Bovine Serum Albumin-Loaded Chitosan/Dextran Nanoparticles: Preparation and Evaluation of Ex Vivo Colloidal Stability in Serum

Chitosan (CS) nanoparticles have several distinct intrinsic advantages; however, their in vivo colloidal stability in biological fluids was not fully explored especially when carrying proteins. The present study aimed to investigate their colloidal stability using an ex vivo physiological model of fetal bovine serum (FBS) and human serum (HS). The stability of bovine-serum-albumin (BSA-) loaded nanoparticles was relatively higher in FBS than that in HS. Particle size of unloaded and BSA-loaded nanoparticles was statistically unchanged up to 24 h after incubation in FBS. However in HS, a significant increase in particle size from 144 ± 17 to 711 ± 22 nm was observed for unloaded nanoparticles and by 2.5-fold for BSA-loaded nanoparticle, at 24 h after incubation in HS. Zeta potential of both nanoparticles was less affected by the components in FBS compared to those in HS. A remarkable swelling extent was experienced for unloaded and BSA-loaded nanoparticles in HS, up to 54 ± 4% and 44 ± 5%, respectively. Morphology of unloaded and BSA-loaded nanoparticles was varied from smooth spherical and rod shape to irregular shape when incubated in FBS; however, form agglomerates when incubated inHS.These findings therefore suggest that HS ismore reactive to cause colloidal instability to the chitosan nanoparticles compared to FBS.


Introduction
Despite several intrinsic and distinct advantages of nanoparticles, the concerns about the health risks of polymeric nanoparticles have been escalating nowadays due to the higher incidence of instability of nanoformulations.Numerous in vitro studies reported that the nanosized particles are biologically more potent than the equivalent micronsized particles of same chemical composition [1,2].The potency and biological activity of polymeric nanoparticles have been connected with several physicochemical (colloidal) characteristics such as the shape of the particles, their surface area, hydrodynamic particle size, agglomeration or flocculation rate, surface potential (zeta potential), and the surface chemistry of nanoparticles [3].These physicochemical characteristics of polymeric nanoparticles are in turn highly affected by the medium in contact such as biological fluids (plasma, serum, saliva, sweat, sebum, tears, etc.).Thus, it was argued that the physicochemical properties and unique kinetics of nanomaterials in biological solutions should be considered prior to various pharmaceutical and pharmacodynamic testings [3][4][5].Besides, the adsorption and desorption affinities of reactive components of biological fluids (such as proteins) onto the nanoparticles surface, type, amount, and conformation of the adsorbed proteins are necessary to be considered when analyzing the biological responses of the nanoparticles [6,7].
Among the biological fluids, fetal bovine serum (FBS), a commonly used supplement for cell culture, and the human serum (HS), the main component of human blood [8], were used in the current research to assess the colloidal stability of nanoparticles.Serum is a complex mixture of different factors and contains a large number of components like growth factors, proteins, vitamins, hormones, trace elements, and other essential and nonessential components [9].Serum Journal of Nanomaterials components would tend to alter the colloidal characteristics (particle size, PDI, and zeta potential) of polymeric nanoparticles and carry their own surface charges (e.g.proteins with negative charges while some growth hormones exhibit positive charges) which may influence the overall physicochemical characteristics of polymeric nanoparticles [10,11].Therefore, determination of protein stability in serum constitutes a powerful and important screening assay.In vivo testing of peptide stability is obviously of more relevance than in vitro but blood samples should be as steriled as possible to ensure maximum activity of proteolytic enzymes and minimal interference with the assay [12].Besides that the blood is a biohazardous material, it must also be heparinised and this may interfere with the assay.
Recently, for preparing the polymeric nanoparticles, chitosan (CS) has been extensively focused on.CS is a natural biopolymer derived from chitin deacetylation [13,14].It has attained a remarkable attention due to its biological properties such as excellent biocompatibility, hydrophilicity, biodegradability, and antibacterial activity [15].Gan and Wang [16] had recommended the CS as the superior particulate polymer for in vivo administration due to its nontoxic nature and degradation by the action of lysozyme in the body.Besides, CS is a good candidate for in vivo and ex vivo applications because it would not accumulate in body tissues [17].
Taken together, the selection of cross-linking agent used in the particle preparation is also an imperative factor to further improve the colloidal stability of CS nanoparticles.Thus, in the present study, CS nanoparticles were prepared by ioniccross linking of CS with dextran sulphate (DS).The criterion for using DS as the cross-linking agent is that it had been reported to produce mechanically more stable nanoparticles compared to the penta-sodium tripolyphosphate (TPP) [18,19].Moreover, CS/DS nanoparticles offer many therapeutic advantages over CS/TPP nanoparticles.Particle size of CS/DS nanoparticles was comparatively small which directly affect the colloidal stability of particulate dispersion [20].Besides the particle size, zeta potential is also the vital indicator to predict and control the stability of colloidal dispersion [21].Hence, the present study was aimed to investigate the colloidal stability of CS/DS nanoparticles in HS and FBS.The colloidal stability was assessed in terms of particle size, zeta potential, PDI, swelling characteristics, and as well as the morphology of CS/DS nanoparticles.

Preparation of Unloaded CS/DS Nanoparticles. CS/DS
nanoparticles were prepared via ionic-gelation method, previously developed by Calvo et al. [26] with some modification.CS was dissolved in 1% v/v glacial acetic acid to produce 0.075% w/v CS solution.Three different concentrations of DS (0.075, 0.1, and 0.125% w/v) were prepared by dissolving DS in distilled water.CS/DS nanoparticles were prepared simultaneously by adding 40 mL of DS solution dropwise in 100 mL of 0.075% w/v CS solution under a constant magnetic stirring at 700 rpm for 30 min.Thereafter, CS/DS nanoparticles were harvested by ultracentrifugation (25000 rpm) using an Optima L-100 XP Ultracentrifuge (Beckman-Coulter, USA) with a rotor NV 70.1 Ti (Beckman-Coulter, USA) at 10 ∘ C for 15 min.

Preparation of BSA-Loaded CS/DS Nanoparticles.
For preparing BSA-loaded CS/DS nanoparticles, BSA was dissolved in PBS solution (pH, 7.4) to produce the concentration of 1 mg/mL.The pH of CS solution was then adjusted to 5.5 by adding either 0.5 M NaOH or 0.5 M HCl.BSA solution was then premixed with 0.075% w/v CS solution and incubated for 30 min at room temperature.DS solution (0.075, 0.1 and 0.125% w/v) was then added dropwise in the reaction mixture under a continuous magnetic stirring (700 rpm) for 30 min to produce BSA-loaded CS/DS nanoparticles.The resultant nanoparticles were harvested by ultracentrifugation at 25000 rpm at 10 ∘ C for 15 min.

Determination of Particle Size, PDI, and Zeta Potential.
Mean particle size, PDI, and zeta potential of unloaded and BSA-loaded CS/DS nanoparticles were measured by using a ZS-90 Zetasizer (Malvern Instruments, UK).For particle size analysis, measurements were performed at 25 ∘ C with a detection angle of 90 ∘ .All measurements were performed in triplicate and results were reported as mean ± standard deviation (S.D).

Entrapment Efficiency (EE).
To determine EE, the BSA-loaded CS/DS nanoparticles were harvested by ultracentrifugation at 25000 rpm using Optima L-100 XP Ultracentrifuge (Beckman-Coulter, USA) with a rotor NV 70.1 Ti (Beckman-Coulter, USA) at 10 ∘ C for 15 min.Supernatants recovered by ultra-centrifugation were then decanted, and the BSA contents were analyzed using Bradford protein assay as per manufacturer's instructions.The samples were then subjected to U.V/Vis spectrophotometer (U.V-1601, Shimadzu, Japan) and analyzed at 595 nm ( max ), and U.V absorbance was recorded.The percent EE of BSA was then calculated indirectly from remaining supernatant using the following equation [22]:

Stability Studies of CS/DS Nanoparticles in Serum.
To assess the colloidal stability of unloaded and BSA-loaded CS/DS nanoparticles, FBS was first incubated at 37 ∘ C in a water bath incubator to simulate physiological media.FBS was then added in the CS/DS nanoparticles dispersion with the volume ratio of 1 : 1 in which the final concentration of 50% v/v of serum was yielded.Subsequently, the resulting mixture was incubated in the water bath incubator at 37 ∘ C for 24 h.The mean particle size, PDI, and surface charge were measured at predetermined time points (0, 10, 30 min, 1, 2, 3, 12, and 24 h).Same experimental protocol was also repeated to determine the colloidal stability of CS/DS nanoparticles in HS.CS/DS nanoparticles were suspended in PBS (pH, 7.4) prior to mixing with FBS or HS.

Swelling Analysis.
To investigate the swelling characteristics of CS/DS nanoparticles, 100 mg of lyophilized powder sample of either unloaded or BSA-loaded CS/DS nanoparticles was immersed in 100 mL of FBS and/or HS at various DS concentrations (0.075, 0.1, and 0.125) for 24 h at room temperature until a swollen equilibrium was achieved.The swollen samples were then collected by filtration, blotted with filter paper for the removal of the surface adsorbed water, and weighed immediately.Then, the swelling ratios of CS/DS nanoparticles were calculated using the following equation [23]: where   and   are the average weights of swollen and dry samples, respectively.Results were reported as mean ± S.D.

Morphological Examination.
With a view to evaluate the effect of FBS and HS on the morphological characteristics of CS/DS nanoparticles, the unloaded and BSAloaded nanoparticles were incubated in both sera for 24 h and were viewed under transmission electron microscope (TEM).
Prior to TEM analysis, the incubated dispersions of CS/DS nanoparticles (in FBS or HS at 24 h after incubation) were diluted to 1 : 10 ratio with phosphate buffered saline (PBS, 0.01 M).The resulting dilution was carried out in order to obtain the clear images (microscopic micrographs) of CS/DS nanoparticles under TEM.To perform the microscopic (TEM) analysis, a drop of diluted nanoparticles dispersion in FBS or HS was placed onto the copper microgrid that was natively stained with phosphotungstic acid and allowed to evaporate and dry at room temperature (25 ± 2 ∘ C).The dried microgrids were then viewed at different TEM resolutions to assess the morphology of unloaded and BSAloaded nanoparticles before and after incubation in FBS and HS.
2.9.Statistical Analysis.The data was presented as mean ± standard deviation (S.D).Data was analyzed using SPSS 17.0 (paired t-test and independent t-test and ANOVA, followed by Tukey's post hoc analysis).For paired t-test,  < 0.05 showed the significant difference between the mean of tested groups.For independent t-test,  < 0.05 showed significant difference between the mean of two independent tested samples.[24].In present study, DS concentrations had less impact on the particle size, PDI, and zeta potential of CS/DS nanoparticles.This was expected because CS/DS weight ratio of different formulations was small and therefore yielded nanoparticles with almost similar colloidal characteristics.The CS/DS weight ratios studied in the present study were 1 : 1, 1 : 1.3, and 1 : 1.7 for 0.075, 0.1, and 0.125% w/v DS, respectively, while, the weight ratios of 5 : 3, 5 : 5, 5 : 10, and 5 : 20 were applied in the previous report.

Results and Discussion
Despite that, all the formulations were within the nanosized range (200∼300 nm).On the other hand, the lower surface charges of BSA-loaded CS/DS nanoparticles as compared to the unloaded nanoparticles were expected to be due to the neutralization of positive charges (−NH 3 + ) on the contour of CS by the negatively charged BSA molecules [22,24,25], and this resulted in a decrease of the overall surface charge of nanoparticles as shown in Figure 1(b).

Entrapment Efficiency (EE).
Data revealed that the EE of BSA was observed to be affected by the DS concentrations.A higher EE was obtained when the CS/DS weight ratio was reduced.The lowest CS/DS weight ratio (1 : 1) produced nanoparticles with the highest EE (95 ± 2%) of BSA.EE for the other DS concentrations wase found to be 91 ± 2% (at 1 : 1.3) and 86 ± 2% (at 1 : 1.7).This was expected as increase in the concentration of DS that would tend to increase the negative charge density on the surface of nanoparticles which may subsequently increase the repulsion energy against the negatively charged BSA molecules which might not be favourable for the protein entrapment process.
Furthermore, EE of BSA was also affected by pH of CS solution.The optimal pH was found to be at 5.5.The highest EE achieved in this study without adjusting the pH of CS solution to 5.5 was only 46 ± 4% (CS/DS weight ratio 1 : 1).This finding was in agreement with Calvo et al. [26] which reported that the greatest loading efficiency could be obtained when protein was dissolved at a pH above its isoelectric point (pH 4.8).At this pH, BSA would predominantly exhibit its highest negative charge and could ionically interact with positively charged −NH 3 + groups on the backbone of CS.BSA also tends to be more negative as the pH of media increases above its isoelectric point [27] which favours the entrapment of BSA into the CS nanoparticles.

Stability in FBS.
Figures 2(a) and 2(b) clearly highlight that the particle size of both unloaded and BSA-loaded CS/DS nanoparticles (prepared from 0.075% w/v of DS) was significantly increased ( < 0.05, independent t-test) at 1 h after incubation and declined from this point, up to 24 h after incubation.Similarly, the zeta potential showed a significant decline ( < 0.05, independent t-test) from +56±5 and +34± 4 mV to +33 ± 3 and +14 ± 2 mV for unloaded (Figure 2(c)) and BSA-loaded nanoparticles (Figure 2(d)), respectively, at 10 min after incubation in FBS.A further decrease in zeta potential of both nanoparticles was also observed over time and this was thought to be caused by their interactions with negatively charged albumin macromolecules (>−20 mV) in the incubation media.The obtained data also suggested that the unloaded nanoparticles were more strongly affected by the components in FBS than the BSA-loaded one.This might be due to stronger ionic interactions between serum albumin and the unloaded nanoparticles which had relatively higher positive surface charges as compared to BSAloaded nanoparticles.The positive surface charge density of BSA-loaded CS/DS nanoparticles was relatively low due to the neutralization action by the BSA during the loading stage [28].Moreover, charge neutralization by the albumin macromolecules in the incubating media further reduced repulsion energy between nanoparticles.This may also progressively facilitate agglomeration process and promote the formation of larger particles [9].The tendency of particles to form aggregates was less prominent in case of BSAloaded nanoparticles because they possessed relatively low zeta potential (positive charges) which subsequently resulted in a lesser degree of interaction with negatively charged components in FBS.Thus, this suggested that BSA-loaded nanoparticles had better colloidal stability in FBS compared to the unloaded ones.This finding is important because colloidal stability in serum for any nanoparticulate system determines the successful delivery of protein/peptides as it prevents particle aggregation or embolism from occurring in the systemic circulation [29].The incubation time is also a critical factor to determine the adsorption pattern of proteins on the solid surfaces of nanoparticles [30].In accordance to that, the particle size and zeta potential of unloaded and BSA-loaded nanoparticles were significantly influenced by the incubation time.Figure 2 clearly shows the changing patterns of the particle size and zeta potential over extended incubation time.A significant difference ( < 0.05, paired t-test) was observed among particle sizes at the different postincubation time points.For example, BSA-loaded CS/DS nanoparticles prepared from 0.075% w/v of DS were found to be enlarged from 225 ± 15 to 416 ± 21 nm at 1 h after incubation in FBS as shown in Figure 2(b).However, the particle size was subsequently decreased to 303 ± 18 nm at 2 h after incubation in FBS.This phenomenon could be explained by the process of association and dissociation of protein molecules on the surfaces of nanoparticles during the course of incubation.These results were also in accordance with previously published study that reported the adsorption, and desorption could also take place during in vivo circumstances [31].Likewise, the continuous fluctuations of the nanoparticles zeta potential in response of continuous attraction-repulsion processes may indicate the positive and negative charge shifting between the particle surface and serum components in order to stabilize the nanoparticles in the serum medium as shown in Figure 2.

Stability in HS.
For in vivo correlation, the stability analysis of unloaded and BSA-loaded CS/DS nanoparticles (produced from DS 0.075% w/v) was further carried out in the HS as presented in Figure 3.The selection of nanoparticles produced from DS 0.075% w/v was based on the criterion that it was observed to be more stable in FBS with respect to its colloidal characteristics (particle size and zeta potential) as shown in Figure 3.The obtained data revealed that a significant increment ( < 0.05, paired t-test) in the particle size and zeta potential of unloaded and BSA-loaded nanoparticles occurred at 10 min after incubation in the HS.The particle size of unloaded and BSA-loaded nanoparticles was increased from 144 ± 17 and 202 ± 20 nm to 382 ± 21 and 367 ± 13 nm, respectively, at 10 min after incubation in HS.The increase in the particle size was expected to be due to the formation of particle aggregates due to an abrupt decline ( < 0.05, paired t-test) in the zeta potential of unloaded and BSAloaded nanoparticles from +61±6 and +50±7 mV to −3±−0.2 and −3±−0.4mV, respectively, after 10 min, incubation in HS.
The particle size of unloaded and BSA-loaded nanoparticles was further increased to 711±22 and 513±21 nm at 24 h after incubation, respectively, when incubated in HS.Conversely, their zeta potential was not changed significantly ( > 0.05, paired t-test) from 10 min to 24 h after incubation as shown in Figures 3(c) and 3(d).However, a significant variation in the particle size and zeta potential of CS/DS nanoparticles regardless of unloaded or BSA-loaded type nanoparticles was expected to be in response of variable interaction between nanoparticles and protein components in HS.Moreover, particle size of both unloaded and BSAloaded CS/DS nanoparticles in FBS and HS was significantly increased ( < 0.05, independent t-test).However, the obtained data further revealed that both the unloaded and BSA-loaded nanoparticles had a higher tendency to form aggregate in HS compared to FBS as shown in Figures 3(a) and 3(b), respectively.For example, in HS, the particle size of BSA-loaded nanoparticles recorded at 1 h after incubation was 492 ± 24 nm, while, in FBS the particle size was 416 ± 21 nm at the same predetermined time point.Similarly, the largest particle size of unloaded nanoparticles in the HS (554 ± 26 nm) was observed to be higher than the one which obtained in the FBS (334 ± 15 nm) at 1 h after incubation.
In addition, the surface charge of unloaded and BSA-loaded CS/DS nanoparticles declined at higher rate and extent in HS compared to FBS during the course of incubation as shown in Figures 3(c) and 3(d).
This diversity could be explained on the basis of disparity into the certain biological components that actively interact with unloaded and BSA-loaded CS/DS nanoparticles in both sera.According to Zelphati et al. [11], antibodies such as globulin and immunoglobulin are more likely abundant in the HS as compared to FBS.Furthermore, FBS is normally derived from calf in which the organ function and enzymes activity were not fully developed and thus resulted in diminished synthesis of antibodies.Meanwhile, the HS is normally sourced from adults in whom the organs and enzymes regulation are functioning optimally and contribute with promising potential to synthesize functional immunoglobulin and globulin.

Swelling Studies.
Figure 4 shows the swelling characteristics of unloaded and BSA-loaded CS/DS nanoparticles in HS and FBS at various CS/DS weight ratios.Generally, the swelling ratio of unloaded and BSA-loaded CS/DS nanoparticles was less affected by the DS concentration.A higher swelling ratio for those nanoparticles prepared at the lowest weight ratio (1 : 1) was expected because high positive charges on the nanoparticles surfaces had provided more crosslinking sites available for bonding with surrounded water (water of hydration).These findings were also in accordance with the previously published data which reported that high surface charge facilitates the strong interaction with water of hydration [32].The resulting interaction between polymer (CS) and water of hydration may facilitate the diffusion of water molecules into the polymer matrices and cause the nanoparticles to swell up.Besides that, a high positive zeta potential of CS/DS nanoparticles promotes a stronger intermolecular electrostatic repulsion which compels the polymer to swell by creating channels into the polymer.Thus, this promotes the penetration of medium between the forming molecules and results in higher swelling ratios.In contrast to that, the CS/DS nanoparticles with higher CS/DS weight ratios (1 : 1.7) have formed a compact infrastructure due to the stronger electrostatic interaction between −NH 3 + groups of CS and −SO 4 − groups donated by DS.The compact structure of nanoparticles will hinder the penetration of incubating medium into the polymer matrices and result in lesser extent of swelling.Lesser extent of swelling could also be explained by the fact that less cross-linking sites available on the surfaces of nanoparticles lead to lower interaction of polymer with water of hydration which did not allow the surrounded water to move inside the polymer channels and thus restrain the swelling of nanoparticles as shown in Figure 4.
Moreover, both unloaded and BSA-loaded nanoparticles had higher percent of swelling in HS as compared with FBS.The variation observed might be in response of varying interaction of nanoparticles with the different components of both sera.Despite that, it could be clearly seen that the unloaded CS/DS nanoparticles had higher swelling extent in both HS and FBS (Figure 4(a)) when compared with BSA-loaded CS/DS nanoparticles (Figure 4(b)).This could be explained on the basis that the BSA-loaded nanoparticles were presented in the swelling media with an extra barrier onto the surface of nanoparticles due to the loaded BSA molecules.This extra layer might prevent the diffusion of water of hydration into the polymer matrices and thus restrain the swelling of BSA-loaded nanoparticles as compared to unloaded one.Secondly, the BSA-loaded nanoparticles have lower cross-linking sites available to interact with surrounded water of hydration because of their lower surface charge caused by neutralization of loaded BSA molecules as compared to unloaded nanoparticles as previously discussed in Section 3.1.

Morphological Examination.
The influence of FBS and HS on the morphology of incubating unloaded and BSAloaded CS/DS nanoparticles was viewed under a TEM.The obtained TEM monograph of unloaded nanoparticles was observed to be smooth spheres before incubating in FBS or HS as shown in Figure 5(a).However, the results depict that the morphology of unloaded CS/DS nanoparticles was generally changed from smooth spheres to irregular one after incubating in FBS as shown in Figure 5(b).Similarly, Figure 5(c) revealed that the unloaded nanoparticles tend to form agglomerates at 24 h after incubation in HS.Taken together, the obtained monograph in Figure 5(c) clearly shows that the unloaded nanoparticles had a higher degree of irregularity as well as aggregation tendency in HS as compared with those incubated in FBS (Figure 5(b)).This finding was in accordance with the results of particle size which explained that the unloaded CS/DS nanoparticles did not show a significant increase in particle size when incubated in FBS at 24 h after incubation (∼200 nm).In contrast to that, the particle size of unloaded nanoparticles was significantly

Figure 3 :
Figure 3: Particle size of unloaded (a), BSA-loaded (b), and zeta potential of unloaded (c) and BSA-loaded (d) CS/DS nanoparticles in HS and FBS (CS 0.075% w/v, BSA 1 mg/mL, mean ± S.D,  = 3).* * Particle size of nanoparticles incubated in HS was significantly different from the ones before incubation and incubated in FBS.Particle surface charge of nanoparticles incubated in HS was significantly different from the ones before incubation and incubated in FBS.

Figure 5 :
Figure5: TEM micrographs of unloaded and BSA-loaded CS/DS nanoparticles before and at 24 h after incubation in human and fetal bovine sera.CS/DS nanoparticles were tend to form aggregates at 24 h after incubation in both FBS and HS (CS 0.075% w/v, BSA 1 mg/mL).