Multi-Material Volumetric Additive Manufacturing of Hydrogels using Gelatin as a Sacrificial Network and 3D Suspension Bath

Abstract

Volumetric additive manufacturing (VAM) is an emerging layerless method for the rapid processing of reactive resins into 3D structures, where printing is much faster (seconds) than other lithography and direct ink writing methods (minutes to hours). As a vial of resin rotates in the VAM process, patterned light exposure defines a 3D object and then resin that has not undergone gelation can be washed away. Despite the promise of VAM, there are challenges with the printing of soft hydrogel materials from non-viscous precursors, including into multi-material constructs. To address this, we use sacrificial gelatin to modulate resin viscosity to support the cytocompatible VAM printing of macromers based on poly(ethylene glycol), hyaluronic acid (HA), and polyacrylamide. After printing, gelatin is removed by washing at an elevated temperature. To print multi-material constructs, the gelatin-containing resin is used as a shear-yielding suspension bath (including HA to further modulate bath properties) where an ink can be extruded into the bath to define a multi-material resin that can then be processed with VAM into a defined object. Multi-material constructs of methacrylated HA and gelatin methacrylamide are printed (as proof-of-concept) with encapsulated mesenchymal stromal cells, where the local hydrogel properties define cell spreading behavior with culture.

Graphical Abstract

Gelatin is used to modulate the viscosity of reactive resins to enable their processing into 3D objects with volumetric additive manufacturing (VAM). Gelatin is also used for sequential suspension bath printing and VAM to allow the printing of multi-material 3D objects with spatial patterning of hydrogels and cells. This advance will expand the utility of VAM printing for biomedical applications.

1. Introduction

Additive manufacturing allows for the fabrication of complex three-dimensional (3D) objects that mimic tissue and organ structures.^[1] Volumetric additive manufacturing (VAM) is an emerging layerless additive manufacturing method that allows the rapid processing of photoreactive resins into 3D structures within seconds, which is much faster than alternative methods such as digital light processing (DLP), stereolithography (SLA), or direct ink writing that can take minutes to hours to print comparable structures.^{[2, 3]} VAM printing involves the irradiation of a rotating resin vial with controlled light exposure in specific projections, resulting in the crosslinking of the resin into a 3D object based on whether voxels have undergone gelation. While much work with VAM printing has involved the curing of elastomeric and ceramic resins^[4], there is great interest in the processing of soft hydrogel materials with additive manufacturing, particularly as hydrogels allow the encapsulation of cells for the production of tissues.

Previous work has explored the use of VAM to print viscous hydrogels containing either cells^[5,6,7] or organoids.^[8] However, past studies have been severely limited in the scope of cytocompatible materials that can be used for VAM. This is because, to date, VAM requires resins that are highly viscous or solid, so that the resin remains stationary in order for a dosage of light to be built up in the same projected area for gelation.^[9] Due to this, the most commonly implemented materials for VAM resins have been based on gelatin, primarily gelatin methacrylamide (GelMA) and more recently gelatin (norbornene) (Gel-NB), since gelatin undergoes thermo-reversible gelation depending on the concentration employed.^[10,11] However, as cells respond to their material microenvironment, it is desirable to be able to print a wide range of soft materials to guide cell behavior, including those that do not meet these requisite rheological properties for VAM. Our group and collaborators have previously investigated the use of sacrificial gelatin to modulate the viscosity of bioinks for the printing of a range of soft materials in extrusion bioprinting.^[12] Such an approach may be useful for the processing of new resins with VAM. One group previously used a gelatin matrix to process a low concentration poly(vinyl alcohol) hydrogel with VAM, but did not explore other hydrogel resins with this approach.^[13]

Beyond the printing of soft materials, the fabrication of multi-material structures that mimic features (e.g., mechanics, biochemical signals) of native tissue heterogeneity is also of interest. Multi-material objects have been made via methods such as photo-patterning,^[14] casting,^[15] electrospinning,^[16] and others. Techniques for multi-material processing with additive manufacturing have also emerged over the past decade with methods such as DLP, Filamented Light (Flight) Biofabrication, extrusion bioprinting, and many other methods.^[17–20] With additive manufacturing, there are specific challenges related to the combination of materials within the specific method used. In each of these examples, the final material properties are based on a combination of the properties of individual components (e.g., mechanics, cellularity) to create heterogenous constructs that may be of use for tissue engineering applications.

Despite these advances in multi-material fabrication with additive manufacturing, there are few reports on the printing of multi-material constructs with VAM with advances in this area primarily from Levato and co-workers. In one example, Größbacher et al combined multiple fabrication strategies, including melt electrowriting to first fabricate tubular scaffolds of polycaprolactone that are then placed within resin vials of GelMA and VAM printed to produce composites of hydrogels containing fibrous stabilizing structures.^[21] This work was the first to combine these two fabrication steps in a sequential process, particularly to enhance the mechanics of soft hydrogels. This same group has also introduced various reactive molecules, including proteins and peptides, directly into VAM printed constructs to pattern them in 3D. ^[13,22] Lastly, Ribezzi recently used jammed GelMA microgel-based as a resin for VAM printing, including leveraging the shear-thinning and healing properties as a bath to receive patterned cells suspensions in methylcellulose or gellan gum and poly(ethylene glycol) combinations. ^[23] Here, the resolution of printing is defined by the particles and challenges with light scattering form particulate structures must be addressed.^[23] These are important advances in the formulation of multi-material structures using VAM printing, but still leave room for additional advances in this area. Particularly, methods that allow the printing of a wider range of hydrogel materials would be useful, particularly in combination with suspension printing to allow multi-material patterning anywhere in 3D space.

Here, we propose the combination of VAM with molecular assembly suspension bath printing to overcome many of the challenges with multi-material printing.^[13,21–25] Suspension bath printing is an extrusion-based printing method whereby an ink is extruded into a shear-yielding and self-healing bath.^[1,26–28] The bath material enables soft, complex 3D structures to be printed which would otherwise collapse in air, as well as introduces controlled heterogeneities to a typically homogenous material. Suspension baths are commonly fluid gels,^[29] molecular assemblies,^[30] and granular supports.^[26,31] As described above, there is a single example of where suspension printing has been combined with VAM, but is limited to a single granular support material. ^[23] We have selected engineered molecular assembly baths and inks to investigate the formulation of multi-material resins.

In our approach, we investigate the use of a sacrificial gelatin matrix to (i) enhance the viscosity of resins for VAM printing and (ii) act as a shear-yielding suspension bath for multi-material printing (see setup in Figure S1). We use the gelatin matrix to print a variety of reactive and non-viscous resins and then use incubation at an elevated temperature to allow the release of gelatin from the printed object.^[12] Modification of the suspension bath properties is investigated with high molecular weight HA to allow for enhanced printability, as well as stability of the bath over time. Finally, we print multi-material acellular and cellular materials with this approach. This method is versatile and allows the printing of a diverse range of materials in more complex multi-material constructs, which we hope expands the material compatibility with VAM printing.

2. Results

2.1. Gelatin Enables the Printing of Non-viscous Hydrogel Precursors via VAM

Since VAM requires high viscosity or solid resins, we hypothesized that the thermo-responsive viscoelastic properties of gelatin could be utilized to print low viscosity reactive macromers and monomers that would otherwise not meet the criteria for VAM printing. In this process, (i) low concentrations of hydrogel precursors can be combined with gelatin at 37°C and cooled to room temperature to increase the solution viscosity and formulate a resin appropriate for VAM printing; (ii) the semi-solid sacrificial gelatin network stabilizes these macromers and monomers during the VAM printing process where the resin is exposed to light via a digital micromirror device while rotating; and (iii) the printed object can be removed from resin that has not undergone gelation and the gelatin can be released with elevated temperature (Figure 1a). This process is feasible since gelatin undergoes gelation through physical interactions when the solution is cooled (i.e., 23°C), and then a rapid change in the network structure and softening occurs when the temperature is increased (i.e., 37°C) (Figure 1b).

Figure 1. Volumetric Additive Manufacturing (VAM) printing of reactive macromers.

(a) Schematic of VAM printing of non-viscous reactive macromers using gelatin as a sacrificial material to modulate viscosity. (b) Rheological profiles of a 2 wt/vol% gelatin solution, during gelation at 23ºC (top) and then showing a >100-fold reduction in modulus when the temperature is increased from 23ºC to 37ºC (bottom). (c) Photorheology of various reactive precursors (20 wt/vol% acrylamide with 0.5 wt/vol% BisA, 2 wt/vol% MeHA, 5 vol/vol% PEGDA) at 23ºC where the shaded region represents where the light is turned on. (d) Fluorescent images of VAM-printed cone-shaped objects from the acrylamide monomer and BisA crosslinker and MeHA and PEGDA macromers containing FITC dextran. Scale bars = 1 mm. (e) Photographs of various VAM printed objects (insets indicate print design) using the acrylamide monomer and BisA crosslinker soaked in Lugol’s solution. Scale bars = 1 mm. (f) Photographs of printed “box frame” objects either without (left) or with (right, 2 wt/vol% gelatin)) VAM printing of the acrylamide monomer and BisA crosslinker. Scale bar = 8 mm. (g) μ-CT of “box frame” object VAM printed from the acrylamide monomer and BisA crosslinker. Scale bar = 1 mm.

Using rheological shear time sweeps, we screened a range of gelatin concentrations to identify the lowest concentration that forms an elastic gel (G’>G”) at room temperature, which would allow for the resin to exhibit solid-like properties and then enable easy removal by subsequent heating (Figure S2). Most of the gelatin concentrations resulted in hydrogels that exhibited shear-thinning behavior (reduction in viscosity with increasing shear-rate), shear-yielding (reduction in storage modulus with increased strain %), and somewhat self-healing (reduction and then partial recovery of storage modulus when tested in high and then low strains), except for the 1 wt/vol% formulation, which exhibited more liquid-like properties. Thus, the lowest investigated gelatin concentration that exhibited hydrogel behavior at room temperature was 2 wt/vol%.

Various hydrogel precursors (20 wt/vol% acrylamide with 0.5% bisacrylamide (BisA), 2 wt/vol% methacrylated hyaluronic acid (MeHA, Figure S3), and 5 vol/vol% poly(ethylene glycol) diacrylate (PEGDA)) were separately combined with 2 wt/vol% gelatin to formulate resins. These macromers and monomers were selected to represent a range of materials of interest in the biomaterials field and that do not have requisite viscosity for VAM printing. Photorheology was performed on these materials to assess the reaction kinetics when exposed to visible light, confirming that each of these hydrogel precursors could be rapidly photocrosslinked into elastic gels, which should be suitable formulations for VAM (Figure 1c). As a proof of concept, all three resins were VAM printed into simple cone structures, illustrating that the formulations identified with photorheology were printable (Figure 1d).

With this proof of concept completed across a range of macromers and monomers, it was identified that the opacity of the acrylamide monomers visually changed during VAM printing, likely a combination of the transition of the small monomer to a hydrogel network and PA-gelatin interactions resulted in increased light diffraction. Thus, this resin was used to further show the printing of more complex structures, such as the CU logo, a disc, an octet, and a box frame (Figure 1e). The printing process took tens of seconds for the sample to be visualized, as illustrated with the polyacrylamide box-frame print (Movie S1). The printed structures were soft, but maintained their stability in aqueous environments, even upon agitation (Movie S2). To illustrate the importance of gelatin, printing was performed with the acrylamide and BisA resin and box-frame print pattern; however, without the addition of gelatin, the overall printed structure was lost (Figure 1f). Without the gelatin present, reacting species likely shift and diffuse as the resin vial rotates and during printing, which then disturbs the desired print patterns. The mechanical properties of printed objects could be tailored by either altering the crosslinker concentration or employing post-crosslinking techniques, both of which changed the compressive moduli of printed discs (Figure S4). The 3D printed structure of the printed box-frame structure was also visualized via μCT and dimensions measured with Fiji (Figure 1g, Movie S3). With printing and swelling, the final PA parts were measured to be ~6% larger than the dimensions of the original CAD model. Thus, the inclusion of gelatin within non-printable hydrogel precursor solutions provides an approach for VAM printing that should be amenable to a wide range of hydrogel systems.

Towards using this system for tissue engineering, we next evaluated how cell viability is influenced by inclusion of a gelatin matrix and then processing with VAM printing. To broadly characterize the potential tissues that could be printed using this process, a range of bovine musculoskeletal cells (chondrocytes, mesenchymal fibrochondrocytes (MFCs), and mesenchymal stromal cells (MSCs)) were included in a resin containing 2 wt/vol% MeHA and 2 wt/vol% gelatin and processed via VAM printing into discs. Cell behavior was then monitored over 7 days (Figure 2). Live/Dead staining showed high cell viability (>90%) across all cell types immediately after printing (day 1) and after 1 week of culture (day 7). Importantly, MFCs and MSCs showed some characteristic spreading behavior after 7 days of culture, while chondrocytes remained rounded, suggesting that cell phenotype is retained within this system. Moving forward, this data shows confidence that these materials and processes can be used for cellular applications.

Figure 2. Cell encapsulation during VAM printing.

(a) Live/dead staining (green: viable, red: dead) and (b) quantified viability for chondrocytes, MFCs, and MSCs encapsulated at 3✕10⁶ cells mL⁻¹ in VAM-printed discs comprised of 2 wt/vol% MeHA and 2 wt/vol% gelatin and cultured for 1 or 7 days. Scale bar = 200 μm. Data are reported as mean ± standard deviation (SD), with n = 4. ns = no significance with p = 0.0842 (two-way ANOVA).

2.2. Gelatin is Released from VAM-Printed Constructs

Although gelatin is important during the printing process, it may be desirable to remove the gelatin from the printed objects to better control for desirable biochemical and biophysical properties. As a first step, the printed object was heated to 37°C to remove it from the remaining gelatin-based resin that has not undergone gelation. However, gelatin could still be retained throughout the crosslinked network. Our approach was to further incubate the object at 37°C over 7 days to allow for gelatin to diffuse from the printed object (Figure 3a). Discs composed of each hydrogel were VAM printed with fluorescent gelatin included. The discs were then placed in PBS and placed on a rocker at 37°C for up to 7 days. Each day, the supernatant was removed and replaced with fresh PBS. The supernatant was then analyzed directly for fluorescence (Figure 3b), as well as for protein release via a BCA assay (Figure 3c). These analyses show that a substantial amount of gelatin (~40-75%) is released within the first 24 hours, a majority of the gelatin content (>85%) is released across the hydrogels by day 3, and then nearly all of the gelatin (~100%) is released within a week. It is interesting that the gelatin release was slowest with the MeHA hydrogel, which could relate to interactions between the HA and gelatin and entanglement; yet, release was very similar for the PA and PEGDA hydrogels. In addition to biochemical assays, fluorescent images were taken of each hydrogel construct over a week with confocal microscope imaging. The fluorescent intensity decreased in all hydrogels over time with only a minimal amount remaining in the PEGDA hydrogel at 7 days and negligible signal was retained in MeHA and PA hydrogels after 3 days. The rapid release of gelatin within 72 hours is important in this process and similar to what was observed previously in extrusion-printed hydrogels.^[12]

Figure 3. Release of sacrificial gelatin from VAM printed objects.

(a) Schematic of the release of gelatin from VAM-printed objects over time through an increase in solution temperature to 37ºC. (b) Release profiles (top) and fluorescent images (bottom) of fluorescent gelatin from VAM-printed discs comprised of various reactive precursors (PA, MeHA, PEGDA). Scale bar = 1 mm. (c) Release profiles of gelatin (quantified through a BCA assay) from VAM-printed discs comprised of various reactive precursors (PA, MeHA, PEGDA). (d) Compressive moduli of VAM-printed discs comprised of various reactive precursors (PA, MeHA, PEGDA) initially and after 1 and 7 days of incubation at 37ºC. Data are reported as mean ± SD, with n = 3. ****p<0.0001, ***p<0.001, ns = no significance with p > 0.2 (two-way ANOVA).

Due to the physical network formed by gelatin, the presence and release of gelatin may also influence biophysical properties of printed constructs. When measured, a decrease in the compressive modulus was observed for all constructs from day 0 to day 7 (Figure 3d). This decrease in mechanical properties is directly correlated to the release of gelatin, as can be seen in a larger drop in mechanical properties between day 0 and day 1 when most of the gelatin is released, and a negligible change between day 1 and day 7. While this decrease in mechanical properties is not unexpected with the release of gelatin, such changes should be considered depending on the application. Furthermore, since the gelatin takes several days to release, this may impact the use of this approach for some applications where the presence of gelatin may be limiting. One of the largest advantages to using VAM is the rapid fabrication, so this incubation period for gelatin release can extend fabrication steps. To circumvent this, other additives that may release faster could be explored and used. However, it should be noted that in many applications the presence of gelatin may not affect the general use of the material.

2.3. Engineering a Gelatin-based Suspension Bath

In addition to using gelatin as a sacrificial material for the VAM printing of low viscosity reactive monomers and macromers due to modulating resin viscosity, gelatin has properties that also enable it to be useful as a suspension bath for 3D printing. Suspension bath printing allows the deposition of an ink anywhere in the 3D volume of a suspension bath, due to bath properties of shear-yielding and self-healing.^[1,25–27] This approach is particularly useful for printing soft, low viscosity materials into 3D and as a consequence allows for the multi-material processing of hydrogel materials based on varied ink and bath properties. If both the ink and the bath contain reactive macromers, this can then be processed with VAM to define the overall shape of a multi-material object after the resin that has not reached gelation is removed (Figure 4).

Figure 4. Sequential suspension bath and VAM printing.

The general approach involves (1) suspension bath printing of an ink (reactive precursor green) into a suspension bath (reactive precursor purple) to formulate a multi-material resin containing biopolymers gelatin and hyaluronic acid (blue) to modulate the resin viscosity to support printing, (2) VAM printing (rotating resin bath within an index matching fluid) to define the 3-dimensional crosslinking of the reactive resin based on light pattern through a digital micromirror device (DMD). The printed object can then be removed from the unreacted resin.

To start, it was important to optimize the properties of both the ink and suspension bath (Figure 5a). Initially, we investigated gelatin alone, which exhibited shear-thinning, shear-yielding, and self-healing behaviors (Figure 5b). However, when used for printing, there was poor fidelity to the print pattern (Figure S5), likely due to the high yielding behavior under strain that can be an issue with baths made from polymer assemblies. Further, gelatin alone as a suspension bath requires precise control of temperature to obtain the ideal viscosity for printing, which we observed to limit reproducibility. In an attempt to change the properties of the suspension bath, high molecular weight HA (HMWHA of ~1.5 MDa, 1 wt/vol%) alone was investigated as a suspension bath (Figure 5b). The HMWHA bath had higher storage moduli, likely due to chain entanglement, as well as a reduction in the yielding behavior. However, when used for printing, although the printed pattern was accurate, the printed ink was not stable over time, likely due to the high viscoelastic properties (Figure S5). This would be particularly problematic during VAM printing, as the vial rotates during the print process. To combine the benefits of both materials, gelatin and HMWHA were combined and used as both the ink and bath materials. Rheologically, this formulation met the desired criteria of a suspension bath and possessed a lower yield point that the individual components alone, likely due to disruption of the gelatin physical assembly as well as the HA entanglement (Figure 5b). Further, suspension prints had high fidelity to the design and were also stable over time (Figure S5). Multiple patterns of inks were then successfully printed in this formulation, including lines, curved patterns, and spiral patterns (Figure 5c). Based on these results, this formulation was used for subsequent studies in both the suspension bath and inks used.

Figure 5. Suspension bath printing into gelatin.

(a) Schematic of the suspension bath printing process using a reactive ink. In this case, gelatin and/or HA are used in both the ink and the bath to alter printability. (b) Rheological characterization of the various suspension baths, including 2 wt/vol% gelatin, 1 wt/vol% HMWHA, and the combination of 2 wt/vol% gelatin and 1 wt/vol% HMWHA. Characterizations include (i) shear-thinning behavior by measurement of the viscosity with increasing shear rate, (ii) strain-yielding behavior by measuring the modulus with increase in % strain, and (iii) self-healing behavior by measuring the modulus through 3 step thixotropic sweeps, where the shaded region is 500% strain and the unshaded regions are at 1% strain. The combination of gelatin and HMWHA yields earlier, increasing the printability of the solution. (c) Images of example suspension prints (inset: print path) where both the suspension bath and ink contain 2 wt/vol% gelatin and 1 wt/vol% HMWHA. The bath contains 2 wt/vol% MeHA macromer, and the ink contains 2 wt/vol% GelMA macromer and 0.05 wt/vol% fluorescent dextran. Scale bar = 1 mm.

2.4. Multi-Material VAM Printing

After successful suspension printing, the next goal was to use VAM printing to process the multi-material resin into a multi-material object with an overall shape defined through the VAM design. Along with the gelatin and HMWHA, both the suspension bath and the ink also included a photoinitiator, a reactive monomer, and a fluorophore for imaging. The same photoinitiator used in the single material hydrogels (LAP, 0.035 wt/vol%) was used in both components for multi-material printing. MeHA was used as the reactive macromer in the suspension bath, whereas GelMA was used as the reactive macromer in the ink. GelMA was not initially used as a single material hydrogel precursor since it has been commonly used in prior hydrogel VAM work and does not require the addition of further gelatin for printing. It is also difficult to measure gelatin release as the macromer itself is comprised of gelatin. However, the characterization of GelMA as a single material VAM resin was completed so that it could be compared to the other hydrogels (Figure S6). Here, columns of the GelMA ink were printed into the MeHA bath (containing HMWHA and gelatin) and then the entire multi-material resin was VAM printed into the shape of a disc (Figure 6a). Although this is a simple design, this provides the proof-of-concept of the multi-material print. Fluorescent z-stack images of the multi-material printed object were obtained, including both a single XY slice and a 3D reconstruction (Figure 6b). This experiment proves that a multi-material print can be achieved via a vat lithography process by combining the techniques of suspension bath and VAM printing. Through confocal microscope imaging, we observe a clear interface between the GelMA and MeHA regions, indicating distinct regions of materials with unique mechanical and chemical characteristics. The potential multi-material hydrogel objects that could be obtained with such an approach is endless.

Figure 6. Multi-material printing with VAM.

(a) Schematic and (b) images of prints of an acellular multi-material VAM print, containing regions of MeHA (2 wt/vol% MeHA, 2 wt/vol% gelatin, 1 wt/vol% HMWHA, and 0.05 wt/vol% rhodamine dextran, used as suspension bath) and GelMA (2 wt/vol% GelMA, 2 wt/vol% gelatin, 1 wt/vol% HMWHA, and 0.05 wt/vol% FITC dextran, used as ink). The top image is a singular slice and the bottom image is a 3D reconstruction. Scale bar = 1mm. (c) Schematic (cells obtained from Biorender) and (d) images of prints of a cellular multi-material VAM print, where both the suspension bath and the ink contained MSCs at 3✕10⁶ cells mL⁻¹. These are max projections of cellular multi-material VAM prints where the cells were stained for actin and the GelMA contained FITC dextran. Scale bar in 2x =1mm. Scale bar in 20x = 150 μm. (e) Quantified spread area and aspect ratio of cells within GelMA and MeHA regions, calculated using ImageJ. Data are reported as mean ± standard deviation (SD), with n ≥ 30. ****p<0.0001 (Welch’s t-test).

With this multi-material printing process developed, the final goal of this study was to create multi-material cellular constructs and observe whether cells behave differently based on local hydrogel microenvironments. Regions of GelMA and MeHA precursors with a cell density of 3 million mL⁻¹ were investigated, based on the extensive use of these hydrogels in the biomaterials field and an understanding of differential cellular behaviors (Figure 6c). Specifically, GelMA possesses both adhesion ligands and sites for enzymatic degradation, whereas HA exhibits units for receptor binding such as CD44 and sites for degradation via hyaluronidases. Fluorescent maximum projection images taken of the cellular multi-material prints show varied spreading based on the local hydrogel formulation (Figure 6d). At both 2x and 20x magnification, it is observed that cells spread much more in GelMA than in MeHA regions, which is likely due to the protease susceptibility of the gelatin-based material as well as the higher amount of cell adhesion ligands present in GelMA than MeHA. Both spread area and aspect ratio were measured, which quantitatively shows that the cells spread significantly more in GelMA than MeHA (Figure 6e). This further indicates that GelMA and MeHA remain as separate regions, and we can spatially control cellular behavior in VAM prints by utilizing suspension bath printing. Live/Dead staining in the multi-material prints showed that cell viability was 89.2% ± 2.1% (Figure S7), which was similar to single material prints as described above. Again, the various potential cellular and material combinations that could be used with this approach are extensive, which will make this method useful in the field of tissue engineering.

3. Conclusions and Future work

We have shown the use of gelatin to both modulate material viscosity to support the printing of non-viscous materials and to allow the suspension printing of inks to formulate multi-material resins for printing into defined objects with VAM. This approach was used with various reactive monomers and macromers, was cytocompatible, and gelatin could be released over time. The inclusion of high molecular weight HA altered the yield behavior of suspension baths to improve print fidelity and stability over time, providing a simple suspension bath for printing based on molecular interactions. This suspension bath allowed for a single light exposure multi-material VAM print, which altered cell behavior locally based on hydrogel formulation. Such an approach allows for the formation of complex scaffolds and constructs useful in tissue engineering.

Although the technology presented here advances the utility of the emergent VAM printing process, future work is needed to apply such an approach to specific biomedical applications. Within tissue engineering, this could be for the development of in vitro models or in translational constructs for new therapies. Having a multi-material printing process where cellular behavior can be controlled locally is important to replicate the heterogenous, complex structures of many tissues. Most tissues during development and then in the adult consist of various regions of complex biochemical and biophysical properties, which can be replicated to some extent with our new approach. Tissue engineering solutions are needed to improve the current standard of care for patients, whether through drug screening or implantable constructs, which can be advanced with this new technology. Additional limitations that can be addressed in future work involves the soft mechanics of VAM printed objects, the large amounts of resin that are needed for printing, and the monitoring of the VAM process during printing to be able to define print durations for optimal designs. Such advances in the technology will only make the process more accessible to a greater community.

4. Experimental Section/Methods

Material Synthesis and VAM Resin Formulation

MeHA Resin: Sodium hyaluronate (HA, 70 kDa mol. wt., LifeCore Biomedical) was used to synthesize MeHA as previously described.^[32] HA was dissolved in deionized water and cooled with ice. Methacrylic anhydride was then added to the solution dropwise while maintaining the pH between 8.5-9.5 for 4 hours and then vigorously stirred overnight. The solution was then precipitated in acetone, and the product dissolved in water, dialyzed for 24 hours, frozen at −80°C, and lyophilized. The degree of modification was determined by dissolving the MeHA at 1 wt/vol% in D₂O and performing ¹H NMR (400MHz Bruker, Figure S3). The resin was formulated as 2 wt/vol% MeHA, 2 wt/vol% gelatin (300 g Bloom, Sigma-Aldrich), and 0.035 wt/vol% lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP, Sigma-Aldrich) dissolved in Dulbecco’s Phosphate Buffered Saline (DPBS, Gibco).

PEGDA Resin:

The resin was formulated as 5 vol/vol% PEGDA (700 Da mol. wt., Sigma-Aldrich), 2 wt/vol% gelatin, and 0.01 wt/vol% LAP dissolved in DPBS.

PA Resin:

The resin was formulated as 20 wt/vol% acrylamide (Sigma-Aldrich), 0.5 wt/vol% N,N’-methylenebis(acrylamide) (BisA, Sigma-Aldrich), 2 wt/vol% gelatin, and 0.025 wt/vol% LAP dissolved in DPBS.

GelMA Resin:

The resin was formulated as 2 wt/vol% GelMA (50% mod, Allevi by 3D Systems), 2 wt/vol% gelatin, and 0.01 wt/vol% LAP dissolved in DPBS.

Rheological Characterization

Rheological characterization was performed on a Discovery HR20 (TA Instruments). The gelatin and high molecular weight HA characterization was performed on a Peltier bottom plate with a 20 mm, 1 degree, cone geometry. Mineral oil was added around the geometry to prevent dehydration of the samples. For temperature sweeps, a strain percent of 1% and angular frequency of 10 rad/s were used. For strain sweeps (0.1% to 1000%), the frequency was 1 Hz. For the thixotropic study, the frequency was held constant at 1 Hz and the strain percentage cycled from a low of 1% to a high of 500%. Lastly, the flow sweeps were performed between shear rates of 0.001 to 1000 s⁻¹. For photorheology, a 20 mm parallel plate geometry was used with a 405 nm light source at an intensity of 15 mW/cm². Runs were performed at room temperature and with a strain percentage of 0.1% and a frequency of 1 Hz. For each run, a baseline was established by leaving the light off until 1 minute and then exposing for 10 minutes.

VAM Printing

The VAM system (Figure S1) consists of a 3 W, 405 nm fiber-coupled laser (Civil Laser) illuminating a Vialux (Vialux) V9001 digital micromirror device (DMD) which is imaged to a vial of rotating photoactive resin with a doubly-telecentric imaging system with a magnification of 1. The cylindrical vial of resin was rotated by a Newport URS150BCC (Newport stage) rotation stage and placed in a cuvette of oil index matched to the resin to reduce refraction. The DMD was synchronized with the rotation stage to display one image per degree. Objects were printed with the various resins with the vials across a range of print designs. After printing, the objects were released from the remaining bath that had not undergone gelation through heating to 37°C.

To generate the set of 2D images required to VAM print a desired 3D object (e.g., box-frame, octet, CU logo), the Object-Space Model Optimization (OSMO) algorithm was used. In OSMO, a model of a desired 3D object is iteratively tuned to result in a printed object closer representing the desired 3D object. In the iterative loop, the filtered back projection (FBP) algorithm first generates a set of 2D projections from the model. Then, the negative values are clipped (i.e., set to zero) and the filtered and clipped 2D projections are used to generate a 3D reconstruction of the model. Next, the reconstructed model is compared to the desired 3D object. For voxels of the reconstructed model that have received too much light (e.g. an out-of-part voxel that has received enough light to polymerize), the corresponding voxels of the model are given a lower value. For voxels of the reconstructed model that have received too little light (e.g. an in-part voxel that has not received enough light to polymerize), the corresponding voxels of the model are given a higher value. This process is repeated until the model generates a set of 2D projections which forms a reconstruction that is similar to the desired 3D object.

Fluorescent Imaging

0.05 wt/vol% fluorescein isothiocyanate-dextran (FITC-dextran, mol. wt. 2,000,000) was added to single material prints for fluorescent imaging and were taken on a SMZ18 stereoscope (Nikon). For multi-material prints, 0.05 wt/vol% FITC-dextran was added to the GelMA ink and 0.05 wt/vol% tetramethylrhodamine isothiocyanate-dextran (Rho-dex, mol. wt. 500,00) was added to the MeHA bath. Fluorescent z-stacks of multi-material prints were taken on an Eclipse Ti2 confocal microscope (Nikon). All images were analyzed and had brightness and contrast adjusted using Fiji (NIH). Individual images from the stacks were transformed into 3D reconstructions.

Micro-CT Imaging

To create 3D reconstructions of the prints, a Xradia 520 Versa 3D X-ray Microscope/Micro CT (Zeiss) was used. For preparation, constructs were soaked in Lugol’s solution overnight and washed in multiple rounds of DPBS to limit diffusion. Once stained, the constructs were suspended in agarose and imaged. Images were analyzed and filtered in the Dragonfly software (Object Research Systems, Canada). Scans were taken with a 0.4x objective, 80 kV of voltage, and a power of 7 W.

Mechanical Characterization

Prior to mechanical testing, printed objects were washed in a heated (37°C) water bath for 15 minutes. For samples that underwent post-curing, after swelling, objects were placed in a solution of 0.05 wt/vol% LAP for 20 minutes to allow for the swelling of more photoinitiator after the initial washing step. The objects were then post-cured with UV light at 15 mW/cm² for 5 minutes. Compression testing of VAM printed constructs was performed on a Q800 Dynamic Mechanical Analyzer (TA Instruments). A preload force of 0.01 N and a compression rate of 0.05 N min ⁻¹ were used. The compressive moduli were calculated from the slopes of the linear region of the stress versus strain curves between 10-20% strain.

Cell Culture and Viability

All cell types (MSCs, MFCs, chondrocytes) were harvested from tissue digests obtained from juvenile bovine knee joints (Research 87, Boston, MA). The culture media was composed of high-glucose Dulbecco’s modified Eagle’s medium, 10 vol/vol% fetal bovine serum and 1 vol/vol% penicillin–streptomycin. The polymers and equipment (syringes, vials, etc.) were sterilized using a germicidal lamp for 30 minutes. The LAP, Rho-dex, and FITC dextran used were sterilized using 0.2 μm sterile filters (Thermo Fisher Scientific). All cell types were encapsulated at a density of 3 million mL⁻¹.

To quantify cell viability, a solution of 2 μM calcein AM and 4 μM EthD-1 (Thermo Fisher Scientific) solution was placed on the constructs for 30 minutes in an incubator at 37°C. Cells were then imaged on the confocal microscope (AXR scanning confocal microscope, Nikon). The z-stacks were analyzed in Fiji. Maximum projections were used to count the number of live and dead cells and calculate viability as the percentage of viable cells compared to the total cell number counted.

Quantification of Gelatin Release

To assess the release of gelatin over time, printed objects containing 0.05 wt/vol% fluorescent gelatin (Thermo Fisher Scientific) were rocked at 37°C for 7 days and supernatant was removed daily for analysis. For the fluorescence measurements (Infinite 200 Pro plate reader, Tecan), the supernatant and standards were analyzed with an excitation wavelength of 490 nm and emission wavelength of 525 nm with 25 flashes. Fluorescent images of the objects were taken on the confocal microscope with consistent settings used across all days. For the BCA assay (Pierce, Thermo Fisher Scientific), the standard protocol was followed.

Suspension Bath Printing

Extrusion printing was performed on a Bio X (Cellink) at room temperature. The ink was composed of 2 wt/vol% GelMA, 2 wt/vol% gelatin, 1 wt/vol% HMWHA (~1.5MDa, LifeCore), 0.035 wt/vol% LAP, and 0.05 wt/vol% FITC dextran. The bath was composed of 2 wt/vol% MeHA, 2 wt/vol% gelatin, 1 wt/vol% HMWHA, 0.035 wt/vol% LAP, and 0.05 wt/vol% rhodamine dextran. Both compositions were mixed overnight at 37°C and were then brought to room temperature. For cellular experiments, both the bath and the ink had a cell density of 3 million mL⁻¹. 30 kPa of pressure and a speed of 3 mm/s were used during extrusion. The container was directly placed into 4°C after printing to solidify.

Multi-material Hydrogel Preparation

For multi-material prints, the suspension baths and inks were prepared as described above. Suspension printing was then performed at room temperature before cooling to 4°C to solidify. After solidification, the vial was used in VAM printing as described above. Then, the solution was heated to 37°C and the printed multi-material object was removed from the excess resin.

Cell Spreading Quantification

To quantify cell spreading in multi-material objects, nuclei were stained with DAPI and actin was stained with Alexa Fluor 647. Constructs were fixed by a 4 wt/vol% paraformaldehyde solution for 15 minutes. Then, a solution of 0.2 vol/vol% DAPI and 0.4 vol/vol% Alexa Fluor 647 was placed over the constructs and then rocked at 4°C overnight. They were then washed in DPBS and imaged on the confocal microscope. Z-stacks were then analyzed in Fiji and the spread area and aspect ratio were calculated using the area and length tools.

Statistical Analysis

Unless otherwise specified, data are reported as mean ± standard deviation and n ≥ to 3. All statistics were performed via Prism 9 (GraphPad by Dotmatics, USA). For comparisons between two data sets, Welch’s t-test was completed (significance determined by p < 0.05). For comparisons between multiple groups or variables, either a one- way analysis of variance (ANOVA, significance determined by α = 0.05) or two-way ANOVA (significance determined by α = 0.05) was completed.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.