Niosomes with Sorbitan Monoester as a Carrier for Vaginal Delivery of Insulin: Studies in Rats

Abstract

To prepare and investigate the potential of the niosomes vaginal delivery system for systemic treatment of insulin is the goal of this study. Two kinds of vesicles with Span 40 and Span 60 were prepared by lipid phase evaporation methods with sonication. The niosomal entrapment efficiency was determined by column chromatography. The particle size and morphology of the vesicles also were evaluated. The results showed optimized niosomes prepared in this study had niosomal entrapment efficiency 26.68 ± 1.41% for Span 40 and 28.82 ± 1.35% for Span 60, respectively. The particle sizes of Span 40 niosomes and Span 60 niosomes were 242.5 ± 20.5 nm and 259.7 ± 33.8 nm, respectively. There were no significant differences in appearance between the two types of vesicles. The hypoglycemic effects, and insulin concentrations after vaginal administration of insulin vesicles to rats were investigated. Compared with subcutaneous administration of insulin solution, the relative pharmacological bioavailability and the relative bioavailability of vaginal administration of insulin vesicles were determined. Compared with subcutaneous administration of insulin solution, the relative pharmacological bioavailability and the relative bioavailability of insulin-Span 60 vesicles group were 8.43% and 9.61%, and insulin-Span 40 niosomes were 9.11% and 10.03% (p > 0.05). Span 60 and Span 40 niosomes were both higher than blank Span 40, Span 60 vesicles, and free insulin physical mixture groups (p < 0.05). The results indicates insulin-Span 60, Span 40 niosomes had an enhancing effect on vaginal delivery of insulin. Although the factors controlling the process for penetration of a portion of vaginally administrated niosomes into bloodstream from vaginal tract is still not fully understood, our results demonstrated that after encapsulation in niosomes of definite type, insulin became an active and efficiently therapeutic agent when administrated vaginally and might be a good carrier for vaginal delivery of protein drugs.

Keywords :

• Insulin
• Niosomes
• Vaginal Delivery

Many of the novel biogenetically engineered peptide and protein drugs are administered parenterally because of their low bioavailability after oral dosing. This has led to an intensive investigation of alternative means of administration including the intranasal, transdermal, buccal, ocular, rectal, and vaginal routes (Owens et al. 2003). Among these alternative routes, vaginal is a potential route for the systemic administration due to its large surface area, rich blood supply, and permeability to a wide range of compounds including peptide and protein, and especially for the treatment of female-related conditions (Richardson et al. 1992a). But due to their poor absorption across the vaginal epithelium and degradation by enzymes in the vaginal lumen, peptides and proteins have low bioavailability after vaginal administration. In general, the bioavailability of vaginally administered peptides may be markedly improved by the use of absorption enhancers (Richardson 1992b) and a polyacrylic acid aqueous gel (Morimoto et al. 1982). In addition, to protect insulin from enzymatic degradation in the absorption site, encapsulating drug in microspheres (Jerry et al. 2001; Richardson et al. 1995; Morishita et al. 1992), liposomes (Kisel et al. 2001) and niosomes (Chattaraj et al. 2003; Varshosaz et al. 2003; Khaksa et al. 2000; Rentel et al. 1999; Yoshida et al. 1992), has been tried.

In recent years, niosomes (nonionic surfactant vesicles) have received much attention for their advantages in biodegradability, absorption enhancement, toxicity reduction, and improved therapeutic effectiveness. Most of all, their chemical characterizations are more stable than liposomes (Uchegbu et al. 1998). Thus, niosomes are used widely as a drug carrier, and especially as peptide and protein drug carriers (Chattaraj et al. 2003; Varshosaz et al. 2003; Rentel et al. 1999; Yoshida et al. 1992; Khaksa et al. 2000).

In our present study, we chose sorbitan monoester (Span 40, Span 60; HLB 6.7, 4.7) to form the niosomes and insulin as a hydrophilic peptide model drug to investigate a potential route for vaginal administration. To optimize the preparation of niosomes in regard to entrapment efficiency, niosomes-containing drug were prepared by lipid hydration method with sonication. The morphology of niosomes was probed by transmission electron microscopy. In addition, the characterization of insulin-loaded niosomes was studied using SDS-polyacrylamide gel electrophoresis. Further, niosomes with insulin were to be examined in vitro release properties and the pharmacological availability of vaginal administration.

MATERIALS AND METHODS

Crystalline porcine insulin (27.4 IU/mg) was purchased from Xuzhou Wanbang Biochemical Company (P.R. China). Sorbitan monoesters (Span 40, HLB = 6.7; Span 60, 4.7), dicetylphosphate (DCP), and cholesterol (CH) were purchased from Sigma Company. Tricine came from Beijing Jinke Reagent Company, blood glucose assay kits were obtained from Zhongsheng High-Tech Bioengineering Company (P.R. China); and insulin radioimmunoassay kits came from the China Institute of Atomic Energy (P.R.China). Protein molecular weight marker (was purchased from Amersham Biosciences, M.W. range: 2512-16949) and cellulose nitrate membrane filters (0.22 μm) from Whatman, Majdstone, UK; Sephadex G-75 (medium) was purchased from Pharmacia (Sweden). Buffer PBS (pH 7.4) was made of 8 g NaCl, 0.2 g KCl, 0.025 g Na₂HPO₄·2H₂O and 0.050 g NaH₂PO₄· 2H₂O per 1L; acetate buffer (pH 4.5) was made of 0.2 M CH₃COONa (430 ml) and 0.2 M CH₃COOH (570 ml). Acetonitrile was the product of Fisher (chromatography grade, USA). All other reagents used in the study were of analytical grade and commercially available.

Preparation of Niosomes

Conventional niosomes were prepared by lipid evaporation method with sonication. The formulations containing nonionic surfactant and cholesterol (300 μmol of total lipid in our experiment) according to the previous study (Varshosaz et al. 2003) resolved in chloroform; the desired volumes were added to a 50 ml round-bottom flask. The flask was attached to a rotary evaporator (Büchi Rotavapor RE 120), lowered into a 50°C water bath, and the organic solvents were evaporated under reduced pressure to form a thin, dry film on the wall of the flask. Any excess organic solvents were removed by leaving the flask overnight in a desiccator under vacuum. The dried lipid film was hydrated when required with 5 ml PBS (pH 7.4) containing insulin 20 IU/ml followed by vigorous shaking in an incubator at 55°C for ∼1 hr to yield multilamellar niosomes. Then, these multilamellar niosomes were sonicated 3 times for 1 min with 2-min intervals under a stream of argon at 50°C using titanium microitp in a UH sonicator (22 kHz, 20 mA, Automatic Science Instrument Co., Tianjin, China). The resultant niosomal dispersion was then left to cool slowly. To study the effect of composition of the vesicles, a series of formulations with different molar ratios were designed (Table 1).

TABLE 1 Effect of lipid composition on encapsulation efficiency of niosomes

Span 40% Mol	Span 60% Mol	Cholesterol (% Mol)	Entrapment efficiency (% ± SD)
70	30	—	17.12 ± 2.27
60	40	—	21.42 ± 1.17
50	50	—	28.82 ± 1.35
40	60	—	11.00 ± 0.82
70	—	30	17.34 ± 1.02
60	—	40	18.35 ± 1.22
50	—	50	26.68 ± 1.41
40	—	60	17.04 ± 0.96

A constant amount of DCP (10% molar ratio) was added to the formulations. SD = standard deviation.

Entrapment Efficiency of Insulin in the Niosomes

The entrapment niosomes were separeated by Sephadex G-75 (1.6 × 70 cm); dilute rate 1.0 ml/min was controlled by pump (LKB Bromma 2070 ultrorac® 11 90 drop; 2138 uvicord s 280 nm ABS-range 0.05; 2210 2-channel recorder 0.02 mm/min; 2232 microperpex S peristaltic pump 30 incp). Then, 2.0-ml aliquot of insulin-containing noisome preparation was loaded onto a Sephadex G-75 column and eluted using PBS, pH 7.4, ambient temperature. Noisome and free insulin were separated and collected, respectively, and free insulin were determined by HPLC analysis (a stability-indicating HPLC assay was used to analyze insulin. The method described by Chinese Pharmacopiea was used. The HPLC system consisted of gold nouveau software workstation, a Beckman 126 NM solvent delivery system, Beckman 508 autosampler with a 100-μ L loop, and Beckman 168 NM PDA detector. The column used was Beckman C18 dp 5 μ m, 4.6 mm × 25 cm (Beckman, USA). The mobile phase consisted of 0.13 mol NaH₂PO₄, 0.1 Na₂SO₄, acetonitrile (pH 3.0) (36.5:36.5:27, v/v/v). The buffer was filtered through 0.22 μm membrane filter (Millipore Corp., Milford, MA, USA) prior to use. The flow rate was 1.5 ml/min. The chromatogram was monitored at a wavelength of 220 nm. Entrapment efficiency (EE%) could be achieved by the following equation:

Transmission Electron Microscopy (TEM)

TEM was a method of probing the microstructure of rather delicate systems such as micelles, liquid crystalline phases, vesicles, emulsions, and also nanoparticles. In this experiment, a drop of the resultant niosomal emulsion was placed on a carbon-coated copper grid, leaving a thin liquid film. The film on the grid was then negatively stained by adding immediately a drop of 1.5% (w/w) phosphotungstic acid (PTA) solution, removing the excess staining solution with a filter paper, and following by a thorough air-drying. The stained film was then viewed on a Hitachi-500 model transmission electron microscope (Hitachi, Japan).

Particle Size of Niosomes

The size distribution of the resultant dispersion was characterized using a laser particle size analyzer based on laser diffraction using Nicomp 370 submicron particle sizer (Nicomp 370 submicron particle sizer, Santa Barbarra, CA, USA). The dispersions were diluted to an appropriate concentration so that the intensity of the transmitted laser beam was within the limits required by the instrument for measurement. The samples were stirred using a magnetic stirrer bead to keep and maintain the sample in suspension. Each formulation was measured three times in triplicate.

Characterization of Niosomes Containing Insulin Preparation

To characterize the properties of insulin-loaded niosomes, we chose the Span 60 niosomes. Protein samples (including protein molecular weight marker; insulin PBS solution; before separation niosomes solution; free insulin separated from column; niosomes solution after disrupted by 1% Triton X-100; insulin-niosomes; blank noisomes) were incubated with SDS sample buffer and were run on a Tris-Tricine system using a vertical electrophoresis apparatus at 150 V. The gel was stained using Komarse light blue® staining kit, and BenchMark™ protein ladder was used as the molecular weight standard.

Physical Stability and in Vitro Release Studies in Simulated Vaginal Fluid

The optimized Span 40 and Span 60 formulations (Surfactant:chol molar ratio was 50:50) were stored in glass vials after purging with nitrogen and kept in a refrigerator (4 ± 1°C), room temperature (R.T. 25 ± 2°C), and 37 ± 1°C for 3 months. The samples from niosomes were withdrawn at definite time intervals; the residual amount of drug in vesicles was determined as described above for the method of entrapment efficiency of insulin in the niosomes. The drug percent retained in vesicles results are shown in Table 2. Further, to investigate the formulations stability or release properties in simulated physiological condition, the drugqtoa release from niosomes was tested in acetate buffer (pH 4.5) at 37 ± 1°C. The drug release results are shown in Figure 1. The release of insulin from Span 60 vesicles in simulated vaginal fluid is slower than from Span 40 vesicles, but had no significant difference (p > 0.05). As can be seen from Figure 1, after 24 h the leakage was approximately 30% of original entrapped insulin, which suggested niosomal delivery system might provide controlled and prolonged release of an adequated insulin in vaginal local treatment.

FIG. 1 Release profile of insulin from niosomes at acetate buffer (pH 4.5) at 37 ± 1°C (n = 3).

TABLE 2 Stability of niosomes (surfactant:cholesterol molar ratio is 50:50) containing insulin expressed as percentage of originally entrapped drug still present in niosomes after three storage conditions at (n = 3, ± SD)

	Surfactant	1 month	2 month	3 month
4 ± 1°C	Span 40	97.12 ± 2.91	96.52 ± 1.27	96.77 ± 2.12
	Span 60	98.75 ± 1.24	97.87 ± 3.12	97.47 ± 2.16
Room temperature
(25 ± 2°C)	Span 40	87.73 ± 1.24	83.27 ± 3.45	65.18 ± 2.72
	Span 60	90.81 ± 1.79	84.83 ± 3.19	67.24 ± 2.20
37 ± 1°C	Span 40	76.78 ± 1.24	53.88 ± 3.46	35.27 ± 2.78
	Span 60	80.72 ± 1.52	63.83 ± 3.72	37.27 ± 1.18

Animals and Model of Experimental Diabetes

A major problem in studies of vaginal absorption in rodent species is the variable structure of the vaginal epithelium at different stages of the estrous cycle. To standardize the thickness of the vaginal epithelium, the animals used in the present study were ovariectomized. Female ovariectomized Wistar rats (body weight, 180–250 g) were used in these experiments. Diabetes was induced by a single intraperitoneal injection of alloxan (150 mg/kg). The rats were used within 2 weeks after alloxan injection when the acute disease phase was completed. In addition, 24 hr prior to the experiments, 100 μl of estradiol solution (40 μg/kg) was administered by subcutaneous injection.

Therapeutic Experiments

Female rats were randomly divided into 7 groups (4 in each group): group 1, vaginal administration of insulin-Span 40 niosomes (10 IU/kg); group 2, vaginal administration of insulin-Span 60 niosomes (10 IU/kg); group 3, vaginal administration of a physical mixture that contained blank Span 40 niosomes and free insulin (10 IU/kg); group 3, vaginal administration of a physical mixture that contained blank Span 60 niosomes and free insulin (10 IU/kg); group 5, vaginal administration of insulin solution (10 IU/kg); group 6, vaginal administration of PBS (pH 7.4); and group 7, subcutaneous administration of insulin solution (1 IU/kg, pH 7.4).

The rats were fasted for 24 h prior to the experiments. Groups of 7 ovariectomized and diabetic rats were anesthetised by intraperitoneal injection of 60 mg/kg pentobarbitone sodium (Beijing Chemical Reagents Company, Beijing) and secured on their backs on boards. The rats were tracheotomized and the caotid artery cannulated. A short length of polythene tubing was attached to a 100 μl syringe and filled with niosomal solution (Richardson et al. 1992, 1989). The tubing was then gently inserted into the vagina and secured in position with a cyanoacrylate adhesive (502®, China) to prevent leakage of insulin. The niosomes-entrapped insulin solution was instilled into the vaginal tract. The dose of insulin was fixed at 10 IU/kg body weight. For control experiments, PBS solution, or free insulin solution, or physical mixture of empty niosomes with free insulin were similarly administered. Insulin solution (1 IU/kg) was subcutaneously administered to group 7 to determine bioavailability. Blood samples were collected from the tail at predetermined time intervals (20, 30, 40, 50, 60, 90, 120, 150, 180, 240, 300, and 360 min) and plasma was separated. The concentration of insulin was determined by radioimmunoassays using a commerical RIA kit according to the protocol provided. Plasma glucose concentrations were measured by a glucose oxidase method.

Relative Pharmacological Bioavailability and Relative Bioavailability Calculations

Relative pharmacological bioavailability calculation: The area above the curves (AAC) was calculated by the trapezoidal method from the blood glucose data. The relative pharmacological bioavailability (Fp) was calculated using the following equation: where the subscripts va and sc refer to vaginal and subcutaneous administration, respectively.

Relative bioavailability calculation: The area under the curve (AUC) was calculated by the conventional trapezoidal method from the plasma insulin level. The relative bioavailability (Fr) was calculated using the following equation: where the subscripts va and sc refer to vaginal and subcutaneous administration, respectively. Student's t-test was used to determine statistical significance.

RESULTS AND DISCUSSION

Entrapment Efficiency and Size of Insulin in the Niosomes

In our present study, chromatography was applied in entrapment efficiency determination due to its simplicity and quick advantages. Column recovery of nonentrapped drug was 97.2%. The elution volume of blank vesicles was 21–30 ml and that of insulin was 36–45 ml. Niosomes and free insulin were separated at 30 ml.

Process variables such as vacuum, hydration medium, hydration time, speed of rotation of flask, and agitational method of size reduction, were optimized to prepare lipid vesicles of insulin. The rotational speed of the flask demonstrated discernible influence on the thickness and uniformity of the lipid film. The speed of 150 rpm yielded a uniform thin, lipid film with vesicular preparation of desired characteristics on hydration. The hydrating temperature used to make niosomes was above the gel-to-liquid phase transition temperature of the system. Various niosomes formulations and the entrapment efficiency of insulin by niosomes are shown in Table 1. The studies of various niosomes consisting of different ratios of Span 40, Span 60, and chol describe the effect of these variables on the degree of entrapment. As can be seen from Table 1, increasing chol level increases the entrapment of drug from 30% to 50%, but with a further increase in the level of chol the entrapment was reduced. This indicates that the chol level beyond a certain level starts disrupting the bilayer structure leading to loss of drug entrapment level (Agarwal et al. 2001), which is known to give the vesicular membrane a more or less ordered structure when above or below transition temperature (Tm), respectively (Manconi et al. 2002). The optimized niosomes prepared in the study had entrapment efficiency 26.68 ± 1.41% for Span 40 and 28.82 ± 1.35% for Span 60, respectively.

In addition, the ionic surfactants in the formulation generally is used to stabilize niosomes by means of an increase of their zeta potential and optimized ion-dipole interaction. Decetylphosphate (DCP) is a charge inducer and was added to all the formulations to increase the vesicles stability in our present study.

The particle sizes of Span 40 niosomes and Span 60 niosomes were 242.5 ± 20.5 nm and 259.7 ± 33.8 nm, respectively. Sorbitan monostearate (Span 60, C₁₈) vesicles were larger than sorbitan monopalmitate (Span 40, C₁₆) niosomes, although the difference was very low. These results are contrary to those found in the literature for unsonicated carboxyfluorescein sorbitan monoester vesicles, where it seems that vesicle size is directly proportional to the surfactant monomer hydrophilicity (Yoshika et al. 1994).

TEM Investigation

The morphology of the niosomes was evaluated. Figure 2 shows the shape of the niosomes entraping with the drug insulin. It was evident that the particles investigated revealed round and homogeneous shading. There were no significant differences in appearance between the two types of vesicles.

FIG. 2 Transmission electron micrographs of niosome particles made of Span 40 (1a), Span 60 (1b)/cholesterol (molar ratio 50:50), the bar in the micrographs represents 1 μ m.

SDS-PAGE Characterization of Niosomes

Figure 3 presents the different samples lanes about Span 60 niosomes in SDS-PAGE. Lane 1 showes the migration patterns for the standard protein; lane 2 is insulin PBS solution; lane 3 is before separation niosomes solution; lane 4 is free insulin separated from column; lane 5 is niosomes solution after disrupted by 1% Triton X-100; lane 6 is insulin-niosomes; lane 7 is blank noisome. SDS-PAGE analysis revealed identical bands for insulin molecular in solution. The insulin PBS solution, niosomes, and entrapped insulin mixture, free insulin separated niosomes and disrupted niosomes showed the band at the same position that indicated the niosomes had encapsulated the insulin. However, the blank niosomes, and insulin-niosomes showed no band, which demonstrated that niosomally encapsulated insulin might be enclosed in the aqueous phase of niosomes and intercalated within the bilayer structure of vesicles.

FIG. 3 Polyacryamide gel electrophoresis of insulin loaded Span 60 niosomes: Lane 1-protein molecular weight marker; lane 2-insulin PBS solution; lane 3-before separation niosomes solution; lane 4-free insulin separated from column; lane 5-niosomes solution after disrupted by 1% Triton X-100; lane 6-insulin-niosomes; lane 7-blank noisome. After electrophoresis, the gel was stained with Coomassie brilliant blue.

Physical Stability at Storage Condition and Insulin Release in Simulated Physiological Fluid

Table 2 shows niosomes were relative stable at 4 ± 1°C storage condition. The drugqtoa leakage percent amounts of original entrapped in niosomes was very small and the drug in vesicles had no significant difference after 3 months compared with that immediately after preparation. In addition, the results of drug retention studies showed higher drug leakage at higher temperature. This might be due to the higher fluidity of lipid bilayers at higher temperature resulting into higher drug leakage.

We know that the pH value of the healthy human vagina ranges between 4.0–5.0, therefor acetic buffer (pH 4.5) was chosen to simulate normal vaginal fluid. As can be seen Figure 1 after 24 hr the drug release was ∼30% of original entrapped insulin, which suggested niosomes delivery system might provide controlled and prolonged release of an adequate drug in vaginal local treatment. The results shown in Figure 1 were that Span 40 niosomes containing insulin were released faster than Span 60 niosomes, especially after 8 hr. This is because the molecules of Span 60 in bilayer structures are in the more ordered gel state at 37°C which is responsible for their less fluidity and permeability (Yoshika et al. 1994).

Therapeutical Effect of Niosomes-Entrapped Insulin

Blood glucose levels were measured using blood glucose assay kits with the glucose oxidase method, and plasma insulin concentrations were assayed using radioimmunoassay kits. These two methods are widely used in the studies on insulin delivery (Khaksa et al. 2000; Kisel et al. 2001; Jerry et al. 2001; Morimoto et al. 1982). The precision and accuracy validations were carried out before the pharmacodynamic and pharmacokinetic studies and these methods proved feasible in our experiments.

Animal group 6, as a blank control, was administered PBS solution, and there were no changes in blood glucose levels and plasma insulin concentration, which still fluctuated around the initial values. This confirmed that diabetic rats had no significant response to blood letting procedure and blood loss.

Figure 4 shows the hypoglycemic effects after vaginal or subcutaneous administration of insulin formulations (7 groups). The maximal hypoglycemic effects with insulin-Span 40, Span 60 niosomes were seen at about 1.5 hr, and the maximal decrease in blood glucose reached 47.49% and 46.66%, respectively. Then 6 hr after vaginal administration, the blood glucose levels of both types of niosomes were still lower than the value at 0 hr. Thus, the niosomes had long-term hypoglycemic effects that may be due to the formation of insulin depot within the vaginal membrane of the rat. Since the insulin is in the aqueous compartment and encapsulated within the vesicles, it is protected from the external environment (Varshosaz et al. 2003).

FIG. 4 Plasma glucose levels after subcutaneous and vaginal insulin administration to rats (n = 4).

Table 3 summarizes the relative pharmacological bioavailabilities in each group. Compared with subcutaneous administration of insulin solution, the relative pharmacological bioavailabilities of Span 40 and Span 60 niosomes were 9.11% and 8.43%, respectively, which were both greater than those of the physical mixture of blank vesicles and free insulin solution (p < 0.05), which indicated that insulin entrapment in niosomes played an important role in the hypoglycemic effect. However, the Span 40 and Span 60 niosomes had no significant difference (p > 0.05). The Fp values in the physical mixture group of Span 40 and Span 60 (5.01%, 4.74%) also were greater than in the control group 5 (vaginal administration of insulin solution) (p < 0.05).

TABLE 3 Relative pharmacological bioavailability and relative bioavailability of insulin vaginal administration to rats (n = 4, ± SD)

	AAC(glucose concentration)	Fp%	AUC(insulin concentration)	Fr%
Subcutaneous control (1 IU/kg)	10577.74 ± 776.23		15870.91 ± 783.56
Insulin solution control (diabetic, 10 IU/kg)	3379.05 ± 453.12	3.19	7019.65 ± 556.35	4.42
Blank Span 40 niosomes and free insulin physical mixture (diabetic, 10 IU/kg)	5294.41 ± 345.89	5.01^a	10891.34 ± 896.35	6.86^a
Blank Span 60 niosomes and free insulin physical mixture (diabetic, 10 IU/kg)	5010.45 ± 421.35	4.74^a	10558.75 ± 889.23	6.65^a
Span 40 niosomes (diabetic, 10 IU/kg)	9638.72 ± 784.56	9.11^a,^b	15911.62 ± 1089.12	10.03 ^a,^b
Span 60 niosomes (diabetic, 10 IU/kg)	8916.35 ± 726.23	8.43^a,^c	15247.28 ± 1120.47	9.61^a,^c

AAC = the area above the curves was calculated by the trapezoidal method from the blood glucose data. AUC = the area under the curve was calculated by the conventional trapezoidal method from the plasma insulin level. Fp (the relative pharmacological bioavailability) = (AAC_va/AAC_sc)/(dose_va/dose_sc). Fr (the relative bioavailability) = (AUC_va/AUC_sc)/(dose_va/dose_sc).

^a p < 0.05, compared with insulin solution control.

^b p < 0.05, compared with blank Span 40 niosomes and free insulin physical mixture.

^c p < 0.05, compared with blank Span 60 niosomes and free insulin physical mixture.

Figure 5 shows the plasma insulin concentration after vaginal delivery of insulin in various formulations, which also confirmed the long-term effects of the insulin-niosomes. As can be seen from Table 3, the relative bioavailabilities of niosomes were 10.03% and 9.61% respectively, which also were greater than those of the blank niosomes and free insulin physical mixture (p < 0.05), compared with subcutaneous administration of insulin solution. The Fp values in the insulin niosomes group and the physical mixture group (6.86%, 6.65%) were all greater than the control group that received vaginal administration of insulin solution (p < 0.05). These results were in agreement with those from hypoglycemic experiments. Insulin entrapment in the niosomes was essential to promote much more insulin penetration into the vaginal membrane. In addition, vaginal administration to the diabetic rats of the same dose of free insulin with empty niosomes produced changes either in blood glucose levels or in insulin levels, which was higher than insulin solution.

FIG. 5 Plasma insulin concentrations after subcutaneous and vaginal insulin administration to rats (n = 4).

Generally, Span 40 and Span 60 are widely considered as absorption enhancers (Fang 2001). A number of possible mechanisms of absorption enhancement have been proposed: epithelial permeability may be increased by the induction of disorder in the phospholipid domain or by the leaching of membrane proteins; the viscosity of the mucus layer may be reduced; tight junctions may be opened by an effect on intercellular calcium ions; and proteolytic enzymes may be inhibited. In addition, several absorption enhancers cause cell loss or damage, which could contribute to a reduction in the epithelial barrier and promote peptide uptake. The document suggested that hydrophilic drugs such as insulin may be absorbed by means of intercellular channels (Richardson 1992; Corbo 1990). Richardson (1992) had investigated the vaginal absorption of insulin in two rat models and in sheep. After vaginal administration of insulin (10 IU/kg) to ovariectomized rats, little change in blood glucose levels was seen. However, co-administration of two surface-active absorption enhancers markedly improved absorption with blood glucose levels falling to 40–60% of initial levels 40–60 min after dosing. In a previous study (Touitou et al. 1978), insulin had been administered to rats in a hydrophilic vehicle containing a surfactant by vaginal and rectal routes. After vaginal administration of insulin in a cetomacrogol 1000 vehicle to 5 diabetic rats, blood glucose concentrations fell to 66%, 51%, and 49% of the initial levels after 1, 2 and 4 hr, respectively.

Span 40 and Span 60 niosomes influence the drug distribution in the body tissue and the drug bioavailability. A previous paper reported vincristine Span 40 niosomes increased the vincristine antitumour activity in S-180 sarcoma and Erlich ascites-bearing mice (Parthasarathi et al. 1994). Span 60 bleomycin niosomes also increased the tumoricidal activity of bleomycin in these two tumour models (Naresh et al. 1996). In a study involving flurbiprofen, Span 60 niosomes, there was increased bioavailability and an increased reduction of carageenan-induced rat paw edema (Reddy et al. 1993).

In the present study, vaginal administration to diabetic rats of insulin entrapped in niosomes produced a higher insulin concentration and Fr and Fp than the blank niosomes and insulin physical mixture. We suggest that factors other than the permeation enhancer effect of surfactants in the vesicles are involved in the enhancement of insulin permeation across vaginal tissues. SDS-PAGE demonstrated that niosomally encapsulated insulin was enclosed in or intercalated within the bilayers of the niosomes. Generally speaking, the vesicles might induce the enhancing effect by, first; adsorption and fusion of drug-loaded vesicles onto the surface of the site leading to a high thermodynamic activity gradient of the drug-mucosa interface; second, the effect of vesicles on mucosa may cause changes in drug permeation kinetics due to an impaired barrier function of the mucosa for the drug (Schreier et al. 1994; Touitou et al. 1994). But the factors controlling the process for penetration of a portion of vaginally administrated niosomes into bloodstream from vaginal tract are still not fully understood.

CONCLUSION

To prepare and investigate the niosomes vaginal delivery system for systemic treatment of insulin were the goals of this study. Two kinds of vesicles with Span 40 and Span 60 were prepared by lipid phase evaporation methods with sonication. The niosomal entrapment efficiency was determined by column chromatography. The results showed that the entrapment efficiencies of the optimized Span 40 niosomes and Span 60 niosomes vesicles were 26.68 ± 1.41% (n = 3) and 28.82 ± 1.35% (n = 3), respectively. The particle sizes of Span 40 niosomes and Span 60 niosomes were 242.5 ± 20.5 nm and 259.7 ± 33.8 nm, respectively. The particle size and morphology of the vesicles also were evaluated. There were no significant differences in appearance between the two types of vesicle. The hypoglycemic effects, insulin concentrations after vaginal administration of insulin vesicles to rats were investigated. Compared with subcutaneous administration of insulin solution, the relative pharmacological bioavailability and the relative bioavailability of vaginal administration of insulin vesicles were determined. Compared with subcutaneous administration of insulin solution, there had no significant difference of the relative pharmacological bioavailability and the relative bioavailability between insulin-Span 60 vesicles group (8.43% and 9.61%) and insulin-Span 40 insulin niosomes (9.11%, 10.03%) (p > 0.05). However, those of Span 60 and Span 40 niosomes were both higher than blank Span 40, Span 60 vesicles and insulin physical mixture groups, and higher free insulin solution (p < 0.05). The results indicated insulin-Span 60, Span 40 niosomes and blank vesicles and free drug physical mixture both had an enhancing effect on vaginal delivery of insulin. Although the factors controlling the process for penetration of a portion of vaginally administrated niosomes into bloodstream from vaginal tract are still not fully understood, the results obtained demonstrated that after encapsulation in niosomes of definite type, insulin became an active and efficiently therapeutic agent when administrated vaginally and might be a good carrier for vaginal delivery of protein drugs.

This work was supported by National “973” planning (G 1999 064).