Swelling and Mechanical Properties of Alginate Hydrogels with Respect to Promotion of Neural Growth

Materials and reagents

Ultrapure sodium alginates with trade name PRONOVA were purchased from Novamatrix, now part of FMC BioPolymer, as guluronic acid rich alginate (LVG) and mannuronic acid rich alginate (LVM), and endogen-free S-alginate were purchased from Sigma (71238; Taufkirchen, Germany). The molecular compositions of LVG and LVM were provided by the supplier, but the composition of S-alginate was quantified for this study by a proton ¹ H-NMR spectroscopy method. ⁴⁷ LVM, LVG, and S-alginates contained 43%, 68%, and 68% guluronic acid residues, respectively. Reagents, including the neurobasal medium, DMEM/F12 (1:1), B-27 supplement, Dulbecco's Phosphate-Buffered Saline without Ca²⁺ and Mg²⁺, and trypsin/EDTA solution were all purchased from Invitrogen (Darmstadt, Germany). CaCl₂ dihydrate, HEPES, TRIS, NaCl, and DNase I originated from Sigma.

Preparation of alginate hydrogels at ambient conditions

Alginates were dissolved in deionized water under permanent stirring at room temperature (RT) for 12 h. The resulting aqueous 0.5%, 1.0%, or 1.5% sols were sterilized by pressing through a 0.45-μm sterile filter. Dried alginate layers were formed on glass coverslips (24×24 mm) when each sol was evenly distributed onto a glass surface and dried at RT for 24 h. Dry alginate layers were stored under humidity-free conditions and used within 10 days after preparation. To form a hydrogel, a glass coverslip with a dry alginate layer was placed into a Petri dish (Ø 5 cm) and overlayed with a 15 mL crosslinking solution; the gelation reaction was carried out at RT for 24 h, and a single interconnected gel slice was formed on the glass surface. The excess crosslinking solution was removed before gravimetric and rheological analyses. Representative images of hydrogels were taken with a digital camera. For the cell culture experiments, hydrogels were prepared under sterile conditions in a laminar flow hood as described previously. ²⁹ For rheological characterization, hydrogels were prepared in plastic Petri dishes (Ø 5 cm) from dry alginate layers derived from 1% alginate sol, and the gelation reaction was carried out at RT for 24 h.

Composition of crosslinking solutions

Alginate hydrogels were prepared through ionic crosslinking, and an aqueous calcium chloride solution was used as a source of Ca²⁺ cations. The concentration of CaCl₂ in the crosslinking solutions ranged from 0 to 10 mM. The pH values of the crosslinking solutions were adjusted to either 5.5 or 7.4 with the HEPES buffer solution. Some of the crosslinking solutions contained NaCl (150 mM, final). Our abbreviation scheme describes the characteristics of the hydrogel's ionic composition in the crosslinking solution; for example, Ca-alginate and Ca-alginate_NaCl hydrogels were generated with CaCl₂ dissolved either in water or in 150 mM NaCl, respectively.

Measurements of alginate hydrogel swelling in gravimetric assay

The weight of each hydrogel (g) was measured on an analytical balance (Sartorius, Göttingen, Germany; 0.1 mg precision). To preserve gel integrity, the cumulative mass of a glass–gel sandwich was measured and the average weight of the glass coverslips (0.22 g) was subtracted to obtain the hydrogel's mass. At least three independent samples were prepared for every experimental condition. The Student's t-test for unpaired observations was applied to make a statistical analysis of the data. The rate of swelling was estimated after comparison of the hydrogel mass with the mass of the 1% alginate sol (taken as 100%).

Rheological characterization of alginate sols and hydrogels

Rheological measurements were carried out using an Anton-Paar Physica MCR301 (Graz, Austria) instrument. The viscosity values of aqueous alginate solutions were recorded under continuous logarithmically increasing shear in a double slit chamber DG26.7 at 20°C. A stress sweep at a constant frequency of 1 Hz was performed to obtain values for the elastic modulus (G′) of alginate hydrogels. Viscosity values were recorded with the help of a measurement cell with parallel plate geometry (PP25) under logarithmically increasing amplitude (0.001–100 Pa) at 37°C. Each hydrogel was transferred onto the measuring device and cut into the shape of a disk (Ø 25 mm). Samples were equilibrated at 37°C for 5 min before measurements.

Culturing of primary neuronal cells

Neuronal cell cultures were prepared from E20 Wistar rat fetuses. Cells were isolated from cortices as described. ⁴⁸ As we reported previously, cleavage of the cell surface molecules during acute brain tissue dissociation using serine protease trypsin alters initiation of neural cell adhesion and growth on soft alginate hydrogels. ²⁹ Therefore, in the present study, we used primary cell spheroids, which were completely recovered from treatment with trypsin. Primary cells were seeded at an initial cell density of 1×10⁶ cells/well in uncovered six-well plates with 2 mL of expansion medium consisting of DMEM/F-12 (1:1) supplemented with 2 mM L-glutamine, 15 mM HEPES, 20 ng/mL of rhEGF (Sigma), and 10 ng/mL of rh-bFGF (Sigma). Subsequently, multiple spheroids formed after reaggregation or/and proliferation of cells on day 3 in vitro. Spheroids were collected, replated onto hydrogel samples, and cultured in a differentiating medium to potentiate differentiation toward neural cells. The differentiating medium consisted of 2% B-27 supplemented neurobasal medium, 0.5 mM L-glutamine, and 1% ampicillin and 1% streptomycin solution. Cell incubation was carried out at 37°C in a humidified 95% air and 5% CO₂ atmosphere.

Immunocytochemistry

The primary monoclonal mouse antibodies were anti-β-tubulin III, anti-GFAP, and anti-Tau-1 (from Chemicon, Schwalbach, Germany) and anti-nestin (R&D Systems, Wiesbaden-Nordenstadt Germany), and secondary antibodies were Alexa Fluor-488 chicken anti-mouse (Molecular Probes/Invitrogen). Labeling with antibodies was performed according to a standard procedure; Hoechst 33342 (10 μg/mL in TBS) was used to label the nuclei. Samples were examined under a fluorescence microscope (Axiovert 40; Zeiss, Jena, Germany) and confocal microscope (LSM-710; Zeiss).

Neurite-bearing spheroids, neurite length, and statistical analysis

Spheroids that were formed from 1×10⁶ primary cortical cells were collected into a sterile tube containing a fresh portion (6 mL) of a differentiating medium, then mixed, and evenly distributed on top of three identical Ca-alginate_NaCl hydrogels (e.g., Ca-alginate_NaCl hydrogels derived from S-alginate crosslinked with a solution containing 2 mM of CaCl₂, 150 mM of NaCl, and at pH 7.4) and cultured subsequently for 48 h. The diameter of spheroids ranged from 50 to 250 μm by day 5 in vitro. Spheroids that extended 10 or more neurite from at least a quarter of its surface contacting the hydrogel were defined as neurite-bearing spheroids; of note, the vast majority of neurite-bearing spheroids extend 50 or more neurite. Single cells as well as aggregates of cells (2–10 cells) were not considered during spheroid counting. The proportion of neurite-bearing spheroids was quantified after the examination of 50 random live spheroids cultured on top of three identical Ca-alginate_NaCl hydrogels. Different types of hydrogels were derived either from S-alginate, LVG, or LVM alginates crosslinked at pH 5.5 or 7.4 with a solution containing different concentrations of aqueous CaCl₂ (i.e., 2, 4, 6, 8, or 10 mM) and 150 mM of NaCl as described above. In each independent cell culture experiment, 30 different hydrogel types were generated and tested for their ability to support neurite outgrowth. Every hydrogel type was generated in triplicate; subsequently, in each independent cell culture experiment, spheroids were examined on 90 individual hydrogels. The final proportion of neurite-bearing spheroids on every hydrogel type was estimated from two independent cell culture experiments after the examination of 100 (in total) individual spheroids on six identical hydrogels. The distance of neurite extension from spheroids was quantified in two independent cell culture experiments 48 h after plating of spheroids on top of different types of soft Ca-alginate_NaCl hydrogels generated using S-alginate, LVG, and LVM with 2 mM of aqueous CaCl₂ at pH 7.4 as described above. Subsequently, 20 random spheroids were examined per one type of a hydrogel and the length of 20–50 random neurite per spheroid was quantified on live phase-contrast images (Zeiss Ph2 Plan-NEOFLUAR 20X/0.5 objective, Axiovert 40; Zeiss) using NeuronJ (www.imagescience.org/meijering/software/neuronj). The Student's t-test for unpaired observations was applied for statistical analysis. Differences between samples with p-value<0.05 were accepted as statistically significant.

Swelling of Ca-alginate hydrogels in aqueous CaCl₂ solution

Swelling is a result of liquid absorption by the polymer, and could be estimated by measuring the hydrogel mass. Dry alginate polysaccharide chains swell in water to produce a sol, but they can swell in the presence of crosslinking di- or multivalent cations, which lead to a hydrogel. Crosslinking of surface-attached dry alginate layers by a 2 mM aqueous CaCl₂ solution resulted in the interconnection of polysaccharide molecules and the formation of a single gel slice (soft hydrogel), which support robust neurite outgrowth in vitro; in contrast, gelation of the alginate sol in an aqueous solution under gentle stirring generated multiple gel fragments of irregular size. The appearance of the soft hydrogels (e.g., fragments or slices) did not alter their high adhesiveness to neurite. These results matched our previous data ²⁹ and demonstrated that dry alginate layers could efficiently form soft hydrogels with the ability to support robust neural growth.

While cultivating neurons on top of soft hydrogels, we often observed variability in the long-term hydrogel stability: some hydrogels disintegrated in the presence of a cell culture medium during the first 3 days, but others persisted for up to 10 weeks and longer. In this study, we generated hydrogels from surface-attached dry layers of different alginates and tested how changes of the ionic strength and pH during the crosslinking influenced mechanical properties of hydrogels. At a constant concentration of calcium (2 mM of CaCl₂), pH 5.5 and either in the absence or presence of NaCl (150 mM) during the crosslinking, the gel mass was proportional to the sol volume (Fig. 1a) and increased linearly with increased alginate concentrations in the sol (Fig. 1b). In the concentration range from 0–10 mM of CaCl₂ in the crosslinking solution, the amount of Ca²⁺ had a nonlinear effect on the growth of a gel mass. A slow linear increase of the hydrogel's mass was correlated with a subsequent decrease in CaCl₂ concentration from 10 to 4 mM, but a much more rapid increase of hydrogel mass was detected after gelation with ≤4 mM of CaCl₂. The formation of a single interconnected hydrogel slice was abolished after Ca²⁺ reached its minimal gel-forming concentration. This swelling profile was observed for S-alginate, LVG, and LVM (Fig. 2). Moreover, the minimal gel-forming concentration of CaCl₂ decreased with an increase of G-content in the alginate, and the level was found to be 0.8±0.2, 1.0±0.2, and 2.0±0.2 mM for S-alginate, LVG, and LVM, respectively (Fig. 2). In the presence of substoichiometric concentrations of calcium ions, larger networks formed when using highly viscous alginates such as LVG and S-alginate (145 and 132 mPa) than when using less viscous LVM (21 mPa) (Fig. 2). The differences in the magnitude of swelling between G-enriched (LVG) and M-enriched (LVM) alginates could be explained by the number of gelling points and their distribution in the primary alginate sequence and by the viscosity of alginates. Indeed, the affinity to bind Ca²⁺ increases with a greater number of G-blocks present in the polymer. ⁶ On the other hand, alginate viscosity is proportional to polysaccharide chain length, and high-viscosity alginates can form larger hydrogels because of the higher probability of crosslinking of the polysaccharide chains and physical entanglement. ⁷

OPEN IN VIEWER

FIG. 1.

Hydrogel volume increases with an increase in (a) the volume of alginate sol at constant alginate concentrations in a sol (1%) and (b) the concentration of an alginate in a sol at a constant sol volume of 600 μL. Hydrogels were generated from dry layers of S-alginate by crosslinking with 2 mM of CaCl₂ at pH 5.5 in the absence (Ca-alginate) or in the presence of 150 mM of NaCl (Ca-alginate_NaCl). Data are represented as means±standard deviations.

OPEN IN VIEWER

FIG. 2.

Swelling of the Ca-alginate hydrogel increases with decreased CaCl₂ concentrations in the crosslinking solution. Hydrogels were generated at pH 5.5 by crosslinking dry alginate layers that originated from 600 μL of a 1% sol or 0.5% sol of (a) S-alginate, (b) guluronic acid rich alginate (LVG), and (c) mannuronic acid rich alginate (LVM). Data are represented as means±standard deviations.

The dissociation constants for mannuronic and guluronic acid monomers are 3.38 and 3.65, respectively. In aqueous solutions with pH 5–9, alginate behaves as a polyanion because carboxylic groups in sodium alginate are fully dissociated. ⁷ In agreement with this fact, we found that the Ca-alginate hydrogel's swelling profile and the minimum gel-forming concentration of calcium ions were not influenced by the pH value (5.5 or 7.4) during the crosslinking (Figs. 2 and 3).

OPEN IN VIEWER

FIG. 3.

NaCl restricts the swelling of hydrogels. Hydrogels were generated at pH 7.4 by crosslinking dry alginate layers with aqueous CaCl₂ solutions in the absence (Ca-alginate) or in the presence of 150 mM of NaCl (Ca-alginate_NaCl); alginate layers originated from 600 μL of 1% sol of (a) S-alginate, (b) LVG, and (c) LVM. Data are represented as means±standard deviations. The gray background distinguishes Ca-alginate_NaCl hydrogels that support robust neurite outgrowth in vitro.

Swelling of Ca-alginate_NaCl hydrogels in aqueous CaCl₂ solution

Like Ca-alginate hydrogels, hydrogels generated in the presence of NaCl (150 mM, Ca-alginate_NaCl) displayed nonlinear swelling profiles in solutions with pH values of 5.5 (not shown) and 7.4, and a decrease in the CaCl₂ concentration to below 4 mM was accompanied by a rapid increase in hydrogel mass (Fig. 3). At a constant concentration of CaCl₂, the Ca-alginate_NaCl hydrogel's mass was proportional to the sol volume and concentration of alginate in the sol (Fig. 1a, b). Na⁺ ions do not have any specific binding sites within the alginate, but interact electrostatically with the negatively charged carboxylic groups of the polysaccharide. ¹⁹ We observed that at a low degree of crosslinking (≤4 mM of CaCl₂), exogenous Na⁺ ensures the generation of less swollen hydrogels (Fig. 3) and shifts the minimum gel-forming concentration of Ca²⁺ ions toward higher Ca²⁺ concentrations for G-enriched alginates (1.4 and 1.6 mM of CaCl₂ in S-alginate and LVG, respectively). Thus, when a high Na⁺/Ca²⁺ ratio exists, Na⁺ could inhibit some G-enriched sequences in the primary alginate structure from being crosslinked by Ca²⁺, but does not alter the fundamental principal of hydrogel formation with Ca²⁺ (CaCl₂≤10 mM). Moreover, as we demonstrated below, only strongly swollen Ca-alginate_NaCl hydrogels with the degree of swelling above 66% supported robust neurite outgrowth in vitro.

Mechanical stability of Ca-alginate_NaCl and instability of Ca-alginate hydrogels in buffered physiological saline

Sodium is a constituting ion of every extracellular medium in the body, and it contributes to the high ionic strength of extracellular fluids, including blood plasma and cerebrospinal fluid. Moreover, the pH value in normal blood and cerebrospinal fluid is tightly regulated between 7.35 and 7.45.^49,50 To predict how long different in situ prepared hydrogels can maintain their dimensions in the presence of these physiological fluids in vivo (e.g., as implants into injured spinal cord) and subsequently support neural growth and regeneration, we quantified hydrogel disintegration in a buffered physiological saline. Ca-alginate and Ca-alginate_NaCl hydrogels generated in the presence of 2 mM of CaCl₂ were immersed in buffered physiological saline (150 mM of NaCl, pH 7.4) supplemented with 2 mM of CaCl₂ and the hydrogel mass was regularly quantified during the subsequent 30 days. Ca-alginate hydrogels derived from LVM completely collapsed, while those derived from LVG and S-alginate lost ≥50% of their mass within the first 30 days (Fig. 4a). Surprisingly, analogous Ca-alginate_NaCl hydrogels demonstrated long-term mass stability during the entire observation period (Fig. 4b).

OPEN IN VIEWER

FIG. 4.

Ca-alginate_NaCl hydrogels demonstrate prolonged volume stability in buffered physiological saline. Ca-alginate and Ca-alginate_NaCl hydrogels were generated at pH 5.5 from dry alginate layers in the presence of 2 mM of CaCl₂ (a) in the absence or (b) in the presence of 150 mM of NaCl and were then incubated in buffered physiological saline (150 mM of NaCl at pH 7.4 and 2 mM of CaCl₂) for 1, 10, and 30 days. Alginate layers were derived from 600 μL of a 1% sol of S-alginate, LVG, and LVM. Data are represented as means±standard deviations.

Macroscopic morphology of hydrogels

Rapid binding of calcium ions to alginate produces macroscopically inhomogeneous structures; conversely, slow gelation kinetics, which commonly occur in the presence of components that compete with alginate for Ca²⁺ complexation such as oligoguluronates and EDTA, lead to more homogeneous gels.^12,51 We noticed remarkable differences in the surface profiles between Ca-alginate_NaCl and analogous Ca-alginate hydrogels. Many of the Ca-alginate_NaCl hydrogels had a smooth and uniform surface, whereas Ca-alginate gels had ruffles and creases of irregular shape and pattern (Fig. 5). The differences in gel appearance were independent of alginate type and became more apparent in gels generated with solutions containing ≤3 mM of CaCl₂ (Fig. 5f, g). Thus, at a high Na⁺/Ca²⁺ ratio, Na⁺ could delay Ca²⁺-mediated chain association, which optimizes the spatial organization of the alginate network.

OPEN IN VIEWER

FIG. 5.

Phase-contrast images of Ca-alginate and Ca-alginate_NaCl hydrogels. Hydrogels were generated at pH 7.4 by crosslinking dry alginate layers of S-alginate with 2–10 mM of aqueous CaCl₂ solutions (a–e) in the absence or (f–j) in the presence of 150 mM of NaCl. Alginate layers were derived from 600 μL of a 1% S-alginate sol. The scale bar is 1 cm.

Rheological characterization of Ca-alginate_NaCl hydrogels

We carried out oscillatory shear measurements in a rheometer and quantified the stiffness of the Ca-alginate_NaCl hydrogels. The stiffness of all tested hydrogels increased with higher levels of Ca²⁺ in the crosslinking solution. In addition, LVM-derived gels were softer than analogous S-alginate- and LVG-derived gels (Fig. 6). For example, the average stiffness of S-alginate- or LVM-derived Ca-alginate_NaCl hydrogels formed in the presence of 2, 4, 8, and 10 mM of CaCl₂ were 0.011, 1.470, 3.665, and 19.230 kPa (S-alginate) and 0.025, 0.622, 3.085, and 7.356 kPa (LVM), respectively.

OPEN IN VIEWER

FIG. 6.

Increase in the storage modulus (G′) of a Ca-alginate_NaCl hydrogel correlates with an increase in the concentration of CaCl₂. Hydrogels were generated at pH 5.5 by crosslinking dry alginate layers of (a) S-alginate, (b) LVG, and (c) LVM with (⋄) 2 mM, (□) 4 mM, (●) 6 mM, and (▴)10 mM of aqueous CaCl₂ solution in the presence of 150 mM NaCl. Measurements were performed at 37°C.

Neurons rapidly extend neurite on soft Ca-alginate_NaCl hydrogels

Neurons have an intrinsic capacity to extend neurite on surfaces they can adhere to. Previously, we described that neurons extend neurite on soft alginate hydrogels prepared with 2 mM of Ca²⁺, but gelation with 10 mM of Ca²⁺ resulted in a stiff and nonadhesive hydrogel; this change was attributed to increased tightness of the alginate network. ²⁹ We tested whether Ca-alginate_NaCl hydrogels are still capable of promoting neural adhesion and growth under growth factor-free conditions and without serum. An array of hydrogels were generated from LVM, LVG, and S-alginate; crosslinking was performed in the presence of a constant concentration of NaCl (150 mM) and different concentrations of CaCl₂ (i.e., 2, 4, 6, 8, or 10 mM) at pH 5.5 or 7.4.

As we reported previously, the ability of a soft alginate hydrogel to promote neurite extension did not depend on the cell culture system because two-dimensional (2D) adherent monolayers of neurons and 3D neural spheroids produced identical results. ²⁹ Thus, 3D spheroids were chosen for the cell culture experiments in this study. Spheroids consisted of living cells, which neither accumulated trypan blue dye nor contained fragmented nuclei (data not shown). Three hours after being plated on soft Ca-alginate_NaCl hydrogels (S-alginate, LVG, and LVM gelled with 2 mM of CaCl₂), multiple neurite extended in all directions and away from the center of the spheroids. The spheroid-derived neurite expressed the neural markers MAP2 and β-tubulin III (Fig. 7a, e, i and b, f, j) and developed clearly defined growth cones (Supplementary Fig. S1; Supplementary Data are available online at www.liebertpub.com/tec). Growth cone migration elongates neurite, while growth cone splitting creates a branch point resulting in two or more branches. Neurite elongation prevailed over neurite splitting on soft Ca-alginate_NaCl hydrogels since vast majority of neurite (Supplementary Fig. S1a), including MAP2-labeled dendrites (Fig. 7a, e, i) did not branch. Neurite predominantly extended from the area of the spheroid that was in direct contact with the underlying hydrogel. Neurite growth was detected on the surface and in the entire hydrogel volume, gentle shaking could not detach the neurite or the neurite-bearing spheroids. Average distance of neurite extension from spheroids was measured 48 h after plating of spheroids on top of hydrogels and corresponded to 261±120, 277±126, and 312±135 μm on S-alginate, LVG, and LVM Ca-alginate_NaCl hydrogels generated in the presence of 2 mM of CaCl₂, respectively. The proportion of neurite-bearing spheroids on top of S-alginate, LVG, and LVM Ca-alginate_NaCl hydrogels generated in the presence of 2 mM of CaCl₂ was 80%, 83%, and 93%, respectively, but only 18%, 21%, and 22% on analogous hydrogels prepared in the presence of 4 mM of CaCl₂. Gelation with ≥6 mM of CaCl₂ abolished the attachment of spheroids to the hydrogels and, consequently, eliminated neurite extension (Fig. 7c, d, g, h, k, l). Unlike the importance of calcium concentrations in a crosslinking solution, the pH values during alginate gelation (5.5 or 7.4) had no influence on spheroid attachment and neurite growth on Ca-alginate_NaCl hydrogels during subsequent cultivation. Spheroid-derived neurite continued to elongate and formed a 3D meshwork by 2 weeks in vitro and express MAP2 and Tau-1 that label dendrites and axons, respectively (Supplementary Fig. S1b, c).

OPEN IN VIEWER

FIG. 7.

Soft Ca-alginate_NaCl hydrogels support neurite outgrowth from primary cortical neurons. Fluorescence images show (a, c, e, g, i, k) MAP2 and (b, d, f, h, j, l) β-tubulin III antibody labeling of rat neural three-dimensional spheroids cultured on top of Ca-alginate_NaCl hydrogels for 48 h. Ca-alginate_NaCl hydrogels were generated at pH 7.4 by crosslinking of dry layers of (a–d) S-alginate, (e–h) LVG, or (i–l) LVM alginates with 2 mM or 6 mM of aqueous CaCl₂ in the presence of 150 mM of NaCl. The β-tubulin III antibody labels neuronal bodies and all of the neurite, whereas the MAP2 antibody labels neuronal bodies and dendrites. Fluorescence images in the panels (a, e, i) are maximum intensity projection of confocal images obtained from 12-μm-thick Z-stacks. Secondary antibodies were labeled with the fluorescence dye Alexa Fluor-488. The scale bar is 100 μm.

In agreement with data in the literature, we observed that multiple nestin- and GFAP-labeled cells migrated from the spheroids on poly-l-lysine-coated surfaces to form an adherent 2D cell monolayer (not shown). In contrast, spheroids preserved their 3D shape on alginate hydrogels, and only neurons extended neurite on the top of the soft hydrogel (Fig. 7a, b, e, f, i, j). Unlike neurons, GFAP-positive cells did not extend their processes outside spheroids neither on soft nor on stiff Ca-alginate_NaCl hydrogels.

Since soft Ca-alginate_NaCl hydrogels can be shaped to take any form or size and do not collapse during gentle mechanical treatment, they can be implanted in vivo by direct placing of a piece of hydrogel into a tissue lesion (e.g., lesion formed after traumatic spinal cord or brain injury). Biocompatible, soft Ca-alginate_NaCl hydrogels promote neural growth and could be tested in the future as neural bridges in vivo.