This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/authorsrights
Transcript
This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies areencouraged to visit:
a Department of Chemistry, Faculty of Science, Universiti Putra Malaysia, 43400, Serdang, Selangor, Malaysiab Laboratory of Molecular Biomedicine, Institute of Bioscience, Universiti Putra Malaysia, Serdang, Selangor, Malaysiac Department of Medicine, Faculty of Medicine and Health Sciences, Universiti Putra Malaysia, 43400, Serdang, Selangor, Malaysiad Centre for Drug Delivery Research, Faculty of Pharmacy, Universiti Kebangsaan Malaysia, Jalan Raja Muda Abdul Aziz, 50300 Kuala Lumpur, Malaysiae Quality Use of Medicines Research Group, Faculty of Pharmacy, Universiti Kebangsaan Malaysia, Jalan Raja Muda Abdul Aziz, 50300 Kuala Lumpur,Malaysia
a r t i c l e i n f o
Article history:Received 3 June 2013Received in revised form 15 July 2013Accepted 20 July 2013Available online 2 August 2013
Palm kernel oil esters nanoemulsion-loaded with chloramphenicol was optimized using response surfacemethodology (RSM), a multivariate statistical technique. Effect of independent variables (oil amount,lecithin amount and glycerol amount) toward response variables (particle size, polydispersity index, zetapotential and osmolality) were studied using central composite design (CCD). RSM analysis showed thatthe experimental data could be fitted into a second-order polynomial model. Chloramphenicol-loadednanoemulsion was formulated by using high pressure homogenizer. The optimized chloramphenicol-loaded nanoemulsion response values for particle size, PDI, zeta potential and osmolality were 95.33 nm,0.238, −36.91 mV, and 200 mOsm/kg, respectively. The actual values of the formulated nanoemulsionwere in good agreement with the predicted values obtained from RSM. The results showed that theoptimized compositions have the potential to be used as a parenteral emulsion to cross blood-brainbarrier (BBB) for meningitis treatment.
Bacterial meningitis is one of the most serious and commoninfections that can occur in the brain or meninges [1]. The symp-toms include fever, nausea, headache, vomiting, neck stiffness, andphotophobia. Meningitis can lead to acute inflammation and subse-quently brain damage [2]. It carries a high risk of death and severeneurologic sequelae, especially when there is delay in diagnosis andantibiotic administration [3]. Meningitis affects 30% of new-bornand young infants and 15–20% of older children [4]. Between 30%and 50% of meningitis survivors will suffer from permanent neuro-logical sequelae [5]. Despite the availability of effective antibiotictreatment, a high mortality rate of up to 30% has been observed[1,6].
Early antibacterial therapy is of utmost importance for treat-ment. Chloramphenicol is a relatively effective drug for bacterial
meningitis and has been used since 1975 [7]. However, the useof chloramphenicol has been administered in increasing dosagein recent years due to the increased incidence of antibiotic resis-tance. In addition, chloramphenicol is a hydrophobic drug whichpoorly dissolves in water. Clinically, this kind of hydrophobic drugdissolves in water in the form of a salt. A higher dosage of drugsolution will therefore be required to ensure that it can efficientlyreach the target cell. A nanoemulsion-based chloramphenicol car-rier could improve the solubility of the drug in the dispersed phaseand drug penetration into target cells due to its extremely smallsize. Therefore, a lower dosage would be needed due to more effi-cient penetration.
The blood-brain barrier (BBB) is a feature that encapsulates thehuman brain and protects brain from harmful compounds [8]. TheBBB is the gate keeper of the brain, and acts as an efflux pump, lim-iting the entry of many structurally divergent lipophilic molecules,such as peptides and many of the drugs used in psychiatry andneurology for the treatment of the brain [9]. BBB acts as a separatorbetween the brain and its blood supply. The criteria required for asuitable parenteral formulation to cross BBB include: particle size
114 S.H. Musa et al. / Colloids and Surfaces B: Biointerfaces 112 (2013) 113–119
less than 100 nm, polydispersity index (PDI) less than 0.25, zetapotential greater than 30 mV or lower than −30 mV and osmolalitybetween 285 and 310 mOsm/kg [10–12].
An emulsion is a mixture of two immiscible liquids with onedispersed in the other [13]. The basic components of an emulsionare immiscible water and oil; a surfactant is needed to decreasethe interfacial tension and maintain the stability of the emulsion.An emulsion can either be in the form of water-in-oil or oil-in-water [14]. A nanoemulsion is an emulsion with a droplet sizeranging from 20 to 200 nm [10]. It has been shown that an emulsionwith a very small particle size can provide greater encapsulationefficiency for delivery [15]. Nanoemulsions have special charac-teristics and can be used for drug delivery due to their extremelysmall size, biocompatibility, relative stability, ability to solubilizehigh quantities of hydrophobic compounds, ability to reduce thetoxicity of cytotoxic drugs, and ability to protect drugs from hydrol-ysis and enzymatic degradation under physiological conditions[16].
Nanoemulsions can be produced through low-energy emulsi-fication or high-energy emulsification methods [15]. Low-energyemulsification relies on the spontaneous formation of tiny oildroplets within mixed oil-water-emulsifier systems when the solu-tion or environmental conditions are altered [17]. An alternativemethod is the fabrication of a nanoemulsion, which is performedusing high-energy emulsification methods [18]. The high energyemulsification approach utilizes a mechanical device to generateintense disruptive forces that break up the oil and water phases toproduce tiny oil droplets. High energy emulsification is carried outby applying a microfluidic homogenizer, ultrasound homogenizer,or high pressure homogenizer [19].
Multivariate statistical techniques such as response surfacemethodology (RSM) have been used for the optimization of ana-lytical procedures for many years. [20]. RSM was first developed byBox and collaborators in the 1950s [21]. The RSM technique wasinspired by the graphical perspective generated regarding the fit-ness of a mathematical model [22]. Mathematical and statisticaltechniques using RSM are in a close relationship with laboratoryexperiments. Modeling and displacing experimental conditions areconducted by linear or polynomial functions to describe the systemunder study. RSM has the ability to determine the relationship andthe interaction between the independent variable and responsesbased on the desired criteria. Moreover, a fewer number of exper-imental trials will be needed to evaluate the interaction if RSM isapplied. Thus, optimizing an experimental process becomes lesstime consuming [23].
In this work, optimization of the composition of achloramphenicol-loaded nanoemulsion with respect to the amountof oil, lecithin and glycerol was carried out. The four responsesconsisted of particle size, PDI, zeta potential and osmolality werestudied to find the best composition for a chloramphenicol-loadednanoemulsion carrier.
2. Materials and methods
2.1. Materials
Palm kernel oil esters (PKOEs) were prepared in our researchlaboratory. Safflower seed oil was purchased from Sigma–AldrichChemie GmbH, Germany. Pure soy bean lecithin (Lipoid S75) waspurchased from Lipoid GmbH, Ludwigshafen-Germany. Glycerolwas purchased from JT Baker, USA. Polysorbate 80 (Tween80)was obtained from Fluka, Sigma–Aldrich Chemie GmbH, Germany.Chloramphenicol was purchased from Euroasias Chemicals PrivateLimited, India. Water was deionized using a Milli-Q filtration sys-tem.
Table 1Coded level for independent variables used in experimental design for nanoemul-sion optimization.
2.2. Determination of the solubility of chloramphenicol in oil
The solubility of chloramphenicol in the mixture of PKOEs withfive different types of oil (sesame oil, soybean oil, sunflower oil,safflower seed oil and pine nut oil) at a ratio of 1:1 (w/w) was deter-mined. The drug (1%) was added into the oil containing lecithin (3%).The solution was kept under moderate magnetic stirring for 24 h toreach equilibrium. The sample was then centrifuged at 4500 rpm for15 min. An aliquot of the supernatant was diluted with methanol.The chloramphenicol content was assayed by ultra-performanceliquid chromatography (UPLC).
2.3. Formulation of the nanoemulsion
Emulsions were prepared using an overhead stirrer (IKA® RW 20Digital, Nara, Japan) at 305–310 rpm. Chloramphenicol (0.5% w/w)was dissolved in the oil phase (PKOEs:safflower seed oil, 1:1 (w/w))containing lecithin (0.5–2.5%) as the surfactant. Tween 80 (0.75%w/w) was then added into the oil phase as a co-surfactant afterthe chloramphenicol was completely dissolved. The oil phase wasadded dropwise into the aqueous phase consisting of glycerol andwas continuously stirred to form a coarse emulsion. The mixturewas stirred for 1 h and was subjected to further processing througha high pressure homogenizer at 1000 psi for six cycles.
2.4. Selection of the co-surfactant
Co-surfactants including Tween 40, Tween 20, and CremophorEL were used instead of Tween 80 in the formulation. All formula-tions were analyzed with respect to the droplet size and PDI.
2.5. RSM experimental design
A three-factor central composite design (CCD) was employedto determine the effect of the amount of oil (4–14%, x1), lecithin(0.5–2.5%, x2), and glycerol (0.5–2.5%, x3) on four response vari-ables: average droplet size (Y1), PDI (Y2), zeta potential (Y3), andosmolality (Y4) of the nanoemulsion. A total of 20 runs weregenerated using Design-Expert® 6.0.6 software (Stat ease Inc.,Minneapolis, USA). Experiments with three independent variablesconsisted of right factorial points and five axial points, and six repli-cates of the center points were carried out [23]. The experimentswere carried out in randomized order to minimize the effects ofunexplained variability in the actual responses due to extraneousfactors [24]. A summary of the independent variables and theircoded levels are presented in Table 1. The results at each pointbased on experimental design (CCD) are presented in Table 2.
2.6. Statistical analysis
Response surface methodology was used to obtain the best for-mulation for the chloramphenicol-loaded nanoemulsion carrierwith respect to the amount of oil (x1), lecithin (x2), and glyc-erol (x3). The main objective was to determine the compositionof the nanoemulsion with the minimum particle size (Y1) andPDI (Y2), optimum zeta potential (Y3), and maximum osmolality
Author's personal copy
S.H. Musa et al. / Colloids and Surfaces B: Biointerfaces 112 (2013) 113–119 115
Table 2The matrix of central composite design (CCD); independent (xi) and response variables (Yi).
Run Type Independent variables, xi(%, w/w) Response variable, Yi
(Y4). The composition with three variables was constructed usingcentral composite design (CCD). A second-order polynomial couldeffectively predict the behavior of the responses to the selectedindependent variables. The generalized response surface model isshown by the following equation:⎡⎣Y = ˇ0 +
3∑i=1
ˇixi +3∑
i=1
ˇiix2i +
2∑i=1
3∑j=i+1
ˇijxixj + s
⎤⎦where ˇ0 is a constant term; ˇi represents the coefficient of thelinear parameters, xi represents the variables and ε is the resid-ual associated to the experiments [20]. In order to analyze and fitall the experimental data to the second order polynomial equa-tion, a multiple regression technique was applied. The significantdifferences between independent variables were investigated byanalysis of variance (ANOVA). Only non-significant (P < 0.05) valueswere involved in constructing a reduced model, while significant(P > 0.05) values were eliminated. For a better understanding ofthe interaction of response variables and independent variables,response surfaces and three dimensional (3D) contour plots of thefitted polynomial regression equation were generated. Numericaland graphical optimizations were conducted to obtain the opti-mum conditions and predicted values for the desirable responsegoals using a response optimizer. This helps with the identificationof the combination of input variable settings followed by the opti-mized solution while it allows the user to compromise between thevarious responses obtained [25].
2.7. Verification of models
Verification of the model was carried out by comparingthe experimentally determined values with the predicted val-ues obtained from the response regression equations [26]. Dataobtained for particle size, PDI, zeta potential, and osmolality forchloramphenicol-loaded nanoemulsions were prepared followingthe recommended composition (Table 2).
2.8. Emulsion particle size, PDI and zeta potential measurement
The particle size and PDI of the nanoemulsion were measuredusing dynamic light scattering, scattered at an angle of 173◦ and atemperature of 25 ◦C. This process was carried out using a MalvernNano ZS90 apparatus (Malvern, UK). The nanoemulsion sample was
diluted with deionized water until it reached the desired concentra-tion. All diluted samples were kept at a count rate of 150–300 kcps.Measurements were repeated three times.
2.9. Emulsion osmolality measurement
The measurement of nanoemulsion osmolality was basedon the freezing-point method as described in the osmometeruser’s manual. After calibration of the osmometer (Model 3320,Advanced Instruments, Inc., USA) with reference standards (100and 3000 mOsm/kg, Advanced Instruments), the osmolality wasrecorded using 0.25 mL of the sample.
3. Results and discussion
3.1. Solubility of chloramphenicol in oil
Fig. 1 shows the solubility of chloramphenicol in the differentoil mixtures. After centrifugation, samples 2 (PKOEs:safflower seedoil) and 4 (PKOEs:soybean oil) did not show any precipitation ordrug leakage at the bottom of the tube. In contrast, a precipitate wasdetected at the bottom of the tube for samples 1 (PKOEs:sesame oil),3 (PKOEs:sunflower oil), and 5 (PKOEs:pine nut oil). However, sincesample 2 could solubilize more chloramphenicol than sample 4, thechoice of a suitable oil phase was a mixture of PKOEs and safflowerseed oil.
Palm kernel oil esters (PKOEs) consist of relatively shorter chainesters that are potentially good carriers for the delivery of activeingredients into the body [27]. From this observation, it showsPKOEs make a good combination with oil containing a high per-centage of linoleic acid. Linoleic acid is an unsaturated fatty acidwith two double bonds. Both safflower seed oil and soybean oilcontain more than 50% linoleic acid as compared to the other threeoils which have a lower linoleic acid component. Han et al. reportedthat safflower seed oil contains 70–87% linoleic acid, which couldhelp in BBB penetration in addition to being rich in vitamin E [28].The addition of linoleic acid helped to increase the solubility ofchloramphenicol.
3.2. Selection of a co-surfactant
The particle sizes of lecithin with the co-surfactants Tween 80,Tween 40, Tween 20 and Cremophor EL were 131.5, 141.8, 155.5and 140.9 nm, respectively. The PDI values for Tween 80, Tween 40,
Author's personal copy
116 S.H. Musa et al. / Colloids and Surfaces B: Biointerfaces 112 (2013) 113–119
Tween 20 and Cremophor EL were 0.135, 0.137, 0.231 and 0.145,respectively. Tween 80 as a co-surfactant provided the smallest par-ticle size as well as PDI. Lecithin, a natural emulsifier which mimicsthe biological system in the brain either by electrostatic or cova-lent binding is known to help particles cross the BBB [29]. However,lecithin by itself was not enough to for a good nanoemulsion sys-tem. It has been reported that the usage of lecithin as an emulsifiermay lead to the formation of lyso-derivatives [30]. The formation oflyso-derivatives should be controlled to prevent hemolytic poten-tial in the body. However, use of a co-surfactant such as Tween80 can minimize this undesired problem. Song et al. stated thatthe usage of Tween 80 in pharmaceutical applications results infavorable encapsulation efficiency [31]. Therefore, Tween 80 wasselected as the co-surfactant in this study.
3.3. Selection of independent variables
The results shows that the best emulsification was achievedby stirring the coarse emulsion for 1 h with an agitation speedof 305 ± 5 rpm and six cycles of passing through the high pres-sure homogenizer at a pressure of 1000 psi. The evaluation wascarried out based on the lower, middle, and upper levels of threeindependent variables. The influence of preparation method (agi-tation speed, mixing time, and the number of cycles) on theparticle size, PDI, and zeta potential was studied. The step wasperformed to avoid the “over-processing” phenomena [32]. Thechloramphenicol-loaded nanoemulsions showed a particle size
less than 200 nm. The accepted range of PDI (less than 0.25),zeta potential between −25 mV and −40 mV and osmolality upto 250 mOsm/kg was maintained by confining the amounts of oil,lecithin, and glycerol at the levels of 4–14%, 0.5–2.5%, and 0.5–2.5%,respectively (data not shown).
3.4. Fitting the response surface models
Table 2 shows the experimental data for each response variableunder different independent variables according to the CCD matrix.Response surface models allow for the prediction of variations inthe particle size, PDI, zeta potential, and osmolality as a functionof the composition of the chloramphenicol-loaded nanoemulsion.The estimated regression coefficients, R2, adjusted R2, regression(P-value), regression (F-value), and standard deviation related tothe effect of the three independent variables are presented inTable 3. Response surface analysis exhibited significant (P < 0.05)regression relationships between the independent variables andthe response variables. The final reduced models included non-significant (P > 0.05) linear terms if it was found that the modelcontaining these variables was still significant (P < 0.05) [33].
The response surface analysis indicated that the second-orderpolynomial response model for osmolality had the highest coef-ficient value (R2 = 0.9990), followed by particle size (R2 = 0.9901),zeta potential (R2 = 0.9415), and PDI (R2 = 0.8736). The results showthat more than 90% of the response variations of the independentvariables (osmolality, particle size, and zeta potential) could be
Table 3Analysis of variance (ANOVA) for the model.
Response variables (Yi) F-value P-value R2 Adjusted R2 Standard deviation
S.H. Musa et al. / Colloids and Surfaces B: Biointerfaces 112 (2013) 113–119 117
Fig. 2. Response surface plot showing the significant (P < 0.05) interaction effect of responses variables (Yi) as a function of independent variables (xi); (a) particle size, (b)zeta potential, and (c) osmolality.
described by RSM model as a function of the main composition.On the other hand, the model obtained for the response variablePDI did not reflect the true behavior of the system, where the modelcould only predict up to 87.36% of the observed response variations.
Analysis of variance (ANOVA) was used to evaluate the coef-ficient significance of the quadratic polynomial models. LargeF-values and small P-values showed significant effects on therespective response variables [34]. From the data obtained, thelinear term of oil amount in the emulsion had the most signif-icant (P < 0.0001) effect on the particle size with a high F-value(518.35) followed by the linear term of the amount of lecithin.However, the glycerol amount was found to have a non-significant(P > 0.05) effect on the particle size as it showed the lowest F-value (2.39). ANOVA revealed that only the amount of oil exerteda significant effect (P < 0.0001, F-value = 46.73) on the PDI of thenanoemulsion, while no linear term (P > 0.05) was observed forthe lecithin and glycerol amounts on this response. Zeta potentialwas mostly influenced by the amount of lecithin with a signifi-cant effect (P = 0.0025, F-value = 17.18), while it was insignificant(P > 0.05) toward the other two variables (the amount of oil andglycerol). All independent variables showed a significant (P < 0.05)effect on the osmolality response. However, based on the F-valueobtained (1269.78), the amount of glycerol had the greatest effecton the osmolality value.
3.5. Response surface analysis
Three-dimensional (3D) graphs of the ANOVA analysis wereplotted to show how independent variables affected the responsevariables. The graphs of the response surface for particle size, zeta
potential, and osmolality were used to interpret the significantinteraction (P < 0.05) of the models.
3.5.1. Particle sizeFig. 2a shows that a smaller particle size could be obtained
at both the highest amount of lecithin (2.5%) and the optimumamount of oil (6.5–9%). However, as the amount of oil was furtherincreased with the same composition of lecithin (2.5%), the particlesize started to increase. This might be due to an insufficient amountof lecithin available to emulsify the oil and the aqueous phase whenthe amount of oil was increased. The plotted model illustrated a lin-ear increase in particle size when the amount of oil was increased.It was shown that increasing the viscosity of the dispersed phase(oil phase) can increase the flow resistance, which may result inproblems with the droplet disruption process. Thus, the break-uprate becomes restricted, causing the formation of larger particles.The model described the effect of the amount of lecithin on theparticle size. As the amount of lecithin decreased, an increase inthe particle size values was detected. This could be due to the fixedamount of the emulsifier (lecithin) leading to incomplete coverageof emulsifier molecules on the newly formed droplets. It is possiblethat this coverage limitation can lead to an increase in the particlesize of an emulsion [16]. The pharmaceutical industry demands theproduction of nanoemulsions with a much smaller particle size asit provides extremely low surface tension for the whole system andthe interfacial tension of oil-in-water droplets [26].
3.5.2. Zeta potentialAs mentioned above (see Fig. 2a), the smallest particle size was
obtained at the highest amount of lecithin and lowest amount of
Author's personal copy
118 S.H. Musa et al. / Colloids and Surfaces B: Biointerfaces 112 (2013) 113–119
oil. Hence, the study was continued, focusing on the area in whichthe highest amount of lecithin (2.5%) and lowest amount of oil (4%)was detected (see Fig. 2b). The results showed a good reading forthe zeta potential with a value of −37.26 mV. This result revealedthat the interaction between the amount of lecithin and oil couldaffect the zeta potential. Choosing a sufficient amount of lecithin incomparison with the amount of oil can cause electrostatic repulsionbetween droplets to provide enough surface charges for stabiliza-tion of the formulation. However, further analysis showed that atthe same amount of oil (4%), the zeta potential became less neg-ative when the amount of lecithin was decreased. Zeta potentialrepresents the surface charge of the emulsion droplet, which mayhave effects on physical and chemical stability [35]. Zeta poten-tial is the potential surrounding the droplet at the plane of shear.It is expected that the corresponding nanoparticles will be coatedby surfactant molecules [36]. The repulsive electrostatic interac-tions between similarly charged emulsion droplets do not allowthem to get closer, which hinders coalescence [37]. Lecithin is ahighly negatively charged phospholipid which can produce emul-sions with a high zeta potential [38]. This could explain whydecreasing the amount of lecithin led to less negative zeta potentialvalues.
3.5.3. OsmolalityFig. 2c shows that the amount of oil is directly proportional
to the increase in the osmolality value. The oil composition wasa combination of safflower seed oil and palm kernel oil esters(1:1). Palm kernel oil, which was extracted from the seeds of palmfruits and is characterized by high contents of lauric and myris-tic acids, represents less than 2% of palm oil [39]. As reported byHan et al., safflower seed oil contains a high percentage of linoleicacid (70–87%) [28]. Linoleic acid is a long chain triglyceride witha C18 structure. These long chain triglycerides contain essentialfatty acids that affect the osmolality value. Thus, increasing theoil amount will increase the amount of fatty acids in the formula-tion. As a result, the osmolality of the nanoemulsion will increase.Lecithin could also affect the osmolality value of the nanoemul-sion. Lecithin is composed of not only phospholipids, but othercompounds including triglycerides, fatty acids, tocopherols, andglycerol [40]. All these compounds can individually contribute tothe osmolality value. Osmolality is the measure of osmoles of soluteper kilogram of solvent. Osmolality is very important in intravenousdelivery, especially in the measurement of pain intensity. The inci-dence of pain on injection and thrombotic complications appear tobe lower with X-ray contrast media of low osmolality [41]. Glyc-erol is an additive that has been used to modify the osmolality ofsolutions for parenteral delivery. Additionally, glycerol is also suit-able for use as a surfactant in the formulation of nanoemulsions.Therefore, increasing the amount of lecithin can alter the osmolalityvalue.
3.6. Optimization of responses
The chloramphenicol-loaded nanoemulsion was optimizedusing Design Expert software. The formulation criteria were aimedat finding the smallest particle size, the smallest PDI, the optimumzeta potential, and the highest osmolality. 3D response surfaces andcontour plots were used to understand the interactions betweenindependent variables and response variables. After analyzing theinteraction from different angles, the optimum chloramphenicolnanoformulation was achieved with a composition of 4% oil, 2.5%lecithin, and 2.25% glycerol. With these optimum conditions, thepredicted values for particle size, PDI, zeta potential, and osmolal-ity were obtained as 95.33 nm, 0.238,−36.91 mV and 200 mOsm/kg,respectively.
Table 4The predicted and actual values of chloramphenicol-loaded nanoemulsion.
Response variables (Yi) Predicted Actual Standarddeviation
The predicted and actual values of the responses were com-pared to check the adequacy of the surface response equation (seeTable 4). The optimized chloramphenicol-loaded nanoemulsionformulated had a particle size, PDI, zeta potential, and osmolality of86.96 nm, 0.234, −33.44 mV, and 198 mOsm/kg, respectively. Theseresults were found to be in a good agreement with the predictedvalues.
4. Conclusion
This study demonstrated that RSM is a useful tool to optimizethe parameters of a chloramphenicol-loaded nanoemulsion andhelps in understanding the relationship between independent vari-ables and response variables. The results showed that the emulsioncomposition could affect the physicochemical characteristics of thenanoemulsion. Second order polynomial regression models weredeveloped using CCD to predict the variation of particle size, PDI,zeta potential, and osmolality. The amounts of oil, lecithin, and glyc-erol showed a linear effect on the responses studied. The amountof oil showed a major significant effect (P < 0.0001) on the par-ticle size and PDI responses, while the amount of lecithin andthe amount of glycerol had major significant effects on the zetapotential (P = 0.0025) and osmolality (P < 0.0001). The optimizedcomposition with 4% oil, 2.5% lecithin, and 2.25% glycerol togetherwith 0.75% Tween 80 provided the most desirable criteria for thechloramphenicol-loaded nanoemulsion.
Acknowledgement
The financial support from Graduate Research Fellowship (GRF)Universiti Putra Malaysia is gratefully acknowledged.
References
[1] M. Weisfelt, D. van de Beek, L. Spanjaard, J.B. Reitsma, J. de Gans, Lancet Neurol.5 (2006) 123–129.
[2] C.L. Bellac, R.S. Coimbra, S. Christen, S.L. Leib, Neurobiol. Dis. 24 (2006) 395–402.[3] C. Franco-Paredes, L. Lammoglia, I. Hernández, J.I. Santos-Preciado, Int. J. Infect.
Dis. 12 (2008) 380–386.[4] J. Klein, R. Feigin, G. McCracken, Pediatr. Infect. Dis. J. 326 (1993) 42–47.[5] M.T. Fitch, D. van de Beek, Lancet Infect. Dis. 7 (2007) 191–200.[6] J. Mishal, A. Embon, A. Darawshe, M. Kidon, E. Magen, Eur. J. Intern. Med. 19
(2008) 421–426.[7] R.F. Lewis, F. Dorlencourt, J. Pinel, The Lancet 352 (1998) 823.[8] K.E. Sandoval, K.A. Witt, Neurobiol. Dis. 32 (2008) 200–219.[9] O.L. De Klerk, A.T.M. Willemsen, F.J. Bosker, A.L. Bartels, N.H. Hendrikse, J.A. den
J. Control. Release 93 (2003) 271–282.[12] S. Ganta, J.W. Paxton, B.C. Baguley, S. Garg, Int. J. Pharm. 360 (2008) 115–121.[13] F. Leal-Calderon, F. Thivilliers, V. Schmitt, Curr. Opin. Colloid Interface Sci. 12
(2007) 206–212.[14] F. Shakeel, W. Ramadan, Colloids Surf. B 75 (2010) 356–362.[15] C. Solans, P. Izquierdo, J. Nolla, N. Azemar, M.J. Garcia-Celma, Curr. Opin. Colloid
Interface Sci. 10 (2005) 102–110.[16] S.M. Jafari, Y. He, B. Bhandari, Food Res. Int. 40 (2007) 862–873.[17] N. Anton, J. Benoit, P. Saulnier, J. Control. Release 128 (2008) 185–199.[18] C. Lovelyn, A.A. Attama, J. Biomater. Nanobiotechnol. 2 (2011) 626–639.[19] R.T. Prakash, P. Thiagarajan, Res. Biotechnol. 2 (2011) 1–13.
Author's personal copy
S.H. Musa et al. / Colloids and Surfaces B: Biointerfaces 112 (2013) 113–119 119
Salleh, Ind. Crop Prod. 29 (2009) 37–44.[28] X. Han, L. Cheng, R. Zhang, J. Bi, J. Food Eng. 92 (2009) 370–376.[29] S. Krol, J. Control. Release 164 (2012) 145–155.
[30] A.G. Floyd, Pharm. Dev. Div. 2 (1999) 134–143.[31] J. Song, F. Shi, Z. Zhang, F. Zhu, J. Xue, X. Tan, L. Zhang, X. Jia, Molecules 16 (2011)
7880–7892.[32] A. Desrumaux, J. Marcand, Int. J. Food Sci. Technol. 37 (2002) 263–269.[33] H. Mirhosseini, C.P. Tan, A.R. Taherian, H.C. Boo, Carbohyd. Polym. 75 (2009)
512–520.[34] E. Bayraktar, Process Biochem. 37 (2001) 169–175.[35] L. Rabinovich-Guilatt, P. Couvreur, G. Lambert, D. Goldstein, S. Benita, C. Duber-
net, Chem. Phys. Lipids 131 (2004) 1–13.[36] M. García-Fuentes, D.D. Torres, M.J. Alonso, Colloids Surf. B 27 (2002)
159–168.[37] H. Mirhosseini, C.P. Tan, N.S.A. Hamid, S. Yusof, Colloids Surf., A 315 (2008)
47–56.[38] S. Benita, M.Y. Levy, J. Pharm. Sci. 82 (1993) 1069–1079.[39] T. Awad, K. Sato, Colloids Surf. B 25 (2002) 45–53.[40] C.R. Scholfield, J. Am. Oil Chem. Soc. 58 (1981) 889–892.[41] W. Klement, J.O. Arndt, Br. J. Aesthet. 66 (1991) 189–195.
![[Salmonella meningitis in children in Libreville. Retrospective study of 9 cases]](https://static.documents.pub/doc/80x56/635da1b01b1c1ace260921f7/salmonella-meningitis-in-children-in-libreville-retrospective-study-of-9-cases.jpg)
