Optimum Surfactant Concentration for Preparation of Amiodarone Loaded Solid Lipid Nanoparticles: Theoretical Estimation Versus Experimental Results by Box-Behnken Design

3.1. Materials

Poloxamer 188, sodium lauryl sulfate (SLS), Tween 80, Witepsol, glyceryl monostearate (GMS), soy lecithin, and Precirol were obtained from Sigma-Aldrich Co. (MO, USA). AMI hydrochloride was provided as a gift. All materials were utilized as received without further purification.

3.2. Determination of Lipid Mixture Melting Point

The lipid mixture, in the optimal formulation ratios (Witepsol: GMS: Lecithin, 15:1.6:1), was heated to 70°C until fully melted and thoroughly mixed. After allowing it to solidify, a capillary tube was dipped into the mixture to collect a sample. The melting point (MP) was determined in open capillaries using an Electrothermal MEL-TEMP apparatus (model 9200), noting the temperature at which the lipid mixture melted.

3.3. Preparation of AMI-loaded Solid Lipid Nanoparticles

Solid lipid nanoparticles (SLNs) containing AMI were prepared using a hot homogenization method, as described by Bhattacharyya and Reddy (13) and Rahman et al. (14), with minor adjustments Various formulations, as specified in Table 1, were prepared. The lipid phase, containing 230 mg Witepsol, 25 mg GMS, and selected amounts of lecithin, was heated to approximately 75°C to melt all components, after which 50 mg AMI was added and dissolved completely. Subsequently, the aqueous phase containing SLS and poloxamer was heated to the same temperature as the lipid phase and combined with the lipid mixture under high shear homogenization (Heidolph, Silent Crusher M, Germany) at 15000 rpm for 10 minutes. The mixture was also stirred at 300 rpm to ensure complete emulsification of all lipids in water and to prevent any un-homogenized areas. Following this, the emulsion underwent further homogenization using a probe sonicator (Bandelin, Germany) at an amplitude of 65% for 4 minutes. The final nanoemulsion was stirred for 1 hour at room temperature to form SLNs, which were then stored at 4°C for subsequent analysis.

3.4. Experimental Design

Given the numerous factors that can influence the outcomes, the variables were categorized into process parameters and formulation parameters, which included surfactant concentrations. To streamline the experimental process, two experimental designs were utilized to minimize the number of runs. Process variables were optimized using FFD, and the most effective preparation procedure identified was then used for the subsequent experimental design to prepare formulations. The quantities of each surfactant were optimized using BBD. The combined use of FFD and BBD enabled the identification of the optimal proportions of each parameter affecting Critical Quality Attributes (CQAs). An initial risk assessment was performed using an Ishikawa diagram, highlighting the process risks and their potential causes (15, 16) Minitab Statistical Software (Minitab Inc., USA) and ANOVA were employed for DOE and statistical analysis. The preparation of optimal formulations was replicated 3 times.

3.5. Evaluation of Diameter and Size Distribution

The average volume diameter of particles and the span value were measured using a particle size analyzer (Wing SALD 2101, Shimadzu, Japan). The size distribution was determined based on the span value using Equation 2. A narrower size distribution indicates the homogeneity of the formulation and low dispersity of particles (20).

D_N% showed that the volume percentage of particles less than D_N% was N% (N = 10, 50, 90).

3.6. Particle Morphology

The morphology of the particles was examined using a Leo 906 transmission electron microscope (TEM), Carl Zeiss AG, at 100kv (Oberkochen, Germany). The optimal formulation was diluted 20-fold with deionized water and stained with 2% uranyl acetate. A drop of the suspension was placed onto a carbon-coated copper grid, allowed to dry, and then observed further (21).

3.7. Drug Loading and Entrapment Efficiency Calculation

Entrapment efficiency (EE)% and drug loading (DL) % were determined indirectly using the ultrafiltration method. An Amicon^® ultrafiltration unit (30 kDa molecular weight cut-off membrane, Millipore, USA) was employed to separate the free drug. The SLN suspension was centrifuged for 10 minutes at 5000 rpm, and the filtrate was analyzed using UV–VIS spectrophotometry at 242 nm (22). Entrapment efficiency % and DL% were calculated using the following Equations 3 and 4 (22):

Where, W_i, W_f, and W_s represent the initial weight of drug added to formulation, weight of unloaded drug, and weight of total solid components, respectively.

3.8. Differential Scanning Calorimetry Analysis

The physical state of the formulation components was investigated using differential scanning calorimetry (DSC) (Shimadzu, Japan). Witepsol, AMI, GMS, lecithin, the physical mixture, and SLNs were examined. Samples were placed in aluminum pans using special pistons, heated at a rate of 10°C/min, and then scanned. The scanning temperature ranged from 25°C to 300°C.

3.9. Which Concentration of Surfactant Was Enough for the Preparation of SLNs?

Based on Martin's Physical Pharmacy and Pharmaceutical Sciences, the area per molecule of surfactants can be calculated (23). Therefore, the surface tension of the surfactant solution versus the logarithmic concentration of the surfactant should be plotted. The Du Nouy ring method was utilized to measure the surface tension at room temperature (25°C). Various concentrations of poloxamer and SLS were prepared in clean beakers. The immersed ring and liquid surface were maintained at a specific distance, and the force required to detach the ring was recorded. The platinum-iridium ring and beakers were cleaned with a solution of potassium chromate and sulfuric acid in water after each measurement (24). The total surface area of the SLNs can be estimated based on the volumetric diameter of the particles. Consequently, by dividing the total volume of the lipid phase by the volume of each particle, the number of particles was determined. The related equations are detailed in the Results section.

4.1. Melting Point Determination of the Lipid Mixture

The lipid phase crucial for effective SLNs remains solid at body temperature. It was hypothesized that the addition of GMS (MP ${\displaystyle \cong }$ 55°C) and lecithin to Witepsol would result in a higher MP for the mixture compared to Witepsol alone, as they possess higher MPs than Witepsol. This hypothesis was verified through the MP determination of the mixture. Incorporating components with higher MPs resulted in an increased MP for Witepsol, the predominant component in the formulation, to 44°C (25).

4.2. Experimental Design

The Ishikawa diagram, shown in Figure 1, identified the CQAs for preparing SLNs. Evaluating potential process risks assists formulators in developing products with the desired characteristics efficiently and cost-effectively. According to the diagram, various risk parameters related to processes and components might influence the considered CQA. Particle size, span value, particle size stability over 2 days, and DL% were identified as CQAs for the development of SLNs.

Figure 1.

Ishikawa diagram indicating CQAs for AMI-loaded SLNs.

4.2.1. Evaluation of Optimal Process Parameters by FFD

Given the numerous parameters involved in the fabrication of SLNs, employing two different designs to minimize the number of runs is recommended. The manufacturing process parameters were initially optimized using FFD to identify independent variables significantly influencing particle size. Fractional factorial design enables efficient resource use by reducing the number of test runs. The optimal preparation conditions identified through FFD were then applied for the remainder of the study. A Pareto chart of the parameters is shown in Figure 2 (left). According to the Pareto chart, the lipid type (A), homogenization time (C), and cooling temperature (E) significantly affected the particle size. The homogenization rate (D) and cooling temperature (E) were the primary factors for preparing stable SLNs. The main effect plot for the process parameters is depicted in Figure 2 (right), demonstrating that using Witepsol as the lipid base, homogenizing for ten minutes, and cooling at room temperature resulted in smaller particle sizes. Additionally, a lower homogenization rate yielded more stable SLNs. Therefore, a homogenization rate of 5000 rpm was selected.

Figure 2.

Pareto charts (left) and main effects plot (right) of FFD are used to find the effective process variables. The upper is the size of the particles after preparation, and the bottom is the particle size after two days. (A), Lipid type;(B), surfactant type ;(C), homogenization time;(D), homogenization rate;(E), cooling temperature.

4.2.2. Evaluation of Optimal Surfactant Concentration by BBD

A three-level BBD of the response surface model was employed to determine the optimal concentration of surfactants. BBD is an efficient design for fitting second-order response surface models, enabling the generation of higher-order response surfaces with fewer runs by applying only 3 levels of factors. The designed formulations and their responses are outlined in Table 2. A p-value of less than 0.05 was considered significant. The BBD study identified the optimal formulation, which featured a smaller, stable particle size and span value, along with a higher DL%. The optimal formulation was number 3, consisting of 230 mg Witepsol, 25 mg GMS, 50 mg AMI, 2 mg SLS, 5 mg poloxamer, and 17.5 mg lecithin in a 10 ml aqueous phase, achieving an EE% of 97.58.

The results, detailed in Table 3, indicated that the quadratic model for size (P-value 0.012), DL% (P-value < 0.05), and one-week size (P-value 0.006) was significant, whereas it was not significant for the span value. The R² value for size, DL%, and one-week size indicated a better correlation between dependent and independent variables compared to the span value. The model for each response highlighted the effects of independent variables and their interactions. A normal probability plot, shown in Figure 3, was used to assess the suitability of the model.

Table 3.Significant Parameters and Regression Analysis of Model Suitability for Responses. (Coded Data)

Factors	Y1: Particle Size		Y2: Span Value		Y3: DL%		Y4: 1-Week Size
	Coefficient	P-Value	Coefficient	P-Value	Coefficient	P-Value	Coefficient	P-Value
Model	-	0.012	-	0.058	-	0.000	-	0.006
a0	312	-	1.101	-	14.4165	-	67127	-
X1	1478	0.190	-	-	-0.0340	0.458	7231	0.157
X2	2195	0.066	-	-	-0.4883	0.000	-985	0.837
X3	3507	0.009	0.361	0.058	-0.3943	0.000	-6081	0.226
${\displaystyle {\mathit{X}}_1^2}$	-	-	-	-	-0.1808	0.022	-25666	0.006
${\displaystyle {\mathit{X}}_2^2}$	-	-	-	-	-	-	-32244	0.001
${\displaystyle {\mathit{X}}_3^2}$	3560	0.046	-	-	-0.1435	0.055	-23200	0.009
X1X2	-	-	-	-	-	-	-	-
X1X3	2631	0.109	-	-	-0.1335	0.062	-	-
X2X3	4442	0.016	-	-	-	-	-	-
Model R2 (%)	81.89		24.87		96.57		85.20

Significant Parameters and Regression Analysis of Model Suitability for Responses. (Coded Data)

Figure 3.

Normal probability plots of BBD for predicted versus actual responses.

4.2.3. Effect of Variables on Responses (Y1-Y4)

Table 4 presents the significant variables and their interactions. The correlation between particle size and independent variables was characterized by a regression coefficient of 81.89%, showing an inverse dependency on the concentrations of SLS, poloxamer, and lecithin. The contour plot demonstrated that increasing lecithin concentration initially led to a decrease in size, but beyond a certain point, it resulted in an increase in size, while a higher concentration of poloxamer led to smaller particles. This revealed that the span value was directly proportional to lecithin concentration, but the model was not sufficiently suitable for the span value. Hence, no contour or surface plot was generated for it. The polynomial equation showed that DL% increased as the concentration of poloxamer decreased and as the concentrations of SLS and lecithin increased, with a regression coefficient of 96.57%. The stability of the particles over time was assessed by measuring the particle size after one week, which had a regression coefficient of 85.20%. Figures 4 (A), (B), and (C) display the contour and surface plots for size, DL%, and one-week size, respectively, with SLS fixed at 7 mg.

Table 4.Equations Related to Effect of Variables on Responses (Y1-Y4)

Responses	Equation
Particle size (Y1)	size=11101-441sls-322poloxamer-1309lecithin+22.78lecithin×lecithin+42.1sls×lecithin+28.43poloxamer×lecithin
Span value (Y2)	span value=0.595+0.0289lecithin
DL% (Y3)	DL%=14.802+0.1318sls-0.03907poloxamer+0.0156lecithin-0.00723sls×sls-0.000919lecithin×lecithin-0.002137sls×lecithin
1-week size (Y4)	1week size =-92078+15819sls+7144poloxamer+4710lecithin-1027sls×sls-206.4poloxamer×poloxamer-148.5lecithin×lecithin

Equations Related to Effect of Variables on Responses (Y1-Y4)

Figure 4.

Effect of poloxamer and lecithin with a fixed amount of SLS (7mg) on particle size (A); DL% (B); one-week size (C). Plots could not be plotted for span value, as the model was not suitable for span value. Left and right are contour plot and surface plot, respectively.

4.3. Particle Morphology

The morphology of the optimal formulation, as determined by BBD (formulation 3), was analyzed using TEM. Figure 5 shows that the particles were spherical and nanometer-sized. There was no observed aggregation of the particles. The average particle size measured by TEM was under 100 nm. Based on the TEM image and scale, the estimated particle diameter was approximately 90 nm, consistent with the results from the particle size analyzer.

Figure 5.

Tem micrograph of optimum formulation of AMI-Loaded SLNs.

4.4. Differential Scanning Calorimetry Analysis

Structural changes in the formulation components can be detected through heat exchange, which is reflected in the DSC thermogram. This thermogram demonstrates how melting (through heat absorption) or crystallization (through heat emission) occurs at specific temperature ranges, thereby determining the structural properties of the samples (25).

Figure 6 shows the thermogram of samples containing Witepsol, GMS, lecithin, AMI, a physical mixture of these components, and the optimal formulation of SLNs. The thermogram revealed that the sharp endothermic peak associated with AMI disappeared in the SLN preparation, indicating that AMI was completely dissolved in the lipid phase. Given that the endothermic peak of AMI occurs above 100°C, it was inferred that the AMI structure remained stable during the SLN preparation process (75°C). The presence of the endothermic peak of Witepsol in the SLNs confirmed the polymorphic form of the lipid, as Witepsol is a major component of the lipid phase. The endothermic peak of GMS was not observed in the SLN formulation. Previous studies have also noted that the sharp endothermic peak of AMI disappeared in the solid self-nanoemulsifying drug delivery system of AMI, indicating a transformation of the crystalline form of the drug upon incorporation into the system (26).

Figure 6.

Differential scanning calorimetry (DSC) thermogram of formulation components and SLNs.

4.5. Which Concentration of Surfactant Was Enough for the Preparation of SLNs?

To identify the appropriate concentration of surfactants for the preparation of SLNs, the surface tension of poloxamer and SLS solutions against their logarithmic concentration was plotted, as shown in Figure 7. Following the principles outlined in Martin's Physical Pharmacy and Pharmaceutical Sciences, the surface excess (Γ), defined as "the amount of amphiphile per unit area of the surface in excess of that in the bulk of the liquid," was calculated using Equation 5 at a constant temperature (23).

Figure 7.

Plot of surface tension versus logarithmic concentration of poloxamer and SLS for evaluation of surface excess (Γ) for surfactants (N = 3).

The ratio of ${\displaystyle \frac{\partial \gamma }{\partial lnc}}$ represented the slope of the straight line prior to the critical micelle concentration (CMC) point. As illustrated in Figure 7, the slope for poloxamer and SLS was -4.569 and -5.049, respectively. Consequently, the Γ value for poloxamer and SLS was 1.843×10^-10 and 2.037×10^-10 mole/cm² at 25˚C, respectively. Given that the unit of surface excess is mol/cm², it was necessary to calculate the total mole of surfactant present in the formulation. Therefore, the total area that could be covered by surfactant molecules was estimated as follows:

Where S_s, m_s, and M_w represent the area covered by a surfactant molecule, the total mass of the surfactant in the formulation, and the molecular weight of the surfactant, respectively. The total moles of poloxamer and SLS in the optimal formulation (number 3) were 30.82 and 6.93 μmol, respectively. Based on Equation 6, the total surface area that both surfactants could cover was calculated as follows:

S_s = surface covered by poloxamer + surface covered by SLS

Poloxamer and SLS could cover a total surface area of 20.12×10⁴ cm². Additionally, the total surface area of all particles needed to be estimated. The volume of each particle was calculated based on the volumetric diameter of the particles, as described in Equation 7.

Where S_t, n_p, S_p, V_l, V_p, ρ_l, m_l, and d represent the total surface area of the particles, the number of particles, the surface area of one particle, the volume of the lipid phase, the volume of one particle, the density of the lipid phase, the mass of the lipid phase, and the volumetric diameter of the particles, respectively. Equation 8 illustrates that the total surface area of the particles is estimated by two factors: The volume of the lipid phase and the volumetric diameter of the particles. According to Equation 8, the total surface area for SLNs containing Witepsol and GMS as lipid components, with particles having a 74 nm volumetric diameter, was calculated to be:

Calculations showed that the total surface area for the amount of lipid used in the formulation was 21.72×10⁴ cm². Higher concentrations of surfactants often result in the formation of smaller particles, but it remains uncertain whether the resulting structure is a micelle or an SLN. Since the surfactants (SLS and poloxamer) would cover 20.12×10⁴ cm², the area occupied by surfactant molecules is less than the surface area of the lipids. Consequently, most molecules are likely situated on the surface of lipid particles to form the emulsion, leaving no free surfactant molecules to form micelle structures. Thus, all particles were SLNs. The remaining particle surface not covered by surfactants could be enveloped by the lipid phase surfactant, which has not been accounted for in this study. It would be beneficial to devise a method for assessing the surface area covered by surfactants in the lipid phase to obtain more precise estimations.

5.1. Conclusions

In summary, AMI-loaded SLNs were successfully prepared using the hot homogenization technique. A dual DOE approach was utilized, where effective process parameters were identified through FFD, and surfactant concentrations were optimized using BBD. The optimal formulation, determined by the BBD results, was thoroughly characterized. A significant aspect of SLN preparation is the concentration of surfactants; hence, an equation for predicting the optimal surfactant concentration was developed. This equation demonstrated that surfactant molecules could sufficiently cover the surface of the particles, suggesting the absence of micelle structures in the formulation. Therefore, this equation could aid in determining the necessary concentration of surfactants in the aqueous phase for preparing SLNs. The derived equation focuses on evaluating the optimal concentration of aqueous phase surfactants. Given the crucial role of lipid-phase soluble surfactants in the formation of SLNs, this area warrants further investigation, and additional studies are needed to develop equations for predicting the required concentrations of lipid-phase surfactants.