Strategic optimization of curcumin-loaded nanostructured lipid carriers via design of experiment approach for the treatment of chronic wound

2.3.1. Solubility assessment of CUR in various solid lipids and liquid lipids

A visual inspection method was employed to assess the solubility of CUR in various solid lipids and liquid lipids. One gram of different solid and liquid lipids was heated in a glass vial to 10°C above the melting point of the solid lipids. The heating was done to maintain the molten state of solid lipids throughout the study. The assessment cannot be performed at room temperature as the solid lipid re-solidifies, hinders equilibrium drug partitioning, and leads to underestimation of solubility. Moreover, this approach also mimics the actual formulation process wherein the drug is dissolved in a lipid blend heated at 10ºC higher than the solid lipids’ melting point. The lipids were agitated at 100 revolutions per minute (rpm) on a magnetic stirrer (Remi 2 MLH, Remi Lab World, India) for 24 hours. CUR was added in increments of 2 mg in various lipids (until saturation was achieved) to determine its solubility. The CUR-NLCs fabrication was done by utilizing the solid lipid and liquid lipid exhibiting maximum solubility of CUR. The solubility study was performed in triplicate [25,26].

2.3.2. Solubility assessment of CUR in surfactants

CUR solubility in various surfactants was established at 25ºC ± 2ºC in 1% w/v surfactant solutions. The surfactant solution demonstrating CUR’s least solubility was further utilized for preliminary batch fabrication. The polydispersity index (PDI), sphericity, drug leaching, and entrapment efficiency of various preliminary batches were considered for the selection of the most suitable surfactant [25,26].

2.3.3. Screening of solid lipid and liquid lipid binary mixture

The solid lipid and liquid lipid selected after conducting the solubility studies were mixed in different ratios, viz., 50:50, 60:40, 70:30, 80:20, 85:15, 90:10, and 95:05 for evaluating their miscibility. Different ratios were prepared by weighing the liquid lipid and solid lipid to equal 1 g. The mixtures were vortexed (Remi CM-101 Plus, Remi Lab World, India) for 1 hour along with intermittent heating at 10ºC above the melting point of the solid lipid. Subsequently, the mixtures were stored at room temperature for 24 hours. The solidified binary mixture was evaluated by smearing it on a Whatman filter paper and observing stains of oil droplets. The mixture ratio depicting no blotting on Whatman filter paper and having more than a 40ºC melting point was selected for batch preparation [27].

2.6.1. Particle size, polydispersity index, and zeta potential (ZP) determination

The particle size and PDI of CUR-NLCs were determined by the Horiba SZ-100 nanoparticle analyzer (Horiba, Japan), which is based on the dynamic light scattering (DLS) principle. The CUR-NLCs were diluted appropriately (~ 10 times) and analyzed at a scattering angle of 90° and 25 °C temperature. Additionally, the ZP of CUR-NLCs was determined after filling the undiluted sample in the zeta cell (equipped with carbon electrodes). The particle size and ZP results of CUR-NLCs were determined using Horiba SZ 100 ^TM software. The sample analysis was done in triplicate [29].

2.6.2. % Drug loading and % entrapment efficiency analysis

The %DL is a measurement of the drug amount incorporated in lipid nanoparticles relative to the total weight of the lipid phase. The %EE is a measurement of the amount of drug entrapped relative to the total quantity of drug added during formulation. It stipulates the % of drug incorporated in nanoparticles and % of free drug remaining in the dispersion medium. The %DL and %EE of CUR-NLCs were determined using Sephadex^® G-50 minicolumns based on the size exclusion chromatography principle. The separation of free CUR was done by centrifugating (Remi R-4C, Remi Lab World, India) the CUR-NLCs sample (2 ml) through the Sephadex^® G-50 minicolumns at 3,000 rpm for 5 minutes. The CUR-NLCs obtained after centrifugation (filtrate) were diluted with methanol: chloroform (2:1) and analyzed using a validated HPLC technique. The analysis was done in triplicate [24,30,31]. The calculation of %DL and %EE was done using the following formula:

$\%\mathrm{DL}=\frac{\mathrm{Drug}\mathrm{entrapped}}{\mathrm{Total}\mathrm{added}\mathrm{Solids}}\times 100$ (Eq. 2)

$\%\mathrm{EE}=\frac{\mathrm{Drug}\mathrm{entrapped}}{\mathrm{Amount}\mathrm{of}\mathrm{drug}\mathrm{added}}\times 100$ (Eq. 3)

2.6.3. Differential scanning calorimetry (DSC) analysis

The DSC analysis of pure CUR, a physical mixture of lipid components, blank-NLCs, and CUR-NLCs was performed to analyze their thermal characteristics by employing DSC7020 (Hitachi, Tokyo, Japan). The aliquots of freshly prepared blank-NLCs and CUR-NLCs were lyophilized (Freezing: −80°C, 8 hours; primary drying: −40°C, 60 mTorr; 24 hours and secondary drying: 20°C, 30 mTorr, 8 hours) to obtain a dry powder. The powdered samples (~5 mg) were weighed, sealed into hermetic aluminum pans, and placed against a blank empty reference cell. The scanning rate was 10°C/minute and the temperature range of 30°C to 250°C was covered. The analysis was conducted under a liquid nitrogen flow of 55 ml/minute to maintain an inert atmosphere [32].

2.6.4. Powder X-ray diffractometry (PXRD)

The PXRD analysis of CUR, lipid mixture, and lyophilized CUR-NLCs was conducted by employing an X-ray diffractometer (D8 Discover-Bruker, Germany) to determine the changes in crystallinity of lipid or drug after NLC preparation. The samples (~500 mg) were exposed to Ni-filtered Cu Kα radiation at 40 kV voltage and 30 mA current. Individual samples were scanned in a diffraction angle (2θ) range from 5° to 55°. Furthermore, using OriginPro Software (OriginLab Corporation, USA).

2.6.5. Transmission electron microscopy (TEM)

The size, shape, and morphology of CUR-NLCs were established by TEM (Talos F200i S/TEM, Thermo Fisher Scientific, Massachusetts, USA). The CUR-NLC dispersion (suitably diluted) was stained negatively using phosphotungstic acid solution (1% w/v), and subsequently placed over a Pelco^® grid (100 mesh formvar^® coated gold grid-3 mm, Ted Pella, California, USA). The TEM imaging was conducted post-drying at 100 kV voltage and 0.3 nm magnification.

2.6.6. In vitro release study and drug release kinetics

The in vitro release study of optimized CUR-NLCs was performed by employing the dialysis bag technique. Prior to the experiment, the dialysis membrane with a molecular weight cut off 12–14 KDa was cut into a desirable size and treated as per Himedia^® protocol. The nanodispersion comprising CUR-NLCs (5 ml) was added in a dialysis tube, sealed, and immersed in 30 ml of release media, viz., phosphate buffer saline (pH 7.4) containing 0.5% w/v Tween 80 and 10% v/v ethanol. Herein, 7.4 pH phosphate buffer saline (PBS) was selected to mimic a chronic wound environment, and Tween 80 and ethanol were added to keep CUR soluble. Chronic wound exudates are in a neutral to alkaline pH range (pH 7.4–8.9). Using pH 7.4 PBS mimics wound bed conditions and provides a physiologically relevant release environment. pH 7.4 PBS is widely used for in vitro release studies owing to its buffering capacity, physiological relevance, and reproducibility [33]. Tween 80, a non-ionic surfactant, promotes the micellar solubilization of CUR, reduces interfacial tension, and thereby increases its solubility. Ethanol was added as a cosolvent to further enhance CUR solubility, prevent its precipitation, and enable reproducible sampling. The concentrations of Tween 80 and ethanol were selected such that they give proper solubility of CUR, help in maintaining sink conditions, without affecting the nanoparticle integrity. The beakers comprising dialysis bag dispersed in release media were kept at 37°C ± 5°C, and 100 rpm (provides proper mixing without damaging dialysis bag) under magnetic stirring (Remi 2MLH, Remi Lab World, India). At predetermined time intervals (0, 0.5, 1, 2, 3, 4, 5, 6, 7, 8, 10, 12, 24, 36, and 48 hours), a 1 ml sample was withdrawn and replenished with 1 ml of fresh media to maintain the sink condition. The analysis of samples was performed using a previously validated HPLC technique. Thereafter, the Drug Dissolution solver (DD solver) software was used to analyze the drug release kinetics. Various mathematical models, viz., first order, zero order, Hixon–Crowell, Higuchi, Korsemeyer–Peppas, and Weibull, were evaluated to determine the drug release mechanism from CUR-NLCs [24]. The model with the lowest F-value and highest R-squared value was the best fit for dissolution profiling. The formula of the models to which the release data was fitted is mentioned below:

First order: F = 100·(1–e^–k1·t) (Eq. 4)

Zero order: F = k₀.t (Eq. 5)

Hixon–Crowell: F = 100.[1– (1–k_HC·t)³] (Eq. 6)

Higuchi: F = k_H·t^0.5 (Eq. 7)

Korsemeyer–Peppas: F = k_KP·tⁿ (Eq. 8)

$\mathrm{Weibull}:\mathrm{F}={\mathrm{F}}_{\max }·\left[1-\mathrm{e}\frac{(\mathrm{t}–\mathrm{Ti}{)}^{\beta }}{\alpha }\right]$ (Eq. 9)

Where, in all models, F is the fraction (%) of drug released in time t, k₀ is a zero-order release constant, k₁ is the first-order release constant, k_His the Higuchi release constant, k_HC is the release constant in the Hixson–Crowell model, k_KP is the Korsemeyer–Peppas release constant, incorporating the structural and geometric characteristics of the formulation, and n is the diffusional exponent describing the release mechanism. F_max is the maximum drug released at infinite time, β (shape parameter) characterizes a curve as exponential, sigmoid, or parabolic, and α (scale parameter) defines the time scale of the process [34].

2.6.7. Stability study

The CUR-NLCs were assessed for stability for 6 months by incubating them at different conditions, viz., 2°C–8°C (refrigerated condition), 30°C ± 2°C/65% ± 5% RH (intermediate condition), and 40°C ± 2°C/75% ± 5% RH (accelerated condition) as per the International Conference on Harmonization (ICH) guidelines. The CUR-NLCs were investigated at stipulated time intervals (initial, 1, 3, and 6 months) to evaluate any changes in particle size, PDI, %EE, and physical appearance [35].

3.4.1. Effect of surfactant and co-surfactant concentration

The preliminary batches were prepared using Solutol HS 15 and Tween 20, at different concentrations to select one among both and an optimum concentration that provides a homogenous CUR-NLCs dispersion, good entrapment efficiency, and better stability. All preliminary batches were characterized to determine their particle size (D90), PDI, and %EE. Various batches of CUR-NLCs (Batch CS1 to CS6) were formulated using Solutol HS 15 at a concentration range of 0.25% to 3%w/v. The lipid content (1% w/v), drug concentration (5 mg), HSH parameters (7,000 rpm for 5 minutes), and HPH parameters (500 bar with 10 cycles) were kept constant in all batches. Similarly, CUR-NLCs (Batch CS7 to CS12) were fabricated using Tween 20 at a 0.25% to 3%w/v concentration range. The results revealed that a lower particle size and higher entrapment efficiency were observed in batches prepared using Solutol HS15 compared to Tween 80. Thus, Solutol HS15 was selected for further trials. Moreover, the effect of Solutol HS15 concentration revealed a decrease in D90 up to 1% w/v; however, an increase in D90 and PDI was observed above 1% w/v concentration. An increase in particle size above 1% w/v concentration might be owing to the larger size micelles formulation in the presence of excess surfactant. A similar trend was observed for %EE, i.e., a decrease in %EE above 1% w/v Solutol HS 15 concentration. A decrease in %EE beyond a certain concentration might be due to the diffusion of the drug from the lipid phase to the surfactant micelles in the aqueous phase [41]. Hence, 1% w/v Solutol HS 15 was selected for further trials (Table 2).

Table 2. Influence of various formulation and processing variables on CUR-NLCs.

^*Data are expressed as mean ± SD, n = 3.

Co-surfactants/emulsifiers have been reported to enhance drug solubilization, provide better emulsification, and increase the uniformity of NLCs [42]. For achieving better stability of CUR-NLCs, various batches were prepared by using different emulsifiers, viz. Phospholipon^® 90H (Batch CCS1 to CCS3), Phospholipon^® 90G (Batch CCS4 to CCS6), Span 80 (Batch CCS7 to CCS9), and Plurol Olique (Batch CCS10 to CCS12), at a concentration range of 0.1% w/v to 0.4% w/v, while keeping 1% w/v Solutol HS 15 concentration. The results of the study revealed that among various co-surfactants, batches prepared with Phospholipon^® 90H demonstrated the lowest PDI and particle size. Moreover, the %EE improved from 74% ± 1.6% (batch CS3) to 90% ± 2.4% (batch CCS2) owing to the addition of Phospholipon^® 90H. The improvement in %EE might be due to increased drug solubilization in the lipid matrix. Thus, Phospholipon^® 90H at 0.2% w/v concentration was selected for further trials (Table 2). Thereafter, to determine the influence of surfactant concentration in the presence of an emulsifier on CUR-NLCs, batches SCC1 to SCC6 were fabricated by keeping a fixed concentration of Phospholipon^® 90H at 0.2% w/v and varying the surfactant concentration from 0.25% to 3% w/v. From the results, it was evident that the batch prepared with 1% w/v Solutol HS 15 and 0.2% w/v Phospholipon^® 90H depicted the highest entrapment efficiency and lowest PDI value (Table 2).

3.4.2. Effect of CUR concentration

Batches DC1, DC2, and DC3 were formulated, comprising 5 mg, 10 mg, and 15 mg, respectively, of CUR. The lipid concentration (1% w/v), Solutol HS 15 (1% w/v), Phospholipon^® 90H, HSH parameters (7,000 rpm for 5 minutes), and HPH parameters (500 bar with 10 cycles) were kept constant in all batches. The results revealed that with an increase in drug concentration from 5 mg to 10 mg, there was an increase in %EE. However, a decrease in %EE was observed when the drug concentration was increased from 10 mg to 15 mg (Table 2). This phenomenon might be due to the saturation solubility of CUR in the lipid matrix, leading to its expulsion. Moreover, the particle size slightly increased with an increase in CUR concentration. The results were in good agreement with the previously reported literature [24].

3.4.3. Effect of homogenization speed

Batches HS1, HS2, and HS3 were fabricated at varied homogenization speeds, viz., 7,000 rpm, 9,000 rpm, and 11,000 rpm, respectively. The lipid concentration (1% w/v), Solutol HS 15 (1% w/v), Phospholipon^® 90H (0.2% w/v), and HPH parameters (500 bar with 10 cycles) were kept constant in all batches. The results described a decrease in particle size with an increase in HSH speed (Table 2). The decrease in particle size can be attributed to increased shear force with an increase in homogenization speed during pre-emulsion formation [25]. Additionally, with an increase in HSH speed, an increase in %EE was observed, which could be due to the fact that at lower rpm, unidirectional and less turbulent flow might have led to a loss of drug. At higher rpm, the generation of proper forces might have contributed to an increase in %EE [43].

3.4.4. Effect of homogenization pressure and cycles

Batches HP1 and HP2 were fabricated with varying homogenization pressures of 300 bar and 500 bar. The homogenization cycles were varied from 5 to 15 cycles. The lipid concentration (1% w/v), Solutol HS 15 (1% w/v), Phospholipon^® 90H (0.2% w/v), and HSH parameters (9,000, 5 minutes) were kept constant in all batches. The results revealed that at 300 bar pressure (5, 10, and 15 cycles), the PDI was found to be >0.4, depicting the formation of polydisperse aggregated particles. Furthermore, when increasing the homogenization pressure to 500 bar, a PDI of < 0.3 was observed at 10 and 15 cycles. Moreover, with an increase in homogenization cycles and pressure, a decrease in particle size of CUR-NLCs was observed (Table 2). This might be owing to the diminution of NLCs by cavitation forces produced inside the HPH at higher pressure and increased cycles [31].

3.5.1. Statistical optimization of CUR-NLCs using Box–Behnken Design

It was evident from the results of preliminary trial batches that among various formulation and process variables estimated, the amount of drug, concentration of emulsifier, and homogenization speed were found to have a noteworthy influence on the size, %EE, and %DL of CUR-NLCs. In order to systematically optimize the critical formulation and process variables, the BBD (3 factors, 3 levels) was applied. The details of independent variables and dependent variables for optimization runs in BBD are displayed in Table 3. The experimental design runs for CUR-NLCs are mentioned in Table 4. Each response was optimized by employing a model that was validated by utilizing ANOVA in Design expert^® software. The results indicated that the quadratic model was the best fit for particle size, %EE, and %DL. For each response, the model’s adequacy was assessed by using p-values, lack of fit, model F-value, and model summary statistics.

Table 3. Box–Behnken design-independent and dependent variables summary.

Table 4. Experimental design runs for CUR-NLCs development.

^*Data are expressed as mean ± SD, n = 3.

For particle size as a response (Y1), the quadratic model was most significant with a model F-value of 21.61, and p-value < 0.05. Moreover, the lack of fit was insignificant with an F-value of 0.99 and a p-value of 0.537, i.e., p > 0.05. The R2 obtained was 0.974, and the difference between the predicted R² (0.7375) and adjusted R² (0.9298) was less than 0.2. The ANOVA table for the quadratic model of response 1, i.e., D90, is represented in Table 5. The mathematical model generated by Design Expert^® software depicting the relationship between X1, X2, and X3 and particle size is described in Equation 10. Furthermore, 3D response surface plots and contour plots depicting the effect of factors on the particle size of CUR-NLCs are demonstrated in Figure 2.

Figure 2. Influence of independent variables on particle size (D90) of CUR-NLCs as evaluated by quadratic model (R2 = 0.974, p < 0.05). [Click here to view]

Table 5. ANOVA summary for the quadratic model assessing the influence of independent factors on D90 of CUR-NLCs.

D90 = +148.33 + 2.50X₁ + 7.37X₂ - 7.63X₃ + 1.75X₁ X₂ + 2.75X₁X₃-9.00X₂ X₃ - 11.17X₁² + 1.58X₂² + 0.5833X₃² (Eq. 10)

where X1, X2, and X3 are the concentration of the drug, co-surfactant concentration, and homogenization speed, respectively. In the quadratic model for D90, X₂, X₃, X₂X₃, and X₁² had a significant effect (p < 0.05) on the particle size of CUR-NLCs. The positive coefficients of X1 imply that with an increase in drug concentration, the particle size increases. This may be attributed to the fact that with an increase in drug concentration, the viscosity of the lipid phase increases, which further leads to an increase in pre-emulsion globule size [24]. Furthermore, it was evident that the X2 variable (co-surfactant concentration) had a positive influence on the particle size of CUR-NLCs [44]. The increase in particle size with an increase in co-surfactant concentration may be attributed to enhanced drug entrapment with increasing co-surfactant concentration. The excess co-surfactant might have formed multilayers around NLCs and provided more space for the incorporation of the drug. Similar findings have been reported by Patel et al. [45]. The negative coefficient of X3 (HSH speed) depicts a diminution in particle size with an increase in its level. This indicated that high shear forces generated at higher HSH speed led to a higher size reduction of CUR-NLCs [26]. The interaction effect of X1X3 had a positive impact on D90. On increasing drug concentration and HSH speed, the particle size of CUR-NLCs increased, which might be due to an increase in viscosity at high drug concentration. The shear forces might be unable to reduce particle size even at increased speed due to increased viscosity, leading to an increase in particle size of CUR-NLCs. The interaction effect of X2X3 had a significant effect on D90. The negative coefficient suggests that an increase in HSH speed, along with co-surfactant concentration, results in a decrease in particle size. This might be because at high speed, the co-surfactant rapidly stabilizes the interfaces formed by droplet disruption [46].

Moreover, for response %EE (Y2), the statistically significant model was the quadratic model with a model F-value of 24.92, and p-value < 0.05. The lack of fit was insignificant with an F-value of 0.28 and a p-value of 0.869, i.e., p > 0.05. The correlation coefficient R² was 0.978, with a predicted R² of 0.8738 and adjusted R² of 0.9389 (difference between adjusted R² and predicted R² < 0.2). The ANOVA table for the quadratic model for response 2, i.e., %EE, is represented in Table 6. In the quadratic model for %EE, X₁, X₁X₂, X₂X₃, X₁², X₂², and X₃² had a significant effect (p < 0.05) on the %EE of CUR-NLCs. Equation 11 represents the influence of X1, X2, and X3 on response Y2, i.e., %EE. The influence of independent variables on %EE is illustrated by contour plots and 3-D response surface plots in Figure 3.

Figure 3. Influence of independent variables on %EE of CUR-NLCs as evaluated by quadratic model (R2 = 0.978, p < 0.05). [Click here to view]

Table 6. ANOVA summary for the quadratic model assessing the influence of independent factors on %EE of CUR-NLCs.

%EE = +83.33 – 5.75X₁ + 0.5000X₂ + 0.2500X₃ + 3.00X₁ X₂ – 1.50X₁ X₃ – 2.00X₂ X₃ – 4.42X₁² – 3.42X₂² – 5.42X₃² (Eq. 11)

The influence of drug concentration on %EE (Y2 response) can be inferred from Figure 3. Initially, with an increase in drug concentration (till a certain point), there was an increase in %EE; however, beyond 12 mg, a decline in %EE was observed. This phenomenon can be justified by the saturation solubility phenomenon, i.e., the expulsion of a drug beyond a saturation point. Furthermore, the positive coefficient of X2 and X3 depicts an increase in %EE with an increase in co-surfactant concentration and HSH speed. This positive impact of X2 can be justified by the fact that the lipophilic nature of the co-surfactant aids the solubility of CUR in the lipid phase, which in turn augments the entrapment of CUR in CUR-NLCs [47]. Additionally, the positive influence of HSH speed on %EE might be due to the generation of proper forces at higher RPM. The drug gets evenly dispersed in the lipid matrix before solidification at higher shear forces. At lower RPM, the unidirectional and less turbulent flow might be insufficient to entrap the drug into the nanocarriers. Moreover, at high shear, the stronger disruptive forces break lipid melt into smaller droplets, which have a higher surface-to-volume ratio and solidify faster on cooling. The lipid matrix entraps more drug because of rapid solidification and reduces the chance of drug diffusion into the aqueous phase [48]. The interaction effect of X1 and X2, i.e., drug concentration and co-surfactant concentration, resulted in an increased %EE. This could be because an increase in co-surfactant concentration improves the solubilization of CUR in the lipid matrix and prevents its leakage into the aqueous phase. Interestingly, increasing co-surfactant concentration and HSH speed individually enhanced %EE; however, their interaction effect was opposite. This might be attributed to the high heat generation (at high HSH speed) and increased permeability of the drug from the interface (at increased co-surfactant), resulting in the diffusion of the drug in the continuous phase.

The influence of X1, X2, X3, and their interactions on %DL (Y3) was deciphered utilizing the quadratic model that had an F-value of 240.37, and p-value < 0.05, and had an insignificant lack of fit (F=1.24; p = 0.4753, i.e., p > 0.05). The R² value was 0.997, with a predicted R² of 0.974 and an adjusted R² of 0.993. The ANOVA table for the quadratic model for response 3, i.e., %DL, is represented in Table 7. In the quadratic model for %DL, X₁, X₂, X₃, X₁X₂, X₁X₃, X₁², X₂², and X₃² had a significant effect (p < 0.05) on the %DL of CUR-NLCs. The influence of the independent variables (X1, X2, and X3 and their interactions) on the dependent variable Y3 is summarized in equation 12, and the response surface plots and contour plots are illustrated in Figure 4.

Figure 4. Influence of independent variables on %DL of CUR-NLCs as evaluated by quadratic model (R2 = 0.997, p < 0.05). [Click here to view]

Table 7. ANOVA summary for the quadratic model assessing the influence of independent factors on %DL of CUR-NLCs.

%DL= +1.67 + 0.4612X₁ – 0.1275X₂ + 0.0688X₃ – 0.1250X₁ X₂ + 0.0525X₁ X₃ + 0.0200X₂ X₃ – 0.2629X₁² – 0.1504X₂² – 0.1429X₃² (Eq. 12)

It is evident from equation 6 that X1 and X3 had a positive influence on Y3 (%DL), while X2 had a negative impact on Y3 (%DL). The increase in %DL with an increase in drug concentration (X1) can be owing to a greater amount of CUR available for solubilization during CUR-NLCs formation. The increase in %DL was found to be linear up to 12 mg drug loading. Beyond that, the relation between drug concentration and %DL was not observed to be linear, which might be due to the saturation solubility of the drug in lipid. An increase in drug concentration beyond the solubility limit in the lipid phase tends to partition out, resulting in drug leaching. The molecular packing and degree of lipid crystallinity strongly impact the drug retention. Crystalline, highly ordered packing reduces the amorphous region to accommodate the drug, while imperfect lipid packing introduced by adding liquid lipid or surfactants provides amorphous regions to accommodate the drug and reduce leaching [49–51]. Moreover, with an increase in co-surfactant concentration, a decrease in %DL was observed. During NLCs fabrication, softening of the drug-lipid melt was observed at higher co-surfactant concentrations, which in turn increased NLCs hardening time, leading to increased probabilities of drug leaching from the lipid matrix [23]. High co-surfactant leads to disruption of solid lipids’ crystalline packing. At optimum levels, a stable amorphous pocket is created for drug accommodation; however, at higher levels, the matrix becomes “fluid-like” and the drug is expelled during cooling. This might result in a lower %DL at higher co-surfactant concentration [52]. Moreover, at higher concentrations of co-surfactant, mixed micelles or a lamellar structure of co-surfactant could have formed outside the lipid particles. The drug might solubilize in these micelles instead of the lipid matrix, resulting in reduced loading in the NLCs. Furthermore, there was a decrease in %DL with an increase in homogenization speed. This can be due to the fact that high homogenization speed creates smaller particles with increased surface area, which allows better entrapment of the drug, leading to higher drug loading. The interaction effect of drug concentration (X1) and co-surfactant concentration (X2) was negative on the %DL. This might be due to the competition of co-surfactant molecules (at higher concentration) with drug molecules to occupy space within the carrier, leading to lesser drug loading. Moreover, the positive X1X3 interaction depicts that with an increase in drug concentration and HSH speed, the drug loading increases. This may be because at high speed, the reduction of particle size leads to increased surface area for drug entrapment, and at high drug concentration, a larger amount of drug is available for loading.

3.5.2. Optimization and model validation

To validate the design model and optimization procedure, constraints were set for each response, and the optimized solution was identified based on the desirability function. According to the desirability, particle size (Y1) was set minimum, %EE (Y2) was set maximum, and %DL (Y3) was set maximum. The optimized batch was fabricated as per the conditions suggested by the software, and the observed and predicted values were compared. The model suggested the optimum conditions (with a desirability score of 1) as follows: Drug concentration (X1) = 10.58 mg, co-surfactant concentration (X2) = 0.177% w/v, and HSH speed (X3) = 9359 rpm. The predicted responses given by the software for the optimized batch were 146.3 nm, 82.148 %, and 1.756 %, for particle size, %EE, and %DL, respectively. In order to check the model’s robustness, the optimized batch was fabricated and analyzed in triplicate. The predicted and observed values of responses Y1, Y2, and Y3 are represented in Table 8, which depicts a good agreement. Thus, the obtained mathematical equations and evolved model can be successfully employed to predict responses within the design space.

Table 8. Comparison of predicted and observed values of optimized CUR-NLCs.

^*Data are expressed as mean ± SD, n = 3.

3.6.1. Particle size, PDI, and Zeta potential

The particle size, PDI, and Zeta potential of optimized CUR-NLCs were found to be 145.2 ± 1.3, 0.233 ± 0.06, and −30.4 ± 2.8 mV, respectively (Fig. 5). The particle size represented by CUR-NLCs was higher than 100 nm, which makes it highly suitable for wound healing applications, as smaller particles may enter the systemic circulation. It has been observed that nanoparticles with a size range of 100–200 nm have been more often used for wound care compared to other nano-dimensions. The optimization of significant formulation factors, such as drug concentration, co-surfactant concentration, and HSH speed that influenced the particle size of CUR-NLCs, helped to attain the desired particle size. Furthermore, the observed PDI was <0.3, which depicts that the optimized CUR-NLCs had uniform particles with a narrow size distribution. The uniformity is ascribed to the optimized HSH and HPH parameters, and the optimal concentration of solutol HS 15 that lessens the interfacial tension and averts droplet coalescence. The zeta potential of optimized CUR-NLCs was observed to be −30.4 ± 2.8 mV, indicating good stability. The lipid polar head groups impart the anionic surface charge, and surfactant adsorption on the particle surface provides stearic stabilization.

Figure 5. Particle size, PDI, and Zeta potential of optimized CUR-NLCs. [Click here to view]

3.6.2. %EE and %DL

The %EE of CUR-NLCs formulation batches as per BBD ranged from 66% ± 2.1% to 85% ± 2.0%. Additionally, the % DL ranged from 0.81% ± 0.03% to 1.98% ± 0.08%. The variation in %EE was attributed to the drug concentration, co-surfactant concentration, and HSH speed. The impact of the above-mentioned factors on %EE has been discussed in section 3.5.1. The optimized CUR-NLCs exhibited a %EE of 81.64% ± 2.0% and %DL of 1.728% ± 0.4%, which were in good agreement with the predicted responses given by the Design Expert^® software.

3.6.3. Differential scanning calorimetry

The DSC analysis was done to characterize the thermal properties and crystallization trend of the NLC system. Figure 6 represents the DSC thermogram of pure CUR, a physical mixture of lipid components, blank NLCs, and optimized CUR-NLCs. The Pure CUR represented an endothermic peak at 177.7°C, depicting its melting point. The physical mixture of lipid components demonstrated a peak at 55.4°C, while blank-NLCs depicted a peak at 57.1°C owing to the presence of co-surfactant and surfactant components. Furthermore, the optimized CUR-NLCs did not exhibit the peak of CUR, confirming the change of its state from crystalline to amorphous, and its molecular inclusion in NLCs [53]. Also, the peak intensity of lipid components decreased in the case of optimized CUR-NLCs thermogram compared to a physical mixture of lipid components and blank-NLCs. This might be due to the distortion of the solid lipid crystalline nature in the presence of surfactant and liquid lipid during the NLCs formation [54].

Figure 6. DSC thermogram of Pure CUR, Physical mixture of lipid components, Blank NLCs, and Optimized CUR-NLCs. [Click here to view]

3.6.4. Powder X-ray diffractometry

The crystallinity of NLCs is influenced by several external and internal factors such as temperature, pressure, lipid type, surfactant type, bioactive component, and so on, which in turn affect the encapsulation and colloidal stability, as well as the release from nanoparticles. The liquid lipid incorporation in NLCs leads to a less ordered structure formation and thereby expands the overall encapsulation efficiency of the drug in NLCs. The XRD diffractograms of CUR, lipid mixture, and lyophilized CUR-NLCs are represented in Figure 7. The PXRD patterns of CUR obtained in the 2θ scale signify its crystalline nature. No distinct peaks of CUR were observed in the diffractogram of lyophilized CUR-NLCs, depicting the change from crystalline to amorphous state, and solubilization and nearly complete entrapment of CUR in the lipid matrix [31,55,56]. The results complied with the findings acquired from the DSC analysis.

Figure 7. PXRD patterns of CUR, Lipid mixture, and Lyophilized CUR-NLCs. [Click here to view]

3.6.5. Transmission electron microscopy

The TEM images of CUR-NLCs demonstrated spherical, homogenous, and smooth-surfaced particles with a size of <150 nm. The particle size observed in TEM imaging was in good agreement with the results obtained by the DLS technique. Moreover, the CUR-NLCs did not reveal any sign of aggregation in the TEM analysis Figure 8.

Figure 8. TEM image of CUR-NLCs. [Click here to view]

3.6.6. In vitro release study and drug release kinetics

The drug release study depicted a biphasic release of CUR from optimized CUR-NLCs. In the first 5 hours, a burst release of more than 30% was observed, and a subsequent prolonged release of ~99% up to 48 hours was observed. The biphasic release pattern might be owing to the imperfect arrangement of liquid lipids and solid lipids in NLCs. The CUR present in liquid lipid might have led to the immediate release, while CUR embedded in solid lipid crystals might be responsible for prolonged release. The degree of lipid crystallinity and drug solubility in the lipid phase determines the location of CUR in the surface and core domains. Higher crystallinity promotes drug expulsion and lower loading, while amorphous pockets of liquid lipids (as in the case of NLC type 1) promote sustained release. Moreover, the particle size and surfactant/co-surfactant concentration also influence drug release by influencing the matrix hydration, wettability, and interfacial area. Smaller particles and higher surface-associated drugs increase the burst fraction. Furthermore, the release kinetics were evaluated by subjecting the release data to various kinetic models by employing the DDSolver software. The release kinetics parameters using different models for CUR-NLCs have been described in Table 9. It was established that among various models, the Weibull model was the best fit for the CUR-NLCs release profile (lowest F and highest R² value) (Fig. 9). The Weibull model describes a combined mechanism of diffusion and dissolution for drug release. This might be due to imperfect NLCs/type 1, which are a type of matrix system with voids that accommodate more drug. A prolonged release up to 48 hours will target the inflammatory phase of wound healing, as CUR is a potent anti-inflammatory and antioxidant agent. It will aid the suppression of pro-inflammatory cytokines, reduce oxidative stress, and decrease neutrophil and macrophage infiltration, thereby preventing chronic inflammation.

Figure 9. Drug release pattern of CUR-NLCs (observed vs. predicted data of Weibull model) (R2 = 0.994) (Mean ± SD; n = 3). [Click here to view]

Table 9. Regression coefficient (R²) and F values of different kinetic models.

3.6.7. Stability studies

The accelerated stability studies of optimized CUR-NLCs were performed as per ICH guidelines, and the changes in particle size, %EE, PDI, and physical appearance were observed. The optimized CUR-NLCs depicted a slight increase in particle size and a minor decrease in %EE in accelerated conditions (40°C ± 2°C/75% ± 5% RH) after 6 months of storage. However, the alterations were not significant in any storage conditions (Table 10). Good stability can be attributed to the proper optimization of significant factors influencing the dependent variables. Moreover, the stearic stabilization provided by surfactant, low PDI, and a zeta potential of −30 mV, which generates good electrostatic repulsion, minimizes aggregation and chances of Ostwald ripening. Moreover, the lipophilicity of CUR also reduces the chances of leaching from the lipid matrix during storage. Thus, it can be concluded that the developed CUR-NLCs were stable and could offer a good shelf life.

Table 10. Stability studies of optimized CUR-NLCs.

^*Data are expressed as mean ± SD, n = 3.