Effect of Temperature on Drug Release: Production of 5-FU-Encapsulated Hydroxyapatite-Gelatin Polymer Composites via Spray Drying and Analysis of In Vitro Kinetics

In this study, 5-fluorouracil(5-FU-) loaded hydroxyapatite-gelatin (HAp-GEL) polymer composites were produced in the presence of a simulated body fluid (SBF) to investigate the effects of temperature and cross-linking agents on drug release. The composites were produced by wet precipitation at pH 7.4 and temperature 37°C using glutaraldehyde (GA) as the cross-linker. The effects of different amounts of glutaraldehyde on drug release profiles were studied. Encapsulation (drug loading) was performed with 5-FU using a spray drier, and the drug release of 5-FU from the HAp-GEL composites was determined at temperatures of 32°C, 37°C, and 42°C. Different mathematical models were used to obtain the release mechanism of the drug. The morphologies and structures of the composites were analyzed by X-ray diffraction, thermal gravimetric analysis, Fourier transform infrared spectroscopy, and scanning electron microscopy. The results demonstrated that for the HAp-GEL composites, the initial burst decreased with increasing GA content at all three studied temperatures. Further, three kinetic models were investigated, and it was determined that all the composites best fit the Higuchi model. It was concluded that the drug-loaded HAp-GEL composites have the potential to be used in drug delivery applications.


Introduction
Hydroxyapatite (HAp,Ca 10 (PO 4 ) 6 (OH) 2 ) is a biomaterial found in bone and teeth, which has an apatite-like structure [1,2]. Because of its excellent chemical stability, biocompatibility, bioactivity, nontoxicity, and osteoconductivity, HAp has been used for the production of synthetic bone materials and bone and teeth implants and in drug delivery applications [2][3][4][5][6][7][8]. Studies have described diverse methods for synthesizing HAp, such as hydrothermal [9,10], sol-gel [11], microemulsion [12], precipitation [13], solid-state reaction [14], and microwave [15] methods. However, the highly brittle and stiff structure of HAp limits its usage in clinical applications [16,17]. These mechanical disadvantages can be overcome by adding a polymer. Thus, many studies have utilized natural polymers, such as collagen, chitosan, gelatin, alginate, and starch-based materials to modify HAp and produce high-quality HAp-bioceramics [8,[18][19][20][21][22][23]. Biopolymers have excellent biocompatibility, more biodegradability, and adequate osteoconductivity [17,24]. Among these polymers, gelatin (GEL) is widely used in drug release systems alone or in composite form [25]. In addition, gelatin has efficient swelling-releasing characteristics and does not produce antigens [16]. However, its drawbacks are high water solubility and low mechanical properties. To prevent its solubility in cell culture, cross-linking is applied [24]. Cross-linking of GEL increases its mechanical and thermal strength. For the chemical cross-linking of gelatin, cross-linking agents, such as glutaraldehyde and formaldehyde, can be applied. Glutaraldehyde (GA) is the most commonly applied agent because of its low cost and high stabilization efficiency [26][27][28][29]. Bera et al. produced HAp-GEL nanocomposites in different ratios and investigated their biocompatibility. It has been concluded that the use of high GEL concentrations covers HAp crystals more intensively and prevents their growth of the crystals [30]. HAp-GEL composite microspheres have been prepared using a wet-chemical method. The porosity of HAp-GEL provides a high ratio of surface area to void regions, making it suitable for use as a drug carrier [17,31]. A composite immunoisolation membrane combining GA-cross-linked GEL and HAp was designed for enclosing microencapsulated insulinoma cells [32]. Ghorbani et al. synthesized HAp-GEL scaffolds, containing dexamethasoneloaded poly(lactide-co-glycolide) microspheres by a freeze casting technique. The results showed that the HAp-GEL scaffolds have the necessary features to regenerate defects, especially at low freezing gradients and high amounts of HAp, and can serve as a basis for future cellular studies to evaluate the effect of these drug-release constructs on cellular differentiation and expression of osteogenic markers [33]. Ture [34] developed a composite wound dressing material consisting of alginate and gelatin with HAp to replace existing wound dressing materials. Examination of the swelling behaviors of the films indicated that the water retention capacities of the material decreased as the amounts of alginate and HAp in the films increased.
Because of the undesired side effects and high costs of treatments systematically carried out with antibiotics, studies regarding the controlled drug release conducted with drugloaded composite materials have gained importance. In the controlled drug release systems, the release profile begins with small changes followed by an initial burst release, after which the release continues at almost a constant rate [35]. In the studies on drug release conducted with bioactive substances, drug release has been observed to be associated with the porosity and permeability of the substance [36]. The porous structure of HAp is similar to that of a bone's inorganic phase because of which HAp has become a preferred substance for drug loading. Studies on HAp-heparin/-GEL trilayer composites showed microspheres to be more convenient for usage in the drug release systems because of their suitability for controlled drug release at long periods of time and their high drug loading efficiency [37]. Teng et al. produced HAp-GEL microspheres by the water emulsion inside oil method and concluded that the microspheres obtained ensured homogeneous drug distribution. The existence of pores increases the drug loading capacity and is convenient for use in controlled drug release [38].
Previous studies have analyzed the loading capacity of drugs such as ibuprofen and tetracycline and their release by HAp-GEL composites containing implant and nanoparticles [35,39]. Chen-Dou et al. examined the drug release profile of HAp-GEL composites loaded with minocycline, which is a type of semisynthetic tetracycline-derived antibiotics. It was concluded that the stable and slow release observed during the study may be related to the starting of the deformation of gelatin and hydroxyapatite [40]. The drug release profiles of gentamicin- [41], insulin- [42], and vitamin D3-(VD3-) [43] loaded HAp-GEL microspheres were examined, and the release mechanism suggested included the absorption of the environment liquid within the pores of the microspheres, dissolution of drug/insulin/VD3 in the liquid, and release of the liquid from the pores together with drug/insulin/VD3. 5-FU used in this study has been widely applied in the treatment of many cancers such as colon, breast, ovary, pancreas, stomach, brain, and skin cancer [44,45]. Encapsula-tion of the drug using natural and synthetic polymers with controlled 5-FU release for the purpose of maximizing the therapeutic properties and minimizing side effects has been studied. Some polymers used for drug release include chitosan [45], gelatin [46], poly(lactide-co-glycolide) [25], poly(d,llactide) [47], poly(methylidenemalonate 2.1.2) [48], and pectin [49] with 5-FU. In addition, researchers have also studied controlled drug release with composite substances such as 5-FU-loaded HAp/PLGA [50], gelatin/chitosan [51], alginate/pluronic F127/eudragitRS 100 [52], and magnetic nanocomposite [53]. Santos et al. produced 5-FU-loaded HAp nanoparticles using a spray drier, and it was concluded that HAp is not sufficient alone for controlled release [54]. The speed of 5-FU release was controlled by changing the amount of the cross-linking agent [46]. In studies on GEL-chitosan microspheres in which GA was used as the cross-linking agent and 5-FU as active principle, the initial burst release ratio was observed to decrease when the cross-linking agent amount was increased [46,51]. The drug release profile of 5-FU-loaded HAp/PLGA composite microspheres was examined, and when the HAp/PLGA ratios were varied, it was observed that the initial burst release decreased with increasing HAp ratio [50].
Drug release mechanisms vary depending on many factors, such as the pore dimension of the drug carrier substance [31], bondability, drug/carrier interaction, and drug/polymer deformation speed. Therefore, the release mechanism needs to be determined before application [55]. Mathematical models can be used for the release kinetics using physical parameters, such as drug diffusion coefficient, and can also contribute to the optimization of the release process [56]. The in vitro drug release mechanism can be determined by measurements taken in phosphate buffer saline (PBS) or simulated body fluid (SBF) environments [57]. The pH of tumor tissues is lower than that of the normal tissues. pH controllable drug delivery system has become a hot topic because of the different pH in the physiological environment (~7.4) and tumor cell (~5.0-7.0). To achieve the drug release, stable drug delivery systems must be prepared first [58][59][60][61][62]. When acidic conditions (pH = 5) were applied, a faster release rate of the drug was observed [63]. Some tumor tissues have different temperatures compared with the host basal temperature because of an increased metabolic rate. Moreover, additional temperature differences can be induced by the external heating of the tumor region, e.g., by ultrasound treatment, magnetic field, or light-sensitive imaging techniques, such as radiology and optical imaging [64][65][66][67]. Thus, the effect of temperature on the drug release profile can be investigated at different temperatures.
In the current study, HAp-GEL composites were fabricated in the presence of SBF, and spray drying was used to obtain drug-loaded composites. 5-FU, which is widely used for the treatment of different cancer types, was selected as the drug [44]. Composites were produced by wet precipitation using GA as the cross-linking agent. To observe the effect of temperature, drug release studies were performed at temperatures of 32°C, 37°C, and 42°C in PBS at pH 4.0. Drug release profiles were evaluated according to three different kinetic models including zero order, first order, and 2 International Journal of Polymer Science Higuchi. The obtained composites were characterized by XRD, TGA, FTIR, and SEM.
To the best of our knowledge, the impact of the different release temperatures of HAp-GEL composites loaded with 5-FU using a spray dryer has not been investigated thus far.

Materials and Methods
Experimental stages consisted of the preparation of SBF and PBS, the production of HAp-polymer composites in SBF with GA (two different ratios as 2% (v/v) and 5% (v/v)) as a cross-linking agent, the encapsulation (drug loading) of 5-FU in HAp-polymer using a spray dryer, and in vitro drug release in a PBS medium at different temperatures.

SBF and PBS Preparation.
To obtain the SBF, the reactants in Table 1 were added to 750 mL deionized water in the given order and were dissolved under constant stirring at 37°C. To prevent an instant pH increase, (CH₂OH)₃CNH₂ was dissolved slowly. Following the addition of the reactants, 1 M HCl was used to adjust the pH to 7.4. The solution was kept at 25°C for 1 day, and afterwards, its volume was made to 1 L by adding deionized water. K 2 HPO 4 (80.2 mL, 1 M) and 19.8 mL 1 M KH 2 PO 4 were prepared and mixed to obtain the PBS solution. The mixture was completed to 1 L with deionized water and was adjusted to pH 4.0 with 1 M HCl solution.

Production of HAp-Polymer Composites in an SBF
Medium. Figure 1 shows the production steps of HAp-GEL polymer composites. Ca(OH)₂-SBF and H₃PO₄ (85%)-GEL-SBF solutions were prepared separately and were mixed using a mechanical mixer for 2 h at 37°C and 400 rpm. Then, the H₃PO₄ (85%)-GEL-SBF solution was fed to the Ca(OH)₂-SBF using a peristaltic pump at a feeding rate of 5 mL/min leading to the formation of HAp crystals in the solution. If necessary, the pH value of the solution was adjusted to 7.4 with the help of 1 M HCl or 1 M NaOH solutions. The obtained solution was mixed at 37°C and 400 rpm for another 2 h and was allowed to rest. After 24 h of aging for the completion of the HAp crystal growth, a GA-deionized water solution (5%/2%) (v/v) was added to the solution for crosslinking and was mixed at 600 rpm for 3 h. Then, the solution was filtered, and precipitated composites were washed with sodium bisulfate (3%) and deionized water to remove all nonreacted GA. After filtering, the precipitate was dried in an oven at 40°C for 24 h. Composites with a 1 : 1 HAp/GEL weight ratio were obtained and cross-linked with a 2% (v/v) and 5% (v/v) GA-deionized water solution.

Encapsulation (Drug Loading) Process in Spray Dryer.
Seventy-five milligrams of drug was dissolved in 300 mL deionized water. This drug solution was mixed at 800 rpm for 30 min. Then, 3 g of polymer-composite sample was added, and mixing was continued. The 5-FU-loaded HAp-GEL composites were prepared using a spray dryer (Yamato ADL310) at 80°C inlet temperature of hot air, 5 mL/min inlet flow of feeding solution, and 0.1 MPa atomizer pressure.  3 International Journal of Polymer Science 2.5. In Vitro Drug Release in the PBS Medium. Twelve milligrams of drug-loaded HAp-GEL sample was weighed and added to tubes containing 20 mL of PBS. In vitro release studies were carried out in a shaking water bath at 200 rpm. The composites were suspended in the PBS medium at 32°C, 37°C, and 42°C and pH 4.0. At predetermined time intervals (5, 10, 15, 30, and 45 min and 1, 2, 3, and 5 h), samples were taken from the solution and were centrifuged for several minutes. Each time after removing the samples, fresh PBS was added to the solution. The 5-FU concentration in the release medium was evaluated using a UV-vis spectrophotometer (Jenway 6305) at a wavelength of 266 nm in triplicate. The released cumulative drug amounts were calculated using the attained absorbance values, and release profiles were formed according to zero order, first order, and Higuchi kinetic models.
(ii) First order (iii) Higuchi model where Q and M t are the cumulative amounts of drug released at time t, C 0 is the initial concentration of the drug, and The zero-order model describes a system where the drug release rate is independent of its concentration. The firstorder model is used to describe the absorption and elimination of some drugs, although it is difficult to understand the mechanism on a theoretical basis. Finally, the Higuchi model describes the drug release as a diffusion process based on Fick's law and is a square-root-of-time-dependent mechanism. This model is often applicable to different geometrics and porous systems [68][69][70][71][72].

Results and Discussions
3.1. XRD Analysis of HAp-GEL Composites. The XRD patterns of the composites cross-linked with different amounts of GA (2% and 5% GA solutions) are presented in Figure 2 and indicate the existence of HAp in the composites. For both samples shown in Figure 2 [73,74]. Moreover, calcium ions and (R-COO)groups of GEL interact with each other by covalent bonds, which leads to the formation of the Ca-GEL complex. This complex stopped the reaction of Ca-P ions, thus inhibiting hydroxyapatite formation [75][76][77]. Additionally, 210 and 110 peaks were bonded together, which could be attributed to the interaction of gelatin and hydroxyapatite [41].

TGA Analysis of the HAp-GEL Composites.
The thermogravimetric analysis of the composite shows a gradual degradation of the organic content. The comparative TGA of HAp-GEL (1 : 1) composites, pure HAp, and pure GEL is given in Figure 3, which shows the weight loss (%) vs. temperature (°C) curves for the HAp-GEL-2% GA and HAp-GEL-5% GA. The GEL phase in the HAp-GEL composites is gradually degraded. The first step of the curve is related to adsorbed water molecules, which disappear at temperatures above 90°C [78]. Around 160°C, the GEL molecules gradually start to degrade and continue to burn steadily till about 300°C, and then the rate of burning gradually decreases. Finally, the pyrolysis of the remaining organic content commences at about 520-530°C. When the temperature reaches 550°C, only the inorganic HAp phase in the composites remained undegraded. This gradual degradation and pyrolysis profiles of the composites give an insight into both intramolecular and intermolecular interactions of GEL in the composites. The initial degradation probably starts with the cleaving of intramolecular bonds of GEL (a denatured protein), resulting in the loss of secondary structures commencing at about 160°C [30].

International Journal of Polymer Science
The GEL ratio was higher in the HAp-GEL 5% GA composite. Thermal, mechanical, and moisture resistance can be increased by forming covalent bonds between aldehydebased cross-linking agents and GEL chains [79]. By increasing the amount of the cross-linking agent, the proportion of GEL remaining dissolved in the SBF and removed by filtration could be reduced during the composite production stage. In this way, the amount of GEL interacting with HAp increased. The increase in the amount of GEL retained in the composite along with the increased amount of crosslinking agent was confirmed by TGA.  [46,51,80,81].

FTIR Analysis of the HAp-GEL Composites.
In Figures 4(b) and 4(c), new 5-FU peaks emerge in the loaded composites, as expected. Some extra bands attributed to 5-FU and ions in the SBF are observed. Around 1400 cm −1 (aromatic ring), which overlaps with the band of CO 3 ions, the bands at 870 cm −1 (CF=CH groups) and 3000-3500 cm −1 (unbonded F group) show the encapsulation of 5-FU [82]. Moreover, the band at 870 cm −1 could be attributed to the FNO 3 structure [83]. Regarding the sample with GA, no significant differences within the sample crosslinked with GA were observed (Figure 4). A comparison among the samples with different amounts of GA showed that the intensities of the amid I and II bands (around 1650 cm −1 and 1550 cm −1 ) increased with the increase in the GA ratio. This situation could be confirmed using the TGA results (Figure 3).

SEM Analysis of the Drug-Loaded HAp-GEL Composites.
SEM images of the drug-loaded HAp-GEL composites are illustrated in Figures 5(a) and 5(b). For each composite, drug-loaded spherical particles were observed. However, similar to the morphology of pure gelatin, adjacent spherical particles were also observed. Agglomerated irregular shapes can be caused by impurities [54,84,85]. In general, a regular particle size distribution was not observed.
SEM images of the unloaded HAp-GEL composites are shown in Figures 5(c) and 5(d). In Figure 5 3.5. Drug Release Profiles of the HAp-GEL Composites. The drug release profiles of the HAp-GEL composites with drug loaded by spray drying were evaluated at three different temperatures (32°C, 37°C, and 42°C) according to three different kinetic models. The release profiles of the HAp-GEL 2% GA and HAp-GEL 5% GA are illustrated in Figure 6. A biphasic release profile is observed in all composites. The drug release data for the 2% GA and 5% GA composites (HAp-GEL-5-FU) at 32°C shows that within the first 5 min, 40% and 35% of 5-FU are released, respectively. For the 2% GA and 5% GA composites, 55% and 50% of 5-FU were released at 37°C and 75% and 60% at 42°C within 5 min, respectively. This rapid initial release rate has been attributed to either the presence of nonencapsulated drug molecules on the surface of the microparticles or drug molecules that are close to the surface (immersed in the polymer matrix) [70]. For the 2% GA and 5% GA composites, the drug was almost completely released from the composites within 60 and 180 minutes, respectively. For HAp-GEL composites, the initial burst increased with increasing temperature. Shirakura et al. studied the drug release behavior of cisplatin in hydrogel nanoparticles with changing temperature, and the highest amount of drug release was observed at 42°C. This result was attributed to the fact that the matrix density decreased because of the swelling of the nanoparticles at higher temperatures, thus facilitating the escape of cisplatin. In addition, the diffusion rate of cisplatin might have been increased with increasing temperature [64].
Similarly, the same results were observed in other temperature-controlled studies. Thus, increasing temperature can cause the escape of 5-FU molecules from HAp-GEL composites, and the diffusion rate of 5-FU may be affected by temperature [65][66][67]. In this study, during the entire release process, the release rate and cumulative release percentage were affected by the temperature, while the diffusion rate of 5-FU was affected by the molecular thermal motion [70,87]. As seen in Figure 6, all HAp-GEL composites showed an initial burst and a 50% drug release in 20-25 minutes. In physiological environments, 5-FU has a halflife of approximately 8-20 minutes [88]. Figure 7 shows the drug release profiles of HAp-GEL composites with different amounts of GA at three different temperatures (32°C, 37°C, and 42°C). At each temperature,   [70,87,89,90]. Because HAp-GEL with a 1 : 1 weight ratio had the highest drug loading efficiency [91], drug release studies were conducted on that sample for different GA ratios. A burst release was observed in all samples during the first hour. For the HAp-GEL composites, the initial burst decreased with increasing GA content at all three temperatures. In other words, the drug release from HAp-GEL-5-FU 5% GA was slower than that from the HAp-GEL-5-FU 2% GA. The diffusion rate and quantity of drug release from the polymer matrix were influenced by the cross-linker ratio [92]. Similarly, Kim et al. and Peng et al. reported that increasing the GA content decreased the initial burst in a GEL microsphere and GEL-HAp nanocomposite, respectively [39,46]. Burst releases may be related to some drugs being weakly bonded to the surface of the composites [93]. According to the work of Santos, the release of 5-FU from spray-dried HAp was accomplished in the first 5 min [54]. Based on the literature, the 5-FU release from GEL is slower and/or less than the drug release from HAp [46,94] Moreover, the addition of GA reduces the drug release because of the increased rigidity in polymer chains [51,93]. Thus, it may be concluded that the addition of GEL and GA to HAp in the presence of an SBF may slow down the burst release. Overall, the release of the anticancer drug occurred in a slow manner, and these HAp-polymer composites could be used for controlled and sustained drug release [92].
Mathematical models of drug release could provide useful information about the mass transfer responsible for drug delivery system. They also reveal the influence of important parameters, such as morphology and loading of encapsulated materials on the release rate [87]. To better understand the release mechanism of 5-FU encapsulated in HAp-GEL, three different fitting equations, including zero order, first order, and Higuchi models, were used to fit the release curve. The kinetic mathematical model and the fitting results for the drug release at different temperatures are shown in Table 2. It is found that the release profile of 5-FU from the drugloaded HAP-GEL polymer composites is best fitted by the Higuchi model with the highest value of regression   International Journal of Polymer Science coefficients (R 2 ). Nevertheless, for HAp-GEL 2% GA at 37°C, the regression coefficient (R 2 = 0:955) is close to 1. This indicates that diffusion is dominated by Fickian diffusion mechanism [82,[95][96][97]. This is commonly observed in calcium phosphates (CaPs) and CaP cements. Moreover, this type of diffusion of 5-FU was also observed in other studies related to gelatin and chitosan polymers [51,82]. For HAp-GEL 2% GA at 32°C and 42°C and HAp-GEL 5% GA at 32°C, 37°C, and 42°C, the regression coefficients (R 2 was 0.817, 0.891, 0.934, 0.747, and 0.905, respectively) were lower than 1. For this reason, the release of 5-FU from the HAp-GEL polymer composites occurred through non-Fickian diffusion mechanisms in most cases. All sample releases are in accordance with the Higuchi model, which is diffusion-controlled [68][69][70][71][72]92].

Conclusion
The current study focused on the preparation of 5-FU-loaded HAp-GEL polymer composites and their in vitro drug release in a PBS medium at different temperatures. The HAp-GEL composites were produced using wet chemical precipitation, and the encapsulation (drug loading) process of 5-FU was conducted in a spray dryer. The obtained composites were characterized by XRD, TGA, FTIR, and SEM.
According to XRD and FTIR analysis, formation of hydroxyapatite crystals was confirmed. To observe the effect of temperature on the drug release mechanism, in vitro drug release studies were conducted at 32°C, 37°C, and 42°C and pH 4.0 in a PBS medium. For the HAp-GEL composites, the initial burst increased with increasing temperature. Two 8 International Journal of Polymer Science different cross-linking solutions of 5% and 2% GA-deionized water (v/v) were used during the preparation of the microparticles to understand the effect of the GA amount. For the HAp-GEL composites, the initial burst decreased with increasing GA content at all three temperatures. These results are in accordance with those of previous studies. The drug release profiles were evaluated according to three different kinetic models including zero order, first order, and Higuchi. In the kinetic model studies, the highest correlation values (R 2 ) were observed for the Higuchi model equation of diffusion, demonstrating that the drug release was proportional to the square root of time and occurred through diffusion. The HAp-GEL polymer composites occurred through non-Fickian diffusion mechanisms, with only one exception. Among the three kinetic studies investigated, it was determined that all composites best fit the Higuchi model. These polymer composites will be biocompatible and biodegradable, thus could be applied to controlled and sustained drug delivery systems.

Data Availability
All raw data used to support the findings of this study are available from the corresponding author upon request.