Alginate hydrogels functionalized with β -cyclodextrin as a local paclitaxel delivery system

Modification of drug delivery materials with beta-cyclodextrins ( β -CyD) is known to increase solubility of poorly water-soluble drugs, protect drugs from degradation and sustain release. In this study, we developed a hydrogel drug delivery system for local paclitaxel delivery using the natural polysaccharide alginate functionalized with β -CyD-moieties. Paclitaxel was chosen due to its ability to form inclusion complexes with cyclodextrins. The rheological and mechanical properties of the prepared hydrogels were characterized, as well as in vitro release of the paclitaxel and in vitro activity on PC-3 prostate cancer cells. Introduction of β -CyD-moieties into the hydrogel reduces the mechanical properties of the gels compared to nonmodified gels. However, gelation kinetics were not markedly different. Furthermore, the β -CyD-modified alginate helped to reduce undesired crystallization of the paclitaxel in the gel and facilitated paclitaxel diffusion out of the gel network. Remarkably, the β -CyD grafted alginate showed increased capacity to complex paclitaxel compared to free HP β CyD. Release of both paclitaxel and degradation products were measured from the gels and were shown to have cytotoxic effects on the PC-3 cells. The results indicate that functionalized alginate with β -CyDs has potential as a material for drug delivery systems.

physiological conditions, and the sol/gel transition is gentle enough that cells and biomolecules will be functional after entrapment in the hydrogel. 5 Furthermore, alginate can be tailored to present different functional groups [6][7][8][9][10] and the mechanical properties can be adjusted. [11][12][13] In addition, the polymer is biocompatible and nontoxic. 14 Different approaches for using alginate as an injectable delivery system are currently being investigated, often in combination with other polymers or components in the quest to achieve optimal delivery capability. [15][16][17][18][19][20] The hydrophilic nature of the hydrogel presents certain challenges for drug delivery. 21,22 Small water-soluble molecules will often quickly diffuse out of the hydrogel matrix, giving a high burst release.
On the other hand, hydrophobic molecules can be difficult to load into the hydrogel and their release in a controlled manner also possesses a challenge. To gain better control over the loading and release of therapeutic agents from a hydrogel, cyclodextrins (CyDs) can be incorporated into the gel. 23,24 CyDs are well known cyclic structures built from glucopyranose units produced by enzymatic treatment of starch. 25 Due to the cyclic structure of the formed CyD molecules, the cavity of the ring structure is hydrophobic while the exterior is hydrophilic. The cavity can then give a local environment in the hydrogel that is better suited for entrapment of hydrophobic molecules and moieties. 24,26,27 Paclitaxel is a highly hydrophobic anticancer drug that can form inclusion complexes with CyDs, including β-CyD. 28,29 It should be noted however that paclitaxel has a limited penetration into the hydrophobic cavity of β-CyD and the obtained inclusion complex is characterized by a low binding constant. 29,30 Given intravenously, paclitaxel is not specifically targeted to the tumor, and thus only low therapeutic levels of the drug can be achieved at the tumor site. 31,32 Moreover, paclitaxel has been shown to have poor penetration across the blood-brain-barrier in preclinical models. 31,33 Hence, paclitaxel use in treatment of brain tumors and brain metastases is limited, although it was shown to be active against various tumor types. 33 Interestingly, it has been shown recently that in addition to the antiproliferative activity, paclitaxel has a capacity to promote antitumor immunity. 34 This makes paclitaxel a very promising candidate for drug delivery systems. To enhance the therapeutic potential of paclitaxel several delivery strategies and formulations of paclitaxel have been developed, 35,36 including only a few injectable hydrogel systems based on alginate. 37,38 CyDs have previously been grafted to alginate by attachment to the carboxyl-or OH-group, [39][40][41][42][43][44][45][46] often for drug delivery applications.
Examples include photoresponsive CyD-grafted alginate hydrogels where UV-light can be used to control the inclusion complex between CyDs and azobenzene moieties, giving control over gel morphology and subsequent release from the gel, 39 release of the antiemetic drug ondansetron by mechanical stimulation of a cyclodextrin-crosslinked alginate gel, 41 and release of the poorly soluble anti-inflammatory agent hydrocortisone acetate from β-CyD grafted alginate hydrogels. 43 We have recently shown that functionalization of alginate hydrogels with β-CyD moieties allows to slow down the release rate of a small model compound methyl orange. 9 Based on this finding, we hypothesized that with the β-CyD-grafted alginate, it might be possible to modulate the release of paclitaxel. While numerous studies have investigated the entrapment of paclitaxel in hydrogel systems, to the best of our knowledge, this study for the first time describes a paclitaxel delivery system based on alginate hydrogels functionalized with β-CyDs. In this study, we focused on, first, the development of an injectable hydrogel system and characterized it with respect to rheological and mechanical properties. Second, in vitro release of the paclitaxel and in vitro cytotoxicity toward PC-3 prostate cancer cells was studied.

| Materials
Alginate (Alg) from Laminaria hyperborea stipe (F G = 0.7, N G > 1 = 11, M w = 126 kDa) was obtained from FMC Health and Nutrition, Sandvika, Norway. For cell studies, ultrapure stipe alginate from Novamatrix™, Norway (UPLVG, F G = 0.67, N G > 1 = 13, M w = 221 kDa) was used. 6-monoazido-6-monodeoxy-β-cyclodextrin was kindly provided from Aalborg University. 47 The oxidized and grafted alginate materials were made in-house from stipe alginate and ultrapure stipe alginate at NTNU as previously described. 9 The following materials were synthesized for this work: partially oxidized alginate  Cells were maintained in exponential growth phase until they were detached by trypsin-EDTA solution, counted, and seeded into experimental plates/flasks.
All the samples and controls were run in triplicates. were incubated with gels at 37 C for 72 hr and cell metabolic activity was assessed using resazurin assay. Drug-free medium was used as a negative control. All the samples and controls were run in triplicates.

| Cytotoxicity of paclitaxel-loaded hydrogels
Gels were examined by optical microscopy before and after experiment.

| Resazurin cell metabolic activity assay
Cells were incubated with 50% diluted AlamarBlue™ reagent at 37 C for 1-8 hr, as it is recommended by a supplier. Then, 100 μL aliquots were transferred into black 96-well plates with clear bottom and fluorescence was read with Spectra Max i3x (Molecular Devices) at 560 nm excitation and 600 nm emission wavelengths. Cell metabolic activity was expressed as percentage of fluorescent intensities relative to a control (cells not exposed to paclitaxel were defined as 100%) after subtraction of non-cell-derived background.

| Statistical analysis
Statistical analysis was performed using Microsoft Office Excel 365 or SigmaPlot Version 14.0 (Systat Software Inc., CA) and differences between groups were compared applying a two-tailed t-test. The significance level was set at 0.05. Differences between more than two groups were tested using one-way (or two-way when necessary) analysis of variance (ANOVA). After ANOVA, the Tukey's post hoc test or a two-tailed t-test was used for comparison. All results are expressed as means ± SD.

| RESULTS AND DISCUSSION
We have previously shown that CyDs grafted to alginate were able to interact with the model compound methyl orange, a small guest molecule that can form inclusion complexes with β-CyD. 9 This new study relies on the hypothesis that introduction of β-CyD moieties into alginate mediates the uptake of highly hydrophobic drugs like paclitaxel and their release from the alginate hydrogel. For this study the alginate was functionalized with β-CyD moieties in a three-step synthesis as previously described. 9 The partial oxidation of alginate performed in the first step is known to lead to noticeable structural changes, such as partial depolymerization and ring opening (cleavage of the C2-C3 bonds). 11,[51][52][53] Such changes in the alginate structure have been shown to affect the gelation capability of the polymer and mechanical properties of the hydrogels fabricated from this polymer. 11,12,40,54 Therefore, we conducted a series of experiments to characterize gelation kinetics and mechanical properties of the hydrogel system based on the modified alginate. The well-established system of alginate gelation, using slowly releasing calcium ions from CaCO 3 particles induced by the slow release of H + from GDL, was used to allow for slow gelation of alginate. 55 Using this technique, it is possible to make soft alginate gels with homogeneous and reproducible structure at physiological conditions, while the end products of the reaction, namely CO 2 and D-glucuronic acid, are nontoxic. 49 Importantly, this method allows to have control over gelation kinetics of alginate by varying the release rate of calcium ions from the calcium carbonate particles. The size of CaCO 3 particles have previously been shown to largely influence the gelation kinetics. 49 For an injectable hydrogel, the gelation time will influence how easy it is to inject the gel, as well as its propensity to gel at the injection site instead of leakage before the gel network can form. 56,57 The rate of gelation was followed by oscillation rheometry show- (1:1) gels (E = 3 kPa, stress at rupture = 0.23 kg), the stress needed to rupture the Alg/POA gels were lower compared to the grafted gels.
The opposite is observed for Ca-saturated gels 12 where Young's modulus decreased for Alg/POAβCyD gels (1:1,~15 kPa) compared to Alg/POA gels (1:1, 24 kPa), while the stress at rupture was approximately the same (~1 kg). This was also observed for mixed hydrogels grafted with other substituents than β-CyD. This indicates that for Ca-unsaturated gels Young's modulus for mixed Alg/POAβCyD gels is not largely affected by grafting, but for Ca-saturated gels the covalent coupling of β-CyD affects the network formation/rearrangement due to steric effects, and may hinder optimal interaction with Ca-ions for the crosslinking units of the alginate.

| Loading of paclitaxel into alginate gels
To load paclitaxel into alginate hydrogels, the drug was solubilized in ethanol and mixed with alginate solution before gelation was initiated.
All the loaded drugs were maintained within the gel, hence, the final concentration of paclitaxel in the gels was 25 μM. The strategy of internal gelation was chosen as this has been shown to give homogeneous distribution of polymer in the gels. 49 Similarly, it was assumed that homogenous distribution of paclitaxel within the gels would also be achieved in addition to facilitate formation of inclusion complexes between the drug and POAβCyD. Here, precipitation of paclitaxel was utilized as an indirect indicator of inclusion complex formation.
Crystallization of paclitaxel is known to occur in both aqueous solution and hydrogels 58 and was clearly observed in the alginate gels to POAβCyD. This is surprising, in particular for HPβ-CyD which was chosen due to its higher water solubility compared to β-CyD, as well as its more frequent use in pharmaceutics due to it being suitable for parenteral injection. 59 The phenomenon of the increase in complexation ability of β-CyD upon grafting has been also observed for β-CyDgrafted hyaluronic acid. 60 Authors reported paclitaxel:cyclodextrin molar ratios of 0.011, 0.00026, and 0.007 for the hyaluronic acid grafted β-CyD, β-CyD, and dimethyl-β-CyD, respectively.
The paclitaxel crystals have been reported to be stable in aqueous environment up to 2 months and are seen as a limitation for achieving a therapeutic effect from the drug. 58 Therefore, the presence of crystals is not desirable, especially for controlled drug release systems.
Here, paclitaxel crystals could be observed in the gels after 21 days after exposure to physiological buffer with same observations of the different gels as mentioned above (Figure 2).

| Release of paclitaxel from alginate gels
The ability of the developed hydrogel system to release paclitaxel was studied in vitro in the physiological buffer ( Figure 3). Paclitaxel concentration was measured based on a standard curve of paclitaxel dissolved in extraction solvent, however the presence of paclitaxel degradation products (7-epi-taxol, 10-deacetyltaxol, and baccatin III) in the release samples was obvious throughout the time of the experiment (21 days) and was measured as areas due to lack of standards. In products, no accurate quantification can be made due to the lack of standards, however, it can be seen that 7-epi-taxol largely follows the release of paclitaxel, whereas 10-deacetyltaxol and baccatin III show higher amounts at the early timepoints (day 1-6) relative to later (day [9][10][11][12][13][14][15][16][17][18][19][20][21]. Interestingly, it has been shown previously that the epimer 7-epi-taxol still shows anti-cancer properties. 61 At day 1, Alg and Alg + HPβ-CyD showed higher concentrations of paclitaxel in physiological buffer than other gels. For Alg samples, the observed increase cannot be explained by higher initial release from this sample, because the areas of the degradation products did not increase accordingly, and vice versa, they were not significantly different from those for Alg/POAβCyD (1:1) and Alg/POAβCyD There are several important factors to consider here: higher concentration of HPβ-CyD than that of β-CyD (7.2 vs. 0.9 mM, respectively) and higher water solubility of HPβ-CyD (>860 mM compared to 16 mM for β-CyD). 62 It was therefore expected to be easier for HPβ-CyD to transport paclitaxel out of the gel via diffusion compared to the β-CyD.
At the end of the release study, the remaining paclitaxel in the gels were measured (Figure 4). All samples contained unreleased paclitaxel, with more in the Alg gels compared to grafted gels. Looking at the average values, about 5% of the loaded drug was found in the Alg/POAβCyD (1:3) hydrogels, about 18% in the Alg/POAβCyD (1:1) hydrogel and about 24% in the Alg hydrogel. For the degradation products found in these gels, only baccatin III and 10-deacetyltaxol were found in the Alg gels whereas also 7-epi-taxol was found in the Alg/POAβCyD gels ( Figure A4, Supporting Information S1).
For the Alg gels mixed with free β-CyD or HPβ-CyD, the variation in the amount of remaining paclitaxel in the gels was very large, compared to the other gels. This may reflect differences in the release of free CyD from the gels and hence drug released or potential differences in initial crystallization ( Figure A2, Supporting Information S1).
On average, the paclitaxel remaining in the Alg gels with free β-CyD was higher (20%) than in the Alg gels with free HPβ-CyD (9%), which again indicates faster release from the gels in presence of free HPβ-CyD and agrees with the observations made in the release study. In these gels, baccatin III, 10-deacetyltaxol, and 7-epi-taxol were also found at day 21 ( Figure A4, Supporting Information S1).
Taken together, our results indicate that more paclitaxel is solubilized in the gels with grafted alginate (Figure 2), and more is released over time, relative to nongrafted alginate and alginate gels with free β-CyD (Figures 3 and 4). In total, 82-95% of the paclitaxel was released into physiological buffer and/or underwent degradation at day 21 for the grafted material ( Figure 4). Inclusion complexes between β-CyDs and paclitaxel are characterized by low association constants as paclitaxel molecules have a limited penetration into the hydrophobic cavity of cyclodextrin. 29,30 On the other hand, one can expect faster release rate of paclitaxel in presence of β-CyDs. Indeed, paclitaxel in vitro release from the different drug delivery systems functionalized with β-CyD moieties has been reported to be 30-100% during the first 24 hr. 60,63,64 Other factors that may influence the release of paclitaxel from the gels could be different mechanical properties of the gels with different concentration of the grafted material (see the Section 3.1). The gels in which POAβCyD is substituted with POA can be considered as a good control for mechanical properties of the gels. For the Alg/POA (1:3), which also showed high concentrations of both paclitaxel and degradation products in physiological buffer, concentration of the remaining drug was found to be~11%, whereas more drug (~26%) was found in the stiffer Alg/POA (1:1) gel ( Figure A5, Supporting Information S1). Stiffer gels are recognized by a higher crosslinking density, that could influence both degradation of the gels 12 as well as gel permeability. Although no direct degradation of the gels was measured in this study, differences in stability were observed as the grafted alginate samples were more difficult to handle than the nonmodified samples. Also, slight decrease in gels weight was found at day 21 ( Figure A6, Supporting Information S1). The permeability of the gels in this study is not known, although one may speculate that lower crosslinking density will lead to more permeable gels. However, for Ca-saturated alginate gels, the opposite is found, as alginates with a low content of guluronic acid were shown to be less permeable to, for example, albumin, than alginate with a higher content of guluronic acid and thus higher crosslinking density. 65 The alginate gel network as such is not expected to influence the release of paclitaxel, as Ca-saturated alginate hydrogels are open for diffusion for larger molecules such as antibodies (IgG, 150 kDa). 13 Hence, paclitaxel, as a nonpolar and small molecule (854 Da), is expected to readily diffuse through the alginate hydrogels. However, differences in release could also be caused by differences in the drug loading procedure (see the discussion above). The crystallization of paclitaxel must also be considered. For a system containing only paclitaxel dissolved in aqueous medium (no gel), only very little of the free drug (less than 10%) has F I G U R E 4 Paclitaxel remaining in the alginate hydrogels after being exposed to physiological buffer for 21 days. Values show amount of paclitaxel found in dissolved gels relative to the amount of the paclitaxel loaded into the hydrogels. Mean ± SD, n = 3 been shown to be released within 10 days in PBS pH 7.4, which is likely due to the low solubilization/crystallization of paclitaxel in water. 66 It should also be noted that the release study was not conducted under sink conditions. The release medium used for drug release studies should ideally obtain sink conditions, that is to say the volume of the medium should be three times higher than the volume required to solubilize the drug to ensure free diffusion out of the drug delivery system. 67,68 The solubility of paclitaxel in physiological buffer has not been tested in this study. But one could expect that paclitaxel solubility in physiological buffer is comparable to the solubility in PBS buffer, which is reported 67 to be within the range of 0.3-10 μg/mL (0.35-12 μM) at 37 C. If all the paclitaxel releases at once from the tested hydrogel, the concentration in physiological buffer will reach 10 μM, which corresponds to the reported solubility limit. However, the release media used in our study also contained traces of ethanol (below 1% (v/v) from the gels), which increases the solubility of paclitaxel.
As degradation of paclitaxel seemed to be a major issue in determining release from the gels, stability studies of paclitaxel were performed in the physiological buffer and water for comparison ( Figure 5). At day 1, about 90% of the paclitaxel was found in both water and physiological buffer. Thereafter, the degradation was especially prevalent in the physiological buffer where only 33%, 11%, and 2% of the initial concentration could be found after 3, 9, and 15 days, respectively. In water, about 80% of the paclitaxel remained in the solution after 3 days and to the end of the study. Hence, limited degradation was seen in water. The degradation products 10-deacetyltaxol and baccatin III 6169 were found in both solutes ( Figure A7, Supporting Information S1), whereas the epimer 7-epitaxol was found only in physiological buffer. This underlines the difficulties in determining release of paclitaxel in relevant fluids. Here, we identified three degradation products, but more degradation products of paclitaxel are reported, 69 hence illustrating the complexity in determining released paclitaxel. Since the degradation of paclitaxel was also observed in water, it was therefore decided to cast paclitaxel-loaded hydrogels at 4 C for further experiments.

| Cytotoxicity of paclitaxel-loaded hydrogels
The effect of the released drug was further studied in in vitro cell culture on the prostate cancer cell line PC-3. For the cell studies, the total concentration of alginate in the gels was increased from 1% (w/v) up to 2% (w/v) while keeping the ratio of Alg/POAβCyD 1:1 (w/w) and paclitaxel loading remained unchanged. This allowed for a drug:cyclodextrin molar ratio in the gels below 0.02 and thus to avoid crystallization of paclitaxel within the gels. Another benefit of increasing the alginate concentration was higher stiffness of the gels and their reduced vulnerability to syneresis in culture medium ( Figure A8, Supporting Information S1).
The potential toxic effect of the grafted polymer by itself was first studied, followed by examining the burst release effects. Thereafter, the cytotoxicity of the paclitaxel-loaded hydrogels on metabolic activity of PC-3 cells was studied. The reasoning behind studying cytotoxicity of the functionalized polymer itself was as follows: Firstly, although cyclodextrins are widely used as excipients in pharmaceutic applications, it has been shown previously that β-CyDs can cause dose-dependent hemolysis and cytotoxicity. 70 Secondly, formation of the Alg/POAβCyD gel is not instant, meaning that upon injection in vivo the surrounding tissue and cells can interact with nongelled polymers. And lastly, possible leakage of POAβCyD from the gels over time 12 can lead to unwanted cytotoxic effects. To address this question, PC-3 cells were exposed to nongelled POAβCyD polymer. Free β-CyD dissolved in POA (β-CyD + POA sample) was used as a control. The cells exposed to nongelled POAβCyD showed viability ≥90% for concentrations up to 1 mM of β-CyD, which equals 5 mg/mL of the polymer (Figure 6a). A decrease in the cells metabolic activity was seen above this concentration in a concentration-dependent manner.
The cytotoxic effect of POAβCyD was nearly two times higher than that one of free β-CyD dissolved in POA solution (β-CyD + POA sample). The half maximal inhibitory concentrations (IC50) were not obtained in this experiment, because high viscosity of the polymers (POAβCyD is highly viscous at concentrations >40 mg/mL) and low water solubility of free β-CyD (water solubility is 16.3 mM 62 ) limited F I G U R E 5 Stability of paclitaxel over time in water (circle) and physiological buffer (triangle) at 37 C further increase of samples concentration. It has been previously reported that substituents can attenuate cytotoxicity of the native CyDs. 71,72 However, our results with POAβCyD show the opposite effect. We hypothesize that the observed effect might be connected to the difference in interaction between free βCyD and grafted βCyD moieties with the cellular cholesterol, since the cytotoxicity of various CyD derivatives is known to depend on their capacity to extract cholesterol from the cell membrane. 70 Next, we formed alginate gels loaded with paclitaxel and studied burst release of the drug in culture media and hydrogels cytotoxicity in vitro. To study the burst release of paclitaxel, Alg and Alg/POAβCyD (1:1) hydrogels were loaded with 25 μM of paclitaxel and incubated in the culture media for 1 hr. The concentration of the released drug was then determined by exposing PC-3 cells to serial dilutions of the collected media and subsequent conversion of the obtained cytotoxicity values into the drug concentration using a standard curve. The standard curve was prepared using free paclitaxel in media ( Figure A1, Supporting Information S1). The effects of the collected medium on metabolic activity of the PC-3 cells are shown on Figure 6b. The release-medium from both grafted and nongrafted alginates decreased cell metabolic activity in a concentration dependent manner. The paclitaxel burst release was found to be 6.2 ± 2.7% for Alg gel and 11.2 ± 2.5% for Alg/POAβCyD (1:1) gel. Under light microscope, the Alg/POAβCyD (1:1) mixed gels did not contain visible crystals of paclitaxel, whereas the Alg gels had clearly visible crystals.
Thus, it is very likely that the Alg gels released less drug during the first hour due to crystal formation of the drug within these gels.
Next, to demonstrate efficacy of the developed delivery system in vitro, free paclitaxel, empty gels and gels loaded with paclitaxel were added to PC-3 cells and cell metabolic activity was determined after 72 hr of co-incubation (Figure 6c-d). Two gel types, namely Alg gels and Alg/POAβCyD (1:1) gels, and three different loadings (0.9, 7.2, and 25 μM) as well as placebo (without drug) were compared. The gels of both grafted and nongrafted alginate without paclitaxel showed no effect F I G U R E 6 Metabolic activity of PC-3 cells. (a) Cells were incubated with alginate functionalized with β-CyD (POAβCyD) solution and free β-CyD mixed with POA solution; (b) cells were exposed to serial dilutions of 1-hr release media for paclitaxel loaded gels (25 μM) for 72 hr; (c) Cells were exposed to the 50 μL gels loaded with 0, 0.9 and 7.2 μM paclitaxel for 72 hr (total volume of medium 4.5 mL); and (d) cells were exposed to the 50 μL gels loaded with 25 μM paclitaxel for 72 hr (total volume of medium 125 mL). Cell metabolic activity is expressed as fluorescent intensities relative to a control (cells not exposed to paclitaxel were defined as 100%) after subtraction of non-cell-derived background. Data are expressed as mean ± SD, n = 2-8. * and ** denote statistically significant differences (p < 0.05 and p < 0.01, respectively) between two groups on the metabolic activity of the cells (Figure 6c). The drug-loaded gels as expected decreased metabolic activity of PC-3 cells in dose-dependent manner relative to the loading concentration of paclitaxel. The Alg gel and Alg/POAβCyD (1:1) mixed gels loaded with 0.9 μM paclitaxel decreased cell viability down to 53.5 ± 4.8% and 60.4 ± 4%, respectively, whereas the gels loaded with 7.2 μM paclitaxel decreased cell viability down to 28.2 ± 1.9% and 24.3 ± 1.9%, respectively. Thus, the difference between two gel types (with and without POAβCyD) was not noticeable when 0.9 and 7.2 μM loadings were tested. These samples did not contain visible crystals of paclitaxel neither inside Alg nor Alg/POAβCyD (1:1) gels. In contrast, upon 25 μM paclitaxel loading, crystallization of the drug was visible inside the Alg gels, but not within the gels comprising grafted β-CyD. The mixed gels of Alg/POAβCyD (1:1) decreased cell metabolic activity to 51.1 ± 19%, whereas the Alg gels reduced it to 71.8 ± 14.3%; ( Figure 6d). The Alg gels still contained undissolved paclitaxel crystals after 72 hr of incubation with cells, whereas the grafted gels remained transparent. Although the difference between Alg and Alg/POAβCyD (1:1) gels was not statistically significant (p = 0.11 (n = 2) and p = 0.2 (n = 2) for two independent experiments), we assume that the observed slightly higher cytotoxic activity of Alg/POAβCyD (1:1) gels could be related to a slightly faster paclitaxel release from these gels as well as higher burst release found in culture media.
In this work, we evaluated the in vitro biological activity of the functionalized alginate hydrogels loaded with paclitaxel on prostate cancer cells. This cell model was chosen because paclitaxel is known to be efficacious against various prostate cancer cells, including PC-3 cells. 73 At present, however, paclitaxel is only approved for the treatment of breast cancer, cancer of the ovaries, nonsmall cell lung cancer, and Kaposi's sarcoma, and is being currently used off-label for the treatment of castration-resistant prostate cancer and some other malignant tumors. 74 In situ forming hydrogel systems can serve as a local depot slowly releasing paclitaxel and may be beneficial for the treatment of patients with prostate tumors. A recent preclinical study demonstrates therapeutic efficacy of local delivery system based on an injectable polymer paste with paclitaxel in LNCaP human prostate cancer xenografts. 75 In addition, the efficacy of the developed paclitaxel delivery system would be interesting to study in models of brain tumor and brain metastases. As mentioned earlier, paclitaxel has strong anticancer activity against glioma cells and many other cancers in vitro. 33 However, paclitaxel is a substrate for the P-gp/ABCB1, and therefore has limited access to the central nervous system. 33 Since local delivery systems have already shown antitumor efficacy against glioblastoma, 4 delivery of drugs to brain tumors is a relevant follow-up of the studied gel system. In addition, other poorly water-soluble chemotherapeutic agents that can form inclusion complexes with β-CyD, for instance cisplatin, doxorubicin, curcumin, camptothecin, and so on, may be considered as potential payloads for the developed depot hydrogel.

| CONCLUSIONS
In this study, we developed a hydrogel-based delivery system employing β-CyD functionalized alginate, where alginate acted as the gelling material and the β-CyD moieties were responsible for formation of inclusion complexes with a poorly soluble drug (paclitaxel). The introduction of β-CyD-moieties into the delivery system generally decreased the mechanical properties of the gels (compared to nonmodified alginate hydrogels) and modulated paclitaxel behavior but did not influence gelation kinetics. The significance of the present work is that it shows increased capacity of β-CyD grafted alginate to complex paclitaxel as compared to free HPβ-CyD. Although the paclitaxel is not an ideal payload for β-CyD-based systems because of its relatively poor retention within β-CyD hydrophobic cavity, and because it is difficult to quantify due to rapid degradation, our findings support that the β-CyD-grafted alginate can modulate release of the paclitaxel from the gels. The β-CyDgrafted alginate prevented crystallization of the paclitaxel by retaining a complexed dispersion of the drug and facilitated paclitaxel diffusion out of the gel network. This effect was observed at paclitaxel:β-CyD molar ratios not exceeding 0.02. Furthermore, the paclitaxel-loaded hydrogels comprising modified alginate were shown to have cytotoxic activity. The results indicate that β-CyD functionalized alginates have potential to be used as a material for drug delivery systems. The hydrogel system could also be applied in future work for delivery of other hydrophobic drugs and molecules that can form inclusion complexes with β-CyD.