Development and optimization of sulforaphane-loaded nanostructured lipid carriers by the Box-Behnken design for improved oral efficacy against cancer: in vitro, ex vivo and in vivo assessments

Figures & data

Table 1. Various independent and dependent variables used in the Box–Behnken design for preparation of SFN-loaded NLCs.

	Levels used, actual (coded factor)
Factor	Low (−1)	Medium (0)	High (+1)
Independent variables
A = Lipid concentration (% w/v)	2	3	4
B = Surfactant concentration (% w/v)	2	2.5	3
C = Sonication time (min)	3	4	5
Dependent variables	Goal
Y1 = Particle size (nm)	Minimize
Y2 = Entrapment efficiency (%)	Maximize
Y3 = Drug loading (%)	Maximize

Figure 1. Solubility of SFN in different (A) liquid lipids, and (B) solid lipids.

Figure 2. Three dimensional (3D) response surface plots showing simultaneous influence of independent variables on response parameters: (A) particle size, (B) entrapment efficiency and (C) drug loading of SFN-loaded NLC formulated within Box–Behnken experimental design.

Figure 3. Linear correlation plots between actual and predicted values and the corresponding residual plots for (A) particle size, (B) entrapment efficiency and (C) drug loading.

Table 2. Observed Box–Behnken experimental runs of SFN-loaded NLCs with their actual and predicted experimental value of Y₁ (particle size), Y₂ (entrapment efficiency) and Y₃ (drug loading).

	Independent variables			Dependent variables
				Y1 (Mean) ± SD		Y2 (Mean) ± SD		Y3 (Mean) ± SD
Runs	A	B	C	Actual	Predicted	Actual	Predicted	Actual	Predicted
NLC1	4	2	4	208.46 ± 9.92	207.65	93.86 ± 4.24	94.41	11.62 ± 2.28	10.84
NLC2	4	2.5	5	194.21 ± 8.74	195.33	85.54 ± 5.46	84.24	12.48 ± 3.12	11.22
NLC3	2	2.5	3	144.35 ± 6.46	145.23	77.54 ± 4.42	78.15	10.72 ± 2.68	11.98
NLC4	3	3	3	159.64 ± 7.24	160.95	90.75 ± 4.28	88.70	10.46 ± 3.24	10.82
NLC5	3	2.5	4	145.38 ± 4.46	145.82	84.94 ± 3.82	84.19	14.82 ± 3.46	14.52
NLC6	3	2.5	4	145.78 ± 4.52	145.82	83.76 ± 3.34	84.19	14.78 ± 3.58	14.52
NLC7	3	2.5	4	146.32 ± 4.76	145.82	84.48 ± 3.66	84.19	14.42 ± 3.72	14.52
NLC8	2	2	4	156.31 ± 7.84	155.36	77.03 ± 4.34	76.80	9.52 ± 3.98	10.37
NLC9	2	3	4	110.42 ± 5.52	112.23	74.92 ± 5.78	74.37	8.56 ± 2.82	9.34
NLC10	3	3	5	147.27 ± 6.78	148.20	78.12 ± 5.92	78.49	10.76 ± 3.38	10.47
NLC11	3	2	5	189.21 ± 9.94	187.90	82.42 ± 6.12	82.47	9.54 ± 3.56	9.89
NLC12	2	2.5	5	132.27 ± 4.24	133.52	63.86 ± 5.98	64.04	8.32 ± 2.86	8.63
NLC13	3	2.5	4	145.32 ± 4.82	145.82	84.56 ± 3.76	84.19	14.32 ± 3.36	14.52
NLC14	4	2.5	3	202.24 ± 9.56	200.99	92.48 ± 4.84	93.30	12.86 ± 4.56	11.75
NLC15	3	2.5	4	146.30 ± 7.12	145.82	83.21 ± 4.52	84.19	14.25 ± 3.68	14.52
NLC16	3	2	3	195.45 ± 8.78	197.52	92.12 ± 3.76	91.25	13.25 ± 4.42	13.64
NLC17	4	3	4	177.56 ± 8.24	175.51	91.58 ± 5.48	90.81	11.28 ± 3.74	11.73

A: Lipid concentration (% w/v); B: Surfactant concentration (% w/v); C: Sonication time (min); Y₁: Particle size (nm); Y₂: Entrapment efficiency (%); Y₃: Drug loading (%).

Table 3. Summary of regression analysis for responses Y₁ (particle size in nm), Y₂ (entrapment efficiency in %) and Y₃ (drug loading in %) for fitting data to different models.

Model	R^2	Adjusted R^2	Predicted R^2	SD	%CV
Response (Y1)
Linear	0.8367	0.7990	0.7551	12.45	–
2F1	0.8424	0.7478	0.5929	13.94	–
Quadratic	0.9999	0.9997	0.9993	3.44	0.27
Response (Y2)
Linear	0.8978	0.8742	0.7898	10.72	–
2F1	0.9120	0.8591	0.5632	9.90	–
Quadratic	0.9961	0.9911	0.9668	5.73	0.87
Response (Y3)
Linear	0.2726	0.1047	0.0894	3.09	–
2F1	0.3384	0.0586	0.5979	3.27	–
Quadratic	0.9921	0.9820	0.9262	1.30	2.49
Y1 = 145.82 + 29.89A – 19.32B – 4.84C + 3.74AB +1.01AC – 1.53BC +6.37A^2 + 11.0B^2 + 16.0C^2
Y2 = 84.19 + 8.76A – 1.26B –5.37C – 0.042AB +1.69AC – 0.73BC – 2.92A^2 + 3.08B^2 – 1.42C^2
Y3 = 14.52 + 1.39A – 0.36B – 077C + 0.16AB +0.51AC +1.0BC – 2.09A^2 – 2.18B^2 – 1.33C^2

Y₁: Particle size (nm); Y₂: Entrapment efficiency (%); Y₃: Drug loading (%); R²: Coefficient of correlation; SD: Standard deviation; CV: Coefficient of variation.

Figure 4. Image showing (A) particle size and particle size distribution, (B) zeta potential and (C) TEM analysis of optimized SFN-loaded NLC formulation.

Figure 5. FT-IR spectra of (A) SFN, (B) Precirol^® ATO 5 and (C) SFN-loaded optimized NLC formulation.

Figure 6. Image showing (A) in vitro drug release profile of plain SFN solution, and optimized SFN-loaded NLC formulation and (B), (C), (D) and (E) represents the release kinetics models for optimized SFN-loaded NLC formulation.

Table 4. In vitro release models fitting in terms of linear regression coefficient (R²).

		Plot
Model	Equation	x-axis	y-axis	R^2
Zero order	Mt = M0 + k0 t	Fraction of drug released	time	0.752
First order	ln Mt = ln M0 + k1 t	Log % drug remaining	time	0.926
Korsmeyer–Peppas model	Mt/M∞ = Kt^n	Log fraction of drug released	log time	0.929
Higuchi matrix model	Mt = M0 + kH t^1/2	Fraction of drug released	$\sqrt{\mathrm{t}\mathrm{i}\mathrm{m}\mathrm{e}}$	0.982
Best fitted model	Higuchi matrix model

Table 5. Stability data of the optimized SFN-loaded NLC formulation at 25 °C ± 2 °C/65% ± 5% RH and 40 °C ± 2 °C/75% ± 5% RH for 90 days (n = 3).

Temp	Time(days)	Phaseseparation	Precipitateformation	Redispersibility	Particlesize ± SD nm	PDI ± SD	% Entrapmentefficiency ± SD	% Drugloading ± SD
25 °C	0	No	No	Yes	145.38 ± 4.46	0.178 ± 0.024	84.94 ± 3.82	14.82 ± 3.46
	30	No	No	Yes	146.42 ± 3.56	0.184 ± 0.031	83.54 ± 4.21	14.57 ± 3.27
	60	No	No	Yes	161.74 ± 5.98	0.189 ± 0.026	80.36 ± 4.76	13.66 ± 4.19
	90	No	No	Yes	196.46 ± 8.64	0.192 ± 0.034	76.47 ± 3.58	12.94 ± 3.78
40 °C	0	No	No	Yes	145.38 ± 4.46	0.178 ± 0.024	84.94 ± 3.82	14.82 ± 3.46
	30	No	No	Yes	163.26 ± 6.72	0.188 ± 0.032	81.48 ± 3.74	13.64 ± 2.34
	60	No	No	Yes	189.78 ± 8.48	0.212 ± 0.046	78.96 ± 4.21	12.42 ± 2.73
	90	No	No	Yes	234.32 ± 11.68	0.248 ± 0.038	74.37 ± 4.67	11.38 ± 3.12

Figure 7. Ex vivo gut permeation study showing: (A) cumulative amount of drug permeated (µg) versus time, (B) Cumulative drug transported (µg cm⁻²) versus time, and (C) P_app (ACP) for plain SFN solution and optimized SFN-loaded NLC formulation. Image (D) and (E) represents confocal laser scanning microscopic image of gut treated with plain rhodamine B solution, and optimized NLC formulation containing rhodamine B, respectively.

Figure 8. Image showing MTT assay for placebo NLC formulation, plain SFN solution and optimized SFN-loaded NLC formulation with (A) MCF-7 human lung adenocarcinoma cancer cells, (B) B16-F10 murine melanoma cells and (C) SW620 human Caucasian colon adenocarcinoma cells. Image (D) represents DPPH and ABTS free radical scavenging assay of free SFN solution and optimized SFN-loaded NLC formulation. All data are mean ± SD (n = 3) with statistical difference at p < .05.

Figure 9. Plasma concentration–time profiles of SFN in rats after oral administration of optimized SFN-loaded NLC formulation and plain SFN suspension.

Table 6. Pharmacokinetic parameters of SFN in optimized NLC formulation and plain SFN suspension (n = 6, mean ± SD).

Pharmacokinetic parameters	Plain SFN suspension	Optimized SFN-NLC
Cmax (ng/ml)	204.21 ± 20.62	772.58 ± 54.36
Tmax (h)	1	4
AUC0→24(ng h/ml)	1086.67 ± 14.82	5475.32 ± 25.46
AUMC0→24 (ng h^2/ml)	6661.10	37581.21
t1/2 (h)	2.54 ± 0.48	7.53 ± 0.86
Kel (h^−1)	0.125	0.091
F	–	5.04

F = Relative bioavailability.

Jackson SJ, Singletary KW. Sulforaphane inhibits human MCF-7 mammary cancer cell mitotic progression and tubulin polymerization. J Nutr. 2004;134:2229–2236.

 PubMed Web of Science ®Google Scholar