MATERIALS AND METHODS

Materials: Enalapril maleate was kindly gifted by Alkem Pharma, Mumbai, India. Paraffin wax was purchased from Himedia labs. Cetyl alcohol, Tween^® 80, potassium dihydrogen orthophosphate, disodium hydrogen orthophosphate and sodium hydroxide were purchased from S.D. Fine Chemicals Ltd. (India). All other chemicals used were of analytical grade. Preparation of Lipospheres

Preparation of lipospheres: Lipospheres were prepared by a method based on the water-in-oil-in-water double emulsion (w/o/w) method reported by Reithmeier et al. (2001) and Cortesi et al. (2002) with few modifications. EM (10, 20 or 30 mg) was solubilized in the 100 μL internal aqueous phase of a w/o/w double emulsion containing Tween^® 20 (3% w/v) as stabilizer to prevent loss of EM to the external phase during solvent evaporation. This aqueous solution of peptide was emulsified in 100 mg of Paraffin wax and cetyl alcohol dissolved in 1.0 mL of methylene chloride under vigorous vortex-mixing for 10 sec. The obtained primary emulsion was further emulsified into 30 mL of a stabilizer (0.1, 0.15 or 0.2% v/v Tween^® 80, 37°C) containing aqueous phase (Stirring speed- 500, 1000 or 1500 rpm) for 1 min. Hardening of the oily internal phase resulting in encapsulation of the peptide was accomplished by pouring emulsion into 100 mL of ice cold water maintained at 4°C and stirred at 300 rpm. After 3-5 h, lipospheres were isolated by filtration, washed with ice cold water and dried at room temperature (25°C) for 24 h. The final product was stored in dessicator at 2-8°C. The full experimental design and layout with coded and actual values of variables for each batch and responses are shown in Table 1. The trials were performed in random order. The other formulation and processing variables were maintained constant during the process.

Characterization
Particle size: Particle size analysis of EM-loaded lipospheres was performed by optical microscopy using a compound microscope (Labomed, India). A small amount of dry lipospheres was suspended in purified water (10 mL). The slide containing lipospheres was mounted on the stage of the microscope and 300 particles were measured using a calibrated ocular micrometer and photographed at a magnification of x400. The process was repeated for each batch prepared.

Morphology: The surface morphology and shape of the lipospheres were analyzed by scanning electron microscopy for selected batches (Leo, VP-435, Cambridge, UK). Photomicrographs were observed at x303 magnification operated with an acceleration voltage of 15 kV and working distance of 10 mm was maintained.

Table 1:	Box-Behnken design layout with coded levels and actual values of variables
*Value in parenthesis indicates coded levels of the variables

Lipospheres were mounted on the standard specimen-mounting stubs and were coated with a thin layer (20 nm) of gold by a sputter-coater unit (VG Microtech, Uckfield, UK).

Drug content: Twenty milligrams of the dried lipospheres were accurately weighed and added to 5 mL of ethyl acetate. The EM separated in phosphate buffer (pH-7.4) was analyzed by HPLC system (Shimadzu LC-10AT vp, binary gradient) equipped with detector (Shimadzu UV-visible SPD-10A vp), software (Spinchrom CFR V. 2.2, Spincotech Pvt. Ltd., Chennai) and Column (Phenomenex Luna, C-18, 5 μm, 25 cmx4.6 mm i. d.) (Walily et al., 1995). Results were expressed as Mean±SD of 3 experiments. The measured responses are shown in Table 2.

In vitro release: In vitro release of EM from lipospheres was evaluated in both acidic buffer (pH-1.2) and phosphate buffer (pH-7.4). Amount of lipospheres equivalent to 20 mg of EM were transferred to the pre-warmed dissolution media (20 mL) and maintained at 37±0.5°C under stirring at 50 rpm. Samples were withdrawn every h up to 12 h and the volume was replaced immediately by fresh phosphate buffer. The sample withdrawn was centrifuged (3000 rpm, 15 min). The EM concentration in the supernatant solution was analyzed by HPLC system as given in drug content Results were expressed as Mean±SD of 3 experiments.

Experimental design: A Box-Behnken experimental design was employed to statistically optimize the formulation parameters of EM microsphere preparation for maximum entrapment and controlled drug release. The Box-Behnken design was specifically selected since it requires fewer treatment combinations than other design in cases involving three or four factors. The Box-Behnken design is also rotable and contains statistical missing corners which may be useful when the experimenter is trying to avoid combined factor extremes. This property prevents a potential loss of data in those cases. Generation and evaluation of the statistical experimental design was performed with the STAT-EASE, design expert, 7.1.1. A design matrix comprising of 16 experimental runs was constructed.

Table 2:	Responses with actual and predicted values

An interactive second order polynomial model was utilized to evaluate both the response variables:

(1)

where, b₀-b₉ are the regression coefficients, X₁-X₃ the factors studied and Y_i is the measured response associated with each factor level combination. To assess the reliability of the model, a comparison be Tween^® the experimental and predicted values of the responses is also presented in terms of %Bias in Table 2.

Bias was calculated by Eq. 2:

(2)

Storage stability: Stability studies were conducted to find out stable product under storage as per ICH guidelines (Q1AR2) for new drug product and Q5C for stability testing of Biotechnological/ Biological products (CPMP/ICH/138/95) (Grimm, 1998).

RESULTS

Experimental designing: For the response surface methodology involving Box-Behnken design, a total of 16 experiments were performed for three factors at three levels each. Table 1 summarizes the experimental runs, their factor combinations and the levels of experimental units used in the study.

Effect of selected formulation variables: In order to determine the levels of factors which yielded maximum entrapment, mathematical relationships were generated be Tween^® the dependent and independent variables.

For estimation of coefficients in the approximating polynomial function (Eq. 1) applying uncoded values of factor levels, the least square regression method was used. A suitable polynomial equation involving the individual main effects and interaction factors was selected based on the estimation of several statistical parameters such as the multiple correlation coefficient (R²), adjusted multiple correlation coefficient (adjusted R²) and the predicted residual sum of squares (PRESS) provided by the design expert software 7.1.1 (Table 3). The mean diameter (Y₁) and entrapment efficiency (Y₂) of lipospheres from the sixteen experiments were used to generate predictor equations for lipospheres with independent variables as peptide amount (X₁), Tween^® 80 concentration (X₂) and stirring speed (X₃). Limit for these variables were selected from preliminary trials. The results of multiple regression analysis and Analysis of Variance (ANOVA) are summarized in Table 4.

As presented in Table 3, the quadratic model was selected as a suitable statistical model for optimized formulation with maximum entrapment because it had the smallest value of PRESS (91.42 for Y₁ and 395.63 for Y₂). PRESS is a measure of the fit of the model to the points in the design. The smaller the PRESS statistic, the better the model fits to the data points (Segurola et al., 1999). From the p-values presented in Table 3, it can be concluded that for both responses the cross product contribution (2FI) of the model was not significant indicating the absence of interaction effects. Furthermore, Mean Diameter (MD) and the percent drug entrapment of EM lipospheres showed R² values of 0.9598 and 0.9687 (Table 4), respectively; indicating good fit and it was concluded that the second order model adequately approximated the true surface.

For estimation of significance of the model, the Analysis of Variance (ANOVA) was applied. Using 5% significance level, a model is considered significant if the p-value is less than 0.05. The results of multiple regression analysis and Analysis of Variance test (ANOVA) are also summarized in Table 4.

Table 3:	Summary of results of (a) model analysis (b) lack of fit (c) R-square analysis for measured responses

Table 4:	Regression analysis data for measured responses

Table 5:	Standardized main effects of the factors on the responses and associated p-values
*Standardized main effects (SME) were calculated by dividing the main effect by the standard error of the main effect

(3)

The predictor equation generated for the mean diameter was found to be significant with an F-value of 28.84 (p<0.0001) and R² value of 0.9352: The Eq. 3 generated revealed that both factors X₂ and X₃ independently exerted a significant influence on the mean diameter. The influence of the main effects on the particle size of the lipospheres was further elucidated using the response surface plot (Fig. 2).

The model generated for encapsulation efficiency was found to be significant with an F-value of 26.36 (p<0.0001) and R² value of 0.9462:

(4)

The model (Eq. 4) indicated that both X₁ and X₂ factors studied exerted independently a significant influence on the encapsulation efficiency. The 3-D plot (Fig. 3) shows that the 3 entrapment efficiency decreased with increase in drug and Tween^® 80 amount.

In Table 5, factor effects of the Box-Behnken model, associated p-values and Standardized Main Effects (SME) values for both responses are presented. A factor is considered to influence the response if the effects significantly differ from zero and the p-value is less than 0.05. Coefficient signs also give an indication of the effect produced (Table 5).

Particle size and yield: All the lipospheres prepared with in the experimental design yielded smooth spherical structures with size in the range of 23.00±0.82-34.57±1.04 μm (Fig. 1; ELS2). The yields of all trials of lipospheres were upto 85% (most of the formulations had yields of more than 65%) which reflects a good efficiency of the preparation method (Table 2).

EM and Tween^® 80 at low level (X₁, -1; X₂, -1) and stirring speed at medium level (X₃, 0) yielded microspheres with highest drug entrapment (78.93±1.36%) with 23.48 μm mean diameter of lipospheres.

Fig. 1:	Photomicrograph of EM lipospheres

Fig. 2:	3D surface curve for the effect of selected variables (X2, X3) on the mean diameter of Microspheres (X2, -1)

Fig. 3:	3D surface curve for the effect of selected variables (X1, X2) on the entrapment of Microspheres (X1, 0)

Fig. 4:	3D surface curve for the effect of selected variables (X1, X2) with medium level of X3 for the desirable response in terms of maximum entrapment and diameter in optimum range

A positive sign indicates a synergistic effect while a negative sign represents an antagonistic effect of the factor on the selected response. SME values were calculated by dividing the main effects by the standard error of the main effects. The SME values in case of Y₁ response indicated that the peptide (EM) had insignificant effect on size of lipospheres whereas Tween^® 80 concentration (SME = 7.97) and stirring speed (SM E= -3.77) significantly affected the size of lipospheres. In case of entrapment efficiency, factor X₁ (SME = -9.49) and X₂ (SME = -4.60) played major role with insignificant effect of factor X₃ (SME = 0.95) (Table 5). This was further investigated by the study of ANOVA. The breakup of source sum of squares (Source SS) in ANOVA indicated that the contribution of factor X₁ (EM) (SSY₂-384.48) is much higher than factor X₂ (Tween^® 80) (SSY₂-90.05) and X₃ (Stirring speed) (SSY₃-3.98) for optimizing the entrapment efficiency. The contribution of factor X₂ (SSY₁-62.27) was higher on the mean diameter of lipospheres than factor X₁ (SSY₁-0.023) and X₃ (SSY₁-13.86).

Factor X₃ affected liposphere size significantly with X₂ (Tween^® 80) whereas factor X₁ affected entrapment efficiency. Tween^® 80 affected both size and entrapment efficiency of lipospheres. The interaction terms X₁X₂, X₂X₃, X₁X₃ and the polynomial terms X₁X₁, X₂X₂ and X₃X₃ indicated insignificant values of individual source sum of squares. In addition, three dimensional response plots were presented to estimate the effects of the independent variables on each response by keeping one factor at constant level.

Using the model generated with both responses (Eq. 3 and 4), the optimization tool in the experimental design software was used to identify a formulation with a maximum entrapment (Fig. 4). It predicted a maximum entrapment of 80.73 and MD of 24.30 μm with a formulation comprising of 10.53 mg EM concentration, 0.1% v/v Tween^® 80 and 1041.45 rpm stirring speed. To confirm the validity of the model, three batches of lipospheres were prepared using this formulation and entrapment was determined. The actual experimental entrapment obtained was 80.62±2.54%. The predicted response and residual value performed at optical values investigated in this study was found to be 80.73% and -0.11, respectively, validating the model generated in this study.

In vitro release study: In vitro release behavior of optimized lipospheres formulation with more than 60% entrapment was investigated in phosphate buffer (pH 7.4) for duration of 12 h. Figure 5, 6 display the release profile of EM from lipospheres. In the prepared formulation, an initial burst of 20.88±1.24% was observed in the first hour due to the drug located on or near the surface of the lipospheres (Fig. 5, 6). All formulations showed an initial burst from 20.88±1.24% to 27.75±1.14% in one hour with additional 73.34±1.02% to 94.68±3.90% in next 12 h. Thus, the formulation could protect the peptide from gastric degradation and would release its drug load slowly at pH 7.4. Release of EM from lipospheres formulation in phosphate buffer (pH 7.4) was faster than that into acidic buffer (pH 1.2) reflecting differences in extent to which the peptide dissolved in the two fluids. A maximum drug release of 15.32±1.06% was observed for optimized formulation in acidic buffer (pH 1.2) after 4 h whereas in phosphate buffer an initial burst followed by sustained release of EM was observed. On the other hand, more than 85% of EM was rapidly released from these formulations within 12 h in phosphate buffer (pH-7.4) and complete release occurred in about 24 h.

Release models such as first order model, Higuchi model and Ritger-Peppas empirical model were applied to the release data (Dredan et al., 1996). Results revealed that peptide was released from lipospheres by a diffusion controlled mechanism following Higuchi matrix model.

Fig. 5:	In vitro release profiles of EM from Lipospheres (ELS2, ELS3, ELS4, ELS5 and ELS6)

Fig. 6:	In vitro release profiles of EM from Lipospheres (ELS7, ELS8, ELS11, ELS13, ELS15 and ELS16)

The value of coefficient of determination (R²) in First order, Higuchi and Ritger-Peppas equation was found to be >0.9 which indicates the diffusion-controlled release.

In vitro stability: HPLC chromatogram of free drug and drug released from lipospheres showed almost identical peaks and pattern similar to free drug indicating stability and intact nature of drug.

Storage stability: Results of stability studies showed that lipospheres lost around 2-4% of protein content in first month under room temperature and around 6-7% in 6 months. This loss was also marginal in case of accelerated conditions where system lost more than 3% drug in 1 month and around 10-12% in 6 months.

DISCUSSION

Lipid based carriers were selected to eliminate the toxic effects associated with the use of polymers as carriers (Rudra et al., 2011). Melt dispersion technique is commonly used for preparation of lipospheres but w/o/w double emulsion method was considered with the aim to possibly reduce the exposure to high temperature of thermolabile compounds, such as proteins and peptides (Baimark, 2009).

When the lipid solution in methylene chloride was used, the aqueous phase coalesced rapidly, especially when the emulsion was prepared by vortex-mixing. So, stabilizer was used to improve the emulsion stability and the encapsulation efficiency in case of the w/o/w-solvent evaporation method. Tween^® 20 and 80 were used as stabilizers in inner and outer aqueous phase, respectively for liposphere formation and emulsion stabilization.

Cetyl alcohol itself exhibits emulsifying capability further stabilizing the primary emulsion (Kamble et al., 2004). Moreover, it also imparts sphericity with smooth surface and modifies the release of the entrapped drug. As being polar lipid, it improves the entrapment of hydrophilic drugs (Maheshwari et al., 2003). The slight loss of solids could be attributed to the losses occurring during various steps of processing such as sticking of the lipid solution, adsorption on the glass wall during solidification or loss of lipospheres during the washing step etc.

The paraffin wax due to its physical properties and behaviour in the intestinal lumen was used to prepare gastro-resistant SLS formulations using the adopted technique (Shivakumar et al., 2007). Since lipospheres produced with paraffin wax alone resulted in poor drug entrapment and release, efforts were made to enhance drug release from the lipospheres by incorporating a polar wax modifier like cetyl alcohol. Study citations reveal that cetyl alcohol has been successfully employed as a wax modifier to modulate drug release from wax microspheres (Maheshwari et al., 2003). Tween^® 80 was used to stabilize the oil in water emulsion by reducing the interfacial tension be Tween^® the hydrophobic wax dispersion and the external aqueous phase, producing an emulsified oily dispersion which resulted in drug loaded lipospheres on cooling. Fatty alcohols like cetyl alcohol and stearyl alcohol have been reported to improve release and entrapment of hydrophilic peptide due to their polar hydrophilic nature (Nasr et al., 2008).

On the basis of above results, factor X₁ (EM) is found to be the main influential factor on the entrapment and factor X₃ (Stirring speed) on the size of liposphere. Factor X₂ exerted combined effect on both size and entrapment efficiency of lipospheres.

Factor X₁ exerted negative influence on entrapment, also supported by the sign of coefficients in the fitted model (Eq. 4). The significant decrease in entrapment with increase in EM concentration may be because of the increase in viscosity of the inner aqueous phase. The increase in EM may improve coagulation of primary emulsion droplets by the increase in viscosity of inner water phase which will accelerate the leakage of inner to outer water phase leading to increase in size with reduced drug load (Ito et al., 2007). Moreover, this effect might also be due to increase in the ratio of EM: Polymer with insufficiency of polymer to effectively coat the drug.

Tween^® 80 concentration (X₂) exerted positive influence on particle size and negative effect on drug entrapment of lipospheres. As for Tween^® 80, its CMC is ~0.014 mol L^-1. So, possible reason for decreased drug entrapment at both medium and high level might be due to formation of sphere shaped micelle at higher concentration of Tween^® 80 than its Critical Micelle Concentration (CMC), whereby sphere shaped micelles are further transformed into cylinder shaped micelle structure also supported by Zhang and Zhu (2004).

Factor X₃ exerted significant effect on size of lipospheres as compared to entrapment efficiency. With the increase in stirring speed from 500 to 1000 rpm, size decreased but with further increase size increased might be due to increased surface free energy of small particles leading to aggregate formation and clumping. At both low and high level of stirring speed (X₃), lower entrapment was found and maximum entrapment was found at medium level.

The findings of release pattern are in agreement with those of Adeyeye and Price (1994) and Giannola and De Caro (1997) who reported that rapid drug release (such as phenytoin and diclofenac sodium) from fatty acid or alcohol-wax microspheres would be expected due to the hydrophilicity and leaching characteristics. Early studies reported that the drug release from matrix systems was affected by the particle size and drug: Polymer ratio (Kirn et al., 1998). These are matrix systems in which the drug molecules are dispersed throughout the particles.

CONCLUSION

The optimized formulation for enalapril maleate was obtained with EM, Tween^® 80 and stirring speed using response surface methodology based on a Box-Behnken design. It was found that the optimized formulation was achieved with 10.53 mg EM concentration, 0.1% v/v Tween^® 80 and 1041.45 rpm stirring speed. The observed responses were close to the predicted values for the optimized formulation. Microencapsulation doesn’t affect the integrity of entrapped drug as determined by HPLC chromatograms. In conclusion, controlled release biocompatible polar lipid based oral delivery system for hydrophilic peptide was successfully developed. Further parameters can be identified by systemic approach for optimum formulation in terms of better long-term stability and to study the therapeutic effects of these particles in vivo.

ACKNOWLEDGMENTS

The authors are thankful to M/s Alkeon Pharma, Mumbai, India, for the gift sample of Enalapril maleate; SIF, AIIMS, New Delhi, India, for Scanning electron micrography and Director, University Institute of Pharmacy, Pt. Ravishankar Shukla University, Raipur (C.G.) India for providing all necessary facilities for carrying out this work and Chhattisgarh council of science and technology for financial assistance (CGCOST/MRP/650).