

37 195 275 25 250 25 30 7.20

38 190 275 25 250 25 35 7.53

31 100 250 25 300 125 7.68

P6: Modified poly 'amide 6, 10; PL: Poly(lactic-co-glycolic acid); PE: Polyethylene oxide PA: Polyacrylic acid; CG: Carrageenan; EC: Ethycellulose; PV: Polyvinylalcohol; XG: Xanthan gum; GL: gelatin; BW: beeswax; ES: Eudragit S100; ER: Eudragit RS 100; ME%: Matrix erosion percentage

Development of an optimal VPD formulation employing Artificial Neural Networks

ANN optimization produced the formulation shown in Table 12 which presented the best matrix integrity (as determined in terms of matrix erosion percentage) leading to a reasonably superior control of drug release with the requisite bioadhesive properties. The sensitivity of each of these polymers against matrix erosion is shown in Figure 22. Table 17: Optimal VPD formulation as predicted by ANN simulations

PA 6,10 PLGA EC PAA PVA

(mg) (mg) (mg) (mg) (mg)

150 400.0 200.0 25 25

Micro-environmental pH variation analysis within the simulated vaginal fluid

It was generally observed that the superficial simulated vaginal fluid immediately adjacent to the immersed VPD exhibited higher pH values than the simulated vaginal fluid immediately surrounding the device. The initial pH, measured as close as possible to the device, upon insertion of the VPD into the titration system (MPT-2) was 4.5±0.01 (N=3). The pH electrode was inserted using a Narashige micro-manipulator and was submerged towards the VPD by careful hydraulic micro-movements to avoid creating any unnecessary turbulent hydrodynamic flow. A slight drop in pH was routinely recorded as the electrode passed in proximity to the VPD with a slightly more acid pH than the entire simulated vaginal fluid (pH 4.48±0.02) around a superficial surface diameter of 5mm. The pH, measured away from the superficial layer was 4.58±0.03 (N=3). The relatively higher pH at the superficial layer may be have been due to the extruding OH^" ions from _mPA 6,10, EC, PAA or PVA. The relatively lower pH observed at the VPD proximity (micro-environmental pH) could most certainly have been due to the break down of PLGA into lactic and glycolic acids. This biphasic response in pH was observed over an experimental period of 30 days. A profile depicting this sequence is shown in Figure 23.

Thermal analysis of the polymer constituents of the intravaginal bioadhesive polymeric device

The thermal stability of the constituent polymers as well as the composite unhydrated VPD was investigated by TMDSC at a temperature range from -35-23O⁰C. The polymers displayed multi-transitional thermal behaviors with multiple T₉, T_m and T₀ values (Table 18) that were attributed to the existence of reversing and non-reversing endothermic signals arising from the transient melting of molecules within each polymer.

Table 18: Critical thermal events evidenced by diverse temperature inflection peaks for the polymer constituents of the intravaginal bioadhesive polymeric device

Critical temperature transition points (⁰C)

Sample Analyzed T₉ Tc T_m

mPA 6,10 163 120; 200 65; 140

PLGA 45-55; 210 230 22; 220

APE-PAA 90 60 30; 130-170

PVA 22; 115 163; 215 30; 180

EC 100 130 170

Hydrated polymer blend 70; 160 210 22; 163; 200

Unhydrated polymer blend 170 220 200

Hydrated VPD 83; 163 160 -10; 38; 85

Unhydrated VPD 150 140; 220 222

T₉=GIaSS transition temperature T_c=Crystallization temperature T_m=IVIelting temperature

Thermal characterization of unhydrated and hydrated physical polymer blends as well as hydrated VPD

TMDSC analysis was also performed on hydrated and unhydrated physical blends of the constituent polymers of the VPD as well as the hydrated VPD in order to determine the effect of compression on the polymer blend. Thermograms obtained on the hydrated and unhydrated physical polymer blends as well as the hydrated and unhydrated VPD are depicted in Figures 24a-d and in Table 18. Overall, there was a distinct similarity between thermal events of the hydrated physical polymer blend and the hydrated VPD (Figures 23a and b). Thermograms presented with regions associated with very low temperatures (-10⁰C and -15°C) for the hydrated samples of the physical polymer blend and the VPD while dehydration was complete at 200⁰C (Figure 24a and c). For the hydrated physical polymer blend (Figure 24a), the onset of the low-temperature endothermic T_m peak (22°C) was attributed to the high moisture content in the physical polymer blend while the apparent T_m peaks (163°C and 200⁰C) resulted from the loss of residual water as heating proceeded. The T_m endotherms were distinctly separated from the total heat flow in the non-reversing signal.

The thermal behavior for the unhydrated physical polymer blend was markedly different from that of the unhydrated VPD (Figure 24b and d). This was attributed primarily to the effect of polymer compression on the physical polymer blend to produce the device. Contrary to the hydrated physical polymer blend, the unhydrated polymer blend showed fewer thermal events (Figure 24b). A single T₉ at 17O⁰C and a T_m peak at 200°C was observed (Figure 24b). Furthermore, the T₉ and T_m peaks that appeared for the hydrated physical polymer blend prior to 170⁰C, were absent in the unhydrated polymer blend (Figure 24a). This may be associated with a baseline transition at ±170°C in the reversing heat flow signal. Overall, the hydrated physical polymer blend presented with lower Tm peaks (22°C and 163°C). These observations were consistent with previous results reported by Frushour (2004), where upon hydrating a physical polymer blend, the T_m peak was reduced well below the onset temperature.

In vitro drug release behaviour

Chromatographic separation of 3'-azido-3'-deoxythymidine and polystyrene sulphonate with methyparaben as an internal standard

A UPLC assay method was used for quantifying the concentration of AZT and PSS released from the VPD. The 3 dimensional chromatographic analysis of blank simulated vaginal fluid revealed complete separation with no interfering peaks at the retention times within the UV wavelength range of 200-400 (AZT/PSS and MP) as shown by the typical representative 3D UPLC profile in Figure 25. Chromatograms depicting the retention times for MP (internal standard), AZT and PSS in simulated vaginal fluid are as shown in Figure 26a and b.

Assessment of the effect of coating on the drug release from the intravagina! bioadhesive polymeric device

The shellac/APE-PAA-coated VPD containing only one of the model hydrophilic drugs, AZT, demonstrated extended drug release when compared to the uncoated devices (28 vs 20 days) (Figure 27a and b). This was due to the shielding effect of the initial shellac undercoat and APE-PAA overcoat applied. Once the APE-PAA coating was hydrated the shellac gradually solubilized in a manner that diffusion channels formed within the coating layer. This facilitated the drug diffusion from the VPD. Furthermore, shellac (used as an undercoat) shielded the device against the ingress of release medium due to its wax-like properties (Sinha, 2003). This may be attributed to its inherent moisture protecting properties, as well as its ability to act as a plasticizer.

Analysis of the drug release behavior from the optimized coated VPD containing AZT and PSS separately and in combination

The substantial matrix integrity imparted by the polymers used to formulate the VPD resulted in the minimization of the rate of matrix disentanglement and consequently prolonged and controlled the release of AZT and PSS from the VPD. Controlled drug release representing zero-order was realized consistently over 40 days for AZT and 72 days for PSS (Figure 28). These results can be attributed to the hydrophobic nature and high compressibility of EC, PLGA and PSS, coupled with the superior matrix resilience of _mPA 6,10. The electrolytic nature of BaSO₄ may have also contributed to the prolongation and control of drug release. The PSS-loaded VPD achieved superior drug release behavior with consistent and controlled release over a period of 72 days (Figure 28d). For the VPD loaded with both PSS and AZT, the release of PSS occurred over 56 days compared to 40 days for AZT (Figure 28b and c). However this was still diminutive in comparison to the 72 days achieved from the PSS-only loaded VPD (Figure 28d). Conversely, the period of AZT release from the VPD was more prolonged than from the AZT-only loaded VPD device (i.e. 40 vs 28 days) (Figure 28a and b). This was clear that the inclusion of PSS had a significant role in controlling the release of AZT from the VPD due to the hydrophobicity of PSS and electrostatic properties (as a sodium salt) arising from its polymeric segmental charge density (Singh et al., 1995). Studies have shown that charge density of polyelectrolytes such as PSS enhance binding interactions and favors oppositely charged compounds, culminating in the lowering of the rate of desorption and diffusion thereby slowing the drug release rate from the polyelectrolyte compound (Singh et al., 1995; Vishalakshi, 1995). The hydrophobicity of PSS results from the presence of strong electrostatic charges and internal linkages (H- and S- bonds) in the residual un-sulfonated aromatic moieties of the PSS molecule (Jiang et al., 2003; Bonifazi et al., 2007). In addition, previous studies conducted on polyelectrolytes such as PSS have shown that the higher the osmotic coefficient and radius of gyration, the greater the ability to control the rate of drug release (Alvarez- Lorenzo et al., 1999; Griffiths et al., 2004; Le Cer et al., 2004; Sen et al., 2007; Thapa et al., 2009; Widen et al., 2009).

Assessment of drug permeation across the pig vaginal tissue

The flux of AZT and PSS across the pig vaginal tissue over time is shown in Figure 29. A relatively constant increase in the rate of flux occurred over the initial 12 hours and thereafter a steady-state was achieved up to 24 hours (Figure 29). This suggests that the mechanism of passive drug transport across the pig vaginal tissue became saturated with drug. The lower flux of PSS reflects its high degree of hydrophobicity coupled with the presence of strong intermolecular charges in the polymer (Knudsen et al., 2004; Pu et al., 2004; Chu et al., 2007). This may therefore have contributed to controlling the permeation of the drug across the vaginal tissue. Overall, it can therefore be proposed that most of the drug was retained within the vaginal tissue. Ex vivo bioadhesMty analysis of the optimized coated and uncoated drug-loaded VPD in the pig model

The devices that were produced were strongly bioadhesive. Textural profile analysis indicated that the uncoated devices had the lowest bioadhesivity (PAF= 1.1976+0.150N; AUC_FD=0.0019±0.0001J) compared to the PAA-coated devices (PAF=3.699±0.0.464N; AUC_FD=0.0098±0.0004J) (Figure 30a and b). This indicates the superiority of APE-PAA as a bioadhesive coating as may be attributed to its hydrophilicity, H-bonding capacity, the high molecular mass and the surface tension properties of the polymer. APE-PAA controlled the extent of interpenetration between the polymer and the vaginal mucosal/epithelial surface. The high hydrophilicity of APE-PAA enabled the formation of strong bioadhesive bonds due to the high water content within the mucosal layer of the pig vaginal tissue. The presence of OH^" and COOH^" groups in APE-PAA may have favored the formation of H-bonds between the entangled APE-PAA chains and the pig vaginal tissue that ultimately resulted in bioadhesion. In addition, the desirable surface tension of PAA facilitated spreading over the epithelial surface of the vaginal mucosal layer thereby enhancing bioadhesion.

Retention of the intravaginal bioadhesive polymeric device within the pig vagina

Analysis of X-ray images (Figure 31) revealed that the devices were maintained in the posterior fornix of the pig vagina for the experimental period up to 30 days. The devices underwent swelling and gradually eroded over time as shown in Figures 31a, b and c which is in accordance with the stipulated design in which the formulation is expected to initially swell in order to facilitate bioadhesion and thereafter gradually erode and release the drug over the vaginal tissue for the required clinical preventative effect.

Chemometric and molecular modeling of the intravaginal bioadhesive polymeric device drug dissolution and diffusion

Postulation of dissolution dynamics and subsequent effect on drug release Chemometric and computational analysis conducted in our laboratories revealed that polymer-polymer and polymer-drug ratios, as well as the ratio between the coating polymers and components of the dissolution medium contributed substantially to drug dissolution kinetics obtained. Figure 32 depicts a step-wise model of the IBPD undergoing dissolution.

Diffusion kinetics depicting the drug flux mechanism during ex vivo studies

The presence of excess simulated vaginal fluid, led to complete dissolution of the caplet in the donor compartment of the Franz Diffusion Cell apparatus during the ex vivo vaginal tissue permeation studies weakening the interactions and physicochemical associations. It was observed that approximately 21% of AZT and 14% of PSS permeated across the pig vaginal tissue in 24 hours. The actual transport of the drugs , (considering that 16% of both drugs i.e. 200mg each in a 1200mg IBPD matrix) from the donor compartment to the receptor compartment was 1.662mg and 1.180mg for AZT and PSS respectively. This indicated that only 3.46% of AZT and 2.46% of PSS permeated through the pig vaginal tissue from the donor compartment to the receptor compartment of the Franz Diffusion Cells. Thus, the total drug transport across the vaginal tissue was computed as 5.92% with equilibrium achieved after 24 hours. A chemometric model depicting the step-wise process of generating diffusion/transport channels perpendicularly to a polymer-strand localized in the IBPD is depicted in Figure 33.

Chromatographic separation and retention times of AZT and PSS from methylparaben (MP) (Internal Standard)

AZT/PSS and MP were eluted at 1.701±0.43 and 1.184±0.18 for AZT and 0.583±0.15 and 1.275±0.13 for PSS after extraction from pig plasma (Figures 34a and b). Assay method validation analysis revealed that the intra-and inter-day precision and accuracy were satisfactory (R²=0.9987 for AZT; R²=0.9998 for PSS). The signal to noise ratio at the lower limit of quantification (25ng/mL) was >10.8 in both cases. For the validation testing of the linear regression model, the weighting factors were selected to be proportional to the reciprocal of the standard deviations which was optimal under the least squares estimation with R²>0.99. The three dimensional chromatographic analysis of blank plasma revealed no interfering peaks at the retention times within the UV wavelength range of 200-400 (AZT/PSS and MP).

AZT and PSS concentrations in the blood and vaginal tissue of the pig model

The drug concentration in the plasma from day 1-28 ranged between 0.012- 0.332mg/mL (AZT) and 0.009-0.256mg/mL (PSS) (N=5) (Figure 35a). This indicates that only a minimal quantity of the AZT and PSS crossed into the systemic circulation. This may be attributed to the hydrophobicity and high matrix resilience of the polymers employed. A steady-state concentration of AZT/PSS was observed from day 14 through day, 28 indicating that no further drug crossed into systemic circulation. In either case, the plasma drug concentration had not yet reached its peak at day 28. The lower concentration value of PSS may be due to the presence of strong intermolecular forces in the polymer that reduced the rate of permeation of the drug into blood compartment (Pu et al., 2004; Chu et al., 2007). The mean drug concentration in the vaginal tissue at day 28 was 1.2148±0.062mg/ml_ (N=5) for AZT and 1.4004±0.071mg/ml_ (N=5) for PSS, while the plasma concentration was 0.332±0.014mg/mL (N=5) for AZT and 0.256±0.013mg/ml_ (N=5) for PSS (Figure 34b). This is an indication that the majority of the drug was retained in the vaginal tissue. For each pig, the vaginal tissue drug concentration at three vaginal sites (namely anterior, middle and posterior) was approximately the same (P>0.05). On average, PSS presented with the highest vaginal tissue drug concentration at the 28^th day (Figure 35b). This indicates that PSS was better retained in the vaginal tissue than AZT. This may be attributed to the high degree of hydrophobicity coupled with the presence of strong intermolecular charges in PSS. The finding corresponds well with the earlier results of drug flux analysis whereby PSS had lower flux values, indicating that its permeation rate across the vaginal tissue was slower compared to AZT. Likewise, the results of the analysis of drug concentration in the plasma revealed that a lower concentration of PSS crossed into the blood compared to AZT. Overall, the results portrayed a direct correlation (R² = 0.99) between the drug concentration in the blood and in the vaginal tissue (Figures 36a and b).

Histopathological evaluation of the vaginal tissue

In this study, negative to mild epithelial hyperplasia was observed (Figure 37a) except for pig number 3, 9 and 14 which presented with moderate to severe hyperplasia (Table 19). This was noted from an increase in the epithelial rete ridges (Figure 36b). The controls showed no epithelial hyperplasia thus showing normal vaginal epithelium (Figure 36c) (Table 19). The reason for the moderate to severe hyperplasia observed in pig number 3, 9 and 14 may have been due to the small vaginas of these pigs. In these pigs, the insertion of the IBPD was achieved with some difficulty resulting in a relatively higher degree of inflammation. Some irritation from the polymeric device was most likely responsible for the hyperplastic epithelial response. Regarding exocytosis, neutrophils were detected in intercellular spaces in the vaginal epithelium, in the process of migrating to the vaginal lumen. The exocytosis varied among the different test groups as well as the placebo, but was not found in the control animals.

The exudate attached to the mucosal surface followed the transepithelial migration and exocytosis of neutrophils into the lumen. This was found in only one of the animals receiving PSS and two that received AZT (Figure 36d). However it was not observed in the placebo and control pigs. Ulceration in the vagina refers to the complete loss of superficial epithelium due to an underlying submucosal inflammatory process and was confirmed only in a few animals from the AZT group (Table 19), and more so in pig number 9 which again may have been caused by the epithelial damage or necrosis during the insertion process due to the small size of the vagina. Mononuclear inflammatory cells (lymphocytes and plasma cells) were observed in the lamina propria and stromal tissue directly underneath the vaginal epithelium (Figure 38e). As shown in Table 19, (Figure 36e), moderate mononuclear inflammation was recorded in PSS (N=1) and AZT-containing IBPDs (N=2) as well as placebo animals (N=1), but was not visible in the controls. Polymorphonuclear cells (eosinophils and neutrophils), were found in the submucosa as well as in the superficial exudate. These cells originated from the circulation and migrated into the tissues from the blood vessels and also by exocytosis into the lumen of the vagina. The mononuclear and polymorphonuclear inflammation indicates that there was mild to moderate chemotaxis of polymorphonuclear cells as well as an antigenic stimulation of mononuclear leukocytes arising from the polymeric devices. This is normal for any foreign object introduced onto an epithelium.

In pig number 9, a foreign body granuloma and inflammation was observed in the focal area of the vaginal epithelium. The foreign material appeared as homogenous eosinophilic staining and was surrounded by numerous macrophages, epithelioid cells and multinucleated foreign body giant cells. This reaction was also associated with ulceration in the epithelium. Possibly the foreign body reaction was related to the traumatic insertion process in this pig. The perivascular inflammatory infiltrates consisted mainly of mononuclear cells with few polymorphonuclear infiltrates dispersed among the mononuclear cells in the submucosa and the wall of the vagina (Figure 36f) (Table 19). It was negative to mild and could be demonstrated in just few individual pigs (PSS: N=1; AZT: N=2; placebo: N=1).

Table 19: Histopathological findings in the vaginal tissue

- Negative + MiId +Moderate + Severe

CONCLUSIONS

This work has displayed the capability of various polymers to develop a bioadhesive intravaginal VPD. The caplet-shaped devices adhered well to the vaginal epithelium revealing the potential for application as a bioadhesive intravaginal drug delivery system. The attempt to prepare a VPD with controlled bioactive release has been successful. Furthermore, results have demonstrated that APE-PAA has superior bioadhesivity than AS-PAA and therefore more suitable for formulating a bioadhesive VPD. Bioadhesivity testing on freshly excised pig vaginal tissue confirmed that the method developed may be useful for measuring the bioadhesivity of intravaginal drug delivery systems on vaginal membranes.

Further studies resulted into robust VPDs with desirable matrix stability and integrity which confirmed the ability of the optimized formulation (PA 6,10, PLGA, PVA, PAA and EC) to superiorly control the release of the model drugs AZT and PSS over a period of' 72 days. The produced VPDs showed the potential of maintaining the acidic micro- environmental pH of the simulated vaginal fluid upon degrading which is a desirable feature in the vagina. The VPDs also displayed a substantially high thermal stability. The chemometric and molecular structural modeling approach qualitatively supported the deduction of the VPD rate of dissolution and has shown that the drug release rate was dependant on the stoichiometric parameters between the polymers, drugs and the simulated vaginal fluid. Furthermore, it was mechanistically deduced that the permeation of drug across the pig vaginal tissue during ex vivo studies was based on an osmotic gradient and depended on the degree of ionization as well as the size and molecular mass of the drug molecules. Ex-vivo bioadhesion and permeation studies revealed that the PAA-coated devices were desirably bioadhesive and a relatively substantial fraction of the drug-load was confined to the vaginal tissue. Thus, the intravaginal polymeric device developed may be suitable for use as a localized intravaginal drug delivery system for most female-related conditions, as an alternative to oral or parenteral administration Results from preliminary in vivo animal studies revealed that the VPD could adhere to the pigs' vaginal tissue for at least 30 days when inserted into the posterior fornix. During the entire period of the study, no signs of inflammation, fever or abnormal discharge were observed from the pig model. Furthermore, there were no abnormal changes in urine colour, eating habits or any loss of body mass in pigs.

Pertaining the in vivo animal studies that were undertaken in pig models, the microscopical examination of the vaginal tissue following the termination of the pigs revealed that most of the histopathological changes (such as inflammation, ulceration) were either negative or mild, indicating that the VPDs were non-toxic. The lower concentrations of AZT and PSS in the systemic circulation of the pig and the higher concentrations in the vaginal tissue, as well as the histopathological findings suggested that the developed VPDs may be suitable for localized intravaginal drug delivery and suitable for the treatment and prevention of HIV and STIs.

REFERENCES

1. Alvarez-Lorenzo C, Gomez-Amoza JL, Martinez-Pacheco R, Souto C, (1999). Microviscosity of hydroxypropylcellulose gel as a basis for prediction of drug diffusion rates. International Journal of Pharmaceutics, 180, 91-103.

2. Anderson R.A., Feathergill X., Diao M., Cooper R., Kirkipatrick P., Spear D. P., Waller C₁ Chany G. F., Doncel F., Herold B. and Zaneveld L.J. D., (2000). Evaluation of poly (styrene-4-sulfonate) as a preventive agent for conception and sexually transmitted diseases, J. Androl., 21, 862-875.

3. Baumgartner, S., Kristl J., Vrecer F., Vodopivec, P. and Zorko, B., (2000). Optimisation of floating matrix tablets and evaluation of their gastric residence time. International Journal of Pharmaceutics. 195, 125-135. 4. Benkop-Schnurch A. and Hornof M., (2003). Intravaginal drug delivery systems: Design, Challenges and Solutions, Am. J. Drug Deliv., 1 , 4, 241-267,

5. Bentley M. E., Marrow K.M. Fullem A., Chesney M., Horton D. S., Zosenberg Z. and Mayer H. K., (2002). Acceptability of a novel vaginal microbicide during safety trial among low-risk women, Fam. Plan. Pers., 32:184.

6. Bilensoy E., Rouf A.M., Vural I., Sen M. and Hincal A.A., (2006). Bioadhesive, thermosensitive, prolonged-release vaginal gel for clotrimazole cyclodextrin complex, AAPSPharmSciTech, 7, 2, 1-13.

7. Bonifazi, D., Enger, O and Diederich F., (2007). Supramolecular [60]fullerene chemistry on surfaces, Chemical Society Reviews, 36, 390-414.

8. - Broumas A. C. and Basara L.A., (2000). Potential patient preference for three day treatment of bacterial vaginosis: Responses to a new suppository form of clindamycin, Adv. Urethrane Sci and Tech. 17, 159.

9. Buddhikot, M., Falkenstein, E., Wehling M. and Meizel, S., (1999). Recognition of a human sperm surface protein involved in the progesterone-initiated acrosome reaction by antisera against an endomembrane progesterone binding protein from porcine liver. Molecular and Cellular Endocrinology. 158, 187-193.

10. Choonara, Y.E., Pillay, V., Singh, N., Khan, RA and Ndesendo, V.M.K., (2008). Chemometric, physicomechanical and rheological analysis of the sol-gel dynamics and degree of crosslinking of glycosidic poymers. Biomedical Materials, 3, 1-15.

11. Chu H, Yeo Y, Chuang KS. (2007). Entry in emulsion polymerization using a mixture of sodium polystyrene sulfonate and sodium dodecyl sulfate as the surfactant. Polym., 48, 8, 2298-2305. 12. D'Cruz, OJ. and Uckun, F.M., (2002). Pre-clinical safety evaluation of novel nucleoside analogue-based dual-function microbicides (WHI-05 and WHI-07). Journal of Antimicrobial Chemotherapy, 50, 793-803.

13. D'Cruz OJ. and UcKun F. M., (2004). Clinical development of microbicides for the prevention of HIV infection, Current Pharmaceutical Design, 10, 3, 315-335.

14. D'cruz, OJ., Erbeck, D. and Uckun F. M., (2005b). A Study of the potential of the pig as a model for the vaginal irritancy of benzalkonium chloride in comparison to the nonirritant microbicide PHI-443 and the spermicide vanadocene dithiocarbamate. Toxicologic Pathology, 33, 465-476.

15. El-Kamel A., Soker M., Naggar V. and Al Gamal S., (2002). Chitosan and sodium alginate-based bioadhesive vaginal tablets, AAPSPharmSciTech, 4, 4, 1r8.

16. Frushour, B. G., (2004). A new thermal analytical technique for acrylic polymers. Polymer Bulletin, 4, 305-314.

17. Garg S., Verman K., Anderson R.A. and Zaneveld L. J., , (2004). Rapidly disintegrating novel bioadhesive vaginal tablets of Polystyrene Sulfonate (PSS), A potential microbicide formulation, Int. Conf. AIDS, July 11-16: Abstract No.TuPeB4656.

18. Gavin E., Sanna V., Juliano C, Benfero C. M. and Giunchedi P., (2002). Bioadhesive vaginal tablets as veterinary system for the controlled release of an antimicrobial drug, acriflavine, AAPSPharmSciTech, 3, 3, 1-7.

19. Giannola L I, De Caro V, Giandalia G, Siragusa MG, Tripodo, Florena AM, Campisi G. (2007). Release of naltrexone on bucal mucosa: Permeation studies, histological aspects and matrix system design. European Journal of Pharmaceutics and Biopharmaceutics, 67, 425-433. 20. Gimeno, E., Moraru, C.I, Kokini, J. L., (2003) Effects of Xanthan gum and CMC on the structure and texture of corn flour pellets expanded by microwave heating. Cereal Chemistry, 81, 100-107.

21. Girish, K.L. aand Dhiren P. S., (2008). Evaluation of Mucilage of Hibiscus rosasinensis Linn as Rate Controlling Matrix for Sustained Release of Diclofenac. Drug Development and Industrial Pharmacy, 34, 807-816.

22. Griffiths PC, Paul A, Khayat 2, et al., (2004). Understanding the Mechanism of Action of Poly(amidoamine)s as Endosomolytic Polymers: Correlation of Physicochemical and Biological Properties. Biomacromolecules, 5, 4, 1422-1427.

23. Iseki, T., Takahashi, M., Hattori, H., Hatakeyama, T. and Hatakeyama, H., (2001). Viscoelastic properties of xanthan gum hydrogels annealed in the sol state. Food and Hydrocolloids, 15, 503-506.

24. Jast B., Li X. and Cleary G., (2003). Recent advances in bioadhesive drug delivery systems, Drug Delivery Polymers, 194-196.

25. Jiang, C, Li, H., Tripp, C. P., (2003). Infrared Method for In Situ Studies of Polymer/Surfactant Adsorption on Silica Powders from Aqueous Solution, Appllied. Spectroscopy, 57, 1419-1424.

26. Justin-Temu M., Damian F., Kinget R. and Van Den Mooter G., (2004). Intravaginal gels as drug delivery systems, J.Women's Health, 31 , 7, 834-843.

27. Khamanga, S.M. and Walker R. B., (2006). Eva luation of Rate of Swelling and Erosion of Verapamil (VRP) Sustained-Release Matrix Tablets. Drug Development and Industrial Pharmacy, 32, 1139 - 1148. 28. Kolawole O.A., Pillay V. and Choonara Y.E., (2007). Novel polyamide 6,10 variants Synthesized by modified interfacial polymerization for application as a rate- modulated monolithic drug delivery System. J. Bioact. Compat. Poly. 22, 281-313.

29. Le Cer RD, Picton L, Argillier JF, Muller G. (2004). Entrapment and release of sodium polystyrene sulfonate (SPS) from calcium alginate gel beads. European Polymer Journal, 40, 12, 2709-2715.

30. Malcolm K.R. and Woolson D.A., (2006). Delivery hope on HIV- Vaginal rings for controlled release of microbicides, Control. ReI., 52-55.

31. Mitchel D., (2006). Focus renewed on HIV microbicides, Int. Health Conf., Canada.

32. Ndesendo V. M. K., Pillay V., Choonara YE, Khan RA, Meyer L, Buchmann E, Rosin iU. (2009). In vitro and ex vivo bioadhesivity analysis of polymeric intravaginal , caplets using physicomechanics and computational structural modeling. International Journal of Pharmaceutics, 370, 1-2, 151-159.

33. Ndesendo V.M.K., Pillay V., Choonara Y.E., Buchmann E., Bayever D. N., Meyer LCR., (2008). Current intravaginal Drug Delivery approaches employed for the prophylaxis of HIV/AIDS and prevention of sexually transmitted infections. AAPS Pharmaceutical Sciences and Technology, 9, 2, 505-520.

34. Notari S., Bocedi A., lppolito G. et al, (2009). Simultaneous determination of 16 anti- HIV drugs in human plasma by high-performance liquid chromatography. Journal of Chromatography B, 831, 258-266.

35. Owen DH, Katz DF., (1999). A vaginal fluid stimulant. Contraception. 59, 2, 91-95.

36. Panttinen P., (2005). Microbicides as an option for HIV prevention, A Report for the International Task Force on Global Public Goods. 37. Pu Q, Ng S, Mok V, Chen SB. (2004). Ion Bridging Effects on the Electroviscosity of Flexible Polyelectrolytes. J Phys Chem B, 108, 37, 14124-14129.

38. Quintanar-Guerrero, D., Villalobos-Garcia, R., Alvarez-Colin, E. and Comejo-Bravo, J. M., (2001). In vitro evaluation of the bioadhesive properties of hydrophobic polybasic gels containing Λ/,/V-dimethylaminoethyl methacrylate-co-methyl methacrylate. Biomaterials, 22, 957-961.

39. Richardson B.A., Martin H. L₁ Stevens C. E., Hillier A. K., Mwatha B. H., Chohan M. P., Nyange M., Mandaliya K., Ndinya-Achola J. and Kreiss J. K., (1998). Use of nonoxynolol-9 and changes in vaginal lactobacilli, J. Inf. Dis., 178, 441-445.

40. Rosenstein I.J., Stafford M.K., Kitchen V.S., Ward H., Weber J.N. and Taylor- Robinson D., (1998). Effect on normal vaginal flora of three intravaginal microbicidal agents potentially active against human immunodeficiency virus type 1 , J. Inf. Dis., 177, 1386-1390.

41. Roy, D. S. and Rohera B. D., (2002). Comparative evaluation of rate of hydration and matrix erosion of HEC and HPC and study of drug release from their matrices. European Journal of Pharmaceutical Sciences, 16, 193-199.

42. Sandri, G., Rossi, S., Ferrari, F., Bonferoni M. C₁ Muzzarelli, C. and Caramella C, (2004). Assessment of chitosan derivatives as buccal and vaginal penetration enhancers. European Journal of Pharmaceutical Sciences, 21, 51-59.

43. Sen AK, Roy S, Juvekar V A. Effect of structure on solution and interfacial properties of sodium polystyrene sulfonate (NaPSS)., (2007). Polymer International, 56, 2, 167-174.

44. Shattock R., (2006). Current status of microbicide design, AIDS Res. Then, 3, 25. 45. Simoes J.A., Citron D. M., Aroutcheva A., Anderson R.A., Chany CJ. , Waller DP., Faro S. and Zanevelde LJ. D., (2002). Two novel vaginal microbicides (Polystyrene Sulfonate and Cellulose Sulfate) inhibit Gardenerella vaginalis and anaerobes commonly associated with bacterial vaginosis, Antimicrob. Agents Chemother, 46, 8, 2692-2695.

46. Singh MP, Lumpkin JA, Rosenblatt J. (1995). Effect of electrostatic interactions on polylysine release rates from collagen matrices and comparison with model predictions. Journal of Controlled Release. 35, 2-3, 165-179.

47. Sinha VR, Kumria R., (2003). Coating polymers for colon specific drug delivery. A comparative in vitro evaluation. Acta Pharm., 53, 41-47.

48. Sriamomsak, P., Thirawong, N., Weerapol, Y., Nunthanid, J. and Sungthongjeen S., (2007). Swelling and erosion of pectin matrix tablets and their impact on drug release behaviour. European Journal of Pharmaceutics and Biopharmaceutics, 67, 211-219.

49. Stone, A., (2002). MICROBICIDES: A new approach to preventing HIV and other sexually transmitted infections. Nature Reviews Drug Discovery, 1 , 977-985.

50. Tambwekar K. R., Gunjan, Verman K., Kandarapu R., Zaneveld L. J. D. and Garg S., (2002). Effect of different bioadhesive polymers on performance characteristics of vaginal tablets, National Institute of Pharmaceutical Education and Research (NIPER), S. A. S. Nagar, Punjab-160062, India,

51. Thapa P, Stevens HNE, Baillie AJ. (2009) In vitro drug release studies from a novel lyophilized nasal dosage form. Kathmandu University Journal of Science and Technology, 5, 1, 71-86. 52. Van Damme L., (2002). Heatlh and Sexuality Microbicides, Special report, 1-8.

53. Van Laarhoven J.A.H., Kruft M.A.B., Vromans H., (2002). In vitro release properties of etonogestrol and ethinyl estradiol from a contraceptive vaginal ring, Int.J. Pharm., 232, 1, 163-173.

54. Verhoeven, E., Vervaet, C. and Remon, J. P., (2006). Xanthan gum to tailor drug release of sustained-release ethylcellulose mini-matrices prepared via hot-melt extrusion: in vitro and in vivo evaluation. European Journal of Pharmaceutics and Biopharmaceutics, 63, 320-330.

55. Vishalakshi B. (1995). The effect of the charge density and structure of the polymer on the dye-binding characteristics of some cationic polyelectrolytes. Journal of Polymer Science: Polymer Chemistry., 33, 365-371.

56. Viriden A, Wittgren B, Larsson A. (2009). Investigation of critical polymer properties fpr polymer release and swelling of HPMC matrix tablets. European Journal of Pharmaceutical Sciences., 36, 2-3, 297-309.

57. Wang Y. and Lee H. C₁ (2004). Effects of intrinsic variables on release of sodium dodecyl sulfate from a female controlled drug delivery system, Int. J. Pharm., 282, 1- 2, 173-181.

58. Wang, Z., Li, X., Su, D., Li, Y., Wu L₁ Wang, Y. and Wu W., (2009). Residue depletion of lmidocarb in swine tissue. J. Agric. Food Chem., 57, 6, 2324-2328.

59. Weber J., Desai K. and Darbyshire J., (2005). The development of vaginal microbicides for the prevention of HIV transmission, PLo Medicine, 2, 5, 142. 60. Wen, X., Wang T., Wang, Z., Li, L., and Zhao C, (2008). Preparation of konjac glucomannan hydrogels as DNA-controlled release matrix. International Journal of Biological Macromolecules, 42, 256-263

61.Zang, H., Dornadula, G., Beumount, M. et al., (1998). Human immunodeficiency virus type 1 in semen of men receiving highly active antiretroviral therapy. New England Journal of Medicine, 339, 1803-1809.

Claims

1. An intravaginal polymeric pharmaceutical dosage form for releasing, in the vagina in a controlled and rate modulated manner, at least one active pharmaceutical composition, the dosage form comprising a suitably shaped combination of biocompatible and biodegradable polymers admixed with the active pharmaceutical composition or compositions such that, in use and over a predetermined time period, the or each active pharmaceutical composition is released from the dosage form as the polymers degrade.

2. An intravaginal polymeric pharmaceutical dosage form as claimed in claim 1 in which the dosage form is in the form of a caplet, alternatively a tablet.

3. An intravaginal polymeric pharmaceutical dosage form as claimed in claim 1 or in claim 2 in which the biocompatible and biodegradable polymers are selected so as to biodegrade at a predetermined rate and, thus, release the active pharmaceutical composition or compositions at a predetermined rate and, thereby, achieve a selected release profile.

4. An intravaginal polymeric pharmaceutical dosage form as claimed in any one of the preceding claims in which the polymers to have bioadhesive properties and adhere, in use, to the vaginal wall, preferably in the region of the posterior fornix, alternatively, the surface of the cervix, until the polymers have degraded and the active pharmaceutical composition or compositions has or have been released.

5. An intravaginal polymeric pharmaceutical dosage form as claimed in any one of claims 1 to 3 in which the polymers to have gelling properties.

6. An intravaginal polymeric pharmaceutical dosage form as claimed in any one of claims 1 to 3 in which the polymers have bioadhesive and gelling properties.

7. An intravaginal polymeric pharmaceutical dosage form as claimed in any one of the preceding claims in which the polymers or at least one polymer of a blend of polymers also functions to inhibit a disease causing microorganism infection.

8. An intravaginal polymeric pharmaceutical dosage form as claimed in claim 7 in which the polymers or at least one polymer of a blend of polymers also functions to inhibit infection by microorganisms causing sexually transmitted diseases, preferably the Human lmmunodefficiency Virus (HIV).

9. An intravaginal polymeric pharmaceutical dosage form as claimed in any one of the preceding claims in which the polymer or polymers are hydrophilic, alternatively hydrophobic.

10. An intravaginal polymeric pharmaceutical dosage form as claimed in any one of claims 1 to 8 in which the poOlymers are a blend of hydrophilic and hydrophobic polymers.

11. An intravaginal polymeric pharmaceutical dosage form as claimed in claim 9 or in claim 10 in which the polymer or polymers are selected from amongst the group comprising poly(acrylic acids) (PAA), poly(lactic acids) (PLA), carageenans, polystyrene sulfonate, polyamides, polyethylene oxides, cellulose derivatives, poly(vinylpyrrolidone) (PVP), polyvinyl alcohol) (PVA), Chitosan, poly(ethy) acrylate, methyl-methacrylate, and chlorotrimethyl-ammoniumethylmethacrilate (PMMA), hydroxyapatite, gum-based polymers such as xanthan gum and their variants or various permutations and derivatives of the said polymer-types.

12. An intravaginal polymeric pharmaceutical dosage form as claimed in claim 11 in which at least one polymer of a blend of polymers is a poly(acrylic acid) (PAA) polymer which, in use, enables the said polymer or polymers to function as a bioadhesive substance..

13. An intravaginal polymeric the polymer or polymers are crosslinked.

14. An intravaginal polymeric pharmaceutical dosage form as claimed in claim 11 in which the crosslinking reagents are selected from a class of biocompatible inorganic or organic salts, used in the crosslinking reactions of the polymer or polymer and pharmaceutical composition(s), and are ionic of either mono-, di-, or trivalent nature, examples of which are sodium chloride, aluminium chloride or calcium chloride.

15. An intravaginal polymeric pharmaceutical dosage form as claimed in any one of the preceding claims in which the dosage form achieves, in use, a rate- modulatable release of an active pharmaceutical composition admixed with the polymer or polymers, said rate modulation being, at least partly, achieved by the the architectural structure of the polymeric network which is, in turn, a function of the different permutations of the hydrophilic and hydrophobic polymers, pharmaceutical composition or compositions, and inorganic and/or organic salts which are selected to achieve a desired rate modulated release of the or each active pharmaceutical composition.

16. An intravaginal polymeric pharmaceutical dosage form as claimed in claim 11 in which a pre-determined rate-modulated release profile is controlled, in use, by the rate of polymeric hydration within the system which is a function of the pKa, concentration and valence of the release rate-modulating chemical substances used.

17. An intravaginal polymeric pharmaceutical dosage form as claimed in any one of the preceding claims in which the dosage form is capable of inducing an acidic pH environment in the vagina upon degradation and thus maintaining a wide spectrum of activity against pathogens.

18. An intravaginal polymeric pharmaceutical dosage form as claimed in any one of the preceding claims in which the polymers are multifunctional polymers, preferably barium sulphate, alternatively hydroxyapatite.

19. A method of manufacturing a pharmaceutical dosage form as claimed in any one of the preceding claims.