Optimization of process parameters by response surface methodology to develop a more bioefficacious nanosuspension of Silybum marianum seed extract

Purpose: To develop a nanosuspension drug delivery system to enhance the dissolution rate of Silybum marianum seeds extract. Methods: Central composite design was used to study the effect of the input variables (stabilizer to plant extract ratio, antisolvent to solvent ratio, stirring time) on the dependent variables (mean particle size, polydispersity index (PDI) and zeta potential). The optimized formulation was characterized by Scanning Electron Microscopy (SEM), Atomic Force Microscopy (AFM), Fourier Transformed Infrared Microscopy (FT-IR) and in vitro dissolution testing. Results: The optimized nanosuspension with mean particle size of 137 nm, PDI of 0.327 and zeta potential of -37 mV was obtained. SEM studies revealed irregular shaped particles. AFM studies showed nanosized particles with good surface characteristics. The optimized formulation showed faster dissolution rate than coarse suspension. Conclusion: Results suggested that nanosuspension has remarkable potential for enhancement of the dissolution properties of poorly soluble S. marianum seed extract. Optimization, Dissolution


INTRODUCTION
Silybum marianum, commonly called milk thistle, is a plant of the Asteraceae family widely used as a therapeutic agent for a variety of diseases including liver diseases, colon and prostate cancer [1]. The purified mixture of flavonoids extracted from the seeds and flowers of Silybum marianum called silymarin. Silybin is the most active component of this mixture (comprised 60 -70 % of the total content of silymarin) is responsible for most of its pharmacological activities [2].
Silybin is an important hepatoprotective flavonoid frequently used for the treatment of various hepatic disorders [3]. However, the clinical applications of silymarin are hindered due to its poor water solubility and low bioavailability [4]. The slightly soluble drugs presented limited dissolution rate in gastrointestinal fluid. Therefore, only a small concentration of the drug becomes available at the target site and in situations where in vivo treatment fails. Thus, in vitro dissolution becomes an important factor in the development of new drugs and the enhancement of the dissolution velocity of poorly water-soluble drugs, to increase their bioavailability which is a challenging task for pharmaceutical preparation workers [5].
Nanosuspension is a carrier free drug delivery system, consisting of nanosized drug particles stabilized by polymers or surfactants [6]. nanosuspension is an ideal pharmaceutical formulation for increasing the solubility and dissolution of poorly water-soluble drugs and phytoconstituents [7]. The reduced particle size offers high surface area which dramatically increased the saturation solubility, dissolution velocity and ultimately translates into improved bioavailability of drug candidates [8]. Other potential advantages of nanosuspensions are high drug loading [9], rapid onset of action, dose reduction, reduced side effects, decreased fasted variability, better stability [10] and enhanced absorption [11]. Due to these unique advantages, nanosuspensions have become a choice of drug delivery system for the pharmaceutical industries and many nanosuspension formulations are now marketed [12]. In addition, various administration routes such as oral, parenteral, nasal, intravenous and pulmonary have been reported for nanosuspensions [13].

EXPERIMENTAL Preparation of plant extract
Seeds of S. marianum were collected from the Natural Product Lab, Department of Chemistry, University of Agriculture, Faisalabad and identified by a plant taxonomist from the Department of Botany, University of Agriculture, Faisalabad. Seeds of S. marianum were first defatted with n-hexane to remove fatty substances, and the defatted plant material was extracted with ethanol. The ethanolic extract was concentrated using rotary evaporator.

Formulation of nanosuspension
Nanosuspension of S. marianum extract was prepared by antisolvent precipitation method. First, plant extract (200mg) was completely dissolved in ethanol which was termed as organic phase. This organic solution was slowly injected into aqueous phase containing sodium lauryl sulphate (SLS) as stabilizer with continuous stirring at mechanical stirrer. The mixture was continuously stirred for six hours at room temperature. The nanosuspension was filtered and stored in transparent plastic bottles [14].

Optimization of process parameters by response surface methodology
For optimization of the parameters of nanosuspensions, Central Composite Design (CCD) was used. Stabilizer to plant extract ratio (S/PE ration, A), antisolvent to solvent ratio (AS/S ration, B) and stirring time (C) were the independent variables, and mean particle size (Y1), PDI (Y2) and zeta potential (Y3) were selected as response variables in this study.

Evaluation of nanosuspension stability
The physical stability of optimized nanosuspension was evaluated at two different temperatures. For this purpose, optimized nanosuspension was stored at 25 ºC (at room temperature) and 4 ºC (in refrigerator) for the period of three months. After three months aliquots of samples were drawn and analyzed for particle size, PDI and zeta potential.

Lyophilization of nanosuspension
Optimized nanosuspension were lyophilized to solid powder form using freeze dryer at -60 ºC for 72 h. The lyophilized powder was collected and used for further solid-state characterization.

Characterization of nanosuspension
The mean particle size and PDI of freshly prepared nanosuspensions were determined by Dynamic Light Scattering method using the Zetasizer (Nano ZS, Malvern Instruments). The surface charge of nanoparticles was analyzed by measuring the zeta potential in a Zetasizer (Nano ZS, Malvern Instruments, UK). Prior to analysis the samples were diluted with de-ionized water. Three dimensional (3D) images of optimized nanosuspension were obtained by using an Atomic Force Microscope (AFM Shimadzu WET-SPM 9600, Tyoto Japan). Surface morphology and topography of the optimized nanosuspension was evaluated by using scanning electron microscope (JEOL, JSM-6400, Japan) at an accelerating voltage of 10.0 KV under different magnifications. The molecular interactions of optimized nanosuspension with stabilizer were recorded on an FT-IR spectrometer (Perkin Elmer Spectrum Version 10.4.3). For this purpose, coarse plant extract, formulated nanosuspension and SLS (stabilizer) were subjected to FT-IR analysis. The spectra were obtained in KBr disk. A small amount of sample was placed on the lens and scans were obtained with scanning range between 4000 -450 cm -1 .

Dissolution study
The in vitro dissolution profile of coarse suspension and optimized nanosuspension were determined in a phosphate buffer (pH 7.4) using USP Apparatus-2 (paddle method). The lyophilized nanosuspension powder (200mg) was placed in 900 mL dissolution medium at 37 o C with constant stirring at 100 rpm. After specific time intervals (10,20,30, 50, 60, 90 and 120 minutes), 5 ml of dissolution medium was collected and replaced with the equal volume of fresh dissolution medium to maintain the constant volume of the dissolution medium. Silybin in S. marianum coarse suspension and nanosuspension was analyzed spectrophotometrically at 288 nm as silybin flavonoids give strong absorbance at this wavelength [15]. The dissolution test was conducted in triplicate and results are expressed as mean ± S.D (n = 3).

Statistical analysis
For process optimization of nanosuspensions, CCD was used and data analyzed using Design Expert Software (version 7.1, Stat-Ease, Inc. USA). Results of dissolution studies were expressed as mean ± SD.

Optimized process parameters for S. marianum nanosuspensions
CCD was employed to study the impact of selected parameters on mean particle size, PDI and zeta potential values of nanosuspensions. The Design-Expert software was used for predicting the best fitted model, and quadratic polynomial regression equations were constructed by means of regression analysis. Response surface plots were generated to explore the main effects and relationship between independent and dependent variables. Total of 20 formulations were prepared according to experimental design and characterized for particle size, PDI and zeta potential. The optimized nano-formulation was selected on the basis of smallest mean particle size, lower PDI and required zeta potential values. The experimental design used for optimization with observed responses for twenty experiments are presented in Table 1. The second order polynomial regression equation showing the impact of S/PE ratio (A), AS/S ratio (B) and stirring time (C) on mean particle size, PDI and zeta potential are represented as in Eqs 1 -3.
Mean particle size (Y1) = 197.78 + 20.12A -38.51B -16.08C + 6.96AB -6.84AC -1.39BC -0.14A 2 -3.60B 2 + 21.25C 2 ……. (1) The positive sign represent a synergistic effect while the negative sign indicates an antagonistic effect of the variables on response. The analysis of variance (ANOVA) has been employed to determine the significance of selected model. The statistical significance of regression models was determined by F-value which is a measurement of variation of experimental data about the mean. The ANOVA Table of quadratic models for mean particle size, PDI and zeta potential of S. marianum nanosuspensions are presented in Table 2, Table 3 and Table 4. The higher model F-values of 35.85 (for mean particle size), 11.90 (for PDI) and 17.41 (for zeta potential) indicated that the proposed quadratic models were good predictors of the experimental data. Values of Prob > F lower than 0.05 (< 0.0001 for Y1, 0.0003 for Y2, < 0.0001 for Y3) revealed that the models were significant.
The Lack of Fit F-value of 0.0783 (for Y1), 0.5859 (for Y2) and 0.4830 (for Y3) implied that lack of fit is insignificant relative to pure error. The insignificant lack of fit confirmed the correctness of selected models. The coeffcient of determination (R 2 ) of 0.9699, 0.9146 and 0.9400 for particle size, PDI and zeta potential showed a good correlation between experimental and predicted values. The value of R 2 adj 0.9363 (Y1), 0.8566 (Y2) and 0.8900 (Y3) were close to R 2 suggesting the high significance of the model. The effect of input variables on response variables was also explored by generating 3D response surface plots. Figure 1 (a) reveals the combined effect of S/PE ratio and AS/S ratio on particle size with stirring time constant at 4 h. The mean particle size was increased by increasing S/PE ratio and decreased by increasing AS/S ratio. The minimum particle size of 125 nm was obtained in formulation S1 prepared with low S/PE ratio (1:1) and high AS/S ratio (15:1). The response surface plot in Figure  1 (b) exhibits the interaction effect of S/PE ratio and stirring time on mean particle size at fixed level of AS/S ratio. Figure 1(b) clearly illustrating that mean particle size was decreased by increasing stirring time and increased by increasing S/PE ratio. The plot in Figure 1 (c) reveals the interaction effect of AS/S ratio and stirring time at fixed level of S/PE ratio. Graph indicates significant reduction in mean particle size by increasing AS/S ratio and stirring time up to an optimum value. Further increase in stirring time, however, increased the particle size.
The interaction effect of S/PE ratio and AS/S ratio on PDI values of S. marianum nanosuspensions at fixed stirring time is given in Figure 2 (a). The graph clearly demonstrates that PDI first decreased and then increased with increasing S/PE ratio. By increasing the AS/S ratio PDI values of S. marianum nanosuspensions were decreased. Figure 2 (b) illustrates the interaction effect of S/PE ratio and stirring time at fixed value of AS/S ratio.      Figure 3(a) shows the interaction effect of S/PE ratio and AS/S ratio on zeta potential values of nanosuspensions at fixed value of stirring time. Zeta potential of S. marianum nanosuspensions were increased by increasing S/PE ratio. However, AS/S ratio has no significant effect on zeta potential values of S. marianum nanoformulations.

Validation of model
To optimize all the responses (mean particle size, PDI, zeta potential) numerical optimization using desirability and graphical optimization technique by using overlay plot was used ( Figure  4). The optimized nanosuspension of S. marianum seeds extract was obtained after setting the constrains on independent and response variables. The levels of input variables which minimize the mean particle size (Y1), PDI (Y2) and maximize zeta potential (Y3) were calculated by Design Expert software for achieving optimized formulation. The optimum values of input variables caluculated from softaware were A (1:1), B (15:1) and C (6 h) to obtain the particle size (139.08 nm), PDI (0.319) and zeta potential (-39 mV). The fresh formulation was prepared and evaluated for particle size, PDI and zeta potential. The optimized formulation with mean particle size of 137 nm, PDI of 0.327 (Figure 5 a) and zeta potential of -37 mV (Figure 5 b) was obtained.

Stability studies
Physical stability of optimized nanosuspension of S. marianum extract was determined at room temperature and 4 ºC for the period of three months. The appearance, mean particle size, PDI and zeta potential were recorded after three months. No visual sedimentation of particles was observed at the two different temperatures after subsequent period. It was found that mean particle size, PDI and zeta potential of S. marianum optimized nanosuspension was slightly increased after the storage period (Table  5). These results indicated that nanosuspension formulation maintained the particle size during storage conditions.

Atomic force microscopic (AFM) features
Atomic Force microscopic features confirmed the mean particle size of optimum formulation in nanometer range. In AFM technique the size of particles evaluated as height of the particles. Particles were spherical in shape with average height 298.26 nm. The height of particles measured in scale A-B were 29.9 and 71.98 nm ( Figure 6). The great difference in sizes of particles indicated non-uniform particle size distribution. Agglomeration of particles was also observed which might be due to drying step during sample preparation.

FT-IR spectra
The FT-IR spectra of S. marianum extract, optimized nanosuspension and SLS (stabilizer used for formulation of nanosuspension) are given in Figure 8(a-c) respectively. A broad peak at 3445.26 cm -1 in recorded spectrum of S. marianum extract was assigned to -OH stretching vibration. A recorded FT-IR spectrum of nanosuspension exhibited the characteristc peak of -OH stretching in this region with slight variation in position and intensity indicating hydrogen bonded interaction between phenolic -OH and stabilizers. Other peaks in spectrum of S. marianum extract was observed at 2917.33 (-CH stretching vibration) and 1638.59 cm -1 (stretching vibration of -CO group). In spectrum of nanosuspension, these peaks were shifted to higher wavenumber. In addition, the more intensity of these peaks in nanosuspension indicated some minor intermolecular interactions between plant bioactive and stabilizer.

Dissolution
The dissolution profile of optimized S. marianum nanosuspension (NS) and coarse suspension (CS) are given in Figure 9. In first ten minutes, approximately 41.41% S. marianum nanosuspension and 21.5% coarse suspension (silybin equivalent) were dissolved in dissolution medium. The dissolution velocity of S. marianum NS and CS was increased with time. The presented results in graph illustarted that the dissolution of S. marianum NS was faster than CS at every time point. After 120 min, 95.78% silybin was released from lyophilized S. marianum nanosuspension. In contrast to this only 58.12% of silybin was released from S. marianum CS into phosphate buffer.

DISCUSSION
In the present study, nanosuspension of S. marianum extract was formulated by using antisolvent precipitation method and different parameters were optimized by using response surface methodology. From 3D response surface plots, increase in mean particle size with increasing S/PE ratio indicated that low concentration of stabilizer was sufficient for providing electrostatic repulsion between nanoparticles. Further increasing the amount of SLS increased the free surfactant in solution forming the SLS micelles. The increased solubility of plant phytoconstituents in SLS micelles promotes Ostwald ripening leading to increase in particle size [16].
Moreover, increase in concentration of SLS above the required value increased the osmotic pressure in the solution, which caused the attraction between colloidal particles, leading to the formation of larger particles [17].
By increasing AS/S ratio, the particle size was decreased due to decreased solubility of plant extract in antisolvent and higher supersaturation rate [18]. The reduction in mean particle size by increasing stirring time is attributed to the increased mechanical energy provided for generating smaller particles. But further increase in stirring time above an optimum value resulted in increase in particle size due to enhanced collisions between nanoparticles, leading to aggregation [19].
The reduction in PDI values by increasing S/PE ratio revealed that higher SLS concentration favors the production of more uniform particle size distribution. But further increase in stabilizer concentration results in Ostwald ripening and growth of nanosized particles. Zeta potential is the potential measured at hydrodynamic shear plane and measured by determining the electrophoretic mobility under an electric field. The electrophoretic mobility depends upon surface charge and concentration of electrolyte. The main mechanism of electrostatic stabilization by surfactants is due to ionic interaction on the drug surface. The high surface charge caused electrostatic repulsion; thus, prevent the nanoparticles from agglomeration and aggregation [20]. It was found that the observed values of response variables were in good agreement with the values predicted by the model. These results showed that that model was reliable for predicting the effect of selected variables on response variables.
The results of dissolution study suggested that enhanced dissolution velocity of S. marianum nanosuspension is due to reduced particle size. The reduced particle size and high surface area of nanoparticles significantly increased saturation solubility and dissolution velocity of nanosuspensions [21]. The saturation solubility was strikingly enhanced by reducing the particle size according to Ostwald-Freundlich equation [22]. The enhanced dissolution rate of nanosuspensions by decreasing particle size was explained by Noyes Whitney equation [23]. The enhanced dissolution rate of nanosuspensions could also be explained due to increased solubility and wettability of poorly soluble flavonoids of plant extract in dissolution medium, due to presence of surface stabilizers [24].

CONCLUSION
Nanosuspension of S. marianum seeds extract has been successfully formulated by anti-solvent precipitation method. The optimized formulation has a mean particle size of 137 nm, PDI of 0.327, zeta potential of -37 mV and is stable for three months at room temperature and in refrigerator. The optimized formulation exhibits remarkable enhancement in dissolution velocity when compared to coarse suspension. Thus, nanosuspensions is potentially a suitable pharmaceutical formulation for poorly soluble herbal extracts.