Enhanced oral absorption of insulin: hydrophobic ion pairing and a self-microemulsifying drug delivery system using a D-optimal mixture design

Figures & data

Figure 1. Physical characterization of HIP complexes. A: Complexation efficiency of INS with different counterions at the indicated counterion to INS molar ratio, (B) partition coefficient of the HIP complex of INS. Values are presented as mean ± standard deviation (n = 3). CE, complexation efficiency; INS, insulin; SOS, sodium n-octadecyl sulfate; DOC, sodium docusate; SDC, sodium deoxycholate; OLE, sodium oleate; HIP, hydrophobic ion pairing.

Figure 2. Characterization of INS–SOS complex. A: Scanning electron microscopy; (B) Fourier transform infrared spectroscopy; (C) differential scanning calorimetry thermograms. (a) INS; (b) SOS; (c) physical mixture of INS and SOS; (d) INS–SOS. Scale bar = 5 μm. INS, insulin; SOS, sodium n-octadecyl sulfate.

Figure 3. Dissociation behavior of the INS–SOS complex. A: Semi-log plots of the remaining INS–SOS complex in different concentrations of NaCl; (B) far-ultraviolet circular dichroism spectra of the INS solution, the dissociated INS, and INS–SOS. Values are presented as the mean ± standard deviation (n = 3). INS, insulin; SOS, sodium n-octadecyl sulfate.

Table 1. Secondary structural compositions of the INS solution, dissociated INS, and INS-SOS.

	α helix	Antiparallel	Parallel	β turn	Others
INS	38.0%	6.9%	7.7%	16.0%	29.8%
Dissociated INS	36.7%	7.4%	7.9%	16.2%	30.6%
INS-SOS	39.4%	7.3%	6.9%	16.1%	25.6%

The quantitative analysis was performed using CDNN secondary structure analysis software. INS, insulin; SOS, sodium n-octadecyl sulfate.

Table 2. Solubility of INS-SOS in various lipid vehicles.

Vehicle	Solubility (mg/mL)
Oil
Capryol 90	6.41 ± 0.17
Capmul MCM	14.40 ± 0.54
Plurol Oleique CC 497	10.56 ± 0.33
Surfactant
Tween 20	0.75 ± 0.06
Labrasol	1.09 ± 0.05
Cremophor EL	0.46 ± 0.02
Cosurfactant
Tetraglycol	45.11 ± 1.24
Transcutol P	36.54 ± 2.87
Brij L4	7.74 ± 0.58

Values are presented as the mean ± standard deviation (n = 3). INS, insulin; SOS, sodium n-octadecyl sulfate.

Figure 4. Pseudo-ternary phase diagram of Capmul MCM (oil), Labrasol (surfactant), and Tetraglycol (cosurfactant). Light gray, dark gray, and dashed areas indicate regions for the self-emulsifying drug delivery system, self-microemulsifying drug delivery system, and experimental domain, respectively.

Figure 5. Experimental design and mathematically analyzed plots. A: Independent and response variables used in the D-optimal mixture design. (B) Three-dimensional response surface plots of each response variable. (C) Overlay plot of the optimized self-microemulsifying drug delivery system formulation. Values in contour lines represent the desirability; X₁: Capmul MCM, X₂: Labrasol, X₃: Tetraglycol, Y₁: droplet size, Y₂: INS stability, Y₃: INS leakage. INS, insulin.

Table 3. Summary of the results of the statistical analysis and model equations for the measured responses.

Model	SD	R^2	R^2 (adj)	p value	Lack of fit p value	PRESS	Remark
Droplet size (nm)
Linear	17.53	0.8807	0.8755	<.0001	<.0001	16,920.15	–
Quadratic	9.25	0.9689	0.9653	<.0001	.0003	5,123.73	–
Special cubic	6.95	0.9829	0.9804	<.0001	.0086	2,928.82	–
Cubic	3.54	0.9959	0.9949	<.0001	.9668	8,13.19	Suggested
INS stability (%)
Linear	5.16	0.7581	0.7476	<.0001	<.0001	1,523.40	–
Quadratic	2.70	0.9383	0.9312	<.0001	.0119	433.10	–
Special cubic	2.29	0.9567	0.9505	.0001	.0659	312.15	–
Cubic	1.77	0.9759	0.9703	<.0001	.4758	207.19	Suggested
INS leakage (%)
Linear	3.84	0.7333	0.7217	<.0001	.0107	831.44	–
Quadratic	2.30	0.9103	0.8999	<.0001	.6858	324.25	–
Special cubic	2.23	0.9176	0.9058	.0616	.7628	312.77	–
Cubic	2.02	0.9372	0.9227	.0131	.9464	262.17	Suggested

INS, insulin; SD, standard deviation; PRESS, predicted residual error sum of squares.

Table 4. Analysis of variance for the quadratic model of the experimental responses.

		Y1 (Droplet size)			Y2 (INS stability)			Y3 (INS leakage)
Source	DF	SS	F value	p value	SS	F value	p value	SS	F value	p value
Model	9	0.01	1044.02	<.0001	4948.13	175.16	<.0001	2382.11	64.70	<.0001
Linear mixture	2	0.01	4154.77	<.0001	3844.05	612.36	<.0001	1863.71	227.80	<.0001
X1X2	1	125.73	10.02	.0030	19.66	6.26	.0166	11.05	2.70	.1083
X1X3	1	64.31	5.12	.0292	26.77	8.53	.0058	11.96	2.92	.0952
X2X3	1	5.06	0.40	.5292	722.38	230.15	<.0001	117.07	28.62	<.0001
X1X2X3	1	0.1678	0.01	.9085	36.66	11.68	.0015	19.87	4.86	.0335
X1X2 (X1 – X2)	1	77.31	6.16	.0175	13.65	4.35	.0436	20.71	5.06	.0302
X1X3 (X1 – X3)	1	2.55	0.20	.6548	32.70	10.42	.0025	13.70	3.35	.0749
X2X3 (X2 – X3)	1	760.77	60.61	<.0001	18.30	5.38	.0205	20.44	5.00	.0312
Residual	39	489.50	–	–	122.41	–	–	159.54	–	–
Lack of fit	24	200.73	0.4344	.9668	76.65	1.05	.4758	69.45	0.4819	.9464
Pure error	15	288.77	–	–	45.76	–	–	90.09	–	–
Total	50	0.01	–	–	5556.42	–	–	2963.55	–	–

X₁: Capmul MCM, X₂: Labrasol, X₃: Tetraglycol. INS, insulin; DF, degree of freedom; SS, sum of square.

Table 5. Predicted and experimental values for the optimized SMEDDS formulation.

	Y1 (Droplet size)	Y2 (INS stability)	Y3 (INS leakage)
Predicted value	107.8	47.56	19.31
Experimental value	115.2 ± 3.7	46.75 ± 3.04	17.67 ± 2.22
Prediction error (%)	6.86	–1.73	–9.28

Values are presented as the mean ± standard deviation (n = 3). SMEDDS, self-microemulsifying drug delivery system.

Figure 6. Semi-log plots of remaining INS in different formulations in simulated gastric fluid (left) and simulated intestinal fluid (right). Values are presented as the mean ± standard deviation (n = 3). INS, insulin; SOS, sodium n-octadecyl sulfate; PM, physical mixture; SMEDDS, self-microemulsifying drug delivery system.

Figure 7. Release profiles of INS, INS–SOS, and the optimized SMEDDS in pH 1.2 medium with 1% Labrasol (left) and pH 6.8 medium (right). Values are presented as mean ± standard deviation (n = 3). Significantly different at p < .05 resulted from post hoc Tukey’s test: *versus INS; ^#versus INS and INS–SOS. INS, insulin; SOS, sodium n-octadecyl sulfate; SMEDDS, self-microemulsifying drug delivery system.

Figure 8. Blood glucose levels in diabetic rats after oral administration of the INS solution (50 IU/kg), INS–SOS solution (50 IU/kg), the optimized SMEDDS at 50 IU/kg (SMEDDS-50IU), and the optimized SMEDDS at 100 IU/kg (SMEDDS-100IU) and after SC administration of the INS solution (5 IU/kg). Values are presented as the mean ± standard deviation. INS, insulin; SOS, sodium n-octadecyl sulfate, SMEDDS, self-microemulsifying drug delivery system; SC, subcutaneous.

Table 6. Pharmacodynamic parameters of INS in a diabetic rat model following oral administration of INS (50 IU/kg), INS-SOS (50 IU/kg), the optimized SMEDDS at 50 IU/kg (SMEDDS-50IU), the optimized SMEDDS at 100 IU/kg (SMEDDS-100IU), and SC injection of INS solution (5 IU/kg).

Parameter	INS (SC)	INS (oral)	INS–SOS	SMEDDS-50IU	SMEDDS-100IU
Dose (IU/kg)	5	50	50	50	100
AAC0–8h (%·h)	439.29 ± 30.00	–44.47 ± 38.33	–23.84 ± 33.05	141.73 ± 50.67	187.07 ± 51.60
BGLmin (%)	22.78 ± 1.92	97.92 ± 1.23	93.60 ± 3.17	44.90 ± 6.39	34.67 ± 10.01
Tmin (h)	3.20 ± 0.84	2.75 ± 0.25	1.08 ± 0.74	1.50 ± 0.35	1.40 ± 0.22
PA (%)	–	–1.01	–0.54	3.23	2.13

Data are expressed as mean ± standard deviation (n = 5–7). INS, insulin; SMEDDS, self-microemulsifying drug delivery system; SC, subcutaneous; AAC_0–8h: area above the blood glucose level curve (% of initial) during 0–8 h; BGL_min: minimum blood glucose level (% of initial levels); T_min: time to reach BGL_min; and PA: relative pharmacological availability versus SC injection of INS.

Table 7. Analysis of variance results of pharmacodynamics parameters.

Parameter	DF	Sum of square	Mean square	F value	p value
AAC0–8h
Administration	4	819,692	204,923	120.77	<.001
Residual	22	37328	1697	–	–
BGLmin
Administration	4	26,613.8	6,653.44	231.94	<.001
Residual	22	631.1	28.69	–	–

AAC_0–8h: area above the blood glucose level curve (% of initial) during 0–8 h; BGL_min: minimum blood glucose level (% of initial levels); DF, degrees of freedom.