Injectable MSC Spheroid and Microgel Granular Composites for Engineering Tissue

Abstract

Many cell types require direct cell-cell interactions for differentiation and function; yet, this can be challenging to incorporate into 3-dimensional structures for the engineering of tissues. Here, we introduce a new approach that combines aggregates of cells (spheroids) with similarly-sized hydrogel particles (microgels) to form granular composites that are injectable, undergo interparticle crosslinking via light for initial stabilization, permit cell-cell contacts for cell signaling, and allow spheroid fusion and growth. One area where this is important is in cartilage tissue engineering, as cell-cell contacts are crucial to chondrogenesis and are missing in many tissue engineering approaches. We develop granular composites from adult porcine MSC spheroids and hyaluronic acid microgels and use simulations and experimental analyses to establish the importance of initial MSC spheroid to microgel volume ratios in granular composites that balance mechanical support with tissue growth. Long-term chondrogenic cultures of granular composites produce engineered cartilage tissue with extensive matrix deposition and mechanical properties within the range of cartilage, as well as integration with native tissue. Altogether, we have developed a new strategy of injectable granular composites that leverages the benefits of cell-cell interactions through spheroids with the mechanical stabilization afforded with engineered hydrogels.

Graphical Abstract

Hydrogel microparticles and mesenchymal stromal cell spheroids are combined into granular composites to engineer cartilage tissue. The spheroids allow for important cell-cell interactions during chondrogenesis and fusion into a cartilage tissue, whereas the granular structure allows for injectability and then stabilization through interparticle crosslinking. This approach advances the use of engineered biomaterials for tissue engineering applications.

1. Introduction

Direct cell-cell interactions (e.g. N-Cadherin) are essential to drive the development of many tissues (e.g., cartilage) and become challenging to incorporate into many 3-dimensional scaffolds for the engineering and repair of tissues.^[1,2] Aggregates of cells, known as spheroids, have become increasingly popular within the tissue engineering field since they possess inherent cell-cell interactions and a potent secretome when compared to cells dispersed on a scaffold or within a hydrogel.^[3-5] Additionally, these high cell dense aggregates are often desirable for the engineering of highly cellular tissues or to mimic developmental processes (e.g., limb bud formation).^[6,7] As individual spheroids are insufficient for engineering tissues due to their size, many studies have investigated the role of spheroids as “building blocks” that can be used alone or in tandem with biomaterials to fabricate macroscale tissues that guide their function.

Biomaterials have been introduced to guide spheroids both internally (incorporated within spheroids) and externally (incorporated outside of spheroids) to introduce mechanical support and guide cell behavior via biomimetic cues (biophysical and biochemical). Internally, challenges with nutrient diffusion, morphogen delivery, and heterogeneous differentiation across individual spheroids have motivated the incorporation of “cell-sized” biomaterial microparticles (i.e. microgels) within spheroids to guide and improve spheroid cell viability, differentiation, and tissue-specific functions.^[8-13] Externally, similar to cell encapsulations, spheroids have been encapsulated within hydrogels, where hydrogel mechanics, adhesive sites, micro-interfaces, and viscoelastic properties have guided spheroid differentiation, cell outgrowth, and paracrine signaling.^[14-17] Additionally, non-hydrogel biomaterials (e.g., polycaprolactone melt electrowriting scaffolds) have been used to guide spheroid cultures and control construct mechanics.^[18-22] There are some challenges with these approaches. For example, scaffold-free spheroid constructs often lack initial stability and may collapse with culture to reorganize into a dense cellular “ball” structure, whereas scaffold-based approaches often limit the fusion of spheroids and ECM deposition.^[23-26] These limitations motivate the development of new, innovative strategies that allow for the use of spheroids as “building blocks” for engineering tissues, while also using permissive biomaterials to guide tissue formation.

Granular hydrogels are an emerging class of biomaterials composed of packed microgels that give rise to unique properties such as injectability due to microgel flow, intrinsic porosity via microgel packing, mechanical support via interparticle crosslinking, and are modular by nature. Such hydrogels have already shown great promise in tissue repair approaches by demonstrating enhanced cell migration and cell infiltration from surrounding native tissues compared to traditional “bulk” hydrogels in both in vitro and in vivo settings given their macroporous structure.^[27-32] Granular hydrogels can also be altered to contain growth factors that elicit chemotaxis for cell infiltration or be chemically modified to present bioadhesive sites to enhance cell migration and behavior.^[27,30,33] Additionally, granular hydrogels have been extensively used as in vitro platforms to investigate biological processes (e.g. angiogenesis, cellular outgrowth, cell condensation) with the incorporation of individual spheroids or cellular suspensions within the pores.^[34-37]

Here, we combine the positive attributes of spheroids and granular hydrogels to engineer tissues through the formation of granular composites (Figure 1a). Specifically, spheroids are mixed with microgels to form granular composites that (i) are injectable for delivery to molds or defects, (ii) introduce cell-cell contacts and spheroid fusion that are conducive to tissue formation, and (iii) undergo interparticle crosslinking via light to immediately stabilize constructs. In review of the extensive prior literature where biomaterials have been incorporated internal or external to cellular spheroids (extensive summary of literature within Table S1), we believe this is an innovative approach that has not been explored previously. As an example, cellular spheroids have been packed as granular baths for 3D printing of vascular structures, but this approach did not include microgels.^[38]

Figure 1. Granular composite approach and component characterization.

a, Schematic overview of granular composite design, where MSC spheroids and NorHA microgels are mixed to enable (i) injectability for delivery to defects or molds, (ii) cell-cell contacts for enhanced chondrogenesis and spheroid fusion for tissue formation, and (iii) interparticle crosslinking via light to stabilize the composites. b, Schematic of granular composites over time, where spheroid fusion and growth result in cartilage tissue throughout the granular hydrogel. Schematics are not accurately scaled. c, Representative image of MSC spheroids (top) and quantification of MSC spheroid size distribution (bottom) 2 days after seeding 1000 cells/spheroid. n=50; scale bar: 200μm. d, Representative day 28 chondrogenic MSC spheroids stained for sGAGs (Alcian blue), chondroitin sulfate, and collagen II; scale bars: 200μm. e, Quantification of microgel size distribution (left) and representative image (right) of microgels fabricated via batch emulsion of NorHA (spun at 280 RPM). n=154; scale bar: 200μm.

To illustrate the utility of this approach, we select cartilage tissue engineering. Cartilage is an ECM dense tissue that develops due to the high cellularity and early cell-cell interactions during limb bud formation that drives cartilage generation, making it an important application for the use of spheroids. The inability of cartilage to regenerate innately after damage has motivated numerous approaches to engineer cartilage for repair due to the prevalence of knee injuries and osteoarthritis within the U.S.^[39,40] One such approach has utilized mesenchymal stromal cells (MSCs) within hydrogels to enhance defect filling;^[41,42] however, nanoporous hydrogels often limit cell-cell interactions and mechanical properties, reducing their efficacy for producing cartilage tissue.^[43,44] N-Cadherin biomimetic peptides and tethered chondroinductive growth factors have been previously incorporated into nanoporous hydrogels to enhance cartilage-specific ECM deposition; however, mechanical properties remain suboptimal to native cartilage.^[25,43,45] Other cartilage tissue engineering strategies have incorporated MEW scaffolds alone or within hyaluronic acid-based hydrogels to enhance mechanical properties regardless of cartilage production.^[46,47] Alternatively, high cell density hydrogels cultures have demonstrated great promise towards reaching native level cartilage mechanics; however, high cell densities can limit hydrogel design due to crosslinking efficiency.^[48] To alleviate this issue, MSC or CH spheroids have been recently used for cartilage tissue engineering by packing spheroids alone, in the presence of enzymatic treatments, or with MEW scaffolds.^{[24,46,49,50]} While these approaches are exceptional as implantable strategies, they all lack the ability to be injected and stabilized post-injection. Thus, our granular composite approach that combines MSC spheroids and norbornene-modified hyaluronic acid (NorHA) microgels into high cell dense granular composites (>200 million cells mL⁻¹) will facilitate the formation of cartilage tissue that is guided by the stabilized granular hydrogel (Figure 1b).

2. Results and Discussion

2.1. Spheroids and Microgels Form Granular Composites When Mixed

Before investigating granular composite cartilage formation, we first sought to understand how individual MSC spheroids grow in chondrogenic medium. MSC spheroids are fabricated by seeding adult porcine MSCs into pyramidal microwells that allow cells to condense by day 2 (D2) (Figure S1a, Supporting Information). These microwells allow for scalable generation of spheroids in the thousands, where the initial MSC seeding density (500-2000 cells/spheroid) controls the MSC spheroid mean diameter (~125-175 μm) and distribution (Figure 1c, Figure S1b, Supporting Information). Spheroid viability is high, with a live cell area >90% after D7, except for the largest MSC spheroids (2000 cells/spheroid) with a live cell area of ~80% (Figure S2, Supporting Information). To incorporate the highest cell density while also maintaining high cell viability, we select the 1000 cell/spheroid seeding density for the remaining studies. MSC spheroids increase ~25% in diameter and ~100% in volume (assuming a perfect sphere) over 28 days (Figure S3a, b, Supporting Information). ECM deposition is confirmed via histological staining for cartilage-specific ECM collagen II, chondroitin sulfate (CS) and sulfated glycosaminoglycans (sGAGs) (Figure 1d).

After characterizing individual MSC spheroids, we next fabricate similar-sized microgels. Due to the presence of hyaluronic acid (HA) within cartilage and the long history of HA in biomedical products (e.g., viscosupplements) and in hydrogel design, we use HA with norbornene modification (i.e., NorHA) to form microgels, which is chosen to enable control over microgel crosslink density and interparticle crosslinking.^[51-53] We fabricate NorHA microgels using batch emulsion, by adding a solution of NorHA, crosslinker, and photoinitiator dropwise into mineral oil while mixing. The emulsion is exposed to UV light for 30 min. for microgel crosslinking and microgel suspensions are obtained by washing from oil (Figure S1c, Supporting Information). Rotation spin speeds alter the mean diameters and distributions of NorHA microgels (e.g., increased speeds reduce NorHA microgel size) (Figure S1d, Supporting Information). To match MSC spheroid diameters, an emulsion spin speed of 280 RPM is used to obtain NorHA microgel diameters of ~140 μm (Figure 1e). NorHA microgel maintain their size (~150μm) over a 1-month period, which is expected as hydrogels of this specific formulation exhibit minimal degradation (e.g., ~30% uronic acid release) during this same period (Figure S4, Supporting Information).

With each granular component characterized, we next investigate how the mixing of spheroids and microgels influences granular composite formation and the connectivity of each component. This is most important to enable microgel interparticle crosslinking for initial construct stabilization. To gain insight into this and to avoid extensive iterative experimental testing, mixing is simulated in Cinema4D (C4D) with rigid body objects. We hypothesize that microgels within the composite require a critical level of connectivity to maintain construct stability, below which the composite will disassemble upon agitation. The mixing of spheroids and microgels within a cylindrical tube is simulated in C4D, where gravity and collisions drive mixing behaviors (Video S1,2, Supporting Information). Voxelation of simulated composites allows the generation of a binary z-stack that is then analyzed in FIJI and the two components are separated (Figure 2a). The 3D object counter then assigns colors to “clusters” of either spheroids or microgels that are physically touching in 3D space, allowing the calculation of a connectivity parameter for either granular component. Connectivity is defined here as the number of particles within the largest cluster divided by the total number of particles within the component (spheroid or microgel total particle number). If all spheroids or microgels are connected, the connectivity parameter would be 1, whereas if there are low numbers in each cluster, the connectivity parameter approaches 0. Similar to percolation theory, this method allows us to understand the connectivity of both spheroid or microgel components to inform how the initial volume ratio impacts stability and tissue formation.^[54] This approach reveals that low ratios (<50:50) of spheroids to microgels result in low spheroid connectivity, whereas high ratios of spheroids to microgels (>50:50) result in decreased microgel contact (i.e., limited interparticle crosslinking) with low microgel connectivity (Figure 2b, c). Further, microgel size polydispersity and spheroid aggregation are introduced into simulations (Figure 2c, Figure S5, Supporting Information). In general, an increase in microgel polydispersity reduced the connectivity of microgels and increased spheroid connectivity in the 35:65 spheroid to microgel volume ratio, whereas increased spheroid aggregation slightly changed microgel connectivity, but increased the spheroid connectivity in the 20:80 and 35:65 spheroid to microgel volume ratio groups. Taken together, we identify 20:80, 35:65, and 50:50 spheroid to microgel volume ratio formulations for further exploration of granular composites, based on the need for high microgel connectivity for construct stabilization and to limit issues of construct collapse over time with spheroids alone.^[23,24]

Figure 2. Spheroids and microgels form granular composites and their ratio dictates connectivity.

a, Schematic workflow of dynamic rigid body simulation connectivity analysis through Cinema4D. Granular composite mixing (green: microgels; pink: spheroids) simulated via gravity and imported into FIJI to slice and voxelize 3D object distributions as binary Z-stacks. Binary Z-stacks analyzed for connectivity through 3D object counter function in FIJI. Connectivity is defined here as the number of particles within the largest cluster divided by the total number of particles within the component (spheroid or microgel total particle number). b, Representative images of granular composite simulations of varying volume ratios of spheroids to microgels. c, Quantified connectivity of monodisperse granular composites of varying spheroid to microgel ratios. n=5, mean ± s.d. Quantification of granular composite connectivity that incorporates microgel size distribution into the simulation of all volume ratios. n=5, mean ± s.d. Quantification of granular composite connectivity that incorporates spheroid aggregation into the simulation for 20:80, 35:65, and 50:50 volume ratios. n=5, mean ± s.d. ****P<0.0001, ***P<0.001, **P<0.01, ns=not significant.

2.2. Granular Hydrogels are Injectable and Mechanically Stabilized with Interparticle Crosslinking

To investigate potential injectability of granular composites, the rheological properties of 20:80, 35:65, and 50:50 spheroid to microgel ratios are measured. Spheroids and microgels are first independently isolated and concentrated via centrifugation and then mixed to form a composite slurry (Figure 3a). Testing is done immediately after mixing to mimic the formulations that would be injected into tissue defects. As controls, either MSC spheroids alone or microgels alone are measured rheologically (Figure 3b), where the microgels exhibit a 10-fold increase in storage modulus and a shift in yielding behavior (G”<G’) to higher strains when compared to spheroids, likely caused by differences in particle interactions under load (Figure 3b). When composites are tested, similar moduli and yielding behaviors are observed across compositions (Figure 3c). Specifically, the quantified values for the storage moduli across groups are similar when microgels are included (except 20:80), whereas the loss modulus increases in composites over either component alone (Figure 3d, e). It should be noted that the jamming conditions (e.g., centrifugation speed) of microgels can alter microgel packing and storage and loss moduli and must be considered when formulating composites and as expected, exposure to light stabilizes jammed microgels (Figure S6a, b, Supporting Information). Cyclic strain behavior of individual components alone and granular composites across all volume ratios demonstrate shear-thinning and self-healing properties (Figure S6c, Supporting Information) and representative composites are easily ejected from a syringe and injected into a model cartilage defect (Video S3, 4, Supporting Information). To determine if injecting these composites influences cell viability, granular composites either casted or injected via 21G needle are evaluated and demonstrate no statistical difference in cell viability between conditions (Figure S7, Supporting Information). Overall, these results indicate that all granular composites and their individual components are injectable and that composite properties are dependent on the formulation, likely due to the differences in properties between the elastic NorHA microgels and the viscoelastic spheroids.^[55]

Figure 3. Granular hydrogels are injectable with properties based on formulation.

a, Schematic of the composite rheology workflow. MSC spheroids and NorHA microgels are jammed independently, mixed into a composite slurry, and transferred onto a rheometer for analysis (note: all experiments performed without interparticle crosslinking). b, Representative strain sweeps (1 Hz) of spheroid only and microgel only granular components. c, Representative strain sweeps (1 Hz) of granular composites at varying volume ratios of spheroids to microgels. Filled objects: G’, Empty objects: G”. d, Quantification of average storage moduli of granular components or granular composites with varying volume ratios. n= 3, mean ± s.d. e, Quantification of average loss moduli of granular components or granular composites with varying volume ratios. n= 3, mean ± s.d. ****P<0.0001, ***P<0.001, **P<0.01, *P<0.05, ns=not significant.

Based on the granular composite simulation findings and the injectability across these formulations, granular composites with 20:80, 35:65 and 50:50 spheroid to microgel ratios are fabricated by transferring composite slurries into a 3D printed mold, exposing to visible light for 3 min. for microgel interparticle crosslinking, and removing from molds for culturing (Figure 4a). Light exposure such as this could be applied clinically using arthroscopic methods where the composite precursor material can be injected into a defect and then exposed to visible light. To visualize granular composites, constructs are cleared with RapiClear^® to evaluate whole MSC spheroids, confocal z-stacks are captured, and 3D reconstructions are performed on Imaris microscopy software (Figure 4b). Upon hydrating composites, the 50:50 granular composite lacks mechanical support (i.e., dissociates), likely due to a low microgel connectivity that is insufficient to form a stable construct (Figure S8, Supporting Information). Despite this, both 20:80 and 35:65 granular composites are mechanically stable after hydration. Expected volume ratios are validated from 3D confocal reconstructions of both groups, resulting in volume ratios within 3-5% on average of anticipated values within granular composites (Figure 4c). Both 20:80 and 35:65 groups maintain a porosity volume of ~20% (Figure 4d). The pore area (750-1250 μm²) is significantly lower in the 35:65 group compared to the 20:80 group, likely due to the increased spheroid concentration and corresponding deformability of the composite during fabrication (Figure 4d).

Figure 4. In vitro granular composite fabrication and characterization.

a, Schematic of granular composite fabrication workflow. Composite slurries are transferred into 3D printed molds (Ø = 4 mm, H= 2 mm), exposed to visible light (20 mW/cm²) for 3 min. to crosslink microgels, and then removed from molds. b, Schematic of granular composite 3D reconstruction workflow for volume validation. Granular composites are formed with MSC spheroids stained with CellTracker Red and NorHA microgels incorporating FITC-dextran for visualization and processed through confocal microscopy (~300 μm z-stack height and reconstruction via Imaris microscopy software) through RapiClear^® tissue clearing; scale bars: 200 μm. c, Quantification of granular composite % total volume for microgel and spheroid components. n=6-8, mean ± s.d. Representative images of 20:80 and 35:65 granular composites and representative 3D reconstructions; scale bars: 200μm. d, Quantification of pore area and % porosity for 20:80 and 35:65 granular composites. n=7, mean ± s.d, ROUT method (Q = 1) determined outliers for pore area and % porosity measurements. Representative confocal slices of granular composites (white) with their porosity highlighted in (black); scale bars: 200 μm. ****P<0.0001, ** P<0.01, *P<0.05, ns=not significant.

2.3. Granular Hydrogels Support MSC Chondrogenesis and Cartilage Formation

As 20:80 and 35:65 spheroid to microgel volume ratio granular composites are both injectable and stable immediately after interparticle crosslinking, these formulations are selected for culture in the presence of chondrogenic medium for up to 56 days (Figure 5a). Given previous reports of spheroid cultures collapsing over time and losing shape fidelity, monitoring of granular composite diameter and shape is used to ensure stability of the constructs with culture, which was maintained over the 56-day culture period for both conditions (Figure 5b). This indicates that the NorHA microgels provide enough mechanical support to prevent MSC spheroids from collapsing during fusion. Additionally, early timepoints (up to 7 days) demonstrate rapid cell migration budding from initial spheroids into the surrounding composite pore space, reminiscent of previous studies demonstrating enhance cell infiltration/migration with granular hydrogels (Figure S9, Supporting Information).^[27,28] Gene expression analysis via qPCR is next used to monitor MSC chondrogenesis, as well as evidence of fibrocartilage (COL1A1) and hypertrophic (COL10A1) markers (Figure S10a, Supporting Information). Over 14 days of culture, granular composites increase expression of chondrogenic markers (SOX9, ACAN, COL2A1) hundreds to thousands in fold change when normalized to undifferentiated MSCs, while fibrocartilage and hypertrophic markers increase only minimally, particularly when compared to changes in chondrogenic markers. All genes apart from COL1A1, which is significantly increased in the 35:65 group, are not influenced by the tested spheroid to microgel volume ratios.

Figure 5. Long-term chondrogenic culture of granular composites.

a, Schematic of granular composites, where spheroid fusion and growth results in cartilage tissue over time when cultured in the presence of chondrogenic factors. b, Quantification of diameters (left) and images (right) of granular composites with culture for 20:80 and 35:65 spheroid to microgel volume ratios. scale bars: 2 mm; n=3, mean ± s.d. c, Quantification of (i) dsDNA, (ii) sGAG, and (iii) collagen contents of granular composites with culture for 20:80 and 35:65 spheroid to microgel volume ratios. n=3, mean ± s.d. d, Uniaxial compression values of granular composites with culture for 20:80 and 35:65 spheroid to microgel volume ratios. n=3, mean ± s.d. e, Representative histological images of granular composites for 20:80 and 35:65 spheroid to microgel volume ratios at day 56 stained for Alcian blue (pH:1, sGAG) and collagen II (IHC); scale bars: 2 mm; inset scale bars: 500 μm. f, Quantification of area fraction % and integrated density (intensity * area) of sGAG and collagen II for granular composites cultured at day 56. n=3, mean ± s.d. **P<0.01, *P<0.05, ns=not significant.

To evaluate the production of cartilage tissue, constructs are formed in molds, removed and cultured for up to 56 days and the biochemical content, biomechanical properties, and histology are monitored. Over time, increases in dsDNA levels remain insignificant in both 20:80 and 35:65 volume ratios, indicating no drastic loss in cell viability; however, by D56, the 35:65 group has significantly higher dsDNA content compared to the 20:80 group, most likely due to the initial cell density difference (Figure 5ci). Regarding ECM content, both groups generally exhibit significant increases in sGAG and collagen over time, but the 35:65 group exhibits ~50% more ECM by D56 over the 20:80 group (Fig. 5cii, iii). We next determine the compressive properties of granular composites via uniaxial compression testing. At D1, the 20:80 granular composites exhibit ~40% greater modulus than the 35:65 group, due to the increased microgel interparticle crosslinks and microgel content. Over time, both groups exhibit increases in compressive moduli until D56, where the 35:65 group (~580 kPa) is significantly greater by ~30% than the 20:80 group (~450 kPa) (Figure 5d). These results are consistent with the observed increases in ECM content and suggest that increased uniaxial compression properties relate to greater ECM deposition. Notably, the increase in cell number in the 35:65 group when compared to the 20:80 group likely contributes to increased sGAG content and compressive moduli. Additionally, the reduction in pore area (while maintaining overall porosity) observed in initial 35:65 constructs may influence overall tissue formation. The compressive moduli values are quite high for engineered cartilage tissue, particularly with the use of adult MSCs (i.e., adult porcine MSCs), likely due to enhanced early cell-cell contacts due to use of spheroids and the ability to elaborate a matrix due to the granular structure. Additionally, the values measured for cartilage formed with granular composites at D56 are within the range previously reported for native hyaline cartilage moduli, specifically 0.1-1.6 MPa.^[56] Representative stress-strain profiles for each condition at various timepoints have been provided to show the linear viscoelastic region (10-20% strain) and that the constructs are stable within the expected loading regimes within joints, which is typically <10% (Figure S14, Supporting Information).^[57-59] Assessment of each biological replicate was evaluated to portray the variability between replicates (Figure S11, 12, Supporting Information).

To further assess ECM deposition and distribution within granular composites, sGAGs and collagens are stained at D28 and D56 (Figure 5e.; Figure S10b, Supporting Information). Quantification of percent area fraction demonstrates that the 35:65 group deposits increased areas of sGAGs and collagen II by ~20-25%, meaning less void spaces without matrix, and ~20-85% more sGAGs and collagen II intensity, meaning greater matrix deposition throughout (Figure 5f). These observations again are consistent with the measured levels of ECM and the greater microgel amounts introduced into the 20:80 constructs that can limit ECM distribution. To ensure a high balance of collagen II/I, often used to indicate hyaline cartilage, granular composites are stained for collagen I and demonstrate very little staining at either D28 or D56 (Figure S10b, Supporting Information). The NorHA microgels used in this study likely have slow degradation due to hydrolysis and enzymatic degradation, which results in ECM (sGAG only) incorporation into microgels over time. Histologically, there are void spaces that remain after 56 days, likely due to microgels that have not yet degraded or filled with ECM. Notably, there is enhanced Alcian Blue staining across the microgels when compared to collagen, which is likely due to their corresponding molecular sizes.

To better evaluate the importance of the granular structure on cartilage formation, controls where the spheroids are directly encapsulated within the NorHA hydrogels are performed, for cultures up to 28 days (Figure S13, Supporting Information). Due to the density difference between NorHA and MSC spheroids, settling of spheroids via gravity during crosslinking produces heterogeneously distributed spheroids within the NorHA hydrogels (Figure S13a, Supporting Information). By day 7, spheroids within both groups result in ~60% live area, as nanoporous hydrogels are shown to often restrict nutrient diffusion (Figure S13b,c, Supporting Information).^[47] After 28 days of chondrogenic culture, both conditions exhibit low biochemical content, with ~20-40% decreased dsDNA, 70-80% decreased sGAG, and 70-75% decreased collagen when compared to granular composites (Figure S13d, Supporting Information). Further, uniaxial compression testing demonstrates low compressive moduli (<60kPa, ~80-85% decrease compared to granular composites), which further supports lower ECM deposition and distribution compared to granular composites (Figure S13e, Supporting Information). Upon histological assessment, spheroids are observed to be loosely connected or isolated with ECM deposition localized to individual spheroids rather than distributed throughout the hydrogels (Figure S13f, Supporting Information). Taken together, granular composites allow spheroids to maintain their viaibility, fuse, and release cells to migrate within the interstitial space, which results in greater functional properties, when compared to spheroids directly encapsulated within hydrogels.

2.4. Integration of Granular Composites with Native Cartilage Tissue

Integration of neocartilage tissue with the surrounding native cartilage tissue is essential for the repair of damaged cartilage and is reported to increase as integration strength improves.^[42,60] To evaluate integration, 35:65 spheroid to microgel ratio composites, granular hydrogels alone, and cartilage plug controls are cultured for 28 days within a cartilage ring ex vivo and integration is tested via push-out tests (Figure 6a, c). Integration push-out tests reveal that after 28 days of culture the 35:65 spheroid to microgel ratio granular composite exhibits increased integration strength (~60 kPa) compared to the granular hydrogel alone control (~20 kPa); however, cartilage plug controls exhibit enhanced integration (~170 kPa) compared to all conditions (Figure 6b). The addition of spheroids within our granular hydrogels demonstrates that neocartilage tissue formed with culture increases the load being applied before failing similar to cartilage plug controls, while granular hydrogels are easily displaced from the center of the cartilage rings (Figure 6c). The enhanced integration strength of cartilage plug controls may be attributed to being press-fitted within the ring as compared to cast (all other conditions). In all cases (including the cartilage plug control) the integration of neocartilage with the surrounding native cartilage tissue are similar to other previously published integration studies; however, still remain low compared to healthy cartilage tissue.^[42,60-62] As integration of neocartilage with native cartilage is challenging, these results motivate future studies that emphasize the need for improved integration with surrounding damaged cartilage. Specifically, approaches such as pre-treatment of the cartilage surface with enzymes (e.g., collagenase) or the addition of adhesive moieties between the construct and the cartilage (e.g., aldehydes) could further improve this integration strength.

Figure 6. Integration of granular composites with surrounding native cartilage.

a, Schematic (top) and reality (bottom) of integration push-out testing setup. Scale bar: 8 mm. b, Integration strength of 35:65 granular composite with microgel only (i.e., granular hydrogel) and cartilage controls 28 days after culture. n= 7-10, mean ± s.d. c, Representative load vs displacement curves for each condition recorded during pushout tests on day 28. d, Transverse brightfield images of each condition on day 28. Scale bar: 4 mm. ****P<0.0001, *P<0.05.

3. Conclusions and Future Outlook

We demonstrate an innovative strategy where MSC spheroids and NorHA microgels are combined into granular composites and illustrate the use of the system to engineer cartilage tissue. This approach leverages desirable features of both spheroids and granular hydrogels, such as injectability through shear-thinning and self-healing properties of granular media, stability through interparticle crosslinking of microgels, cell-cell contacts to promote MSC chondrogenesis, and the formation of cartilage tissue with culture due to spheroid fusion and growth. Importantly, the use of spheroids allows for high number of cells to be included when compared to cells encapsulated within hydrogels with nanoporous structures or cell suspensions introduced within the pores of granular hydrogels where porosity is limited. Simulated mixing of granular composites using rigid body objects reveals that the initial spheroid to microgel volume ratio influences the connectivity of components within granular composites, which is supported through experimental testing of compositions with high microgel interconnectivity that result in rapid initial construct stabilization and shape maintenance with culture. These granular composites possess ~20% porosity initially, which fills with cartilage matrix during culture and results in mechanical properties that are greater than 200-fold from initial properties and are similar to those of native hyaline cartilage. Controls where spheroids are directly encapsulated within hydrogels further demonstrate the advantage of granular composites over traditional ‘bulk’ hydrogels for enabling cartilage growth, as poor cell viability, limited matrix distribution, and lower functional properties are observed due to the hydrogel nanoporous structure that limits spheroid fusion and nutrient diffusion.

Of note, evidence of microgels remains after culture, providing future opportunities to enhance microgel degradation to further support ECM distribution and reduce the potential of mechanical failure within our composites at microgel-spheroid interfaces from prolonged applied forces, as others have observed in hydrogels.^[63] While the lack of microgel degradation could negatively impact mechanical properties, this only occurs when the imperfecta does not integrate with the surrounding material; however, if integrated properly, these materials exhibit enhanced toughness compared to the single material alone.^[64,65] While no chemical bioadhesives were integrated within our system, spheroids are intrinsically adhesive to tissue given their pericellular matrix deposition and are currently used alone in clinical therapies for focal lesion repair applications that show promising post-operative results and no issues of adhesion for over 2 years.^[6] Our composite only contains partial adhesion due to mixing with microgels, which we quantified as higher than microgels alone, and will likely be protected within focal lesion fillings that are shielded from surrounding articular cartilage loading.

When comparing to prior work in cartilage tissue engineering, there are numerous advantages to this approach. The shear-thinning and self-healing properties of the granular composites enables injectability, which is missing with many biomaterial and non-biomaterial approaches, such as with non-viscous hydrogel precursors or cell suspensions that may disperse prior to crosslinking. Although there are other examples of injectable materials that could be retained within defects, such as supramolecular assemblies, these systems lack a macroporous structure that may limit cell-cell contacts and ECM deposition.^[66] Additionally, prior work with MSC spheroids exhibit comparable mechanical properties to our approach, but the lack of injectability and limited initial construct stabilization will limit clinical translation.^{[24,50,67,68]} Although construct stability with spheroids can be introduced through seeding on macroporous lattice scaffolds, this approach is again not injectable and suffers from poor spheroid seeding and uniformity.^[5,60,69,70] MSCs may also be encapsulated directly within microgels, supporting injectability, but this often results in fibrocartilage production and limited cell-cell contacts.^[30,31,71] Together, it becomes apparent that this new granular composite approach for engineered cartilage tissue is very promising to advance the field. Future work will be needed to assess within pre-clinical cartilage defects.

4. Experimental Section

Materials and Instruments.

Sodium hyaluronate (HA, mol. wt. = 66 kDa) was purchased from Lifecore Biomedical (Chaska, MN), lithium phenyl-2,4,6trimethylbenzoylphosphinate (LAP) was purchased from Colorado Photopolymer Solutions (Boulder, CO), RapiClear^® 1.49 (RC149001) was purchased from SunJin Lab Co. (Taiwan), and all immunohistochemistry supplies were purchased from Agilent (Santa Clara, CA) except for DAB chromogen stains (Sigma). Unless otherwise specified, all other reagents were purchased from Sigma-Aldrich and Fischer Scientific. Confocal microscopy: Leica SP5 II microscope. Rheology: TA Instruments AR2000ex. Uniaxial compression testing: TA Instruments Q800 DMA. Fluorescence intensity: Tecan Infinite 200 Pro with 480 nm excitation and 520 nm emission wavelengths for Picogreen assay, absorbance at 525 nm for sGAG DMMB assay, and absorbance of 560 nm for hydroxyproline assay. Histology imaging: Nikon SMZ18 stereoscope. PCR and qPCR: BIORAD CFX 96 real-time system.

Polymer synthesis and characterization.

Norbornene-modified HA (NorHA) was synthesized as previously described.^[60] Briefly, sodium HA was dissolved in DI water and mixed with Dowex 50W x 200 proton exchange resin (3:1 ratio, resin:HA) for 2 hrs., titrated to pH ~7.02-7.05 with tetrabutylammonium hydroxide (0.2 M), frozen and lyophilized. HA-TBA methylation peak was determined, normalized to the methyl peak on HA (δ~2.0-1.8) and measured (δ~0.7-.9; 17.21) via ¹H-NMR (Figure S15a, Supporting Information). HA-TBA was modified with norbornene through benzotriazole-1-yl-oxy-tris-(dimethylamino)-phosphonium hexafluorophosphate (BOP) coupling. Briefly, HA-TBA and 5-norbornene-2-methylamine was dissolved with anhydrous dimethyl sulfoxide (DMSO), BOP was cannulated into the reagent solution and allowed to react for 2 hrs. under N₂. The reaction was quenched with cold DI water, placed on dialysis (6-8 kDa mesh cutoff) for 2 days with DI water and salt and a subsequent 3 days with DI water, frozen and lyophilized. The extent of norbornene modification was determined via ¹H-NMR to be ~20-22% (δ~5.8-6.2, 2H) of the disaccharide repeat units of HA when normalized to the methyl peak on HA (Figure S15b, Supporting Information).

Microgel fabrication and size characterization.

NorHA microgels were fabricated via water-in-oil emulsion. A 3% NorHA precursor solution was prepared with 0.05% I2959, 10 mM dithiothreitol (DTT), and 1x phosphate buffer saline (PBS), added to oil phase (98% light mineral oil/ 2% Span 80) dropwise, allowed to stir for 30 sec., and crosslinked with a UV (320-390 nm) omnicure lamp at 20 mW cm⁻² for 30 min. Oil phase stir speed was varied to obtain varying size distributions. The reaction was carried out in a 100mL glass beaker (Ø:4.5cm; H: 6.5cm) with a stir bar (L: 4cm; W: 1mm). Microgels were washed by discarding oil phase, soaking in 1% tween 20 solution for 10 min., and washing with 1x PBS, which was repeated 5x. Microgels were collected, centrifuged at 15,000 x g for 5 min. and stored at 4 °C before use in experiments. Fluorescein isothiocyanate (FITC)-dextran (MW: 2 million Da) was added into the NorHA precursor solution to allow visualization of microgels for size characterization. Microgels were visualized with a fluorescent microscope, thresholded, and sizes (ferret diameter) were measured via analyze particle function (FIJI).

Microwell fabrication.

Microwell pyramid arrays were fabricated from a poly(dimethyl)siloxane (PDMS) molding process to generate spheroids. PDMS was mixed in a 10:1 polymer to crosslinker ratio, degassed, added into commercial microwell plates (StemCell Technologies, AggreWell^™ 400) and placed into an oven at 80°C for 2 hrs. to form a negative mold. Negative molds were removed, surface treated with Trichloro(1H,1H,2H,2H-perfluorooctyl)silane for 45 min. via vapor deposition after O₂ plasma activation with plasma etcher (Plasma Etch) for 3 min. Positive PDMS molds were created by adding uncured PDMS into each 6 well, degassing, positioning the negative PDMS mold, and curing in the oven at 80 °C for 2 hrs. PDMS positive molds were soaked in isopropyl alcohol for 1 hr., left in DI water overnight, and sterilized under a UV germicidal tissue culture hood for 25 min. Each well was rinsed with 1x PBS and an anti-adherence rinsing solution (StemCell Technologies, Aggrewell^™) was added into each well. Plates were centrifuged at 2000 x g for 5 min to remove air bubbles. The wells were aspirated, rinsed with basal medium, and placed in an incubator (37 °C/ 5% CO₂) with chondrogenic medium until cell seeding.

Cell isolation and expansion.

Cells were obtained from adult Yucatan mini-pigs (porcine), specifically from femoral tissue discarded from other studies (no approval needed for use). Femoral bone marrow was extracted and MSCs were isolated via plastic adherence during culture in Dulbecco’s modified eagle medium (DMEM) with 10% fetal bovine serum (FBS), and 1% penicillin/streptomycin (P/S) according to previous literature.^[72] All donors were seeded at 5000 cells cm⁻², with alpha modified eagle medium (α-MEM) supplemented with 10% FBS, 1% P/S, and fibroblast growth factor (FGF, 1 ng mL⁻¹).

Spheroid formation, size and viability characterization, and culture.

Spheroid Formation. Porcine MSCs were rinsed with 1x PBS, trypsinized, and resuspended in chondrogenic medium at varying densities, such that the total volume in each well was 5 mL. Cells were allowed 48 hrs. to condensate into spheroids. Chondrogenic medium consisted of DMEM + Glutamax, 10% FBS, 1% P/S, 1% ITS, 1 x 10⁻³ M sodium pyruvate, 50 μg ml⁻¹ ascorbic acid 2-phosphate, 40 μg ml⁻¹ L-proline, 1 x 10⁻⁷ M dexamethasone, and 10 ng ml⁻¹ transforming growth factor-β3 (TGF-β3). Spheroid Viability. Spheroid viability was evaluated at days 1 and 7 via Calcein AM (Invitrogen, 2 μM) and ethidium homodimer (Invitrogen, 4 μM) by staining for 3 hrs. at 4 °C to allow diffusion throughout the spheroid. Cell viability was quantified from confocal stacks and reported as % live cell area, which was calculated by the ratio of Calcein AM stained area to the combined Calcein AM and ethidium homodimer stained area. Spheroid Culture. MSCs were seeded at 1000 cells spheroid⁻¹ in chondrogenic medium and spheroid growth were monitored (ferrets diameter) for 28 days of culture. Chondrogenic medium was replenished every 2-3 days, imaged with a brightfield phase contract microscope every 7 days, and quantified via analyze particle function (FIJI). Spheroid volume was calculated assuming a perfect sphere.

Immunofluorescent staining.

Spheroids were stained at varying culture periods (day 1, 7, 28) by fixing in 10% formalin at 4 °C overnight, blocking with 1% bovine serum albumin (BSA), and washing with 1x PBS. Samples were blocked, and anti-chondroitin sulfate mouse monoclonal antibodies (Sigma SAB4200696, 1:200) and anti-collagen II mouse monoclonal antibodies (DSHB II-II6B3, 1:100) were introduced overnight at 4 °C. Secondary antibody staining was performed using goat anti-mouse polyclonal IgG (H+L) AlexaFluor^™ 647 (Invitrogen A-20990, 1:1000) for 2 hrs. at 25 °C. Samples were washed, permeabilized in 2% triton X-100 overnight at 25 °C, and cleared in 2 hrs. of adding RapiClear^® solution. Confocal z-stacks (~150 μm thickness) were taken to capture chondroitin sulfate and collagen II matrix deposition.

Granular composite fabrication, viability and long-term chondrogenic cultures.

MSC spheroids were disrupted from the microwells via mixing with pipette, washed with 1x PBS, and transferred to a conical tube on ice. Spheroid suspensions were centrifuged at 300 x g for 20 sec. and NorHA microgels were centrifuged at 15,000 x g for 5 min. to determine total volume. 2.5 mM DTT and 0.05% LAP photoinitiator was added to the microgels to allow for interparticle crosslinking, vortexed, and centrifuged at 15,000 x g for 5 min. The spheroids and microgels were then added together at varying volume ratios, manually mixed with a spatula, centrifuged for 20 sec at 300 x g, and transferred into 3D printed molds (Ø: 4 mm, H: 2 mm) to photocrosslink with a visible light (400-500 nm) omnicure lamp at 20 mW cm⁻² for 3 min. Granular composites were removed from 3D printed molds and placed within a 24-well non-adherent TC plate with chondrogenic medium. Viability was assessed immediately after crosslinking via Calcein AM (Invitrogen, 2 μM) and ethidium homodimer (Invitrogen, 4 μM) by staining for 3 hrs. at 4 °C to allow diffusion throughout the construct. Cell viability was quantified from confocal stacks and reported as % live cell area, which was calculated by the ratio of Calcein AM stained area to the combined Calcein AM and ethidium homodimer stained areas. Long-term chondrogenic cultures were maintained up to 56 days and the medium was replenished every 2-3 days. Gross granular composite diameters were manually measured at varying culture periods (day 1, 28, 56) via measure tool (FIJI).

Spheroid encapsulated NorHA hydrogel control fabrication.

MSC spheroids were condensed as previously described and subsequently incorporated within NorHA prepolymer solution (identical formulation as described previously) at 20:80 and 35:65 spheroid: polymer ratios and photocrosslinked within cut syringe tips (~4.5mm diameter) for 3 minutes under visible light (identical power intensity to granular composites). Spheroid encapsulated NorHA hydrogels were then cultured for up to 28 days in chondrogenic medium as previously described. Briefly, on day 7 spheroid viability was conducted by staining samples with Calcein AM (Invitrogen, 2 μM) and ethidium homodimer (Invitrogen, 4 μM) for 1 hr. at 37 °C. On day 28, histological IHC and Alcian blue staining, biochemical content, and uniaxial compression testing were conducted identically to long-term granular composite chondrogenic cultures.

Granular composite 3D reconstructions, volume, and porosity analysis.

To image granular composites, MSCs were stained with CellTracker Red CMTPX and (FITC)-dextran was added into NorHA microgels as previously described above. Granular composites were fabricated, immediately fixed with 10% formalin overnight at 4 °C, permeabilized with 2% triton X-100 overnight at 25 °C and cleared with RapiClear^® solution for 2 hrs. before imaging. Timepoint cell migration studies were conducted identically except for culture time within chondrogenic medium (day 0, 1, 3, 7). Confocal z-stacks (~300 μm) were imaged, and 3D reconstructions were created in Imaris 9.8 Microscopy Software. Total volume of each phase was exported and used to validate % total volume of the granular composite components. % Total volume is reported as volume of microgels or spheroids normalized to the total volume of microgels and spheroids * 100. Pore volume was excluded from this calculation. Porosity area and % porosity of granular composites was manually measured via slices from confocal z-stacks (FIJI). % Porosity is reported as the % area of binary pores within the slice normalized to the total % area of both granular composite and pores.

Mechanical testing.

Rheological Testing. Granular composites were transferred onto the base platform while a 20 mm acrylic parallel plate was lowered to a gap height of 1 mm. Oscillatory rheological time sweeps (1% strain, 1 Hz) were performed to obtain the average storage (G’) and loss (G”) moduli. Oscillatory rheological strain sweeps (1-100% strain, 1 Hz) were performed to demonstrate shear-thinning properties. Oscillatory strain on/off sweep (500%/0% strain, 1 Hz) were performed to demonstrate self-healing properties. Granular microgels or spheroids were tested with oscillatory rheological time sweeps. Uniaxial Compression Testing. Granular hydrogels and granular composites of varying volume ratios were crosslinked into cylindrical disks (Ø: 4 mm, H: 2 mm). Granular hydrogels crosslinked at varying exposure times (0.5, 1, 3, 5 min.; 400-500 nm 20 mW cm⁻²) were tested to achieve an optimal crosslink time. Granular composites were tested at varying culture periods (1, 28, 56 days). Granular hydrogels and day 1 granular composites were compressed (0.05 N min⁻¹) with a preload of 0.001 N. Day 28 and 56 granular composites were compressed to (0.5 N min⁻¹) with a preload 0.05 N. Compressive moduli were calculated and reported as the stress-strain slope between 10-20% strain. Day 28 spheroid hydrogels were tested and analyzed identically to granular composites at the same timepoint.

Biochemical Content Analysis.

Composite Digestion. Granular composites and spheroid NorHA hydrogels were digested as previously described with slight modifications.^[60] Briefly, samples were homogenized in 100 μg mL⁻¹ of proteinase K, and 1 mg mL⁻¹ hyaluronidase (750-3000 U mL⁻¹) RNAse free solution. Samples were homogenized (Fisher Scientific, Bead Mill 24; 3.55 m s⁻¹; three 5 min. cycles) with a 5 mm S.S. bead and placed in a 65 °C water bath to digest for 24-48 hrs. Picogreen dsDNA Biochemical Analysis. Total dsDNA content was measured with a Picogreen dsDNA assay kit (Invitrogen, P7589) according to manufacturer’s protocol and previous literature.^[73] Samples were diluted 1:10 (100 μL) compared to standards to fit within the working curve. Dimethylmethylene Blue (DMMB) sGAG Biochemical Analysis. Total sGAG content was measured via DMMB colorimetric assay as previously published.^[73] Briefly, chondroitin sulfate (Fischer Scientific, AAJ6034106) standards were created and samples were diluted 1:20 compared to standards to fit within the working curve. Hydroxyproline Collagen Biochemical Analysis. Total collagen content was measured with a hydroxyproline assay kit (Abcam, ab222941) according to manufacturer’s protocol. Briefly, 10 N NaOH was added to each sample, heated to 120 °C for 1.5 hrs., and neutralized with 10 N HCl once samples reach 25 °C. Collagen content was calculated by hydroxyproline: collagen ratio 7.14:1 as previously reported in literature.^[73] All biological replicates are reported individually in Figure S9, Supporting Information.

Histology and Immunohistochemistry.

Histology. Granular composite and spheroid hydrogels were fixed overnight at 4 °C in 10% formalin, dehydrated with ethanol (50-100 v/v%), embedded with high-melt paraffin wax, and sectioned (5 μm) prior to staining. All samples were deparaffinized with citrisolv detergent and dehydrated before staining. sGAG matrix deposition was visualized with Alcian blue staining (Newcomer Supply, 1% pH:1). Samples were mounted in a toluene-based mounting solution and sealed with clear nail polish. Immunohistochemistry. Samples were stained as previously described.^[60] Collagen I, II, and chondroitin sulfate were visualized via staining for anti-collagen type I mouse monoclonal antibodies (Sigma MAB3391, 1:100), anti-collagen II mouse monoclonal antibodies (DSHB II-II6B3, 1:100), anti-chondroitin sulfate mouse monoclonal antibodies (Sigma SAB4200696, 1:200) overnight at 4 °C., followed by biotinylated secondary antibody staining and Streptavidin HRP for 10 min. DAB chromogen staining (1:25, A to B chromogens) was used to visualize stains by incubating for 10 min at 25 °C (Millipore). Samples were mounted in an aqueous-based mounting solution and sealed with clear nail polish. Histology images were white-balance corrected and % area fraction and integrated density was calculated (FIJI).

Gene expression analysis.

RNA Isolation, Polymerase Chain Reaction (PCR), and quantitative PCR (qPCR).

Samples were flash frozen with liquid N₂ and sample RNA was extracted on ice via TRIzol reagent method described previously with slight modifications.^[74] Briefly, samples were homogenized for 30 sec. at 3.55 m s⁻¹ cycle speed in TRIzol reagent, followed by phase separation with chloroform and precipitation with isopropanol. RNA was washed 2x with ethanol and resuspended in 60 μL of RNAse-free water. Total RNA was reverse transcribed with a high-capacity cDNA reverse transcription kit (Applied Biosystems, 4368814) according to manufacturer’s instructions. qPCR was carried out using SYBR green master mix (Applied Biosystems, A25742) according to manufacturer’s protocols. Primer forward and reverse sequences are reported in Table S2. qPCR was measured in duplicate and fold change was reported as ΔΔCT method normalizing to GAPDH housekeeping gene and then undifferentiated MSCs.

Cinema4D connectivity simulations and analysis.

Connectivity simulations. To simulate mixing of composites, microgels and spheroids were simulated as rigid body spheres and the ratio of spheroids and microgels were varied to analyze connectivity of each component. Each particle was assigned a dynamic body tag to introduce rigid body dynamic physics for mixing simulations. A cylindrical tube and 2 planes were used to contain mixing simulation and were given collision body tags. Simulations were run for 70 frames per second, baked, and the current state was converted to an object that was exported as a .stl file. To introduce microgel size distributions, a step mograph effector was created and a spline was matched to an experimental size distribution as demonstrated in Supplemental Figure 4. To introduce spheroid aggregation, a force mograph effector was introduced to the spheroids. For all simulations, gravity dynamics were set to parameters: Time Scale = 100% and Gravity = 10 cm. Simulation analysis. Files were uploaded onto FIJI’s 3D Viewer, voxelized, smoothing and fill holes binary functions (FIJI). Connectivity was analyzed with 3D Objects Counter (FIJI) that determines the number of connected objects within a 3D space.^[75] Simulations were repeated 5 times with different seeds for each composite to introduce variability. A more detailed simulations methods section can be found in the Supporting Information.

Integration push-out tests.

Push-out tests.

Cartilage rings (Ø: 8 mm) were harvested from juvenile bovine femoral condyles via biopsy punch. The bone layer was removed to obtain only cartilage tissue. A 4 mm hole was introduced into the center of each cartilage disk via biopsy punch to create a ring that can hold composites and controls. Granular composites and hydrogels were casted and cartilage plugs were press-fit into the cartilage rings. All conditions were subjected to 15 min. of incubation at 37°C to allow spheroid fusion before interparticle crosslinking was initiated with visible light for 3 min @ 20 mW/cm². All conditions were cultured in 4 mL of chondrogenic medium as defined above and medium was changed every 1-2 days. On day 28, integration samples were subjected to push-out tests using the uniaxial compression settings on a HR20 (TA instruments) rheometer with a custom 3D printed push out mold (Ø: 0.19 mm) setup. Each ring was compressed (0.2 mm s⁻¹) until a displacement of 3.5 mm was reached. Representative Load vs. Displacement graphs only plot until failure occurs. Integration strength was calculated by dividing load at failure by frontal surface area of constructs (4 mm).

NorHa bulk and microgel swelling, mass loss and degradation.

Microgel swelling.

NorHA microgels were fabricated and washed as previously stated above. Microgels were jammed at 15k x g for 3 min and excess liquid was removed. Using a spatula, 0.3mL of microgels were transferred to a new Eppendorf tube where 1mL of 1X PBS was added. Samples were vortexed and placed in 37°C shaking incubator. Samples aliquots were taken at each timepoint for image visualization and swelling manual measurements (FIJI). Bulk NorHA hydrogels were made at similar formulations, pipetted into mold and photocrosslinked for 3 min. Wet weights were taken at each timepoint and hydrogels were frozen and lyophilized to remove water content to obtain the dry weight. Mass ratio is reported as (wet mass - dry mass)/wet mass. Uronic acid release was determined by collecting the supernatant at each time point and determining the concentration using a uronic acid assay as previously described.^[76]

Statistical analysis.

All statistical tests were performed in GraphPad Prism 9. Comparison between two groups were analyzed with a two-tailed Student’s T test (if normality is assumed) or a Mann-Whitney test (if normality is not assumed). Normality calculated with D’Agostino & Pearson test; α = 0.05. Comparison between groups > 2 were analyzed with one-way or two-way ANOVA and multiple comparisons between groups were analyzed with Tukey multiple comparisons test with an α = 0.05 and 0.95 confidence interval. Bar graph descriptors (mean ± standard deviation, s.d.) are displayed within the figure panel caption. Biological replicates are stated in figure captions. All statistical test p, q, t, and DF values and a more detailed description of statistical rational are reported in Table S3.

Data Availability Statement

All data and analysis generated during this study can be found within the main text, supplemental text, and supplemental videos. All the data that support the findings of this study are available from the corresponding author upon request. Gene expression data was omitted from repositories (i.e. Gene Expression Omnibus, Array Express) due to low-throughput assay (<20 genes not accepted). Raw and calculated gene expression data are available from the corresponding author upon request. All statistical parameters for all analyzed data are available in Supplemental Table 2 within the Supplemental Information document.