A robust systematic design: Optimization and preparation of polymeric nanoparticles of PLGA for docetaxel intravenous delivery

Materials Science & Engineering C 104 (2019) 109950

Contents lists available at ScienceDirect

Materials Science & Engineering C journal homepage: www.elsevier.com/locate/msec

A robust systematic design: Optimization and preparation of polymeric nanoparticles of PLGA for docetaxel intravenous delivery Pedram Rafieia,b, Azita Haddadia, a b

T

⁎

Division of Pharmacy, College of Pharmacy and Nutrition, University of Saskatchewan, Saskatchewan, Saskatoon, Canada Department of Pharmaceutical Sciences, Leslie Dan Faculty of Pharmacy, University of Toronto, Toronto, Canada

A R T I C LE I N FO

A B S T R A C T

Keywords: Poly (lactide-co-glycolide) (PLGA) Polymeric nanoparticles Experimental factorial design Taguchi design Docetaxel Emulsification-solvent-evaporation Poly (vinyl alcohol) (PVA)

Poly (lactide-co-glycolide) (PLGA) is a biocompatible, biodegradable, and non-toxic polymer used in a variety of biomedical and pharmaceutical applications. Polymeric nanoparticles prepared from PLGA have been extensively used as delivery vehicles of various chemotherapeutic agents. The variability of PLGA polymer and nanoparticle fabrication process potentially results in variability of particle characteristics. Nanoparticle characteristics determine nanoparticles' performance when used as drug delivery systems. Having control on nanoparticle's characteristics grants control over the fate of nanoparticles and the associated drug. Here, L16 Taguchi experimental design was used to evaluate the effect of polymer characteristics and fabrication variables on PLGA nanoparticles. The design was used to determine an optimized preparation condition for PLGA nanoparticles as an intravenous delivery system for docetaxel. An emulsification-solvent-evaporation method was used to fabricate nanoparticles. Docetaxel concentration, organic phase:aqueous phase ratio, polymer molecular weight, polymer terminus, lactide:glycolide ratio, and Poly(vinyl alcohol)(PVA) concentration were selected as main determinants. First two factors were evaluated at 4 levels and the rest at 2 levels. Particle-important characteristics including size, polydispersity index (PDI), surface charge (zeta potential), and docetaxel loadingefficiency were determined. Factors affecting nanoparticle characteristics were ranked according to level of effectiveness. Factors that affected nanoparticle properties with statistical significance were identified. Models to predict nanoparticle characteristics were built. An optimized fabrication method was identified and used to prepare PLGA nanoparticles for docetaxel delivery.

1. Introduction As a member of taxan family of antineoplastic agents, docetaxel exerts its cytotoxic effects on microtubules [1]. 10-deacetyl-baccatin III isolated from Taxus brevifolia or Taxus baccata (trees of Taxus family), is used to produce docetaxel semi-synthetically [2]. Docetaxel disrupts the network of microtubules which is necessary for mitotic cell division and contributes to inhibition of cell proliferation [3]. It contributes to inhibition of cell division by interrupting tubulin disassembly and results in cell death [4]. Docetaxel has over 90% plasma protein binding. However, upon intravenous administration, docetaxel is widely distributed to various tissues due to its large volume of distribution [3]. Docetaxel has poor water solubility. Therefore, a combination of ethanol and Polysorbate 80 (Tween™ 80) are used as co-solvents in the commercial formulation (Taxoter®) [2]. Evidence suggests that ingredients in the formulation (tween 80/ethanol) can cause allergic

reactions, non-specific distribution, and other issues such as fluid retention and edema [5–8]. Accordingly, nanoparticles as drug delivery platforms have been developed to address disadvantages of conventional chemotherapy. This has led to the development of cancer nanomedicine [9–13]. Polymeric nanoparticles prepared from poly (lactideco-glycolide) (PLGA) have been extensively used as delivery vehicles of a wide-spectrum of drugs including docetaxel [14–17]. Loading of drugs in nanoparticles impacts their pharmacokinetic and biodistribution profile in the body [18–20]. Nanoparticles may enhance a drug's systemic circulation residence time or modify its tissue distribution [21]. They may elevate the drug's delivery to particular organs or decrease the drug's distribution to other organs. Nanoparticle properties determine nanoparticles' performance when used as drug delivery systems [22]. For instance, average size acts as an important characteristic in determining the fate of nanoparticle in the body. Rapid clearance from systemic circulation can happen due to filtration in

⁎ Corresponding author at: College of Pharmacy & Nutrition, University of Saskatchewan, Room 3D01.1, D Wing Health Sciences Building, 107 Wiggins Road, Saskatoon, SK S7N 5E5, Canada. E-mail address: [email protected] (A. Haddadi).

https://doi.org/10.1016/j.msec.2019.109950 Received 10 November 2018; Received in revised form 6 June 2019; Accepted 5 July 2019 Available online 07 July 2019 0928-4931/ © 2019 Published by Elsevier B.V.

Materials Science & Engineering C 104 (2019) 109950

P. Rafiei and A. Haddadi

kidneys or entrapment in organs of reticuloendothelial system (RES) if nanoparticles have very small or very large size, respectively [23]. Clearance mechanisms such as renal or RES can be evaded by nanoparticles having suitable size range [24], and result in longer systemic residence time. Nanoparticle's surface characteristics influence its fate in the body as well. After systemic administration, nanoparticles with positive charge have greater chance of uptake by endothelial cells which have negative charge compared to nanoparticles with negative or neutral charge [25–27]. Due to unique features and characteristics, PLGA has attracted substantial interest for the fabrication of polymeric nanoparticles with drug delivery purpose [28,29]. The polymer is biocompatible, biodegradable, and non-toxic [30,31] and has been approved by the US Food and Drug Administration (FDA) for human use [32]. PLGA nanoparticles have been used for the delivery of a wide spectrum of materials (e.g., small molecule, macromolecule, peptide, protein, etc.) and offer both sustained release and targeted drug delivery [33]. They also offer the possibility of nanoparticle surface engineering [34]. The polymer can assume different physicochemical characteristics based on the type and number of monomers present in the PLGA structure. Therefore, nanoparticles prepared from PLGA polymer can assume variable features and characteristics based on the type of polymer and nanoparticle fabrication method. Because the fabrication methods create nanoparticles with unique properties, the formulation conditions that lead to nanoparticles with most intended characteristics need to be identified. This is usually done by optimization techniques. The conventional approach to optimize a process like development of a drug formulation has been involved with evaluating the effect of one variable while keeping others constant to see if the variable significantly affects the process. Such approaches are usually strenuous and uneconomical, [35,36]. Meanwhile, much simpler screening of experimental parameters has been made possible through application of experimental factorial designs with reduced number of experiments (i.e., a fraction of the full factorial combinations) to draw valuable conclusions and establish a relationship between process variables and the outcome of interest [37]. Among others, Taguchi method has been widely used to analyze many factors with few runs while factors are analyzed independently of one another being weighted equally and therefore offer a balanced design [38,39]. In addition, the estimation of one factor is not influenced by the effect from another factor. In the present work, an experimental design based on Taguchi robust model was used as a systematic design for the evaluation of the effect of several nanoparticle preparation variables (factors) on nanoparticle-important characteristics. These factors included the initial docetaxel concentration in organic phase (dispersed phase), ratio between organic phase:aqueous phase, PLGA polymer molecular weight (Mw), PLGA polymer terminus, PLGA polymer lactide:glycolide ratio, and PVA concentration in aqueous phase (continuous phase) on important PLGA nanoparticle characteristics including particle size, polydispersity index (PDI), surface electric charge (zeta potential), and drug loading efficiency. Accordingly, two factors were evaluated at four levels and four factors were evaluated at two levels (Table 1). The design was also used to build models for prediction of important particle characteristics based on nanoparticle fabrication conditions. Finally, response optimization was performed to obtain an optimized PLGA nanoparticle formulation loaded with docetaxel with suitable properties as a sustained-release drug delivery system for intravenous administration.

Table 1 Variables (factors) along with their corresponding values and trial levels in the Taguchi experimental design used in fabrication of docetaxel-loaded PLGA nanoparticles through emulsification-solvent-evaporation technique. Level Code

factor

1

2

3

4

A B C D E F

Drug concentration (mg/ml) Organic/aqueous phase ratio Polymer molecular weight (dl/g) Polymer terminus Lactide:glycolide ratio PVA concentration

0.25 1:2 0.15–0.25 Acid 50:50 2.2%

0.5 1:3 0.55–0.75 Ester 75:25 9%

1 1:4

1.5 1:5

lactide:glycolide (L:G) monomer ratio of 50:50, and inherent viscosity of 0.15–0.25 dl/g in hexafluoroisopropanol (HFIP), 2) PLGA polymer with ester terminal group, L:G monomer ratio of 50:50, and inherent viscosity of 0.15–0.25 dl/g in HFIP, 3) PLGA polymer with acid terminal group, L:G monomer ratio of 50:50, and inherent viscosity of 0.55–0.75 dl/g in HFIP, 4) PLGA polymer with ester terminal group, L:G monomer ratio of 50:50, and inherent viscosity of 0.55–0.75 dl/g in HFIP, 5) PLGA polymer with acid terminal group, L:G monomer ratio of 75:25, and inherent viscosity of 0.15–0.25 dl/g in HFIP, 6) PLGA polymer with ester terminal group, L:G monomer ratio of 75:25, and inherent viscosity of 0.15–0.25 dl/g in HFIP, 7) PLGA polymer with acid terminal group, L:G monomer ratio of 75:25, and inherent viscosity of 0.55–0.75 dl/g in HFIP, 8) PLGA polymer with ester terminal group, L:G monomer ratio of 75:25, and inherent viscosity of 0.55–0.75 dl/g in HFIP. Poly (D,L-lactide-co-glycolide)-poly (ethylene glycol) (PLGA-PEG) di-block co-polymer (50:50 PLGA attached to mPEG 5000, 15% wt) was purchased from Evonik Degussa Corp. (Birmingham, AL, USA). Docetaxel was purchased from LC laboratories (Woburn, MA, USA). All reagents were analytical grade or above and used as received, unless otherwise stated.

2.2. Experimental design L16 Taguchi robust design was used in the preparation, optimization, and evaluation of the effect of different factors on particle-important characteristics of docetaxel-loaded PLGA nanoparticles. Table 1 exhibits six factors and their corresponding levels. Two factors including initial drug concentration and organic/aqueous phase ratio were evaluated at four levels while polymer Mw, polymer terminus, and PVA concentration were evaluated at two levels. To examine the factors under investigation, 16 different combinations of factor/levels were followed based on the Taguchi array (Table 2). To reach the precision required for the ultimate statistical conclusion, all experimental runs were replicated three times. Response optimization was applied to identify the combination of nanoparticle preparation factors/levels that jointly optimize nanoparticle characteristic of interest. Response optimization was done for nanoparticle size, zeta potential, PDI, and drug loading efficiency. Optimized docetaxel-loaded PLGA nanoparticle formulation was then prepared. The L16 Taguchi design used here for the optimization of the PLGA NP preparation process is in fact a `fractional` factorial design. The 16run study is a fraction of the full factorial design. For such design (i.e., two factor at 4 levels and four factors at 2 levels), the full factorial design (all combination of factor/levels) would consist of 42 × 24 = 256 runs. Here, L16 Taguchi design uses a selected fraction (subset) of the experimental runs of a full factorial design (i.e., 16 out of 256) to acquire the most informative features of the process under study. In addition, the design has allowed for conduction of a systematic statistical analysis that demonstrates the factors having significant

2. Materials and methods 2.1. Materials Poly (D,L-lactide-co-glycolide) (PLGA) polymers of different characteristics were purchased from Absorbable Polymers International (Pelham, AL, USA): 1) PLGA polymer with acid terminal group, 2

Materials Science & Engineering C 104 (2019) 109950

P. Rafiei and A. Haddadi

Table 2 Combination of different factor/levels of Taguchi design studied in preparation of docetaxel-loaded PLGA nanoparticles [L16(4**2 2**4) Array] and corresponding characteristics demonstrated by fabricated nanoparticles. Run #

Drug conc. (A)

Org./aq. ratio (B)

Polymer Mw (C)

Terminus (D)

L:G (E)

PVA conc. (F)

Size

PDI

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

0.25 0.25 0.25 0.25 0.5 0.5 0.5 0.5 1 1 1 1 1.5 1.5 1.5 1.5

1:2 1:3 1:4 1:5 1:2 1:3 1:4 1:5 1:2 1:3 1:4 1:5 1:2 1:3 1:4 1:5

0.15–0.25 0.15–0.25 0.55–0.75 0.55–0.75 0.15–0.25 0.15–0.25 0.55–0.75 0.55–0.75 0.55–0.75 0.55–0.75 0.15–0.25 0.15–0.25 0.55–0.75 0.55–0.75 0.15–0.25 0.15–0.25

Acid Acid Ester Ester Ester Ester Acid Acid Acid Acid Ester Ester Ester Ester Acid Acid

50:50 50:50 75:25 75:25 75:25 75:25 50:50 50:50 75:25 75:25 50:50 50:50 50:50 50:50 75:25 75:25

2.2% 9% 2.2% 9% 2.2% 9% 2.2% 9% 9% 2.2% 9% 2.2% 9% 2.2% 9% 2.2%

151.1 ± 19.3 108.6 ± 1.8 159.8 ± 11.4 105.1 ± 5.5 191.4 ± 21.1 105.0 ± 2.6 135.1 ± 1.6 93.5 ± 6.9 117.2 ± 8.6 171.9 ± 7.1 100.2 ± 1.5 113.4 ± 4.5 119.8 ± 2.0 163.1 ± 8.9 104.0 ± 8.1 118.4 ± 2.2

0.175 0.256 0.113 0.130 0.219 0.158 0.122 0.154 0.148 0.118 0.177 0.198 0.095 0.180 0.290 0.243

Zeta

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.080 0.014 0.019 0.035 0.070 0.022 0.004 0.061 0.019 0.015 0.015 0.037 0.015 0.033 0.059 0.022

−22.8 −29.8 −25.3 −17.8 −19.3 −18.9 −31.1 −28.7 −32.8 −32.5 −18.2 −26.0 −18.4 −21.1 −29.8 −36.5

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

2.3 4.5 5.1 1.2 4.8 1.0 1.7 3.7 4.8 0.3 0.3 2.7 2.7 3.6 1.7 0.6

Drug loading efficiency (%)

Drug loading (%)

79.6 41.5 86.4 43.2 96.2 60.0 50.8 35.1 46.6 49.7 29.9 42.9 43.7 41.4 25.2 37.8

0.199 0.104 0.217 0.108 0.481 0.300 0.254 0.176 0.467 0.497 0.299 0.429 0.653 0.622 0.369 0.568

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

2.6 3.6 4.2 3.3 4.0 3.1 6.0 3.4 3.1 0.8 3.6 1.6 4.2 4.1 1.1 1.6

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.006 0.091 0.094 0.084 0.201 0.156 0.302 0.171 0.310 0.083 0.360 0.161 0.589 0.619 0.019 0.249

added to the precipitate. Again, the mixture was vortexed for 30 s, sonicated for 30 min, and centrifuged for 20 min. The supernatant was separated and mixed with supernatant 1. The mixture of supernatants was evaporated. 1 ml methanol was added to the residue to precipitate the polymer, vortexed for 30 s and centrifuged for 20 min at 4000 rpm. Docetaxel was then quantified in supernatant using tandem mass spectrometry method. Then, nanoparticles drug loading and encapsulation efficiency of the preparation method were calculated as follows:

effect on NP-important characteristics. Furthermore, the analysis and optimization has resulted in the development of models that help predict NP characteristics. Technically, these models could be used to prepare NP formulations with desired characteristics. 2.3. Statistical analysis Minitab® statistical software, version 16 (Minitab, State College, PA, USA) was used to obtain the L16 Taguchi design and to perform statistical analysis on the results. Multiple regression analysis, analysis of variance (ANOVA), and F test have been used to build prediction models for nanoparticle important characteristics. A p value of < 0.05 was considered statistically significant throughout the study unless otherwise stated.

Drug loading (%) = (weight of drug in particles/weight of particles) × 100 Drug loading efficiency (%) = (weight of drug in particles/initial weight of drug added) × 100

2.4. Preparation of nanoparticles

2.6. Analysis of quantification of docetaxel

Several techniques have been developed to produce polymeric nanoparticles from preformed polymers including emulsification-solventevaporation method which is one of the most common and the oldest technique used for fabrication of PLGA nanoparticles [29]. Here, a modified emulsification-solvent-evaporation technique was used to prepare docetaxel-loaded nanoparticles. Briefly, drug and PLGA polymer were dissolved in ethyl acetate to have a solution of 10% (w/v) PLGA and 0.25, 0.5, 1, or 1.5 mg/ml docetaxel. The solution was then added to 2.2 or 9% (w/v) Poly (vinyl alcohol) (PVA) at different organic:aqueous phase ratios of 1:2, 1:3, 1:4, or 1:5. The mixture was then vigorously shaken and subsequently sonicated. Obtained nano-suspension was then left to stir for 2 h for organic solvent evaporation. Nanoparticles were then obtained after washing/centrifugation steps, resuspended in 1% (w/v) sucrose, lyophilized, and kept at −20 °C.

A mass spectrometry (MS/MS) method was developed and validated in our lab and used to quantitatively determine docetaxel in PLGA nanoparticles being explained elsewhere [40]. Briefly, fragmentation pattern of docetaxel was determined as follows. Solution of docetaxel and paclitaxel (internal standard (IS)) in methanol containing 0.1% (v/ v) formic acid were directly infused into the ionization source using a model 11 Plus syringe pump (Harvard Apparatus, MA, USA) at a flow rate of 10 μl/min into a Hybrid Triple Quadrupole/Linear Ion trap mass spectrometer (AB Sciex 4000 QTRAP MS/MS system, Framingham, MA, US). The Turbo Ion-spray Source was set to 5500 V. Curtain gas, nebulizer, and heater gas pressure were 30, 40, and 40 psi respectively. The de-clustering potential (DP) of docetaxel and paclitaxel was 46 and 55 respectively. Collision energy of docetaxel and paclitaxel was 21 and 93 respectively. Collision cell exit potential was 18 for both materials. Other instrument parameters include: interface heater = ON (150 °C), collisionally activated dissociation (CAD) = 5, exit potential = 10. Ultimately, product-ion scans (MS/MS) were performed using positive electrospray ionization (+ESI) with appropriate set mass (parent m/z). According to the obtained pattern of docetaxel and IS fragmentation, 4 fragments of docetaxel were chosen to be used in Multiple Reaction Monitoring (MRM) transitions and docetaxel quantitation. Injection of samples to the mass spectrometer was done using an Agilent Quaternary pump (1200 series) and Agilent G1329A (1200 series) auto sampler (Santa Clara, CA, USA) through a pre-column guard (Eclipse XDB-Rapid resolution, C18, 2.1 × 30 mm, 3.5 μ, Agilent, Santa Clara, CA, USA) with a mobile phase flow rate of 200 μl/min (Methanol, 0.1% formic acid, isocractic elution). Quantitation procedure was performed on corresponding docetaxel and IS MRM graphs

2.5. Nanoparticle characterization Nanoparticles characteristics including mean diameter/size distribution profile, polydispersity index (PDI), and zeta potential were measured using a Malvern Zetasizer (Nano-series, Nano ZS, model ZEN3600, Malvern Instruments, Worcestershire, UK). NPs were suspended in water prior to the measurements at 25 °C. Nanoparticle drug loading and method encapsulation efficiency was evaluated through extraction of docetaxel from freeze-dried PLGA nanoparticle formulations. Briefly, acetone (1 ml) was added to drug-loaded nanoparticles (5 mg) to dissolve the polymer and drug. The mixture was vortexed for 30 s, sonicated for 30 min, and centrifuged for 20 min at 4000 rpm. The supernatant was separated and preserved (supernatant 1). Acetone was 3

Materials Science & Engineering C 104 (2019) 109950

P. Rafiei and A. Haddadi

coloured product. After incubating the plate for 1 to 4 h, absorbance at each well was measured at 490 nm using a 96-well plate reader. Cell viability was calculated using the following equation:

using Analyst software version 1.6. 2.7. Optimized nanoparticle formulation

Cell viability (%)

2.7.1. Preparation and characterization The optimized nanoparticle formulation was prepared according to the optimization results. Briefly, drug and PLGA polymer were dissolved in ethyl acetate to have a solution of 10% (w/v) PLGA and 1.5 mg/ml docetaxel. The solution was then added to 2.2% (w/v) PVA solution at organic:aqueous phase ratios of 1:3. The mixture was then vigorously shaken and subsequently sonicated. Obtained nano-suspension was then left to stir for 2 h for organic solvent evaporation. Nanoparticles were then obtained after washing/centrifugation steps, re-suspended in 1% (w/v) sucrose, lyophilized, and kept at −20 °C for further use. A surface-modified version of the optimized nanoparticle formulation was also prepared using a di-block co-polymer of PLGA and PEG to obtain PEGylated PLGA nanoparticles. Optimized nanoparticle formulations were then characterized in terms of average size, PDI, zeta potential, and loading characteristics as mentioned above.

= (Absorbance of treated cells/Absorbance of untreated cells) × 100 3. Results and discussion 3.1. L16 Taguchi design Table 2 demonstrates the combination of six different factors and their corresponding levels based on the L16 Taguchi design and the important particle characteristics obtained from each run including particle size, PDI, zeta potential, drug loading, and drug loading efficiency of the method. Nanoparticles demonstrate average size between 93.5 ± 6.9 nm and 191.4 ± 21.1 nm. The range of zeta potential exhibited by nanoparticles was between −17.8 ± 1.2 mV and −36.5 ± 0.6 mV. Fabricated nanoparticles demonstrated a uni-distributed population having PDI values between 0.095 ± 0.015 and 0.290 ± 0.059. Nanoparticle preparation at 16 different combinations of factors/levels resulted in a range of drug loading efficiency between 25.2 ± 1.1% and 96.2 ± 4.0%. The outcome of statistical analysis performed on each characteristic individually is also provided below.

2.7.2. Docetaxel release profile To have an estimate of the rate and pattern of release of docetaxel encapsulated inside nanoparticles, in vitro drug release test was performed on optimized docetaxel-loaded PLGA and PEGylated PLGA nanoparticles. Briefly, 20 mg nanoparticle was re-suspended in 4 ml phosphate buffered saline (PBS) and placed at 37 °C in a shaker-incubator. At designated time-points, the suspension was centrifuged at 15000 rpm for 10 min and the supernatant was harvested. Nanoparticles were re-suspended in fresh PBS in the original tube for further incubation. The supernatant was then extracted twice and the extract was subjected to quantitative analysis by mass spectrometer.

3.1.1. Particle size The main effect plots for the means of particle size versus different factor/level in Fig. 1A depict how particle size is affected by factors at different levels. The difference between maximum and minimum mean particle sizes (range) is highest in case of PVA concentration (43.8 nm) followed by organic:aqueous phase ratio (37.3 nm), lactide:glycolide ratio (11.0 nm), polymer Mw (9.2 nm), polymer terminus (7.3 nm), and drug concentration (5.6 nm). Based on the variation of mean particle size (range), PVA concentration had the largest effect on particle size, followed by organic:aqueous phase ratio, lactide:glycolide ratio, polymer Mw, polymer terminus, and drug concentration. Supplementary Fig. 1A represents the normal probability plots of residuals for particle size which demonstrates the normality of particles size and constant variance throughout the response range. Linear regression analysis (Table 3) demonstrates that both PVA concentration and organic:aqueous phase ratio affect particle size with statistical significance (p < 0.05). Also, the results of ANOVA and F test performed on obtained data indicate that the fitted model for particle size can predict particle size with statistical significance (p < 0.05) (Table 4). The R2 from regression analysis was equal to 0.899 and the regression equation was calculated to be:

2.7.3. Cytotoxicity of loaded docetaxel in nanoparticles Hela cell (American Type Culture Collection, Manassas, VA, USA) lines were cultivated in high glucose Dulbecco's modified Eagle's medium (DMEM) culture media supplemented with 10% fetal bovine serum (FBS) and 1% antibiotic (100 unit/ml penicillin, 100 μg/ml streptomycin, and 0.25 μg/ml amphotericin B) at 37 °C in a humidified incubator with 5% CO2. The cells were maintained in an exponential growth phase by periodic sub-cultivation. Hela cells were seeded into 96-well plates at a density of 5000 per well and incubated for 24 h (humidified 37 °C, and 5% CO2). Then, the growth medium was removed from the wells and replaced with 90 μl of fresh medium. Cells were then treated with 10 μl of the following: 1. Free drug solution in culture media at 25, 50, 100, 200, and 400 ng/ ml docetaxel concentrations. 2. PLGA or PEGylated PLGA nanoparticle suspension in phosphate buffer (pH = 7.4) with equivalent docetaxel concentrations of 25, 50, 100, 200, and 400 ng/ml. 3. Plain PLGA or PEGylated PLGA nanoparticle suspension (control).

Particle Size = 189–2.01 (A) − 12.4 (B) + 9.18 (C) + 7.26 (D) + 11.0 (E) − 43.8 (F) To establish the relationship between the response of interest (i.e., particle size) and the factors that significantly affect particle size (i.e., Organic/Aqueous phase ratio and PVA concentration), response surface graphs have been plotted (Supplementary Fig. 2). The figure indicates that maximum response (i.e., particle size) is obtained at low PVA concentration (level one) and 1:2 organic:aqueous phase ratio (level one), and vice versa. It is also evident that increasing the volume of aqueous phase decreases particle size (Supplementary Fig. 2). The amount of aqueous phase in the suspension played an important role in determining the final size of the nanoparticles. In samples with a high volume of aqueous phase, nano-droplets were distributed throughout a larger volume of liquid and therefore had less chance of adhesion. The rate of organic solvent diffusion from nano-droplets to the surrounding aqueous medium and the final evaporation from nano-suspension also decreased particle size. In a nano-suspension with a large volume of aqueous

Blank wells were also included in the assay to exclude the potential absorbance of culture media and PLGA or PEGylated PLGA nanoparticles. Blank samples were as follows: 1. Wells with culture media without cells. 2. Wells with culture media without cells treated with plain PLGA or PEGylated PLGA nanoparticle suspension. After cell treatment, plates were incubated for 6, 12, 24, and 48 h. The percentage of cell viability was assayed by the number of surviving cells after incubation. At each time-point, medium in each well was replaced with 100 μl fresh medium containing 10 μl CellTiter 96® AQueous One Solution Cell Proliferation Assay Reagent (Promega, WI, USA). The compound in the reagent was bio-converted by cells into a 4

Materials Science & Engineering C 104 (2019) 109950

P. Rafiei and A. Haddadi

Fig. 1. Main effect plot for A) particle size (larger S/N is better), B) particle PDI (larger S/N is better), C) particle zeta potential (Nominal is best), and D) drug loading efficiency (larger S/N is better) at different factor-levels.

presence of PVA on the surface contributes to fewer nano-droplet interactions, adhesion, or fusion. Accordingly, using higher PVA concentration contributed to localization of larger number of PVA molecules on nano-droplets and resulted in lower particle size. The study by Murugesan et al. [44] also demonstrated that increasing the concentration of stabilizing agent reduces nanoparticle size. The ratio of monomers (L:G) also affected the size of nanoparticles. Lactic acid is chemically more hydrophobic than glycolic acid. Consequently, the 50:50 (L:G) PLGA polymer was less hydrophobic than the 75:25 (L:G) PLGA polymer. Therefore, the organic phase

phase, the rate of organic solvent diffusion from nano-droplet to the aqueous medium was also higher [41–43]. Therefore, nano-droplets solidified faster and had less chance of adhesion. Consequently, nanoparticles retained their sizes. Our results agree with those of Murugesan and colleagues [44] who demonstrated that the amount of aqueous phase during nanoparticle preparation has an inverse relationship with nanoparticle size. PVA concentration and nanoparticle size demonstrated an inverse relationship. PVA as a non-ionic spatial stabilizer is considered to be located on nano-droplets' surface during nanoparticle preparation. The

Table 3 Regression analysis of various particle characteristics versus nanoparticle preparation variables. Particle size

PDI

Predictor

Coefficients

SE Coefficients

T

p

Coefficients

SE Coefficients

T

p

Constant Drug concentration (A) Organic/aqueous phase ratio (B) Polymer Mw (C) Polymer Terminus (D) Lactide:glycolide ratio (E) PVA concentration (F)

189.30 −2.008 −12.413 9.183 7.258 10.992 −43.850

20.89 2.725 2.725 6.094 6.094 6.094 6.094

9.06 −0.74 −4.55 1.51 1.19 1.80 −7.20

0.000 0.480 0.001 0.166 0.264 0.105 0.000

0.28163 0.00990 0.00636 −0.0819 −0.0296 0.00758 0.00508

0.06573 0.00857 0.00857 0.01917 0.01917 0.01917 0.01917

4.28 1.15 0.74 −4.27 −1.55 0.40 0.27

0.002 0.278 0.477 0.002 0.156 0.702 0.797

Zeta potential

Constant Drug concentration (A) Organic/aqueous phase ratio (B) Polymer Mw (C) Polymer terminus (D) Lactide:glycolide ratio (E) PVA concentration (F)

Drug loading efficiency

Coefficients

SE coefficients

T

p

Coefficients

SE coefficients

T

p

−34.103 −1.0446 −1.2404 −0.810 9.848 −2.098 2.540

5.308 0.6925 0.6925 1.549 1.549 1.549 1.549

−6.42 −1.51 −1.79 −0.52 6.36 −1.35 1.64

0.000 0.166 0.107 0.613 0.000 0.209 0.135

97.94 −9.506 −8.044 −2.034 9.672 10.037 −19.949

14.25 1.860 1.860 4.158 4.158 4.158 4.158

6.87 −5.11 −4.33 −0.49 2.33 2.41 −4.80

0.000 0.001 0.002 0.636 0.045 0.039 0.001

Data in bold represent statistically significant difference between factors (p > 0.05) 5

Materials Science & Engineering C 104 (2019) 109950

P. Rafiei and A. Haddadi

Table 4 ANOVA tables of fitted model for various particles characteristics.

Particle size

PDI

Zeta potential

Drug loading efficiency

Source

Degree of freedom

Sum of squares

Mean square

F

p

Regression Residual error Total Regression Residual error Total Regression Residual error Total Regression Residual error Total

6 9 15 6 9 15 6 9 15 6 9 15

11,885.1 1337.1 13,222.2 0.033466 0.013236 0.046702 486.552 86.332 572.883 5486.88 622.52 6109.41

1980.9 148.6

13.33

0.001

0.005578 0.001471

3.79

0.036

81.092 9.592

8.45

0.003

914.48 69.17

13.22

0.001

prepared from 50:50 (L:G) PLGA had less hydrophobic integrity than the organic phase prepared from 75:25 (L:G) PLGA. During sonication, organic phase droplets (with lower hydrophobicity) more easily broke into smaller droplets, which lowered nanoparticle average size. On the other hand, because of their lower hydrophobic integrity, the organic solvent escaped nano-droplets prepared from 50:50 (L:G) PLGA faster, which caused the nanoparticles to solidify more rapidly. This decreased the chance of droplet adhesion and fusion during the solvent evaporation period. Increasing the polymer Mw also increased average particle size. The PLGA polymer with low Mw resulted in an organic phase with low viscosity. During sonication, organic phase droplets with low viscosity broke into small droplets more easily than droplets with higher viscosity (higher Mw). Therefore, low Mw PLGA provided nanoparticles with a lower average size compared to high Mw PLGA. Koopaei and colleagues [45] demonstrated that organic phase with higher viscosity leads to nanoparticles with a higher average size, which agrees with our result.

preparation, and vice versa (Supplementary Fig. 3A). Viscosity of organic phase helped determine the average size of nanoparticles. PLGA polymers with higher Mw led to more viscose droplets (in the preparation emulsion) that broke into smaller droplets more difficultly. This kept the size of nano-droplet at a certain range, which decreased the diversity and variability in nanoparticles' size and kept the PDI low. Conversely, when low molecular weight polymer was used for nanoparticle preparation, organic phase had a lower viscosity than when high molecular weight polymer is used. During sonication, organic-phase droplets with lower viscosity sequentially broke into smaller particles more easily. Consequently, low molecular weight polymer resulted in nanoparticles with smaller average size. The sequential breakdown of droplets resulted in a nano-suspension with nano-droplets possessing a wide range of sizes that increased the PDI. On the other hand, acid terminal functional group and low molecular weight resulted in maximum PDI response while ester terminus and high molecular weight provided minimum PDI values (Supplementary Fig. 3B).

3.1.2. Particle Polydispersity Index (PDI) As exhibited in Table 2, nanoparticles prepared by the modified emulsification solvent evaporation technique demonstrate PDIs < 0.3, which is indicative of a narrow distribution profile among particle population. Fig. 1B demonstrates the main effect plot for the means of particle PDI versus factor/levels. The figure demonstrates how average PDI of nanoparticle populations is changed at various factors/levels. Polymer Mw had the highest range (0.0819) followed by drug concentration (0.0419), polymer terminus (0.0297), organic:aqueous phase ratio (0.022), lactide:glycolide ratio (0.0076), and PVA concentration (0.0051). Accordingly, polymer Mw had largest effect and drug concentration had second largest effect on PDI. The randomness in the variance around the response range for particle PDI and their normal distribution (Supplementary Fig. 1B) was further demonstrated from the analysis. Further analysis of the data demonstrates that polymer Mw was the factor affecting the particle PDI with statistical significance (p < 0.05) (Table 3). Based on the results from regression analysis, the fitted model could significantly (p = 0.036) (Table 4) explain about 70% (R2 = 0.717) of variations in PDI based on the variations in factors under study. The regression analysis resulted in below equation:

3.1.3. Particle zeta potential Table 2 exhibits the zeta potentials of PLGA nanoparticles from 16 different experimental runs. Obtained particles possess zeta potentials with negative value within a range of −16 to −37 mV. Fig. 1C exhibits the average zeta potential at each factor/level depicting how zeta potential varies by factors at different levels. The ranges between minimum and maximum nanoparticle zeta potential were calculated between levels of each factor and were as follows (in decreasing order): polymer terminus (9.85 mV), organic:aqueous phase ratio (3.96 mV), drug concentration (3.43 mV), PVA concentration (2.54 mV), lactide:glycolide ratio (2.10 mV), and polymer Mw (0.81 mV). Accordingly, polymer terminus had the highest magnitude of effectiveness on the surface zeta potential of nanoparticles while organic:aqueous phase ratio has the second largest effect. This is in agreement with the fact that acid terminus can easily become ionized and expose more negative charges to the surface of nanoparticles. The fitted model to explain variations in particle zeta potential was statistically significant (p < 0.05) (Table 4). The regression equation was calculated to be (R2 = 0.849):

Particle Zeta Potential = −34.1 − 1.04 (A) − 1.24 (B) − 0.81 (C) + 9.85 (D) − 2.10 (E) + 2.54

Particle PDI

(F)

= 0.282 + 0.00990 (A) + 0.00637 (B) − 0.0819 (C) − 0.0297

Normal distribution and variance throughout the range of residuals and fitted values again satisfied requirements of the regression analysis done on particles zeta potential (Supplementary Fig. 1C). Compared to other factors in the model, only polymer terminus was able to define variations in particle zeta potential with statistical significance (p < 0.05) (Table 3). To understand the relationship between the response of interest

(D) + 0.0076 (E) + 0.0051 (F) The relationship between particle PDI and the factor that significantly affects particle PDI (i.e., polymer Mw) is depicted in Supplementary Fig. 3 (response surface graphs). Accordingly, it is evident that maximum PDI is obtained when low Mw PLGA polymer (level 1) and high docetaxel concentration (level 4) are used for nanoparticle 6

Materials Science & Engineering C 104 (2019) 109950

P. Rafiei and A. Haddadi

(i.e., particle zeta potential) and the factor significantly affecting it (i.e., polymer terminus), response surface graphs were plotted (Supplementary Fig. 4). Accordingly, the maximum value of particle zeta potential was obtained when ester-terminated polymer (level 2) and a 1:2 ratio of organic:aqueous phase (level 1) were used during the nanoparticle fabrication procedure and that polymers with acid terminus at 1:5 organic:aqueous phase ratio resulted in highest surface negativity (Supplementary Fig. 4A). Also, the minimum value of particle zeta potential was obtained when acid-terminated polymer (level 1) and a low PVA concentration (level 1) were used during nanoparticle fabrication and that using a high concentration of PVA and ester-terminated polymer resulted in nanoparticles with less surface negative zeta potential (Supplementary Fig. 4B). Generally, zeta potential of PLGA nanoparticles can be attributed to many determinants. An important factor is considered to be the role of polymer terminus arranged on the surface of nanoparticles which can potentially be ionized. For instance, hydrolysis of carboxylic acid (in case of acid-terminated polymer) or ester bond (in case of ester-terminated polymer) can contribute to different surface potentials. Surface zeta potential is also affected by the amount and type of stabilizer used during nanoparticle preparation. Traces of PVA (stabilizer) can potentially remain on the surface of PLGA nanoparticles even after consecutive washing steps. Associated/loaded drug in nanoparticles can also affect zeta potential being dependent on physicochemical characteristics of the drug. Docetaxel and PLGA are hydrophobic molecules. Consequently, non-covalent bonds such as Van der Waals or hydrogenbonds between the two molecules (being different at different docetaxel concentrations) can potentially result in different arrangements of polymer terminus on the surface of nanoparticles and result in different surface zeta potentials. Other studies dealing with preparation and characterization of PLGA nanoparticles loaded with docetaxel have demonstrated similar ranges of negative zeta potential [46–48]. Anyhow, surface characteristics such as zeta potential can affect the fate of nanoparticles in the systemic circulation [49]. Positively charged nanoparticles have a higher rate of cell-uptake and clearance when compared to neutral and negatively charged formulations [25]. Presence of negatively charged serum proteins can lead to formation of aggregates after intravenous administration of positively charged nanoparticles [50]. These aggregates can secondarily be entrapped inside RES organs or result in embolism in blood capillaries [51–54]. Opsonization of nanoparticles has been correlated to the surface charge as well.

The analysis resulted in a statistically significant model (p < 0.05) to define drug loading efficiency (Table 4). Interestingly, in the fitted model, five factors (including drug concentration, organic:aqueous phase ratio, polymer terminus, lactide:glycolide ratio, and PVA concentration) turned out to be statistically significant in contributing to the final model defining drug loading efficiency (Table 3). Response surface graphs (Supplementary Fig. 5) demonstrate the relationship between the response of interest (i.e., drug loading efficiency) and drug concentration, organic/aqueous phase ratio, and PVA concentration. It is evident that minimum drug loading efficiency is obtained when 1:5 ratio of organic:aqueous phase (level 4) and highest drug concentration (level 4) are used during nanoparticle preparation, and vice versa (Supplementary Fig. 5A). Also, highest drug concentration (level 4) and high PVA concentration (level 2) resulted in minimum drug loading efficiency (Supplementary Fig. 5B). Accordingly, highest drug loading efficiencies were obtained at organic:aqueous phase ratio of 1:2 (level 1) and low PVA concentration (level 1), and vice versa (Supplementary Fig. 5C). Studies involving preparation and characterization of docetaxelloaded PLGA nanoparticles have reported a range of drug loading encapsulation efficiencies from 16% to 76% [46–48] which was different based on the type of PLGA and nanoparticle preparation method. The modified emulsification-solvent-evaporation used here resulted in a range of drug loading efficiency between 25% and 96%. Drug loading efficiency of nanoparticle preparation method was reduced parallel to the increase in drug concentration. This trend is attributed to the fact that polymer matrix (forming nanoparticles) can potentially become saturated by drug molecules when higher concentration of drug is used. In addition, another trend is the evident decrease in drug loading efficiency parallel to the increment in the amount of aqueous solution. This can be attributed to the fact that when the amount of aqueous solution is increased, more drug molecules can escape from organic phase to aqueous phase and as a result reduce the overall drug loading efficiency. Murugesan and colleagues [44] also evaluated the effect of initial docetaxel concentration and stabilizer concentration on drug loading efficiency. They demonstrated that drug loading efficiency has an inverse relationship with the concentration of stabilizing agent and docetaxel during preparation. In other words, increasing the concentration of stabilizer or docetaxel contributes to reduction in drug loading efficiency of nanoparticle preparation method. They also demonstrated that increasing the amount of aqueous phase during nanoparticle preparation results in reduction of drug entrapment efficiency.

3.1.4. Drug loading efficiency Drug loading efficiencies from sixteen different runs of the design were determined using the developed analysis method [40]. Nanoparticles demonstrate drug loading efficiencies as low as 25% and as high as 96% (Table 2). Main effect plot for the means of drug loading efficiency is exhibited in Fig. 1D. The range between maximum and minimum drug loading efficiency was calculated for each factor and determined organic:aqueous phase ratio to be the factor with highest range (26.78%). Ranges of drug loading efficiency for other factors were as follows: drug concentration (25.61%), PVA concentration (19.95%), lactide:glycolide ratio (10.04%), polymer terminus (9.67%), and polymer Mw (2.03%). Organic:aqueous phase ratio was considered to be the factor with highest magnitude of effect on drug loading efficiency while drug concentration had the second largest effect. The normal probability plot of residuals (Supplementary Fig. 1D) demonstrated a constant variance throughout the response range for drug loading efficiency of nanoparticle preparation method. Statistical analysis to fit a model for docetaxel loading efficiency of the method resulted in below regression equation with an R2 equal to 0.898:

3.2. Optimization of nanoparticle preparation Response optimization was applied to identify the combination of factors/levels during nanoparticle preparation that jointly optimize nanoparticle characteristic of interest. Response optimization was performed for nanoparticle size, PDI, zeta potential, and drug loading efficiency. Fig. 2 demonstrates optimization plots of nanoparticle characteristics under study. Based on the obtained data, optimized PLGA nanoparticle fabrication conditions relating to individual nanoparticle characteristics were determined to be as follows: Nanoparticle size: Composite desirability near 1 was obtained at level 3.8977 of drug concentration, 3.0965 of aqueous:organic phase ratio, 1.1200 of polymer Mw, 1.0 of polymer terminus, 1.0263 of L:G ratio, and 1.0 of PVA concentration. PDI: Composite desirability near 1 was obtained at level 3.9556 of drug concentration, 2.0950 of aqueous:organic phase ratio, 1.6775 of polymer Mw, 1.0 of polymer terminus, 1.0263 of L:G ratio, and 1.0148 of PVA concentration. Zeta Potential: Composite desirability near 1 was obtained at level 3.9410 of drug concentration, 1.80 of aqueous:organic phase ratio, 1.0001 of polymer Mw, 1.60 of polymer terminus, 1.0 of L:G ratio, and 1.0148 of PVA concentration.

Drug Loading Efficiency = 97.9–9.51 (A) − 8.04 (B) − 2.03 (C) + 9.67 (D) + 10.0 (E) − 19.9 (F) 7

Materials Science & Engineering C 104 (2019) 109950

P. Rafiei and A. Haddadi

Fig. 2. Response optimization plot showing the optimized factor/level setting for A) average size, B) PDI, C) zeta potential, and D) drug loading Efficiency.

Drug Loading Efficiency: Composite desirability near 1 was obtained at level 3.6021 of drug concentration, 1.4470 of aqueous:organic phase ratio, 1.0492 of polymer Mw, 1.0279 of polymer terminus, 1.0164 of L:G ratio, and 1.0 of PVA concentration. Taken collectively the factor/levels at which suitable composite desirability level was obtained regarding individual nanoparticle characteristics, overall fabrication conditions for an optimized docetaxel-loaded PLGA nanoparticle formulation were determined as: Docetaxel Concentration: Level 4 (1.5 mg/ml) Organic/Aqueous phase ratio: Level 2 (1:3) PLGA Molecular weight: Level 1 (0.15–0.25 dl/g) PLGA polymer Terminus: Level 1 (Acid terminated) Polymer lactide/glycolide ratio: Level 1 (50:50) PVA Concentration: Level 1 (2.2%)

which are followed by swelling and erosion of the polymer matrix leading to drug diffusion out of nanoparticles [48]. During the incubation period, 25% and 49% of the loaded drug was released from PLAG nanoparticles and PEGylated PLGA nanoparticles, respectively. The difference in the drug release pattern is attributed to the presence of PEG molecules which lead to better nanoparticle hydration [55]. In addition, the PEG structure can sometimes become entrapped inside nanoparticles matrix [56] and again contribute to water attraction and more pronounced hydration of the nanoparticles. According to Abouelmagd [57] and colleagues, release profile of poorly water-soluble drugs such as paclitaxel could be different in various media. They studied the release profile of paclitaxel from PLGA nanoparticles in PBS and in PBS/fetal bovine serum (FBS). They observed that the extent of drug release was higher when FBS was present (98%) compared to PBS (34%). They attributed this difference to the role of serum proteins offering solubilizing effect to the drug. They concluded that measuring drug release in PBS could result in underestimation of the extent of drug release (our release test).

3.3. Optimized nanoparticle formulation 3.3.1. In vitro characteristics The optimized fabrication method provided docetaxel-loaded PLGA nanoparticles with average diameter of 123.6 ± 9.5 nm, PDI of 0.245 ± 0.041, and zeta potential of −28.3 ± 1.25 mV. The method resulted in 0.56 ± 0.025% drug loading and 37.25 ± 1.6% and drug loading efficiency. The optimized preparation method provided surface-modified (PEGylated) version nanoparticles with larger average diameter, loading characteristics, and higher zeta potential, as compared to optimized docetaxel-loaded PLGA nanoparticle formulation. PEGylated nanoparticle formulation exhibited 186.7 ± 2.9 nm average diameter, −25.9 ± 3.5 mV zeta potential, and PDI of 0.103 ± 0.10. The optimized nanoparticle formulation demonstrated characteristics suitable for an intravenous sustained-release drug-delivery system. Average particle size was considered small enough (being below 200 nm) to evade filtration in organs of RES and large enough to evade filtration in the kidneys [23]. Nanoparticles had a uni-distributed population with negative surface zeta potential which allow for long circulating times for nanoparticles in the blood [49,51].

3.3.3. Cytotoxicity of loaded docetaxel in nanoparticles Fig. 4 demonstrates cytotoxic activity of docetaxel loaded in PLGA and PEGylated PLGA nanoparticles on Hela cancer cells compared to that of the free drug in a set of concentrations (25–400 ng/ml) at various incubation times. Hela cells were also treated and incubated with plain nanoparticles to determine cytotoxicity of empty nanoparticles. The results showed that plain nanoparticles did not influence cell viability. At six hours of incubation, the free docetaxel solution demonstrated cell viability of 81–94% (at various concentrations). Cell viability for docetaxel-loaded PLGA and PEGylated PLGA nanoparticles were 88–107% and 79–99%, respectively. At 12 h of incubation, the free drug solution demonstrated 34–54% cell viability, while PLGA and PEGylated PLGA nanoparticles respectively exhibited 29–62% and 32–74% cell viability. After 24 h of incubation, the free docetaxel solution, docetaxel-loaded PLGA nanoparticles, and docetaxel-loaded PEGylated PLGA nanoparticles resulted in 5–31%, 3–32%, and 3–38% cell viability, respectively. Incubation of cells with the free docetaxel solution, docetaxel-loaded PLGA and docetaxel-loaded PEGylated PLGA nanoparticles for 48 h resulted in 2–7%, 2–9% and 3–25% cell viability, respectively. In contrast to the free drug solution, the cytotoxicity of docetaxel loaded in nanoparticles is attributed to a series of mechanisms [8] as follows. Nanoparticles can potentially interact with cells and get stationed near the cell membrane. This could result in an elevated drug concentration adjacent to the cell membrane and result in a concentration gradient that promotes the drug's influx into the cell [58].

3.3.2. Docetaxel release profile The optimized nanoparticle formulation demonstrated a sustained in vitro release profile, which was expected from PLGA nanoparticles being suitable for intravenous sustained-release purpose [20]. As evident in Fig. 3, both the optimized nanoparticle formulation and the surface-modified (PEGylated) formulation exhibit a burst release during the first 24 h, followed by a sustained drug release typical of a biphasic release pattern. The initial burst release is generally attributed to the liberation of the drug from nanoparticle's surface and outermost layers 8

Materials Science & Engineering C 104 (2019) 109950

P. Rafiei and A. Haddadi

Fig. 3. Docetaxel release profile from PLGA nanoparticles and PEGylated PLGA nanoparticles (1,6,12, and 24 h, then 2,3,4 and 5 days). Reproduced with permission from reference [21].

nanoparticles in Hela cells. Accordingly, they demonstrated that, upon contact (incubation) with PLGA nanoparticles, Hela cells engage in both association and internalization acts. However, the association between the nanoparticles and cells was more pronounced compared to the internalization of nanoparticles.

Considering the possibility that nanoparticles could station near the cells and contribute to elevated drug concentration at the cell membrane, the efflux transport mechanisms such as P-glycoprotein (P-gp) pumps, could become saturated leading to pronounced drug uptake and cytotoxicity [59,60]. Another possibility is nanoparticle up take by cells. Cell-internalization could further bypass the effect of P-gp pumps. Sims and colleagues [61] evaluated the uptake and transport of PLGA

Fig. 4. Effect of plain PLGA and PEGylated PLGA nanoparticles (PLGA-PEG), free docetaxel, and docetaxel loaded in PLGA and PEGylated PLGA nanoparticles (PLGA-PEG) on the viability of Hela cells (Human Cervix Carcinoma) after incubation for A) 6 h, B) 12 h, C) 24 h, and D) 48 h (n = 3). 9

Materials Science & Engineering C 104 (2019) 109950

P. Rafiei and A. Haddadi

4. Conclusion

[13] N.R. Jabir, et al., Nanotechnology-based approaches in anticancer research, Int. J. Nanomedicine 7 (2012) 4391–4408. [14] R. Dinarvand, et al., Polylactide-co-glycolide nanoparticles for controlled delivery of anticancer agents, Int. J. Nanomedicine 6 (2011) 877–895. [15] W. Tao, et al., Docetaxel-loaded nanoparticles based on star-shaped mannitol-core PLGA-TPGS diblock copolymer for breast cancer therapy, Acta Biomater. 9 (11) (2013) 8910–8920. [16] R. Pradhan, et al., Docetaxel-loaded polylactic acid-co-glycolic acid nanoparticles: formulation, physicochemical characterization and cytotoxicity studies, J. Nanosci. Nanotechnol. 13 (8) (2013) 5948–5956. [17] R. Agrawal, et al., Polyelectrolyte coated polymeric nanoparticles for controlled release of docetaxel, J. Biomed. Nanotechnol. 8 (1) (2012) 19–28. [18] M. Hamidi, et al., A pharmacokinetic overview of nanotechnology-based drug delivery systems: an ADME-oriented approach, Crit. Rev. Ther. Drug Carrier Syst. 30 (5) (2013) 435–467. [19] X. Song, et al., Pharmacokinetics and disposition of various drug loaded biodegradable poly(lactide-co-glycolide) (PLGA) nanoparticles, Curr. Drug Metab. 11 (10) (2010) 859–869. [20] R. Pedram, H. Azita, Pharmacokinetic consequences of PLGA nanoparticles in docetaxel drug delivery, Pharmaceutical Nanotechnology 5 (1) (2017) 3–23. [21] P. Rafiei, A. Haddadi, Docetaxel-loaded PLGA and PLGA-PEG nanoparticles for intravenous application: pharmacokinetics and biodistribution profile, Int. J. Nanomedicine 12 (2017) 935–947. [22] J.D. Heidel, M.E. Davis, Clinical developments in nanotechnology for cancer therapy, Pharm. Res. 28 (2) (2011) 187–199. [23] S.M. Moghimi, A.C. Hunter, T.L. Andresen, Factors controlling nanoparticle pharmacokinetics: an integrated analysis and perspective, Annu. Rev. Pharmacol. Toxicol. 52 (2012) 481–503. [24] N.T. Huynh, et al., The rise and rise of stealth nanocarriers for cancer therapy: passive versus active targeting, Nanomedicine (Lond) 5 (9) (2010) 1415–1433. [25] J. Wang, M. Sui, W. Fan, Nanoparticles for tumor targeted therapies and their pharmacokinetics, Curr. Drug Metab. 11 (2) (2010) 129–141. [26] E. Frohlich, The role of surface charge in cellular uptake and cytotoxicity of medical nanoparticles, Int. J. Nanomedicine 7 (2012) 5577–5591. [27] Y.H. Choi, H.K. Han, Nanomedicines: current status and future perspectives in aspect of drug delivery and pharmacokinetics, J Pharm Investig 48 (1) (2018) 43–60. [28] R.A. Jain, The manufacturing techniques of various drug loaded biodegradable poly (lactide-co-glycolide) (PLGA) devices, Biomaterials 21 (23) (2000) 2475–2490. [29] C.E. Astete, C.M. Sabliov, Synthesis and characterization of PLGA nanoparticles, J Biomater Sci Polym Ed 17 (3) (2006) 247–289. [30] L.S. Nair, C.T. Laurencin, Biodegradable polymers as biomaterials, Prog. Polym. Sci. 32 (8–9) (2007) 762–798. [31] M.S. Shive, J.M. Anderson, Biodegradation and biocompatibility of PLA and PLGA microspheres, Adv. Drug Deliv. Rev. 28 (1) (1997) 5–24. [32] D.N. Kapoor, et al., PLGA: a unique polymer for drug delivery, Ther. Deliv. 6 (1) (2015) 41–58. [33] S. Rezvantalab, et al., PLGA-based nanoparticles in cancer treatment, Front. Pharmacol. 9 (2018) 1260. [34] D. Ding, Q. Zhu, Recent advances of PLGA micro/nanoparticles for the delivery of biomacromolecular therapeutics, Mater. Sci. Eng. C Mater. Biol. Appl. 92 (2018) 1041–1060. [35] B. Singh, et al., Optimizing drug delivery systems using systematic “design of experiments”. Part II: retrospect and prospects, Crit. Rev. Ther. Drug Carrier Syst. 22 (3) (2005) 215–294. [36] B. Singh, R. Kumar, N. Ahuja, Optimizing drug delivery systems using systematic “design of experiments”. Part I: fundamental aspects, Crit. Rev. Ther. Drug Carrier Syst. 22 (1) (2005) 27–105. [37] S. Saraf, Process optimization for the production of nanoparticles for drug delivery applications, Expert Opin Drug Deliv 6 (2) (2009) 187–196. [38] R.S. Rao, et al., The Taguchi methodology as a statistical tool for biotechnological applications: a critical appraisal, Biotechnol. J. 3 (4) (2008) 510–523. [39] M. Hamidi, et al., Taguchi orthogonal array design for the optimization of hydrogel nanoparticles for the intravenous delivery of small-molecule drugs, J. Appl. Polym. Sci. 126 (5) (2012) 1714–1724. [40] P. Rafiei, D. Michel, A. Haddadi, Application of a rapid ESI-MS/MS method for quantitative analysis of docetaxel in polymeric matrices of PLGA and PLGA-PEG nanoparticles through direct injection to mass spectrometer, Am. J. Anal. Chem. 6 (2015) 164–175. [41] K.C. Song, et al., The effect of type of organic phase solvents on the particle size of poly(d,l-lactide-co-glycolide) nanoparticles, Colloids Surf. A Physicochem. Eng. Asp. 276 (1–3) (2006) 162–167. [42] A. Budhian, S.J. Siegel, K.I. Winey, Haloperidol-loaded PLGA nanoparticles: systematic study of particle size and drug content, Int. J. Pharm. 336 (2) (2007) 367–375. [43] N. Sharma, P. Madan, S. Lin, Effect of process and formulation variables on the preparation of parenteral paclitaxel-loaded biodegradable polymeric nanoparticles: a co-surfactant study, Asian J Pharm Sci 11 (3) (2016) 404–416. [44] S. Murugesan, et al., PEGylated poly(lactide-co-glycolide) (PLGA) nanoparticulate delivery of docetaxel: synthesis of diblock copolymers, optimization of preparation variables on formulation characteristics and in vitro release studies, J. Biomed. Nanotechnol. 3 (1) (2007) 52–60. [45] M. Noori Koopaei, et al., Docetaxel loaded PEG-PLGA nanoparticles: optimized drug loading, in-vitro cytotoxicity and in-vivo antitumor effect, Iran J Pharm Res 13 (3) (2014) 819–833. [46] T. Musumeci, et al., PLA/PLGA nanoparticles for sustained release of docetaxel, Int. J. Pharm. 325 (1–2) (2006) 172–179.

Offering several advantages, polymeric nanoparticles of PLGA have been one of the most promising drug delivery platforms in cancer nanomedicine. However, the variable nature of PLGA polymer along with the variability in nanoparticle fabrication process can potentially result in nanoparticles with different characteristics. Upon administration of nanoparticles to the body some factors determine the fate of nanoparticles as well as the loaded drug. These factors are considered to be the ones attributed to the body and those attributed to the nanoparticle itself. Factors attributed to the nanoparticle are generally related to nanoparticle's characteristics such as size, shape, and surface which can be predicted and controlled through nanoparticle preparation procedure. Therefore, it is understood that controlling nanoparticle characteristics can provides control over the fate of nanoparticle and its payload in the body. In the present work, an experimental design based on Taguchi robust method was used for optimization of PLGA nanoparticles loaded with docetaxel. The L16 Taguchi design used here allowed us to identify factors that significantly influence nanoparticleimportant characteristics. Analysis of the data resulted in models (statistically significant) with good robustness as suitable means to predict particle characteristics even from combination of factors/levels that has not been studied experimentally. The proposed optimized method of preparation has been used to fabricate an optimized docetaxel-loaded PLGA nanoparticle formulation. The optimized nanoparticle formulation demonstrated suitable properties for intravenous sustained drug delivery purpose. The optimized nanoparticle formulation was further applied in a pharmacokinetic and biodistribution study and the corresponding data is presented elsewhere [21]. Acknowledgements This work was funded by a research grant from the Natural Sciences and Engineering Research Council of Canada (NSERC). Authors declare no conflict of interest. Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.msec.2019.109950. References [1] R.S. Herbst, F.R. Khuri, Mode of action of docetaxel - a basis for combination with novel anticancer agents, Cancer Treat. Rev. 29 (5) (2003) 407–415. [2] P. Zhao, D. Astruc, Docetaxel nanotechnology in anticancer therapy, ChemMedChem 7 (6) (2012) 952–972. [3] C. Takimoto, M. Beeram, Microtubule stabilizing agents in clinical oncology, the taxanes, in: T. Fojo (Ed.), The Role of Microtubules in Cell Biology, Neurobiology, and Oncology, Human Press, Totowa, USA, 2008, pp. 395–419. [4] R. Hitt, C. Rodríguez, M. Schwab (Ed.), Docetaxel, in Encyclopedia of Cancer, Springer-Verlag, Berlin Heidelberg, 2011, pp. 1148–1150. [5] S. Drori, G.D. Eytan, Y.G. Assaraf, Potentiation of anticancer-drug cytotoxicity by multidrug-resistance chemosensitizers involves alterations in membrane fluidity leading to increased membrane permeability, Eur. J. Biochem. 228 (3) (1995) 1020–1029. [6] van Zuylen, L., J. Verweij, and A. Sparreboom, Role of formulation vehicles in taxane pharmacology. Investig. New Drugs, 2001. 19(2): p. 125–41. [7] A.J. ten Tije, et al., Pharmacological effects of formulation vehicles : implications for cancer chemotherapy, Clin. Pharmacokinet. 42 (7) (2003) 665–685. [8] V. Sanna, et al., Novel docetaxel-loaded nanoparticles based on poly(lactide-cocaprolactone) and poly(lactide-co-glycolide-co-caprolactone) for prostate cancer treatment: formulation, characterization, and cytotoxicity studies, Nanoscale Res. Lett. 6 (1) (2011) 260. [9] E. Blanco, et al., Nanomedicine in cancer therapy: innovative trends and prospects, Cancer Sci. 102 (7) (2011) 1247–1252. [10] B. Haley, E. Frenkel, Nanoparticles for drug delivery in cancer treatment, Urol. Oncol. 26 (1) (2008) 57–64. [11] M. Conti, et al., Anticancer drug delivery with nanoparticles, In Vivo 20 (6A) (2006) 697–701. [12] K. Cho, et al., Therapeutic nanoparticles for drug delivery in cancer, Clin. Cancer Res. 14 (5) (2008) 1310–1316.

10

Materials Science & Engineering C 104 (2019) 109950

P. Rafiei and A. Haddadi

[55] H. Zhang, et al., Preparation of poly(lactide-co-glycolide-co-caprolactone) nanoparticles and their degradation behaviour in aqueous solution, Polym. Degrad. Stab. 91 (9) (2006) 1929–1936. [56] T. Riley, et al., Physicochemical evaluation of nanoparticles assembled from poly (lactic acid)−poly(ethylene glycol) (PLA−PEG) block copolymers as drug delivery vehicles, Langmuir 17 (11) (2001) 3168–3174. [57] S.A. Abouelmagd, et al., Release kinetics study of poorly water-soluble drugs from nanoparticles: are we doing it right? Mol. Pharm. 12 (3) (2015) 997–1003. [58] C. Fonseca, S. Simoes, R. Gaspar, Paclitaxel-loaded PLGA nanoparticles: preparation, physicochemical characterization and in vitro anti-tumoral activity, J. Control. Release 83 (2) (2002) 273–286. [59] M.L. Amin, P-glycoprotein inhibition for optimal drug delivery, Drug Target Insights 7 (2013) 27–34. [60] Y. Shirasaka, T. Sakane, S. Yamashita, Effect of P-glycoprotein expression levels on the concentration-dependent permeability of drugs to the cell membrane, J. Pharm. Sci. 97 (1) (2008) 553–565. [61] L.B. Sims, et al., Enhanced uptake and transport of PLGA-modified nanoparticles in cervical cancer, J Nanobiotechnology 14 (2016) 33.

[47] F. Esmaeili, et al., Cellular cytotoxicity and in-vivo biodistribution of docetaxel poly (lactide-co-glycolide) nanoparticles, Anti-Cancer Drugs 21 (1) (2010) 43–52. [48] C.G. Keum, et al., Practical preparation procedures for docetaxel-loaded nanoparticles using polylactic acid-co-glycolic acid, Int. J. Nanomedicine 6 (2011) 2225–2234. [49] J.W. Yoo, E. Chambers, S. Mitragotri, Factors that control the circulation time of nanoparticles in blood: challenges, solutions and future prospects, Curr. Pharm. Des. 16 (21) (2010) 2298–2307. [50] J.S. Zhang, F. Liu, L. Huang, Implications of pharmacokinetic behavior of lipoplex for its inflammatory toxicity, Adv. Drug Deliv. Rev. 57 (5) (2005) 689–698. [51] D.E. Owens Iii, N.A. Peppas, Opsonization, biodistribution, and pharmacokinetics of polymeric nanoparticles, Int. J. Pharm. 307 (1) (2006) 93–102. [52] T. Cedervall, et al., Detailed identification of plasma proteins adsorbed on copolymer nanoparticles, Angew Chem Int Ed Engl 46 (30) (2007) 5754–5756. [53] A. Gessner, et al., Nanoparticles with decreasing surface hydrophobicities: influence on plasma protein adsorption, Int. J. Pharm. 196 (2) (2000) 245–249. [54] S. Nagayama, et al., Time-dependent changes in opsonin amount associated on nanoparticles alter their hepatic uptake characteristics, Int. J. Pharm. 342 (1–2) (2007) 215–221.

11