- Email: [email protected]
Pyrazinamide-loaded poly(lactide-co-glycolide) nanoparticles: Optimization by experimental design
Journal of Drug Delivery Science and Technology xxx (2015) 1e7
Contents lists available at ScienceDirect
Journal of Drug Delivery Science and Technology journal homepage: www.elsevier.com/locate/jddst
Pyrazinamide-loaded poly(lactide-co-glycolide) nanoparticles: Optimization by experimental design Dinh-Duy Pham a, b, **, Elias Fattal a, Nicolas Tsapis a, * a b
^tenay-Malabry, France Univ Paris-Sud, Institut Galien Paris-Sud, CNRS UMR 8612, LabEx LERMIT, Cha University of Medicine and Pharmacy, Faculty of Pharmacy, Pharmaceutics Department, Ho Chi Minh City, VietNam
a r t i c l e i n f o
a b s t r a c t
Article history: Received 4 June 2015 Received in revised form 12 July 2015 Accepted 12 July 2015 Available online xxx
Pyrazinamide (PZA) plays a unique role in shortening Mycobacterium tuberculosis (TB) treatment compared to other first-line anti-TB drugs. We optimize PZA encapsulation into poly(lactide-coglycolide) (PLGA) nanoparticles through Taguchi experimental design in a strategy to improve drug bioavailability, reduce the dosing frequency and improve patient adherence to prescribed therapy, one of the critical obstacles in the control of TB epidemics. PZA-loaded PLGA nanoparticles were prepared by the double emulsion method. The Taguchi method, a statistical design with an orthogonal array, was implemented to optimize the formulation parameters. The solvent type (dichloromethane and ethyl acetate), the PZA/PLGA weight ratio and the organic/aqueous phase volume ratio were chosen as significant parameters affecting the particle size and polydispersity index (PDI). Taguchi method proved to be a quick, valuable tool in optimizing the formulation. The optimized experimental values for the nanoparticle size and PDI were about 170 nm and ˂ 0.1. Optimized PLGA nanoparticles possessed a zeta potential of about 1 mV, an encapsulation efficacy of 7e8% and a drug loading of 3.1%. They are suitable to be tested in vivo for TB treatment. © 2015 Elsevier B.V. All rights reserved.
Keywords: Nanoparticles Pyrazinamide Poly(lactide-co-glycolide) Taguchi Experimental design
1. Introduction Pyrazinamide (PZA) is an important first-line antitubercular (anti-TB) drug used in combination with other anti-TB drugs for the treatment of drug susceptible TB and multi-drug resistant TB (MDR-TB) [1]. PZA plays an essential role in shortening Mycobacterium tuberculosis treatment to the current standard of 6 months, often referred to as short course chemotherapy [2]. Indeed, PZA enters M. tuberculosis in lesions by passive diffusion, is converted to pyrazinoic acid (POA) by pyrazinamidase and is then excreted by a weak efflux pump. The protonated POA diffuses back into the bacilli under acidic conditions and accumulates due to the poor efficacy of the efflux pump, causing membrane damage [3]. Therefore PZA is more active against old non-growing bacilli than against young actively growing tubercle bacilli. PZA is usually given at a dose of
de Pharmacie, 5 * Corresponding author. Univ Paris Sud, UMR CNRS 8612, Faculte ment, 92296 Ch^ Rue Jean-Baptiste Cle atenay-Malabry, France. ** Corresponding author. University of Medicine and Pharmacy, Faculty of Pharmacy, 41-43 Dinh Tien Hoang, District 1, Ho Chi Minh City, Vietnam. E-mail addresses: [email protected] (D.-D. Pham), nicolas.tsapis@ u-psud.fr (N. Tsapis).
20e25 mg/kg daily corresponding to 1.5e2 g depending of patient weight [4]. Although PZA bioavailability is higher than 90%, a high oral dose over long periods of time leads to hepatotoxicity [2,5]. PZA lung delivery therefore appears as an interesting alternative to the oral route since it avoids the hepatic first-pass effect and might help concentrating the drug at its site of action and therefore reduce hepatic side effects. PZA could be delivered directly to the lungs as a solution by nebulization but it requires drug stability in solution [6]. Dry powder inhalers (DPI) can be considered as they allow to deliver efficiently large amounts of drugs [7e9]. However, recent work has shown that the amount of PZA found in alveolar macrophages after PZA powder delivery to the lungs is rather low due to premature dissolution [10]. In the recent years, nanoparticle-based drug delivery systems have been applied for anti-TB drugs with the goal of improving drug bioavailability, reducing the dosing frequency and preventing non adherence to prescribed therapy, a critical obstacle in the control of TB epidemics [11,12]. Nebulization of nanoparticles encapsulating PZA or administration of these nanoparticles as powders made of Trojan particles [13,14] could represent an interesting alternative. Indeed, due to their small size, nanoparticles can be phagocytosed by macrophages and therefore
http://dx.doi.org/10.1016/j.jddst.2015.07.006 1773-2247/© 2015 Elsevier B.V. All rights reserved.
Please cite this article in press as: D.-D. Pham, et al., Pyrazinamide-loaded poly(lactide-co-glycolide) nanoparticles: Optimization by experimental design, Journal of Drug Delivery Science and Technology (2015), http://dx.doi.org/10.1016/j.jddst.2015.07.006
2
D.-D. Pham et al. / Journal of Drug Delivery Science and Technology xxx (2015) 1e7
facilitate drug entrance in these diseased cells [12,15]. Poly(lactideco-glycolide) (PLGA) is one of the most common biodegradable and biocompatible polymer used to formulate nanoparticles (NPs). Encapsulation of anti-TB drugs into NPs could help maintaining drug levels in various tissues for long periods of time, but also could reduce adverse effects while exhibiting similar efficacy at reduced dose compared to conventional drug administration [12,15e17]. Moreover, PLGA is an interesting polymer since it exerts a rather low toxicity towards the pulmonary epithelium and does not induce a strong immune response both in vitro on cell culture models [18e22] and in vivo in rodents [23]. PLGA NPs are often formulated using the double emulsion solvent evaporation method to encapsulate water soluble drugs [24]. The physicochemical properties of NPs are affected by various parameters such as drug/ polymer ratio, surfactant concentration, volume ratio of organic phase and outer aqueous phase and sonication time to obtain desired NP size with maximum drug entrapment [25,26]. The effect of the mentioned parameters is complex and optimization of these factors is time and labor consuming. Taguchi method using statistical experimental design represents an easy and rapid strategy to develop an optimal formulation using minimum raw materials [27]. It uses the table of orthogonal arrays and analysis of variance (ANOVA) as analysis tools. ANOVA can estimate the influence of a factor on NP properties. While conventional statistical experimental design determines the optimal conditions based on the measured values of the characteristic properties, Taguchi method determines the experimental conditions having the least variability as the optimum condition [27]. In this study, we investigated parameters affecting the size of PZA-loaded PLGA nanoparticles prepared by the double emulsion solvent evaporation method. Parameters were optimized using the Taguchi method, by considering equal contributions of variables. The resulting formulation was characterized by evaluating NP size, zeta potential, encapsulation efficacy and drug loading. 2. Materials and methods 2.1. Materials Pyrazinamide (PZA) with specified purity greater than 99%, poly(vinyl-alcohol) (PVA) (87e89% hydrolyzed, MW 30,000e70,000) and sodium chloride (NaCl) were obtained from SigmaeAldrich (France). Poly(D,L-lactide-co-glycolide) (PLGA 75:25) Resomer RG756 was purchased from BoehringereIngelheim (Germany). Dichloromethane (DCM) and Ethyl acetate (EA) in analytical grade and acetonitrile in high-performance liquid chromatography (HPLC) grade were all obtained from Carlo Erba Reagents (France). Water was purified using a RIOS/MilliQ system from Millipore (France). 2.2. Preparation of pyrazinamide-loaded nanoparticles PLGA nanoparticles encapsulating PZA were prepared by double emulsion solvent evaporation method. Briefly, 1 mL of an aqueous drug solution containing 10 mg PZA was first emulsified into 10 mL of solvent (DCM, EA or mixture of DCM:EA 50:50) containing PLGA (drug/polymer: 1:1e1:5 by weight) for 1 min by probe sonication (Digital Sonifier®, Branson, France; 116 W, 20 kHz) over an ice bath. The primary emulsion was poured into 10e200 mL of an aqueous solution including 0.5e2% PVA and 0.5% NaCl. This mixture was then sonicated (300 W, 20 kHz) for 3 min to form the second waterin-oil-in-water emulsion (W/O/W). The secondary emulsion was stirred continuously 6 h with a magnetic stir bar for complete removal of solvent at room temperature. PZA-loaded PLGA nanoparticles were recovered by centrifugation (40 000 rpm, 30 min),
washed twice with water to remove the excess of PVA from the nanoparticles. Finally, the NP suspension was freeze-dried with trehalose (25 mg/mL) as cryoprotectant and stored in the freezer.
2.3. Design of experiment Taguchi has developed a method based on “orthogonal array” experiments which reduce experimental variance for the experiment with “optimum settings” of control parameters. By combining design of experiments with optimization of control parameters, the Taguchi method allows to obtain optimal results [28]. In this study, to identify the best formulations, four important factors that influence the PLGA NP size and polydispersity index were the followings: solvent type, PZA/PLGA weight ratio, organic phase/outer aqueous phase volume ratio and PVA concentration. Three different levels for each parameter (Table 1) and a fractional factorial design, namely, a standard L9 orthogonal array were considered (Table 2). The effects of the proposed experiments on the responses were then analyzed by the Design Expert software (trial version 8.0.6, Stat-Ease, Minneapolis, USA), to obtain independently the main effects of these factors, followed by the analysis of variance (ANOVA) to determine which factors were statistically significant. The optimum conditions were determined by desirability function.
2.4. Nanoparticle characterization 2.4.1. NP size distribution and zeta potential The drug-loaded PLGA NPs were analysed for their size and polydispersity index (PI) using a Zetasizer ZS (Malvern Instruments, France), based on photon correlation spectroscopy. Size measurements were performed in triplicate following a 1/100 v/v dilution of the NP suspension in purified water at 25 C at a fixed angle of 173 . The polydispersity index range was comprised between 0 and 1. Zeta potential was measured using the same instrument with a disposable zeta cuvette at 25 C following a 1/50 v/v dilution in a 1 mM KCl solution. For each sample, the mean diameter/Zeta potential ± standard deviation of three determinations were calculated applying multimodal analysis.
2.4.2. Encapsulation efficacy (EE%) and drug loading (DL%) within PLGA nanoparticles The amount of PZA encapsulated within PLGA nanoparticles was determined by measuring the amount of non encapsulated PZA in the aqueous solution recovered after ultracentrifugation and washing of the particles. PZA was analyzed by UV spectrophotometer at 268 nm. The drug encapsulation efficiency was expressed as the percentage of drug entrapped with respect to the initial amount, and the drug loading was expressed as the amount of drug entrapped per gram of polymer þ drug. Table 1 Definition and trial levels of studied variables in Taguchi's L9 orthogonal array experiment. Studied variables
Levels 1
2
3
Solvent type PZA/PLGA weight ratio Organic phase/aqueous phase volume ratio PVA concentration (%)
DCM 1:1 1:1 0.5
DCM:EA 1:2.5 1:10 1
EA 1:5 1:20 2
Please cite this article in press as: D.-D. Pham, et al., Pyrazinamide-loaded poly(lactide-co-glycolide) nanoparticles: Optimization by experimental design, Journal of Drug Delivery Science and Technology (2015), http://dx.doi.org/10.1016/j.jddst.2015.07.006
D.-D. Pham et al. / Journal of Drug Delivery Science and Technology xxx (2015) 1e7
3
Table 2 Particle size and polydispersity index of different PLGA NPs. Runs
Solvent type
PZA/PLGA weight ratio
Organic phase/aqueous phase volume ratio
PVA concentration (%)
Average particle size (nm)
1 2 3 4 5 6 7 8 9
EA EA DCM:EA DCM DCM DCM:EA DCM EA DCM:EA
1:2.5 1:1 1:1 1:2.5 1:5 1:2.5 1:1 1:5 1:5
1:1 1:20 1:10 1:10 1:20 1:20 1:1 1:10 1:1
2 1 2 1 2 0.5 0.5 0.5 1
146 215 79 133 392 152 467 88 387
EE% ¼
initial amount of drugðmgÞ free amount of drugðmgÞ initial amount of drugðmgÞ 100
DL% ¼
initial amount of drugðmgÞ free amount of drugðmgÞ total amount of polymerðmgÞ þ encapsulated drugðmgÞ 100
2.4.3. Scanning electron microscopy The morphology and shape of nanoparticles were examined by scanning electron microscopy (SEM) using a MERLIN microscope (Carl Zeiss) and operating at 3 kV with a filament current of about 0.1 mA. Liquid samples were deposited on a carbon conductive double-sided tape (Euromedex, France). They were coated with a palladiumeplatinum layer of about 4 nm, using a Cressington sputter-coater 208HR with a rotary planetary-tilt stage, equipped with a MTM-20 thickness controller. 3. Results and discussions 3.1. Taguchi array design and analysis of variance The Taguchi orthogonal array design is used to identify the optimal conditions and to select the parameters having the principal influence on the average particle size and polydispersity index. Table 2 shows the structure of Taguchi orthogonal array design and the results of measurement of the average particle size and polydispersity index. The experimental results were then analyzed by the Design Expert software to extract independently the main effects of these factors, followed by the analysis of variance (ANOVA) to determine which factors were statistically significant. The average particle size and polydispersity index of PLGA nanoparticles were considered as input into the Design Expert software for further analysis. To ensure a good model, test for significance of the regression model and test for significance on individual model coefficients were performed. An ANOVA table is commonly used to summarize the tests performed. Tables 3 and 4
± ± ± ± ± ± ± ± ±
4 3 4 3 28 40 52 1 29
Polydispersity index 0.10 0.54 0.56 0.13 0.17 0.25 0.51 0.17 0.22
± ± ± ± ± ± ± ± ±
0.03 0.13 0.11 0.04 0.09 0.04 0.03 0.02 0.24
present the ANOVA tables for the Taguchi method. When the input data were analyzed by using ANOVA, results showed that PVA concentration did not affect the average particle size and PDI. Therefore, this parameter does not appear in the ANOVA analysis. For the average particle size, Table 3, the value of “Prob > F” for the model is lower than 0.05, which indicates that the model is significant. Solvent type and organic phase/outer aqueous phase volume ratio are significant model terms, whereas PZA/PLGA weight ratio can be considered not significant. This not significant variable can be removed and may result in an improved model. The R2-value calculated is 0.99, reasonably close to 1, which is acceptable. It implies that about 99% of the variability in the data is explained by the model. The predicted R2-value of 0.71 is not as close to the adjusted R2 of 0.94 as one might normally expect. This may indicate a large block effect or a possible problem with the model and/or data. However, the difference of both values is not too large (about 0.2), hence, we still used this model for further analysis. The adjusted R2-value is particularly useful for comparing models with different number of terms. “adeq. precision” measures the signal to noise ratio. A ratio greater than 4 is desirable and indicates an adequate signal. Although the PZA/PLGA weight ratio has P-value greater than 0.05, this variable was still kept into the model because the model becomes significant keeping this parameter. For the polydispersity index, the original response values were transformed by the following equation: y0 ¼ (y þ k)l, where constant (k) ¼ 0 and lambda (l) ¼ 1.82. The transformed values were then analyzed similarly as average particle size. Transformation was needed as the error (residuals) was a function of the magnitude of the response (predicted values). In the ANOVA table (Table 4), the value of “Prob > F” for the model is lower than 0.05, which indicates that the model is significant. In this case, solvent type, PZA/PLGA weight ratio and organic phase/outer aqueous phase volume ratio all are significant model terms. The R2-value calculated is 0.99, reasonably close to 1, which is acceptable. The predicted R2-value of 0.99 is almost in reasonable agreement with the adjusted R2 of 0.99. “adeq. precision” greater than 4 is desirable and indicates an adequate signal. Fig. 1 shows the various effects of the independent variables on the average NP size and the polydispersity index. The NP size is found to decrease as DCM is mixed with EA or replaced by EA. NP size exhibits a minimum for the organic phase/outer aqueous phase
Table 3 ANOVA table for the average particle size (R2 ¼ 0.99; adjusted R2 ¼ 0.94; adeq precision ¼ 12.13 (> 4). Source
Sum of squares
Degrees of freedom
Mean square
F-value
P-value (Prob > F)
Model A-Solvent type B-PZA/PLGA weight ratio C-Organic phase/aqueous phase volume ratio Residual Cor Total
170001.5 51347.35 34441.46 84212.7 2510.056 172511.6
6 2 2 2 2 8
28333.59 25673.67 17220.73 42106.35 1255.028
22.5761 20.4567 13.7214 33.5501
0.0430 0.0466 0.0679 0.0289
Significant
Please cite this article in press as: D.-D. Pham, et al., Pyrazinamide-loaded poly(lactide-co-glycolide) nanoparticles: Optimization by experimental design, Journal of Drug Delivery Science and Technology (2015), http://dx.doi.org/10.1016/j.jddst.2015.07.006
4
D.-D. Pham et al. / Journal of Drug Delivery Science and Technology xxx (2015) 1e7
Table 4 ANOVA table for the polydispersity index (R2 ¼ 0.99; adjusted R2 ¼ 0.99; adeq precision ¼ 121.07 (> 4). Source
Sum of squares
Degrees of freedom
Mean square
F-value
P-value (Prob > F)
Model A-Solvent type B-PZA/PLGA weight ratio C-Organic phase/aqueous phase volume ratio Residual Cor Total
0.1601 0.0035 0.1554 0.0012 0.0000 0.1601
6 2 2 2 2 8
0.0267 0.0018 0.0777 0.0006 0.0000
2724.38 179.59 7933.69 59.85
0.0004 0.0055 0.0001 0.0164
Significant
Fig. 1. Effects of the independent variables on the average particle size and polydispersity index.
volume ratio of 1:10. The polydispersity index increases significantly when using a mixture of DCM and EA as well as when the organic phase/outer aqueous phase volume ratio decreases, whereas this index decreases strongly as PZA/PLGA weight ratio decreases. Computer optimization process by Design Expert software and a
desirability function determined the effect of the levels of independent parameters on the responses. All responses were fitted to the linear model. The constraints of independent variables and the responses are presented in Table 5. Based on the data of the particle size and polydispersity index and on the ANOVA analysis, Design Expert software proposed
Please cite this article in press as: D.-D. Pham, et al., Pyrazinamide-loaded poly(lactide-co-glycolide) nanoparticles: Optimization by experimental design, Journal of Drug Delivery Science and Technology (2015), http://dx.doi.org/10.1016/j.jddst.2015.07.006
D.-D. Pham et al. / Journal of Drug Delivery Science and Technology xxx (2015) 1e7
5
Table 5 Constraints of independent variables and responses.
Solvent type PZA/PLGA weight ratio Organic phase/aqueous phase volume ratio PVA concentration (%) Average particle size (nm) Polydispersity index
optimum solutions for formulating PZA-loaded PLGA nanoparticles (Table 6). Optimization was performed by using a desirability function to obtain the levels of investigated factors, which minimized the particle size and polydispersity index (Table 6). The predicted optimal conditions of solution 1 with a desirability factor equal to 87% were chosen to prepare PZA-loaded PLGA nanoparticles since they yield simultaneously the lower PDI and the lower size. Nanoparticles were then prepared using ethyl acetate as a solvent, a 1:2.5 PZA/PLGA weight ratio a 1:1 organic/aqueous phase volume ratio and 1% PVA as a surfactant. The experimental result of average particle size showed an acceptable agreement with the model (Table 7). There is as large difference between the predicted and experimental polydispersity index. However, the experimental polydispersity index is lower than the model, proving that nanoparticles are rather monodisperse. We believe the high value might arise from the transformation of the PDI needed for predictions. 3.2. Nanoparticle characterization The optimal formulation was then prepared in duplicate and characterized. The physicochemical properties of the optimal formulation of PZA-loaded PLGA NPs are shown in Table 8. Particle size distribution and zeta potential are presented in Fig. 2. SEM images show rather monodisperse nanoparticles below 200 nm in agreement with DLS measurements (Fig. 3). PLGA nanoparticles fabricated by double emulsion solvent evaporation method have been commonly applied for encapsulating hydrophilic molecules [29]. Typically, the size of the PLGA nanoparticles is in the range of 100e250 nm. Here, as the classical solvent (dichloromethane) was replaced by a partially water-miscible organic solvent (ethyl acetate), the nanoparticle size was lower than 200 nm [30]. Particle size is in a suitable range for macrophage uptake after pulmonary delivery [31]. The zeta potential of nanoparticles was found about 1 mV (Fig 2) in agreement with values reported in the literature for PVA coated nanoparticles [18,30]. Particles being
Goal
Limited range
In range Equal to 1:2.5 Equal to 1:1 Term not in model Minimize Minimize
DCM e EA 1:1e1:5 1:1e1:20 79e467 0.10e0.56
Table 7 Predicted and experimental results for particle size and polydispersity index of PZAloaded PLGA nanoparticles.
Prediction Experimental Error %
Average particle size (nm)
Polydispersity index
170 ± 35 173 ± 4 1.8
0.11 ± 0.01 0.05 ± 0.03 54.5
neutral should interact with lung cells or mucus [19,21]. The encapsulation efficiency of PZA in the optimized formulation was about 8%, leading to a drug loading of 3.1%, in the range usually found in the literature [32]. For encapsulating small hydrophilic molecules in PLGA nanoparticles, the double emulsion solvent evaporation method often suffers from limited encapsulation efficiency because of the drug rapid partitioning to the external aqueous phase [30]. Optimization of the drug loading could be obtained for example by playing with the difference of osmotic pressure between in the internal and external aqueous phases of the double emulsion.
4. Conclusion The Taguchi L9 method appears to be a fast, simple and valuable tool in optimizing the various parameters for the preparation of PZA-loaded PLGA nanoparticles. The organic solvent was found to be the main parameter having significant effects on the particle size and polydispersity index and ethyl acetate was chosen. The optimal formulation was prepared using a 1:2.5 PZA/PLGA weight ratio, a 1:1 organic/aqueous phase volume ratio and 1% PVA as a surfactant. A good agreement was obtained between the predicted optimized Taguchi method and the experimental data. The suitability of these PZA-loaded nanoparticles for pulmonary delivery to treat TB should be tested in the future.
Table 6 Optimum solutions for formulating PZA-loaded PLGA nanoparticles. The column highlighted in blue corresponds to the chosen solution.
Please cite this article in press as: D.-D. Pham, et al., Pyrazinamide-loaded poly(lactide-co-glycolide) nanoparticles: Optimization by experimental design, Journal of Drug Delivery Science and Technology (2015), http://dx.doi.org/10.1016/j.jddst.2015.07.006
6
D.-D. Pham et al. / Journal of Drug Delivery Science and Technology xxx (2015) 1e7
Table 8 Characteristics of PZA loaded PLGA nanoparticles. Lot#
Particle size (nm)
PDI
Zeta potential (mV)
Encapsulation efficacy (%)
Drug loading (%)
1 2
170 ± 2 175 ± 3
0.03 ± 0.01 0.07 ± 0.00
1.08 ± 0.14 1.10 ± 0.06
8.8 ± 4.7 7.1 ± 0.8
3.4 ± 1.7 2.8 ± 0.3
Fig. 2. Optimized Nanoparticles Formulation. The top graph represents Particle Size Distribution, the bottom graph represents the Zeta Potential distribution.
Fig. 3. Typical SEM images of optimized nanoparticle formulation.
Acknowledgments D. D. Pham fellowship is supported by Project 322, Ministry of Education and Training, Vietnam. Institut Galien Paris-Sud is a member of the Laboratory of Excellence LERMIT supported by a grant from ANR (ANR-10-LABX-33). References [1] WHO, Global Tuberculosis Control, 2011. [2] D.A. Mitchison, P.B. Fourie, The near future: improving the activity of rifamycins and pyrazinamide, Tuberc. Edinb. 90 (3) (2010) 177e181. [3] Y. Zhang, et al., Mode of action of pyrazinamide: disruption of Mycobacterium tuberculosis membrane transport and energetics by pyrazinoic acid, J. Antimicrob. Chemother. 52 (5) (2003) 790e795. [4] Pyrazinamide, Tuberculosis 88 (2) (2008) 141e144.
[5] R.M. McCune Jr., R. Tompsett, Fate of Mycobacterium tuberculosis in mouse tissues as determined by the microbial enumeration technique. I. The persistence of drug-susceptible tubercle bacilli in the tissues despite prolonged antimicrobial therapy, J. Exp. Med. 104 (5) (1956) 737e762. [6] H.I. Seifart, D.P. Parkin, P.R. Donald, Stability of isoniazid, rifampin and pyrazinamide in suspensions used for the treatment of tuberculosis in children, Pediatr. Infect. Dis. J. 10 (11) (1991) 827e831. [7] H.W. Frijlink, A.H. DE Boer, Dry powder inhalers for pulmonary drug delivery, Expert Opin. Drug Deliv. 1 (1) (2004) 67e86. [8] D.D. Pham, et al., Formulation of pyrazinamide-loaded large porous particles for the pulmonary route: avoiding crystal growth using excipients, Int. J. Pharm. 454 (2) (2013) 668e677. [9] D.D. Pham, E. Fattal, N. Tsapis, Pulmonary drug delivery systems for tuberculosis treatment, Int. J. Pharm. 478 (2) (2015) 517e529. [10] D.D. Pham, et al., Pulmonary delivery of pyrazinamide-loaded large porous particles, Eur. J. Pharm. Biopharm. 94 (2015) 241e250. [11] S. Gelperina, et al., The potential advantages of nanoparticle drug delivery systems in chemotherapy of tuberculosis, Am. J. Respir. Crit. Care Med. 172 (12) (2005) 1487e1490. [12] F. Andrade, et al., Nanotechnology and pulmonary delivery to overcome resistance in infectious diseases, Adv. Drug Deliv. Rev. 65 (13e14) (2013) 1816e1827. [13] N. Tsapis, et al., Trojan particles: large porous carriers of nanoparticles for drug delivery, Proc. Natl. Acad. Sci. U. S. A. 99 (19) (2002) 12001e12005. [14] C. Gomez-Gaete, et al., Dexamethasone acetate encapsulation into Trojan particles, J. Control. Release 128 (1) (2008) 41e49. [15] P. Muralidharan, et al., Inhalable nanoparticulate powders for respiratory delivery, Nanomedicine 11 (5) (2015) 1189e1199. [16] R. Pandey, S. Sharma, G.K. Khuller, Oral poly(lactide-co-glycolide) nanoparticle based antituberculosis drug delivery: toxicological and chemotherapeutic implications, Indian J. Exp. Biol. 44 (6) (2006) 459e467. [17] R. Pandey, et al., Poly (DL-lactide-co-glycolide) nanoparticle-based inhalable sustained drug delivery system for experimental tuberculosis, J. Antimicrob. Chemother. 52 (6) (2003) 981e986. [18] S. Mura, et al., Influence of surface charge on the potential toxicity of PLGA nanoparticles towards Calu-3 cells, Int. J. Nanomed. 6 (2011) 2591e2605. [19] S. Mura, et al., Biodegradable nanoparticles meet the bronchial airway barrier: how surface properties affect their interaction with mucus and epithelial cells,
Please cite this article in press as: D.-D. Pham, et al., Pyrazinamide-loaded poly(lactide-co-glycolide) nanoparticles: Optimization by experimental design, Journal of Drug Delivery Science and Technology (2015), http://dx.doi.org/10.1016/j.jddst.2015.07.006
D.-D. Pham et al. / Journal of Drug Delivery Science and Technology xxx (2015) 1e7 Biomacromolecules 12 (11) (2011) 4136e4143. [20] E. Fattal, et al., Lung toxicity of biodegradable nanoparticles, J. Biomed. Nanotechnol. 10 (10) (2014) 2852e2864. [21] N. Grabowski, et al., Toxicity of surface-modified PLGA nanoparticles toward lung alveolar epithelial cells, Int. J. Pharm. 454 (2) (2013) 686e694. [22] N. Grabowski, et al., Surface coating mediates the toxicity of polymeric nanoparticles towards human-like macrophages, Int. J. Pharm. 482 (1e2) (2015) 75e83. [23] Santiago, L.A., et al., Compared in vivo toxicity in mice of lung delivered biodegradable and non biodegradable nanoparticles. Nanotoxicology, 2015. [in press]. [24] S. Malathi, S. Balasubramanian, Synthesis of biodegradable polymeric nanoparticles and their controlled drug delivery for tuberculosis, J. Biomed. Nanotechnol. 7 (1) (2011) 150e151. [25] D.D. Ankola, et al., Preparation, characterization and in silico modeling of biodegradable nanoparticles containing cyclosporine A and coenzyme Q10, Nanotechnology 21 (6) (2010) 065104. [26] A. Tripathi, R. Gupta, S.A. Saraf, PLGA nanoparticles of anti tubercular drug:
[27]
[28]
[29] [30]
[31]
[32]
7
drug loading and release studies of a water in-soluble drug, Int. J. Pharm. Tech. Res. 2 (3) (2010) 2116e2123. K.D. Kim, et al., Optimization of parameters for the synthesis of zinc oxide nanoparticles by Taguchi robust design method, Colloid Surf. A Physicochem. Eng. Asp. 311 (1e3) (2007) 170e173. J.-C. Lu, S.-L. Jeng, K. Wang, A Review of Statistical Methods for Quality Improvement and Control in Nanotechnology, vol. 41, American Society for Quality, Milwaukee, WI, ETATS-UNIS, 2009, p. 17. F. Danhier, et al., PLGA-based nanoparticles: an overview of biomedical applications, J. Control. Release 161 (2) (2012) 505e522. E. Cohen-Sela, et al., A new double emulsion solvent diffusion technique for encapsulating hydrophilic molecules in PLGA nanoparticles, J. Control. Release 133 (2) (2009) 90e95. V. Kanchan, A.K. Panda, Interactions of antigen-loaded polylactide particles with macrophages and their correlation with the immune response, Biomaterials 28 (35) (2007) 5344e5357. E. Fattal, N. Tsapis, Nanomedicine technology: current achievements and new trends, Clin. Transl. Imag. Rev. Nucl. Med. Mol. Imag. 2 (2014) 77e87.
Please cite this article in press as: D.-D. Pham, et al., Pyrazinamide-loaded poly(lactide-co-glycolide) nanoparticles: Optimization by experimental design, Journal of Drug Delivery Science and Technology (2015), http://dx.doi.org/10.1016/j.jddst.2015.07.006
Contents lists available at ScienceDirect
Journal of Drug Delivery Science and Technology journal homepage: www.elsevier.com/locate/jddst
Pyrazinamide-loaded poly(lactide-co-glycolide) nanoparticles: Optimization by experimental design Dinh-Duy Pham a, b, **, Elias Fattal a, Nicolas Tsapis a, * a b
^tenay-Malabry, France Univ Paris-Sud, Institut Galien Paris-Sud, CNRS UMR 8612, LabEx LERMIT, Cha University of Medicine and Pharmacy, Faculty of Pharmacy, Pharmaceutics Department, Ho Chi Minh City, VietNam
a r t i c l e i n f o
a b s t r a c t
Article history: Received 4 June 2015 Received in revised form 12 July 2015 Accepted 12 July 2015 Available online xxx
Pyrazinamide (PZA) plays a unique role in shortening Mycobacterium tuberculosis (TB) treatment compared to other first-line anti-TB drugs. We optimize PZA encapsulation into poly(lactide-coglycolide) (PLGA) nanoparticles through Taguchi experimental design in a strategy to improve drug bioavailability, reduce the dosing frequency and improve patient adherence to prescribed therapy, one of the critical obstacles in the control of TB epidemics. PZA-loaded PLGA nanoparticles were prepared by the double emulsion method. The Taguchi method, a statistical design with an orthogonal array, was implemented to optimize the formulation parameters. The solvent type (dichloromethane and ethyl acetate), the PZA/PLGA weight ratio and the organic/aqueous phase volume ratio were chosen as significant parameters affecting the particle size and polydispersity index (PDI). Taguchi method proved to be a quick, valuable tool in optimizing the formulation. The optimized experimental values for the nanoparticle size and PDI were about 170 nm and ˂ 0.1. Optimized PLGA nanoparticles possessed a zeta potential of about 1 mV, an encapsulation efficacy of 7e8% and a drug loading of 3.1%. They are suitable to be tested in vivo for TB treatment. © 2015 Elsevier B.V. All rights reserved.
Keywords: Nanoparticles Pyrazinamide Poly(lactide-co-glycolide) Taguchi Experimental design
1. Introduction Pyrazinamide (PZA) is an important first-line antitubercular (anti-TB) drug used in combination with other anti-TB drugs for the treatment of drug susceptible TB and multi-drug resistant TB (MDR-TB) [1]. PZA plays an essential role in shortening Mycobacterium tuberculosis treatment to the current standard of 6 months, often referred to as short course chemotherapy [2]. Indeed, PZA enters M. tuberculosis in lesions by passive diffusion, is converted to pyrazinoic acid (POA) by pyrazinamidase and is then excreted by a weak efflux pump. The protonated POA diffuses back into the bacilli under acidic conditions and accumulates due to the poor efficacy of the efflux pump, causing membrane damage [3]. Therefore PZA is more active against old non-growing bacilli than against young actively growing tubercle bacilli. PZA is usually given at a dose of
de Pharmacie, 5 * Corresponding author. Univ Paris Sud, UMR CNRS 8612, Faculte ment, 92296 Ch^ Rue Jean-Baptiste Cle atenay-Malabry, France. ** Corresponding author. University of Medicine and Pharmacy, Faculty of Pharmacy, 41-43 Dinh Tien Hoang, District 1, Ho Chi Minh City, Vietnam. E-mail addresses: [email protected] (D.-D. Pham), nicolas.tsapis@ u-psud.fr (N. Tsapis).
20e25 mg/kg daily corresponding to 1.5e2 g depending of patient weight [4]. Although PZA bioavailability is higher than 90%, a high oral dose over long periods of time leads to hepatotoxicity [2,5]. PZA lung delivery therefore appears as an interesting alternative to the oral route since it avoids the hepatic first-pass effect and might help concentrating the drug at its site of action and therefore reduce hepatic side effects. PZA could be delivered directly to the lungs as a solution by nebulization but it requires drug stability in solution [6]. Dry powder inhalers (DPI) can be considered as they allow to deliver efficiently large amounts of drugs [7e9]. However, recent work has shown that the amount of PZA found in alveolar macrophages after PZA powder delivery to the lungs is rather low due to premature dissolution [10]. In the recent years, nanoparticle-based drug delivery systems have been applied for anti-TB drugs with the goal of improving drug bioavailability, reducing the dosing frequency and preventing non adherence to prescribed therapy, a critical obstacle in the control of TB epidemics [11,12]. Nebulization of nanoparticles encapsulating PZA or administration of these nanoparticles as powders made of Trojan particles [13,14] could represent an interesting alternative. Indeed, due to their small size, nanoparticles can be phagocytosed by macrophages and therefore
http://dx.doi.org/10.1016/j.jddst.2015.07.006 1773-2247/© 2015 Elsevier B.V. All rights reserved.
Please cite this article in press as: D.-D. Pham, et al., Pyrazinamide-loaded poly(lactide-co-glycolide) nanoparticles: Optimization by experimental design, Journal of Drug Delivery Science and Technology (2015), http://dx.doi.org/10.1016/j.jddst.2015.07.006
2
D.-D. Pham et al. / Journal of Drug Delivery Science and Technology xxx (2015) 1e7
facilitate drug entrance in these diseased cells [12,15]. Poly(lactideco-glycolide) (PLGA) is one of the most common biodegradable and biocompatible polymer used to formulate nanoparticles (NPs). Encapsulation of anti-TB drugs into NPs could help maintaining drug levels in various tissues for long periods of time, but also could reduce adverse effects while exhibiting similar efficacy at reduced dose compared to conventional drug administration [12,15e17]. Moreover, PLGA is an interesting polymer since it exerts a rather low toxicity towards the pulmonary epithelium and does not induce a strong immune response both in vitro on cell culture models [18e22] and in vivo in rodents [23]. PLGA NPs are often formulated using the double emulsion solvent evaporation method to encapsulate water soluble drugs [24]. The physicochemical properties of NPs are affected by various parameters such as drug/ polymer ratio, surfactant concentration, volume ratio of organic phase and outer aqueous phase and sonication time to obtain desired NP size with maximum drug entrapment [25,26]. The effect of the mentioned parameters is complex and optimization of these factors is time and labor consuming. Taguchi method using statistical experimental design represents an easy and rapid strategy to develop an optimal formulation using minimum raw materials [27]. It uses the table of orthogonal arrays and analysis of variance (ANOVA) as analysis tools. ANOVA can estimate the influence of a factor on NP properties. While conventional statistical experimental design determines the optimal conditions based on the measured values of the characteristic properties, Taguchi method determines the experimental conditions having the least variability as the optimum condition [27]. In this study, we investigated parameters affecting the size of PZA-loaded PLGA nanoparticles prepared by the double emulsion solvent evaporation method. Parameters were optimized using the Taguchi method, by considering equal contributions of variables. The resulting formulation was characterized by evaluating NP size, zeta potential, encapsulation efficacy and drug loading. 2. Materials and methods 2.1. Materials Pyrazinamide (PZA) with specified purity greater than 99%, poly(vinyl-alcohol) (PVA) (87e89% hydrolyzed, MW 30,000e70,000) and sodium chloride (NaCl) were obtained from SigmaeAldrich (France). Poly(D,L-lactide-co-glycolide) (PLGA 75:25) Resomer RG756 was purchased from BoehringereIngelheim (Germany). Dichloromethane (DCM) and Ethyl acetate (EA) in analytical grade and acetonitrile in high-performance liquid chromatography (HPLC) grade were all obtained from Carlo Erba Reagents (France). Water was purified using a RIOS/MilliQ system from Millipore (France). 2.2. Preparation of pyrazinamide-loaded nanoparticles PLGA nanoparticles encapsulating PZA were prepared by double emulsion solvent evaporation method. Briefly, 1 mL of an aqueous drug solution containing 10 mg PZA was first emulsified into 10 mL of solvent (DCM, EA or mixture of DCM:EA 50:50) containing PLGA (drug/polymer: 1:1e1:5 by weight) for 1 min by probe sonication (Digital Sonifier®, Branson, France; 116 W, 20 kHz) over an ice bath. The primary emulsion was poured into 10e200 mL of an aqueous solution including 0.5e2% PVA and 0.5% NaCl. This mixture was then sonicated (300 W, 20 kHz) for 3 min to form the second waterin-oil-in-water emulsion (W/O/W). The secondary emulsion was stirred continuously 6 h with a magnetic stir bar for complete removal of solvent at room temperature. PZA-loaded PLGA nanoparticles were recovered by centrifugation (40 000 rpm, 30 min),
washed twice with water to remove the excess of PVA from the nanoparticles. Finally, the NP suspension was freeze-dried with trehalose (25 mg/mL) as cryoprotectant and stored in the freezer.
2.3. Design of experiment Taguchi has developed a method based on “orthogonal array” experiments which reduce experimental variance for the experiment with “optimum settings” of control parameters. By combining design of experiments with optimization of control parameters, the Taguchi method allows to obtain optimal results [28]. In this study, to identify the best formulations, four important factors that influence the PLGA NP size and polydispersity index were the followings: solvent type, PZA/PLGA weight ratio, organic phase/outer aqueous phase volume ratio and PVA concentration. Three different levels for each parameter (Table 1) and a fractional factorial design, namely, a standard L9 orthogonal array were considered (Table 2). The effects of the proposed experiments on the responses were then analyzed by the Design Expert software (trial version 8.0.6, Stat-Ease, Minneapolis, USA), to obtain independently the main effects of these factors, followed by the analysis of variance (ANOVA) to determine which factors were statistically significant. The optimum conditions were determined by desirability function.
2.4. Nanoparticle characterization 2.4.1. NP size distribution and zeta potential The drug-loaded PLGA NPs were analysed for their size and polydispersity index (PI) using a Zetasizer ZS (Malvern Instruments, France), based on photon correlation spectroscopy. Size measurements were performed in triplicate following a 1/100 v/v dilution of the NP suspension in purified water at 25 C at a fixed angle of 173 . The polydispersity index range was comprised between 0 and 1. Zeta potential was measured using the same instrument with a disposable zeta cuvette at 25 C following a 1/50 v/v dilution in a 1 mM KCl solution. For each sample, the mean diameter/Zeta potential ± standard deviation of three determinations were calculated applying multimodal analysis.
2.4.2. Encapsulation efficacy (EE%) and drug loading (DL%) within PLGA nanoparticles The amount of PZA encapsulated within PLGA nanoparticles was determined by measuring the amount of non encapsulated PZA in the aqueous solution recovered after ultracentrifugation and washing of the particles. PZA was analyzed by UV spectrophotometer at 268 nm. The drug encapsulation efficiency was expressed as the percentage of drug entrapped with respect to the initial amount, and the drug loading was expressed as the amount of drug entrapped per gram of polymer þ drug. Table 1 Definition and trial levels of studied variables in Taguchi's L9 orthogonal array experiment. Studied variables
Levels 1
2
3
Solvent type PZA/PLGA weight ratio Organic phase/aqueous phase volume ratio PVA concentration (%)
DCM 1:1 1:1 0.5
DCM:EA 1:2.5 1:10 1
EA 1:5 1:20 2
Please cite this article in press as: D.-D. Pham, et al., Pyrazinamide-loaded poly(lactide-co-glycolide) nanoparticles: Optimization by experimental design, Journal of Drug Delivery Science and Technology (2015), http://dx.doi.org/10.1016/j.jddst.2015.07.006
D.-D. Pham et al. / Journal of Drug Delivery Science and Technology xxx (2015) 1e7
3
Table 2 Particle size and polydispersity index of different PLGA NPs. Runs
Solvent type
PZA/PLGA weight ratio
Organic phase/aqueous phase volume ratio
PVA concentration (%)
Average particle size (nm)
1 2 3 4 5 6 7 8 9
EA EA DCM:EA DCM DCM DCM:EA DCM EA DCM:EA
1:2.5 1:1 1:1 1:2.5 1:5 1:2.5 1:1 1:5 1:5
1:1 1:20 1:10 1:10 1:20 1:20 1:1 1:10 1:1
2 1 2 1 2 0.5 0.5 0.5 1
146 215 79 133 392 152 467 88 387
EE% ¼
initial amount of drugðmgÞ free amount of drugðmgÞ initial amount of drugðmgÞ 100
DL% ¼
initial amount of drugðmgÞ free amount of drugðmgÞ total amount of polymerðmgÞ þ encapsulated drugðmgÞ 100
2.4.3. Scanning electron microscopy The morphology and shape of nanoparticles were examined by scanning electron microscopy (SEM) using a MERLIN microscope (Carl Zeiss) and operating at 3 kV with a filament current of about 0.1 mA. Liquid samples were deposited on a carbon conductive double-sided tape (Euromedex, France). They were coated with a palladiumeplatinum layer of about 4 nm, using a Cressington sputter-coater 208HR with a rotary planetary-tilt stage, equipped with a MTM-20 thickness controller. 3. Results and discussions 3.1. Taguchi array design and analysis of variance The Taguchi orthogonal array design is used to identify the optimal conditions and to select the parameters having the principal influence on the average particle size and polydispersity index. Table 2 shows the structure of Taguchi orthogonal array design and the results of measurement of the average particle size and polydispersity index. The experimental results were then analyzed by the Design Expert software to extract independently the main effects of these factors, followed by the analysis of variance (ANOVA) to determine which factors were statistically significant. The average particle size and polydispersity index of PLGA nanoparticles were considered as input into the Design Expert software for further analysis. To ensure a good model, test for significance of the regression model and test for significance on individual model coefficients were performed. An ANOVA table is commonly used to summarize the tests performed. Tables 3 and 4
± ± ± ± ± ± ± ± ±
4 3 4 3 28 40 52 1 29
Polydispersity index 0.10 0.54 0.56 0.13 0.17 0.25 0.51 0.17 0.22
± ± ± ± ± ± ± ± ±
0.03 0.13 0.11 0.04 0.09 0.04 0.03 0.02 0.24
present the ANOVA tables for the Taguchi method. When the input data were analyzed by using ANOVA, results showed that PVA concentration did not affect the average particle size and PDI. Therefore, this parameter does not appear in the ANOVA analysis. For the average particle size, Table 3, the value of “Prob > F” for the model is lower than 0.05, which indicates that the model is significant. Solvent type and organic phase/outer aqueous phase volume ratio are significant model terms, whereas PZA/PLGA weight ratio can be considered not significant. This not significant variable can be removed and may result in an improved model. The R2-value calculated is 0.99, reasonably close to 1, which is acceptable. It implies that about 99% of the variability in the data is explained by the model. The predicted R2-value of 0.71 is not as close to the adjusted R2 of 0.94 as one might normally expect. This may indicate a large block effect or a possible problem with the model and/or data. However, the difference of both values is not too large (about 0.2), hence, we still used this model for further analysis. The adjusted R2-value is particularly useful for comparing models with different number of terms. “adeq. precision” measures the signal to noise ratio. A ratio greater than 4 is desirable and indicates an adequate signal. Although the PZA/PLGA weight ratio has P-value greater than 0.05, this variable was still kept into the model because the model becomes significant keeping this parameter. For the polydispersity index, the original response values were transformed by the following equation: y0 ¼ (y þ k)l, where constant (k) ¼ 0 and lambda (l) ¼ 1.82. The transformed values were then analyzed similarly as average particle size. Transformation was needed as the error (residuals) was a function of the magnitude of the response (predicted values). In the ANOVA table (Table 4), the value of “Prob > F” for the model is lower than 0.05, which indicates that the model is significant. In this case, solvent type, PZA/PLGA weight ratio and organic phase/outer aqueous phase volume ratio all are significant model terms. The R2-value calculated is 0.99, reasonably close to 1, which is acceptable. The predicted R2-value of 0.99 is almost in reasonable agreement with the adjusted R2 of 0.99. “adeq. precision” greater than 4 is desirable and indicates an adequate signal. Fig. 1 shows the various effects of the independent variables on the average NP size and the polydispersity index. The NP size is found to decrease as DCM is mixed with EA or replaced by EA. NP size exhibits a minimum for the organic phase/outer aqueous phase
Table 3 ANOVA table for the average particle size (R2 ¼ 0.99; adjusted R2 ¼ 0.94; adeq precision ¼ 12.13 (> 4). Source
Sum of squares
Degrees of freedom
Mean square
F-value
P-value (Prob > F)
Model A-Solvent type B-PZA/PLGA weight ratio C-Organic phase/aqueous phase volume ratio Residual Cor Total
170001.5 51347.35 34441.46 84212.7 2510.056 172511.6
6 2 2 2 2 8
28333.59 25673.67 17220.73 42106.35 1255.028
22.5761 20.4567 13.7214 33.5501
0.0430 0.0466 0.0679 0.0289
Significant
Please cite this article in press as: D.-D. Pham, et al., Pyrazinamide-loaded poly(lactide-co-glycolide) nanoparticles: Optimization by experimental design, Journal of Drug Delivery Science and Technology (2015), http://dx.doi.org/10.1016/j.jddst.2015.07.006
4
D.-D. Pham et al. / Journal of Drug Delivery Science and Technology xxx (2015) 1e7
Table 4 ANOVA table for the polydispersity index (R2 ¼ 0.99; adjusted R2 ¼ 0.99; adeq precision ¼ 121.07 (> 4). Source
Sum of squares
Degrees of freedom
Mean square
F-value
P-value (Prob > F)
Model A-Solvent type B-PZA/PLGA weight ratio C-Organic phase/aqueous phase volume ratio Residual Cor Total
0.1601 0.0035 0.1554 0.0012 0.0000 0.1601
6 2 2 2 2 8
0.0267 0.0018 0.0777 0.0006 0.0000
2724.38 179.59 7933.69 59.85
0.0004 0.0055 0.0001 0.0164
Significant
Fig. 1. Effects of the independent variables on the average particle size and polydispersity index.
volume ratio of 1:10. The polydispersity index increases significantly when using a mixture of DCM and EA as well as when the organic phase/outer aqueous phase volume ratio decreases, whereas this index decreases strongly as PZA/PLGA weight ratio decreases. Computer optimization process by Design Expert software and a
desirability function determined the effect of the levels of independent parameters on the responses. All responses were fitted to the linear model. The constraints of independent variables and the responses are presented in Table 5. Based on the data of the particle size and polydispersity index and on the ANOVA analysis, Design Expert software proposed
Please cite this article in press as: D.-D. Pham, et al., Pyrazinamide-loaded poly(lactide-co-glycolide) nanoparticles: Optimization by experimental design, Journal of Drug Delivery Science and Technology (2015), http://dx.doi.org/10.1016/j.jddst.2015.07.006
D.-D. Pham et al. / Journal of Drug Delivery Science and Technology xxx (2015) 1e7
5
Table 5 Constraints of independent variables and responses.
Solvent type PZA/PLGA weight ratio Organic phase/aqueous phase volume ratio PVA concentration (%) Average particle size (nm) Polydispersity index
optimum solutions for formulating PZA-loaded PLGA nanoparticles (Table 6). Optimization was performed by using a desirability function to obtain the levels of investigated factors, which minimized the particle size and polydispersity index (Table 6). The predicted optimal conditions of solution 1 with a desirability factor equal to 87% were chosen to prepare PZA-loaded PLGA nanoparticles since they yield simultaneously the lower PDI and the lower size. Nanoparticles were then prepared using ethyl acetate as a solvent, a 1:2.5 PZA/PLGA weight ratio a 1:1 organic/aqueous phase volume ratio and 1% PVA as a surfactant. The experimental result of average particle size showed an acceptable agreement with the model (Table 7). There is as large difference between the predicted and experimental polydispersity index. However, the experimental polydispersity index is lower than the model, proving that nanoparticles are rather monodisperse. We believe the high value might arise from the transformation of the PDI needed for predictions. 3.2. Nanoparticle characterization The optimal formulation was then prepared in duplicate and characterized. The physicochemical properties of the optimal formulation of PZA-loaded PLGA NPs are shown in Table 8. Particle size distribution and zeta potential are presented in Fig. 2. SEM images show rather monodisperse nanoparticles below 200 nm in agreement with DLS measurements (Fig. 3). PLGA nanoparticles fabricated by double emulsion solvent evaporation method have been commonly applied for encapsulating hydrophilic molecules [29]. Typically, the size of the PLGA nanoparticles is in the range of 100e250 nm. Here, as the classical solvent (dichloromethane) was replaced by a partially water-miscible organic solvent (ethyl acetate), the nanoparticle size was lower than 200 nm [30]. Particle size is in a suitable range for macrophage uptake after pulmonary delivery [31]. The zeta potential of nanoparticles was found about 1 mV (Fig 2) in agreement with values reported in the literature for PVA coated nanoparticles [18,30]. Particles being
Goal
Limited range
In range Equal to 1:2.5 Equal to 1:1 Term not in model Minimize Minimize
DCM e EA 1:1e1:5 1:1e1:20 79e467 0.10e0.56
Table 7 Predicted and experimental results for particle size and polydispersity index of PZAloaded PLGA nanoparticles.
Prediction Experimental Error %
Average particle size (nm)
Polydispersity index
170 ± 35 173 ± 4 1.8
0.11 ± 0.01 0.05 ± 0.03 54.5
neutral should interact with lung cells or mucus [19,21]. The encapsulation efficiency of PZA in the optimized formulation was about 8%, leading to a drug loading of 3.1%, in the range usually found in the literature [32]. For encapsulating small hydrophilic molecules in PLGA nanoparticles, the double emulsion solvent evaporation method often suffers from limited encapsulation efficiency because of the drug rapid partitioning to the external aqueous phase [30]. Optimization of the drug loading could be obtained for example by playing with the difference of osmotic pressure between in the internal and external aqueous phases of the double emulsion.
4. Conclusion The Taguchi L9 method appears to be a fast, simple and valuable tool in optimizing the various parameters for the preparation of PZA-loaded PLGA nanoparticles. The organic solvent was found to be the main parameter having significant effects on the particle size and polydispersity index and ethyl acetate was chosen. The optimal formulation was prepared using a 1:2.5 PZA/PLGA weight ratio, a 1:1 organic/aqueous phase volume ratio and 1% PVA as a surfactant. A good agreement was obtained between the predicted optimized Taguchi method and the experimental data. The suitability of these PZA-loaded nanoparticles for pulmonary delivery to treat TB should be tested in the future.
Table 6 Optimum solutions for formulating PZA-loaded PLGA nanoparticles. The column highlighted in blue corresponds to the chosen solution.
Please cite this article in press as: D.-D. Pham, et al., Pyrazinamide-loaded poly(lactide-co-glycolide) nanoparticles: Optimization by experimental design, Journal of Drug Delivery Science and Technology (2015), http://dx.doi.org/10.1016/j.jddst.2015.07.006
6
D.-D. Pham et al. / Journal of Drug Delivery Science and Technology xxx (2015) 1e7
Table 8 Characteristics of PZA loaded PLGA nanoparticles. Lot#
Particle size (nm)
PDI
Zeta potential (mV)
Encapsulation efficacy (%)
Drug loading (%)
1 2
170 ± 2 175 ± 3
0.03 ± 0.01 0.07 ± 0.00
1.08 ± 0.14 1.10 ± 0.06
8.8 ± 4.7 7.1 ± 0.8
3.4 ± 1.7 2.8 ± 0.3
Fig. 2. Optimized Nanoparticles Formulation. The top graph represents Particle Size Distribution, the bottom graph represents the Zeta Potential distribution.
Fig. 3. Typical SEM images of optimized nanoparticle formulation.
Acknowledgments D. D. Pham fellowship is supported by Project 322, Ministry of Education and Training, Vietnam. Institut Galien Paris-Sud is a member of the Laboratory of Excellence LERMIT supported by a grant from ANR (ANR-10-LABX-33). References [1] WHO, Global Tuberculosis Control, 2011. [2] D.A. Mitchison, P.B. Fourie, The near future: improving the activity of rifamycins and pyrazinamide, Tuberc. Edinb. 90 (3) (2010) 177e181. [3] Y. Zhang, et al., Mode of action of pyrazinamide: disruption of Mycobacterium tuberculosis membrane transport and energetics by pyrazinoic acid, J. Antimicrob. Chemother. 52 (5) (2003) 790e795. [4] Pyrazinamide, Tuberculosis 88 (2) (2008) 141e144.
[5] R.M. McCune Jr., R. Tompsett, Fate of Mycobacterium tuberculosis in mouse tissues as determined by the microbial enumeration technique. I. The persistence of drug-susceptible tubercle bacilli in the tissues despite prolonged antimicrobial therapy, J. Exp. Med. 104 (5) (1956) 737e762. [6] H.I. Seifart, D.P. Parkin, P.R. Donald, Stability of isoniazid, rifampin and pyrazinamide in suspensions used for the treatment of tuberculosis in children, Pediatr. Infect. Dis. J. 10 (11) (1991) 827e831. [7] H.W. Frijlink, A.H. DE Boer, Dry powder inhalers for pulmonary drug delivery, Expert Opin. Drug Deliv. 1 (1) (2004) 67e86. [8] D.D. Pham, et al., Formulation of pyrazinamide-loaded large porous particles for the pulmonary route: avoiding crystal growth using excipients, Int. J. Pharm. 454 (2) (2013) 668e677. [9] D.D. Pham, E. Fattal, N. Tsapis, Pulmonary drug delivery systems for tuberculosis treatment, Int. J. Pharm. 478 (2) (2015) 517e529. [10] D.D. Pham, et al., Pulmonary delivery of pyrazinamide-loaded large porous particles, Eur. J. Pharm. Biopharm. 94 (2015) 241e250. [11] S. Gelperina, et al., The potential advantages of nanoparticle drug delivery systems in chemotherapy of tuberculosis, Am. J. Respir. Crit. Care Med. 172 (12) (2005) 1487e1490. [12] F. Andrade, et al., Nanotechnology and pulmonary delivery to overcome resistance in infectious diseases, Adv. Drug Deliv. Rev. 65 (13e14) (2013) 1816e1827. [13] N. Tsapis, et al., Trojan particles: large porous carriers of nanoparticles for drug delivery, Proc. Natl. Acad. Sci. U. S. A. 99 (19) (2002) 12001e12005. [14] C. Gomez-Gaete, et al., Dexamethasone acetate encapsulation into Trojan particles, J. Control. Release 128 (1) (2008) 41e49. [15] P. Muralidharan, et al., Inhalable nanoparticulate powders for respiratory delivery, Nanomedicine 11 (5) (2015) 1189e1199. [16] R. Pandey, S. Sharma, G.K. Khuller, Oral poly(lactide-co-glycolide) nanoparticle based antituberculosis drug delivery: toxicological and chemotherapeutic implications, Indian J. Exp. Biol. 44 (6) (2006) 459e467. [17] R. Pandey, et al., Poly (DL-lactide-co-glycolide) nanoparticle-based inhalable sustained drug delivery system for experimental tuberculosis, J. Antimicrob. Chemother. 52 (6) (2003) 981e986. [18] S. Mura, et al., Influence of surface charge on the potential toxicity of PLGA nanoparticles towards Calu-3 cells, Int. J. Nanomed. 6 (2011) 2591e2605. [19] S. Mura, et al., Biodegradable nanoparticles meet the bronchial airway barrier: how surface properties affect their interaction with mucus and epithelial cells,
Please cite this article in press as: D.-D. Pham, et al., Pyrazinamide-loaded poly(lactide-co-glycolide) nanoparticles: Optimization by experimental design, Journal of Drug Delivery Science and Technology (2015), http://dx.doi.org/10.1016/j.jddst.2015.07.006
D.-D. Pham et al. / Journal of Drug Delivery Science and Technology xxx (2015) 1e7 Biomacromolecules 12 (11) (2011) 4136e4143. [20] E. Fattal, et al., Lung toxicity of biodegradable nanoparticles, J. Biomed. Nanotechnol. 10 (10) (2014) 2852e2864. [21] N. Grabowski, et al., Toxicity of surface-modified PLGA nanoparticles toward lung alveolar epithelial cells, Int. J. Pharm. 454 (2) (2013) 686e694. [22] N. Grabowski, et al., Surface coating mediates the toxicity of polymeric nanoparticles towards human-like macrophages, Int. J. Pharm. 482 (1e2) (2015) 75e83. [23] Santiago, L.A., et al., Compared in vivo toxicity in mice of lung delivered biodegradable and non biodegradable nanoparticles. Nanotoxicology, 2015. [in press]. [24] S. Malathi, S. Balasubramanian, Synthesis of biodegradable polymeric nanoparticles and their controlled drug delivery for tuberculosis, J. Biomed. Nanotechnol. 7 (1) (2011) 150e151. [25] D.D. Ankola, et al., Preparation, characterization and in silico modeling of biodegradable nanoparticles containing cyclosporine A and coenzyme Q10, Nanotechnology 21 (6) (2010) 065104. [26] A. Tripathi, R. Gupta, S.A. Saraf, PLGA nanoparticles of anti tubercular drug:
[27]
[28]
[29] [30]
[31]
[32]
7
drug loading and release studies of a water in-soluble drug, Int. J. Pharm. Tech. Res. 2 (3) (2010) 2116e2123. K.D. Kim, et al., Optimization of parameters for the synthesis of zinc oxide nanoparticles by Taguchi robust design method, Colloid Surf. A Physicochem. Eng. Asp. 311 (1e3) (2007) 170e173. J.-C. Lu, S.-L. Jeng, K. Wang, A Review of Statistical Methods for Quality Improvement and Control in Nanotechnology, vol. 41, American Society for Quality, Milwaukee, WI, ETATS-UNIS, 2009, p. 17. F. Danhier, et al., PLGA-based nanoparticles: an overview of biomedical applications, J. Control. Release 161 (2) (2012) 505e522. E. Cohen-Sela, et al., A new double emulsion solvent diffusion technique for encapsulating hydrophilic molecules in PLGA nanoparticles, J. Control. Release 133 (2) (2009) 90e95. V. Kanchan, A.K. Panda, Interactions of antigen-loaded polylactide particles with macrophages and their correlation with the immune response, Biomaterials 28 (35) (2007) 5344e5357. E. Fattal, N. Tsapis, Nanomedicine technology: current achievements and new trends, Clin. Transl. Imag. Rev. Nucl. Med. Mol. Imag. 2 (2014) 77e87.
Please cite this article in press as: D.-D. Pham, et al., Pyrazinamide-loaded poly(lactide-co-glycolide) nanoparticles: Optimization by experimental design, Journal of Drug Delivery Science and Technology (2015), http://dx.doi.org/10.1016/j.jddst.2015.07.006