Resveratrol loaded functionalized nanostructured lipid carriers for breast cancer targeting: Systematic development, characterization and pharmacokinetic evaluation

Colloids and Surfaces B: Biointerfaces 181 (2019) 756–766

Contents lists available at ScienceDirect

Colloids and Surfaces B: Biointerfaces journal homepage: www.elsevier.com/locate/colsurfb

Resveratrol loaded functionalized nanostructured lipid carriers for breast cancer targeting: Systematic development, characterization and pharmacokinetic evaluation

T

Neelam Pooniaa,b, Jasjeet Kaur Narangc, Viney Latherd, Sarwar Bege, Teenu Sharmae, ⁎ Bhupinder Singhe,f, Deepti Panditag, a

Department of Pharmaceutics, Jan Nayak Ch. Devi Lal Memorial College of Pharmacy, Sirsa, 125055, Haryana, India I. K. Gujral Punjab Technical University, Jalandhar, Punjab, India Department of Pharmaceutics, Khalsa College of Pharmacy, Amritsar, Punjab, India d Amity Institute of Pharmacy, Amity University, Sector-125, Noida, 201313, India e University Institute of Pharmaceutical Sciences, UGC-Centre of Advanced Studies, Panjab University, Chandigarh, 160 014, India f UGC-Centre of Excellence in Applications of Nanomaterials, Nanoparticles and Nanocomposites, Panjab University, Chandigarh, 160 014, India g Amity Institute of Molecular Medicine & Stem Cell Research (AIMMSCR), Amity University, Sector-125, Noida, 201313, India b c

ARTICLE INFO

ABSTRACT

Keywords: Resveratrol Nanostructured lipid carriers Optimization Breast cancer Targeting

Resveratrol (RSV) has shown to possess anti-cancer potential in various studies; however, its poor water solubility, extensive first-pass metabolism, and photostability issues have limited its clinical application. Therefore, the aim of the current investigation was to formulate and optimize a nanostructured lipid carriers (NLCs) based parenteral formulation of RSV for its effective delivery to breast cancer cells. NLCs loaded with RSV (RSV-NLCs) were formulated by the modified solvent injection technique and were systematically optimized using a three level-three factor Box-Behnken design. The optimized RSV-NLCs exhibited an optimum particle size of 88.3 ± 3.1 nm and high entrapment efficiency of 88.0 ± 2.6%. These optimized NLCs were further investigated for the targeting potential using folic acid as the targeting moiety and cell cytotoxicity experiments revealed high cytotoxic effects of folate modified NLCs (RSV-FA-NLCs) compared to unmodified NLCs on MCF-7 cells with high levels of over-expressed folate receptors suggesting the high potential of targeted NLCs in enhancing the therapeutic concentration of RSV to breast cancer cells. In vivo pharmacokinetic studies demonstrated a nine-fold increase in AUC values obtained with RSV-FA-NLCs (57.92 ± 4.15 μg h/mLh) in comparison to free RSV (6.37 ± 1.16 μg h/mLh). The promising results from this investigation corroborated the tremendous potential of lipidic nanocarriers in augmenting the therapeutic potential of RSV.

1. Introduction Breast cancer is a potentially fatal disease and remains a considerable cause of morbidity and mortality worldwide. American Cancer Society estimates approximately 268,670 new cases of breast cancer and 41,400 cancer deaths to occur in the United States in 2018 [1]. Conventional chemotherapeutic agents viz. cyclophosphamide, doxorubicin, paclitaxel, docetaxel, etc. used for the treatment of breast cancer are associated with severe side effects and leads to damage of normal cells [2]. In this regard, natural compounds having anti-cancer potential with wide accessibility and fewer adverse effects have been investigated as a new class of drugs [3]. Since last two decades, re-

⁎

searchers have identified various phytochemicals showing promising anti-cancerous properties owing to their antioxidant potential and their capability to interact with basic cellular mechanisms [4,5]. Resveratrol (RSV) is a natural phytoalexin found in various edible natural products like grapes, berries, red wine etc. and well known for its wide variety of health benefits such as anti-oxidant, anti-inflammatory, anti-mutagenic and anti-proliferative effects etc [6]. The chemopreventive activity of RSV was first identified by Jang and his coworkers in late nineties [7]. Indeed, RSV also possesses chemotherapeutic activities and numerous in vitro experiments focused on exploring the molecular mechanisms lying behind this action have demonstrated that this natural compound suppresses proliferation,

Corresponding author. E-mail address: [email protected] (D. Pandita).

https://doi.org/10.1016/j.colsurfb.2019.06.004 Received 23 February 2019; Received in revised form 14 May 2019; Accepted 3 June 2019 Available online 17 June 2019 0927-7765/ © 2019 Elsevier B.V. All rights reserved.

Colloids and Surfaces B: Biointerfaces 181 (2019) 756–766

N. Poonia, et al.

metasization, invasion and induces apoptosis in various cancer cell lines via interacting with different molecular targets such as P53, caspases-9, 8, 7, 3, P21, MAPK, MMPs-2, 7, 9, VEGF etc [8]. However, RSV has low water solubility and undergoes extensive hepatic first-pass metabolism which contributes to its poor absorption in vivo [9]. Also, the photolabile nature of RSV is another hurdle in its clinical application [10]. Therefore, formulation scientists are involved in exploring novel drug delivery systems which could provide maximum benefits of RSV as chemotherapeutic agent via overcoming its limitations. For instance, RSV loaded gelatin nanoparticles exhibited very rapid and more efficient cellular uptake, antiproliferative efficacy in NCI-H460 cells than free RSV [11]. In another study, Acrysol K150 based nanoemulsion of RSV showed improved in vitro cytotoxicity against MCF-7 cells as well as in vivo chick chorioallantoic membrane assay depicted the potentiation of its anti-angiogenic activity compared to free RSV [5]. Similarly, encapsulation of RSV in zein-pectin nanoparticles exhibited higher in vitro antioxidant activity and antiproliferative activity when tested using hepatocarcinoma cells compared to the free drug [12]. These nanotechnology-based delivery approaches relied on passive intracellular accumulation of drug from the developed nanocarrier owing to the enhanced permeation and retention (EPR) effect observed in the tumor microenvironment. However, in spite of wide applicability of passive targeting strategy, variability in EPR effect among different patients, tumor models and even in different areas of a particular tumor are observed [13–15]. Another strategy for drug delivery referred as “active targeting” has been accomplished which combats drawbacks of passive targeting approach. This strategy exploits the binding capability of receptors overexpressed in cancer cells to various ligands or antibodies attached on the surface of nanocarriers [15]. It implements accumulation of bioactives within cancer cells via receptor-mediated endocytosis, maximizing therapeutic efficacy and minimizing undesired toxic effects [16]. Folate receptor, highly overexpressed in various human malignancies such as breast cancer, ovarian, colon cancer etc., is one of the valuable therapeutic targets for active targeting. It has been extensively investigated by various researchers for its application in in vivo delivery of bioactives [17,18]. Systematic development of nanopharmaceutical products furnishing thorough understanding of their product and process performance requires implementation of the principles of Quality by Design (QbD) [19]. It facilitates the development of drug products with predefined objectives for attaining the target end-product performance. QbD requires use of statistical experimental designs for establishment of mathematical relationship between the critical material attributes and critical quality attributes, which further helps in the identification of optimum formulation [20]. The variables are initially subjected to factor screening study using first order experimental designs (like, fractional factorial designs, Taguchi design and Plackett-Burman design) to demarcate the important ones, followed by factor optimization study using central composite design using second-order experimental designs like Box-Behnken design, optimal design, etc [21,22]. In the present study, the elegance of exploiting folate receptor for targeting RSV to the breast cancer cells was investigated. For this, RSV loaded nanostructured lipid carriers (NLCs) were systematically optimized using the design of experiments to obtain the optimum formulation (RSV-NLCs). The folic acid-stearic acid conjugate was synthesized and employed to develop folate-targeted NLCs i.e. FA-RSVNLCs. The prepared nanoformulations were characterized for size, morphology and drug release profile. Cell viability study was done to evaluate the potential of the developed system against folate receptorpositive breast cancer cells. Further, pharmacokinetic studies were carried out in healthy Wistar rats after intravenous administration of developed NLCs and free drug to assess their pharmacokinetic parameters.

2. Material and methods 2.1. Materials RSV was obtained from Biotivia LLC, USA. Stearic acid was resourced from Merck Schuchardt, Germany. DPPH (1,1-diphenyl-2-picrylhydrazyl), oleic acid and folic acid were supplied by Sigma Aldrich, USA. Poloxamer 188 was provided by Pluronic® BASF Corp., USA. Phospholipon® 90 G was received as gift sample from Lipoid GMBH, Germany. D-trehalose dihydrate was procured from Sisco Research Laboratories Pvt. Ltd., India. Dialysis cellulose membrane (MW cut-off 12 kDa) and Amicon® Ultra-4 Centrifugal filter tubes were purchased from Sigma Aldrich, USA, and Millipore, Carrigtwohill Co., Ireland, respectively. 2.2. Screening of solid lipid The solubility of the drug in various solid lipids was estimated semiquantitatively by determining the melting point transition of drug mixed with the different lipids using differential scanning calorimetric (DSC) analysis technique as done by Gadgil et al. [23]. Physical mixtures of drug and with each of the lipid viz., stearylamine, precirol® ATO 5 and stearic acid in the ratio of 1:5 were sealed in standard aluminum pans and were screened between 30–350 °C at a heating ramp of 10 °C min−1 under nitrogen atmosphere using DSC-4000, Perkin Elmer, USA. 2.3. Synthesis of folic acid-stearic acid conjugate Referred to the previous literature, the synthesis of the folic acidstearic acid conjugate was carried out as reported by Yuan and coworkers [24]. Briefly, stearic acid dissolved in dimethylformamide was reacted for overnight with excess folic acid in the presence of 1-Ethyl-3(3-dimethylaminopropyl) carbodiimide and pyridine at room temperature. The synthesized conjugate was further precipitated after adding distilled water into the reaction mixture and the resultant dispersion was dialyzed against distilled water for consecutive three days, to remove unreacted products. Finally, the synthesized precipitate was filtered through a 0.45 μm millipore filter and freeze-dried for further use. The synthesized conjugate was subjected to Fourier transform infrared spectroscopic analysis (Shimadzu Corporation, Japan) using a disc of potassium bromide and 1H NMR (Avance III, 400 MHz Bruker, Germany) spectrum of sample dissolved in ˜5 wt.% (CD3)2SO was also obtained for analysis. 2.4. Preparation of nanostructured lipid carriers (NLCs) RSV-NLCs were formulated using a modification of the commonly used solvent injection technique described elsewhere [25,26]. Briefly, a varied amount of solid lipid i.e. stearic acid along with 25 mg of oleic acid (liquid lipid), 25 mg of Phospholipon® 90 G and 10 mg of RSV were dissolved in 5 ml of ether and were rapidly injected through an injection needle (30G1/2 Precision Glide needle) into a magnetically stirred (100 rpm) 20 ml aqueous solution at 40 ± 2 °C containing 1.5% w/v poloxamer in order to form lipid dispersion. The mixed system was allowed to evaporate at 40 ± 2 °C and was further homogenized (Heidolph instruments GmbH and Co. KG, Germany) at varied speeds for 1 h. The obtained RSV-NLCs dispersion was lyophilized for 24 h after addition of cryoprotectant solution (15% w/v trehalose). Folic acid modified NLCs (RSV-FA-NLCs), were prepared in a similar manner by replacing stearic acid by the folic acid-stearic acid conjugate.

757

Colloids and Surfaces B: Biointerfaces 181 (2019) 756–766

N. Poonia, et al.

Table 1a Results obtained during factor screening studies on RSV-loaded NLCs using a 11-factor-12-run Plackett-Burman Design layout. Factor 1

Factor 2

Factor 3

Factor 4

Factor 5

Factor 6

Factor 7

Factor 8

Factor 9

Factor 10

Factor 11

Particle size (nm)

EE (%)

1 −1 −1 −1 1 −1 −1 1 1 1 1 −1

1 −1 −1 1 −1 −1 1 −1 1 1 −1 1

−1 −1 1 1 1 −1 −1 −1 1 −1 1 1

−1 1 −1 1 1 −1 1 −1 −1 1 1 −1

−1 −1 1 −1 1 −1 1 1 −1 1 −1 1

1 1 1 −1 −1 −1 −1 −1 −1 1 1 1

−1 1 −1 −1 −1 −1 1 1 1 −1 1 1

1 −1 1 1 −1 −1 1 1 −1 −1 1 −1

1 1 1 −1 1 −1 1 −1 1 −1 −1 −1

−1 1 1 1 −1 −1 −1 1 1 1 −1 −1

1 1 −1 1 1 −1 −1 1 −1 −1 −1 1

177 ± 18.68 175 ± 11.42 432 ± 9.91 290 ± 12.23 87 ± 3.22 447 ± 8.78 251 ± 9.67 91 ± 4.45 162 ± 7.48 441 ± 17.12 178 ± 10.22 136 ± 11.56

46.6 24.2 14.7 98.2 28.2 12.1 94.6 87.3 94.1 44.5 86.7 96.1

Description of factors

Coding by design

1 2 3 4 5 6 7 8 9 10 11

Concentration of soya lecithin (10-40 mg) Homogenization rate (15,000-25,000 rpm) Homogenization time (30 min-90 min) Stirring rpm (50-350 rpm) Phase ratios (1:3 to 1:6) Quantity of cryoprotectant (2-15%) Time of agitation (15- 45 min) Amount of poloxamer (0.1-2%) Injection (instant or dropwise) Amount of liquid lipid (40-5 mg) Amount of solid lipid (95-60 mg)

A B C D E F G H J K L

1.45 2.08 1.64 1.98 2.17 1.09 3.89 1.12 2.21 3.23 4.67 2.21

of the central point. The independent variables of experimental design and their values at coded levels and the composition of various formulations are presented in Table 2a.

Table 1b Various variables employed during factor screening and their description. Factor

± ± ± ± ± ± ± ± ± ± ± ±

2.7. Optimization and validation Numerical optimization technique was employed to obtain the optimum NLCs formulation having a minimum particle size and maximum percent entrapment efficiency. Further, validation of the chosen response surface methodology was accomplished by developing eight check-point formulations and comparing their observed responses with the predicted values using linear correlation plots. 2.8. Characterization of developed NLCs

2.5. Factor screening studies

2.8.1. Particle size and zeta potential analysis NLCs samples were diluted with unionized water and were further subjected to particle size and zeta potential measurement using a Zetasizer (Malvern, United Kingdom). Measurements were done in triplicate at a scattering angle of 90°.

An eleven-factor and twelve-run study containing Plackett-Burman design (PBD) was employed for factor screening for identifying the important formulation and process variables critically influencing the product quality. Table 1a and 1b illustrates the design matrix as per the selected design along with description of low and high levels for each of the studied material attributes (MAs) and process parameters (PPs). Linear polynomial models generated using the design were analyzed to delineate the main effects for each of the studied response variables through half-normal plots and Pareto charts.

2.8.2. Percent entrapment efficiency and drug loading efficiency The amount of free drug in the NLCs dispersion was determined by ultrafiltration of the system using Amicon® Ultra-4 centrifugal filter tubes as described by Pandita et al. [10]. Briefly, NLCs dispersions were transferred to the centrifugal filter tubes and were centrifuged at 10,000 rpm (5590 × g) for 1 h. The obtained filtrate was diluted in methanol in a ratio of 50:50 and was injected into the HPLC system (Shimadzu LC-2010C HT, Japan) equipped with LC-20AT pump and SPD-20A UV/Vis detector and the chromatographic separation was performed as described in our previous work [10]. The percent entrapment efficiency and drug loading efficiency were calculated using the following equations:

2.6. Systematic optimization using Box-Behnken design In the present study, response surface methodology was employed to determine the effect of three independent formulation variables including the amount of solid lipid (X1), time of agitation (X2) and homogenization rate (X3) on the particle size (Y1) and percent entrapment efficiency (Y2) of RSV-NLCs. Using 33 (3 variables-3 levels) BoxBehnken design, a set of 15 runs was designed including 12 points representative of midpoints of each edge of the cube and three replicates

Entrapment efficiency (%) Total weight of drug added Weight of drug unentrapped = × (100) Total weight of drug added

Table 2a Independent and dependent variables of Box-Behnken design for optimization of RSV-NLCs. Independent variables

X1: Amount of solid lipid (mg) X2: Homogenization speed (rpm) X3: Time of agitation at magnetic stirrer (min.) Dependent variables Y1: Particle size (nm) Y2: Entrapment efficiency

Drug loading efficiency (%) Total weight of drug added Weight of drug unentrapped = × 100 Total weight of nanoparticles

Levels Low (-1)

Medium (0)

High (1)

60 15,000 15

77.5 20000 30

95 25,000 45

2.8.3. Transmission electron microscopy Transmission electron microscopy (Morgagni 268D, FEI Co., The Netherlands) was used for elucidation of the morphology of RSV-FANLCs. The diluted formulation was deposited on a copper grid followed by negative staining with 2% w/v phosphotungstic acid. Images were acquired via observing dried sample under the microscope at the 200 kV efficient voltage.

Constraints Minimize Maximize

758

Colloids and Surfaces B: Biointerfaces 181 (2019) 756–766

N. Poonia, et al.

2.8.4. DSC analysis Thermograms of RSV, stearic acid, physical mixture, and the freezedried RSV-NLCs and RSV-FA-NLCs were obtained using DSC-4000 Perkin Elmer, USA. The samples were separately sealed in aluminum pans and were scanned between 30–350 °C at a heating rate of 10 °C min−1 under an inert nitrogen atmosphere (50 mL min−1).

3. Results and discussion Lipid based nanosystems such as solid lipid nanoparticles and NLCs have attracted researchers’ interest owing to their small size, capability of easy scale up and use of physiological lipids during their fabrication [31]. Citing these advantages, these nanosystems have been reported in literature for improving RSV’s light stability, solubility and oral bioavailability [10,32,33]. However, as per author’s knowledge it is the first report wherein systematic optimization of RSV loaded NLCs employing QbD approach have been carried out for active targeting of developed nanocarrier to folate receptor positive cancer cells.

2.9. In vitro release studies The in vitro release of RSV from developed RSV-NLCs and RSV-FANLCs was determined by dialysis method [27]. Briefly, freeze-dried NLCs equivalent to 1 mg of RSV was redispersed in 1 ml PBS (pH 7.4) and then the dispersion was placed in a pre-activated dialysis membrane (MW cut off: 12,000 Da). Dialysis bag was immersed into 200 mL PBS (pH 7.4) which was continuously stirred on a magnetic stirrer at a speed of 100 rpm and kept at 37 °C. At different time intervals, aliquots of 1 mL were withdrawn and replenished with fresh medium to maintain the sink conditions. The amount of RSV was assessed using HPLC and the mechanism by which RSV was released from these lipid-based nanosystems was predicted.

3.1. Selection of lipid components The solubility of the drug in the lipid carrier is known to be an important precondition to obtain high percent entrapment efficiency [34]. Solubility of RSV in stearic acid; oleic acid and mixture of solid lipid and liquid (in ratio of 3:1) was also assessed and found to be 112.05 ± 1.87; 18.13 ± 0.25 and 98.09 ± 0.98 mg/g respectively. Further preliminary studies were carried out to select the most appropriate solid lipid core using DSC technique for formulating RSV-NLCs with desired percent entrapment efficiency. Fig. S1 (Supplementary data) represents the overlay of DSC thermograms of pure RSV, physical mixtures of RSV with stearylamine, precirol® ATO 5 and stearic acid. Although the characteristic endothermic peak corresponding to the melting point of RSV (261 °C) was evident in the physical mixtures with stearylamine and precirol® ATO 5 but only a hump was observed with stearic acid. This could be due to higher miscibility of RSV in stearic acid compared to other lipids and thus it was selected for further investigations. In addition, organic solvent was also employed in the formulation of NLCs to facilitate the solubility. Oleic acid was selected as liquid lipid since RSV, stearic acid and oleic acids were soluble in common solvent ether.

2.10. Cell cytotoxicity The cell cytotoxicity of pure RSV, RSV-NLCs, and RSV-FA-NLCs along with pure paclitaxel (reference) was assessed against MCF-7 and A549 cell lines using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT) assay [28]. Briefly, both MCF-7 and A549 cells (5 × 105) were cultured separately in 96-well plates in serum DMEM at 37 °C for 24 h incubation period. The DMEM was replaced with serumfree DMEM after 24 h and the cells were then exposed to various test samples of different concentrations (2–64 μg/mL) for 72 h. Then MTT (0.5 mg/mL) was added and the plates were allowed to incubate for 4 h at 37 °C. Finally, 100 μL of DMSO was added to dissolve the developed formazan crystals and the absorbance was recorded at 570 nm using 630 nm as reference.

3.2. Optimization of NLCs Preliminary batches of RSV-NLCs were prepared via solvent injection technique to identify possible factors having the significant influence on particle size and percent entrapment efficiency of developed nanoparticles. Based on the results of preliminary formulation studies (Placket Burmann design), three factors i.e., amount of solid lipid, homogenization speed and time of agitation were found to have great influence on the properties of RSV-NLCs by solvent injection technique, as depicted by Pareto charts of both the responses (Fig. 1). Further, BoxBehnken design was utilized for optimizing process variables and the experimental runs and the observed responses for the fifteen formulations are presented in Table 2b. Based on the experimental runs, the factor combination resulted in different size and percent entrapment efficiency values ranging from 87.25 to 412.4 nm and 64.02 to 98.04%, respectively. The data were analyzed with Design Xpert software 10.0 and quantitative effects of the three influencing factors at different levels on both responses i.e. particle size and percent entrapment efficiency could be predicted using following polynomial equations, respectively:

2.11. Pharmacokinetic studies Pharmacokinetics studies were performed in female Wistar rats (220–240 g). The experimental protocol was carried out according to Committee for the Purpose of Control and Supervision of Experiments on Animals, India and was duly approved by the Institutional Animal Ethics Committee, Jan Nayak Ch. Devi Lal Memorial College of Pharmacy, India (approval no.: JCDMCOP/IAEC/06/16/34). The healthy disease free animals were procured from Lala Lajpat Rai University of Veterinary and Animal Sciences, Hisar, India. The pharmacokinetic profile of intravenously administered free RSV, RSV-NLCs and RSV-FA-NLCs at a dose 2 mg/kg body weight was determined [29]. For blood sampling, sparse sampling methodology reported by Nair et al., was followed shown in Table S1, Supplementary data [30]. Various groups received different solutions as per the following administration outline: group 1, RSV solution; group 2 and 3, RSV-NLCs and RSV-FA-NLCs dispersion respectively, by intravenous bolus injection via the tail vein. Free drug solution was prepared in a mixture of ethanol and polysorbate 80 (1:1). Blood samples of 400 μL per sampling point were collected in heparinized capillary tubes from retro-orbital sinus under mild anesthesia at predetermined intervals and plasma was separated via centrifugation at 3000 rpm for 10 min. Extraction of the drug from plasma samples was carried out by solid phase extraction technique as reported in our previous article [10]. The pharmacokinetic parameters were calculated using Kinetica software 5.0. ANOVA was performed to evaluate differences between the treatments and p-value < 0.05 was considered to be significant.

Y1=+94.50-27.44X1-21.03X2+120.98X3+8.49X1X2-1.86X1X37.63X2X3 32.45X12+118.16X22+50.11X32+65.37X12X2-123.54 X12X3 Y2=+84.35 + 4.00X1+4.52X2-4.52X3-1.41X1X2-4.44X1X30.51X2X3+10.89X12-9.23X22-2.57X32-8.77X12X2+7.46 X12X3 Summary of the results of the regression analysis of observed responses are presented in Table 2c indicating high statistical significance for parameters, size (< 0.0035) and percent entrapment efficiency (< 0.0121). 759

Colloids and Surfaces B: Biointerfaces 181 (2019) 756–766

N. Poonia, et al.

Fig. 1. Depiction of half-normal plots along with Pareto charts of particle size (A) and percent entrapment efficiency (B) of RSV-NLCs as per Plackett Burmann design.

3.2.1. Response surface analysis The response surface analysis was carried out for studying the cause-and-effect relationship between independent and dependent variables. Fig. 2A and 2B illustrate the 3D-plots for the response variables i.e. particle size and percent entrapment efficiency, respectively with the presence of interactions among the factors. The 3D-response surface plot for particle size indicated the curvilinear relationship between the concentration of solid lipid and homogenization rate. At higher levels of homogenization rate, an increase in the concentration of solid lipid showed an initial decrease in the values of particle size up to the intermediate levels, followed by an increasing trend at high levels. Increase in size may be due to the higher

Table 2c Results of statistical analysis for particle size and percent entrapment efficiency. Source

Sum of squares

df

Mean square

F value

p-value

Y1 Regression residuals Total R2=0.9951 Y2 Regression residuals Total R2=0.9887

1.399E+005

11

12721.96

55.40

0.0035

1327.97

11

120.72

23.80

0.0121

Table 2b Factor optimization of RSV-NLCs showing response variables corresponding to formulations prepared as per 33 Box-Behnken design. Formulation code

Amount of solid lipid

Homogenization rate

Time of agitation

Particle size (nm)

Entrapment efficiency (%)

F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15

77.5 60 60 95 77.5 95 77.5 77.5 95 60 77.5 77.5 95 60 77.5

25,000 20000 15,000 25,000 20000 20000 25,000 20000 20000 25,000 15,000 20000 15,000 20000 15,000

15 45 30 30 30 15 45 30 45 30 45 30 30 15 15

128.4 ± 13.41 130.4 ± 12.22 180.3 ± 14.17 197.1 ± 13.91 87.25 ± 14.13 355.1 ± 22.32 97.65 ± 11.41 102.12 ± 11.46 252.0 ± 18.71 88.80 ± 10.86 412.4 ± 24.94 94.14 ± 10.12 191.44 ± 8.97 131.8 ± 12.34 155.2 ± 14.31

82.1 ± 1.14 96.2 ± 2.13 84.7 ± 1.75 84.5 ± 1.98 81.53 ± 5.13 98.04 ± 2.1 72.05 ± 1.05 87.01 ± 7.54 95.04 ± 2.59 79.03 ± 1.42 64.02 ± 1.13 84.53 ± 6.75 95.81 ± 1.14 81.43 ± 1.07 72.04 ± 0.98

760

Colloids and Surfaces B: Biointerfaces 181 (2019) 756–766

N. Poonia, et al.

Fig. 2. Response surface plots for the effect of different factors on (A) particle size, and (B) percent entrapment efficiency.

viscosity of lipid-solvent diffusion phase which hindered diffusion rates of solute molecules as observed previously by Hejri et al. [35]. Also, increased collisions and aggregation phenomenon in higher lipid phase could be another reason for the development of larger sized particles [36]. Similarly, at higher levels of solid lipid concentration, particle size exhibited a declining pattern with an increase in homogenization speed, followed by a sharply increasing trend. The initial declining pattern may be observed owing to good homogenization of emulsion at the high rate of shear. But after an optimal level of homogenization speed, further increase in homogenization rate might have resulted in flocculation leading to the formation of higher sized NLCs [10,37]. Fig. 2A illustrates a “rising ridge” shaped response surface plot between the solid lipid content and the time of agitation, where a hump was observed at the high level of time of agitation with an increase in the solid lipid content from low to high levels. Moreover, the response surface graph between homogenization rate and time of agitation illustrates a linear increase in the values of particle size with increase in time of agitation both at extremely low and high levels of homogenization rate. The 3D-response surface plots for entrapment efficiency are depicted in Fig. 2B and it was observed that at low levels of homogenization rate, increase in the values of solid lipid concentration indicated a slightly declining trend, followed by an increasing pattern.

The increase in the percent entrapment efficiency may be attributed to the increase in the availability of lipophilic ambiance which leads to the accommodation of higher drug molecules and also reduces the escaping tendency of drug molecules in the external phase [38]. On the contrary, entrapment efficiency was found to initially rise and further depicted a declining trend in increasing the values of homogenization rate from low to high levels. Fig. 2B depicts the relationship between the concentration of solid lipid and time of agitation, where mild to the moderate influence of both the factors were observed on the values of entrapment efficiency. The influence between homogenization rate and time of agitation was also investigated. A curvilinear response was seen with maximal values of entrapment efficiency at intermediate levels of both independent variables. 3.2.2. Search for the optimum formulation and validation of the experimental design The optimum NLC formulation was determined with the assistance of numerical optimization method employing desirability function technique by “trading off” various responses with the goals, i.e., minimization of particle size and maximization of the entrapment efficiency, to obtain the desirability function value close to 1 (Table 2d). Fig. 3 represents the overlay plot indicating yellow colour region as the design

Table 2d Optimum solutions obtained using numerical optimization with different values of desirability function. Factor A (Amount of solid lipid)

Factor B (Homogenization rate)

Factor C (Time of agitation)

Particle size

Entrapment efficiency (%)

Desirability function value

84 95 60

19735 15874 19094

31 21 45

86.87 87.25 137.66

86.88 97.39 96.36

0.998 0.919 0.791

761

Colloids and Surfaces B: Biointerfaces 181 (2019) 756–766

N. Poonia, et al.

Fig. 3. Overlay plot with the design space demarcated by yellow region presenting the optimized RSV-NLCs by a flag. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article).

space with flagged point as the optimum formulation with its composition and responses. A solid lipid concentration of 84%, homogenization speed of 20,000 rpm and agitation time of 31 min resulted in RSVNLCs with particle size and percent entrapment efficiency of 88.31 nm and 87.99% respectively. Further, validation studies were performed by preparing eight check-point formulations and comparing observed responses with the predicted values using linear correlation plots. The values of percent prediction error for the response variable ranged between 6.38 and 3.33%, with overall mean ± SD as 0.13% ± 2.19, while the values of R ranged between 0.982 and 0.999, thus warranting excellent goodness of fit of the data (p < 0.001). Moreover, the residual plots also indicated quite regulated pattern of data distribution indicating the high degree of prognosis of the employed experimental design approach.

size distribution with minimum tendency to agglomerate [40]. High values of percent entrapment efficiency, i.e., 87.99 ± 2.56 and 86.34 ± 1.98% were obtained for RSV-NLCs and RSV-FA-NLCs, respectively. The zeta potential values i.e. -31.27 ± 4.11 and -42.15 ± 3.76 mV for RSV-NLCs and RSV-FA-NLCs respectively indicated that prepared NLCs had sufficient charge to inhibit aggregation due to electric repulsion as reported elsewhere [38]. Likewise, the lipid based nanoparticles developed by Neves and his co-workers with zeta potential around −30 mV depicted high physical stability in long term stability studies [32]. TEM micrograph showed that the majority of produced NLCs possessed relatively spherical shape with no visible signs of aggregation (Fig. 4A). In addition, NLCs presented sizes below 100 nm, which were in good agreement with Malvern Zetasizer measurements. Thermal analysis of pure RSV; the physical mixture of stearic acid and RSV; RSVNLCs and RSV-FA-NLCs was performed for characterization of druglipid interactions and crystallinity of RSV before and after formulation in nanocarrier matrices. The pure RSV and the physical blend of lipid and drug showed a sharp endothermic peak at 265.43 °C corresponding to the melting point of drug representing the crystalline nature of RSV. In contrast, lyophilized RSV-NLCs and RSV-FA-NLCs did not show the endothermic peak (Fig. 4B). The absence of melting endotherm of RSV in lyophilized NLCs suggests its conversion to the amorphous state and anticipates its complete encapsulation in the lipid matrix.

3.3. Physiochemical characterization of optimized NLCs The physiochemical characterization of RSV-NLCs and RSV-FANLCs produced based on the optimized parameters, was conducted in terms of average hydrodynamic diameter, polydispersity index (PDI) and zeta potential (Table 3a). RSV-FA-NLCs depicted smaller particle size compared to the RSV-NLCs owing to the enhanced hydrophilicity of their surface due to the presence of stearic acid-folic acid conjugate. A similar trend was also observed by Zhang et al. [39]. Developed NLCs depicted low PDI values, i.e., 0.298 and 0.375, indicating homogenous

Table 3a Values of particle size, PDI and zeta potential of optimized RSV-NLCs and RSV-FA-NLCs. Formulation

Particle size (nm)

PDI

Zeta potential (mV)

Entrapment efficiency (%)

Drug loading (%)

RSV-NLCs RSV-FA-NLCs

88.31 ± 3.12 82.44 ± 2.78

0.298 ± 0.219 0.375 ± 0.124

−31.27 ± 4.11 −42.15 ± 3.76

87.99 ± 2.56 86.34 ± 1.98

8.54 ± 0.67 8.05 ± 0.75

*PDI: polydispersity index. 762

Colloids and Surfaces B: Biointerfaces 181 (2019) 756–766

N. Poonia, et al.

i.e., high drug release upto 35.12 and 28.17%, respectively in the first 2 h followed by sustained release upto 99.95 and 97.65%, till 72 h, respectively (Fig. 4C). This demonstrates that the produced nanoformulations exhibited an initial burst release followed by a sustained release phase as observed in various previous NLCs based investigations [41,42]. The initial burst release could be ascribed to the drug adsorbed on the surface of nanoparticles or to the presence of liquid lipid in the outer shell containing a dissolved drug which is released initially while the drug within lipid matrix is released gradually for a longer period of time. The release profile of free RSV was also examined which showed upto 49.37% in 2 h and 98.08% release in 12 h. The slower release profile of RSV from NLCs could be due to its lipophilic nature facilitating higher affinity for lipid material [38], and thus demonstrates NLCs as a suitable carrier for sustained drug delivery of RSV. The release data were also analyzed using different mathematical models i.e. zero order, first order, Higuchi and Korsmeyer–Peppas by plotting the cumulative percent drug release versus time, log cumulative percent drug remaining versus time, cumulative percent drug release versus the square root of time and log cumulative percent drug release versus log time respectively, to define the release kinetics and was well fitted with the Korsmeyer-Peppas model, high r2 values of 0.904 and 0.984 were obtained for RSV-NLCs and RSV-FA-NLCs, respectively depicting good linearity. The release exponent “n” was observed to be less than 0.5 which appears to indicate the Fickian diffusion phenomenon which occurs by usual molecular diffusion of the drug [43]. 3.5. Cytotoxicity The standard cytotoxicity assay of free paclitaxel (anti-cancer drug), free RSV, unmodified and RSV loaded modified NLCs against MCF-7 and A549 cells were carried out in order to assess the sensitivity towards folate receptor positive and negative cancer cells. The proliferation inhibition effect was found to increase in concentration-dependent manner for 72 h when treated with different formulations (Fig. 5A and 5B). The obtained results of IC50 values depicted that RSVFA-NLCs and RSV-NLCs proved to be significantly more cytotoxic (p < 0.05) to both cells than free RSV (Table 3b). Higher anti-proliferative effect of NLCs is attributed to the membrane affinity of lipidic components and their small size providing enhanced endocytotic uptake within tumor cells compared to the passive diffusion of free RSV through the cell membrane [39]. Also, sustained release of drug from nanoparticles facilitating constant exposure of drug to cancer cells could possibly be another reason for the higher activity. In a previous study, RSV loaded solid lipid nanoparticles have demonstrated cytostatic effect against NCTC2544 cell lines owing to sustained release from its carrier with potential benefits in skin cancer [27]. After 72 h of incubation, the cell cytotoxicity of free paclitaxel on MCF-7 cells was significantly greater than the anti-proliferative effect of free RSV, RSV-NLCs, and RSV-FA-NLCs (P < 0.05); also, RSV-FA-NLCs contributed to the higher reduction in cell viability than RSV-NLCs. The greater cytotoxic response of RSV-FA-NLCs compared to RSV-NLCs is presumably due to sustained release of RSV in the extracellular compartment and direct receptor-ligand interaction involving folate receptors overexpressed in MCF-7 cells which promoted internalization process [44]. It was also evident that although RSV-FA-NLCs exhibited stronger cytotoxic effects towards MCF-7 cells; its effects were lesser on A549 cells with negligible expression of folate receptors. The obtained results were consistent with levels of folate receptor expression in the tested cell lines and the difference in response of folate modified formulation towards both cells could be attributed to the non-existence of ligand-receptor interaction in case of A549 cells leading to decreased cytotoxicity. Although, the results of RSV-FA-NLCs were not comparable to standard anti-cancer drug paclitaxel but it could be concluded that the nanoformulation was highly cytotoxic compared to free RSV. Thus, from obtained results, it could be deduced that RSV-FA-NLCs demonstrated superior cytotoxic effect against MCF-7 cells via receptor-

Fig. 4. (A) TEM micrograph of RSV-FA-NLCs; (B) DSC thermograms of RSV (a), RSV-FA-NLCs (b), RSV-NLCs (c) and RSV with stearic acid (d); (C) In vitro drug release profile of RSV from different formulations.

3.4. In vitro release studies To predict the release profile upon systemic administration, in vitro release study of RSV from RSV-NLCs and RSV-FA-NLCs was performed at physiological conditions (pH 7.4). From the obtained results, it was evident that RSV-NLCs and RSV-FA-NLCs exhibited initial burst release, 763

Colloids and Surfaces B: Biointerfaces 181 (2019) 756–766

N. Poonia, et al.

Fig. 5. The in vitro cell viability of (A) MCF-7 and (B) A549 cells at various drug concentrations of the drug after 72 h treatment. (C) Plasma concentration-time profile of RSV in rats with intravenously administered free RSV, RSV-NLCs and RSV-FA-NLCs; inset shows representative HPLC chromatogram of RSV in rat plasma.

Table 3b Cytotoxicity of free drug and NLCs-based formulations demonstrated using MCF-7 and A549 cells. Formulation

Free Paclitaxel Free RSV RSV-NLCs RSV-FA-NLCs

Table 3c Pharmacokinetic parameters of free RSV, RSV-NLCs and RSV-FA-NLCs using in vivo studies in rats.

IC50 values MCF-7 (μg/mL)

A549 (μg/mL)

5.2 27.3 21.9 11.1

4.1 26.4 21.2 18.4

Parameters

Free RSV

RSV-NLCs

RSV-FA-NLCs

Cmax (μg/mL) Clearance (L/h) t1/2 (h) MRT (h) AUC(0-t) (μg/mLh)

4.01 ± 0.54 0.379 ± 0.05 0.98 ± 0.10 1.85 ± 0.11 6.37 ± 1.16

3.52 ± 0.21 0.053 ± 0.004 10.38 ± 0.76 17.52 ± 0.98 27.11 ± 3.92

4.19 ± 0.421 0.028 ± 0.003 12.04 ± 0.83 19.75 ± 1.14 57.92 ± 4.15

3.6. Pharmacokinetic studies

mediated endocytosis mechanism and could be employed as the effective delivery vehicle for treatment of folate receptor-positive cells. Likewise, in a previous study folate receptor-targeted NLCs developed for effective delivery of curcumin to breast cancer cells depicted active targeting ability and were highly cytotoxic to MCF-7 cells [45]. From the results of the present study, it could be concluded that developed RSV-FA-NLCs had the high potential for further studies, towards the treatment of breast cancer, owing to its specific targeting potential.

In the present study, sparse sampling methodology was employed for obtaining blood samples wherein complete profile is not sampled from a single animal and each animal accounted for three to four blood samples only. This methodology ensures adequate recovery of rodents after blood sampling owing to sufficient long time interval between successive sampling from each animal [30]. Blood samples were obtained and the amount of RSV in the plasma at different time intervals

764

Colloids and Surfaces B: Biointerfaces 181 (2019) 756–766

N. Poonia, et al.

was calculated using the calibration curve; y = 72029x-169.2 (R2 = 0.998) and the obtained chromatogram is depicted Fig. 5C (inset). Fig. 5D depicts the plasma concentration-time profile of RSV after single dose intravenous administration of free RSV, RSV-NLCs, and RSV-FA-NLCs and it was observed that RSV concentration was measurable till 48 h in case of RSV-NLCs and RSV-FA-NLCs, whereas free RSV almost completely disappeared within 6 h. Pharmacokinetic parameters also implicated that RSV-FA-NLCs exhibited the highest area under curve (AUC) value followed by RSV-NLCs, and RSV (Table 3c). The extended plasma RSV profile observed in case of NLCs could be due to the slow release of RSV from lipid matrix, and also the presence of poloxamer 188 on its surface might have prevented their uptake by reticulo-endothelial system prolonging their residence time in systemic circulation. Similarly, Vijayakumar and coworkers also reported a significantly higher AUC in case of RSV loaded stealth solid lipid nanoparticles compared to the free RSV after parenteral administration, by avoidance of reticulo-endothelial system [46]. The results of the present study also suggested that encapsulation of drug within carrier could lead to significant improvement in pharmacokinetic parameters. A difference in RSV plasma concentrations between both NLCs was also observed. In particular, RSV-FA-NLCs (57.92 ± 4.65 μg h/mLh) demonstrated nine-fold increase in AUC values in comparison to free RSV (6.37 ± 1.165 μg h/mLh), while RSV-NLCs (27.11 ± 3.92 μg h/ mLh) demonstrated four times increment in AUC value. Similarly, higher t1/2 and MRT i.e 12.04 ± 0.83 h and 19.75 ± 1.14 h, respectively was also observed in case of RSV-FA-NLCs compared to RSV-NLCs and free RSV. Elimination rate constant and clearance were also observed to be significantly (P < 0.05) reduced in comparison to that of RSV solution which further supports longer residence time of RSV-FANLCs within systemic circulation. The possible reason for the higher AUC, t1/2 and MRT could be the slower release of RSV from these NLCs compared to RSV-NLCs as depicted in in vitro release studies, which have accounted for lower elimination rate constant and clearance. A similar trend of improved pharmacokinetic parameters in case of solid lipid nanoparticles permitting slower release of clozapine was observed by Manjunath et al. [47]. The results of the present study suggested that RSV-FA-NLCs could postpone the elimination of RSV and could be an effective carrier for achieving high therapeutic efficacy.

Acknowledgments The authors are thankful to I. K. Gujral Punjab Technical University, Jalandhar (Punjab); and Management, JCD Vidyapeeth, Sirsa (Haryana) for their support and Biotivia LLC, USA for providing the gift sample of RSV. We are thankful to Guru Jambheshwar University of Science and Technology, Hisar and All India Institute of Medical Sciences, New Delhi for sample analysis support. Prof. Bhupinder Singh (Bhoop), gratefully acknowledges the generosity of M/s Stat-Ease Inc., Minneapolis, USA, for providing one perpetual licence and ten single annual licence of Design Expert® software, version 9.0, as an award for his unparalleled contribution in the area of QbD-based drug delivery research work. Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.colsurfb.2019.06.004. References [1] L.S. Rebecaa, D.M. Kimberly, J. Ahmedin, CA-Cancer J. Clin. 68 (2018) 7–30. [2] G.L. Finch, L.A. Burns-Naas, Encyclopedia of Toxicology, third ed., Academic Press, Oxford, 2014, pp. 630–641. [3] Q.S. Zhao, L.L. Hu, Z.D. Wang, Z.P. Li, A.W. Wang, J. Liu, Mater. Sci. Eng. C 72 (2017) 185–191. [4] M. Kampa, A.P. Nifli, G. Notas, E. Castanas, Rev. Physiol. Biochem. Pharmacol. 159 (2007) 79–113. [5] S. Pund, R. Thakur, U. More, A. Joshi, Colloids Surf. B Biointerfaces 120 (2014) 110–117. [6] V. Benetou, P. Orfanos, P. Lagiou, D. Trichopoulos, P. Boffetta, A. Trichopoulou, Cancer Epidemiol. Biomarkers Prev. 17 (2008) 387–392. [7] M. Jang, L. Cai, G.O. Udeani, K.V. Slowing, C.F. Thomas, C.W. Beecher, H.H. Fong, N.R. Farnsworth, A.D. Kinghorn, R.G. Mehta, R.C. Moon, J.M. Pezzuto, Science 275 (1997) 218–220. [8] M. Gonzalez-Vallinas, M. Gonzalez-Castejon, A. Rodriguez-Casado, A. Ramirez de Molina, Nutr. Rev. 71 (2013) 585–599. [9] N. Summerlin, E. Soo, S. Thakur, Z. Qu, S. Jambhrunkar, A. Popat, Int. J. Pharm. 479 (2015) 282–290. [10] D. Pandita, S. Kumar, N. Poonia, V. Lather, Food Res. Int. 62 (2014) 1165–1174. [11] S. Karthikeyan, N. Rajendra Prasad, A. Ganamani, E. Balamurugan, Biomed. Prev. Nutr. 3 (2013) 64–73. [12] X. Huang, Y. Dai, J. Cai, N. Zhong, H. Xiao, D.J. Mcclements, K. Hu, Food Hydrocolloid. 64 (2017) 157–165. [13] R.K. Jain, T. Stylianopoulos, Nat. Rev. Clin. Oncol. 7 (2010) 653–664. [14] Y.H. Bae, K. Park, J. Control. Release 153 (2011) 198–205. [15] S.P. Egusquiaguirre, M. Igartua, R.M. Hernández, J.L. Pedraz, Clin. Transl. Oncol. 14 (2012) 83–93. [16] J.E.N. Dolatabadi, Y. Omidi, Trends Analyt. Chem. 77 (2016) 100–118. [17] S. Xu, B.Z. Olenyuk, C.T. Okamoto, S.F. Hamm-Alvarez, Adv. Drug Deliv. Rev. 65 (2013) 121–138. [18] Y. Liu, Y. Pu, L. Sun, H. Yao, B. Zhao, R. Zhang, Y. Zhang, J. Biomed. Nanotechnol. 21 (2016) 1393–1403. [19] S. Beg, G. Sharma, K. Thanki, S. Jain, O.P. Katare, B. Singh, Int. J. Pharm. 493 (2015) 466–482. [20] P. Negi, B. Singh, G. Sharma, S. Beg, O.P. Katare, J. Microencapsul. 32 (2015) 419–431. [21] R. Kapil, S. Dhawan, S. Beg, B. Singh, Drug Dev. Ind. Pharm. 39 (2013) 466–480. [22] B. Singh, K. Raza, S. Beg, Chemical Ind. Digest 12 (2013) 70–76. [23] P. Gadgil, J. Shah, D.S.L. Chow, Int. J. Pharm. 544 (2018) 265–277. [24] H. Yuan, J. Miao, Y.Z. Du, J. You, F.Q. Hu, S. Zeng, Int. J. Pharm. 348 (2008) 137–145. [25] D. Pandita, A. Ahuja, V. Lather, B. Benjamin, T. Dutta, T. Velpandian, R.K. Khar, AAPS Pharm. Sci. Tech. 12 (2011) 712–722. [26] D. Pandita, A. Ahuja, T. Velpandian, V. Lather, T. Dutta, R.K. Khar, Pharmazie 64 (2009) 301–310. [27] K. Teskac, J. Kristl, Int. J. Pharm. 390 (2010) 61–69. [28] C. Chen, H. Hu, M. Qiao, X. Zhao, Y. Wang, K. Chen, X. Guo, D. Chen, Int. J. Pharm. 480 (2015) 116–127. [29] S. Bollam, P. Kandadi, S. Thumma, K. Veerabrahma, AAPS Pharm. Sci.Tech. 12 (2011) 422–430. [30] S.N. Nair, M. Mhatre, S. Menon, S. Shailajan, J. Pharmaceut. Biomed. 100 (2014) 190–198. [31] M.R. Gasco, Adv. Drug Deliv. Rev. 59 (2007) 377–378.

4. Conclusion As the poor pharmacokinetic profile of RSV tends to affect its therapeutic efficacy, the present study was successfully undertaken to prepare RSV loaded NLCs to optimize these systematically using response surface methodology. Further, folic acid-stearic acid conjugate was synthesized and folic acid-decorated NLCs were developed for active targeted delivery of RSV to cancer cells. The developed RSV-FANLCs offered promising perspectives in terms of solubility, sustained release, and targeting potential, assisting effective delivery of RSV to breast cancer cells. The results provide supplementary evidence that RSV-FA-NLCs exhibit targeting potential with great benefits in folate receptor-positive breast cancer cells. However, future studies must be carried out to ratify the therapeutic potential of the targeted NLCs in breast tumor-bearing animals. Declaration of Competing Interest The authors declare that they have no conflict of interest.

765

Colloids and Surfaces B: Biointerfaces 181 (2019) 756–766

N. Poonia, et al. [32] A.R. Neves, M. Lúcio, S. Martins, J.L. Lima, S. Reis, Int .J. Nanomedicine 8 (2013) 177–182. [33] A. Neves, S. Martins, M. Segundo, S. Reis, Nutrients 8 (2016) 131–138. [34] R.K. Subedi, K.W. Kang, H.K. Choi, Eur. J. Pharm. Sci. 37 (2009) 508–513. [35] A. Hejri, K. Khosravi, K. Gharanjig, M. Hejazi, Food Chem. 141 (2013) 117–123. [36] J. Zhang, Y. Fan, E. Smith, J. Pharm. Sci. 98 (2009) 1813–1819. [37] M. Almousallam, C. Moia, H. Zhu, Int. Nano Lett. 5 (2015) 241–248. [38] A.A. Abdelbary, M.H. AbouGhaly, Int. J. Pharm. 485 (2015) 235–243. [39] X.G. Zhang, J. Miao, Y.Q. Dai, Y.Z. Du, H. Yuan, F.Q. Hu, Int. J. Pharm. 361 (2008) 239–244.

[40] M. Ferreira, L.L. Chaves, S.A. Lima, S. Reis, Int. J. Pharm. 492 (2015) 65–72. [41] L. Jia, J. Shen, D. Zhang, C. Duan, G. Liu, D. Zheng, X. Tian, Y. Liu, Q. Zhang, Int. J. Biol. Macromol. 50 (2012) 523–529. [42] H. Shen, S. Shi, Z. Zhang, T. Gong, X. Sun, Theranostics 5 (2015) 755–771. [43] M. Ferreira, L.L. Chaves, S.A.C. Lima, S. Reis, Int. J. Pharm. 492 (2015) 65–72. [44] G. Zhang, F. Liu, E. Jia, L. Jia, Y. Zhang, Drug Deliv. 23 (2016) 1393–1397. [45] M. Lin, L. Teng, Y. Wang, J. Zhang, X. Sun, Drug Deliv. 23 (2016) 1420–1425. [46] M.R. Vijayakumar, K.Y. Vajanthri, C.K. Balavigneswaran, S.K. Mahto, N. Mishra, M.S. Muthu, S. Singh, RSC Adv. 6 (2016) 74254–74268. [47] K. Manjunath, V. Venkateswarlu, J. Control. Release 107 (2005) 215–228.

766