Granular Biomaterials Made From Ground Tissue ECM For Flowable Graft Applications

Introduction: Osteoarthritis (OA) is a degenerative joint disease that afflicts millions of patients in the US [1]. Current solutions for younger patients with OA who do not qualify for total knee replacements are primarily anti-inflammatory drugs that don’t restore the afflicted cartilage and bone tissue. There is a need for a flowable graft that recapitulates native cartilage and bone, and that can be used to create realistic in vitro models for OA drug development, or even as an implantable tissue filler. A top-down approach to biomaterial design, such as repurposing existing extracellular matrix (ECM), can be an efficient way to capture the full array of ECM components. By processing decellularized ECM into microparticles, the structural architecture of native tissue can additionally be preserved as a microenvironment for infiltrating cells. These ECM particles can be packed into a hydrogel base at high density [2] to achieve a viscous yet extrudable granular biomaterial that serves as a flowable graft. Furthermore, the resulting shear-thinning properties make this granular ECM biomaterial a promising option for 3D bioprinting [3], which enables specific size and shape customization for both in vitro models and implants. In this study, we present a hyaluronic acid-based granular cartilage ECM biomaterial ink, with compositional, rheological, and mechanical characterization, and proof-of-concept 3D printed constructs and cellular responses. There are many considerations in the translation of tissue-based biomaterials that we account for in our granular biomaterial, including decellularization methods, viral inactivation, manufacturing scheduling, and animal donor variability.

Materials and

Methods: Cartilage and bone ECM was sourced from AlloSource (human; Centennial, CO) and a local abattoir (porcine), and decellularized using viral inactivation (VI) in 0.2M HCl and 1.0M NaOH. A subset of porcine cartilage underwent an additional SDS detergent treatment. Tissues were lyophilized, pulverized, and sorted to < 100um diameter via a micro sieve (Electron Microscopy Sciences, Hatfield PA). Particle size distribution of particles suspended in water was measured in a laser diffraction particle size analyzer (Mastersizer 3000). DNA content was quantified using a PicoGreen assay and DAPI staining. Biomaterial ink formulations consisted of 10mg/mL hyaluronic acid (HA) (~20% thiolation) packed with 0.2g/mL cartilage lyophilized particles. A room temperature rheological shear sweep (0.01 to 250 s-1) was performed within 5 mins of mixing (AR-G2, TA Instruments). Molded cylindrical biomaterial constructs were polymerized for 30 min at 37°C. For mechanical testing, constructs were compressed (unconfined) to 40% strain at quasi-static 0.1%/s loading (Bose ElectroForce 5500 system). For particle density analysis, constructs were stained with Ghost Dye, imaged via confocal microscopy, and thresholded on Image J to calculate area of particles to total area. Biomaterial inks were 3D printed on a pneumatic extrusion printer with 18-20 G nozzles (BioX, Cell Ink Life Sciences) and varying pressure and flow rate parameters. Pre-stained particle inks were 3D printed into a molded Sylgard 184 polydimethylsiloxane (PDMS) tissue-on-chip model and imaged using confocal microscopy. Primary bovine chondrocytes were cultured with cartilage particles and hyaluronic acid in proof-of-concept PDMS microfluidic chips, and LIVE/DEAD stained at 2 different time points.

Results, Conclusions, and Discussions: In past work, we utilized an SDS process to decellularize tissues [1], with a preliminary viral inactivation (VI) process to streamline manufacturing. We removed the SDS step to fit our process within an 8-hour manufacturing window, and now process donors individually to comply with human donor safety requirements and to assess donor variability. This new process sufficiently reduces DNA content across donors (Fig 1A) to below the industry standard of 50 ng/mg dry tissue [4] for porcine cartilage and bone, and human cartilage. Ground bone matrix (GBM) and demineralized bone matrix (DBM) also demonstrated acceptable dsDNA content without further processing. SDS was removed from our process also to eliminate risk of cytotoxicity from residual SDS. In qualitatively assessing our DNA content using DAPI staining (Fig 1B), we found areas of high pixel saturation on particles treated with SDS, likely due to nonspecific binding of the DAPI stain to residual SDS. We found a minimal effect of SDS on dsDNA content, but a decrease in viscosity (Fig 1D) and polymerized compressive modulus (Fig 1E) was observed in VI Only compared to VI + SDS porcine cartilage. Particle size distribution was characterized for porcine cartilage particles (Fig 1C), demonstrating a peak at around 127µm particle diameter and range all the way up to 400µm. This contrasts with the lyophilized particles, which are size sorted to < 100µm diameter, suggesting that particles swell when hydrated. Hydrated particle size distribution can inform our granular biomaterial particle density and 3D bioprinting resolution. Particle density for porcine cartilage granular biomaterial was quantified to be around 80% (Fig 1F), suggesting that a percolation network between particles has formed [2]. At this high packing density, particles are essentially re-assembled into a tissue graft. Proof-of-concept osteochondral scaffolds (Fig 1G) and PDMS “on chip” models (Fig 1H) were successfully 3D bioprinted. Over a 14-day period, chondrocytes cultured with cartilage particles on this PDMS chip demonstrated viability, proliferation, and adherence to cartilage particles (Fig 1I). Future work will be done to characterize porcine and human bone and cartilage granular biomaterials, and cellular response to these biomaterials for translation as flowable grafts.

Acknowledgements (Optional): Acknowledgements: This work was supported by NIH U01 grant. We also acknowledge AlloSource for providing human donor tissue.

References: [1] Xu, Y et al. Sci. Rep., 11:1-10, 2021. [2] Barthold, J.+, Adv. Funct. Mat., 31:2103355, 2021. [3] Barthold, J.+, Biofab., 14:025021, 2022. [4] Crapo, P., Badylak, S., et al. Biomat., 32:12:4, 2011.