Screening of process variables using Plackett–Burman design in the fabrication of gedunin-loaded liposomes

Abstract

This study is to screening the formulation and process variables that produce significant effect on the gedunin-loaded liposome formulations by using quality-by-design approach. Placket–Burman screening design was used to screen the most influencing formulation and process variables. Mean vesicle size, zeta potential, entrapment efficiency, and loading capacity were found in the range of 112–990 nm, −19.39 to −39.20 mV, 45.25–87.60%, and 3.54–10.47%, respectively. Differential scanning calorimetry (DSC) and powder X-ray diffraction (XRD) result suggested that Gedunin encapsulated within liposome as amorphous state. The analysis of Pareto chart represented that selected independent variables had a most significant effect on dependent variables.

Keywords:

• Gedunin
• liposome
• Plackett–Burman design
• quality by design

Introduction

Microencapsulation is a process in which core materials are encapsulated within a wall. It can release their core material contents at the controlled manner (Mozafari et al. 2008). Liposomes are spherical vesicles mainly composed of phospholipid bilayer membranes, which either encapsulating water-soluble drugs in their core cavity or to solubilize lipid-soluble drugs in their bilayer membrane (Felber et al. 2012). During the last few decades, liposomes got to success for drawing interest to many researchers as a potential and versatile carrier for drug-delivery systems. Liposomes is one of the most reliable lipid-based colloidal carriers, which extremely biocompatible, low antigenicity, low toxicity, controlled and targeted release drugs at the site of action (Allen and Cullis 2013) and provide protection against external unfavorable conditions such as light, pH, and enzymes (Shehata et al. 2008).

Gedunin [C₂₈H₃₄O₆] is an important triterpenoid limonoid bioactive isolated from various meliaceous plants. Gedunin [GDN] possess various promising in vitro and in vivo pharmacological effects such as antimalarial (Roy and Saraf 2006), antiallergic (Ferraris et al. 2012), anticancer (Brandt et al. 2008), antifungal (Pandey et al. 2014), vasodilator (Pramely and Raj 2012), antisecretory effect and protects the gastric mucosa (Lakshmi et al. 2010), and filaricidal activity against human lymphatic filarial parasite (Mishra et al., 2011). Along with wide range of promising pharmacological activity, their insolubility or low aqueous solubility and stability limit their clinical applicability (Roy and Saraf 2006).

All available techniques for the preparation of liposomes have some limitation. There are many process and formulation parameters that directly and indirectly influenced the quality of liposomes. Nowadays, principles of quality by design (QbD) adopt to ensure the quality of drug as it relates their safety and efficacy (ICH 2005). Design of experiment (DoE) is one of the most common and frequently used methodologies for investigate a phenomenon to gain better understanding or improve performance. It is a method used for the determination of relationship between the various factors influenced the process and the outputs of the process. Experimental design can be applied to define the objective of investigation and decide the number and nature of experimental variables, nature of the response, and minimum number of experimental runs (Jones 2009). Experimental design includes an initial factor screening (usually by Plackett–Burman and fractional factorial design) followed by optimization (usually by Box–Behnken and central composite design) (Gupta et al. 2016).

The present work focused to prepare GDN-loaded liposome by a modified thin-film hydration technique. Plackett–Burman screening design (PBD) was used to screen the effect of various process and formulation factors. In this study, the selected formulation and process factors were drug concentration (DC), lipid concentration (LC), cholesterol/lecithin ratio (C/L), chloroform/methanol ratio (C/M), volume of organic solvent (O/S), volume of aqueous vehicle (AV), rehydrated solution pH (pH), water bath temperature (WBT), rotary evaporator rotation time (RT), rotary evaporator rotation speed (RS), and ultrasonication time (UST). The effects of the process and formulation factors on mean vesicle size (VS), zeta potential (ZP), drug encapsulation efficiency (EE), and drug-loading capacity (LC) were investigated. The GDN-loaded liposomes were further evaluated by transmission electron microscopy (TEM), differential scanning calorimetry (DSC), powder X-ray diffraction (XRD), and in vitro drug release study.

Materials and methods

Materials

Gedunin (GDN) was purchased from Lanospharma Laboratories Co. Ltd. Chongqing, China having 98% percentage purity. Cholesterol and soya lecithin were purchased from HiMedia Laboratories Pvt. Ltd. Mumbai, India. Chloroform and methanol were purchased from Loba Chemie Pvt. Ltd. Mumbai, India. All other chemicals and solvents used for the study were of analytical grade.

Instrumentations

The liposome was produced by using rotary evaporator (IKA RV 10 digital, Staufen, Germany) and a probe sonicator (Frontline Sonicator, Mumbai, India). Cooling centrifuge (Remi Instruments Ltd. Mumbai, India) was used to remove larger lipid aggregates and titanium particles. For release study, diffusion cell apparatus (EMFD-08 orchid scientific & Innovative India pvt. ltd. Nasik, Maharashtra India) was used. For shaking, the solution used vortex shaker (IKA Werke GmbH & Co. KG, Staufen, Germany). The Malvern Zetasizer (Nano ZS 90, Malvern Instruments, UK) was used for the estimation of vesicle size, polydispersity index and zeta potential value of liposome formulations. The UV-visible double-beam spectrophotometer (Lambda 25, PerkinElmer, Waltham, MA) with matched quartz cells (1 cm) was used for determining the encapsulation efficiency. Transmission electron microscope (Zeiss EM 10, Germany) was used for surface morphology study. XRD (PANalytical 3 kW X’pert Powder, UK) and DSC (Shimadzu DSC-60 Systems, Japan) were used for characterizing physical properties of formulations.

Methods

Preparation of GDN-loaded liposomes

The GDN-loaded liposomes were prepared by conventional thin-film hydration method as described previously with slight modification (Gradauer et al. 2012). Briefly, the required amount of GDN and lipid (cholesterol/lecithin) were mixed in the organic solvent (chloroform/methanol). This mixture was poured into a clean, dry round-bottom flask and dried under the reduced pressure by using rotary evaporator (IKA RV10 Digital Rotary Evaporator, Germany) at 60 °C with 100 rpm to form a thin lipid film. Further, this thin lipid film was holding under vacuum overnight to remove traces of the organic solvent. The dried thin lipid film was rehydrated with the addition of phosphate buffer saline solution (PBS; 10 mM) and then vortexes to maintain the temperature at 60 °C. The resulting mixture contained MLVs (multilamellar vesicles) suspension, which was further ultrasonication with probe sonicator (Frontline Sonicator, Mumbai, India) at room temperature in order to produce SUVs (small uni-lamellar vesicles). After that, resulting, SUVs suspension was centrifuged by using ultracooling centrifugation machine (REMI Cooling Centrifuge, Mumbai, India) to remove larger lipid aggregates and titanium particles released from the Sonicator probe. Then, SUVs suspensions were filtered through 0.22-μm membrane filters to get small size liposomes (Wang et al. 2010).

Experimental design (Plackett–Burman design)

Nowadays, various statistical models are frequently used to screen the large number of factors with fewer runs of experiments. In this research work, we have adopted Plackett–Burman screening design (PBD) to develop the GDN-loaded liposome formulation by modified thin film hydration method. An 11-factor PBD at two levels (high and low) was applied for the preliminary screening of the main effects of eleven variables (Table 1) on vesicle size (VS), polydispersity index (PDI), zeta potential (ZP), drug encapsulation efficiency (EE) and drug-loading capacity (LC) of GDN-loaded liposome formulations (Table 2). The low and high levels of the variables were selected for the basis of previous experiments and literature reviews. PBD experiment was done by using Design-Expert® 9.0.6.2 software (Stat-Ease Inc., Minneapolis, MN) (Gupta et al. 2016, Ravanfar et al., 2016).

Table 1. The independent variables and levels of Plackett–Burman screening design (PBD).

Code	Independent variables	Abbreviation	Low level	High level	Units
A	Drug concentration	DC	25	50	mg
B	Lipid concentration	LC	250	500	mg
C	Cholesterol/lecithin ratio (w/w)	C/L	0.25	1
D	Chloroform/methanol ratio (v/v)	C/M	0.5	2
E	Volume of organic solvent	OS	10	20	mL
F	Volume of aqueous vehicle	AV	25	50	mL
G	Rehydrated solution pH	pH	6	8
H	Water bath temperature	WBT	30	60	°C
J	Rotary evaporator rotation time	RT	1	3	hrs
K	Rotary evaporator rotation Speed	RS	50	100	rpm
L	Ultrasonication time	UST	1	10	mins

Table 2. Outline and observed response of Plackett–Burman screening design batches (average ± standard deviation, n =3).

Batchcode	A	B	C	D	E	F	G	H	J	K	L	VS (nm)	PDI	ZP (-mV)	EE (%)	LC (%)
1	50	250	0.25	0.5	20	25	8	60	1	100	10	278 ± 5.00	0.212 ± 0.019	20.25 ± 0.97	58.45 ± 0.70	10.47 ± 0.62
2	25	500	1	0.5	20	50	8	30	1	50	10	282 ± 8.72	0.187 ± 0.010	38.15 ± 1.67	73.4 ± 0.79	3.54 ± 0.49
3	25	500	0.25	2	20	25	8	60	3	50	1	682 ± 17.06	0.197 ± 0.015	34.72 ± 1.27	80.51 ± 1.31	3.86 ± 0.58
4	25	500	1	2	10	25	6	60	1	100	10	292 ± 7.94	0.134 ± 0.007	39.2 ± 0.71	76.31 ± 0.84	3.67 ± 0.49
5	50	500	1	0.5	10	25	8	30	3	100	1	990 ± 18.02	0.537 ± 0.035	35.71 ± 1.04	81.89 ± 1.18	7.57 ± 0.48
6	25	250	1	0.5	20	50	6	60	3	100	1	451 ± 9.16	0.418 ± 0.014	23.35 ± 0.70	51.41 ± 0.52	4.89 ± 0.12
7	25	250	0.25	0.5	10	25	6	30	1	50	1	322 ± 8.89	0.301 ± 0.016	19.39 ± 0.55	45.25 ± 0.75	4.33 ± 0.38
8	50	500	0.25	2	20	50	6	30	1	100	1	612 ± 7.82	0.181 ± 0.005	33.28 ± 0.87	83.2 ± 0.60	7.68 ± 0.63
9	25	250	0.25	2	10	50	8	30	3	100	10	112 ± 4.36	0.431 ± 0.012	20.79 ± 0.23	55.61 ± 0.62	5.27 ± 0.22
10	50	250	1	2	10	50	8	60	1	50	1	787 ± 12.12	0.172 ± 0.009	22.75 ± 1.28	56.11 ± 1.03	10.09 ± 0.41
11	50	500	0.25	0.5	10	50	6	60	3	50	10	258 ± 13.07	0.214 ± 0.007	35.71 ± 0.66	87.6 ± 0.60	8.05 ± 0.31
12	50	250	1	2	20	25	6	30	3	50	10	211 ± 11.36	0.182 ± 0.014	25.35 ± 0.84	61.71 ± 0.83	11 ± 0.50

In this design, the variables were correlated by using following polynomial equation. (1) where Y is the response (dependent variable), A₀ is the constant, A_1, A_2, A_3, A₄ …….A_n is the coefficient of the response and X_1, X_2, X_3, X₄………… X_n is representing the effect of each variable ordered within −1, +1 (Kuchekar and Pawar 2014).

On the basis of Pareto chart, analysis of variance (ANOVA) results and three dimensional (3D) surface plots, we were established the variables which exhibited significant main effects on the selected responses.

Characterization of gedunin-loaded liposomes

Determination of vesicle size (VS), polydispersity index (PDI), and zeta potential (ZP)

Liposomes vesicle size (VS) and polydispersity index (PDI) were measured by photon correlation spectroscopy (PCS) using a Malvern Zetasizer (Nano ZS 90, Malvern Instruments, UK) at 25 °C. The liposomes dispersion was diluted appropriately with ultra-purified water before measurement to overcome the opalescence. All measurement was conducted at a 90° scattering angle. The zeta potential (ZP) was determined by measuring the electrophoretic mobility through Malvern Zetasizer. The temperature and field strength were applied at 25 °C and 23 V/cm, respectively. The Helmoltz–Smoluchowski equation was generally applied for conversion into the ZP. All the above measurements were conducted in triplicate (Petralito et al. 2014).

Determination of encapsulation efficiency (EE) and loading capacity (LC)

The encapsulation efficiency (EE) and loading capacity (LC) of GDN in liposomes were determined by UV-Visible spectrophotometry as described earlier with slight modification (Aditya et al. 2014, Yang et al. 2014). Briefly, GDN-loaded liposomes suspension was centrifuged at 12,000 rpm for 20 min at 4 °C. After centrifugation, unencapsulated GDN was settled down and formed pellets due to their hydrophobicity and crystallinity. The supernatant was discarded, and GDN pellet was dissolved in methanol and diluted appropriately. Estimation of unencapsulated GDN was carried out with UV-Visible spectrophotometer (Lambda 25, PerkinElmer) at 220 nm and calculation was done by using following equations:

Transmission electron microscopy (TEM) study

The surface morphology of the liposome vesicles was analyzed by using transmission electron microscope (TEM). A water-diluted dispersion of the liposomes was placed on a carbon-coated copper grid, which permitted the adsorption of the liposome vesicles. Excess water was removed by filter paper and staining was done by using phosphotungstic acid (1% w/v). Excess amount of staining solution was removed and the grid was air-dried at room temperature, further observed by the TEM (Zeiss EM 10, Germany) (Petralito et al. 2014).

Powder X-ray diffraction (XRD) study

Powder X-ray diffraction (PXRD) measurements of pure Gedunin (GDN), cholesterol, soy lecithin, lyophilized empty, and drug-loaded liposomes were carried out with X-ray diffractometer (PANalytical 3 kW X’pert Powder, UK) using Cu K radiation as X-ray source. The samples were placed in sample holder and scanned from 2θ to 50θ with a scan angular speed 2θ min⁻¹. Operating voltage was 40 kV, and current was 30 mA (Das et al. 2011).

Differential scanning calorimetric (DSC) study

Thermograms of the pure gedunin (GDN), cholesterol, soy lecithin (phosphatidylcholine), lyophilized empty, and drug-loaded liposomes were recorded by using a differential scanning calorimetric (DSC, Shimadzu DSC-60 Systems, Japan). Selected samples (5 mg) were placed on standard aluminum pans and sealed nonhermetically. Then, sealed pans were kept under the isothermal condition at 25 °C for 10 min. Calibration of the instrument was done by using indium (purity >99.999%) on the basis of melting point and heat of fusion. Scanning was performed in the temperature range between 30 and 280 °C at heating rate 10 °C/min, and nitrogen purge of 20 mL/min. In this study, an empty sealed aluminum pans with a lid was taken as a reference (Mishra et al. 2014, Petralito et al. 2014).

In vitro drug release study

The in vitro GDN release from liposome formulation was assessed by the diffusion cell apparatus (EMFD-08 orchid scientific & Innovative India Pvt. ltd. Nasik, Maharashtra, India) using dialysis membrane (molecular weight cutoff 10,000 Da). Before starting experiments, dialysis membrane was kept in double distilled water for 24 h. The liposome suspension containing 2 mg/mL of GDN was placed in the donor compartment, and the receptor compartment was filled with release media (0.1 N hydrochloric acid). The entire arrangement was maintained temperature at 37 ± 0.5 °C by continuous stirring at 50 rpm by magnetic stirrer. After 2 h, release medium was replaced with phosphate buffer pH 7.4, and study was extended next 24 h. At regular time intervals, samples were withdrawn from the receptor compartment and exact volume of a medium was added to the same compartment to maintain the constant volume throughout the study. Samples were further diluted appropriately and amount of GDN released was measured by UV-visible spectrophotometer analysis. The drug release data were reported as the mean ± standard deviation (SD) of three replicates (n = 3) (Kuchekar and Pawar 2014).

Statistical analysis

All measurements were repeated three times. Design-Expert software (version 9.0.6.2; trial version, Stat-Ease, Minneapolis, MN) was utilized for statistical analysis and graph plotting. The statistical analysis was performed at a significant level of P values <0.05.

Result and discussion

Plackett–burman screening design

The prediction of the main effect of independent variables on the dependent variables is a very critical point to the development of GDN-loaded liposomes. PBD was widely applied screening method for identifying the most influencing significant independent variables with few runs. Here, 11 factors have selected as independent variables that may affect the dependent variables (responses) (Dejaegher and Heyden 2011). Independent factors and their levels were shown in Table 1. The outcome of the observed dependent variables of PBD formulation was shown in Table 2. A relationship between dependent an independent variables were identified by the help of polynomial equations for individual dependent variables. The outcome of the Table 3 indicates that the determination coefficients (R²) were found larger than 0.9, which shown that over 90% of the response variation could be explained by this model. The high value of R² was also confirmed the goodness of fit into the model. The F value for each regression model was obtained to be a very high value (69.23–1108.21) with low probability (0.0009–0.0143), representing that the regression model is significant with a 95% confidence level.

Table 3. Summary of analysis of variance (ANOVA) of Plackett–Burman screening design batches.

	Regression parameters of dependent variables
	Vesicle size (VS)			Zeta potential (mV)			Entrapment efficiency (EE)			Drug-loading capacity (LC)
Factors	R^2	F value	P values(Prob >F)	R^2	F value	P values(Prob>F)	R^2	F value	P values(Prob>F)	R^2	F value	P values(Prob>F)
Model	0.9968	69.23	0.0143^a	0.9996	629.40	0.0016^a	0.9998	1108.21	0.0009^a	0.9981	117.28	0.0085^a
A: DC		65.16	0.0150^a		4.73	0.1617		787.52	0.0013^a		880.68	0.0011^a
B: LC		60.03	0.0163^a		5239.81	0.0002^a		8690.46	0.0001^a		139.95	0.0071^a
C: C/L		36.92	0.0260^a		301.71	0.0033^a		34.95	0.0274^a
D: C/M					9.06	0.0949		87.05	0.0113^a		7.59	0.1104
E: OS		3.95	0.1852					12.74	0.0703		6.21	0.1303
F: AV		4.91	0.1572					3.76	0.1921		1.95	0.2971
G: pH		63.86	0.0153^a		11.12	0.0794					1.43	0.3545
H:WBT		3.16	0.2176		7.97	0.1060		31.75	0.0301^a		2.76	0.2386
J: RT					4.95	0.1560		246.72	0.0040^a
K: RS		2.45	0.2579		8.86	0.0968					1.79	0.3130
L: UST		382.61	0.0026^a		76.39	0.0128^a		78.91	0.0124^a		13.15	0.0684

Significant level ^a = P< 0.05.

Effect of independent factors

Vesicle size

The mean vesicle size of GDN-loaded liposomes was obtained from the range of 112–990 nm (Table 2). The Pareto chart (Figure 1(a)) indicates that the independent variables drug concentration (DC), lipid concentration (LC), cholesterol/lecithin ratio (C/L), rehydrated solution pH (pH), and ultrasonication time (UST) possesses significant influence on the mean vesicle size. In the Pareto chart, effects above the Bonferroni limit are most likely significant, effects above the t-value limits are possibly significant (should be considered if they are not already selected) and effects below the t-value limits are not likely to be significant. The ANOVA results also confirm that all the five factors show P values less than 0.05, which indicating that all the five factors having significantly different from zero and 95% confidence level (Table 3). The polynomial equation (Equation. (2)(2) ) which shows the correlation between the independent variables on the dependent variable (VS) was given below (Kuchekar and Pawar 2014). (2)

Figure 1. (a) Pareto chart of the standardized effects of independent variables on vesicle size (b) Response surface plot for vesicle size.

The ultrasonication time (UST) effects cross the Bonferroni limit, that is, it showed great influence on the mean vesicle size (Figure 1(a)). It also has the least P values (0.0026) with respect of all other factors (Table 3). The negative coefficient value of UST in Equation. (2)(2) suggested that the increase in the timing of ultrasonication decreased the mean vesicle size. The higher absolute value of the coefficient suggested that the higher magnitude of effect on the factor in the response variable (Shah et al. 2013). Ultrasonication was responsible for breaking the coarse emulsion globules into the nanoemulsion globules. Increase the timing of ultrasonication place more sonication energy to the liposomal dispersions, which reduced the size of nanoemulsion globules (Das et al. 2011). Lipid concentration (LC) and cholesterol/lecithin ratio (C/L) showed possibly significant influencing on mean vesicle size. Increase the LC and C/L ratio mean vesicle size also increased because sonication energy distribution in the dilutes dispersion is good as compare to concentrated dispersion (Sahu et al. 2014). The positive coefficient value of DC and pH in Equation. (2)(2) indicates that the increase in the DC and pH increased the mean vesicle size (Mishra et al. 2014). Other factors rather than these five factors are insignificant influence on the mean vesicle size.

The response surface plot (Figure 1(b)) further confirms the direct relationship between the UST and mean vesicle size. On the basis of above findings, the input factor UST should be considered for further optimization studies to get the appropriate values.

Zeta potential (ZP)

Zeta potential is the charge get hold on the particle surface in a dispersed system, which supports the physical stability to the system (Das and Choudhury 2011). It is already reported that for the assurance of good stability of the dispersed system the zeta potential value should be ±30 mV (Das et al. 2011, Freitas and Muller 1998) The average zeta potential (ZP) of prepared liposome batches was found to be in the range of −19.39 to −39.2 mV (Table 2). According to the Pareto chart (Figure 2a) independent variables lipid concentration (LC), cholesterol/lecithin ratio (C/L) and ultrasonication time (UST) possess significant influence on the average ZP. The ANOVA results for ZP also confirmed the above observation because p values less than 0.05 (Table 3). The polynomial equation for ZP was shown in Equation. (3)(3) . (3)

Figure 2. (a) Pareto chart of the standardized effects of independent variables on zeta potential (b) Response surface plot for zeta potential.

The positive coefficient values of independent variables LC (B), C/L ratio (C), and UST (L) in Equation. (3)(3) suggested that increase the value of these factors influences with positive effect on ZP (Pradhan et al. 2014). Liposomal dispersion system exhibited negative ZP values because the presence of cholesterol and lecithin on their surface (Szczes 2013). Increase in the ultrasonication time, the smallest vesicle size with homogenous size distribution and highest zeta potential were obtained (Awada et al. 2015). Other factors are producing insignificant influence on the ZP.

Further confirmation was done through the response surface plot (Figure 2(b)) that indicates the direct relationship between factors (LC and C/L ratio) and ZP. On the basis of above findings, the input factors LC and C/L ratio should be considered for further optimization studies to get the appropriate values.

Entrapment efficiency (EE)

The average entrapment efficiency of GDN-loaded liposomes was obtained in the range of 45.25–87.60% (Table 2). Based on the Pareto chart (Figure 3(a)) independent variables drug concentration (DC), lipid concentration (LC), rotary evaporator rotation time (RT), chloroform/methanol ratio (C/M), cholesterol/lecithin ratio (C/L), rotary evaporator water bath temperature (WBT), and ultrasonication time (UST) possesses significant influence on the dependent variable entrapment efficiency (EE). According to the Bonferroni and t-value limit on Pareto chart, factors LC, DC and RT shows almost certainly significant influence on EE, while C/M, C/L, WBT, and UST shows possibly significant influence on EE. The ANOVA result confirmed the above observation (Table 3). The correlation between the independent variables on the dependent variable (EE) was described by polynomial equation, which shown in Equation. (4)(4) . (4)

Figure 3. (a) Pareto chart of the standardized effects of independent variables on entrapment efficiency (b) Response surface plot for entrapment efficiency.

In Equation. (4)(4) , the positive coefficient values of independent variable DC (A), LC (B), C/M ratio (D), WBT (H), RT (J), and UST (L) suggested that increase the value of these factor influences with positive effect on EE. On the other hand, negative coefficient values on a factor C/L ratio (C) suggested that decrease the value increase the EE. Entrapment efficiency significantly increases with the increasing of drug concentration (DC) because EE is directly depended on the DC. Increase the LC in formulations EE is significantly increased because the higher amount of lipid was present in the formulation for encapsulating the drug (Das et al. 2011). EE significantly increased with increased the RT because the contact time of drug and lipid was increased. When the contact time increases than more amounts of drug encapsulated into lipids, which led to high EE (Pradhan et al. 2014). C/L ratio shows negative influence on EE means to increase the ratio decrease EE. Lecithin is a lipophilic surfactant helps to soluble drug in lipid and decreases their concentration solubility of drug in lipid might be affected (Mishra et al. 2014). Other factors OS, AV, and pH shown insignificant effects on the EE.

Further confirmation was done through the response surface plot (Figure 3(b)) which indicates the direct relationship between factors (DC, LC, and C/L RT) and EE. On the basis of above findings, the input factors DC, LC, and RT should be considered for further optimization studies to get the appropriate values.

Drug-loading capacity (LC)

The average drug-loading capacity of GDN loaded liposomes was obtained from the range of 3.54–11.00% (Table 2). According to the Pareto chart (Figure 4(a)) independent variables drug concentration (DC) and lipid concentration (LC) possesses significant influence on the dependent variable drug loading capacity. On the basis of Bonferroni and t-value limit factors DC and LC shows almost certainly and possibly significant influence on drug-loading capacity, respectively. The ANOVA result also confirmed the above observation becoause of P values less than 0.05 (Table 3). The correlation between the independent variables on the dependent variable (DL) was described by polynomial equation, which shown in Equation. (5)(5) . (5)

Figure 4. (a) Pareto chart of the standardized effects of independent variables on drug loading (b) Response surface plot for drug loading.

In Equation. (5)(5) , the positive coefficient value of DC (A) suggested that increase the value of DC also increases the drug-loading capacity. Furthermore, negative coefficient value of LC (B) suggested that increase the value of LC decrease the drug-loading capacity. Every lipid has certain drug-loading capacity, after reaching that drug-loading capacity cannot increase. Further increase the LC, drug-loading capacity was significantly decreased when DC was fixed during formulation. This effect was produced due to reduction of drug-to-lipid ratio in formulations (Das et al. 2011, Hao et al. 2012, Pradhan et al. 2014). Drug concentration increases at their optimum level drug-loading capacity also increase, after that addition of drug led to increase free drug in formulations (Mishra et al. 2014.) All other independent variables were shown insignificant effects on the drug-loading capacity.

The response surface plot (Figure 4(b)) further confirms the direct relationship between independent variables (DC and LC) and drug-loading capacity. On the basis of above findings, the input factors DC and LC should be considered for further optimization studies to get the appropriate values.

Transmission electron microscopy (TEM) study

Transmission electron microscope (TEM) was generally used to morphological evaluation analysis of vesicles or particles. TEM image of GDN-loaded liposome in Figure 5 clearly indicate that vesicles have the spherical or oval shape. This image also reveals that vesicle size of the shown vesicles was less than 200 nm.

Figure 5. TEM image of gedunin-loaded liposome.

Powder X-ray diffraction (XRD) study

Powder X-ray diffraction patterns of pure Gedunin (GDN), cholesterol, soy lecithin, lyophilized empty, and drug-loaded liposomes were shown in Figure 6. The GDN showed distinct sharp and intense peaks at 2^θ values of 18.97, 25.25, 38.75, and 48.52 in the diffractogram. These peaks indicate that the GDN was crystalline nature. These peaks of GDN were absent in GDN-loaded liposome diffractogram which suggested that GDN was present in the amorphous form within the liposome. If GDN was present outside the liposome, it would occur crystallization because its poor aqueous solubility and might be affected the diffraction patterns of GDN-loaded liposome. But no changes occur in diffractogram which indicate that GDN was successfully encapsulated into the liposome. The X-ray diffraction patterns of empty and GDN-loaded liposome was quite similar indicating that presence of GDN did not affect the nature of liposome. In case of bulk cholesterol diffractogram peaks obtained between 2^θ values of 14.67–26.56. These peaks were also obtained in empty and GDN-loaded liposome diffractogram but having lower intensity, which suggest a less ordered structure of cholesterol in liposome.

Figure 6. XRD graph of (a) pure gedunin (GDN), (b) cholesterol, (c) soy lecithin, (d) lyophilized empty liposome and (e) drug-loaded liposome.

Differential scanning calorimetric (DSC) study

Differential scanning calorimetry (DSC) is a useful technique to provide the information about melting and recrystallization behavior of crystalline compounds (Mishra et al. 2014). DSC thermogram of gedunin (GDN), cholesterol, soy lecithin, empty liposome, and drug-loaded liposomes were shown in Figure 7. Gedunin show sharp peak at 216 °C match to melting temperature of GDN (Lakshmi et al. 2010). DSC thermogram of cholesterol showed endotherm at 149 °C and lecithin showed endotherm at 114 °C and 178 °C. In empty liposome formulation, the endotherm of lipids was slightly shifted as comparison to bulk lipids due to decrease in their crystallinity. In case of drug-loaded formulation, same observation was found which probably due to the presence of drug in lipid layer. The endotherm of GDN was totally absent in liposome formulation. It indicates that GDN either form an amorphous dispersion in the lipid matrix or soluble in the lipid matrix. XRD results also support this observation.

Figure 7. DSC thermograph of (a) pure gedunin (GDN), (b) cholesterol, (c) soy lecithin (phosphatidylcholine), (d) lyophilized empty liposome and (e) drug-loaded liposome.

In vitro drug release study

In vitro drug release graphs of different batches 1–6 and 7–12 of liposome formulations are shown in Figure 8 (a,b), respectively. The release study result indicated the prolonged drug release from different batches of liposome formulations. The effect of independent variables on GDN release from liposome formulations after 24 h was revealed from an in vitro release graph. Formulation batch 1 containing LC 250 mg, C/L ratio 1:4, and DC 50 mg shown maximum (98.97%) and formulation batch 4 containing LC 500 mg, C/L ratio 1:1 and DC 25 mg shown minimum (67.50%) percent of GDN released at the 24 h, respectively. This result suggested that LC and C/L ratio played major role in release of GDN from liposome formulations. Increase the LC (250–500 mg) and C/L ratio (1:4 to 1:1) decreased the percent of GDN release from liposome formulations.

Figure 8. In vitro drug release study (a) PBD formulation 1–6, (b) PBD formulation 7–12.

Conclusions

GDN-loaded liposome formulation was successfully prepared by slight modified thin film hydration method. The PBD-QbD approach was used to screen the independent variables and to understand the effect of most significant factors on the dependent variables of GDN-loaded liposome formulations. Within the factors screening studied, three formulation factors drug concentration (DC), lipid concentration (LC), cholesterol/lecithin ratio (C/L), and two process factors rotary evaporator rotation time (RT), ultrasonication time (UST) were found to have significant effects on responses. The present study concluded that PBD could be useful to screen most influencing variables like DC, LC, C/L, RT, and UST that might be used for further optimization. This study proved that PBD was an efficient tool to recognize the independent factors that affecting the response variables and to identify the most significant factor.

Acknowledgements

The authors are thankful to Director, University Institute of Pharmacy, Pt Ravishankar Shukla University Raipur, (C. G.) for providing necessary infrastructural facilities. One of the authors Anil Kumar Sahu is thankful to UGC-BSR for JRF and extends her gratitude towards the supervisor for guidance and support. Author also wants to thank the library of Pt. Ravishankar Shukla University for providing e-resources available through UGC-INFLIBNET.

Disclosure statement

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article.