In-Situ Assembly of Ca–Alginate Gels with ... - ACS Publications

Sep 8, 2015 - popular gels successfully utilized as polymer scaffolds. Here we propose a ... culture a precise control over properties and internal st...
1 downloads 0 Views 8MB Size
Download PDF
Article pubs.acs.org/Langmuir

In-Situ Assembly of Ca−Alginate Gels with Controlled Pore Loading/ Release Capability Alena S. Sergeeva,†,‡ Dmitry A. Gorin,‡ and Dmitry V. Volodkin*,†,§ †

Fraunhofer Institute for Cell Therapy and Immunology (Fraunhofer IZI), Am Muehlenberg 13, Potsdam, 14476, Germany Saratov State University, Astrakhanskaya 81, Saratov, 410012, Russia § Lomonosov Moscow State University, Department of Chemistry, Leninskiye gory 1-3, Moscow, 119991, Russia Langmuir 2015.31:10813-10821. Downloaded from pubs.acs.org by UNIV OF BRITISH COLUMBIA on 01/03/19. For personal use only.

‡

S Supporting Information *

ABSTRACT: Development of tailor-made porous polymer scaffolds acting as a temporary tissue-construct for cellular organization is of primary importance for tissue engineering applications. Control over the gel porosity is a critical issue due to the need for cells to proliferate and migrate and to ensure the transport of nutrition and metabolites. Gel loading with bioactive molecules is desired for target release of soluble signals to guide cell function. Calcium−alginate hydrogels are one of the most popular gels successfully utilized as polymer scaffolds. Here we propose a benchtop approach to design porous alginate gels by dispersion of CaCO3 vaterite crystals in sodium alginate followed by the crystal elimination. CaCO3 crystals play a triple role being (i) cross-linkers (a source of calcium ions to cross-link gel network), (ii) pore-makers (leaching of crystals retains the empty pores), and (iii) reservoirs with (bio)molecules (by molecule preloading into the crystals). Pore dimensions, interconnectivity, and density can be adjusted by choosing the size, concentration, and packing of the sacrificial CaCO3 crystals. An opportunity to load the pores with biomolecules was demonstrated using FITC-labeled dextrans of different molecular masses from 10 to 500 kDa. The dextrans were preloaded into CaCO3 vaterite crystals, and the subsequent crystal removal resulted in encapsulation of dextrans inside the pores of the gel. The dextran release rate from the gel pores depends on the equilibration of the gel structure as concluded by comparing dextran release kinetics during gelation (fast) and dextran diffusion into the performed gel (slower). Macromolecule binding to the gel is electrostatically driven as found for lysozyme and insulin. The application of porous gels as scaffolds potentially offering biomacromolecule encapsulation/release performance might be useful for alginate gel-based applications such as tissue engineering.

■

INTRODUCTION Design of porous polymer scaffolds acting as a temporary tissue-construct for cellular organization is one of the main objectives in tissue engineering and regenerative medicine. Biocompatible alginate hydrogels are known as promising materials for a wide range of biomedical applications serving as artificial extracellular matrices for biological studies,1−3 allowing cartilage repair,4,5 bone regeneration,5,6 wound healing,5 nerve tissue engineering,7 providing drug delivery,2,5 and preventing surgical adhesion.8 Versatile applications of alginate hydrogels are possible due to tailor-made structure and tunable properties (mechanical strength, swelling behavior, controlled mesh size) of alginate gels which are provided by the specificity of divalent metal ions to differently arranged α-L-guluronic and β-Dmannuronic acid monomers of alginate backbone.2,5,9 For cell culture a precise control over properties and internal structure of the hydrogel scaffolds is of crucial importance due to the need for cells to accommodate the polymer matrix, migrate, and proliferate. Mass transport through the alginate gel matrix is also defined by both gel physicochemical properties and structure (mesh size of gel network and microscale porosity).9,10 Thus, the gel porosity has to be tuned in the © 2015 American Chemical Society

scale comparable with cell dimensions to ensure cell migration and diffusion of metabolites and nutrition. Encapsulation of bioactive molecules into the gel is desirable feature allowing providing cells with necessary soluble signals such as growth factors and cytokines which define a cell fate. Alginate hydrogels with a variety of geometries and topological features have been proposed such as thin films patterned with microwells,11 gel grids,12 three-dimensional sponge-like3 structures, and gels with tube-like13 or spherical14 pores. Hydrogels produced using simple benchtop approaches have anisotropic properties and uncontrolled internal architecture (randomly distributed pores of different dimensions).3 In other cases there is a need to utilize special techniques such as lithography and printing. Thus, there is a need to find an easy way to produce alginate hydrogel scaffolds with controlled porosity and desirable feature of encapsulation and release of biomolecules. Received: April 27, 2015 Revised: August 11, 2015 Published: September 8, 2015 10813

DOI: 10.1021/acs.langmuir.5b01529 Langmuir 2015, 31, 10813−10821

Article

Langmuir

Average size of the crystals has been calculated from optical microscopy images (transmission mode on confocal laser scanning microscope) and SEM images from the particle diameter measured (at least 25 particles have been counted). Both types of synthesized CaCO3 crystals are vaterite polymorphs as indicated from the typical porous internal structure and spherical size19−21 (Figure 1a,b). The shape and porosity are typical features allowing distinguishing between the polymorphs of CaCO3. Calcite or aragonite (other two polymorphs of CaCO3) are nonporous cubic crystals and elongated needles, respectively. Loading of CaCO3 Crystals with FITC-Dextran. FITC-dextrans (MW 10, 70, 500 kDa) were encapsulated into the sacrificial CaCO3 vaterite crystals (average size 8 μm) by one of two approaches: solvent evaporation15,22,23 or coprecipitation.24 Solvent evaporation. The molar ratios of FITC-dextrans to CaCO3 were 250:1, 3500:1, and 50000:1 for the FITC-dextrans with MWs of 10 kDa, 70 kDa, and 500 kDa, respectively. The high excess is necessary to achieve the maximal loading of FITC-dextrans into the nanoporous CaCO3 vaterite. The loading by solvent evaporation was done as follows. A 20 mg sample of the already prepared and dried CaCO3 vaterite crystals was dispersed in 0.5 mL of 1 mg·ml−1 FITCdextran solution (25 °C). The dispersion of CaCO3 vaterite crystals in FITC-dextran was mixed at room temperature over 15 min by constant shaking using a Vortex-Genie 2 (model G-560E, Scientific Industries Inc., USA) resulted in permeation of FITC-dextran molecules inside the nanoporous vaterite crystals. After shaking, the dispersion was dropped onto the cover glass and dried in air at 70 °C over 1 h. The dried CaCO3 vaterite crystals loaded with FITCdextrans were stored in Eppendorf microtubes in darkness at room temperature. Coprecipitation, FITC-dextrans were integrated into the internal structure of CaCO3 vaterite during the crystal growth. The synthesis was performed at room temperature as follows. A 2.5 mL aliquot of FITC-dextran solution (0.2−0.4 mg·ml−1) was placed in a glass beaker (total volume of 10 mL). Under a constant rotation induced by a thermomagnetic stirrer, 0.615 mL of 1 M Na2CO3 and 0.615 mL of 1 M CaCl2 were added at once to the beaker with FITC-dextran. Mixing time was 10 s. Then the stirring was stopped, and the suspension was incubated for another 3 min at room temperature to complete the crystal growth. The formed CaCO3 vaterite were rinsed with Millipore water three times. Finally, the suspension was dropped onto a cover glass and dried in air at 70 °C during 1 h. The dried CaCO3 crystals loaded with FITC-dextrans were stored in Eppendorf microtubes in darkness at room temperature. Fabrication of Alginate Gel with the Empty Pores or Loaded with FITC-Dextran. To fabricate porous alginate gel, CaCO3 crystals (8 or 33 μm in average size) were mixed with 3−5% alginic acid sodium salt solution in water. Note that to encapsulate FITC-dextrans into the pores of the forming gel, CaCO3 crystals preloaded with FITC-dextrans should be used. Concentration of CaCO3 crystals in the suspensions (20 or 200 mg·ml−1) provides a sufficient amount of Ca2+ ions to cross-link all the carboxylic −COO− groups of alginate molecules (the 2 time excess of Ca2+ to COO− was always provided). In addition, sodium chloride (0−1 M NaCl) was added to the alginate solution on this step if needed in order to examine the effect of salt on pore formation. A fresh suspension of CaCO3 crystals in alginate solution (i) was spread over a flat glass substrate to form a thin film or (ii) was placed in an Eppendorf microtube. Aiming to obtain high porosity and interconnected pores, CaCO3 crystals were packed in alginate solution via strong centrifugation (10000 rpm, 3 min). Then, HCl at a concentration in the range 0.01−1 M (0.36−3.65%, w/w) was added on top of the suspension of CaCO3 vaterite in sodium alginate to eliminate the crystals leading to formation of the porous gel. The molar ratio of HCl to CaCO3 was chosen to be at least 2 to ensure a full dissolution of the crystals. The fabricated hydrogels were stored at room temperature in 0.1 M CaCl2 solution to prevent the gel degradation with time caused by the cross-linking Ca2+ ion release. Confocal Laser Scanning Microscopy (CLSM). CLSM experiments were performed using Zeiss LSM 510 Meta installation (Zeiss, Germany). Oil-immersion objectives with 40× and 63× magnifications

In this work composite alginate gels with stable and welldefined micropores have been fabricated by means of one-step approach based on leaching of CaCO3 crystals dispersed in alginate solution. Vaterite CaCO3 crystals have been chosen as a source of Ca2+ ions cross-linking the gel. The crystals play a role of sacrificial templates allowing the pore formulation and also encapsulation of the material of interest into the pores of the forming gel. Loading of the pores is achieved due to the crystal solubility at mild conditions (acidic medium or EDTA) and their loading capability.15−18 We focus on pore structure, stability as well as encapsulation capability of the porous hydrogels. Confocal fluorescent scanning microscopy (CLSM) has been employed to examine the gel internal structure and loading/release of macromolecules (dextrans) encapsulated into the pores of the gel.

■

EXPERIMENTAL SECTION

Materials. Alginic acid sodium salt from macrocystis pyrifera (MW 12-80 g·mol−1, viscosity of 2% solution at 25 °C approximately 250 cps, plant cell culture tested, Sigma-Aldrich, CAS Number 9005-38-3), calcium carbonate (CaCO3) crystals with the average diameters of 8 ± 2 and 33 ± 9 μm, hydrochloric acid fuming 37% (HCl, for analysis, Merck Chemicals, CAS Number 7647-01-0), Rhodamine 6G chloride (Rho 6G, dye content approximately 99%, Sigma-Aldrich, CAS Number 989-38-8), sodium chloride (NaCl, ≥99.5%, bioreagent, suitable for cell culture, ROTH, CAS Number 7647-14-5), phosphate buffer saline tablets (PBS, Sigma-Aldrich), Trizma base (TRIS, Primary Standard and Buffer, ≥99.9% (titration), crystalline, Sigma, CAS Number 77-86-1), FITC-dextrans (MW 10, 70, 500 kDa, SigmaAldrich, CAS Number 60842-46-8), sodium carbonate (Na2CO3, ACS reagent, anhydrous, >99.5%, Sigma-Aldrich, CAS Number 497-19-8), calcium chloride dihydrate (CaCl2, ACS reagent, ≥99%, SigmaAldrich, CAS Number 10035-04-8). All chemicals were used without further purification; all solutions were prepared using Millipore water produced in a three-stage Millipore Milli-Q Plus 185 purification system and having a resistivity higher 18.2 MΩ·cm. Synthesis of CaCO3 Crystals. To fabricate two types of alginate gels possessing pores of different dimensions, CaCO3 vaterite with average diameters of 8 ± 2 and 33 ± 9 μm (later referred to as 8 and 33 μm CaCO3 crystals, respectively) were synthesized based on the approach established elsewhere.19−21 To fabricate 8 μm CaCO3 vaterite, Millipore water (2.5 mL) was placed in the glass beaker (total volume of 10 mL) and stirred slowly at an agitation speed of approximately 200 rpm using a thermomagnetic stirrer (IKA-Werke GmbH & Co. KG, Germany). Then, 0.625 mL of 1 M Na2CO3 and 0.615 mL of 1 M CaCl2 solutions were added simultaneously into the beaker under continuous rotation. The suspension with the nucleating CaCO3 crystals was stirred over 30 s. Then rotation was stopped and the suspension was incubated for 180 s to complete the growth of CaCO3 vaterite. Resulted crystals were separated by centrifugation (10000 rpm, 1 min), rinsed three times with Millipore water, and airdried at 70 °C during 1 h. The fabrication of the crystals as described was done at room temperature. Larger carbonate crystals were produced at lower temperature because it allows growing crystals of larger size without an increase of the calcite form of CaCO3. To produce 33 μm CaCO3 crystals, all the initial salt and water solutions used were preliminary cooled down to 7.5 °C. At first, 3 mL of 1 M CaCl2 was added to 9 mL of Millipore water under constant stirring by a thermomagnetic stirrer, then 3 mL of 1 M Na2CO3 was rapidly added to the solution at ambient conditions (agitation speed approximately 160 rpm, the mixing time 30 s), followed by incubation during 25 min at 7.5 °C. Then incubation was continued at room temperature (15 min). Resulting CaCO3 crystals were separated by centrifugation using a microliter centrifuge Heraeus Biofuge Pico (Kendro Laboratory Products, Germany) (10000 rpm, 1 min), rinsed three times with Millipore water, and dried in air at 90 °C overnight. The dry crystals were stored at room temperature in an Eppendorf microtube. 10814

DOI: 10.1021/acs.langmuir.5b01529 Langmuir 2015, 31, 10813−10821

Article

Langmuir (numerical apertures are 1.3 and 1.4, respectively) were used to capture the optical transmittance and fluorescence images. Standard filter settings for excitation and emission of FITC were applied for laser sources with the wavelengths of 488 and 633 nm, respectively, for the experiments where FITC-dextrans were used. To examine the internal structure of the highly porous alginate gels, the samples were stained with Rhodamine 6G: fabricated gel was placed into 0.1 mg·ml−1 Rho 6G water solution for 24 h to allow time for dye permeation over the whole sample. The stained gel was rinsed with water and placed in the tightly closed ibidi chamber (a 35 mm imaging dish with an ibidi Standard Bottom, low walls, and an imprinted 500 μm relocation grid, Cat. No. 80151) to avoid evaporation leading to gel drying and shrinking. In the closed ibidi chamber, gel structure was examined using a laser source with the wavelength 543 nm by a Z-stack imaging mode of CLSM. The obtained CLSM images were analyzed by means of Zeiss LSM Image Browser and ImageJ software. Profile graphs were plotted using Microsoft Excel software. The size of micropores and CaCO3 crystals was measured as the width of the fluorescent or transmittance profile at a half of its intensity maximum. Macromolecular Release from the Gel Pores during the Gelation and Diffusion Study into the Performed Gel. Dextrans, lysozyme, and insulin were chosen as model macromolecular probes for mass transport in the gel examined by CLSM. The proteins were conjugated with FITC at protein:FITC molar labeling ratio 5:1 for lysozyme and 7:1 for insulin to avoid labeling of any protein molecule with more than two FITC molecules to minimize chemical changes by the fluorescent dye. The labeling reaction was performed for 2 mg mL−1 protein solution in carbonate buffer (pH 9.0−9.2) during 4 h followed by dialysis against water and further TRIS buffer (pH 7.2− 7.5). Release and diffusion experiments were performed in the tightly closed ibidi μ-Dishes35 mm Grid-500 (a 35 mm imaging dish with an ibidi Standard Bottom, low walls, and an imprinted 500 μm relocation grid, Ibidi, Germany, Cat. No. 80156) fixed on the objective table of the CSLM microscope in order to avoid solvent evaporation and gel movement. To examine macromolecular mass transport properties of porous alginate gels, that is, the transport through the gel network, two alternative experiments were done: (I) release of the pre-encapsulated into CaCO3 templates macromolecules outside from the pore into the gel, and (II) diffusion of the macromolecules from the solution of a big volume into the gel. Both experiments were carried out in online mode using CLSM installation. For macromolecular release from the pores during its formation, the gel was fabricated in situ on a cover glass placed on the microscope table, so that all the steps of gelation process were observed by CLSM (namely, dispersion of CaCO3 templates with pre-encapsulated macromolecules in alginate solution, dissolution of templates, and gel cross-linking, release of the encapsulated macromolecules outside the pores into the gel). Release of the CaCO3 pre-encapsulated macromolecules from the pores was assumed to be started at the time moment when CaCO3 template dissolves as visually observed. For study on macromolecular diffusion into the performed gel, the already prepared alginate gel was placed in the ibidi μ-Dish35 mm Grid-500 and exposed to 1 mg mL−1 solution of a chosen macromolecule followed by immediate monitoring of fluorescence outside and inside the gel. The volume of the solution was in excess, at least 10 times higher as compared with the corresponding hydrogel sample volume.

Figure 1. (a,b) SEM images of the synthesized CaCO3 vaterite crystals (average diameter of 8 μm) preloaded with FITC-dextran, MW 10 kDa. (c,d) Schematics of the approach to fabricate porous alginate hydrogel: (c) the sacrificial CaCO3 vaterite crystals with preencapsulated (bio)molecules are mixed with sodium alginate solution; (d) addition of HCl provides the CaCO3 dissolution leading to the release of Ca2+ ions cross-linking the gel network, so that the pore is formed after the CaCO3 crystal removal. The dextran molecules preencapsulated into the CaCO3 crystals can be released from the pore.

formed in gel are empty, which can be proven by comparison of fluorescence outside the Rho−stained gel and inside the pore (Figure 2c). In both cases the level of fluorescence corresponds to the background signal meaning that the pores are empty. Porosity has been found to be influenced by the way of dispersion of CaCO3 crystals in alginate solution. In the case when CaCO3 crystals were sedimented by centrifugation, the formed gel had interconnected pores resulting from the contacts between the closely packed CaCO3 crystals (interconnected pores are marked by white arrows in Figure 2a). If the crystals were not packed before gelation but just dispersed in alginate, the pores are separated from each other (Figure 3). The position of pores in the alginate gel coincides with the position of initially dispersed sacrificial carbonate crystals. This gives an advantage to tune pore density and pore interconnectivity by suitable dispersion of the CaCO3 crystals that is indispensable for tissue engineering applications. So, if it is necessary to infiltrate the gel with biological cells, it is

■

RESULTS AND DISCUSSION Tunable Porosity of Gels. According to the general approach described in the experimental part (Figure 1), alginate hydrogels with different internal architecture and composition were prepared. All the hydrogels fabricated at conditions examined here had pores after CaCO3 crystals removal and demonstrated a wide range of pH stability, from very acidic conditions (HCl or citric acid, pH 0) to neutral pH at physiological conditions (TRIS or PBS, pH 7.4). If there is no pre-encapsulated material in the CaCO3 crystals, the pores 10815

DOI: 10.1021/acs.langmuir.5b01529 Langmuir 2015, 31, 10813−10821

Article

Langmuir

Figure 3. Optical transmittance images of 8 μm CaCO3 crystals dispersed in 5% alginate solution (at crystal concentration of 20 mg/ mL) containing 1 M NaCl (a) followed by addition of HCl (b). Yellow arrows indicate the pores affected with the released CO2 bubbles; red arrows indicate the pores not affected with CO2 bubbles and retaining dimensions of the removed CaCO3 crystals.

achieve high local concentration of calcium ions for faster alginate cross-linking, a rather high concentration of HCl (≥0.01M) was used. The identical size of pores and crystals means that there is no dominating force in the balance between osmotic pressure generated by the released Ca2+ ions and the surface tension which should result in a energetically more favorable closure (collapse) of the pores. One can assume that there is a high integrity of the gel network cross-linked very fast (second range) by the abruptly released Ca2+ ions during the template elimination. Indeed, the distance x which ions/ molecules may travel within one second (t = 1) can be found from the Einstein diffusion equation28,29

Figure 2. Fluorescent (a) and optical transmittance (b) images of the highly porous alginate gel prepared using 33 μm CaCO3 crystals packed in 3% alginate solution (initial suspension concentration is 200 mg/mL) and 0.1 M HCl; the formed gel is stained with Rho 6G. The fluorescent profile (c) is taken as shown with the yellow interrupted line in image a. White arrows in image a indicate interconnections between the pores in the alginate gel.

x=

2Dt

(1)

where D is diffusion coefficient for a given ion/molecule, and t is diffusion time. Thus, according to calcium diffusion coefficient (D(Ca2+) = 7.54 × 10−6 cm2·s−1),30 Ca2+ ions doped from the dissolving CaCO3 vaterite crystals are able to cross-link alginate molecules over the area of about 40 μm2 within only 1 s, providing enough stable cross-linked junctions to the gel network almost instantaneously. To estimate the ion diffusion as described above, we used the simple relation between the distance traveled by the ions and ion diffusion coefficient. Real distribution of the diffused ions is more complex and depends on the multiple parameters such as concentration gradient, temperature, viscosity, etc. We assume that there are no diffusion limitations for calcium ions to diffuse through the

important to pack CaCO3 crystals as close as possible to obtain a maximal number of interconnections between the pores. The pore dimensions in the prepared gels have been found to be identical to the precursor crystals (the same size, shape, and position) (Figure 2−4). This can be explained by the mechanism of pore formation. CaCO3 vaterite crystals play a multiple role in the hydrogel fabrication process. The solubility of calcium carbonate25−27 provides an opportunity to eliminate the sacrificial CaCO3 crystals at slightly acidic conditions (pH < 7). Thus, addition of HCl induces the dissolution of calcium carbonate, accompanied by the release of Ca2+ ions binding alginate molecules. To accelerate the dissolution process and 10816

DOI: 10.1021/acs.langmuir.5b01529 Langmuir 2015, 31, 10813−10821

Article

Langmuir

have further considered if ionic strength may affect the gel formation and its structure because of the presence of salt at physiological conditions. Influence of Salt. While in the pure alginate gel the pores have been found identical to the precursor CaCO3 crystals (the same size, shape, and distribution), the addition of NaCl (in physiological solution the content of NaCl is 0.9%) to the alginate can affect the pores. To understand if NaCl induces changes of alginate gel structure, first the prepared hydrogel was placed in 6 M NaCl solution. However, no influence on the gel structure has been found. Then, NaCl was added directly to the alginate solution before gelation. Above the critical salt concentration (1 M NaCl in 5% alginate) phase separation has been observed, and no stable pores have been found (a lack of pores at all, or it is impossible to distinguish them). The gel kept the porous structure if up to 1 M of NaCl was added to alginate solution. It is significant to emphasize that the pores are partially affected in the presence of NaCl. Indeed, there are swollen pores in certain places (indicated with yellow arrows in Figure 3), while the other pores have the same size as sacrificial CaCO3 crystals (red arrows in Figure 3). Optical microscopy (online monitoring of the pore formation) has evident that the observed swelling is attributed to the CO2 gas bubbles which were formed inside the pores during CaCO3 dissolution (data not shown). Gas bubbles enlarging the pores mechanically appeared in some places because the solubility of CO2 was decreased in the presence of NaCl salt.27 In fact, the release of CO2 takes place even when CaCO3 crystals are decomposing into ions,25 but usually CO2 molecules uniformly diffuse outside to the surrounded gel without any bubbles. The microscale gas bubbles were observed only if the alginate solution contained NaCl. In this context the solubilization rate of CO2 was reduced, so that CO2 molecules caused the formation of gas bubbles swelling the pores mechanically. In principle, the other mechanism of the pore enlargement can be proposed. The addition of NaCl to alginate may reduce the cross-linking of the gel network via the screening of charges of the carboxylic group when the competition between Ca2+ and Na+ ions (and protons) for the available binding sites on alginate chains.31,32 Most probably, this process does not take place, otherwise all the pores should be affected to the same or similar extent; however, this is not the case (bubble formation is out of control Figure 3b). Moreover, the binding constant of alginate to Ca2+ is rather high (Ka(Alg/Ca2+) ≈ 1.07 × 104 M−1)33 as compared with that to Na+ (Ka(Alg/Na+) ≈ 2.25 × 100 M−1) and to H+ (Ka(Alg/H+) ≈ 5.00 × 102 M−1).34 So, only the solubility of CO2 plays a role in bubble formation affecting pores. The gas bubbles have been found mainly in places with higher CaCO3 concentration that also supports the proposed mechanism of bubble formation. Thus, the addition of salt to alginate affects the kinetics of gelation via formation of CO2 gas bubbles enlarging the pores, so that dimensions of the pores are not identical to the sacrificial CaCO3 crystals anymore. The process of bubble generation is completely uncontrolled, and the only average pore size can be tuned by means of doping alginate with NaCl. For example, the addition of 0.5 M NaCl to 5% alginate solution caused an average pore size increase by 18% when 0.1 M HCl was used for gelling. Without NaCl, the pores have the size identical to the calcium carbonate crystals under otherwise equal gelling conditions.

Figure 4. (a−h) CLSM-images showing an opportunity to encapsulate FITC-dextran molecules into the pores of alginate gel: (a−d) CaCO3 crystals preloaded with 70 kDa FITC-dextran dispersed in 3% alginate followed by the addition of 0.01 M HCl; no FITC-dextran molecules inside the pores; (e−h) CaCO3 crystals preloaded with 500 kDa FITC-dextran dispersed in 3% alginate followed by the addition of 0.01 M HCl; the molecules were saved inside the pores for a long time. Fluorescence images d and h were obtained through 5 min after the release beginning; the fluorescent profiles i and j are taken through the yellow interrupted lines in images g and h, respectively. Yellow interrupted circles in images b and f serve as guide lines indicating boundaries of the pores formed in the alginate gel.

porous hydrogel which has a mesh size much larger than the dimensions of the ions. To better understand the formation of the porous gel and to get an option to control the porosity, we 10817

DOI: 10.1021/acs.langmuir.5b01529 Langmuir 2015, 31, 10813−10821

Article

Langmuir Loading of the Pores of Gel. As mentioned in the introduction, the nanoporous nature of vaterite calcium carbonate allows the encapsulation of fragile biomolecules at mild conditions.15,17,35 In this part of the work, the alginate hydrogels were prepared using the CaCO3 crystals preloaded with FITC-dextrans of different molecular masses to demonstrate an opportunity to load the pores of the gel. To show the principle of the loading, dextrans have been pre-encapsulated into 8 μm-sized crystals by coprecipitation because this approach provides the uniform distribution of the macromolecules over the whole volume of CaCO3 vaterite (Figure 4). Solvent evaporation also allows trapping the dextran molecules into the volume of the crystals (Figure S2); however, the molecule distribution is mostly on the edges of the crystals. Both approaches should be applicable because they aim to load the molecules of interest into the internal volume of the crystals. CLSM monitoring of the gelling process was started before addition of the gelation agent (0.01 M HCl) and was finished after a few hours because the release of the polymer molecules pre-encapsulated in pores might be slow. Obtained results demonstrate that FITC-dextran molecules might be trapped inside the pores of the gel if the hydrodynamic radius is large enough to prevent a spontaneous diffusion of molecules from the pores to the surrounded gel (Figure 4). The release of FITC-dextrans after elimination of the CaCO3 crystals was found to be directly dependent on molecular mass of the molecules (Figure 5), which is typical for homogeneous

spontaneous drug diffusion (see Einstein diffusion eq 1) through the liquid filled gel matrix. Because of the high dissolution rate, it is a challenge to determine a model fitting the demonstrated burst release (Figure 5).39,40 Under identical conditions, high-molecular weight FITC-dextran (MW of 500 kDa, Stokes radius Rh is 15.90 nm)37 was captured inside the pores for a much longer time (Figure 5). According to the Einstein relation eq 1, for 500 kDa FITC-dextran having diffusion coefficient D of 43 μm2/s,41 a distance traveled per second will be about 9 μm if there is free diffusion out from the pore. However, 500 kDa FITC-dextran was trapped inside the pores for a long time in contrast to the low-molecular weight molecules (MW of 10 and 70 kDa) quickly dissolving through the gel matrix. Note that the distribution of the long-chain FITC-dextran molecules in the pore (Figure 4h) was slightly changed if compared with the CaCO3 crystal (Figure 4g) which can be attributed to the optical effects taking place in the pores filled uniformly with a solvent and in the solid crystals. We have also examined the effect of alginate concentration used to prepare the gel on the release performance of 500 kDa dextran. There is no effect observed for the gels prepared using either 3% or 5% alginate solution (Supporting Information, Figure S1). This most probably means that the mesh size of the gel does not depend on the initial alginate concentration (in the examined range). So, the dextran release determined by the sterical hindrance for the uncharged dextran molecules, will only depend on the molecule dimensions and gel swelling state. We used dextrans to avoid gel−molecule interactions, focusing on identification of the release cutoff based on the molecular mass avoiding gel− molecule interaction, for instance by electrostatics. One can conclude that according to the hydrodynamic radii of 70 and 500 kDa dextrans37,38 a cut off is in the range of 7−16 nm. This means that large enough FITC-dextran molecules are stopped to diffuse freely through the alginate gel network, and the release kinetics is changed from the burst to more complicated dissolution profile (Figure 5), which is beyond zero and first order approximation40,42 but is well fitted to the Korsmeyer− Peppas model42−45 defined in general by the function f=

Ct = kt n + b C∞

(2)

Here, the fractional release of the drug (Ct/C∞) is defined by the constant k incorporating structural and geometrical characteristics of the drug dosage form, and by the exponent n indicating different release mechanisms.42 Parameter b takes into account the possible burst effect (otherwise, b = 0 if the burst is absent).42 To determine n, the only initial interval (Ct/ C∞) < 0.6 of the release curve should be used.42 As applied to the release curve for 500 kDa dextran (describing by the equation y = 900 x−0.9 with a determination coefficient R2 = 0.88), the |n| value lies in the range 0.5−1 meaning that the mass transfer exhibits non-Fickian (anomalous) diffusion.42,46 However, it should be taken into account that the fitting model applied does not consider all the factors affecting the diffusion through the gel network. In this way, the diffusion of FITCdextran could be Fickian but affected by a variety of factors, including the swelling behavior of alginate gel due to the osmotic pressure generated by dextran molecules diffusing through the gel network.36,46 Indeed, the swelling state of the hydrogel can be assumed as constant if there is no diffusion processes within the gel (no pressure of diffusing molecules on

Figure 5. Effect of the molecular weight on the release of FITCdextran molecules from the pores of the alginate hydrogel. Start of the release process is assumed at the initial instant t0 when the CaCO3 crystal is just dissolved. An average fluorescence intensity inside the pore at the time moment t0 is taken as 100%. The interrupted line for dextran 500 kDa corresponds to fitting of the experimental data using the Korsmeyer−Peppas model. The bars on the graph correspond to the standard deviations.

matrices where the release is controlled only by a spontaneous diffusion.36 As evident from the release curves and CLSM images, during the contact with the gel (release environment), low-molecular weight FITC-dextrans (MW of 10 and 70 kDa, Stokes radii Rh are about 1.86 and 6.49 nm, respectively37,38) were instantly dissolved (no fluorescence inside the pores, Figure 4d) meaning that the hydrogel matrix is not a barrier for these dextrans, so the release kinetics were essentially ruled by a 10818

DOI: 10.1021/acs.langmuir.5b01529 Langmuir 2015, 31, 10813−10821

Article

Langmuir the gel network). However, all the examined dextran molecules are small enough to diffuse through the gel network, so that they will swell the gel network during diffusion, which will develop the opposite restrictive force from the gel.46 So, an equilibrium state of gel will be only reached when this elastic restoring force of alginate chains becomes equal to the osmotic pressure driven by the FITC-dextran molecules.46 Until this moment, the front of the shock diffusant wave will penetrate into the polymer matrix causing the osmotic pressure and network rearrangement.46 To summarize, dextran release kinetics is determined mainly by the sterical hindrance for molecules that depends on a density of the gel network (mesh size). Interactions of FITCdextran with the alginate chains should not affect the release mechanism owing to the almost neutral charge of the molecules. Perhaps, the diffusion of small FITC-dextrans outside of the pore can be fortified by the internal osmotic pressure generated by molecules/ions releasing during the CaCO3 dissolution. Moreover, dextran molecules themselves provide some pressure on the pore during the release that could affect the pore formation process. The values of the generated osmotic pressure may be rather high (in the range of MPa) due to small size of the releasing ions; however, the pressure generated is not enough to enlarge the formed pores. The effect of osmotic pressure on pore formation is addressed in our separate study. However, the dimensions of CaCO3 crystals and pores were identical even if FITC-dextran was pre-encapsulated into the sacrificial CaCO3 crystals by a complete pore filling. This means that the gel is still very resistant, and even the internal osmotic pressure induced by both the released ions and the preloaded FITC-dextran molecules is insufficient to let the pores swell. Further we would like to understand if the interaction between molecules to be encapsulated into the gel pores (and further released) and the alginate gel network may affect the release rate. This can be expected because dextrans, as mentioned above, do not possess any significant charge and their interaction with the gel is expected to be solely dependent on the gel porosity. Moreover, it should be noted that the polydispersity index of the dextran molecules used (as of many natural polymers) is rather high and reported to be 0.7−0.9 (for dextrans of MW from 8 to 2000 kDa).47 Therefore, the estimated cutoff (7−16 nm) and molecular weights of dextrans can only provide a rough idea for release kinetics. For more deep understanding of the release rate, macromolecules with narrow size distribution such as proteins should be used. In this respect, proteins are also very attractive model macromolecules because a defined protein net charge makes them good candidates to examine an effect of electrostatic interactions between diffusing molecules and the alginate gel network. In addition, proteins are often used as models for biologically relevant and biologically active macromolecules such as growth factors, cytokines, hormones, etc. Therefore, we have studied macromolecule−gel interaction by monitoring macromolecular diffusion into the performed alginate gels. This complements and makes analysis of the interaction more simple if compared to that performed for dextrans as described above (by loading of the gel pores with macromolecules followed by dextran release analysis immediately after CaCO3 removal). It is of note that for diffusion into preformed gel an effect of high osmotic pressure generated and reorganization of the gel structure during the CaCO 3

dissolution can be revealed. Schematics of the experiment on protein diffusion into the preformed gel is shown in Figure S3. Two proteins having at neutral pH opposite signs of their net charges have been examined, namely positively and negatively charged lysozyme (pI 10.5, 14 kDa) and insulin (pI 5.2, 6.5 kDa). Dextrans of 10 and 70 kDa were also used for the same analysis. Alginate gel possesses a negative charge due to the low pKa of carboxylic groups of the alginate molecules, which is 3.65 and 3.38 for guluronic and for mannuronic acid blocks, respectively.48 The probes used were labeled by FITC, enabling monitoring their diffusion into the gel by CLSM. Figure 6a

Figure 6. (a) Fluorescence of FITC-labeled macromolecules of lysozyme, insulin, and dextrans (10 and 70 kDa) into preformed alginate gel as a function of exposure time; 100% of fluorescent signal corresponds to the signal in solution (1 mg·ml−1) of the used macromolecules. The interrupted lines are the fitting curves to guide the eye. (b−e) Optical transmittance and fluorescence images of the alginate gels exposed to solutions of the macromolecules tested. The scale bars on all the CLSM images are 20 μm.

shows diffusion profiles for these macromolecules by plotting the fluorescence registered into the gel (100% corresponds to fluorescence in the macromolecule solution) as a function of time of exposure of the probe solution to the gel at static conditions (no stirring). Lysozyme (oppositely charged compared to alginate gel) diffuses very fast into the gel and even accumulates inside, which can be driven by electrostatic interactions. Diffusion of insulin is also rather fast (scale of minutes); however, it does not reach the same concentration as in solution most probably due to no electrostatic interaction with the gel (both are negatively charged). About 60% of the protein fluorescence registering in solution is observed within the gel, which may indicate that a rather small amount of insulin molecules can 10819

DOI: 10.1021/acs.langmuir.5b01529 Langmuir 2015, 31, 10813−10821

Article

Langmuir diffuse through the gel pores easily and fill a significant part of the gel internal volume. Importantly, after rinsing the gel filled with proteins, all the insulin molecules were washed out (no fluorescence was registered inside the gel). In contrast, lysozyme molecules were kept inside the gel for at least 3 h even after multiple rinsing, showing a reduction of the fluorescent signal from the gel by 19%. These results prove that electrostatics plays an important role in the macromolecule−gel interaction, and the gel pores are large enough and do not suppress transport of nanometer-sized macromolecules. The 10 kDa dextran has a similar diffusion profile as insulin (overlap to high extent) demonstrating also about 60% loading (Figure 6a). However, larger dextran (70 kDa) diffuses slower and its fluorescence in the gel is also about 50−60% compared to that in solution. The slower diffusion may be attributed to a larger size of the dextran. Interestingly, the diffusion kinetics of the dextrans into the gel is slower than release kinetics from the pores into the gel (Figure 5) observed after dissolution of CaCO3 with pre-encapsulated dextrans. This allows one to assume that macromolecules cannot effectively be captured by the gel during the gel formation and even 500 kDa dextran releases rather fast (Figure 5). We believe that this is related to some time necessary for the gel to form an equilibrium structure during its formation. Obviously, proteins encapsulated in the pores such as dextrans in this study should diffuse slower than dextrans due to strong impact of electrostatics in the molecule−gel interaction as proven above for lysozyme and insulin. In future work we would like to investigate how to improve an ability of the gel to capture macromolecules into the pores during the gelation process and also to better tune macromolecule release kinetics.

Thus, the approach described here combines the CaCO3 vaterite loading capability with the templating of porous hydrogel and can be used to achieve the tailor-made architecture of alginate gel loaded with (bio)molecules of interest. One can create the highly porous 3D scaffolds with interconnected pores, applicable for cell culture, if densely packed CaCO3 crystals of large diameter are used for gelation. At the same time, it is possible to fabricate the scaffold as a thin gel film with the closed freestanding pores, in which the bioactive molecules might be encapsulated. The latter structure has promising application in target drug delivery, because preencapsulated inside-the-pores materials might be released from the pores at a controlled manner by some external signal such as light. A control over the scaffold porosity and release of the differently charged (bio)molecules from the scaffold are of high importance for cell growth into/onto the scaffold, therefore these topics have to be studied as the next steps of the actual research.

SUMMARY This work demonstrates a benchtop approach to design porous alginate hydrogels with tuned internal porosity using dispersion of the CaCO3 vaterite crystals in sodium alginate solution. HClmediated dissolution of the sacrificial crystals resulted in crosslinking of the alginate gel network by the released Ca2+ ions. The pores were formed instead of the eliminated crystals; the porosity (size, shape, concentration, and distribution of the pores) is identical to the sacrificial CaCO3 crystals, so that the alginate hydrogel structure can be tailored, depending on the aim. Moreover, pre-encapsulation of the macromolecules into CaCO3 vaterite crystals provides loading of the pores of gel with these molecules after dissolution of the crystals. The absence of charge for dextran molecules means that release kinetics outside of the pores is determined by only the sterical hindrance for the given molecules that depends on the density of the gel network (mesh size of the gel network and the hydrodynamic radius of the molecule). On the basis of behavior of the FITC-dextrans of different molecular weights, a cutoff to suppress macromolecular release from the pores just after gelation was estimated to be in the range 7−16 nm. The dextran release rate from the gel pores is enhanced as concluded by a comparison of dextran release kinetics during the gelation and dextran diffusion into the performed gel. Most probably the equilibration of the gel structure during gelation that takes time and the forming gel is much more permeable than the finally prepared one. Electrostatic interactions play a crucial role in protein−gel interaction as shown by diffusion/ release of lysozyme and insulin into/out of the gel.

Notes

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.langmuir.5b01529. Figure S1, normalized fluorescence from the gel pores; Figure S2, CLSM images; Figure S3, schematics of the experiments and additional fluoresence images (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*Phone: +49-331-58187-327. Fax: +49-331-58187-399. E-mail: [email protected].

■

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The work was partially supported by Alexander von Humboldt Foundation (AvH Fellowship and Sofja Kovalevskaja Program) and by DAAD (German Academic Exchange Service, “Mikhail Lomonosov” Program).

■

REFERENCES

(1) Rowley, J.; Madlambayan, G.; Mooney, D. J. Alginate hydrogels as synthetic extracellular matrix materials. Biomaterials 1999, 20, 45− 53. (2) Augst, A. D.; Kong, H. J.; Mooney, D. J. Alginate Hydrogels as Biomaterials. Macromol. Biosci. 2006, 6, 623−633. (3) Sharma, C.; Dinda, A. K.; Mishra, N. C. Fabrication and characterization of natural origin chitosan- gelatin-alginate composite scaffold by foaming method without using surfactant. J. Appl. Polym. Sci. 2013, 127, 3228−3241. (4) Wang, C.-C.; Yang, K.-C.; Lin, K.-H.; Liu, H.-C.; Lin, F.-H. A highly organized three-dimensional alginate scaffold for cartilage tissue engineering prepared by microfluidic technology. Biomaterials 2011, 32, 7118−7126. (5) Sun, J.; Tan, H. Alginate-Based Biomaterials for Regenerative Medicine Applications. Materials 2013, 6, 1285−1309. (6) Cipitria, A.; Lange, C.; Schell, H.; Wagermaier, W.; Reichert, J. C.; Hutmacher, D. W.; Fratzl, P.; Duda, G. N. Porous scaffold architecture guides tissue formation. J. Bone Miner. Res. 2012, 27, 1275−1288. (7) Wang, M.-D.; Zhai, P.; Schreyer, D. J.; Zheng, R.-S.; Sun, X.-D.; Cui, F.-Z.; Chen, X.-B. Novel crosslinked alginate/hyaluronic acid

10820

DOI: 10.1021/acs.langmuir.5b01529 Langmuir 2015, 31, 10813−10821

Article

Langmuir hydrogels for nerve tissue engineering. Front. Mater. Sci. 2013, 7, 269− 284. (8) Esser, E.; Tessmar, J. K. V. Preparation of well-defined calcium cross-linked alginate films for the prevention of surgical adhesions. J. Biomed. Mater. Res., Part B 2013, 101, 826−839. (9) Gombotz, W. R.; Wee, S. F. Protein release from alginate matrices. Adv. Drug Delivery Rev. 2012, 64, 194−205. (10) Golmohamadi, M.; Wilkinson, K. J. Diffusion of ions in a calcium alginate hydrogel-structure is the primary factor controlling diffusion. Carbohydr. Polym. 2013, 94, 82−87. (11) Ozawa, F.; Ino, K.; Arai, T.; Ramón-Azcón, J.; Takahashi, Y.; Shiku, H.; Matsue, T. Alginate gel microwell arrays using electrodeposition for three-dimensional cell culture. Lab Chip 2013, 13, 3128. (12) Zawko, S. A.; Schmidt, C. E. Simple benchtop patterning of hydrogel grids for living cell microarrays. Lab Chip 2010, 10, 379. (13) Despang, F.; Börner, A.; Dittrich, R.; Tomandl, G.; Pompe, W.; Gelinsky, M. Alginate/calcium phosphate scaffolds with oriented, tubelike pores. Materialwiss. Werkstofftech. 2005, 36, 761−767. (14) Roberts, J. R.; Ritter, D. W.; McShane, M. J. A design full of holes: functional nanofilm-coated microdomains in alginate hydrogels. J. Mater. Chem. B 2013, 1, 3195. (15) Schmidt, S.; Volodkin, D. Microparticulate biomolecules by mild CaCO3 templating. J. Mater. Chem. B 2013, 1, 1210. (16) Volodkin, D. CaCO3 templated micro-beads and -capsules for bioapplications. Adv. Colloid Interface Sci. 2014, 207, 306−324. (17) Volodkin, D. Colloids of pure proteins by hard templating. Colloid Polym. Sci. 2014, 292, 1249−1259. (18) Sergeeva, A. S.; Gorin, D. A.; Volodkin, D. V. Polyelectrolyte Microcapsule Arrays: Preparation and Biomedical Applications. BioNanoSci. 2014, 4, 1−14. (19) Volodkin, D. V.; Petrov, A. I.; Prevot, M.; Sukhorukov, G. B. Matrix Polyelectrolyte Microcapsules: New System for Macromolecule Encapsulation. Langmuir 2004, 20, 3398−3406. (20) Volodkin, D. V.; Larionova, N. I.; Sukhorukov, G. B. Protein Encapsulation via Porous CaCO 3 Microparticles Templating. Biomacromolecules 2004, 5, 1962−1972. (21) Volodkin, D. V.; Schmidt, S.; Fernandes, P.; Larionova, N. I.; Sukhorukov, G. B.; Duschl, C.; Möhwald, H.; von Klitzing, R. OneStep Formulation of Protein Microparticles with Tailored Properties: Hard Templating at Soft Conditions. Adv. Funct. Mater. 2012, 22, 1914−1922. (22) Schmidt, S.; Behra, M.; Uhlig, K.; Madaboosi, N.; Hartmann, L.; Duschl, C.; Volodkin, D. Mesoporous Protein Particles Through Colloidal CaCO 3 Templates. Adv. Funct. Mater. 2013, 23, 116−123. (23) Schmidt, S.; Uhlig, K.; Duschl, C.; Volodkin, D. Stability and cell uptake of calcium carbonate templated insulin microparticles. Acta Biomater. 2014, 10, 1423−1430. (24) Petrov, A. I.; Volodkin, D. V.; Sukhorukov, G. B. ProteinCalcium Carbonate Coprecipitation: A Tool for Protein Encapsulation. Biotechnol. Prog. 2005, 21, 918−925. (25) Goss, S. L.; Lemons, K. A.; Kerstetter, J. E.; Bogner, R. H. Determination of calcium salt solubility with changes in pH and PCO2 simulating varying gastrointestinal environments. J. Pharm. Pharmacol. 2007, 59, 1485−1492. (26) Coto, B.; Martos, C.; Pena, J.; Rodriguez, R.; Pastor, G. Effects in the solubility of CaCO3: Experimental study and model description. Fluid Phase Equilib. 2012, 324, 1−7. (27) Liu, Y.; Hou, M.; Yang, G.; Han, B. Solubility of CO2 in aqueous solutions of NaCl, KCl, CaCl2 and their mixed salts at different temperatures and pressures. J. Supercrit. Fluids 2011, 56, 125−129. (28) Miller, C. C. The Stokes-Einstein Law for Diffusion in Solution. Proc. R. Soc. London, Ser. A 1924, 106, 724−749. (29) Einstein, A. Ü ber die von der molekularkinetischen Theorie der Wärme geforderte Bewegung von in ruhenden Flü ssigkeiten suspendierten Teilchen. Ann. Phys. 1905, 322, 549−560. (30) Hazel, J.; Sidell, B. A method for the determination of diffusion coefficients for small molecules in aqueous solution. Anal. Biochem. 1987, 166, 335−341.

(31) Plazinski, W. Sorption of lead, copper, and cadmium by calcium alginate. Metal binding stoichiometry and the pH effect. Environ. Sci. Pollut. Res. 2012, 19, 3516−3524. (32) Seely, G. R.; Hart, R. L. The Binding of Alkaline Earth Metal Ions to Alginate. Macromolecules 1974, 7, 706−710. (33) Fang, Y.; Al-Assaf, S.; Phillips, G. O.; Nishinari, K.; Funami, T.; Williams, P. A.; Li, L. Multiple Steps and Critical Behaviors of the Binding of Calcium to Alginate. J. Phys. Chem. B 2007, 111, 2456− 2462. (34) Lamelas, C.; Avaltroni, F.; Benedetti, M.; Wilkinson, K. J.; Slaveykova, V. I. Quantifying Pb and Cd Complexation by Alginates and the Role of Metal Binding on Macromolecular Aggregation. Biomacromolecules 2005, 6, 2756−2764. (35) Volodkin, D. V. CaCO3 templated micro-beads and -capsules for bioapplications. Adv. Colloid Interface Sci. 2014, 206, 437−454. (36) Grassi, M.; Grassi, G. Mathematical Modelling and Controlled Drug Delivery: Matrix Systems. Curr. Drug Delivery 2005, 2, 97−116. (37) Armstrong, J.; Wenby, R.; Meiselman, H.; Fisher, T. The Hydrodynamic Radii of Macromolecules and Their Effect on Red Blood Cell Aggregation. Biophys. J. 2004, 87, 4259−4270. (38) Lawrence, J.; Wolfaardt, G.; Korber, D. R. Determination of Diffusion Coefficients in Biofilms by Confocal Laser Microscopy. Appl. Enviroment. Microbiol. 1994, 60, 1166−1173. (39) Huang, H.; Brazel, C. On the importance and mechanisms of burst release in matrix-controlled drug delivery systems. J. Controlled Release 2001, 73, 121−136. (40) Kim, C.-K.; Lee, E.-J. The controlled release of blue dextran from alginate beads. Int. J. Pharm. 1992, 79, 11−19. (41) Kommedal, R.; Milferstedt, K.; Bakke, R.; Morgenroth, E. Effects of initial molecular weight on removal rate of dextran in biofilms. Water Res. 2006, 40, 1795−1804. (42) Costa, P.; Lobo, J. M. S. Modeling and comparison of dissolution profiles. Eur. J. Pharm. Sci. 2001, 13, 123−133. (43) Razavilar, N.; Choi, P. In-vitro Modeling of the Release Kinetics of Micron and Nano-Sized Polymer Drug Carriers. Int. J. Drug Delivery 2013, 5, 362−378. (44) Dash, S.; Murthy, P. N. L.; Chowhury, P. Kinetic modeling on drug release from controlled drug delivery systems. Acta Polym. Pharm. 2010, 67, 217−223. (45) Fu, Y.; Kao, W. J. Drug release kinetics and transport mechanisms of non-degradable and degradable polymeric delivery systems. Expert Opin. Drug Delivery 2010, 7, 429−444. (46) De Kee, D.; Liu, Q.; Hinestroza, J. Viscoelastic (Non-Fickian) Diffusion. Can. J. Chem. Eng. 2005, 83, 913−929. (47) Masuelli, M. A. Dextrans in Aqueous Solution. Experimental Review on Intrinsic Viscosity Measurements and Temperature Effect. Journal of Polymer and Biopolymer Physics Chemistry 2013, 1, 13−21. (48) Rehm, B. Alginates: Biology and Applications: Biology and Applications; Springer: 2009.

■

NOTE ADDED AFTER ASAP PUBLICATION This paper was published on the Web on September 16, 2015, with one missing author affiliation. The corrected version was reposted on September 21, 2015.

10821

DOI: 10.1021/acs.langmuir.5b01529 Langmuir 2015, 31, 10813−10821