Hyaluronic Acid-Based Bioink for Anisotropic Neural Tissue Cryobioprinting

Abstract

In this study, we present the development of a cryobioink designed to fabricate anisotropic scaffolds that support both neural and muscle cell alignment. Given the critical role of cellular organization in nerve fibers and neuromuscular junctions, we employed a vertical cryobioprinting-enabled ice-templating technique to create scaffolds with aligned microchannels. These channels facilitated cell alignment, which is important in modeling neural and neuromuscular tissues. By integrating hyaluronic acid-methacrylate (HAMA) with gelatin methacryloyl and the necessary cryoprotective agent melezitose, we showcased that the cryobioink could preserve cell viability during freezing/thawing processes, even at low temperatures employed during cryobioprinting. We optimized HAMA concentration to enhance neural cell viability and alignment, and successfully constructed anisotropic scaffolds featuring distinct sections that contained muscle and neural cells, establishing a model for neuromuscular junctions. The resulting models provide a versatile platform for studying nerve fibers and neuromuscular dysfunctions, offering potential advancements in neural regeneration research.

Introduction

The structural organization of cells is crucial to the functions of anisotropic tissues such as the muscle and the nerve fibers (1,2). Considering the numerous dysfunctions associated with nerve fibers (3,4) and the cholinergic synapses in muscle functions (the neuromuscular junctions) (5), researchers have explored various methods to construct reliable models for both nerve fibers (3,6) and neuromuscular junctions (7,8). In both cases, cell alignment is a key factor.

While patterning techniques have allowed for easy control over neural network structures in planar models—providing valuable insights into the relationship between structure and functions (9,10)—the transition to three-dimensional (3D) models poses greater challenges in structural control and cellular alignment (11). Among the various methods for creating anisotropic structures that support cell alignment in 3D, ice-templating stands out as a cost-effective and promising approach (12,13). Unidirectional freezing based on ice-templating, or freeze-casting, involves exposing one side of an insulated solution to a steep cooling gradient, creating aligned microchannels within the scaffold as the solution solidifies (14). These microchannels facilitate axonal growth and alignment (15,16). Accordingly, ice-templating has been broadly used to align neural cells (6,17–19). However, in these cases, cells are seeded into the scaffolds after their fabrication to avoid damage during freezing and thawing. This post-seeding approach limits control over cellular arrangements and prevents the precise positioning of multiple cell types within specific regions of the anisotropic scaffold.

Recently, ice-templating has been combined with 3D bioprinting, with the bioinks optimized to maintain the viability of embedded cells during the fabrication procedure (20–25). Nonetheless, preserving cell viability during cryobioprinting at temperatures lower than −15 °C remains a challenge due to osmotic pressure and ice-crystallization (20,26). Notably, the freezing temperature directly affects the freezing height, which influences the size of the resulting anisotropic scaffolds produced (21). There is also a need for a cryobioink formulation that better-mimics the properties of neural tissue. The most common cryoprotective agent (CPA) used for minimizing ice crystal-formation during the freezing and thawing processes is dimethyl sulfoxide (DMSO) (27), although osmotic damage during such processes can be as well prevented by using higher-molecular-weight CPAs, like disaccharides and trisaccharides (20,28).

Hyaluronic acid (HA), a linear polysaccharide found throughout the nervous system, is one of the most physiologically relevant extracellular matrix (ECM) components (29,30). With its ability to mimic the brain’s ECM, HA is widely utilized in neural regeneration research (31–33). HA-based hydrogels support dendrite and axon growth, promote synapse-formation, and enhance electrical activities, aiding in the repair of central nervous system (CNS) injuries (34). HA also has broad applications in the peripheral nervous system (PNS), where it can reduce scar-formation, promote fibrin matrix-development, and facilitate regeneration of the peripheral rat sciatic nerve (35–37), thereby improving nerve growth, conduction velocity, and myelination.

To harness these properties, we synthesized HA-methacrylate (HAMA) and combined it with gelatin methacryloyl (GelMA) and melezitose to create a cryobioink suitable for neuronal tissue engineering applications. By optimizing the HAMA concentration, we improved neural cell viability and achieved unidirectional cell alignment within a two-segmented anisotropic scaffolding system containing both muscle and neural cells (Figure 1).

Figure 1. The cryobioprinting process.

Schematics showing the main steps to build the neuromuscular junction model using cryobiofabrication.

Our study presents an enabling approach in which a bioink with pre-embedded neural cells is frozen to achieve favorable neural cell viability and alignment. Notably, this work represents the first use of cryobiofabrication to simultaneously align neural and muscle cells, showcasing the cryobioink’s versatility in producing multi-section filamentous structures using a multi-material cryobioprinter. We observed acetylcholine receptor-clustering at the muscle-neural interface, indicating the formation of functional neuromuscular junctions. This finding highlights the construct’s potential as a robust platform for studying neuromuscular junction functions and dysfunctions. Combined with ice-templating and cryobioprinting techniques, this formulation offers significant promise for developing aligned nerve fibers and neuromuscular junction models, with important implications for disease modeling and tissue engineering.

Materials and Methods

GelMA-synthesis

Unless otherwise mentioned, all materials were purchased from Sigma-Aldrich (USA). GelMA was synthesized following our previously published protocol (38,39). Briefly, type-A gelatin from cold-water fish skin (10 g, Mw = 60 kDa) was dissolved in 100 mL of Dulbecco’s phosphate-buffered saline (DPBS), stirring at 50 °C for 1 h to dissolve the gelatin completely. 8 mL of methacrylic anhydride was added drop by drop using a syringe pump, and the emulsion was stirred for 2.5 h at 50 °C. After diluting twice in DPBS, the mixed solution was dialyzed in distilled water at 40 °C for 7 days using a dialysis membrane (Mw cut off (MWCO): 12–14 kDa, Spectrum Chemical, USA), changing the water every 12 h. The dialyzed GelMA was filtrated using a 0.22-μm Stericup-GP sterile vacuum filtration system (Millipore, USA), frozen, lyophilized, and stored at 4 °C until used.

HAMA-synthesis

HAMA was synthesized following the protocol described previously (40). Briefly, 1 g of sodium hyaluronate (1,500 kDa, HAworks, USA) was dissolved in 100 mL of deionized water at 4 °C. 1 mL of methacrylic anhydride was added to the sodium hyaluronate solution, and the reaction was performed at 4 °C for 24 h, keeping the pH between 8 and 10 using 1-M sodium hydroxide. The mixture was dialyzed using a 12−14-kDa dialysis membrane in 4 °C deionized water for 5 days, changing the water every 12 h. The dialyzed HAMA was filtrated using a 0.22-μm Stericup-GP sterile vacuum filtration system, frozen, lyophilized, and stored at 4 °C until used.

Cell culture

Three types of cells, namely C2C12 mouse skeletal myoblasts (CRL-1772, ATCC, USA), NG108–15 neuroblastoma cells (HB-12317, ATCC), and SH-SY5Y neuroblastoma cells (CRL-2266, ATCC), were cultured separately. The C2C12 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Thermo Fisher Scientific, USA) supplemented with 1% (v/v) antibiotic-antimycotic (anti-anti, Thermo Fisher Scientific) and 10% (v/v) fetal bovine serum (FBS, Thermo Fisher Scientific). The NG108–15 cells were cultured in DMEM supplemented with 1% (v/v) hypoxanthine, aminopterin, thymidine (HAT, ATCC), 1% (v/v) penicillin/streptomycin (P/S, Thermo Fisher Scientific), and 10% (v/v) FBS. The SH-SY5Y cells were cultured in DMEM supplemented with Glutamax (Glutamax medium, Gibco) along with 1% (v/v) P/S and 10% (v/v) FBS.

The cells were cultured in a 5% CO₂, humidified incubator at 37 °C. The media were changed every other day, and the cells were passaged with 0.05% trypsin-ethylenediaminetetraacetic acid (trypsin-EDTA, Thermo Fisher Scientific) when a confluency of approximately 80% was achieved.

Cryobioink formulation

The cryobioinks were formulated by dissolving GelMA and HAMA at the desired concentrations in Hanks’ balanced salt solution (HBSS, Sigma-Aldrich)/FBS (1:1 v:v) solution, supplemented with 12% (w/v) of D-melezitose hydrate and 0.3% (w/v) of lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP, Advanced Biomatrix, USA). To study the effect of DMSO, the corresponding cryobioinks were prepared with a further addition of 10% (v/v) DMSO. Finally, we evaluated the cryoprotective effect of poly(ethylene glycol) (PEG) by incorporating a 10% (w/v) concentration into the bioink.

Mold freeze-casting

The cryobioink formulation was optimized using a casting method. A cylinder of Clear Resin (Formlabs, USA) with 2 mm of diameter and 10 mm of height was designed with PreForm software (Formlabs), printed with Form 3B+ printer (Formlabs), and used to produce the negative mold with polydimethylsiloxane (PDMS, Ellsworth). The mold was attached to a customized freezing plate, and the cryobioink was added from the top. The temperature of the freezing plate was controlled by adjusting the input voltage. Once the infilled cryobioink was fully frozen (after few minutes), the sample was exposed to UV light (time of exposure: 30 s; power density: 16.5 W cm⁻², S2000 ELITE, Excelitas, USA) to crosslink the polymer solution. It was then placed in the culture medium, which thawed within a few seconds, resulting in the anisotropic hydrogel scaffold.

Mechanical tests

The samples were prepared using a cylinder PDMS mold with a diameter of 6 mm and a height of 10 mm and then assessed for their compressive mechanical properties using a Low-Force Testing System (3400 Series, Instron, USA). The data acquired were analyzed using Prism (GraphPad Software, USA).

Swelling tests

The hydrogel was prepared using the cryobioink with GelMA and HAMA concentrations of 5% (w/v) and 0.25% (w/v), respectively. To obtain a stable fluorescence, 10% (w/w) of the GelMA used was conjugated with rhodamine B following a standard protocol (41). The pillars were fabricated using PDMS mold as described above, and both frozen and unfrozen samples were maintained in DPBS. Using a fluorescence microscope (Zeiss, Germany), the images were acquired at 0, 0.5, 3, and 6 h post-fabrication.

Degradation assay

Two cryobioinks were prepared for the degradation assay, one using only GelMA (5%, w/v) and the other one using both GelMA (5%, w/v) and HAMA (0.25%, w/v). The frozen and unfrozen samples were prepared using the PDMS mold and kept in DPBS at 37 °C for the experimental duration. The samples were weighed using a standard laboratory microbalance once the excess liquids were removed from their surfaces.

Channel size measurements

The samples were prepared as described for the swelling test. The images were acquired using a confocal fluorescence microscope (LSM880, Zeiss) and analyzed using ImageJ (National Institutes of Health, USA).

Cryobiofabrication and culturing

The samples were prepared using the technique described in the mold freeze-casting section. After detaching from the flask, the cells were centrifuged to remove the supernatant, and the pellet was resuspended in the cryobioink with a concentration of 5–6×10⁶ cells mL⁻¹. The day after cryobiofabrication, the medium was entirely replaced and subsequently changed every other day.

The C2C12 samples were maintained in DMEM with 10% (v/v) FBS and 1% (v/v) anti-anti, NG108–15 samples in DMEM supplemented with 1% (v/v) HAT, 1% (v/v) P/S, and 10% (v/v) FBS, while SH-SY5Y sa,ples in Glutamax medium supplemented with 1% (v/v) P/S and 10% (v/v) FBS. The co-cultured C2C12/SH-SYS5 samples were maintained in Glutamax medium supplemented with 1% (v/v) P/S and 10% (v/v) FBS.

The differentiation of the SH-SYS5 samples was started between 10 and 14 days post cryobiofabrication in Glutamax medium supplemented with 1% (v/v) FBS, 1-μM retinoic acid, and 1% (v/v) P/S. The differentiation of the C2C12 samples was started between 10 and 14 days post cryobiofabrication in DMEM with 2% (v/v) horse serum and 1% (v/v) anti-anti. The differentiation of the segmented C2C12/SH-SY5Y samples was started between 10 and 14 days post cryobiofabrication in Glutamax medium supplemented with 1% (v/v) FBS, 1-μM retinoic acid, and 1% (v/v) P/S.

Cell viability assay

To assess cell viability, the samples were stained with calcein-AM at 0.5 μL mL⁻¹ and ethidium homodimer-1 at 2 μL mL⁻¹ (Thermo Fisher Scientific). Unless otherwise specified, all cell viability tests were conducted at a cooling plate temperature of −30 °C. Images were captured using a Nikon Eclipse-Ti inverted fluorescence microscope (Japan), and cell viability was quantified by counting live and dead cells with ImageJ. Each study group was tested in triplicate at a minimum.

Immunostaining

After 10–14 days in the differentiation medium, the cells were fixed in 4% paraformaldehyde in DPBS for 15 min at room temperature. The samples were kept overnight at 4 °C in 1% (v/v) bovine serum albumin (BSA) supplemented with 0.2% (v/v) Triton X-100 to permeabilize the membrane and block the non-specific binding. Next, the samples were incubated overnight at 4 °C with primary antibodies diluted in the blocking buffer solution, followed by three washes with DPBS and subsequent exposure to secondary antibodies. Rabbit anti-β-tubulin (monoclonal, Synaptic System, Germany) was used as the primary antibody for staining the neural processes in SH-SY5Y cells at a dilution of 1:200, while mouse anti-fast myosin skeletal heavy chain antibody (Abcam, USA) was used to identify muscle development, with nuclei counterstained using 4’,6-diamidino-2-phenylindole (DAPI). Goat anti-rabbit AlexaFluor 594 and goat anti-mouse AlexaFluor 488 were used as secondary antibodies (1:400 dilution, Life Technologies, USA). Staining using α-Bungarotoxin antibody conjugated with Alexa Flour 555 (BTX, Thermo Fisher Scientific) was performed in order to determine the presence of acetylcholine receptors (AChR’s) in myotubes. The constructs were examined and imaged using a fluorescence microscope (Nikon). When required, a confocal fluorescence microscope (LSM880, Zeiss) was also utilized for additional imaging. Using the formula for corrected total cell fluorescence (CTCF), defined as CTCF = integrated density - (area of selected cell × mean fluorescence of background readings), as outlined in (42), the fluorescence intensity of BTX was calculated using Excel (Microsoft, USA).

Coaxial nozzle fabrication

For the multimaterial printing, we fabricated coaxial nozzles, as described previously by us (43). Briefly, the 16G needle was used as the shell, and the 20G for the core.

Cryobioprinting

A laboratory-designed mechanical (motor-driven) multimaterial extrusion bioprinter was used for 3D vertical cryobioprinting. To enable cryobioprinting, a customized freezing plate was used, and the temperature was adjusted by changing the input voltage (20,21). The samples were crosslinked using UV light (time of exposure: 30 s; power density: 16.5 W cm⁻²) immediately after cryobioprinting.

Statistical analyses

Statistical analyses were conducted using Prism. All data are presented as means ± standard deviations. For comparisons between two groups, an unpaired t-test was applied, while one-way analysis of variance (ANOVA) was used for comparisons involving three or more groups. Two-way ANOVA was employed to evaluate the simultaneous effects of two factors. Differences were considered statistically significant when p < 0.05 (*).

Results and discussion

Our initial hypothesis posits that HA, as a polysaccharide, may contribute to preserving cellular viability during the cryobiofabrication process. Literature has shown that saccharides can mitigate osmotic damages during the freezing and thawing processes (20,28), and polysaccharides aid in reducing osmotic stresses and inhibiting ice-recrystallization, both of which are critical for maintaining cell viability during cryopreservation (44). Specifically, polysaccharides in cryoprotective media stabilize cell membranes, prevent disruption of bilayer structures, and help regulate osmotic pressures, all essential for cell-survival during freezing (45).

To evaluate the impact of HAMA on preserving the viability of embedded neural cells during unidirectional freeze-casting, we incorporated different concentrations of HAMA into a GelMA-based bioink. GelMA, derived from gelatin, has a protein-based composition rather than a polysaccharide-based one. As a result, it lacks the same capacity to maintain cell viability under extreme conditions, such as at temperatures below −15 °C. To demonstrate the role of HAMA in preserving cell viability, we designed a control bioink formulation identical to the experimental bioink, differing only in the absence of HAMA. In this control formulation, GelMA was the sole polymer, allowing us to evaluate and highlight the protective role of HA under the critical freezing conditions studied.

The scaffolds were cryobiofabricated using a cooling plate set to −30 °C. Figure 2a presents live and dead staining images of NG108–15 cells across various cryobioink formulations at different time points post-cryobiofabrication. The quantification results of cell viability demonstrated that on day 1, the cell viability values increased with higher concentrations of HAMA (Figure 2b), reaching four times the viability for the cryobioink containing only GelMA when the formulation had 1% HAMA. This enhanced cell viability at elevated HAMA concentrations could be attributed to the polysaccharide nature of HA.

Figure 2. Effect of HAMA concentration on cell viability and scaffold mechanical properties.

a) Fluorescence micrographs showing the live (green)/dead (red) staining of encapsulated NG108–15 cells in different cryobioink formulations on days 1, 3, and 7 of culture post-cryobiofabrication. b) Corresponding quantification results of cell viability values with different HAMA concentrations (one-way ANOVA,* p<0.05, ** p<0.01, *** p<0.001, n≥3). c) Quantifications of compressive moduli of the frozen and unfrozen samples, produced with different cryobioink formulations (two-way ANOVA, ** p<0.01, *** p<0.001, **** p<0.0001, n≥4).

Over the rest of the culture period, the cell viability values increased steadily, reaching approximately 58%, 81%, 83%, and 66% on day 7 for cryobioinks containing 0%, 0.25%, 0.5%, and 1% HAMA, respectively. It may seem counterintuitive that cellular vitality with 1% HAMA was significantly lower than those of the 0.25% and 0.5% HAMA conditions on day 7. However, while this phenomenon was not extensively explored in this study, it could be attributed to two possible reasons. First, higher HAMA concentrations would enhance the scaffold’s mechanical properties (Figure 2c), which diverge from those of the native neuronal tissues (38). Neuronal cells are susceptible to discrepancies in stiffness (46). Thus, it is reasonable to argue that although a higher concentration of HAMA may better-preserve the cells during freezing and thawing, the neuronal cells would display improved behaviors in matrices with stiffness values more closely aligned with those of the native nerve tissues, such as the constructs made with 0.25% and 0.5% HAMA.

Second, even though HAMA is a promising polymer for biological applications, the non-adhesive nature of HA limits its use in applications requiring cell spreading (40). Literature extensively documents cases where the presence of HA complicates neuronal adhesion and growth. For example, it was reported that HA led to reduced adhesion and survival of dissociated embryonic spinal cord cells, with the few attached neurons exhibiting limited process-outgrowth (47). Similar challenges in neuronal cell adhesion were also observed by another study (48). As such, hydrogels containing HA often require the addition of supplemental cell-adhesive peptides (49). Indeed, reports showed that laminin and collagen were necessary to enhance neurite extension from neonatal rat dorsal root ganglia in HA hydrogels (50). For this reason, HAMA is often combined with GelMA, which introduces cell-interactive functional groups to the HA hydrogel matrix, improving the cell adhesion properties of the resulting polymer (51,52). Therefore, in our case, increasing the concentration of HAMA increases the non-adhesive contribution of the hybrid polymer.

In summary, although 1% HAMA preserves cell viability more effectively during the freezing and thawing processes, it is less effective than 0.25% and 0.5% HAMA in promoting neural cell attachment and spreading. This is probably due to its mechanical properties differ more from nervous tissue and an increase in the non-adhesive component.

After establishing the optimal range of HAMA concentration, we incorporated 12% melezitose, a trisaccharide previously used to preserve cell viability during freezing and thawing steps in our recently developed cryobioprinting technology (20–23). Figure 3a and 3b illustrate cell viability results, demonstrating that adding the trisaccharide melezitose indeed contributed to preserving cell viability and spreading in the days of culture post-cryobiofabrication. By day 7, the percentage of live cells in cryobioink formulations containing melezitose reached approximately 94%, compared to around 84% in formulations comprising only GelMA and HAMA. Moreover, from Figure 3a, we could appreciate the neural cell alignment achieved on day 7 of culture due to the vertically unidirectional pores present within the constructs.

Figure 3. Effect of melezitose on cell viability.

a) Fluorescence micrographs showing the live (green)/dead (red) staining of encapsulated NG108–15 cells in different cryobioink formulations on days 1, 3, and 7 of culture post-cryobiofabrication. b) Corresponding quantification results of cell viability values with and without 12% of melezitose (one-way ANOVA,* p<0.05, ** p<0.01, *** p<0.001, n≥3). c) Fluorescence micrographs showing the live (green)/dead (red) staining of encapsulated NG108–15 cells in different cryobioink formulations on day 7 of culture post-cryobiofabrication. d) Corresponding quantification results of cell viability values with different HAMA concentrations (one-way ANOVA,* p<0.05, ** p<0.01, n≥3).

Finally, we examined whether melezitose alone and melezitose combined with 1% HAMA could provide the same level of protection as the two optimal cryobioink formulations identified previously. Figures 3c and 3d display the cell viability results at 7 days of culture post-cryobiofabrication for GelMA with melezitose without HAMA and GelMA with melezitose and different HAMA concentrations. These results clearly highlighted the significance of HAMA, confirming that the two best cryobioink formulations were 5% GelMA + 12% melezitose + 0.25% or 0.5% HAMA. Consistent with the findings in Figure 2b, cell viability values in the 0.25% and 0.5% HAMA groups remained comparable.

Considering that DMSO is commonly used as a cryoprotective agent, we also evaluated its contribution to prevent cell death within our cryobioink formulations. However, as shown in Figure S1a and S1b, at 7 days of culture post-cryobiofabrication, the best results still remained for the conditions without DMSO in our experiments. While DMSO effectively minimizes ice crystal-formation (27), it can also have a cytotoxic effect that compromises cell viability (53), particularly when cells are incubated in a non-frozen state for extended periods (54). This trend was further supported by the data from the samples without HAMA, as illustrated in Figure S1c and S1d.

As shown in Figures 4a and 4b, the freezing heights of the samples were significantly affected by the cooling plate temperature and the presence of DMSO. When DMSO was included, the freezing height reached approximately 2.3 mm and 3 mm, respectively, for a cooling plate temperatures of −15 and −30 °C. However, without DMSO, the freezing height significantly increased, reaching approximately 3.5 mm and 7 mm for the same temperatures. This difference could be explained by DMSO’s dual role; while it reduces ice crystal-formation, it also slows down the freezing process (20,54).

Figure 4. Effect of temperature and DMSO on freezing height and microchannels size.

a) Hydrogel constructs cryobiofabricated with different cryobioink formulations and cooling plate temperatures. White color indicates the frozen segment, while blue-dyed segment was unfrozen. b) Quantifications of frozen heights of the samples with and without DMSO at different cooling plate temperatures (two-way ANOVA, * p<0.05, **** p<0.0001, n≥4). c) Fluorescence micrographs showing the boundary between the frozen and unfrozen segments of the sample (GelMA 5% HAMA 0.25% melezitose 12%, cooling plate temperature: −30 °C). GelMA was chemically labeled with rhodamine B to aid visualization. d) Calcein-stained NG108–15 cells at day 3 of culture post-cryobiofabrication. e) Fluorescence micrographs of the microchannels in the samples produced with different cryobioink formulations and cooling plate temperatures. GelMA was chemically labeled with rhodamine B to aid visualization. f) Quantifications of the corresponding microchannel sizes (two-way ANOVA, ns p>0.05, ** p<0.01, *** p<0.001, n≥6).

The freezing height is a critical parameter as it dictates the height of the anisotropic section of the scaffold, which is the region where the microchannels form. Figure 4c illustrates the boundary between the frozen and unfrozen portions of the scaffold. In the frozen area, aligned microchannels were present, extending from the bottom to the top as vertical ice columns form, causing phase-separation. This process halts when the cryobioink can no longer freeze. The effect of these microchannels on directing cellular displacement is shown in Figure 4d. In the frozen segment, the cells were aligned along the direction of the microchannels, which guided their growth. In contrast, the cells were randomly distributed as isolated spots in the unfrozen part, with no noticeable spreading or alignment observed. These results clearly highlight the importance of the freezing height, especially for the fabrication of a multi-segmented models, such as one for the neuromuscular junction (as illustrated schematically in Figure 1), where the formation of microchannels must also reach the upper section to ensure proper cell alignment.

The different cooling temperatures affect not only the freezing height and cell viability but also the diameter of the microchannels formed during the freezing process. We measured the diameters at the two different cooling plate temperatures, both with and without DMSO (Figure 4e). The results in Figure 4f revealed that the microchannels were larger at the higher cooling plate temperature and in formulations without DMSO. This observation could be attributed to the varying rates and sizes of ice crystal-propagation under these conditions, which are influenced by both temperature and the presence of DMSO. The ideal microchannel size should closely match the soma sizes of the neural cells, which range between 15 and 25 μm (55), to ensure optimal cellular alignment. If the microchannels are too small, the cells will not have enough space to grow and proliferate. On the other hand, if the microchannels are too large, the cells will not be sufficiently confined to encourage unidirectional growth, leading to clustering as the cells stack on top of each other. As demonstrated in Figure 4f, when a cooling plate temperature of −30 °C was used without DMSO, the average microchannel size was approximately 28 μm, an appropriate size for promoting neural cell alignment.

However, if achieving a compact cellular alignment is not necessary, or even a specific freezing height, it is possible to achieve high cell viability as early as day 1 post-cryobiofabrication using a cooling plate temperature of −15 °C (Figure S1e and S1f). The higher cell viability observed on the 1^st day with a cooling plate temperature of −15 °C, compared to −30 °C, can be explained by the fact that temperatures between −15 °C and −60 °C are particularly harmful to cells (26).

Previously, we demonstrated that incorporating DMSO into our bioink resulted in a time-dependent decrease in cell viability, likely due to its cytotoxic effects. Although DMSO is one of the most commonly used CPAs, alternative compounds with improved biocompatibility, such as PEG, have been proposed in the literature. PEG has been shown to enhance post-thaw cell-recovery by mitigating ice-formation and osmotic stress (56,57). Moreover, recent studies have indicated that the cryoprotective efficacy of PEG is highly dependent on its molecular weight (58).

In our study, PEG 200-Da was selected as it demonstrated superior performance compared to higher-molecular-weight PEG (58). Cell viability analyses performed at 7 days post-cryobiofabrication revealed that bioink supplemented with 10% (w/v) PEG exhibited higher cell viability relative to bioink containing 10% (v/v) DMSO (Figure S1g and S1h). However, the viability remained comparable to that of the control bioink (GelMA 5% + HAMA 0.25% + melezitose 12% and without PEG), suggesting that its incorporation did not provide a significant advantage in this specific application.

Based on the observations made so far, it is evident that the optimal cryobioink composition among all tested conditions was GelMA 5% + HAMA 0.25% + melezitose 12%, where cell viability could be maintained at approximately 95% at 7 days of culture post-cryobiofabrication. Nonetheless, using a HAMA concentration of up to 0.5% was also feasible, yielding results comparable to those of HAMA at 0.25% despite slightly reduced viability. These findings highlight the crucial role of HA in preserving cellular viability during cryofabrication, emphasizing its potential as a key component in the fabrication of anisotropic scaffolds using this technique.

Given its critical role in neural tissue, HA has been widely utilized in neural tissue engineering for various applications, including reducing cell-mediated hydrogel degradation (59), enhancing mechanical properties (60), protecting neural precursor cells from damage by reactive oxygen species (61), improving survival and proliferation rates of neural precursors (62), promoting regeneration (63), and supporting neurite growth, differentiation, and proliferation (64). Additionally, it has been employed in the fabrication of anisotropic scaffolds through unidirectional freezing (65). For example, it was demonstrated that the incorporation of Puramatrix into the HA-based hydrogel resulted in an aligned fibrous structure. However, different from our approach, cells were not present during the fabrication of the anisotropic scaffold, and the resulting cellular alignment was not verified. Similarly, a gelatin and HA-based ink was developed for 3D printing combined with cryogelation, producing scaffolds with interconnected micropores and enhanced mechanical properties (66). However, their procedure, which involves printing on a substrate at 4 °C followed by exposure to −18 °C for 16 hours, does not permit the incorporation of cells during the printing process. This prolonged method requires significantly more time (16 hours versus 3–5 minutes) and necessitates post-fabrication cell seeding, potentially affecting cellular distribution and integration within the scaffolds.

Moreover, considering the microchannel size (Figure 4f), the increased freezing height (Figure 4b), the strong cell alignment observed (Figure 3a), and the high cell viability of nearly 95% after 7 days of culture post-cryobiofabrication (Figure 3b), we confirmed that using a cooling plate temperature of −30 °C would be optimal for our intended purpose. Indeed, although lower temperatures would further increase the freezing height, the current level is already sufficient for working with multiple sections. Moreover, lower temperatures could reduce the size of the microchannels below that of the neuronal soma, thereby hindering proper cell proliferation and alignment.

Animal models have been invaluable for studying various diseases, but significant differences exist in the molecular mechanisms between species, such as mice and humans, even when similar disease phenotypes are observed (67). A key limitation is the inability of animal models to fully replicate the complexity of human physiology, diseases, and disease-progression (68,69), particularly in neurodegenerative disorders (70). To overcome the challenges posed by the species barrier, after optimizing the cryobioink using one of the most common animal cell lines (NG 108–15), we also adopted the two most promising formulations to the human-derived SH-SY5Y cell line.

Given that the human cell line SH-SY5Y requires more time to proliferate and align than NG 108–15, we extended the cell viability assessment to 14 days. Figure 5a presents live and dead staining images for up to 14 days of culture post-cryobiofabrication, while the quantitative data in Figure 5b showed no significant difference between the two formulations for this duration evaluated. To assess alignment efficiencies, we measured the alignment errors relative to the longitudinal axes of the anisotropic scaffolds (Figure 5c). In both formulations, the average alignment errors were approximately 3 degrees. After achieving aligned organization, the cells were differentiated into neurons using the differentiation medium. Once differentiated, the samples were fixed and immunostained with β-tubulin to visualize dendritic arborization. As shown in Figure 5d, the differentiated neuron-like cells exhibited neuronal processes aligned with the microchannels.

Figure 5. Viability, alignment, and differentiation of neural cells.

a) Fluorescence micrographs showing live (green)/dead (red) staining of encapsulated SH-SY5Y cells in different cryobioink formulations on days 1, 3, 7, 10, and 14 of culture post-vertical cryobiofabrication. b) Corresponding quantification results of cell viability values (unpaired t-test, ns p>0.05, n≥7). c) Quantifications of alignment errors relative to the longitudinal axes of the scaffolds on day 14 of culture post-cryobiofabrication (unpaired t-test, ns p>0.05, n≥7). d) Fluorescence micrograph showing SH-SY5Y cells embedded in the anisotropic scaffold containing 0.25% HAMA, cultured for 14 days in the differentiation medium, immunostained for β-tubulin (red), with nuclei counterstained with DAPI (blue).

Although this study focuses on neural tissue engineering, the bioink formulation and fabrication technique can also be applied to other tissue types. As mentioned previously, we chose HA for two reasons: first, to achieve a biomimetic approach, given that it is one of the major components of the ECM in the neural tissue; second, for its polysaccharide nature, which is particularly advantageous for preserving cell viability in this application. This aspect is independent of the cell type, as it is more related to the intrinsic properties of HA than to its interaction with specific cell types. Regarding the biomimetic aspect, HA is also found in other tissues where cellular alignment is crucial, such as in vascular (50,71), tendon (72,73), corneal (74,75), and cardiac (76,77) tissues. This versatility underscores its potential applicability in being used as scaffolds designed for these and other alignment-critical applications.

To better-understand the stability of the constructs, we assessed the degradation of both frozen and unfrozen samples, considering the contribution of HAMA. As shown in Figure S2a, adding 0.25% HAMA significantly improved the hydrogel’s resistance to degradation compared to samples without HAMA. This effect was most pronounced in the frozen samples. Without HAMA, the non-frozen hydrogels were completely degraded after 10 days in DPBS at 37 °C. However, for the frozen hydrogels, degradation was only ~20% even after 2 months. There was no significant difference between the frozen samples with and without 0.25% HAMA. These results suggest that increasing the HAMA concentration or freezing the samples promotes the sample’s resistance to degradation. The stability of HA-based biomaterials is largely determined by the efficiency of HA dissolution, which is critical for preserving their structural integrity (78). However, HAMA hydrogels have demonstrated the ability to maintain their structural stability over prolonged periods, exhibiting less degradation under physiological conditions (79,80). This enhanced stability positions HAMA hydrogels as more robust and durable compared to their HA-based counterparts, which are prone to more rapid degradation.

Moreover, it was observed that the samples subjected to the freezing process also exhibited higher resistance to swelling (Figure S2b and S2c). Swelling is a characteristic property of hydrogels, enabling the storage of liquids within the polymer matrices. However, hydrogels often undergo deformations upon immersion in an aqueous environment, leading to partial loss of the predesigned shapes and sizes created during processes such as bioprinting. To this end, the ability of the cryobiofabricated constructs to withstand swelling could be advantageous in preserving their structural fidelity for tissue engineering applications.

As noted earlier, a higher freezing height facilitates the fabrication of anisotropic scaffolds with segments to co-align two or more types of cells, such as a cholinergic synapse responsible for muscle functioning in the neuromuscular junction where the alignments of the muscle and neuronal cells have to coexist. Since neuromuscular junction dysfunction is the basis of several muscle atrophies (5), different methods have been studied to obtain reliable in vitro neuromuscular junction models over the years (7,8,81). Nonetheless, in such a model, cellular alignment is pivotal not only for better-mimicking the in vivo structures but also for achieving corresponding functionality (5,82–84).

Before assessing the co-culture of the two cell types (C2C12 and SH-SY5Y), we assessed the behaviors of C2C12 myoblasts in the two most optimal cryobioink formulations obtained thus far. As shown in Figure 6a and 6b, the C2C12 viability results obtained on different days of culture post-cryobiofabrication were consistent with those observed for the neural cells. The results obtained with C2C12 cells further highlight the crucial role of HAMA. When comparing our findings to those reported in the literature using the same cells (i.e., C2C12) 7 days after cryofabrication, it becomes clear that, despite employing a more cell-damaging temperature in our study (i.e., −30 °C), the outcomes are comparable to those achieved at −15 °C with a bioink lacking HAMA (20). Moreover, Figure 6c shows the well-aligned differentiation of C2C12 cells into multinucleated myotubes within the anisotropic scaffold microchannels.

Figure 6. Viability, alignment, and differentiation of muscle cells.

a) Fluorescence micrographs showing live (green)/dead (red) staining of encapsulated C2C12 cells in different cryobioink formulations on days 1, 3, and 7 of culture post-vertical cryobiofabrication. b) Corresponding quantification results of cell viability values (unpaired t-test, * p<0.05, n≥3). c) Fluorescence micrograph showing myotubes aligned along the microchannels of the anisotropic scaffold cultured for 7 days in differentiation medium, immunostained with myosin skeletal heavy chain (green), with nuclei counterstained with DAPI (blue).

Accordingly, we cryobiofabricated the two-segment model with C2C12 cells at the bottom and SH-SY5Y cells at the top, as illustrated in the schematic in Figure 1. To evaluate the healthy co-existence of the two cell types, we monitored the viability values of the dual-segmented sample post-cryobiofabrication (Figure 7a). Between 7 and 14 days of culture, both cell types revealed viability ratios of 80–90% (Figure 7b and 7c). Similar results were observed in the dual-segmented co-culture of C2C12 and NG108–15 cells (Figure S3). After 10 additional days of culture, C2C12 and SH-SY5Y cells were successfully co-differentiated into myotubes and neurons. Figure 7d displays an immunofluorescence micrograph where the co-alignment of myotubes (stained with myosin skeletal heavy chain antibody) and neurons (stained with β-tubulin antibody) along the anisotropic scaffold microchannels is clearly visible.

Figure 7. Cryobiofabrication of the neuromuscular junction model.

a) Fluorescence micrographs showing live (green)/dead (red) staining of encapsulated C2C12 and SH-SY5Y cells in multimaterial constructs containing two vertical segments on days 1, 3, 7, 10, and 14 of culture post-vertical cryobiofabrication. b) Corresponding quantification results of cell viability values of the bottom segment (C2C12) (unpaired t-test, * p<0.05, n≥7). c) Corresponding quantification results of cell viability values of the top segment (SH-SY5Y) (unpaired t-test, * p<0.05, n≥5). d) Fluorescence micrograph showing the junction of the two segments immunostained for myotubes (myosin skeletal heavy chain, green), neurons (β-tubulin, red), and nuclei (blue). e) Fluorescence micrograph of a two-segmented sample with neurons and myotubes immunostained for AChRs (BTX, red), myotubes (myosin skeletal heavy chain, green), and nuclei (blue). f) Quantitative CTCF analyses on the BTX fluorescence signals for the samples containing on the myotubes and the two-segmented samples containing the neuromuscular junctions (unpaired t-test, *** p<0.001, n≥4).

Moreover, we investigated the presence of acetylcholine receptors (AChRs) in muscle cells, which are indicators of neuromuscular junction-formation (85). In the co-culture of neurons and muscle cells, to verify the presence of AChR clusters in differentiated C2C12 cells, we employed bungarotoxin (BTX) as a marker for AChR, along with antibody against skeletal myosin heavy chain to highlight the myotubes. Figure 7e clearly shows the presence of AChR clusters (marked in red) in the myotubes (marked in green). These clusters are almost absent in samples with only muscle cells (Figure S5). From the quantitative analyses of the signals conducted via corrected total cell fluorescence (CTCF) measurements (Figure 7f), it could be concluded that there was a significant difference in the formation of AChR clusters in the myotubes when they were co-cultured with neurons, suggesting the presence of neuromuscular junctions.

Furthermore, we tested our cryobioink formulation in cryobioprinting, a process of particular interest due to its numerous advantages over traditional bioprinting techniques. One major limitation of conventional bioprinting is the lack of shelf-ready availability for bioprinted constructs, which often require onsite fabrication at the point of use. This presents significant challenges for clinical applications, as the delicate nature of cells and the time-consuming bioprinting process make onsite fabrication difficult. Cryobioprinting overcomes this barrier by enabling the production of shelf-ready, cell-laden constructs that retain both their viability and functionality (20). This makes them ideal for use in tissue engineering, regenerative medicine, and drug screening, among others.

Another key benefit of cryobioprinting is its ability to print directly along the Z-axis without the need for support structures, streamlining the fabrication process and enhancing the versatility of the technique. During cryobioprinting, rapid freezing enables extrusion along the vertical direction without requiring a dedicated support structure. This capability is often lacking in traditional extrusion bioprinting of soft hydrogels due to insufficient mechanical properties of these materials. In contrast, integrating cryogenic directional freezing with bioprinting allows for overcoming this limitation, as the freezing process provides support for building the vertical structures along the Z-axis (21). Printing along the Z-axis allows us to expose one face of an insulated solution to a steep cooling gradient. The heat-transfer caused by the unidirectional temperature gradient solidifies the solution, forming aligned ice crystals and creating a scaffold with oriented microchannels. This approach preserves the principles of freeze-casting while leveraging the versatility of bioprinting. On the other hand, coaxial nozzles have been demonstrated to extrude, simultaneously or sequentially, at least two different bioinks during a single bioprinting session without needing to change the nozzle. For example, we utilized coaxial nozzles to fabricate models of intestinal villi and hair follicles (86), while we also achieved the creation of vertically heterogeneous, core-shell, or hollow structures with coaxial nozzles combined with cryobioprinting (21).

To this end, we finally designed a coaxial nozzle, illustrated in Figure 8a, highlighting its dual-channel capability. The top image shows the external appearance of the nozzle, while the bottom image provides a closed-up view, emphasizing the concentric configuration that enables the extrusion of two distinct cryobioinks. In Figure 8b, a multimaterial bioprinter setup is depicted, which combined with coaxial nozzles, enabled the vertical cryobioprinting of freestanding filaments with two segments using two different cryobioinks (as illustrated in Figure 1). On the left side of the bioprinter, there is a cooling mechanism connected to the cooling plate to maintain the appropriate cryobioprinting conditions. The temperature of the cooling plate can be readily controlled via an input voltage. A zoomed-in section demonstrates the cryobioprinting process, with a freestanding vertical filament being deposited. The cryobioprinted structure has two distinct segments, as seen in the inset, where the top segment is blue and the bottom is red. In Figure 8c, a fluorescence micrograph shows the boundary between the two segments, where the top section incorporated blue fluorobeads into the cryobioink while the bottom included red fluorobeads. This validates the ability of the coaxial nozzle to achieve precise segmental differentiation. The continuity of the unidirectional microchannels across the two segments were validated in Figure S4. Compared to mold-based freeze-casting, vertical cryobioprinting using coaxial nozzles offers significant versatility, as the method allows for convenient and rapid designs of target structures without the need to create dedicated PDMS molds. For demonstration purposes, it was shown that cryobioprinting also enables the generation of vertical or other freestanding structures with multiple segments (Figure 8d) or segments connected in a branching configuration (Figure 8e), showcasing the versatility of the cryobioprinting method to fabricate complex geometries, such as segments connected at varying angles.

Figure 8. Cryobioprinting using coaxial nozzles.

a) Photographs showing a coaxial nozzle produced using a 16G external nozzle and a 20G for the core one. b) Multimaterial bioprinter setup. c) Fluorescence micrograph showing the boundary between the bottom segment (containing red fluorobeads) and the top segment (containing blue fluorobeads). d) Fluorescence photograph showing a three-segmented filament produced with the multimaterial bioprinter. e) Fluorescence photograph showing a Y-shaped structure produced with the multimaterial bioprinter. In all cases the entireties of the constructs were within the freezing height limit and thus were all frozen.

Conclusions

In this study, we have presented pioneering efforts on alignment of neural cells using the ice-templating technique with pre-embedded cells. To accomplish this, we fine-tuned the cryobioink formulations to minimize cellular damages during the freezing/thawing processes. Our findings revealed the positive impact of HA, a polysaccharide prevalent in the nervous system, on preserving cellular viability in such an application. We demonstrated that nearly 100% cellular viability could be achieved within 1 week of the culture post-cryobiofabrication. Following the optimizations of the cryobioink and characterizations of the anisotropic scaffolds, we established the foundation for developing a proof-of-concept neuromuscular junction model. This model involved the co-alignment of neural and muscle cells, organized into two distinct segments and connected at the interface, where the presence of acetylcholine receptor clusters was confirmed in the myotubes, suggesting the formation of a neuromuscular junction. Subsequently, we showcased the printability of our optimized cryobioinks using the vertical cryobioprinting technique with a multimaterial cryobioprinter for future expanded applications. In conclusion, the formulation of the bioink aims to enhance the fabrication of neural fiber models and neuromuscular junctions, proposing to advance our understanding of their functionality and associated disorders.