Release Behavior of Folic Acid Grafted Hollow Hydroxyapatite as Drug Carrier

Based on the formation of carbodiimide compounds between carboxyl and primary amines, hollow microspheres arising from the folic acid (folate-FA) grafted onto the surface of the modified hydroxyapatite were successfully prepared. The hollow morphology and composition of the FA-grafted hydroxyapatitemicrospheres were confirmed by scanning electronmicroscopy (SEM) and other characterizations. Brunauer-Emmett-Teller (BET) assay revealed the specific surface area and average pore size of the microspheres were 34.58m/g and 17.80 nm, respectively. As a drug carrier, the kinetic investigation of doxorubicin (DOX) loaded shows that the adsorbed behavior of drug on the adsorbent is more suitable to be described with pseudo-first-order model. Furthermore, the release rate can reach 83% at pH 5.7, which is greater than the release of 39% at pH 7.4, indicating an excellent performance of controlled drug release for response pH.The release mechanism of DOX coincides with Fickian diffusion as a result of KorsmeyerPeppas model analysis and the release phenomena can be well explained by Fickian diffusion second law.


Introduction
In recent years, hydroxyapatite has received great attention in tissue engineering and biomedical areas due to its similarity to the mineral component of bone tissues [1][2][3], excellent biocompatibility [4][5][6], and bioactivity as well as osteoconductive feature without causing local or systemic toxicity and inflammation [7,8].As for the applications of biological medicine fields, the hydroxyapatite has been extensively investigated as control-released carriers of drug, such as antibiotics [9][10][11], 5-flurouracil [12], and protein drugs [13,14].
As an ideal carrier of drug, good loading and controlreleased ability as well as the active targeting are always expected by people.Especially, as the first-line drug for a wide range of cancers, doxorubicin hydrochloride (DOX), an effective antibiotic, is usually applied to killing cancer cells or inhibiting their proliferation by means of chemotherapy way [15,16].However, long-term or repeated therapeutic action can create nonselective cytotoxicity because of the lack of targeted ability for cancer cells and then leads to severe side effects.In order to overcome these issues and improve therapeutic efficacy, many researchers have focused on such a kind of explorations to modify hydroxyapatite particles with specific ligands as a drug control-released carriers.It is well known that folic acid (folate-FA) is one of the receptor mediated targeting moieties [17][18][19] and its receptor shows an overexpression on the surfaces of cancer cells, such as breast, brain, kidney, and ovarian.Hence, the drug carriers covalently conjugated with FA could achieve selective targeting to cancer cells [20].For the various drugs, the pathway to prepare drug carriers with the FA-modified method can be different, and the performance of the carriers obtained can also vary.Regarding the prepared process, the hydroxyapatite can be either functionalized by polyethylene glycol (PEG) in advance or directly functionalized with folic acid [21] or aminated folic acid [22], and the relevant assessments showed that these modified microparticles could be a great candidate for biomedical application.Of course, the microparticles mentioned above do not possess the hollow morphology that has been noticed.
Although the hydroxyapatites with nanoscale have attracted more attention of researchers owing to the highly specific surface area and large capacity for drug loading.But, it may be expected that the porous hollow hydroxyapatite microparticles possess greater potential as one of the Advances in Polymer Technology promising drug carriers based on the advantages of nanosized hydroxyapatite.Therefore, a series of hydroxyapatite microparticles with porous structure have been prepared using a great diversity of methods or strategies, for instance, previously reported emulsion preparation [23], sol-gel synthesis [24], frequently used hydrothermal fabrication [25][26][27], and template method [28], as well as microwave-assistant method [29,30] and others [31][32][33][34].The relevant studies and characterizations have also illustrated a better performance of this kind of microparticles as drug carriers [35].In general, for the preparation of hollow hydroxyapatite by hydrothermal approach, the vaterite CaCO 3 is often selected as the sacrificial template.Under the alkaline condition, the hydrated Ca 2+ cations on the surface of the hollow CaCO 3 templates can react with PO 4 3− added into the solution and then create a replacement of CO 3 2− by PO 4 3− .As the reaction proceeds, finally, the hydroxyapatite was formed and possessed the shape and hollow structure of template microparticles.Certainly, the white Ca(OH) 2 emulsion can be also used as a sacrificial templates, and the shape of the as-prepared hollow hydroxyapatites retains the polyhedral structure of Ca(OH) 2 solid [36].
In this work, in order to explore the drug loading and release performance of FA-grafted hollow hydroxyapatite, the hollow hydroxyapatite microspheres were firstly fabricated using simple hydrothermal approach [37] and then modified the as-prepared microspheres with (3aminopropyl)triethoxysilane (APTES) to achieve the amination for the hydroxyapatite and finally conjugate primary amines to carboxyl groups in the activated folic acid by forming carbodiimide compounds to realize the folate grafting of the modified hydroxyapatite using the related research as the basic principle [38][39][40][41][42]. SEM, BET, and relevant measurements characterized the morphology, porous sizes, and composition of the FA-grafted hydroxyapatite particles.In addition, as a drug carrier, its adsorption behavior of drug and in vitro release kinetics were investigated.Meanwhile, the adsorption kinetic equation of loaded drug and the mechanism of release drug were evaluated.
. .Preparation of FA-Gra ed Hollow Hydroxyapatite Microspheres.First, the hollow vaterite CaCO 3 microspheres were prepared similar to the previous report with proper modifications [43,44], described in detail as follows: 5mL of PEI (1%) and SDS (0.015mol⋅L −1 ) were separately poured into 50 mL of CaCl 2 solution (0.1 mol⋅L −1 ) under vigorous magnetic stirring, after 10 minutes, 50mL of Na 2 CO 3 solution (0.1 mol⋅L −1 ) was added dropwise into the above mixed solution, the mixture was heated to 40 ∘ C under continuous agitation and kept for 2 h, the white suspension obtained was centrifuged for 5 min at 5000rpm, and the sample was washed with distilled water and then placed in the oven and dried at 80 ∘ C for 24 h.Next, the as-obtained CaCO 3 0.1g was dispersed in 50mL distilled water and raised temperature to 60 ∘ C after ultrasound for 10min, and then 50mL of Na 2 HPO 4 solution (0.03mol⋅L) was added with the rate of 1.5 mL⋅min −1 under stirring at 400rmp.After that, the pH of the reaction system was adjusted to 11 using NaOH solution and reacted persistently for 2 h.The suspension centrifuged was washed and dried at 65 ∘ C for 24h.
The fabricated 0.2 g of hydroxyapatite was dispersed in 100 mL of the solution corresponding to a ratio of absolute ethanol to distilled water 1:1 (volume ratio), the suspension was agitated for 30 min with a magnetic stirrer at room temperature, and then 0.18 g of APTES was added dropwise.After the continuous reaction for 24 h at room temperature, an aminated hydroxyapatite was obtained.The centrifugate was washed adequately enough with absolute ethanol, dried under vacuum, and added to the activated folic acid solution, which was prepared that 0.2g FA was dissolved in DMF/DMSO (28 mL 3: 1) solution contained 0.062g DCC and 0.032g NHS with stirring for 4 h.The generated mixture was reacted with N 2 protection in the dark for 24 hat room temperature.The centrifuged product was washed with water and absolute ethanol and freeze dried, FA-grafted porous hollow hydroxyapatite microspheres were obtained.
. .Loading and In Vitro Release of Drug.The adsorption test of DOX on the as-fabricated hydroxyapatite and FA-grafted hydroxyapatite was carried out as follows: 5mg hydroxyapatite and FA-grafted hydroxyapatite were separately added into 50mL DOX solution with different initial concentration.The adsorption has taken place in an orbital shaker at 37 ∘ C. At designated time intervals, 3ml of the suspension was separated by centrifuging at 12000rpm for 20min, and the DOX concentration was determined from the supernatant by UV spectrophotometer at a wavelength of 480nm, and the sediment was returned to the adsorption system.The amount of adsorbed DOX (mg/g) was calculated according to where G(g) is the adsorbent material quantity and c 0 (mg/L) and  e (mg/L) are the initial and equilibrium concentrations of DOX in the solution, respectively.V(L) is the solution volume.The influence of adsorption time was monitored for each material.The microspheres of maximum adsorption of drug was dried under vacuum at ambient temperature and stored until use.
As the in vitro drug releases, the drug-loaded microspheres, 10 mg, were placed in dialysis tube (MWCO 12-14kDa) and immersed in 25 mL of phosphate buffer solutions (PBS, pH=5.7 or 7.4) and shaken with the constant 200rpm at 37 ∘ C. The released solution (2 mL) was removed at fixed time intervals and analyzed the DOX concentration with a UV spectrophotometer at a wavelength of 233 nm, and then the same volume of fresh PBS medium was added to the release system.The released percentage of drug was calculated and the kinetics of release for the drug-loaded FAgrafted hydroxyapatite were investigated in the PBS solution with pH 5.7 and 7.4.
. .Characterization.The morphology of the FA-grafted hydroxyapatite was observed by Scanning electron microscopy (SEM, Hitachi S-3400N) at an accelerating voltage of 5-10 kV.Powder X-ray diffraction patterns (X/Pert Pro 3040/60) were used to analyze the structural phase composition with Cu Kradiation at 40kV and 40mA.Meanwhile, nitrogen adsorption-desorption isotherms were obtained with an surface area and porosity analyzer (ASAP 2020HD88, Micromeritics), after the sample was degassed under vacuum at 80 ∘ C for 12 h.Fourier transform infrared (FTIR) spectra were carried out on a Thermo Nicolet 6700 spectrometer with the KBr method.UV-vis absorption measurements were carried out with a Shimadzu UV-360 spectrophotometer.Thermogravimetric analysis (TGA) was conducted in a TA instrument SDT Q600 with a heating rate of 10 ∘ C⋅min −1 under N 2 atmosphere.The diameter size distribution of the samples was determined using a Zetasizer Nano-ZS90 dynamic light scattering instrument.Raman of the sample was measured using the Raman microspectrometer (DXR ThermoFisher) with an excitation wavelength of 532 nm.

Results and Discussion
. .Morphological and Composition Characteristics.The SEM images of the as-prepared hydroxyapatite and FA-grafted hydroxyapatite are shown in Figure 1.A uniform spherical morphology can be clearly observed, and some broken shapes with hollow interior indicate that the as-prepared microspheres have a porous hollow structure, and the average diameters are around 2.0 m.From the dynamic light scattering measurement, the mean size of the hydroxyapatite is 2034 nm with PDI (polydispersity index) value of 0.173 and the FA-grafted hydroxyapatite is 2063 nm with PDI of 0.157, respectively.These results are well consistent with the SEM observation, and the porous structure of FA-grafted hydroxyapatite has not been significantly altered (seen in Figure 1(b)).Furthermore, the local magnification image shows that the short rod-shaped microparticles of about 400 nm compose the hollow shape, which is in accord with related research [45,46].
The nitrogen adsorption/desorption isotherms and the corresponding Barrett-Joyner-Halenda (BJH) pore size distribution is displayed in Figure 2. Clearly, the curve is classified as a type IV isotherm with a hysteresis loop, and the specific surface area of Brunauer-Emmett-Teller (BET) is 34.58m 2 /g, the cumulative pore volume is 0.17cm 3 /g, and the average pore size is 17.80 nm.By contrast, the hydroxyapatite is assigned severally to 35.07m 2 /g, 0.18cm 3 /g and 18.89nm.The BET data of the FA-grafted hydroxyapatite do not create a significant difference, also illustrating that the FA grafting did not alter the porous hollow structure of the as-prepared sample.The porous and hollow structures of the FA-grafted hydroxyapatite are favorable for drug loading and release because they provide the ideal physical space.Figure 3 shows the FTIR spectrum of the FA-grafted hydroxyapatite, in comparison with FA and aminated hydroxyapatite.In Figure 3(c), the characteristic adsorption peaks at 564 cm −1 and 606 cm −1 are attributed to the triply degenerated vibration of P-O-P bond in PO 3− 4 groups, and the intense absorption bands at 1095 cm −1 and 1036 cm −1 are assigned to the antisymmetric stretching vibration of P-O bond, while the weak peak at 960 cm −1 belongs to symmetric stretching of P-O bond.The obviously broad band at 3442 cm −1 is related to the overlap for the symmetric stretching of OH (around 3570 cm −1 ) and adsorbed water.In addition, the adsorption peaks at 2927 cm −1 and 2830cm −1 are ascribed to the symmetric and asymmetric stretching of CH 2 groups, and the CH 2 deformation vibration is also observed at 1425 cm −1 .As the bands observed at 2928, 2834, and 1430 cm −1 as well as 969 cm −1 in Figure 3(b), in comparison of the corresponding peaks in Figure 3(c), these peaks are shifted to the lower wave numbers, indicating that a certain influence was created due to the FA introduced.It is worth noting that the intense band located at 1650 cm −1 of NH 2 in the aminated sample has not been found in Figure 3(c).Instead, a characteristic bending vibration (CNH) of amide II, arising from the amidation between COOH and NH 2 , is observed at 1586 cm −1 .Furthermore, the band at 645 cm −1 results from the bending vibration of O=C-N, and the band at 715 cm −1 is ascribed to the out-of-plane bending vibration of N-H.As Raman spectrum of the FA-grafted hydroxyapatite (see Figure S1 in the Supplementary Material for comprehensive image analysis), the special signals correspond to the symmetric bending (v 2 ) and stretching (v 1 ) vibration as well as the antisymmetric bending (V 4 ) and stretching (v 3 ) modes of the PO 3− 4 coincident with the reported work [47,48].The stretching vibration of C=O (amide II) especially appears at 1616 cm −1 , and the bands at 1582, 664 and 712 cm −1 belong to the bending mode (CNH) of amide II, the bending mode of O=C-N, and the out-of-plane bending mode of N-H, respectively.These results are consistent with the analysis of IR and also illustrate to form hydroxyapatite and achieve the grafting of FA onto hydroxyapatite.X-ray diffraction patterns of the aminated and FA-grafted hydroxyapatite is presented in Figure 4.In contrast to the hexagonal hydroxyapatite (JCPDS card No. 09-0432), all the featured planes can be well indexed, and no other planes of crystalline phase were detected, hinting that the as-prepared samples have a better crystallinity.
Thermal stability analysis of the as-prepared samples was carried out, as shown in Figure 5.The weight losses of 3.19% and 3.96% below 128 ∘ C were attributed to evaporation of residual water in the two samples, respectively.The aminated sample created the weight loss of 5.29% at 198-389 ∘ C, while in the curve of the FA-grafted hydroxyapatite the sharp weight loss of 28.02% at 245-490 ∘ C was due to the decomposition of carbodiimide compounds formed on the surface of the

FA-grafted hydroxyapatite Hydroxyapatite Pseudo-first-order
Pseudo-second-order Pseudo-first-order Pseudo-second-order  as-prepared sample, so implying that the grafted results of folic acid resulted in around 22.73% mass loss.
. .Adsorption Kinetics.Adsorption capacity of drug on the FA-grafted hydroxyapatite as a function of time is shown in Figure 6(a).One can find that the amount of adsorbed drug increases with the increase of initial DOX concentration, and an adsorption plateau was observed after the adsorption for 12h.As compared with the hydroxyapatite, in Figure 6(b), at the same drug concentration and time, the adsorption capacity of the former was 2-3 times as much as the latter, revealing that the adsorption rate was the same multiple of the latter.In other words, the adsorption performance of the FA-grafted hydroxyapatite is far better than that of the hydroxyapatite.According to the specific surface measuring of the FA-grafted hydroxyapatite and the hydroxyapatite, their values have no prominent difference; hence, the specific surface is not a main factor for leading to the apparent difference in adsorption capacity.For the as-prepared FA-grafted hydroxyapatite with porous hollow structure, a lot of carboxyl groups exist on the surface of the porous microspheres, and these groups may create strong interaction with the amino in the DOX molecule.However, as the unfunctionalized hydroxyapatite, the OH groups in porous hollow hydroxyapatite can only provide relatively low affinity towards the drug molecules.Therefore, the FA-grafted hydroxyapatite shows a higher adsorption capacity and rate at the same concentration.As for the jumps of adsorption capacity for both the samples between 500 and 1000 mints, the more possible reason arises from the physical adsorption at later stage of adsorption.
To confirm the kinetic equation of adsorption process of drug, the classical pseudo-first-order and pseudo-secondorder kinetic model are used, and the model equations are as follows: where c (mg/L) and c 0 (mg/L) are assigned severally to the solution concentration of drug at time t and the initial concentration, and k 1 (min −1 ) and k 2 (L⋅mg −1 ⋅min −1 ) are the rate constant of corresponding model equation.The calculated results are listed in Table 1.As compared to k 2 , one can see that k 1 of the FA-grafted hydroxyapatite is close to not only a constant but also the hydroxyapatite.In the meantime, the values of correlation coefficient associated with the pseudofirst-order are better than that of the pseudo-second-order except for the individual R 2 of the hydroxyapatite; it means that the adsorbed behavior of drug on the adsorbents is more suitable to be described with the pseudo-first-order model.  . .Drug Release.The releases of the DOX-loaded FAgrafted hydroxyapatite in PBS solution are plotted in Figure 7.It can be found that both of the DOX releases at pH value of 5.7 and 7.4 at 37 ∘ C exhibited a similar trend, and the release plateaus were all reached after 28 hours, indicating that the drug-loaded FA-grafted hydroxyapatite showed a slow and sustained release behavior.Meanwhile, nearly 83% of the loaded DOX was released in the PBS solution with pH 5.7.This amount of release was higher than that of the approximately 39% in the PBS solution with pH 7.4, implying that the DOX-loaded FA-grafted hydroxyapatite had a pHresponsive characteristic for controlling release.This pHresponsive release was mainly attributed to the higher solubility of the DOX-loaded FA-grafted hydroxyapatite in PBS solution with a low pH value than that in PBS with a high pH value.Certainly, the interaction decreasing between DOX molecular and the carriers due to the protonation of the amino and carboxyl groups in the acidic PBS solution may play partly a role.It is worth noting that, in the PBS solution with pH 5.7, the DOX-loaded FA-grafted hydroxyapatite underwent a rapid release in the first 6 h, while in the PBS solution with pH 7.4 the rapid release was about in the first 4 h; after that, a slow release was all displayed the range of next time to 48 h.It revealed that the as-prepared FA-grafted hydroxyapatite has excellent controlled-release property of drug.In the initial stage of the DOX release, the DOX molecules adsorbed close to the surface diffused easily into the bulk solution, and then the DOX molecules, which was bound to the FA-grafted hydroxyapatite matrix and located below the surface, performed to diffuse through the FA-grafted hydroxyapatite matrix and pores in the matrix, so in this case a slow and sustain-controlled release was observed.
The drug, DOX, is loaded into the porous hollow hydroxyapatite through the adsorption pathway, and the drug release may be regarded as a liking release behavior of matrix system.Hence, Korsmeyer-Peppas mathematical model can be used to analyze the mechanism of the drug release.The model equation is [49] where  t / ∞ is the fraction of drug release at time t, and a is the release rate constant, and n is the release exponent.
The experimental data obtained from in vitro drug release are linearly fitted in Figure 8.In terms of the calculated results, the correlation coefficients, R 2 , are 0.9654 and 0.9251 in the PBS solution with pH 5.7 and 7.4, respectively, and the n values are correspondingly 0.24 and 0.28.The high R 2 value hints a good correlation degree which is well consistent with Korsmeyer-Peppas model.The value of n is all less than 0.5, indicating that the process of drug release belongs to Fickian diffusion mechanism.
As for Fickian diffusion second law, the overall process of release in a certain thickness device can be divided into early and late stage, and in the early stage the release amount is proportional to the square root of time, whereas in the late stage the release amount is an exponential function of time due to the decrease of drug concentration in the device.Therefore, when the release rate data in the initial stage are taken the natural logarithm and plotted against lnt and the slope of the fitted straight line should be 0.5.Thereupon the related release data were analyzed, as shown in Figure 9; the high values of R 2 are all obtained, and at pH 7.4 the slope is 0.4609; it is nearly 0.5; that is, the release behavior accords with Fickian second law.Meanwhile, the slope is only 0.2547 at pH 5.7; there is a larger difference value than 0.5, implying that the release did not well conform to Fickian second law.This might be attributed to the dissolution of the loaded drug in the low pH condition, which led to a little anomalous behavior of drug diffusion in the first 6 hours.

Conclusions
SEM and the relevant characterizations have demonstrated that the FA-grafted hydroxyapatite microspheres with porous hollow structure were successfully prepared, and FA grafted onto the surface of modified hydroxyapatite did not alter the porous hollow structure of the as-prepared hydroxyapatite.For the loading of drug DOX, the FA-grafted hydroxyapatite showed a higher loading capacity compared to the hydroxyapatite due to the strong interaction between the carboxyl group on the surface of FA-grafted microspheres and the amino in the DOX molecule, and the adsorption kinetics of drug for the two microspheres are well consistent with pseudo-first-order model.The drug release of the DOXloaded FA-grafted microspheres in the solution of pH 5.7 and pH 7.4 showed a rapid release in the initial stage and then sustained release in the next time, revealing that an excellent property of pH-controlled release, and the release behavior could be well described with Korsmeyer-Peppas mathematical model and fitted in with Fickian diffusion second law.

Figure 6 :
Figure 6: Adsorption capacity of the samples at the various drug concentrations as a function of time.(a) FA-grafted hydroxyapatite; (b) hydroxyapatite.

Figure 7 :
Figure 7: The release curves of DOX-loaded FA-grafted hydroxyapatite in PBS at pH 5.7 and 7.4.

Figure 8 :
Figure 8: Natural logarithm of the release fraction of drug as a function of ln .

Figure 9 :
Figure 9: Natural logarithm of the drug release rate as a function of lnt in the early stage of drug release.

Table 1 :
The fitted results of k 1 and k 2 and correlation coefficients for two models.