The Effect of Polymer Microstructure on Encapsulation Efficiency and Release Kinetics of Citropin 1.1 from the Poly(ε-caprolactone) Microparticles

Cationic antimicrobial peptides represent a promising therapeutic option against multidrug-resistant bacteria for the treatment of local infections. However, due to their low stability and potential toxicity, there are limited possibilities for their application in clinical practice. In this study, different poly(ε-caprolactone) (PCL) microparticles (MPs) loaded with citropin 1.1 (CIT) were investigated in order to demonstrate the effect of the polymer microstructure on the encapsulation efficiency (EE) and kinetics of the peptide release from the newly developed devices. The characteristics of the new systems in terms of surface morphology, particle size, EE and zeta potential analysis, as well as the haemolytic activities of the peptide were investigated. The in vitro release kinetics of CIT from the MPs was also investigated. CIT loading was favoured by a high content of negative charged linear polymer chains in the PCL structure. The presence of non-charged, amorphous macrocycle domains results in faster degradation of the PCL matrix. Depending on the crystallinity of the PCL, the peptide release exhibited a near-zero-order or near-first-order profile with no “burst release”. The results indicated that CIT-loaded PCL MPs could potentially be a promising drug delivery system (DDS) for the treatment of local infections.


Introduction
Cationic antimicrobial peptides (AMPs) represent a promising group of active pharmaceutical ingredients (APIs) for a new therapeutic option against multidrug-resistant bacteria in the treatment of local infections. Furthermore, depending on their structure and origin, AMPs may also show antitumor, angiogenic, immunomodulatory and wound healing activities [1,2].
Despite their promising therapeutic potential, as low molecular weight cationic molecules, peptides have relatively low stability in vivo due to a more favourable interaction with the anionic

Double Emulsion Microspheres' Preparation
The CIT (1 mg) was dissolved in PBS (1 mL) and the solution was added to the 5 mL (1.5% w/v or 3.0% w/v) of the PCL/DCM solution to make a primary (W/O) emulsion using a homogenizer (SilentCrusher, Heidolph Instruments GmbH & Co, Schwabach, Germany) for 1 min at 10,000 rpm. The primary (W/O) emulsion was added dropwise to the 8 mL of 5% (w/v) PVA/distilled water solution and homogenized for 5 min at 10,000 rpm to produce a final double (W/O/W) emulsion. The 60 mL of 1% (w/v) PVA/distilled water solution was then added dropwise and stirred mechanically for 4 h at 400 rpm at room temperature until the organic solvent was thoroughly evaporated. Finally, the mixture was washed further with distilled water under centrifugation to thoroughly remove PVA. Then, the MPs were preserved at 4 • C.

Determination of the Encapsulation Efficiency
The microspheres were dissolved in DCM and the resulting solution was added to 1 mL of PBS and stirred for 24 h. The mixture was then centrifuged for 10 min at 35,000 rpm and the supernatant aqueous phase containing CIT was then quantified by the high-performance liquid chromatography (HPLC) method, according to previously described procedures [10]. The samples were prepared in triplicate. The peptide EE was calculated using the following equation (Equation (1)): EE = (measured drug content/theoretical drug content) × 100 (1)

In Vitro Studies of Citropin 1.1 Release From the Microspheres
The CIT-loaded PCL microspheres were incubated with 1mL of PBS (pH 7.4) at 37 • C. The samples were centrifuged at 3,500 rpm for 10 min and a supernatant from each run was removed for analysis at the selected time intervals, after which this was replaced with a new PBS solution. The quantity of CIT released from the PCL microspheres was analysed using the HPLC method. The samples were prepared in triplicate.
Release data were analyzed on the basis of zero-order and first-order mathematical models as well as the Higuchi model and the Korsmeyer-Peppas model (Table 1) [11][12][13]. Table 1. Mathematical models of kinetic or release mechanism.

Mathematical Model Equation
Zero-order model F=kt n (F < 0.6) F: fraction of drug released up to time (t), F 0 : initial concentration of drug, k: constant of the mathematical models, n-exponent of Korsmeyer-Peppas model.

Hemolytic Activity of the Peptide and Calculation of the Therapeutic Index
The haemolytic activity of CIT was measured as the amount of haemoglobin released according to Reference [14] with a small own modification. Human Red Blood Cells (hRBC) with EDTA was rinsed three times with saline (0.9% NaCl) using centrifugation for 10 min at 800 x g and was then re-suspended in saline. Serial dilution of the peptides (64, 128, 192, 256 and 512 µg·mL −1 ) was performed in saline on 96-well plates. A stock solution of hRBC was then added to obtain a final volume of 100 µL with a 4% concentration of erythrocytes (v/v). The control wells for haemolysis of 0 and 100% contained hRBC suspended in saline and 1% Triton-X 100, respectively. The plates were then incubated for 60 min at 37 • C and then centrifuged at 800 x g for 10 min at 4 • C. Following centrifugation, The therapeutic Index (TI) of the CIT in our investigations was calculated using the following equation (Equation (2)): where MHC: minimal hemolytic concentration of CIT to hRBC; MIC: minimal inhibitory concentration of CIT to methilicin-resistant S. aureus (MRSA) strains, commonly known as the major pathogen in surgical wound infections.

Measurements
The polymerization products were characterized by means of 1 H and 13 C NMR (Varian 300 MHz) at room temperature, with CDCl 3 as solvent. The infrared Fourier transform (FT-IR) spectra were measured from KBr pellets (Perkin Elmer spectrometer, Perkin Elmer, Warsaw, Poland).
The morphological assessment of the MPs was investigated by scanning electron microscope (SEM) FEI Quanta 250 FEG (FEI Inc., Eindhoven, The Netherlands).
The average particle size and zeta potential of the microspheres were determined by a dynamic light scattering (DLS) technique using a Zetasizer Nano ZS instrument (Malvern Instruments, Westborough, MA, USA) equipped with red laser at a wavelength of 633 nm and scattering angle of 173 • at 25 • C.

Morphology, Size and Zeta Potential of MPs
The results of the SEM investigations revealed that spherical, non-porous and smooth PCL MPs were successfully obtained. There were no morphological differences between AMPs-loaded and placebo MPs prepared using both the single-and double-emulsion methods. The crucial parameter for the preparation of non-porous MPs was the appropriate amount of the aqueous phase in the final emulsion. Moreover, moderate evaporation of DCM during the last stage of the process is also required to obtain smooth PCL MPs [15]. Figure 1A presents the SEM micrograph of the CIT-loaded PCL MPs obtained using the optimum amount of the aqueous external phase to diffuse DCM in the emulsion. Figure 1B shows the SEM micrograph of the MPs prepared using the same procedure as reported in the 'Experimental' section but with a smaller amount of the aqueous external phase. Consequently, the fast evaporation of an organic solvent causes leads to irregular, high porous and rough PCL MPs being obtained.
The size and surface charge of the polymeric MPs are the key factors responsible for their cellular uptake and toxicity effect in vivo. The mean diameter of the MPs prepared using a single emulsion method ranged from 3.69 to 3.92 µm with a narrow particle size distribution (0.22-0.24) in the case of both matrices ( Table 2). The mean diameter of the placebo and the CIT-loaded MPs, prepared using a double-emulsion solvent evaporation method, was slightly higher and ranged from 3.90 to 5.56 µm with a dyspersity values of 0.13-0.38.
MPs are able to prevent the toxic effects associated with the electrostatic interaction between negative-charged glycoproteins molecules embedded in the hRBC membrane and cationic APIs [16]. Moreover, surface charge affects the stability of the MPs in suspensions. Zeta potential measurements revealed that all MPs have negative surface charge (from −19.7 to −12.2 mV), which could be explained by the presence of the carboxylic end-group of the linear PCLs chains (Table 2) [17]. The size and surface charge of the polymeric MPs are the key factors responsible for their cellular uptake and toxicity effect in vivo. The mean diameter of the MPs prepared using a single emulsion method ranged from 3.69 to 3.92 μm with a narrow particle size distribution (0.22-0.24) in the case of both matrices ( Table 2). The mean diameter of the placebo and the CIT-loaded MPs, prepared using a double-emulsion solvent evaporation method, was slightly higher and ranged from 3.90 to 5.56 μm with a dyspersity values of 0.13-0.38.
MPs are able to prevent the toxic effects associated with the electrostatic interaction between negative-charged glycoproteins molecules embedded in the hRBC membrane and cationic APIs [16]. Moreover, surface charge affects the stability of the MPs in suspensions. Zeta potential measurements revealed that all MPs have negative surface charge (from −19.7 to −12.2 mV), which could be explained by the presence of the carboxylic end-group of the linear PCLs chains (Table 2) [17].

Encapsulation Efficiency of CIT
The EE of the peptide is a function of the characteristics of the peptide and the MPs microstructure. This depends on the electrostatic and hydrophobic interactions between polymer and cationic peptides. In our experiments, the preparation of CIT-loaded MPs was successfully achieved using a double-emulsion method, which is widely used for the encapsulation of water-soluble APIs [18][19][20].
Moreover, we noted some differences in the EE of CIT between MPs originating from polymers with different microstructures. In the case of PCL-1, the EE was in the range of 42-47% and was lower than that of PCL-2 (51-52%). This is probably due to the differences in the polymer microstructure, which consists of non-charged macrocycles and negatively charged linear PCL chains. The presence of the negatively charged molecules in the MPs structure allows for an electrostatic interaction between the carboxyl end-group of the linear polymer and the cationic peptides, which results in enhanced EE. PCL-1 with Xc of 50.5% that have a high content of non-charged macrocycles in their structures (ca. 35% [9]), thus the EE in this case was lower. In contrast, we noted higher EE values for MP-PCL-2 (Xc = 56.3%) with a lower content of macrocyclic chains (19% [9]).

Encapsulation Efficiency of CIT
The EE of the peptide is a function of the characteristics of the peptide and the MPs microstructure. This depends on the electrostatic and hydrophobic interactions between polymer and cationic peptides. In our experiments, the preparation of CIT-loaded MPs was successfully achieved using a double-emulsion method, which is widely used for the encapsulation of water-soluble APIs [18][19][20].
As shown in Table 2  Moreover, we noted some differences in the EE of CIT between MPs originating from polymers with different microstructures. In the case of PCL-1, the EE was in the range of 42-47% and was lower than that of PCL-2 (51-52%). This is probably due to the differences in the polymer microstructure, which consists of non-charged macrocycles and negatively charged linear PCL chains. The presence of the negatively charged molecules in the MPs structure allows for an electrostatic interaction between Nanomaterials 2018, 8, 482 6 of 10 the carboxyl end-group of the linear polymer and the cationic peptides, which results in enhanced EE. PCL-1 with X c of 50.5% that have a high content of non-charged macrocycles in their structures (ca. 35% [9]), thus the EE in this case was lower. In contrast, we noted higher EE values for MP-PCL-2 (X c = 56.3%) with a lower content of macrocyclic chains (19% [9]).

The In Vitro Kinetics Release of CIP from PCL MPs
The in vitro kinetic release of CIT from the obtained MPs was determined at pH 7.4, 37 • C over about 72 h. The in vitro release peptide profiles are presented in Figure 2. The ordinate of the plot was calculated based on the cumulative amount of CIT released. MPs were obtained from two PCL matrices with varying crystallinity. Our main intention was to examine how the kinetics of CIT release depends on the crystallinity of the polymeric matrices. It has, in fact, been reported that PCL degradation is very slow in an aqueous medium because of its semi-crystallinity and hydrophobicity. However, water can penetrate easily into the amorphous regions of the polymer matrix, facilitating the release of the drug by diffusion [21][22][23]. A comparison of CIT release from MPs with varying amounts of the peptide was also analyzed.

The In Vitro Kinetics Release of CIP from PCL MPs
The in vitro kinetic release of CIT from the obtained MPs was determined at pH 7.4, 37 °C over about 72 h. The in vitro release peptide profiles are presented in Figure 2. The ordinate of the plot was calculated based on the cumulative amount of CIT released. MPs were obtained from two PCL matrices with varying crystallinity. Our main intention was to examine how the kinetics of CIT release depends on the crystallinity of the polymeric matrices. It has, in fact, been reported that PCL degradation is very slow in an aqueous medium because of its semi-crystallinity and hydrophobicity. However, water can penetrate easily into the amorphous regions of the polymer matrix, facilitating the release of the drug by diffusion [21][22][23]. A comparison of CIT release from MPs with varying amounts of the peptide was also analyzed.  It was found that the difference in the observed release rate for the MPs was mainly attributed to the difference in the X c of the PCL matrices. The rate of in vitro peptide release increased as the X c of the matrices decreased. The percentage of the peptide released after 33 h of incubation was about 59.7% for the MP-PCL-2-CIT-0.5, 71.4% for the MP-PCL-2-CIT-1.0, 84.3% for theMP-PCL-1-CIT-0.5 and 90.3% for the MP-PCL-1-CIT-1.0. It was also observed that the release rates varied depending on the amount of CIT encapsulated in the MPs. The release was slower for MPs with a higher amount of peptide, which could be attributable to the high hydrophobicity nature of the CIT (ca. 56% of hydrophobic amino acids in sequence). MPs consisting of different initial amounts of the peptide released 100% of CIT in about 72, 55, 48 and 33-48 h for MP-PCL-2-CIT-0.5 (51.5% of the initial amount of peptide), MP-PCL-2-CIT-1.0 (50.5% of the initial amount of peptide), MP-PCL-1-CIT-0.5 (46.8% of the initial amount of peptide) and MP-PCL-1-CIT-1.0 (42.2% of the initial amount of peptide).
The release data points were subjected to zero-order and first-order mathematical models, as well as the Higuchi model and the Korsmeyer-Peppas model, to evaluate the kinetics and release mechanisms of the CIT from the obtained MPs (Table 3).
In the Korsmeyer-Peppas model, the value of n characterizes the release mechanism of the peptide. According to the Korsmeyer-Peppas model, in the case of spheres, 0.43 ≤ n corresponds to a Fickian diffusion mechanism, 0.43 < n < 0.85 to non-Fickian transport, n = 0.85 to case II transport and n > 0.85 to super case II transport [11][12][13][14]. It was shown that the CIT release from MP-PCL-2-CIT-1.0 occurred in one-step. The peptide was released with a near-zero-order kinetics (R 2 = 0.9878). This suggests that CIT release from the MP-PCL-2-CIT-1.0 is a highly controlled process. Interestingly, the peptide was released in two steps from MP-PCL-2-CIT-0.5. However, CIT was also released in accordance to the near-zero-order kinetics form MPs in both phases. The controlled peptide release profiles were obtained with no significant "burst release". High R 2 values (0.9708 and 0.9835) were obtained for the near-zero-order kinetics model. The MPs obtained from the PCL-1 show a slightly less controlled release profile. The CIT release exhibited a more near-first-order release profile. The This suggests that the CIT release was governed mainly by super case II transport. However, in the case of phase II of MP-PCL-2-CIT-0.5 (in phase I), the n value was 0.79, which indicates that the anomalous transport (non-Fickian) dominates in these systems. Table 3. Analysis data of CIT release from MPs. As presented in Table 3, the data was also suited to the Higuchi model, with correlation coefficients of 0.9714-0.9887. According to this model, the liquid penetrates the MPs and dissolves the embedded CIT, thus the peptide release seems to be a process predominately controlled by diffusion. As commonly known, drug release from nano-or microparticles generally follows diffusion/degradation or a combination of diffusion and degradation-mediated release phenomena. It was shown that CIT was released from MP-PCL-2-CIT-0.5 and MP-PCL-2-CIT-1.0 according to near-zero-order kinetics in case II transport or super case II transport. In turn, in the cases of MP-PCL-1-CIT-0.5 and MP-PCL-1-CIT-1.0, the peptide was released according to near-first-order kinetics by super case II transport. Analysis of the models suggested that peptide release depended more on the erosion of the PCL than on the diffusion process. However, as is known, mathematical models do not provide an exact insight into the mechanism of peptide release.

Hemolytic Activity of Peptides and Calculation of the Therapeutic Index
Membrane-active peptides with a high content of hydrophobic amino acids in their sequences could interact with eukaryotic cells and provide some toxic effects in vivo. The mechanisms of their interaction depend on the type of the target cells. In the case of hRBC, the toxic effect is a result of the pore formation of the peptide in cell membranes, whereas AMPs induce apoptosis in Human White Blood Cells (hWBC) [24]. In our study, we chose hRBC as the test cells to determine the TI of CIT according to Chen et al. [25] TI is a widely used parameter for determining the AMPs specificity towards prokaryotic cells [24]. TI is described as the ratio of minimal haemolytic concentration (MHC) and minimal inhibitory concentration (MIC), thus larger TI values indicate the greater specificity of the peptides toward the bacteria being tested. In our investigation, we considered the TI in relation to the MIC of the methicillin-resistant S. aureus (MRSA) strains. S. aureus is commonly known to be the major pathogen in surgical wound infections. The MIC values of CIT toward MRSA were in the range of 1-16 µg·mL −1 , as previously reported [26]. The results of toxicological assays (Figure 3) revealed 9% haemolysis at 128 µg·mL −1 and 74% haemolysis at 512 µg·mL −1 .
To calculate TI we chose the highest reported MIC value (16 µg·mL −1 ), and the TI value of the CIT in relation to MRSA strains was eight. The results revealed that CIT indicates a relative high specify toward MRSA, hence it could be used as a potential therapeutic agent against multidrug-resistant bacteria in surgical infection. towards prokaryotic cells [24]. TI is described as the ratio of minimal haemolytic concentration (MHC) and minimal inhibitory concentration (MIC), thus larger TI values indicate the greater specificity of the peptides toward the bacteria being tested. In our investigation, we considered the TI in relation to the MIC of the methicillin-resistant S. aureus (MRSA) strains. S. aureus is commonly known to be the major pathogen in surgical wound infections. The MIC values of CIT toward MRSA were in the range of 1-16 μg·mL −1 , as previously reported [26]. The results of toxicological assays (Figure 3) revealed 9% haemolysis at 128 μg·mL −1 and 74% haemolysis at 512 μg·mL −1 .
To calculate TI we chose the highest reported MIC value (16 μg·mL −1 ), and the TI value of the CIT in relation to MRSA strains was eight. The results revealed that CIT indicates a relative high specify toward MRSA, hence it could be used as a potential therapeutic agent against multidrugresistant bacteria in surgical infection.

Conclusions
Poly(ε-caprolactone) (PCL) based micropartices (MPs) loaded with citropin 1.1 (CIT) were successfully prepared and investigated. MPs were obtained from PCL with varying crystallinity and different contents of CIT. It was found that CIT was released from the obtained MPs with rather controlled kinetics. In some cases, the peptide was released from the carriers with a near-zero-order release kinetics. It is also worth noting that peptide "burst release" was not observed. Importantly, the MPs loaded with CIT exhibited antimicrobial activity during c.a. 72 h of the degradation process. The development of the obtained MPS should be of great interest in the delivery systems of antimicrobial agents.
Author Contributions: U.P. conceived, designed, and directed the studies, performed the preparation, characterization of the microparticles, and wrote the manuscript. M.S., E.O., and W.K. performed the peptide release studies (tests and their interpretation) and helped write the paper. S.B. performed hemotoxicity assays (tests and their interpretation). All authors read and approved the final manuscript.
Funding: This work was financially supported by the National Science Centre (Poland), PRELUDIUM 6 research

Conclusions
Poly(ε-caprolactone) (PCL) based micropartices (MPs) loaded with citropin 1.1 (CIT) were successfully prepared and investigated. MPs were obtained from PCL with varying crystallinity and different contents of CIT. It was found that CIT was released from the obtained MPs with rather controlled kinetics. In some cases, the peptide was released from the carriers with a near-zero-order release kinetics. It is also worth noting that peptide "burst release" was not observed. Importantly, the MPs loaded with CIT exhibited antimicrobial activity during c.a. 72 h of the degradation process. The development of the obtained MPS should be of great interest in the delivery systems of antimicrobial agents.