3.2. Formation and Characteristics of Blank and Drug-Loaded Polymeric Nanoparticles
The first characteristic that should be performed for nanoparticle carriers is the evaluation of macroscopic aspects/homogeneity by visual observation of the sample. All samples appeared as turbid, but macroscopically homogenous. Blank NP formulations were milky-white or opaque. The Tyndall effect was easily observed in ERL NPs, and also in PLGA NPs. It was more difficult, but possible to observe the Tyndall effect in PCL NPs. Under the Tyndall effect, the long wavelengths are transmitted to a higher extent than the short wavelengths, and the latter are scattered more strongly. The Tyndall effect is observed in nano-suspensions when the diameters of dispersed particles are in the range from 40 nm to 900 nm. Interestingly, all curcumin-loaded formulations showed different colors. Dispersions of curcumin-loaded ERL, PLGA and PCL NPs were red, yellow-orange and yellow, respectively. Curcumin demonstrates keto-enol tautomerism. The bis-α, β-unsaturated β-diketone exists in equilibrium with its enol tautomer [29
]. These forms have different colors. The bis-keto form predominates in acidic and neutral aqueous solutions and in the cell membrane, whereas the enol form is in the majority of aqueous solutions at pH above 8 and in the solid state. The first pKa (7.5 to 8.5) changes curcumin color from yellow to red. In the pH range between 3 and 7, curcumin acts as a potent donor of hydrogen atoms, while, at pH 8, this compound acts mainly as an electron donor, similar to many phenolic antioxidants. The α,β-unsaturated carbonyl is a good Michael acceptor and undergoes nucleophilic addition [29
]. Diverse colors of different types of polymeric NPs suggest that the quantities of the keto and enol tautomers are different in all these formulations. Because the color of curcumin depends on pH [32
], the pH of NP dispersions were measured. The pH of PLGA NPs was 3.7–3.85. The anions of acidic groups of PLGA may be proton acceptors, and this may facilitate the ionization of a small fraction of curcumin molecules, which can serve, in this case, as proton donors and may explain the yellow-orange color, which suggests the presence of both, keto and enol forms. PCL NPs had pH 5.25–5.55. PCL is a non-charged hydrophobic polymer, and, in this hydrophobic environment, the keto form seems to predominate. The red color observed in curcumin-loaded ERL NPs was similar to that observed at basic pH (9.5–10) by Pourreza and Golmohammadi [32
], but the measured pH, rather than being basic as would be expected, was in the range 5–5.2. It is at basic pH that curcumin exists in the anionic form [30
]. However, due to the fact that ERL is a cationic polymer, it may change the conformation of curcumin and promote ionization of curcumin to produce electrostatic interactions between the anionic form of curcumin and cations of quaternary ammonium groups of ERL. When mixed with acetone, the supernatant from centrifugation of curcumin-loaded ERL NPs containing the non-encapsulated curcumin fraction became yellow, whereas the NP fraction, after mixing with acetone, aggregated and the color became orange-brown. Because of the variable color of different samples, it was necessary to dilute them as much as 1000 times during quantification by UV to provide the same environment so that all the samples would produce yellow color similar to that of the standard.
Another very important parameter which must be taken into account when formulating nanoparticles is the possibility of obtaining a solid state form. As colloidal systems, NPs are thermodynamically unstable and may be susceptible to aggregation after extended periods of storage as a suspension. Long term stability is an important challenge in the development of NPs. Moreover, solid state form may improve the stability of substances that are prone to degradation and unstable in an aqueous environment. Freeze-drying is an attractive approach for achieving long term stability, as suspensions can be converted into solid state materials with greater physical stability than liquids. Lyophilized formulations also provide easy handling including during shipping and storage [33
]. However, it has been shown that stress generated during freeze-drying may adversely impact the properties of NPs [34
]. For this reason, the redispersibility of polymeric NPs after freeze-drying was evaluated. Lyophilized particles showed a cotton-like texture. Blank NPs had a white or grey color; curcumin-loaded ERL, PLGA and PCL NPs powder appeared red, orange-yellow and yellow-brown, respectively. After reconstitution of each formulation in water the color was the same as that before freeze-drying.
The particle size affects the biopharmaceutical properties of the nanocarriers and drug release [35
]. DLS measurements confirmed that polymeric nanoparticles with sizes in the range of hundreds of nanometers were successfully obtained using the single emulsion-solvent evaporation method (Figure 2
). ERL NPs showed the smallest particle size (231 ± 8 nm) of all tested polymers (Figure 2
a). Curcumin loading did not exert a significant effect on particle size (p
= 0.0744). Although ERL NPs were successfully re-dispersed after freeze-drying, a significant increase in particle size, by 100–120 nm, for both blank and curcumin-loaded NPs was observed (p
< 0.0001). The largest particle size was obtained for PCL NPs (512 ± 50 nm) (Figure 2
b). Both lyophilization and curcumin loading affected the particle size of PCL NPs (p
= 0.0001 and p
= 0.0005, respectively). After lyophilization of blank PCL NPs, the particle diameter was doubled and the presence of a small quantity of aggregates was observed. Interestingly, incorporation of curcumin improved the redispersibility of drug-loaded PCL NPs compared with blank PCL NPs. The re-dispersed curcumin-loaded particles showed an average size of 641 nm and were macroscopically homogenous. PLGA NPs were characterized by a diameter of 332 ± 7 nm (Figure 2
c). Although the size of reconstituted blank and curcumin-loaded PLGA NPs was not markedly different from that before freeze-drying (plyophilization
= 0.6617, ploading
= 0.8834 and pinteraction
= 0.9666), the presence of a small quantity of aggregates was observed. These aggregates did not interfere with DLS measurements due to rapid sedimentation. The DLS measurements are in agreement with the visual aspect of the samples and observation of the Tyndall effect.
The polydispersity index is the ratio of size deviation (or width of size distribution) to mean particle diameter [35
]. High polydispersity index values indicate large variations in particle size, whereas polydispersity index values smaller than 0.3 suggest that particles are monodisperse. All formulations were characterized by homogenous size distribution with polydispersity index values in the range between 0.169 ± 0.004 (blank PLGA NPs) and 0.289 ± 0.069 (blank PCL NPs) (Figure 3
). Neither curcumin loading (p
= 0.9689), nor freeze-drying (p
= 0.0926), affected the size distribution of ERL NPs (Figure 3
a). Curcumin loading significantly decreased the polydispersity index of PCL NPs (p
= 0.0200) (Figure 3
b). Lyophilization did not have a significant effect on the size distribution of PCL NPs (p
= 0.0805). However, the polydispersity index of blank PCL NPs after freeze-drying should be interpreted with caution because high values of this parameter and the presence of aggregates indicate that the sample is not suitable for DLS analysis. The improvement of the redispersibility of PCL NPs by curcumin loading is reflected not only by a lower particle diameter, but also by a more homogenous size distribution. Curcumin loading did not affect the size distribution of PLGA NPs (p
= 0.2335) (Figure 3
c). Freeze-drying slightly increased the polydispersity index of PLGA NPs (p
The zeta potential data of blank NPs reflects the charges of the raw polymers. Indeed, the polycationic Eudragit RL, bearing positive charges, conferred by the quaternary ammonium groups (8.8–12%), presented the highest zeta potential (+38.7 ± 4.0 mV) (Figure 4
a). The zeta potential of ERL NPs was not influenced either by curcumin loading (p
= 0.5168) or by lyophilization (p
= 0.3164). As reported many times, PCL and PLGA, being mostly uncharged polymers, displayed negative zeta potential values, close to neutrality. PCL NPs can be considered to be neutral to the zeta potential of −3.6 ± 7.3 mV (Figure 4
b). This is not surprising because the PCL molecules do not contain any ionizable groups in their structure. The zeta potential was not modified either by lyophilization (p
= 0.0805) or by curcumin loading (p
= 0.7338). However, due to the limitations of laser Doppler velocimetry, the neutral zeta potentials are only estimated values and should be interpreted with caution. PLGA NPs exhibited a slightly negative zeta potential of −10.5 ± 1.8 mV (Figure 4
c), which may be due to the ionization of lactic and/or glycolic acid moieties in PLGA molecules. Curcumin loading did not influence the zeta potential of polymeric NPs (p
= 0.4588). Interestingly, the zeta potential in modulus of re-dispersed PLGA NPs increased significantly after freeze-drying (p
= 0.0002). This may be attributed to the rearrangement in surface coverage by PVA molecules and possibly the removal of some PVA molecules from the NP surface and larger exposure of anionic groups of PLGA to the external environment.
It has been reported that in order to effectively stabilize the NPs during freeze-drying and to ensure their adequate reconstitution, suitable excipients are required [34
]. It is possible that PVA plays the role of a stabilizing agent and cryoprotectant during the lyophilization of NPs. PVA is an emulsion stabilizer and stabilizes the polymeric NPs not only during the emulsification step and in solution, but also in the solid state. PVA is known to bind to the surface of PLGA NPs in an irreversible manner. The PVA binding affects the hydrophilicity/hydrophobicity of a particle’s surface and its digestibility [36
]. Boury et al. [37
] have shown that PLGA microparticles, surface coated with PVA, displayed higher water wettability compared with a reference hydrophobic PLGA film surface. Thus, the PVA adsorption may enable wetting and facilitate the redispersion of polymeric NPs after lyophilization. The addition of other cryoprotectants should be considered for more effective freeze-drying, particularly in the cases of PLGA and PCL NPs.
NP morphology was examined by SEM. Both blank and curcumin-loaded ERL NPs appeared as fused particle assemblies (Figure 5
a,b). Because samples were successfully re-dispersed after freeze-drying, as indicated by DLS measurements and visual observation; the presence of poorly formed particles in SEM images can possibly be explained by particle fusion during sputter coating. Moreover, local heating from the SEM beam was found to destroy the particles and alter their morphology to a higher extent than in the case of either PCL or PLGA NPs. PLGA and PCL NPs had spherical shapes (Figure 5
c–f); however, in some cases, deformations were observed due to the presence of neighboring particles. The particles generally appeared as either individual particles or assemblies of non-fused particles; some of them were connected by “bridges”, with a broad range of sizes distributed throughout the sample. The particles displayed a rough surface. No shape modification was observed in the presence of curcumin, independent of the polymer used (Figure 5
3.3. Nanoparticle Recovery, Curcumin Loading and Encapsulation Efficiency
The formulation of NPs involved a purification step and other operations that require material transfer, which can lead to possible loss of NPs. For this reason, it is important to determine NP recovery, also referred to as NP yield. To calculate the NP recovery, only the mass of polymer used for the fabrication of NPs and, in the case of curcumin-loaded particles, the mass of curcumin, was taken into account. NP recovery is shown in Table 2
. Curcumin loading did not affect the NP yield. Approximately 64–67% of PCL and PLGA NPs were recovered. The production yield of ERL NPs was significantly smaller compared with both PLGA and PCL. The smaller yield observed for ERL may be explained by the presence of an important fraction of NPs with a small size that did not undergo sedimentation during the centrifugation process. Indeed, the DLS measurements confirmed that ERL NPs were characterized by the smallest size. Moreover, ERL contains hydrophilic ammonium groups that are present as salts and make this polymer more permeable to water [38
]. Thus, the possible penetration of water into the particle structure can affect particle density and decrease the difference in density between the particles and the external aqueous phase, thereby reducing the sedimentation rate.
NP recovery was determined by dividing the mass of NPs by the mass of polymer carrier and curcumin, whereas drug loading was defined as a ratio of curcumin mass divided by the mass of the polymer carrier and curcumin (Section 2.3
). It is likely that the NPs also contain PVA that was not taken into account in the determination of either DL or NP recovery. It has been demonstrated that PLGA particles with a size below 1 µm have a constant PVA surface density of about 1.8 mg/m2
that is independent of the PVA concentration in the continuous phase of the manufacturing process. The PVA contents were approximately 24 mg/g and 28.2–32.5 mg/g for 0.53 µm and 0.35–0.37 µm particles, respectively [36
When developing an effective nanoparticulate delivery system, drug loading is a key parameter, as low loading often limits the use of such systems, because a substantial amount of the formulation must be administered to achieve the therapeutic effect. For this reason, high drug encapsulation efficiency is important for developing successful formulations. The type of polymer had an important, statistically significant effect on curcumin encapsulation. PCL NPs were characterized by the highest EE from all tested polymers, with 99 ± 0.2% of curcumin incorporated within the NPs. The EE in PLGA NPs (90 ± 1.5%), although significantly smaller than that of PCL NPs, can also be considered very high. Interestingly, the EE was markedly reduced in the case of ERL NPs to 61.7 ± 0.7%. The theoretical loading for all drugs should be 9.1%. The loading obtained for PCL and PLGA NPs are close to this value; however, the loading in ERL NPs was considerably smaller (5.61 ± 0.43%). This is in good agreement with the encapsulation efficiency. There are a few possible explanations for the lowest EE and loading observed for ERL NPs. The affinity of curcumin to ERL may be lower than the affinity to PLGA or PCL. The considerably low yield of ERL NPs may contribute to a low EE because a significant fraction of curcumin could have been lost with the ERL polymer. Another possible explanation for the lower encapsulation of curcumin in ERL NPs could be the fact that some of curcumin molecules underwent the aza-Michael addition to quaternary ammonium groups of ERL. The aza-Michael addition (conjugate reaction of various amines with α,β-unsaturated carbonyl compounds that provides β-amino carbonyl ingredients) does not require a basic catalyst and can occur in water, in contrast with the ‘classic’ Michael reaction. Nevertheless, even ERL NPs were capable of encapsulating quantities of curcumin that were high enough to produce curcumin concentration of at least 500 µg/mL after reconstitution. This concentration is more than 500 times higher than that reported to exert anticancer effects (0.74 µg/mL) against MCF-7 cells [7
3.4. In Vitro Release Studies
The in vitro release studies of curcumin from polymeric NPs were performed to determine the influence of the polymer on the release kinetics. Among the three tested polymers, only ERL NPs released the total amount of encapsulated curcumin, whereas, in the case of either PCL or PLGA NPs, only 55 ± 2% and 47 ± 2% of curcumin, respectively, were released after 24 h (Figure 6
). PLGA and PCL NPs showed the same kinetics of curcumin release, as there was no statistically significant difference either in the percentage of released curcumin at any time point or in the release rate constant (Table 3
). A burst release was observed, with about 43 ± 3% and 52 ± 1% of curcumin released from PLGA and PCL NPs, respectively, within the first hour. After an initial burst release phase, only a very small amount of curcumin (less than 2% and 5% for PCL and PLGA NPs, respectively) was released between 1 h and 24 h. It is generally accepted that the initial phase of rapid drug release from NPs is the fraction weakly bound to or adsorbed on the NP surface. Then, the delayed release is probably due to drug diffusion from the matrix and/or matrix erosion [16
]. The retention of curcumin in the PCL and PLGA NPs may be explained by the high affinity of this hydrophobic compound to those polymers. The retention of hydrophobic active pharmaceutical ingredients, such as promazine or chlorpromazine, in the nanocarriers has already been reported. The percentage of the compound that remained within the nanocarrier depended on its affinity to the nanocarrier and the entrapment efficiency [39
]. It was impossible to extract the curcumin that remained in PCL NPs after release studies. A quantity of 77 ± 5% of curcumin was recovered from PLGA NPs (this includes the released curcumin and curcumin that was extracted from the NPs after release studies using acetone (dilution 1:1) and dichloromethane (dilution 1:10)).
ERL NPs showed a different release profile than the other polymers, because the percentage of released curcumin was markedly different from that released from either PCL or PLGA NPs at any time point (p
< 0.0001). The release of curcumin from ERL NPs can be considered to be immediate, with 57 ± 7%, 81 ± 6% and 91 ± 5% of curcumin released after 5 min, 30 min and 60 min, respectively. The fact that the mass balance was obtained and the total amount of curcumin was recovered may indicate that the aza-Michael reaction either did not occur or was reversible. The higher percentage of released curcumin observed for ERL NPs may be explained by the fact that this polymer, although generally considered to be hydrophobic, contains the ammonium groups that are present as salts, making this polymer permeable to water [38
]. This may facilitate the penetration of the release medium through the polymer matrix, thereby facilitating curcumin release. The release of a higher percentage of curcumin from ERL NPs can be linked with its lower EE compared with PLGA or PCL NPs. Moreover, the fact that ERL NPs released the total amount of encapsulated curcumin may compensate for the lower EE, as indicated by the amount of curcumin released at infinity.
Interestingly, the suspension of curcumin raw material displayed the lowest release rate, with the constant k significantly smaller than that observed for the NP formulations. Curcumin was released gradually and no burst release was observed. The percentage of curcumin released from the suspension of raw material was significantly lower than that released from NPs up to 2 h (p
< 0.0001). The slow release from suspension may be attributed to the large particle size (particles were visible to naked eye) and rapid sedimentation of suspended curcumin, which settled on the bottom of Eppendorf tube and was more difficult to mix than the NPs. Moreover, the contact area of the curcumin suspension with the external aqueous phase was much smaller than that of NPs due to the smaller surface-to-volume ratio. As a hydrophobic material, curcumin probably shows lower wettability than the NPs covered by the surfactant (PVA). It has been reported that the dissolution rate of curcumin from dripping pills is higher than that of pure curcumin because the wettability and solubility of curcumin are improved by surfactants (Poloxamer 188 and Cremophor RH40) [40
To ensure sink conditions during the release studies, the solubility of curcumin in 0.1% w
Tween 80 in PBS was evaluated. The solubility of curcumin after 24 h at 37 °C was 28 ± 5 µg, and it did not change significantly after 48 h. The kinetics of curcumin release in our study were markedly different than those observed in PLGA NPs by Khalil et al. [41
] , who observed prolonged release over a period of 9 days. The difference may be attributed to experimental conditions. Khalil et al. used PBS (0.01 M, pH 7.4) without the addition of surfactants, such as Tween 80, that could potentially increase the solubility of curcumin. Moreover, they used a relatively high concentration of curcumin (1.5 mg in 12 mL of PBS). This concentration is markedly higher than the solubility of curcumin, and, for this reason, the sink conditions were probably not achieved. In another study by Kasinathan et al. [20
], the release of curcumin from PCL NPs has been described. After an initial burst release, the amount of released drug was low and steady. The initial burst release was attributed to the curcumin adsorbed on the NP surface. Although, in their study, the release was more prolonged than in our study, it is very unlikely than the sink conditions were achieved because NPs containing 2.5 mg of curcumin were transferred into 2 mL of PBS [20
According to Biopharmaceutics Classification System (BCS), curcumin belongs to class II, with low solubility in water but high intestinal permeability [40
]. Therefore, rapid release is a desired characteristic for oral administration of this compound, because enhancement of the dissolution rate could increase its bioavailability which is limited by poor solubility.
3.5. Cytocompatibility Studies
The cytocompatibility of free curcumin, blank and curcumin-loaded polymeric NPs to intestinal Caco-2 cells was evaluated using the MTT assay. In living cells, mitochondrial dehydrogenase oxidizes the MTT to formazan product. Damaged or dead cells show reduced or no dehydrogenase activity [42
]. It is generally accepted that if cell viability is higher than 80%, the compound can be considered as non-toxic to cells [42
]. In our study, free curcumin, at a concentration equal to or lower than 5 µg/mL, did not affect the viability of Caco-2 cells (Figure 7
a). However, at a concentration of 50 µg/mL (135.7 µM), free curcumin significantly impacted the metabolic activity of Caco-2 cells (46 ± 2% of viable cells) (p
< 0.0001). It was impossible to test higher concentrations of curcumin, because curcumin absorbance interfered with the MTT test. These results confirmed those obtained by Wahlang et al. [42
], showing the toxicity of curcumin at 265 µM (98 µg/mL) with only 50% Caco-2 cell viability after 24 h of exposure.
Blank NPs did not have any significant effect on the viability of Caco-2 cells with 92.0 ± 2.1%, 86.5 ± 6.3% and 102.5 ± 12.6% cell viability for PLGA, PCL and ERL NPs, respectively, at the highest equivalent curcumin concentration tested (Figure 7
b–d). It is known that polymers that are insoluble in water generally do not interact physically or chemically with living cells unless the surface of the material has very sharp projections or a high density of cationic moiety. However, as a result of the degradation, biodegradable polymers always release low molecular weight compounds into the outer environment, and, in some cases, the physiology of the cell is disturbed by these foreign compounds [44
]. The degradation products of PLGA or PCL are generally considered to be non-toxic. Among tested polymers, only ERL has a high density of cationic moiety. Although the zeta potential of ERL NPs in water was highly positive, it dropped drastically in complete medium with −5.81 ± 0.55 mV and −4.68 ± 0.35 mV for blank and curcumin-loaded ERL NPs, respectively. It has been shown that the charge of chitosan-based particles decreased or even inverted from highly positive to slightly negative in media with high ionic strength and after an increase in pH from slightly acidic to 7.4 [17
]. Such a change in NP charge from positive to negative as a result of pH modification reduced the cytotoxicity of chitosan NPs towards Caco-2 cells [45
]. Furthermore, a change in pH from 5 to 7.4 and an increase in the ionic strength of the medium might reduce the ionization of quaternary ammonium groups of ERL molecules. Moreover, positively charged particles can be surrounded by negatively charged components of the medium and this can also affect their zeta potential and interaction with cells. Furthermore, it has been shown that adsorption of albumin onto ERL NPs results in a considerable increase in particle size and polydispersity index [23
Apart from PLGA NPs, no statistical differences were shown between blank or curcumin-loaded NPs showing that NPs are protective against toxicity from high concentrations of curcumin. The delayed release of curcumin from PCL NPs could explain the decrease in the cytotoxicity of this compound. In the case of long lasting releasing NPs, curcumin was presented gradually in terms of concentration and time allowing cells to adapt to stress conditions. However, the cytotoxicity and the interaction between cells and curcumin-loaded-NPs depend also on particle–cell interactions and particle uptake by cells.