Multiscale Engineered Heterogeneous Hydrogel Composites for Digital Light Processing 3D Printing

After obtaining human dermal fibroblast cells (HDFs) (ATCC) from multiple donors, the cells were cultured in an aseptic condition in 5% CO₂ at 37 °C with Dulbecco-modified eagle medium (DMEM) high glucose (Cytiva-Hyclone), supplemented with 1% penicillin/streptomycin (100 U/100 μg/mL; Gibco). Human mesenchymal stem cells (hMSCs) (ATCC) were also obtained from multiple donors and cultured in previously described conditions with α-Minimum Essential Medium (α-MEM) (Cytiva-Hyclone), supplemented 1% penicillin/streptomycin (100 U/100 μg/mL; Gibco). After every 2 days, the media was changed with fresh media, and after obtaining 80% confluency, cells were passaged using 0.5% trypsin-EDTA (Gibco) at approximately 75,000 cells/cm² for expansion.

To form photo-cross-linked, cell-laden microgels, DMEM was used as a solvent to prepare the precursor, and HDFs were dispersed in the precursor with a final cell density of 5 × 10⁵ cells/ml in 7.5% GelMA. The generated droplets were photo-cross-linked using the previously specified UV intensity and exposure time. As the temperature must be maintained within a cell-friendly range and cell density may vary, the fluidity of cell-laden precursors could be compromised, which affects the monodispersity of the droplets. Finally, the cell-laden microgels were collected by centrifugation (500g for 10 min), redispersed in the medium, and transferred to a 96 well-plate containing DMEM and incubated at 37 °C for 1, 3, 5, and 7 days.

The cytocompatibility of the cell-encapsulated microgels was analyzed using the Alamar Blue assay. The assay was performed on days 1, 3, 5, and 7 according to the manufacturer's protocol. Following incubation, the reading was taken at 570 and 600 nm with a plate reader (Tecan Group Ltd., Tecan Infinite 200Pro M Plex, Switzerland). The percent reduction was calculated after each time point, with the values on days 3, 5, and 7 normalized by the values obtained on day 1.

hMSCs were first stained using CellTracker fluorescent probes (Green-CMFDA, Invitrogen) according to the manufacturer's instructions. hMSC-laden microgels were fabricated and collected following the previously described method using a precursor with a final cell density of 1.8 × 10⁶ cells/ml in 7.5% GelMA. To prepare composite bioink, hMSC-laden microgels were mixed with a 12.5% GelMA bulk. Subsequently, hMSC-laden hydrogel composites with a 10 mm diameter and 2.5 mm thickness were fabricated using DLP and transferred to a 24 well-plate containing α-MEM and incubated at 37 °C for 1, 3, and 5 days. At each time point, the brightfield and fluorescence images of hMSC-laden hydrogel composites were acquired with a microscope at 10× magnification (BioTek Lionheart LX, Agilent).

For a bioink to achieve proper functionality for DLP printing, the ink viscosity should allow for sufficient spreading across the building platform at printing temperatures. Optimal spreading is typically observed in an ink viscosity range of 10⁰ to 10⁴ mPa·s., Low temperatures increase the viscosity of GelMA-based bioinks, while excessive temperatures can lead to denaturation of polymer chains and negatively affect cell viability during bioprinting.− Preliminary experiments were performed to assess the preprint fluidity of different composite ink formulations based on varying concentrations of GelMA within the microgels. With all other variables held constant, varying polymer concentrations of the microgels alone did not significantly affect the viscosity at printing temperatures. All experimental groups remained within a suitable viscosity above 30 °C (FigureB), which would prevent the physical gelation of GelMA during printing and provide cytocompatible conditions for any bioprinting applications.

We then wanted to explore the effect of microgel composition on the rheological and mechanical properties of the printed composites. Studies have shown that embedding softer inclusions within a polymer network reduces the overall stiffness but improves yield strength and failure properties. Microgels with greater stiffness than the surrounding matrix result in an increased composite stiffness but reduced yield strength. Composite bioinks containing microgels with varying GelMA concentrations were applied to DLP 3D printing to fabricate heterogeneous hydrogel composites. Uniaxial compression to 60% strain was performed on printed constructs. In composites with a μgel/bulk volume ratio of 1:5, the addition of microgels reduced the compressive modulus of the resulting structure relative to the monolithic ink, regardless of microgel polymer concentration. Microgels acted as defects in the system, disrupting cross-linking throughout the network and weakening structural stability. This effect became more pronounced as the difference between polymer concentrations in the separate components increased. Adding microgels containing 12.5% GelMA to a bulk precursor with an identical amount reduced compressive modulus by approximately 25%. Meanwhile, the addition of microgels containing 7.5% GelMA to the same ink reduced the compressive modulus by 55% (FigureC). A potential explanation for this can be found in the microgel's low stiffness, which prevents it from effectively resisting permanent deformation within the composite.

The results of the mechanical tests indicated a different effect on failure properties. Failure was observed at similar strains for each group, indicating the properties of the matrix dominated this aspect (Figure S1A↗). This is likely due to the low volume fraction of microgel inclusions. Unsurprisingly, a decrease in failure stress was observed as polymer concentration in microgels decreased. As the compressive moduli of composites were reduced, their ability to resist deformation due to applied stress was decreased. Rheological strain sweeps indicated that although the storage modulus of composites containing different microgels was initially similar, the yield strain increased as polymer concentration in the microgels decreased (FigureB).

In 3D bioprinting, fidelity and printability measurements are performed to determine the geometric accuracy of printed structures. Fidelity, measured by comparing printed line width to that of the specified geometry, describes the ability to reproduce computer-designed structures with high accuracy and precision. The ideal fidelity measurement is 1, while deviations indicate under or overcross-linking. Circularity equations can be applied to different shapes for a separate geometric accuracy assessment by defining a relationship between perimeter and area. The adherence of printed shapes to this relationship determines its printability. At both the micro- and macroscales, fidelity and printability together evaluate the overall printing performance. ⁸⁸

In the printing of 2D shapes, when GelMA concentration in the bulk was constant, changing microgel formulation had no significant effect on fidelity and printability. Moreover, at this bulk formulation, all experimental groups showed reduced line uniformity when printing shapes with a line width of 0.5 mm. As the microgel polymer concentration increased, this fluctuation in fidelity gradually diminished (FiguresG and S1B↗). This phenomenon can be attributed to the presence of soft microgels, which introduce discontinuities within the matrix of the composite. Softer structures have shown limited compatibility with DLP due to their inability to support the weight of layers during printing. Further reduction of stiffness through the incorporation of microgels confirmed this relationship as printed structures became less accurate.

Overall, in comparison to monolithic bulk hydrogels with different GelMA concentrations (Figure S2A–C↗), the incorporation of microgels at lower volume ratios significantly impacted compressive moduli but did not show a similar effect on viscosity and geometric accuracy of DLP-printed shapes. This indicates the potential for designing composite bioinks with modular compressive moduli without affecting printability.

In addition to varying polymer concentration in the microgels, we wanted to understand how making similar changes to the bulk precursor would affect composite properties. As polymer concentration increases, chain entanglements become more extensive, resulting in an increased cross-link density. Additionally, the entanglements may restrict the movement of the molecular chain and increase the shear modulus of the composite, thereby affecting the ink fluidity. This was assessed by performing a rheological temperature sweep on bioink precursors prepared using either 10, 12.5 or 15% GelMA. Similar to the previous results, changing bulk formulations within this range did not have a significant effect on preprint viscosity (FigureE).

To explore how bulk polymer concentration affected the stiffness of composites, different bulk formulations were mixed with 12.5% GelMA microgels at a μgel/bulk volume ratio of 1:5. Results of uniaxial compression to 60% strain indicated that, as expected, the concentration of GelMA in the ink directly affected its stiffness. After cross-linking, hydrogel composites with 12.5 and 15% GelMA in the bulk had comparable storage moduli and similar yield strains, while both properties were reduced with 10% GelMA (FigureE). Similarly, the use of 10% GelMA in the bulk led to significant reduction in compressive modulus compared to 12.5%, while an increase to 15% did not show a significant effect (FigureF). We then studied the effect of polymer concentration in the bulk on the printing performance of composite bioinks. Following the results of previous characterizations, the reduction of GelMA concentration in the bulk from 12.5 to 10% affected the printing of smaller structures, presumably due to insufficient mechanical stability of the composite bioink after cross-linking (FiguresG and S3B↗). As a result, part of the printed construct was damaged during printing due to surface tension between the building platform and the bioink reservoir. On the other hand, while increasing bulk GelMA concentration to 15% did not result in any failed printing attempts, fidelity and printability were less ideal (FiguresG and S3B↗). This is caused by increased viscosity leading to bioink aggregation in localized areas during printing, thereby increasing the risk of overcrosslinking.

In DLP, photoabsorbers improve the resolution of printed structures by absorbing UV light energy and preventing cross-linking in areas outside of projected image layers. Therefore, photoabsorber concentration dramatically influences the degree of photopolymerization for ink. Tartrazine, a synthetic dye, is commonly used as a photoabsorber for DLP bioprinting due to its high absorbance of 405 nm light and its high cytocompatibility within a wide range of concentrations.In this experiment, both microgel and bulk GelMA concentrations were maintained at 12.5%, and three tartrazine concentrations (0.5, 1, and 2 mM) were explored for their effects on cross-linking kinetics, stiffness of printed composites, and fidelity and printability. , ^{90 91}

Rheological testing was performed to explore how varying photoabsorber concentrations would affect gelation kinetics of the composite bioinks. As expected, increasing the concentration of tartrazine prolonged the gelation time for the composite bioink, and correspondingly reduced the storage modulus of the cross-linked sample. At 0.5 mM tartrazine, the composite bioink exhibited a rapid increase in storage modulus within the first 10 s of UV exposure and reached a plateau after 25 s. Upon increasing tartrazine concentration to 1 mM, the onset of cross-linking was delayed and the rate of increase was slightly reduced. However, the exposure time required to achieve the maximum storage modulus was unaffected. When tartrazine concentration was increased to 2 mM, the gelation of the composite bioink was significantly delayed (FigureB). Moreover, the storage modulus was unable to reach similar levels after nearly 2 min of UV exposure. This is likely due to excess tartrazine absorbing a significant portion of UV energy, delaying the time at which a noticeable increase in storage modulus occurs. A similar trend was also observed when investigating how tartrazine concentration affected the maximum cure depth. Under the same UV intensity and exposure time, composite bioinks with 0.5 mM tartrazine resulted in cure depths exceeding 1000 μm. When using 2 mM tartrazine, a layer of only 250 μm was thoroughly cured after 1 min (FigureC).

Uniaxial compression tests to 60% strain were performed to evaluate the impact of tartrazine concentration on the mechanical strength of the printed structures. When using 0.5 mM tartrazine, the compressive modulus of the printed construct was around 250 kPa (FigureD). As the tartrazine concentration increased, a corresponding decline in modulus was observed, ultimately decreasing by 80% when tartrazine concentration was increased to 2 mM (FigureD). It is notable that with 2 mM tartrazine, the printed structures exhibited the same failure stress as those with 0.5 mM Tartrazine (Figure S4A↗). This may be attributed to the 2 mM tartrazine concentration causing a decrease in cross-link density, resulting in a softer printed construct. This reduction in stiffness with 2 mM tartrazine resulted in a composite able to undergo greater deformations at slower rates, displaying the same failure stress as stiffer constructs. This result was confirmed by the failure strain results, where an increase in tartrazine concentration resulted in a 20% increase in failure strain (Figure S4A↗).

Varying tartrazine concentrations led to significant effects on the fidelity and geometric accuracy of printed structures (FiguresH, and S4B↗). Composites containing 1 mM Tartrazine achieved optimal fidelity and printability, forming constructs that highly resembled the CAD models. When the concentration was increased to 2 mM, excessive tartrazine absorbed the energy from UV exposure, preventing the initiation of photo-cross-linking. This composite bioink formulation was unable to print uniform shapes at either line thickness and instead formed randomly separated hydrogel fragments. With 0.5 mM tartrazine, shapes were printed with severely reduced fidelity and printability. While features of both honeycombs and grids were formed, significant overcrosslinking was observed during all printing attempts.

In DLP printing, increasing UV intensity transfers more energy to the same unit area, cross-linking constructs more rapidly. On the other hand, insufficient UV intensity can result in incomplete cross-linking, reducing the mechanical strength of the printed constructs and significantly compromising printing performance. An appropriate UV intensity can provide the proper energy to facilitate rapid cross-linking of layers to maintain sufficient mechanical properties without compromising fidelity and printability.

To explore the effects of UV intensity on the mechanical properties and printability of the composite bioinks, we initially investigated how variations would affect cross-linking kinetics, cure depth, and compressive modulus. The GelMA concentrations in both microgels and bulk were maintained at 12.5%, with tartrazine held at a constant 1 mM. Initially, three different UV intensities were selected based on the power levels of a LumenX+ bioprinter. Specifically, 40, 60, and 80% power correspond to 4, 7.8, and 11 mW/cm², respectively. Rheological assessments indicated that at all UV intensities, a significant rise in storage modulus occurred within 30 s after UV exposure. However, 40% power did not supply enough energy for the composite to reach a storage modulus plateau within 2 min (FigureE). Typically, DLP printing is performed using exposure times on the range of 0.5 to 30 s per layer. Within that range, this low amount of UV energy would result in incomplete cross-linking and a significant reduction in stiffness. The cure depth experiment further confirmed this limitation. After 1 min of UV at 80% power, a layer of approximately 600 μm was cured, while 40% power only cured around 250 μm (FigureF).

Mechanical testing through uniaxial compression to 60% strain determined that a reduction in UV exposure affected the moduli of printed composites. The use of 40% and 60% power led to structures with moduli that were significantly reduced compared to constructs printed at 80%. Within the limited exposure time, the energy delivered at these power levels was insufficient to induce adequate cross-linking, resulting in a less stable network. Similar to previous results, softer constructs were able to withstand the same compressive force as stiffer composites by undergoing greater deformation. Failure strain was inversely proportional to UV exposure, as the constructs with a reduced stiffness were able to undergo slightly increased levels of deformation (Figure S5A↗). No direct relationship between UV exposure and failure stress was observed.

Printing of 2D shapes was performed to determine the optimal level of UV exposure for geometric accuracy. (FiguresH, and S5B↗) Using 40% power did not result in successful printing of any attempted shape. This is due to insufficient cross-linking in printed layers, hindering their ability to support the weight of the following layers as the structure was built, resulting in print failure. The 60% power level failed to print any shapes with a 0.5 mm line width but was able to print shapes with a 1 mm line width to an extent, although the fidelity and printability were not ideal. In contrast, 80% power resulted in successful prints of all four shapes with good fidelity and printability. These results allow for the optimization of cross-link density through UV exposure rather than varying material-related parameters.

To study the effect of microgel volume fraction on the heterogeneous hydrogel composite system, we prepared composite bioinks with three different μgel/bulk volume ratios (1:5, 1:1, 3:1). This ratio directly affects the microstructure of composites, as interactions between these two components become more or less pronounced. At a μgel/bulk volume ratio of 1:5, the bulk dominated the heterogeneous hydrogel composite, with the microgels acting as defects embedded within the continuous phase and influencing the bulk's behavior. At 3:1, the microgels occupied the majority of the volume in the hydrogel composite system. In this case, the properties of the composite bioink resembled those of a granular hydrogel with interstitial fluid filling the interparticle voids.

The 1:5 volume ratio was intended to represent the effect of a small amount of microgel inclusions on the composite properties. At 1:1 volume ratio, the microgel and bulk occupy the same volume in the system and were expected to have equal impacts on the properties of the hydrogel composite. A 3:1 volume ratio was prepared to investigate the properties of the composite bioink when microgel serves as the primary component. At this ratio, microgels comprised 75% of the total volume, which is the maximum packing density achievable for spherical particles. Beyond this limit, softer gel particles may undergo deformation during packing. SEM was utilized to examine the composites' microstructure and the interfacial architecture between microgels and bulk (Figure S6↗). At a 1:5 volume ratio, a smoother, more continuous bulk matrix was observed, with microgels throughout and pores that formed during lyophilization. As the microgel content increased, the continuity of the bulk became disrupted. At a 3:1 volume ratio, the composite was mainly composed of microgels, with the bulk appearing as a thin, mesh-like coating over the surface of packed microgels.

Temperature sweeps showed that at 1:5 volume ratio, the bulk had a more dominant influence on its properties. The fluidity of the composite remained unchanged compared with the bioink without any microgel. A similar effect was observed at the 1:1 volume ratio, which showed comparable viscosities. In the composite bioink with a 3:1 volume ratio, the microgels dominated the viscous properties. Even after the temperature exceeded 40 °C, the viscosity of the 3:1 composite was not within a printable range (Figure S7A↗).

The gelation tests demonstrated that as microgel volume fraction increased, the time required for the storage modulus to reach its maximum was prolonged. Compared with the composite bioinks with 1:5 and 1:1 volume ratios, those with a 3:1 ratio had a slightly higher storage modulus prior to UV irradiation, which is aligned with the results of the temperature sweeps. Moreover, the composite bioink with a 3:1 volume ratio required more exposure time before an increase in modulus was observed, which we hypothesized was caused by microgels preventing the formation of a uniform network and slowing the gelation kinetics of the composite. The 3:1 volume ratio also resulted in the smallest difference in modulus before and after UV exposure. An opposite trend was observed in the composite bioink with a 1:5 volume ratio. This bioink showed a lower viscosity before UV exposure and reached a storage modulus of 100 Pa within 20 s of exposure, compared to more than 45 s at a 3:1 volume ratio (FigureC).

The compressive moduli of printed constructs were negatively correlated with the microgel volume fraction. Uniaxial compression tests to 60% strain determined that as the volume fraction of microgels increased, the compressive modulus of printed composites decreased significantly. Volume ratios of 1:5 μgel/bulk produced structures with an average modulus of approximately 140 kPa, while 3:1 volume ratios led to softer structures of ∼66 kPa (FigureD). Failure stresses of the composite structures followed a similar trend. However, the failure strain was independent of the volume ratio, indicating that microgels did not affect the deformability of printed constructs, but influenced its energy absorption capacity under the same levels of deformation (Figure S7B↗). The granular nature of the packed microgels in printed constructs with a 3:1 volume ratio resulted in lower failure stress. In contrast, composite bioinks with lower microgel content could absorb more energy as well as withstand and distribute greater forces. Additionally, the higher microgel volume fraction affected the degradation rate of the composites. Composites with a 1:1 volume ratio degraded within 4 days, whereas those with a 1:5 ratio showed improved stability comparable to monolithic bulk hydrogels (Figure S8↗).

Printability assessments supported these findings, showing that 1:5 and 1:1 μgel/bulk volume ratios allowed the fabrication of structures with reduced geometrical accuracy. On the other hand, 3:1 μgel/bulk volume ratio significantly reduced the preprint fluidity of the composite as well as its capability to produce specified printing features. (FiguresE, and S7C↗) Fidelity and printability of 2D shapes were optimal at the 1:5 ratio, however, acceptable print quality was obtained when printing with a 1:1 ratio. Upon increasing the microgel fraction to 75%, none of the attempted prints was successful.

Up to this point, the microgels used to fabricate the heterogeneous composites were formed using UV-triggered covalent cross-linking. This cross-linking process reduces the number of reactive methacrylate groups in the microgels, limiting their ability to interact and integrate with the bulk during the printing of composites. To address this and attempt to increase the interaction of microgels with the surrounding matrix, microgels were physically cross-linked through incubation at 4 °C without prior UV exposure and combined with bioink precursors to prepare composites (FigureA). When physically cross-linked microgels were incorporated into the bulk, more functional groups were available to react with the matrix under UV light exposure (FigureA,B), strengthening μgel-bulk interactions. Additionally, the mechanical strength of the printed constructs was enhanced due to the stronger bonding between microgel and bulk. Uniaxial compression to 60% strain confirmed that at similar volume ratios, printed constructs containing physically cross-linked microgels had higher compressive modulus and failure stresses than those containing photo-cross-linked microgels.

Rheological assessments of gelation kinetics showed that a similar trend was observed with physically cross-linked microgels as compared to those that were photo-cross-linked. As microgel volume increased, a longer exposure time was required for an increase in storage modulus to be observed. With physically cross-linked microgels, an increase in maximum storage modulus was achieved when volume fraction was kept constant (FigureC). Assessments of compressive modulus (FigureD) and failure properties (Figure S9A↗) showed that the volume fraction-dependent reduction in stiffness observed previously was consistent. However, the increase in available methacrylate groups led to an increase in the average compressive modulus of composites at all volume ratios. This is likely due to increased interactions with the surrounding bulk.

The addition of physically cross-linked microgels slightly improved the print resolution of composites compared to photo-cross-linked gels (FigureE). However, fidelity and printability were both still negatively impacted by increases in microgel content. At 1:5 and 1:1 μgel/bulk volume ratios, composite bioinks with physically cross-linked microgels showed similar performance to those containing photo-cross-linked microgels. However, the printability and fidelity were improved at the 3:1 volume ratio, indicating stronger interactions between embedded microgels and the surrounding bulk (FiguresE, and S7C, S9B↗).