Human Insulin solution (50 IU/ml) was prepared in tris buffer pH 7.4 (10 mmol) containing 0.8 mg/ml m-cresol and mixed with solution of HP-β-CD (24 mg/ml) in tris buffer pH 7.4 (10 mmol) in a beaker and stirred for 10 to 15 min on magnetic stirrer. Then zinc chloride solution (0.4 mg/ml) was added in the form of micro-droplets to the above insulin HP-β-CD solution with continuous stirring, after complete addition solution was stirred for further 10 to 15 min to obtain opalescent solution of insulin. The above insulin-zinc mixture was mixed with solution of stearic acid (0.5% w/v), α tocopherol acetate (0.04% w/v) and SPC (0.04% w/v) in t-butanol:water mixture (9:1) solution in a beaker with stirring for 5 min. The resulting opalescent solution was frozen at −40° for 3 h in a tubular glass vial and freeze dried using table top laboratory freeze dryer (Lanconco®, USA) for 36 h. After complete drying, vacuum was released with nitrogen gas to allow purging of vials with nitrogen which were then immediately capped with butyl rubber stoppers and aluminum seal. Sealed vials were stored at 2-8°.

Insulin analysis was performed by gradient controlled HPLC equipped with UV/Vis detector (Agilent 1200 series RRLC®) and a reversed phase C₁₈, 1.8 μm, 4.6×50 mm column. The gradient system consisted of mobile phase-A, mixture of buffer and acetonitrile (82:18), and mobile phase-B, mixture of buffer and acetonitrile (50:50). The flow rate was 1 ml/min with oven temperature of 40° and detection wavelength at 214 nm. The nanoparticles obtained after lyophilization were dissolved in a mixture of 0.01N HCl and ethanol (99%) with solvent ratio of 4:1, respectively. The solution was filtered though 0.22 μm syringe disc filter (Ultipor®N66, Pall life Sciences) and injected.

Samples of nanoparticles were diluted with 0.22 μm filtered tris buffer pH 5.5 solution, then by dynamic laser light scattering (Brookhaven Instruments BI 90 Particle sizer, Holtsvilee, New York) at room temperature. The laser power is 5 mW HeNe vertically polarised. The morphology of the human insulin nanoparticles was viewed using a conventional scanning electron microscope (JSM 5400, Joel, Japan) at an accelerating voltage of I5 kV. One drop of the nanoparticle suspension was placed on a graphite surface. After air drying, the sample was coated with gold using Ion Sputter.

The entrapment efficiency was measured by dispersing human insulin nanoparticles in 0.01N HCI and mixed for 5 min on vortex mixer (Remi®, India). The mixture was ultracentrifuged in order to isolate entrapped insulin from unentraped insulin. The supernatant was removed and nanoparticle sediment was washed by distilled water. Sediment was dissolved in a mixture of 0.01N HCl and ethanol (99%) with solvent ratio of 4:1, respectively. The solution was sonicated and filtered through 0.22 μm syringe disc filter (Ultipor®N66, Pall life Sciences). The filtered solution was quantitatively analyzed for human insulin by RP-HPLC. The entrapment efficiency (%) was determined by human insulin content obtained as percentage of initial amount used in the formulation.

The release rate of human insulin from the nanoparticles were determined by suspending weighed amount of lyophilized nanoparticles (equal to 100 IU of human insulin) in 50 ml, pH 7.4 phosphate buffer USP. During the experiment (3 h) samples were shaken horizontally in a constant temperature shaker (Remi®, India) at 37±1° and 50 stokes per minute. At scheduled time intervals, 1 ml of sample was removed and replaced with 1 ml of fresh release medium. The sample was ultracentrifuged and supernatant was collected and quantitatively analyzed for human insulin content by RP-HPLC.

The protection of insulin nanoparticles was evaluated in α-chymotrypsin with specific activity of 51 U/mg protein at pH 7.8 (optimal pH for enzyme activity) and 37±1°. The nanoparticles were suspended in 0.05 M phosphate buffer pH 7.8 to which α-chymotrypsin had been added. α-Chymotrypsin was also added to insulin solution (as control). The final concentration of the insulin and the enzyme were 15 U/ml and 26 U/ml. Samples were removed after 30 min and 0.5 ml 0.1% TFA was added to inactivate traces of enzyme present. Subsequently, 0.5 ml of ethanol 99% was added to each tube and sonicated for 10 to 15 sec. Then the volume was adjusted to get the human insulin concentration of 0.5 to 0.6 IU/ml with a mixture of 0.01N HCl and ethanol (4:1). The insulin solution was filtered though 0.22 μm syringe disc filter. The filtered solution was quantitatively analyzed for human insulin content by RP-HPLC.

Biological activity of entrapped human insulin in the nanoparticles was assessed using ELISA technique. Insulin loaded nanoparticles (equal to 3 units of human insulin) were dissolved in a mixture of phosphate buffer pH 7.4 and ethanol 99% with solvent ratio of 4:1. The solution was mixed for 5 min and then sonicated for 15 sec and volume adjusted to 5 ml and filtered through 0.22 μm syringe disc filter. Second dilution was made by diluting 50 μl of filtrate to 200 ml with phosphate buffer pH 7.4 to final concentration of about 150 mU/l. An aliquot of sample was withdrawn from the solution and insulin content was analyzed by ELISA as per standard protocol. Results were obtained by reading the optical density at 450 nm using plate reader (BioRad®, model 680 microplate reader) and results were calculated using Microplate Analyst 3.0.2 software.

Male Wistar rats (200-280 g) were housed in a light and temperature controlled environment. All the animal procedures were reviewed and approved by the ethical committee for animal studies. Diabetic state was induced with single injection of 40 mg/kg streptozocin (STZ) in citrate buffer (pH 4.5) via intraperitoneal route. The plasma samples were analyzed for glucose content by glucose oxygenase method (Reagent: Pinnacle Biotechnology Ltd, Mumbai) and measuring the optical density at 505 nm (Start 21 Autoanalyser). Rats were considered diabetic and included in the study when plasma glucose levels were near and above 300 mg/dl. Diabetic rats were divided into three groups, each containing 6 rats. Group-1, placebo nanoparticles (equal to 20 IU/kg of insulin) dispersed in 0.5 ml purified water was adiministered intragastrically to each diabetic rat as a control; Group-2: human insulin injection (Huminsulin-R® 40 IU/ml injection, Elly Lilly) was administered subcutaneously to each diabetic rat at a dose level of 2 IU/kg. This group served as reference standard; Group-3: human insulin nanopartic1es dispersed in 0.5 ml purified water (equal to 20 IU/kg of human insulin) was administered intragastrically to each diabetic rat. After administration of the dosage forms, 0.5 ml of purified water was administered. Prior to and at specified time intervals over 12 h period blood samples were collected from retroorbital vein in capillary tubes for serum human Insulin estimation and fluoride oxalate impregnated capillary tubes for glucose estimation. The serum samples were analyzed for the insulin content by using ELISA technique (Mercodia® Insulin ELISA kit). Optical density at 450 nm was measured within 30 min on microplate reader (BioRad®, Model 680) and human insulin concentration was calculated using Microplate Analysis 3.0.2 software. The plasma samples were mixed with glucose oxygenase reagent and the samples were incubated at 37° for 10 min for colour development. Optical density of the samples was measured at 505 nm. The glucose concentration was auto-calculated using Star 21 glucose autoanalyzer. Human insulin pharmacokinetic parameters such as C_max, T_max, T_½ and AUC_0-t were estimated from the serum human insulin concentration versus time profile. Semilogarithmic plot of the serum human insulin concentration versus time was constructed for estimation of pharmacokinetic parameters such as elimination rate constant, K_e and plasma half life, T_½.

Blood glucose concentrations were determined in triplicate prior to dosing and the mean concentration was considered as 100% level. All following concentration-time data were expressed as a fraction of the base line, considering the fact that blood glucose concentrations over 12 h following intragastric administration of placebo nanopartic1es to diabetic rats (control groups) were not significantly different from the base line (assuming a flat base line). The mean±SD of each concentration-time point in each treatment group was calculated and compared.

The mean of percent blood glucose concentrations at each time point was subtracted from 100% and the area above the percent blood glucose-time curve (AAC_0-12h) were estimated by the trapezoidal rule reported by Touitou and Rubinstein12.