NTX, PF-127, and TAβCD were purchased from Sigma-Aldrich (Schrelldorf, Germany). Analytical grade ethanol and inorganic salts were obtained from Merck (Darmstadt, Germany). Cellulose acetate membrane (molecular weight cutoff 12,000 Da) was obtained from Sigma (Germany). All other chemical reagents used were of pharmaceutical grade.

The PF-127 solutions were prepared by the cold method.20 A weighed amount of PF-127 was slowly added to a cold aqueous solution (5–10°C) with gentle mixing until complete dissolution of the polymer. The solutions used included a phosphate-buffered solution of the polymer at different pH values. When inorganic salts (NaCl, Na₂SO₄, Na₂HPO₄) were used, they were dissolved in distilled water. The polymer was then slowly added to the cold water as described.

An appropriate amount of NTX (10 mg/mL) was added to each formulation. All gels were kept overnight at 4°C until a homogenous solution was obtained.

The binary system of NTX/TAβCD in different molar ratios (1:1, 1:2, and 1:4) was prepared by the kneading method, ie, TAβCD was wetted in a ceramic mortar with ethanol:water 50% (v/v) solution until a paste was obtained. The required amount of NTX was then added slowly whilst grinding, and the slurry was kneaded for about 45 minutes. During this process, an appropriate quantity of solvent was added in order to maintain a suitable consistency. The product was then dried at 40°C over 24 hours and added to the poloxamer solution.

The optimum NTX/TAβCD complex was characterized by thermal analysis, performed using a differential scanning calorimeter (DSC-60, Shimadzu Co., Kyoto, Japan). Thermograms of the different samples (inclusion complex, physical mixture, and pure substances) were obtained from differential scanning calorimetry equipped with a thermal analysis data system. Weighted samples (6–8 mg) were contained in sealed aluminum pans and scanned at a rate of 10°C/min, between 0°C and 300°C, using nitrogen as a purging gas. The thermal analysis was carried out over 300°C during preliminary runs, but only thermal events observed over 240°C involved decomposition of the materials.

In this study, diffusion cells were used to evaluate NTX release from the poloxamer hydrogels.21 The diffusion cells are comprised of two compartments separated by a cellulose acetate membrane (cutoff 12,000 Da), which does not constitute a barrier against NTX diffusion. The donor compartment was filled with 1 mL of formulation, and the receiver compartment was filled with pH 7.4 phosphate buffer (30 mL). Temperature was maintained at 37°C for all experiments. At the specified time, aliquots (500 μL) of medium on the receiver side were taken and replaced with an equal volume of fresh phosphate buffer, in order to maintain sink conditions. The collected samples were analyzed by ultraviolet spectroscopy at 281 nm. Each experiment was performed in triplicate.

The effect of PF-127 concentration (20%, 25%, 30%, and 35%) on drug release at 37°C is shown in Figure 2. It is apparent that the drug release rate decreases as the PF-127 concentration increases. Increasing the concentration of PF-127 from 20% (w/v) to 35% (w/v) decreased the cumulative percent drug released after eight hours from 80% to 50%. As shown in Figure 2, the highest concentration of PF-127 could sustain NTX release at about 55% in 10 hours and 70% in 48 hours, which is away from our desirable sustained profile release.

A hydration layer surrounds PF-127 molecules at low temperatures in aqueous solution. However, when the temperature is raised, the hydrophilic chains of the copolymer become desolvated as a result of the breakage of the hydrogen bonds that had been established between the solvent and these chains. This phenomenon favors hydrophobic interactions among the poly oxypropylene domains, leading to increased chain entanglement and gel formation.22 This entanglement is more marked at higher concentrations of PF-127, yielding an increase in gel strength and consequently a decrease of the drug release rate.27 We should mention that 30% and 35% PF-127 hydrogels were very viscous and have very low syringability, so are not suitable vehicles for subcutaneous injection, so 25% PF-127 which has similar release pattern (P > 0.05) was selected for the next experiments.

The effect of pH on NTX release was studied at vehicle pH values of 5.5, 7.4, and 8.5 using 25% PF-127 phosphate-buffered solutions (Figure 3). The results showed that pH 5.5 and 8.5 produced faster release of NTX among the studied formulations. This indicates that when the pH of polymeric solution is lower or higher than 7.4, diffusion of NTX is increased. It may be concluded that ionization of hydroxyl groups of polyethylene blocks could increase interchain movement and presence of hydronium or hydroxide ions between these chains, and have effective roles in polymeric chain distance and easier diffusion of drug molecules.

Most pharmaceutical PF-127 gels are formulated with salts, either for buffering or ionic strength adjustment.28 It has been shown that the diffusion of drugs within the gel network is significantly affected by salts. We have examined the effect of NaCl, Na₂SO₄, and Na₂HPO₄ (1% w/v) on drug release from PF-127 gels (Figure 4). The results showed no significant effect of any of the salts on drug release rate (Figure 3). This is very different from earlier results where the same salts were shown to decrease diffusion of propranolol across a membrane from F127 gels containing salts.29 We believe that this discrepancy can be explained as follows. The water structure causes salts to reduce the critical micelle concentration and temperature of PF-127, make the poly(ethylene oxide)-poly(ethylene oxide) chain interactions stronger due to a decrease in hydrogen bonding with water, and thus make the poly(ethylene oxide) chain network tighter. Although the inclusion of salts results in a tighter micelle network, it also increases the water uptake rate into the gel due to an osmotic effect.30 The net result is that salts do not have a measurable effect on drug release rate in prepared NTX hydrogel.

The release profiles of NTX and NTX/TAβCD binary systems in different molar ratios incorporated into 25% w/v PF-127 formulations are presented in Figure 5. The data show that hydrophobic TAβCD slows the release rate, and this effect is dependent on the amount of TAβCD in the matrices. Cumulative release of the drug after 48 hours approached 85% for PF-127 hydrogel alone, compared with less than 60% in a gel containing1:4 NTX/TAβCD complex. This retardation of the drug release rate with TAβCD may be attributed to formation of complexes in the kneaded product.

Problems in controlling drug release from polymeric carriers are common, particularly for matrices composed of hydrophilic polymers, eg, PF-127 hydrogels. In such cases, drug release from the carrier may be rapid and characterized by a burst release.21 Narasimhan and Langer31 showed that burst release was controlled by the solubility of drug, the drug diffusion coefficient and the initial drug distributions within the polymeric carrier. Retarded drug release, through a dissolution-limited mechanism, is a common objective for the use of many poorly water-soluble cyclodextrin complexes.32

Poorly water-soluble alkylated cyclodextrin derivatives, such as TaβCD, are useful as slow-release carriers for water soluble drugs.33,34 Furthermore, the incorporation of these agents into polymeric drug delivery systems frequently permits a greater degree of controlled drug release rate and time.35

TAβCD forms noncovalent and poorly water soluble complexes with NTX and reduced drug solubility in aqueous medium and total concentration of diffusible species, resulting in a retarded release. In vitro release experiments showed that the formulations containing drug/TAβCD in molar ratios of 1:2 and 1:4 with no significant difference were able to prolong and control NTX release for more than 144 hours. The optimum formulation showed more than three-week stability at refrigerator temperature and 72 hours at 37°C.

The solid binary systems were analyzed by means of differential scanning calorimetry to detect possible altered thermal properties with regard to the pure substances. When guest molecules are incorporated in the TAβCD cavity, their melting, boiling, and sublimation points generally shift to a different temperature or disappear at the temperature range within which the cyclodextrin lattice is decomposed.36 The differential scanning calorimetry curves for NTX, TaβCD, and respective drug-carrier combinations in a molar ratio of 1:2 are shown in Figure 6. The thermal curve of NTX indicated a mixture of amorphous states and a characteristic broad endothermic peak at 85.19°C. The TAβCD thermogram displayed an endothermic peak at 190.77°C due to the fusion process of this crystalline molecule. The appearance of two endothermic peaks corresponding to the fusion of both components was also evident in the thermogram of the physical mixture, as if this differential scanning calorimetry curve was the superimposition of those for the components analyzed separately. The shape of the NTX melting peak was unaffected by blending with TAβCD, and hence the drug maintained its original amorphous state in the physical mixture. The complete disappearance of the drug endothermic peak was found in the kneaded preparation. The absence of the NTX fusion peak indicated the existence of interactions between drug and TAβCD in the solid state, and could be considered as an indication for the formation of real inclusion complexes. The sustained release of drug from the hydrogel containing NTX/TAβCD binary system supports this theory.