Poly [D,L-lactide-co-glycolide (PLGA)] 7525 7E (lactide/glycolide ratio, 75/25; molecular weight, 113,000 Da; inherent viscosity, 0.74 dl/g) was obtained from Lakeshore Biomaterials, Inc. (Birmingham, AL, USA). Butyl stearate was purchased from TCI Shanghai Development Co., Ltd. (Shanghai, China). Methylene dichloride was purchased from Beijing Chemical Works (Beijing China), and polyvinyl alcohol (PVA; Mw 13,000–23,000) was supplied by Sigma-Aldrich (Merck KGaA, Darmstadt, Germany). Isoperidone was synthesized in our laboratory. Sodium-dodecyl sulphate (SDS) was purchased from Baotaike Bio-Technology Co., Ltd, Shenyang, China.

Healthy female beagle dogs aged 1–2 years and weighing 12±2 kg were supplied by Yadong Laboratory Animal Research Center (Nanjing, China). The animals were kept in iron cages at 20±4°C, with 45–65% relative humidity and a 12 h light/dark cycle. They were fed with granulated feed and had free access to water. All animal experiments were approved by Institutional Animal Care and Use Committee of Jilin University (Changchun, China).

An O/W emulsion solvent evaporation method was used to prepare the microspheres (12). To prepare the oil phase, the isoperidone, additives and PLGA were briefly dissolved in 13.8 ml methylene dichloride (DCM). This was slowly added to 1.7 liters 0.5% (w/v) PVA aqueous solution (pre-saturated with 0.5 ml DCM) for 1.5 min and homogenized at 211 × g for another 3.5 min at 20±0.5°C. The mixture was stirred at 20±0.5°C for 4 h until the microspheres hardened. The microspheres were then screened for sizes of 25–154 µm using sieves, washed three times with 1 liter ddH₂O and lyophilized with 10% (w/w) mannitol as lyoprotectant. The formulas of the isoperidone microspheres are presented in Table I.

Prior to imaging, dried microspheres were mounted onto metal stubs. Conductive double-sided adhesive tape coated it with a thin layer of gold under a vacuum was used and analyzed with an accelerating voltage of 20 kV. The microspheres were frozen in liquid nitrogen for 30 min and cut with a knife blade. For the fluorescence study, an epoxy resin embedded cutting method was used to acquire sections (epoxy resin:curing agent, 6:1 v/v) of F1 and F4. Microspheres (20 mg) were embedded in a 200 µl medium at 40°C for 2 h and cut with a knife blade. The surfaces of both the microspheres and sections were observed by scanning with an electron microscope at an original magnification of ×1,000 for F1 and ×750 for F4 (JXA-840; JEOL, Tokyo, Japan).

A LS13 320 laser particle size analyzer (Beckman-Coulter, Inc., Brea, CA, USA) was used to determine the mean particle diameter and particle size distribution. The particle size was expressed as the volume-weighted mean particle diameter (µm).

To determine the encapsulation efficiency of isoperidone in the microspheres, 2 mg of the lyophilized microspheres were dissolved in 1 ml acetonitrile. Next, the resulting solution was diluted in methanol:ultrapure water (80:20, v/v) to 20 ml, filtered through a 0.45 µm membrane filter (Thermo Fisher Scientific, Inc., Waltham, MA, USA) and the isoperidone content was analyzed by high-pressure liquid chromatography (HPLC). The HPLC system consisted of a Waters 1525 binary pump and a Waters 2487 Dual Absorbance Detector (Waters Corporation, Milford, MA, USA) set at 280 nm. An Agilent Extend C18 column (4.6×250 mm, 5 µm; Agilent Technologies, Inc., Santa Clara, CA, USA) was used for drug separation. A mixture of methanol:water:triethylamine [80:19.5:0.5 (v/v/v)], adjusted to pH 10.22 with acetic acid, was used as the mobile phase, as previously described (21). The flow rate was set at 1 ml/min. Chromatography was performed at 25°C with an injection volume of 20 µl. The drug loading percentage and encapsulation efficiency were calculated as follows: Encapsulation efficiency (%)=(drug loading determined by HPLC/theoretical drug loading) ×100 (1).

In vitro drug release was measured in PBS (pH 7.4) containing 0.02% (w/v) sodium azide and 0.1% (w/w) SDS. Approximately 1 mg microspheres were suspended in 35 ml PBS in a 50 ml centrifuge tube. The tube was shaken horizontally at 100 rpm in a shaking bath, which was maintained at 37±0.5°C. Following centrifugation at 844 × g for 5 min at 25°C, 20 ml PBS was removed from the tube at 4 h, and on days 1, 2, 4–42, 46, 50–82. The medium removed from the tubes was replaced with an equal amount of fresh PBS. The collected samples were filtered through the 0.45 µm filter and subjected to further HPLC analysis, as described above.

To examine accelerated drug release, the temperature of the shaking bath was maintained at 45±0.5°C. The sampling times were 4 h, and days 1, 2, 3, 4–15. The same steps as in the 37±0.5°C in vitro drug release studies were followed. The experiment was repeated for surface observation of microspheres using SEM at an original magnification of ×1,000 for F1 and ×750 for F4 at 1 and 7 days.

Microspheres of 30, 40 or 50% theoretical DL with 3% (w/w) butyl stearate or without, were scanned on a DSC (Mettler-Toledo GmbH, Greifensee, Switzerland). Microspheres of 30% theoretical DL with 3% stearic acid (w/w) were also scanned as a control. Scans were carried out at 10°C/min over the temperature range of −50 to 200°C. The rate of nitrogen was set at 50 ml/min.

Laser-scanning confocal microscopy (LSM 710, Germany) was used to monitor the distribution of green fluorescence of isoperidone excited at 488 nm. Prepared microspheres were placed on a glass slide, wet with a drop of deionized water and covered by a coverslip. The glass slide was set on the platform of the microscope to find the microsphere surface. The microsphere was scanned by laser-scanning confocal microscopy with a 4 µm scanning width until the green fluorescence disappeared.

Pharmacokinetic studies of F1 and F4 were performed on female beagle dogs weighing 12±2 kg (n=3). For each dog, the microspheres were suspended in a 5 ml aqueous suspension in a penicillin bottle, consisting of 0.9% (w/w) 5-carboxymethyl cellulose sodium (CMC-Na), 0.127% (w/w) sodium phosphate monobasic dihydrate, 0.1% (w/w) anhydrous citric acid, 0.8% (w/w) sodium chloride and 0.086% (w/w) sodium hydroxide. The resulting suspension was drawn into a 5 ml syringe and injected slowly and deeply into the hip muscle of each dog. Blood samples were collected into heparinized tubes before (0 h), at 4 h after, and each day for 40 days after the injection. They were separated by centrifugation at 2,110 × g for 5 min at 4°C. Plasma samples were collected and stored at −80°C until analysis.

Dosage of isoperidone was calculated as: 6 mg/426.5 × (526.5/70 kg) × 0.28×1.87×30=1.66 mg/kg (2), where 6 mg is the dose of the paliperidone tablets for humans, 426.5 is the molecular weight of paliperidone, 526.5 is the molecular weight of isoperidone, 70 kg is the body weight of an adult, 0.28 is the oral bioavailability of paliperidone, 1.87 is the dosage conversion factor, and 30 is the number of days.

A liquid extraction method was used to process the plasma prior to analysis by HPLC-MS/MS (Thermo TSQ Quantum Access triple-quadrupole mass spectrometer; Thermo Fisher Scientific, Inc.), which was equipped with a positive ion electrospray ionization (ESI), and an Thermo Scientific™ Excalibur™ workstation (version 1.4 Sur1; Thermo Fisher Scientific, Inc.). A mixture of n-hexane:methylene dichloride:isopropanol (2:1:0.1, v/v/v) was used to extract the solvent. The internal standard was a clozapine solution (20 ng/ml). The clozapine solution (10 µl) and sodium hydroxide solution (25 µl; 0.1 mol/l) was added to 50 µl plasma. The mixture was extracted with 3 ml extracting solvent. Following vortexing for 3 min, the system was centrifuged at 492 × g for 10 min at 4°C. The supernatant was separated, dried by flowing nitrogen at 35°C, redissolved in 100 µl mobile phase [acetonitrile:water, 60:40 (v/v), containing 0.4 mmol/l ammonium acetate and 0.01% (w/w) acetic acid] and analyzed. Quantification was performed in multiple-reaction monitoring mode of the transitions m/z 427.31→207.18 for paliperidone. The nebulizer pressure was 50 psi and the flow rate was 0.2 ml/min.

SPSS 13.0 statistical software (SPSS Inc., Chicago, IL, USA) was used to perform the statistical analyses. Student's t-test was used to compared differences between two groups. Each experiment was repeated three times. Data were presented as the mean ± standard error of the mean. P<0.05 was considered to indicate a statistically significant difference.

Table I shows the actual drug loading, encapsulation efficiency, mean particle diameter and particle size distribution of the microspheres prepared with different additives. The drug encapsulation efficiencies were between 83.63±9.56 and 100.85±3.81%, indicating that they were not significantly affected by butyl stearate and stearic acid. The mean particle size of the microspheres ranged from 80.35 to 89.97 µm and the particle size distribution was narrow.

Fig. 2 shows the spherical morphology of the microspheres. The SEM results of the accelerated testing (Fig. 2A) revealed that on day 1 a number of pores surrounded by everted matrices appeared on smooth surfaces of microspheres F4. However, fewer pores were observed in the core of microspheres with 3% butyl stearate (Fig. 2B). This phenomenon was not observed in microspheres F1. Along with the degradation of PLGA, the microspheres broke down into pieces on day 7. The pieces of F4 aggregated and became more compact, compared with F1. When the microspheres were cut after being frozen in liquid nitrogen, shell structures were observed for both F1 and F4 (Fig. 2B). While the pores were closed in the shell for F4, they were inside for F1.

Fig. 3A shows the release profiles of isoperidone of F1, F2, F3, F4, F5 and F6. The profiles of all of the microspheres could all be divided into three parts, excluding F4: i) 0–8 day lag phases; ii) 8–14 day phases; and iii) 16 day quick release phases and subsequent slow release phases. Isoperidone was released from F4 with near-zero-order release kinetics. When the additional content of butyl stearate was relatively low, such as 1 or 2%, the periods of quick release phase were reduced, while the total release periods were only slightly altered. The release periods were prolonged by increasing the butyl stearate content from 1 to 3%. However, the rate of drug release increased when the butyl stearate content was above 3%. This was especially the case for F6 with 5% butyl stearate, whose quick release phase was faster than F1. The in vitro release period for the F7 and F8 formulas was greatest at 3 weeks due to the fast release period from 8 to 16 days (Fig. 3B).

In the accelerated testing, the periods of drug release decreased to 15 days (Fig. 3C). Butyl stearate still decreased the drug release rate and F5 had the longest drug release period compared with F3, F4 and F6. These results suggested that butyl stearate could increase the thermodynamic stability of the microspheres, which has been previously reported (22).

Three types of peaks were observed in the thermograms: Butyl stearate melting peaks, PLGA glass-transition peaks (Tg) and isoperidone crystallization peaks (Fig. 4). The integral of the isoperidone crystallization peak was large when DL was elevated from 30 to 50%, or when 3% butyl stearate was added. The addition of 3% stearic acid did not increase the area of the isoperidone crystallization peaks of microspheres with 30% DL.

Isoperidone emitted green fluorescence when it was excited at 488 nm due to similar structures to benzisoxazole (23). The laser-scanning confocal microscope recorded the green fluorescence intensity of each scanning layer. The green fluorescence emission intensity was relatively weak and absorbed by PLGA. This led to a darkening of the fluorescence by the scanning depth in microspheres (F1; Fig. 5A). From 12 to 44 µm scanning depth, the center of the circular scanning surface of F4 was brighter than the surrounding part, which darkened during the scanning process (Fig. 5B). Fig. 5C shows the fluorescence of sections from F1 and F4, indicating that the drug content close to the cores was higher in F4, compared with F1.

A very low initial burst release was observed for F1 and F4 (Fig. 6). Following the initial burst release, the drug release became very slow with the lag phase lasting for approximately 4 days. The fast release phase lasted from 4 to 22 days for F1 and from 4 to 26 days for F4. On days 16 and 20, there was F4 release was significantly higher, compared with F1 (P<0.05). The results were consistent with the in vitro drug release results, and implied that 3% butyl stearate may have produced more sustained release microspheres.