The Impact of the Ionic Cross-Linking Mode on the Physical and In Vitro Dexamethasone Release Properties of Chitosan/Hydroxyapatite Beads

In this study, the effect of the ionic cross-linking mode on the ability to control physical properties and in vitro release behavior of the dexamethasone (DEX) drug from chitosan (CS) and chitosan/hydroxyapatite (CS/HA) beads was investigated. CS solutions without and with HA and DEX were dripped into two coagulation solutions, prepared with a non-toxic ionic crosslinker (sodium tripolyphosphate, TPP) and distilled water, one at pH = 9.0 and other at pH = 6.0. Optical microscopy (OM) and scanning electron microscopy (SEM) results showed changes on the surface topology of the beads, with a reduction of roughness for beads prepared at pH = 6.0 and an increase for the one prepared at pH = 9.0. The diameter and sphericity of the beads prepared at pH = 6.0 proved more uniform and had a larger pore size with a good interconnectivity framework. Attenuated total reflectance-Fourier transform infrared spectroscopy (ATR-FTIR) suggested a higher crosslinking degree for beads prepared at pH = 6.0, corroborated by X-ray diffraction profiles (XRD) analysis that indicated a decrease in the crystalline structure for such beads. In in vitro drug release data, all beads presented a sustained release during the studied period (24 h). The drug release rate was affected by the pH of the coagulation solution used in the preparation of the beads. The in vitro kinetics of the release process was of the Peppas–Sahlin model, controlled by both diffusion and relaxation of polymer chains or swelling (anomalous transport mechanism). Our results suggest that DEX-loaded CS/HA beads, crosslinked in TPP coagulation solution at pH = 9.0, led to a decrease in the DEX release rate and prolonged the release period. Thus, this composition might have prospective as a functional material for bone and cartilage tissue engineering.


Introduction
Polymeric material suitable for biomedical applications must be biocompatible and biodegradable [1]. Due to its biodegradability generating non-toxic degradation products, nonimmunogenic, noncarcinogenic, cytocompatibility, biodegradable, and commercial availability, chitosan (CS) has been gaining increasing importance in tissue engineering, wound healing, and drug delivery [2]. In addition, the cytocompatibility of CS due to its chemical structure, which has a similar

Beads Morphology
Optical microscopy studies revealed that the pH of TPP solution (coagulation solution) used to fabricate CS beads was found to affect their morphology ( Figure 1).

Molecules 2019, 24, x FOR PEER REVIEW 3 of 20
Optical microscopy studies revealed that the pH of TPP solution (coagulation solution) used to fabricate CS beads was found to affect their morphology ( Figure 1). Beads prepared in coagulation solution at pH = 6.0 (CS-TPP6, CS/HA-TPP6, CS/DEX-TPP6, and CS/HA/DEX-TPP6) had a smooth surface. On the other hand, surface roughness was visible on beads prepared at pH = 9.0 (CS-TPP9, CS/HA-TPP9, CS/DEX-TPP9, and CS/HA/DEX-TPP9). This could be attributed to the higher ionic crosslink density within the beads fabricated in the coagulation solution at pH = 6.0. This is consistent with previous reports [20][21][22][23]. In addition, the diameter and sphericity of the beads prepared in coagulation solution at pH = 6.0 (Table 1) have not changed statistically with the incorporation of HA, DEX, and the blend HA/DEX. These data were evaluated using analysis of variance (ANOVA), and the significance of the model was verified with the F test. In the significant models, the averages were compared with Tukey's test, with a significance level of 95% (p < 0.05) using Sisvar 5.6.
Surface and cross-sectional SEM images of CS, CS/HA, CS/DEX, and CS/HA/DEX beads crosslinked in coagulation solution at pH = 9.0 (TPP9) and pH = 6.0 (TPP6), followed by freezing and lyophilization, using the ice particulates as the porogen materials, are shown in Figure 2. All beads crosslinked at TPP9 had a rougher surface, whereas the ones crosslinked at TPP6 seemed to be significantly densified with negligible pores, corroborating with the results of optical microscopy presented in Figure 1.  Beads prepared in coagulation solution at pH = 6.0 (CS-TPP6, CS/HA-TPP6, CS/DEX-TPP6, and CS/HA/DEX-TPP6) had a smooth surface. On the other hand, surface roughness was visible on beads prepared at pH = 9.0 (CS-TPP9, CS/HA-TPP9, CS/DEX-TPP9, and CS/HA/DEX-TPP9). This could be attributed to the higher ionic crosslink density within the beads fabricated in the coagulation solution at pH = 6.0. This is consistent with previous reports [20][21][22][23]. In addition, the diameter and sphericity of the beads prepared in coagulation solution at pH = 6.0 (Table 1) have not changed statistically with the incorporation of HA, DEX, and the blend HA/DEX. These data were evaluated using analysis of variance (ANOVA), and the significance of the model was verified with the F test. In the significant models, the averages were compared with Tukey's test, with a significance level of 95% (p < 0.05) using Sisvar 5.6. Surface and cross-sectional SEM images of CS, CS/HA, CS/DEX, and CS/HA/DEX beads crosslinked in coagulation solution at pH = 9.0 (TPP9) and pH = 6.0 (TPP6), followed by freezing and lyophilization, using the ice particulates as the porogen materials, are shown in Figure 2. All beads crosslinked at TPP9 had a rougher surface, whereas the ones crosslinked at TPP6 seemed to be significantly densified with negligible pores, corroborating with the results of optical microscopy presented in Figure 1. Analyses of cross-sections of the beads showed a pore structure where pore size, size distribution, and interconnectivity were affected by coagulation solution pH ( Figure 2). The beads prepared in coagulation solution at pH = 6.0 (CS-TPP6, CS/HA-TPP6, CS/DEX-TPP6, and CS/HA/DEX-TPP6) showed a structure with more open pores and more uniformly distributed, i.e., with a better interconnectivity framework, especially the ones prepared with DEX (CS/DEX-TPP6) and HA/DEX (CS/HA/DEX-TPP6). Although pore sizes were not measured in this study, it could be inferred that they were smaller for the beads prepared with HA (CS/HA/DEX-TPP6 and CS/HA/DEX -TPP9). This morphology should be suitable as controlled drug delivery devices and could make access to fluids more difficult and retarding both the diffusion process and release rate of drugs. In addition, it could also allow cell penetration and migration, tissue in-growth and vascularization, and nutrient supply within the bone graft, shortening the healing time, and enhance clinical effect [24][25][26]. Furthermore, the existence of the HA phase in the beads will improve the biomineralization and bioactivity [27,28]. The decrease in the porosity with HA addition to chitosan may be because the HA occupied more space in ice in the frozen mixture resulting in the lower porosity. This is in accordance with previously reported literature [25,[29][30][31][32][33].
Finally, HA particles were scattered in the CS bead uniformly (Figure 3a). The presence of the elements, calcium (Ca), phosphorus (P), nitrogen (N) proper of calcium phosphates and the element sodium (Na) and also P typical of crosslinking of the beads with TPP, and the presence of chlorine (Cl) and sodium (Na) peaks that are perhaps related to the residues of the PBS solution used to washing beads (Figure 3b).

Surface
Cross-sectional Surface Cross-sectional Analyses of cross-sections of the beads showed a pore structure where pore size, size distribution, and interconnectivity were affected by coagulation solution pH ( Figure 2). The beads prepared in coagulation solution at pH = 6.0 (CS-TPP6, CS/HA-TPP6, CS/DEX-TPP6, and CS/HA/DEX-TPP6) showed a structure with more open pores and more uniformly distributed, i.e., with a better interconnectivity framework, especially the ones prepared with DEX (CS/DEX-TPP6) and HA/DEX (CS/HA/DEX-TPP6). Although pore sizes were not measured in this study, it could be inferred that they were smaller for the beads prepared with HA (CS/HA/DEX-TPP6 and CS/HA/DEX -TPP9). This morphology should be suitable as controlled drug delivery devices and could make access to fluids more difficult and retarding both the diffusion process and release rate of drugs. In addition, it could also allow cell penetration and migration, tissue in-growth and vascularization, and nutrient supply within the bone graft, shortening the healing time, and enhance clinical effect [24][25][26]. Furthermore, the existence of the HA phase in the beads will improve the biomineralization and bioactivity [27,28]. The decrease in the porosity with HA addition to chitosan may be because the HA occupied more space in ice in the frozen mixture resulting in the lower porosity. This is in accordance with previously reported literature [25,[29][30][31][32][33].
Finally, HA particles were scattered in the CS bead uniformly (Figure 3a). The presence of the elements, calcium (Ca), phosphorus (P), nitrogen (N) proper of calcium phosphates and the element sodium (Na) and also P typical of crosslinking of the beads with TPP, and the presence of chlorine (Cl)
After the CS crosslinking process, spectral changes were detected ( Figure 4a). The first spectral changes observed were in the 3600-2700 cm −1 region, where the OH stretching (3440 cm −1 ) and NH2 symmetric stretching (3362 cm −1 ) and NH2 asymmetric stretching (3294 cm −1 ) absorption bands involved in hydrogen bonds became less prominent and broadened with crosslinking process, especially for CS beads crosslinked in TPP solution at pH = 6.0 (CS-TTP6, CS/HA-TPP6, CS/DEX-TPP6, and CS/HA/DEX-TPP6). Likewise, a broadening and decrease in the intensity of the CH asymmetric stretching vibration peak (2875 cm −1 ) were observed. The reason for these variations is the break of some hydrogen bonds in the CS structure during the crosslinking process and the establishment of new H-bonds between CS and TPP [34]. Moreover, the band at 1650 cm −1 (C=O stretching in amide group, -NHCOCH3 due to the partial deacetylation of chitosan, amide I vibration) shifted to 1640 cm −1 (CS-TPP9) and 1630 cm −1 (CS-TPP6). The band at 1585 cm −1 (N-H bending in primary amine groups, -NH2, overlapping the amide II vibration) shifted to 1556 cm −1 (CS-TPP9) and
After the CS crosslinking process, spectral changes were detected ( Figure 4a). The first spectral changes observed were in the 3600-2700 cm −1 region, where the OH stretching (3440 cm −1 ) and NH 2 symmetric stretching (3362 cm −1 ) and NH 2 asymmetric stretching (3294 cm −1 ) absorption bands involved in hydrogen bonds became less prominent and broadened with crosslinking process, especially for CS beads crosslinked in TPP solution at pH = 6.0 (CS-TTP6, CS/HA-TPP6, CS/DEX-TPP6, and CS/HA/DEX-TPP6). Likewise, a broadening and decrease in the intensity of the CH asymmetric stretching vibration peak (2875 cm −1 ) were observed. The reason for these variations is the break of some hydrogen bonds in the CS structure during the crosslinking process and the establishment of new H-bonds between CS and TPP [34]. Moreover, the band at 1650 cm −1 (C=O stretching in amide group, -NHCOCH 3 due to the partial deacetylation of chitosan, amide I vibration) shifted to 1640 cm  The shift and intensity of the bands were altered by reaction conditions. The shift and intensity of the bands were higher for the CS beads crosslinked in the TPP solution at pH = 6.0 (TTP6) in comparison with the ones prepared at pH = 9.0 (TPP9) (Figure 4a). The reason may be attributed to the degree of protonation that was higher for CS-TPP6 than for CS-TPP9, 18 and 14, respectively, as estimated by comparing the ratio of the peak area at 1650 cm -1 (NH3 + vibrational band) to the peak area at 1380 cm -1 (which was unaffected by the crosslinking process) (A1650/A1380) ( Figure 5) [35]. Thus, for CS crosslinked in TPP solution at pH = 6.0, tripolyphosphate and H + ions predominate and react with amino protonated groups (-NH3 + ) chitosan mainly by ionic crosslinking. On the other hand, CS crosslinking in the TPP solution of pH = 9.0, a smaller number of chitosan amino groups was protonated. In this solution, tripolyphosphate and OH − ions were present, and they could  The shift and intensity of the bands were altered by reaction conditions. The shift and intensity of the bands were higher for the CS beads crosslinked in the TPP solution at pH = 6.0 (TTP6) in comparison with the ones prepared at pH = 9.0 (TPP9) (Figure 4a). The reason may be attributed to the degree of protonation that was higher for CS-TPP6 than for CS-TPP9,~18 and 14, respectively, as estimated by comparing the ratio of the peak area at 1650 cm −1 (NH 3 + vibrational band) to the peak area at 1380 cm −1 (which was unaffected by the crosslinking process) (A1650/A1380) ( Figure 5) [35]. Thus, for CS crosslinked in TPP solution at pH = 6.0, tripolyphosphate and H + ions predominate and react with amino protonated groups (-NH 3 + ) chitosan mainly by ionic crosslinking. On the other hand, CS crosslinking in the TPP solution of pH = 9.0, a smaller number of chitosan amino groups was protonated. In this solution, tripolyphosphate and OH − ions were present, and they could competitively react with the -NH 3 + of chitosan by ionic crosslinking and deprotonation, respectively [22,23,[36][37][38]. Therefore, the ionic crosslinking density of chitosan beads prepared at pH = 6.0 can be higher than at pH = 9.0. Other researchers [23,[39][40][41][42][43][44][45] observed similar results.
Finally, the spectra of CS/HA, CS/DEX, and CS/HA/DEX beads exhibited characteristic peaks typically present in chitosan (Figure 4a). The effect of HA and DEX characteristic bands on the beads could be distinguished clearly on CS/HA, CS/DEX, and CS/HA/DEX beads spectra, since the typical ATR-FTIR bands of HA and DEX were overlapping with those of CS. Even so, the peak at 1020 cm −1 corresponded to the P-O bond in the HA bulk spectrum (Figure 4b) could be observed on CS/HA spectra (Figure 4a), indicating the presence of HA in the CS beads prepared. In addition, characteristic peaks at the 3600-2700 cm −1 region for CS/DEX beads were broadened, suggesting an increase in the hydrogen bonds between DEX drug and CS polymeric chains corroborate with the literature [5,49].
The assignments and corresponding band position for CS, TPP, HA, and DEX are described in Table 2 and are in agreement with the literature [10,42,43,[50][51][52][53][54].  The spectra of chitosan beads prepared at pH = 6.0 (CS-TPP6, CS/HA-TPP6, CS/DEX-TPP6, and CS/HA/DEX-TPP6) also exhibited a new band at 1210 cm −1 (Figure 4a), consistent with the literature [46][47][48]. This suggests that chitosan might bind with TPP ions more easily at lower pH conditions (pH = 6.0). Hence, the above mentioned changes observed in the FTIR spectra of chitosan after reaction with crosslinking agent TPP confirm that pH of coagulation solution affected the chemical structure of chitosan beads in agreement with the literature [23,39,43].
Finally, the spectra of CS/HA, CS/DEX, and CS/HA/DEX beads exhibited characteristic peaks typically present in chitosan (Figure 4a). The effect of HA and DEX characteristic bands on the beads could be distinguished clearly on CS/HA, CS/DEX, and CS/HA/DEX beads spectra, since the typical ATR-FTIR bands of HA and DEX were overlapping with those of CS. Even so, the peak at 1020 cm −1 corresponded to the P-O bond in the HA bulk spectrum (Figure 4b) could be observed on CS/HA spectra (Figure 4a), indicating the presence of HA in the CS beads prepared. In addition, characteristic peaks at the 3600-2700 cm −1 region for CS/DEX beads were broadened, suggesting an increase in the hydrogen bonds between DEX drug and CS polymeric chains corroborate with the literature [5,49].

XRD Analysis
The X-ray diffraction patterns of non-crosslinked chitosan (CS) and crosslinked CS beads at coagulation solution pH The diffraction pattern of CS, as shown in Figure 6, exhibited two major crystalline peaks at 10.8 • and 20.0 • , which confirms the partial crystallinity of the polymer, in agreement with previously reported results [55][56][57][58]. These peaks (10.8 • and 20.0 • ) are typical fingerprints of chitosan related to hydrated and anhydrous crystals, respectively [56]. After crosslinking process, the intensity of these peaks in the XRD patterns significantly decreased (Figure 6), mainly for the beads crosslinked at coagulation solution pH = 6.0 (CS-TPP6, CS/HA-TPP6, CS/DEX-TPP6, CS/HA/DEX-TPP6). The peak at 10.8 • , besides presenting lower intensity also presented a little shift in the diffraction angles as shown in Figure 6. The peak at 20.0 • , assigned to chitosan chains aligned through intermolecular interactions, became wider and weaker for the beads crosslinked at TPP6 (Figure 6), which corresponds to so-called amorphous halo [23,43]. These results indicated that crystallinity of the chitosan decreased after its crosslinking with TPP especially for the beads crosslinked at coagulation solution pH = 6.0 (TPP6). This is because the CS beads crosslinked at pH = 6.0 (TPP6) were completely ionic-crosslinking dominated and those beads crosslinked at pH = 9.0 (TPP9) were dominated by deprotonation accompanied with slightly ionic crosslinking, corroborating ATR-FTIR data, in agreement with other reports in the literature [35,37,45].  [59][60][61][62], can be clearly seen in Figure 6. It can also be seen that the broad peak assigned to chitosan at 20.0° becomes wider and weaker with the incorporation of HA. It suggests that the addition of HA obviously affects the crystallinity of chitosan, resulting in a more amorphous structure, possibly due to reduced interactions between chitosan molecules by the presence of HA. Other researchers observed the same behavior [63][64][65][66].
XRD spectra of the CS/DEX-TPP9, CS/DEX-TPP6, CS/HA/DEX-TPP9, and CS/HA/DEX-TPP6 beads (Figure 6), show peaks at 2 value of 13.6° and 16.0°, which are reflections of dexamethasone (DEX), confirming the presence of DEX crystalline particles at these beads. The incorporation of DEX into CS and CS/HA beads has not affected the crystallinity of chitosan.

In Vitro Drug Release
The in vitro release behavior of dexamethasone (DEX) from the CS and CS/HA beads prepared at coagulation solution pH = 9.0 (TPP9) and pH = 6.0 (TPP6) was investigated through UV-visible spectroscopy in PBS up to 24 h at pH 7.4 and 37 °C (Figure 7). According to release profiles, all beads presented a sustained release during the studied period (24 h), but the drug release rate was affected by the pH of the coagulation solution used in the preparation of the beads. The beads prepared at pH  [59][60][61][62], can be clearly seen in Figure 6. It can also be seen that the broad peak assigned to chitosan at 20.0 • becomes wider and weaker with the incorporation of HA. It suggests that the addition of HA obviously affects the crystallinity of chitosan, resulting in a more amorphous structure, possibly due to reduced interactions between chitosan molecules by the presence of HA. Other researchers observed the same behavior [63][64][65][66].
XRD spectra of the CS/DEX-TPP9, CS/DEX-TPP6, CS/HA/DEX-TPP9, and CS/HA/DEX-TPP6 beads (Figure 6), show peaks at 2θ value of 13.6 • and 16.0 • , which are reflections of dexamethasone (DEX), confirming the presence of DEX crystalline particles at these beads. The incorporation of DEX into CS and CS/HA beads has not affected the crystallinity of chitosan.

In Vitro Drug Release
The in vitro release behavior of dexamethasone (DEX) from the CS and CS/HA beads prepared at coagulation solution pH = 9.0 (TPP9) and pH = 6.0 (TPP6) was investigated through UV-visible spectroscopy in PBS up to 24 h at pH 7.4 and 37 • C (Figure 7). According to release profiles, all beads presented a sustained release during the studied period (24 h), but the drug release rate was affected by the pH of the coagulation solution used in the preparation of the beads. The beads prepared at pH = 6.0 (CS/DEX-TPP6 and CS/HA/DEX-TPP6) showed a faster release than the one prepared at pH = 9.0 (CS/DEX-TPP9 and CS/HA/DEX-TPP9). Around 50% of DEX drug was released from the CS/DEX-TPP6 and CS/HA/DEX-TPP6 beads in 42 min and 1 h 36 min, respectively, whereas the same amount of drug was released from the CS/DEX-TPP9 and CS/HA/DEX-TPP9 beads in 2 h 40 min and 3 h, respectively (Figure 7). Beads of CS/DEX-TPP6 and CS/HA/DEX-TPP6 released 100% of the drug within 24 h, whereas about 72% and 67% of the loaded drug from CS/DEX-TPP9 and CS/HA/DEX-TPP9 were released, respectively, within this time. Since the beads prepared at TPP9 exhibits higher crystalline and smaller pore size, according to XRD and SEM data, compared to those prepared at TPP6, the access to release medium was more difficult; thus, both the diffusion process and release rate of drugs were retarded [67].
Molecules 2019, 24, x FOR PEER REVIEW 10 of 20 = 6.0 (CS/DEX-TPP6 and CS/HA/DEX-TPP6) showed a faster release than the one prepared at pH = 9.0 (CS/DEX-TPP9 and CS/HA/DEX-TPP9). Around 50% of DEX drug was released from the CS/DEX-TPP6 and CS/HA/DEX-TPP6 beads in 42 min and 1 h 36 min, respectively, whereas the same amount of drug was released from the CS/DEX-TPP9 and CS/HA/DEX-TPP9 beads in 2 h 40 min and 3 h, respectively ( Figure 7). Beads of CS/DEX-TPP6 and CS/HA/DEX-TPP6 released 100% of the drug within 24 h, whereas about 72% and 67% of the loaded drug from CS/DEX-TPP9 and CS/HA/DEX-TPP9 were released, respectively, within this time. Since the beads prepared at TPP9 exhibits higher crystalline and smaller pore size, according to XRD and SEM data, compared to those prepared at TPP6, the access to release medium was more difficult; thus, both the diffusion process and release rate of drugs were retarded [67]. Figure 7 also shows that the incorporation of HA in the CS beads led to a decrease in the DEX release rate and prolonging the release period. The addition of HA in CS indeed reduced the water absorption due to the reduction of polar hydroxyl groups and amino groups in CS/HA beads [68]. As shown in Scheme 1, hydroxyl ions on the surface of HA, Ca10(PO4)6(OH)2, might interact with the amino and hydroxyl ions of chitosan by the formation of hydrogen bonds [63,69]. This is in agreement with the XRD data ( Figure 6), where it was observed a decrease in chitosan crystallinity with the incorporation of HA, suggesting a hydrogen-bonding interaction between chitosan and hydroxyapatite. These interactions result in the increased physical crosslinking density, and subsequently, due to the decrease in the pore sizes of beads (Figure 2), the water absorption of CS/HA beds is decreased. Therefore, the DEX was strongly kept in the network of the CS/HA beads, causing a decrement in the diffusion of the drug through the beads. Mahdavinia et al. [69] and Asadian-Ardakani, et al. [70] reported identical data for another drug.  Figure 7 also shows that the incorporation of HA in the CS beads led to a decrease in the DEX release rate and prolonging the release period. The addition of HA in CS indeed reduced the water absorption due to the reduction of polar hydroxyl groups and amino groups in CS/HA beads [68]. As shown in Scheme 1, hydroxyl ions on the surface of HA, Ca 10 (PO 4 ) 6 (OH) 2 , might interact with the amino and hydroxyl ions of chitosan by the formation of hydrogen bonds [63,69]. This is in agreement with the XRD data ( Figure 6), where it was observed a decrease in chitosan crystallinity with the incorporation of HA, suggesting a hydrogen-bonding interaction between chitosan and hydroxyapatite. These interactions result in the increased physical crosslinking density, and subsequently, due to the decrease in the pore sizes of beads (Figure 2), the water absorption of CS/HA beds is decreased. Therefore, the DEX was strongly kept in the network of the CS/HA beads, causing a decrement in the diffusion of the drug through the beads. Mahdavinia et al. [69] and Asadian-Ardakani, et al. [70] reported identical data for another drug. Scheme 1. Hydrogen bonds between chitosan and hydroxyapatite [63].

Kinetics of Release
To analyze the release mechanism of the dexamethasone (DEX) from chitosan (CS) and the chitosan/hydroxyapatite (CS/HA) beads, prepared at two coagulation solutions (pH = 9.0 and pH = 6.0), over release behavior, the experimental drug release data were fitted to three kinetic models: zero-order, Higuchi, and Peppas-Sahlin [71][72][73][74]. DDSolver (Software Version, Company, City, Country) [75], an add-in program for modeling and comparison of drugs, was used for the analyses. The criteria for selecting the most appropriate model were based on the best goodness of fit of the experimental results, which is based on the statistically higher values of the adjusted coefficient of determination (r 2 ), the lower value of Akaike information criterion (AIC), and the largest value of model selection criterion (MSC) [75]. Table 3 summarizes the values obtained for r 2 , AIC, and MSC considering the CS and CS/HA beads crosslinked at pH = 9.0 (CS/DEX-TPP9 and CS/HA/DEX-TPP9) and pH = 6.0 (CS/DEX-TPP6 and CS/HA/DEX-TPP6). The results show that the DEX release from CS and CS/HA beads, prepared at both pH, showed good fitting with Peppas-Sahlin model, as is exhibited in Figure 8, where higher values of r 2 , lower values of AIC, and higher values of MSC were recorded (Table 3). This model is a semi-empirical model that describes the drug release kinetic from hydrophilic polymers and is applicable for different geometric shapes. This kinetic model points to a release mechanism that relies on the effects of Fickian diffusion and relaxation of polymer chains or swelling (anomalous transport mechanism). Thus, neither absolute Fickian diffusion (as a result of pure drug diffusion) nor zeroorder (because of polymer chain relaxation) was the predominant mechanism in this case. This can be assumed as a combination of both contributions of polymer relaxation and drug diffusion factors that determine the drug release mechanism [74,76,77]. According to these results, it can be concluded that coagulation solution pH does not show a significant influence in the drug transport mechanism (anomalous in all the cases).

Kinetics of Release
To analyze the release mechanism of the dexamethasone (DEX) from chitosan (CS) and the chitosan/hydroxyapatite (CS/HA) beads, prepared at two coagulation solutions (pH = 9.0 and pH = 6.0), over release behavior, the experimental drug release data were fitted to three kinetic models: zero-order, Higuchi, and Peppas-Sahlin [71][72][73][74]. DDSolver (Software Version, Company, City, Country) [75], an add-in program for modeling and comparison of drugs, was used for the analyses. The criteria for selecting the most appropriate model were based on the best goodness of fit of the experimental results, which is based on the statistically higher values of the adjusted coefficient of determination (r 2 ), the lower value of Akaike information criterion (AIC), and the largest value of model selection criterion (MSC) [75]. Table 3 summarizes the values obtained for r 2 , AIC, and MSC considering the CS and CS/HA beads crosslinked at pH = 9.0 (CS/DEX-TPP9 and CS/HA/DEX-TPP9) and pH = 6.0 (CS/DEX-TPP6 and CS/HA/DEX-TPP6). The results show that the DEX release from CS and CS/HA beads, prepared at both pH, showed good fitting with Peppas-Sahlin model, as is exhibited in Figure 8, where higher values of r 2 , lower values of AIC, and higher values of MSC were recorded (Table 3). This model is a semi-empirical model that describes the drug release kinetic from hydrophilic polymers and is applicable for different geometric shapes. This kinetic model points to a release mechanism that relies on the effects of Fickian diffusion and relaxation of polymer chains or swelling (anomalous transport mechanism). Thus, neither absolute Fickian diffusion (as a result of pure drug diffusion) nor zero-order (because of polymer chain relaxation) was the predominant mechanism in this case. This can be assumed as a combination of both contributions of polymer relaxation and drug diffusion factors that determine the drug release mechanism [74,76,77]. According to these results, it can be concluded that coagulation solution pH does not show a significant influence in the drug transport mechanism (anomalous in all the cases). order (because of polymer chain relaxation) was the predominant mechanism in this case. This can be assumed as a combination of both contributions of polymer relaxation and drug diffusion factors that determine the drug release mechanism [74,76,77]. According to these results, it can be concluded that coagulation solution pH does not show a significant influence in the drug transport mechanism (anomalous in all the cases).

Preparation of Chitosan/Hydroxyapatite Beads (CS/HA)
Chitosan/hydroxyapatite beads were manufactured by ionotropic gelation followed by lyophilization. First, a 2% (w/v) chitosan solution was prepared by dissolving chitosan in a 1% (v/v) aqueous solution of glacial acetic acid. The polymer solution was maintained under constant mechanical stirring (IKA, RW 20, Staufen, Germany) at 75 rpm, at room temperature (25 ± 1 • C) until it became a homogeneous solution. Thereafter, HA powder with a concentration of 20% (w/v) was added to the chitosan solution, which was maintained under mechanical stirring at 75 rpm until it became a homogeneous mixture. The CS/HA suspension was dripped, using a syringe pump (Pump 11 Pico Plus Elite, Harvard Apparatus, Holliston, MA, USA) and 22G × 1 2 gauge needles at 25 mL/h, into two coagulation baths (TPP/distilled water solutions). One at pH = 9.0 (coded as TTP9) and other at pH = 6.0 (coded as TPP6) with the same TPP concentration in both baths concentrations (0.1 M). The beads were left in the precipitating solutions for 24 h, at 75 rpm, and 25 ± 1 • C to crosslinking of the beads. Then, the beads were washed with distilled water and PBS solution to a neutral pH (7.4). Afterward, frozen in a freezer (Brastemp clear -410, São Bernardo dos Campos, SP, Brazil), at approximately −18 • C for 24 h and lyophilized in a lyophilizer (L108, AISI304-LIOTOP stainless steel, São Carlos, SP, Brazil) at −56 • C for 48 h. Subsequently, they were dried at room temperature. The schematic representation of the fabrication of the CS/HA beads through two coagulation baths (TTP9 and TPP6) is shown in Figure 9. CS beads were also fabricated for comparison purposes. into two coagulation baths (TPP/distilled water solutions). One at pH = 9.0 (coded as TTP9) and other at pH = 6.0 (coded as TPP6) with the same TPP concentration in both baths concentrations (0.1 M). The beads were left in the precipitating solutions for 24 h, at 75 rpm, and 25 ± 1 °C to crosslinking of the beads. Then, the beads were washed with distilled water and PBS solution to a neutral pH (7.4 Figure 9. CS beads were also fabricated for comparison purposes.

Preparation of Chitosan/Hydroxyapatite/Dexamethasone Beads (CS/HA/DEX)
The CS/HA/DEX beads were prepared by the same method described in Section 3.2. Initially, the drug dexamethasone (DEX) was dissolved in ethanol by dispersing 10% (based on the CS mass) (10 mg mL) under mechanical agitation at 75 rpm at room temperature (25 ± 1 °C). Subsequently, added to the CS/HA solution and maintained under mechanical stirring until homogeneity of the system and the following steps were similar to those presented in Section 3.2.

Optical Microscopy (OM)
The dimensions of the beads were determined using an optical microscope (Hirox-KH 7700, Tokyo, Japan). About 10 beads were placed on a dry glass slide, positioned at the base of the microscope. The images were made by applying the 5040Z lens and a 100X magnification. The measurements of the beads were carried out using a 5040LOW lens with magnification 20X using the 2D_MEASURE software (2.3.1 Version, Tokyo, Japan). The volumetric diameter (Dv), surface

Preparation of Chitosan/Hydroxyapatite/Dexamethasone Beads (CS/HA/DEX)
The CS/HA/DEX beads were prepared by the same method described in Section 3.2. Initially, the drug dexamethasone (DEX) was dissolved in ethanol by dispersing 10% (based on the CS mass) (10 mg mL) under mechanical agitation at 75 rpm at room temperature (25 ± 1 • C). Subsequently, added to the CS/HA solution and maintained under mechanical stirring until homogeneity of the system and the following steps were similar to those presented in Section 3.2.

Optical Microscopy (OM)
The dimensions of the beads were determined using an optical microscope (Hirox-KH 7700, Tokyo, Japan). About 10 beads were placed on a dry glass slide, positioned at the base of the microscope. The images were made by applying the 5040Z lens and a 100× magnification. The measurements of the beads were carried out using a 5040LOW lens with magnification 20× using the 2D_MEASURE software (2.3.1 Version, Tokyo, Japan). The volumetric diameter (Dv), surface diameter (Ds), volume (V), surface area (As), and sphericity (S) of the samples were calculated by equations: (1) (3) where A is the area of the particle, and P is the perimeter.

Scanning Electron Microscopy (SEM)
The surface and cross-sectional morphology of the beads was observed using scanning electron microscopy (SEM) Phenom-World, Pro-X800-07334 (Eindhoven, The Netherlands). The beads were placed in double carbon tape. The central part of the beads was sectioned with a surgical blade to examine the cross-sectional morphology. SEM images of each bead were taken by applying an electron beam accelerating voltage of 15 kV, with a depth of focus of 1 mm and a resolution of 30 nm. The beads surface deposition of HA was investigated using SEM with energy dispersive spectrum (EDS) (EDAX, Oxford, UK).

Attenuated Total Reflectance-Fourier Transform Infrared Spectroscopy (ATR-FTIR) Analysis
The beads (CS, CS/HA, CS/DEX and CS/HA/DEX), finely grinded using a mortar and pestle, and their pure components (TPP, HA, and DEX) were analyzed by attenuated total reflectance-Fourier transform infrared spectroscopy (ATR-FTIR) using a Perkin Elmer 400 FTIR Spectrometer (Perkin Elmer, MA, USA) equipped with attenuated total reflectance (ATR) accessory employing the Zn-Se crystal. Spectra were recorded over the range of 4000 to 600 cm −1 at 4 cm −1 resolution, and each spectrum represents 64 co-added scans referenced against an empty ATR cell spectrum.

X-Ray Diffraction (XRD)
X-ray diffraction profiles (XRD) were recorded on a Shimadzu diffractometer model XRD-7000 (Shimadzu, Tokyo/Kyoto, Japan) with Ni-filtered Cu-Kα radiation, a voltage of 40 kV and a current of 30 mA, to identify the composition and crystallinity of the beads (CS, CS/HA, CS/DEX, and CS/HA/DEX), prepared into two coagulation baths (TPP6 and TPP9). The scanning range was from 2θ = 2 • to 60 • at a scan speed of 0.033 • s −1 . For this characterization, the beads were cryogenically ground and then placed in the sample holder of the diffractometer.

In Vitro Drug Release
The amount of dexamethasone (DEX) released from CS and CS/HA beads was measured using an ultraviolet-visible (UV-VIS) spectrophotometer (Shimadzu, Model 1800, Kyoto, Japan). First, the bead samples (exactly heavy, with masses corresponding to a concentration of 25 µg/mL) were incubated in a volumetric flask in a 50 mL PBS solution (pH = 7.4), as a release medium, and kept under constant stirring of 150 rpm in an IKA shaker model KS4000i (Werke, Germany) at 37 • C ± 0.5 • C to simulate body temperature. Subsequently, an aliquot of the release medium (4 mL) was withdrawn at fixed time intervals (0, 0.5, 1, 2, 3, 4, 5, and 24 h), and replaced with 4 mL of a fresh release medium (at 37 • C ± 0.5 • C) each time. The collected samples were then analyzed for DEX content by measuring the absorbance at 242 nm (λmax of DEX in phosphate buffer solution pH 7.4 measured using a UV-VIS spectrophotometer). All the release experiments were carried out in triplicate, and the average values were used for further data treatment and plotting. The drug concentration was calculated according to a standard curve, and Equation (6) determined the cumulative release. The corresponding drug-release profiles were represented through plots of the cumulative percentage of drug release (calculated from the total amount of DEX contained in each matrix) versus time.
Cummulative release(%) = n i=0 C i V 0 m ×100 where V 0 is the sampling volume (4.0 mL), Ci is the concentration (mg/L) of the released drug collected at time t, and m is the mass of the drug incorporated in the polymer (1.25 mg).

Release Kinetics
In vitro release data were analyzed by zero-order (Equation (7)), Higuchi (Equation (8)), and Peppas-Sahlin (Equation (9)) kinetic models [71][72][73][74][78][79][80], using the DDSolver [75], an add-in program for modeling and comparison of drugs [81]. This study was carried out to investigate the mechanism of dexamethasone (DEX) release from CS and CS/HA beads, prepared into two coagulation solutions (TPP9 and TPP6). The criteria for selecting the most appropriate model were based on the best goodness of fit of the experimental results, based on the statistically (i) higher values of adjusted coefficient of determination R 2 ); (ii) lower value of Akaike information criterion (AIC), and (iii) largest value of model selection criterion (MSC) [75,82].
where Q t is the amount of drug released at time t; Q 0 the amount of drug in the solution at t = 0; (usually, Q 0 = 0) and k 0 the zero-order release constant.
Q t = k H t 1/2 (8) k H representing the Higuchi rate constant.
Q ∞ is the maximum amount of the drug released measured at the equilibrium, k 1 and k 2 are the release constants, and m is the release exponent.

Conclusions
Beads of chitosan (CS) and chitosan/hydroxyapatite (CS/HA) loaded with steroid anti-inflammatory agent, dexamethasone (DEX), crosslinked into two sodium tripolyphosphate (TPP) coagulation solutions, one at pH = 9.0 and other at pH = 6.0, with the same concentration of TPP in both solutions (0.1 M), were fabricated by freezing and lyophilization method. The morphology of the beads was influenced by the coagulation solution pH. The beads crosslinked at neutral pH had a rougher surface, whereas the ones crosslinked at acid pH had a smooth surface. The diameter and sphericity of the beads prepared at acid pH proved uniform. Likewise, at this pH, SEM data suggested a larger pore size with a good interconnectivity framework. The crosslinking degree was higher for the beads prepared at acid, as suggested by FTIR-ATR data, corroborating with XRD analysis that indicated a decrease in the crystalline structure for the beads prepared at pH = 6.0. According to in vitro drug release data, all beads presented a sustained release during the studied period (24 h). However, the ones prepared with HA at basic pH (pH = 9.0) led to a decrease in the DEX release rate and prolonging the release period.
The in vitro kinetic of DEX from CS and CS/HA beads, prepared at both pH, was best described using the Peppas-Sahlin model. This kinetic model points to a release mechanism that relies on the effects of Fickian diffusion and relaxation of polymer chains or swelling (anomalous transport mechanism). Thus, neither absolute Fickian diffusion (as a result of pure drug diffusion) nor Zero-order (because of polymer chain relaxation) was the predominant mechanism in this case. This can be assumed as a combination of both contributions of polymer relaxation and drug diffusion factors that determine the drug release mechanism. All of these results demonstrate that DEX-loaded CS/HA beads, crosslinked at TPP coagulation solution at basic pH, may be a promising drug carrier for bone tissue engineering.