Ribbon-shaped microgels as bioinks for 3D bioprinting of anisotropic tissue structures

Abstract

Granular microgels are attractive bioinks for bioprinting due to their injectability, printability, modularity, and enhanced macroporosity compared to conventional nanoporous hydrogels. Despite the potential of microgels for bioprinting, most previous work has relied on spherical microgels and produced isotropic tissues, whereas many native tissues are inherently anisotropic. While emerging studies have explored non-spherical microgels for bioprinting, there remains a need for bioinks that support cell alignment and tunable niche cues. Microribbons (μRB) are anisotropic ribbon-shaped microgels, but the potential of μRBs as bioinks for printing 3D anisotropic tissues remains unexplored. Here, we report the development of μRBs with tunable stiffness as bioinks for extrusion-based bioprinting and demonstrate that μRB bioinks maintain excellent printability and align during extrusion. μRB bioinks support alignment of MSCs and endothelial cells, with greater alignment as μRB stiffness increases. Increasing μRB stiffness also accelerates mesenchymal stromal cell osteogenesis in 3D. Finally, we demonstrate the potential of μRB bioinks for modeling breast cancer-bone metastasis, which features spatial patterning of multiple cell types to model cancer cell invasion at the tissue interface. Together, these results establish ribbon-shaped microgels as a new class of anisotropic bioinks, offering a versatile platform to support a broad range of bioprinting applications.

Graphical abstract

Highlights

• •Ribbon-shaped microgels aligned during bioprinting in a stiffness-dependent manner.
• •Gelatin μRBs were a versatile bioink platform and exhibited rheological characteristics suitable for bioprinting.
• •Cells oriented along individual μRBs and within organized printed constructs.
• •μRB bioinks supported stiffness-modulated MSC osteogenesis and multi-tissue cancer-bone modeling.

1. Introduction

Bioprinting is an advanced fabrication technology that enables precise spatial patterning of cells and biomaterials [[1], [2], [3]]. Unlike conventional bulk fabrication methods, bioprinting allows for the creation of complex tissue structures with defined architecture and has been widely used in regenerative medicine and disease modeling [[4], [5], [6], [7], [8], [9]]. Conventional extrusion-based bioinks, typically made from hydrogel precursors such as gelatin, hyaluronic acid, and alginate [[10], [11], [12], [13], [14]], result in nanoporous hydrogels that can hinder cell survival, proliferation and migration [15,16]. To overcome this limitation, granular hydrogels have emerged as a promising alternative solution. They are formed by jamming microgels via centrifugation or vacuum filtration and then crosslinking to form a macroporous scaffold [[17], [18], [19]]. Granular hydrogels are injectable and offer macroporosity and modularity that conventional hydrogels lack, making them well-suited for bioprinting applications [20,21].

Currently, most granular hydrogels are assembled from spherical microgels, which form isotropic tissue structures with identical properties in all directions. [22,23]. However, many biological tissues, such as bone, blood vessels and tendons, are inherently anisotropic, exhibiting direction-dependent structures and functions [24,25]. To better mimic the anisotropic features of native tissues, recent studies have begun exploring microgels with non-spherical shapes as bioinks, for guiding directional cellular and extracellular matrix (ECM) organization [[26], [27], [28]]. For example, rod-shaped microgels have been shown to enhance endothelial cell sprouting in vitro and promote cell infiltration and tissue integration in vivo [28]. The shear stress of extrusion-based bioprinting can induce alignment of collagen fibrils, resulting in cell alignment and tissue organization [29]. Strand-shaped microgels fabricated by extruding bulk hydrogels through grids also offer bioinks with high aspect ratio, but they rely on entanglement for scaffold formation with limited tunability [27]. While electrospun nanofibers can promote cell alignment, the resulting scaffolds are nanoporous and cells need to be seeded after fabrication, making them unsuitable for cell-laden bioprinting [[30], [31], [32]]. Thus, there remains a need for bioinks that can support homogeneous cell distribution during printing, while guiding cell alignment in 3D and can crosslink into macroporous scaffolds with anisotropic properties.

To address this need, we turned to gelatin microribbons (μRB), ribbon-shaped microgels produced by wet-spinning [33]. To fabricate gelatin μRBs, wet-spun gelatin microfibers were transferred into acetone, causing asymmetrical collapse of microfibers into flattened ribbon shape. In contrast to fibers, with circular cross-sections, the flattened surface area of ribbons enhances contact area between adjacent strands and improve mechanical and shock absorbing properties. We have previously shown that various natural ECM-derived polymers (i.e. gelatin, hyaluronic acid, chondroitin sulfate) and synthetic polymers (i.e. PLGA, polyethylene glycol diacrylate) can be fabricated into photocrosslinkable μRBs using wet-spinning, offering modularity and enabling broad tunability of material composition and ribbon geometry [[33], [34], [35], [36]]. μRBs also support homogeneous cell encapsulation and the resulting 3D cell-laden scaffolds are inherently macroporous, making them attractive candidates for use as bioinks. Using conventional mold fabrication methods, our group has shown that μRBs can support the culture of diverse cell types and regeneration of multiple anisotropic tissues, including cartilage [34], bone-tendon interface [37], and smooth muscle [38]. For example, when mesenchymal stromal cells were encapsulated within multi-layered μRBs of varying compositions, they underwent zonal differentiation, generating cartilage with stratified structures that mimic superficial to deep zone transition [34]. Notably, individual μRBs have been shown to promote alignment of human adipose-derived stromal cells on their surface [33]. Similarly, smooth muscle cells seeded on aligned gelatin μRB scaffolds formed aligned smooth muscle tissues, demonstrating structural anisotropy [38]. In addition to tissue engineering applications, spatially patterned μRBs have also been used to model 3D cancer metastasis at the bone-bone marrow tissue interfaces [39]. Importantly, crosslinked μRB scaffolds demonstrate superior shock-absorbing mechanical properties [33,40,41]. However, all previous studies on μRBs have relied on mold-based fabrication, and μRBs have never been used for bioprinting. The potential of μRBs as bioinks for printing 3D tissues with anisotropic features remains unexplored.

Here, to demonstrate the tunability of μRB bioinks for various applications, we fabricated gelatin μRBs with tunable stiffnesses [42], and assessed their alignment and printability. We choose to vary μRB stiffness as a main parameter for two reasons. First, stiffness of spherical microgels has been shown to strongly influences key properties of granular hydrogel bioinks, including cell viability, printing fidelity, and structural stability [21]. Second, substrate stiffness is a key regulator of cell fate, including differentiation, matrix production, and migration [[43], [44], [45]]. The μRB bioinks with tunable stiffness were thoroughly characterized by assessing the rheological properties, printability, shape fidelity, cytocompatibility, and structural stability. As a proof of concept, we demonstrated three applications using μRB bioinks with tunable stiffness: guiding cell alignment, enhancing stromal cell osteogenesis [44], and spatial patterning to model cancer invasion at tissue interface. Together, these results establish μRBs as a new class of microgel-based bioinks for bioprinting, offering tunable and modular features to support a broad range of bioprinting applications.

2. Methods

2.1. Preparation of gelatin microribbon (μRB) and microgel (μGel) bioink

Materials. Type-A gelatin, dimethyl sulfoxide (DMSO), Methacrylic acid N-hydroxysuccinimide ester (MA-NHS), glutaraldehyde, L-lysine hydrochloride, and lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP) were purchased from Sigma-Aldrich. All chemicals were used as received.

Synthesis of the gelatin microribbons by wet spinning. To prepare a spinnable solution, type-A gelatin (20 g) was stirred in 80 ml DMSO at 60 °C overnight, forming a viscous solution of gelatin. To carry out the wet-spinning process to produce microribbons, the gelatin/DMSO solution was transferred into a 50-ml syringe, loaded into a syringe pump, and ejected through a 16-gauge needle at 5 ml h⁻¹ at room temperature into a tank of anhydrous ethanol (4 L), which was situated 6 feet under the syringe outlet and stirred by a 5 cm long magnet at 1100 rpm. The microfibers were collected and transferred into acetone overnight to form μRBs. To integrate methacrylate groups, which enable photocrosslinking, the μRBs were stirred in a mixture of MA-NHS and methanol. To obtain μRBs with tunable stiffness, methacrylated μRBs were internally crosslinked overnight in methanol (200 ml) at room temperature with varying amounts of glutaraldehyde equivalent to reacting with 5 %, 10 %, or 15 % of amine groups on gelatin. To quench any residual glutaraldehyde, the aldehyde-fixed μRBs were neutralized with a 20X molar excess of L-lysine hydrochloride in PBS at 25 °C for 12 h, then washed 3 times with DI water. This step minimized any potential cytotoxicity from unreacted aldehyde groups.

Synthesis of the gelatin microgels by fragmentation. 20 % Gelatin aqueous solution was transferred into a 5 ml syringe and incubated at 4 °C for 30 min. The bulk gelatin hydrogel was extruded with 18G, 20G, and 22G needles into cold water. The collected μGels were transferred into a mixture of MA-NHS and methanol. Subsequent crosslinking, washing, and quenching steps were conducted in the same manner as the preparation of stiff μRBs to ensure fair comparison.

Characterizations of the gelatin microribbons. Fourier-transform infrared spectroscopy (FTIR) analysis was conducted using a Nicolet iS50 FTIR spectrometer (Thermo Scientific). Lyophilized samples were measured in a 4000-750 cm-1 window, with 1 cm-1 resolution.

For scanning electron microscopy (SEM) imaging, lyophilized samples were mounted on SEM studs using Ag paste, followed by Au/Pd (60:40 ratio) sputter-coating. Images were acquired using an Apreo S LoVac SEM (Thermo Fisher Scientific), with an accelerating voltage of 5 kV, 50 pA beam current, and an Everhart-Thornley Detector. The Young's modulus was measured by BioScope Resolve atomic force microscope (Bruker) using the standard force volume with an 8 μm × 8 μm matrix, with 3 matrices for each group. Sample measurements were conducted at a ramp rate of 1 Hz, a 5 μm ramp size, and up to 100 nN trigger threshold. The modulus was determined by fitting the curve in NanoScope Analysis (Bruker) using the Hertzian model. To characterize the stability of bioprinted μRB scaffolds at 37 °C, μRB bioinks with varying stiffnesses were fabricated into acellular scaffolds and incubated in PBS for 14 days or were used to encapsulate MSCs and cultured in osteogenic medium for 28 days. Scaffold stability was characterized using brightfield microscopy and tracking wet weight over time.

Preparation and bioprinting of the μRB bioink. Prior to 3D bioprinting, the μRB bioink was compacted by centrifugation at 4300×g for 10 min. To visualize the width of μRBs and the porosity of their scaffolds, the μRBs were stained with FITC (0.2 μg ml⁻¹, Sigma-Aldrich) for 30 min. The μRBs were then imaged using a Leica Stellaris 5 confocal microscope, and the μRB width and μRB bioink porosity were determined using FIJI. 3D printing was performed with a lab-made 3D printer kindly provided by Prof. Mark Skylar-Scott's lab. The μRB bioinks were loaded into 3 ml syringes and extruded with 18G or 20G tapered tip.

2.2. Characterization of μRB bioink and bioprinting

Extrudability of μRB bioinks by injection force and rheology. The method was carried out as previously described [[46], [47], [48], [49]]. Briefly, to measure the injection force, 3 ml syringes loaded with the μRB bioinks were installed on the MTS system to record the force applied while pushing the plunger at a constant extrusion speed of 2 ml h⁻¹. To measure the rheological properties of the μRB bioinks, samples were loaded onto an ARG2 stress-controlled rheometer (TA Instruments) equipped with a 20 mm diameter parallel plate geometry with a 1.0 mm gap height, which is ten time larger than the thickness of μRB, as suggested by the method from National Institute of Standards and Technology [50]. For strain sweep tests, samples were subjected to a strain range from 1 % to 500 % at a frequency of 1 Hz at 25 °C. The self-healing test was measured alternatively at both a high-magnitude strain (500 %) and a low-magnitude strain (1 %). Samples were subjected to shear stress-sweep tests over a stress range of 0.01–1000 Pa at a frequency of 1 Hz at 25 °C.

Printability assessment of μRBs bioinks. To determine filament uniformity in each image of the 2D grid pattern, 15 thickness measurements were taken along each of the printed filaments and compared to the specified thickness in the printer input. stl files. To quantify printability, FIJI was used to determine the perimeter and area of each square printed within the 2D grids [51,52].

Equation (1) is used to calculate printability

$$\mathrm{Pr}=L2/A$$ (1)

where L is the perimeter and A is the area. To measuring hanging length, filaments were extruded through the needle to find the filament collapse length, where the longest hanging lengths were recorded. For the height accuracy, height analysis of printed grids with 6, 9, and 12 layers were performed by comparing the height to the. stl files.

2.3. In vitro studies of cell alignment, differentiation, and invasion

Cell Culture. Normal bone-marrow derived human mesenchymal stem cells (MSCs, Lonza) were cultured in DMEM media supplemented with 10 % (v/v) fetal bovine serum, 1 % (v/v) penicillin-streptomycin, and basic fibroblast growth factor (10 ng ml⁻¹). All experiments with MSCs were conducted at passage 6 or earlier. Human umbilical vein endothelial cells (HUVECs, Lonza) were cultured in Endothelial Cell Growth Media (R&D Systems). Breast cancer cell lines MDA-MB-231 (ATCC) and MCF-7 (ATCC) were lentivirus transduced to express GFP and cultured in DMEM media (4.5 g L⁻¹ glucose) supplemented with 10 % (v/v) fetal bovine serum and 1 % (v/v) penicillin-streptomycin. All cells were incubated in a humidified atmosphere at 37 °C and 5 % CO2. All media components were purchased from Gibco.

MSC alignment on individual μRBs. 10 k MSCs were mixed with 3.3 mg of μRBs in a 48-well plate and allowed to attach overnight in MSC growth media. MSCs were labeled with Calcein AM (Corning, 4 mM) for 30 min and imaged using a Keyence BZ-X810 fluorescence microscope. Cell alignment was quantified using FIJI to measure the angle relative to the long axis of each μRB. Cell alignment in printed constructs was measured relative to the printing direction.

μRB and cell alignment analysis. MSC-incorporated μRB bioinks were printed into single-layer sheets, crosslinked, and incubated in growth media overnight. Then, HUVEC suspensions at the density of 10,000 cells ml⁻¹ were seeded on the scaffolds, and the scaffolds were transferred into the 1:1 mixture growth media of MSC and HUVEC growth media for 7 days. μRB alignment analysis was performed on the fluorescent images of FITC-labeled μRBs. For analyzing cell alignment of HUVEC and MSC, immunostaining of actin (ActinRed™ 555 ReadyProbes, ThermoFisher) and VE-cadherin (1:400, 2500 T, Cell Signaling) were performed on the scaffolds. Goat anti-rabbit IgG AlexaFluor 647 (1:500, A-21245, Invitrogen) was used as secondary antibody. The images were processed with ImageJ with the “OrientationJ” plugin [53], and the statistics were then analyzed by package ‘circular’ in R. The detailed method was as previously described [26].

Osteogenic Differentiation. Scaffolds were cultured in Ultra-Low Attachment plates (Corning) in osteogenic media for 28 days in a humidified atmosphere at 37 °C and 5 % CO2. Media was refreshed every two to three days. Osteogenic media was made from DMEM (1 g L⁻¹ glucose, Corning) supplemented with 10 % (v/v) MSC-qualified fetal bovine serum (Gibco), 1 % (v/v) penicillin-streptomycin (Gibco), β-Glycerophosphate disodium salt (10 mM), Dexamethasone (100 nM), and ascorbic-2-phosphate (50 μg ml⁻¹).

Live/Dead Staining. Scaffolds were washed in PBS and stained with Calcein AM (Corning, 4 mM) and Ethidium homodimer-1 (Chemodex, 2 mM) for 30 min at 37 °C and 5 % CO2. After incubation, images were taken using a Keyence BZ-X810 fluorescence microscope.

Cell Metabolic Activity. Bioprinted scaffolds were individually cultured in an Ultra-Low Attachment 96-well plate (Corning) in osteogenic media for 28 days. Each week, PrestoBlue reagent (ThermoFisher) was added to each well in fresh media and incubated for 1 h at 37 °C and 5 % CO2. Fluorescence intensity was measured at 560 nm excitation and 600 nm emission in a plate reader to determine cell metabolic activity. After the assay, scaffolds were washed two times with osteogenic media. Six replicates were measured for each condition.

DNA Quantitation. Scaffolds were lyophilized and rehydrated in papain solution (125 μg mL⁻¹ papain with 10 mM cysteine in Tris-borate-EDTA buffer). Once hydrated, samples were homogenized and incubated at 60 °C for 16 h. After cooling, digests were centrifuged at 10,000×g for 5 min and DNA content in the supernatant was measured using the Quant-iT PicoGreen dsDNA Assay (ThermoScientific) following the standard protocol.

Histology and Immunofluorescence Staining. Scaffolds were fixed in 4 % paraformaldehyde (PFA) for 30 min on an orbital shaker, then washed two times in PBS and one time in water. Samples were incubated in optimal cutting temperature (OCT) compound (Fisher Scientific) for 72 h on an orbital shaker and then frozen in tissue embedding molds at −80 °C. Frozen sections were cut on a Microtome Cryostat (Leica) at 10 μm thickness and stored at −20 °C before use.

Alizarin Red S (ARS) staining of cryosectioned slides was performed by removing OCT in distilled water for 5 min, staining in ARS solution for 2 min, and washing with distilled water. After drying, slides were mounted using Permount (Fisher Scientific). Aniline Blue staining was performed by removing OCT in PBS for 30 s and distilled water for 5 min, staining in Weigert's Iron Hematoxylin for 10 min, rinsing in warm running tap water for 10 min, staining in 5 %–5 % phosphotungstic-phosphomolybdic acid for 15 min, staining in 2.5 % aniline blue for 5 min, differentiating in 1 % acetic acid for 1 min, and rinsing in distilled water for 1 min. Sections were dehydrated in two changes of anhydrous alcohol and three changes of clearing reagent before mounting with Permount. All histology slides were imaged with a 20× objective on the Leica Aperio CS2 Slide Scanner and quantified for percent positive stain area using QuPath color deconvolution. Alkaline phosphatase (ALP) staining was performed on whole fixed scaffolds. Scaffolds were incubated in Fast Blue RR Salt dissolved in Naphthol AS-MX phosphate alkaline solution (0.25 % w/v, pH 8.6, Sigma-Aldrich) for 30 min at room temperature. After staining, scaffolds were washed in water two times and imaged with a Keyence BZ-X810 fluorescence microscope.

Immunofluorescence staining on cryosectioned slides was performed by removing OCT in PBS for 5 min, permeabilizing in 0.3 % Triton X-100 for 15 min, and blocking in 3 % bovine serum albumin (BSA) with 0.1 % Triton X-100 for 1 h at room temperature. Primary antibody for osteocalcin (1:200, 23418-1-AP, Proteintech), GFP (1:100, 50430-2-AP, Proteintech) was diluted in 1 % BSA with 0.1 % Triton X-100 and incubated overnight at 4 °C. After 3 washes with PBS, slides were incubated with secondary antibody, goat anti-rabbit IgG AlexaFluor 647 (1:500, A-21245, Invitrogen), goat anti-rabbit IgG AlexaFluor 488 (1:500, A32731, Invitrogen) for 1 h at room temperature. Slides were washed three times with PBS and imaged with a Keyence BZ-X810 fluorescence microscope or Leica Stellaris 5 confocal microscope. Mean fluorescence intensity quantification was completed using FIJI.

Scaffold cross-sectional area quantification. Scaffolds were imaged with brightfield microscopy with a Keyence BZ-X810 microscope. The cross-sectional area was quantified using FIJI software by drawing a selection around each scaffold and measuring the area.

Scaffold weight change quantification. Scaffolds were removed from culture media and excess liquid was removed with a Kimwipe. The wet weight of the scaffolds was measured at each time point and percent change was calculated using Equation (2):

$$\frac{W_{28}-W_0}{W_0}$$ (2)

where W₀ and W₂₈ are the wet weight of each scaffold on day 0 and day 28 of culture, respectively.

Bioprinting of tissue-engineered bone and breast cancer constructs. MSC bioink was formulated at a density of 15 million cells ml⁻¹ in the stiff μRBs and bioprinted through an 18G needle into a 4-by-4 grid with pores measuring 1 mm × 1 mm. The scaffolds were differentiated into tissue-engineered bone over 28 days in osteogenic media. The MSCs were labeled with CellTracker Red (Thermo Fisher). MDA-MB-231 and MCF7 bioink was formulated at a density of 10 million cells ml⁻¹ in the soft μRBs and bioprinted into the pores of the tissue-engineered bone scaffold to form the tissue construct. The GFP-expression breast cancer cell lines in the construct were imaged using a Leica Stellaris 5 confocal microscope. After 14 days of coculture, the constructs were fixed in 4 % PFA for 30 min, then washed two times in PBS and one time in distilled water. After fixation, the constructs were embedded into acrylamide gels by crosslinking 5 % acrylamide with 0.1 % LAP photoinitiator sandwiched between two glass slides under UV light. The embedded gels were incubated in optimal cutting temperature (OCT) compound (Fisher Scientific) for 72 h on an orbital shaker and then frozen in tissue embedding molds at −80 °C. The frozen samples were sectioned on a Microtome Cryostat (Leica) at 50 μm thickness and stored at −20 °C before immunofluorescence staining of GFP. Confocal microscopy of stained sections was taken with a Z-stack spanning 500 μm. Cell invasion quantification was performed using FIJI for 5 compartments per group.

3. Results and discussion

3.1. Anisotropic μRB bioinks with tunable stiffness

Scaffolds formed by traditional spherical microgel-based bioinks generally exhibit inferior mechanical properties compared to bulk hydrogels, due to limited surface contact between the particles, which restricts effective inter-particle crosslinking. There remains a need for granular microgels that not only exhibit excellent bioink properties, but also support forming macroporous scaffolds with mechanical properties suitable for regenerating load-bearing tissues [54,55]. In contrast to microfibers with round cross section, the μRBs feature a flattened geometry (Supplementary Fig. S1), and crosslinked scaffolds exhibit cartilage-mimicking shock-absorbing mechanical properties, as evidenced in previous study showing that μRBs was more flexible than microfibers under cyclic strain testing [33]. Compared with scaffolds formed by isotropic gelatin microgels (μGel), which represent the typical building blocks for granular bioinks, scaffolds formed by μRBs showed enhanced mechanical stability and compressive strength under cyclic mechanical loading. After five cycles of 30 % compression, the maximum load of μGel scaffolds decreased by 8 %, whereas the maximum load of μRB scaffolds remains largely intact (Supplementary Fig. S2). μGel scaffolds showed a rapid loss of modulus as the concentration decreased. While jammed or highly concentrated μGels (≥12 wt%) produced scaffolds with relatively high compressive modulus, μGels at 10 wt% or lower failed to form stable scaffolds due to insufficient inter-particle crosslinking. In contrast, μRBs readily formed stable scaffolds across a wider range of concentrations, likely because their elongated, flattened geometry promotes interlocking and provides greater contact area for inter-particle connectivity (Supplementary Fig. S3A and B) [33].

Varying microgel stiffness can significantly impact its bioink properties, such as printability, mechanical stability, and cell viability. Microribbons were fabricated by wet-spinning a gelatin precursor solution (20 % w/w), which were subsequently crosslinked with glutaraldehyde to stabilize the ribbon shape (Fig. 1A). To fabricate gelatin μRB with varying stiffnesses (soft, medium, stiffness), three concentrations of glutaraldehyde were used (5 %, 10 %, and 15 % of total amine groups on gelatin). Gelatin μRB were further modified using MA-NHS chemistry to introduce methacrylate groups on μRB, confirmed by FITR, to allow photocrosslinking of μRBs after bioprinting (Supplementary Fig. S4). To create viscous and self-supporting μRB bioinks, the dispersed μRBs were mixed with cells and compacted by centrifugation. This jamming process resulted in μRB bioinks with a gel-like and self-supporting properties, as shown in Fig. 1B. Then, we assessed how tuning the μRB bioink stiffness modulate the ribbon morphology, as well as porosity, and mechanical properties of the crosslinked scaffolds. To facilitate visualization of μRB morphology and scaffold porosity, FITC-labeled gelatin was used, and scaffolds were imaged using confocal microscopy, where μRBs exhibit flat and anisotropic features. μRBs with increasing stiffness also exhibit reduced width, and the average width of soft, medium, and stiff μRBs were 168 μm, 108 μm, and 83 μm, respectively (Fig. 1E). Tuning μRB stiffness affects the distribution of μRB width but has minimal impact on their length. Initially, continuous wet-spinning produced strands several meters long. For printing, these strands were cut into shorter segments using a blender, yielding lengths between 200 and 1000 μm. Length distributions were comparable across the soft, medium, and stiff groups. (Supplementary Fig. S5B). The intra-crosslinking density of μRBs affects the swelling ratio of gelatin, resulting in a gradient of both stiffness and width.

Fig. 1.

Fabrication, characterization, and alignment of μRB-based bioinks with tunable stiffness. (A) Schematic illustrating the fabrication of μRBs with tunable stiffness and their formulation into bioinks for 3D printing. The printing process promotes μRB alignment, enabling the formation of anisotropic scaffolds for tissue engineering. (B) Visual comparison of diluted μRB suspension vs. jammed μRB bioink. (C) Morphology of individual μRBs with different stiffness (scale bar = 100 μm). (D) Measurement of individual μRB stiffness using atomic force microscopy (AFM) (n = 350 per group). (E) Quantification of μRB width across groups with varying stiffness from confocal images (n = 30 per group). (F) Confocal images showing aligned μRBs within printed filaments (Scale bar = 100 μm). (G) Fluorescence images tracking the movement of FITC-labeled μRBs during extrusion (Scale bar = 400 μm). (H) Quantification of μRB alignment (n = 10 per group). Data are presented as mean ± S.D. and p-values were determined by one-way analysis of variance (ANOVA) with Tukey's multiple comparisons test (D, E) or Watson-Wheeler test (H); ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.005, ∗∗∗∗p < 0.001.

Given the inherent anisotropic nature of ribbon-shaped microgels, we next explored their potential for bioprinting applications aiming to mimic extracellular matrix (ECM) anisotropy, a feature commonly observed in tissues such as vasculature, tendon, and muscle [37,38,[56], [57], [58]]. Anisotropic ink fillers, such as methylcellulose or alginate rods have previously been shown to align parallel to print direction due to the shear stress and extension flows associated with a tapering nozzle [59,60]. In agreement with these prior observations. We observed that μRB bioinks exhibited spontaneous alignment during the extrusion process (Movie 1), with the degree of alignment increasing as the μRB stiffness increased observed by fluorescence microscopy and SEM (Fig. 1F and Supplementary Fig. S6). This enhanced alignment in stiffer μRB bioinks is likely driven by shear and elongational flows in the nozzle, which impose drag forces along the μRB long axis (Fig. 1G). To further quantify μRB alignment following extrusion, we used the ‘OrientationJ’ plugin in ImageJ for image analysis and performed statistics using the ‘circular’ package in R [26]. The results confirmed enhanced alignment in stiffer μRB bioinks, with statistically significant differences observed among all groups (p-value = 0.0027) (Fig. 1H). While anisotropic particles such as nanofibers or microstrands have been shown to align under shear [26,27], our results demonstrate for the first time that microgel stiffness strongly modulates this process: stiffer μRBs resist deformation and maintain orientation under flow, whereas softer μRBs partially bend, resulting in weaker alignment.

Supplementary video related to this article can be found at doi:10.1016/j.bioactmat.2025.12.040

The following is/are the supplementary data related to this article:

Video s2

3.2. μRB bioinks exhibit injectability, self-healing, and shear-thinning properties

To prepare μRB bioinks, μRBs were first suspended as a 7.5 wt% slurry to ensure fully hydration of μRBs, then centrifuged at 4300×g for 10 min to remove extra liquid. The jammed bioinks with three soft, medium, and stiff μRBs bioinks were then imaged to obtain their porosity, which all exhibited interconnecting microporous structures (Fig. 2A). Due to the higher swelling ratio, larger width, and better deformability of soft μRBs, the resulting supernatant had less volume to remove and the bioink has lower porosity after jamming than medium and stiff groups (Fig. 2B and C). Although the crosslinked scaffolds formed from the soft μRB slurry at 7.5 wt% exhibited the highest compressive modulus due to their lower porosity, their modulus became comparable to those of the medium and stiff groups after jamming (Supplementary Fig. S7) [26]. This was likely also due to the reduced packing efficiency of the medium and stiff μRBs, which are less deformable during jamming process, and the higher porosity of their scaffolds. These findings suggest that the compressive modulus of bioprinted granular microgels is impacted by both intra- and inter-microgel crosslinking, rather than solely by the stiffness of individual microgels.

Fig. 2.

Assessment of injectability, rheological properties, and printability of μRB bioinks with varying microgel stiffness.

(A) Confocal images showing the Z-cross section and 3D projection of μRB bioinks (scale bar = 200 μm). (B) Quantification of excess PBS volume removed during the jamming process to prepare μRB bioink (n = 3 per group, scale bar = 200 μm). (C) Porosity of the resulting crosslinked μRB bioinks (n = 200 per group). (D) Shear rate sweep showing shear thinning properties of μRB bioinks. (E) Strain sweep from 1 % to 500 % strain. (F) μRBs bioinks exhibit self-healing behavior under alternating 1 % and 500 % strain cycles. (G) Schematic of compression testing to measure injection force for extruding μRBs bioink. (H) Force-displacement curves during injecting μRB bioinks through 18G and 20G needles. (I) Maximum injection force recorded from the force curves (n = 3 per group). Data are presented as mean ± S.D. and p-values were determined by one-way (B, C) or two-way (I) analysis of variance (ANOVA) with Tukey's multiple comparisons test; ∗∗p < 0.01, ∗∗∗p < 0.005, ∗∗∗∗p < 0.001.

Next, to evaluate how μRB bioinks respond to varying strain and shear stress conditions during extrusion, rheological analyses were performed including shear rate sweep, strain sweep, and self-healing tests. In the frequency sweep, all groups demonstrated gel-like behavior, and increasing stiffness of μRB increased the storage modulus of the bioinks (Supplementary Fig. S8A). The medium and stiff groups were 1.5 and 3 times higher than the soft group, respectively, which corresponded to the Young's modulus of individual μRBs. The shear-rate sweep tests showed that all bioinks exhibited excellent shear-shinning properties, as shown by the power-law index (n). This is calculated from fitting the curve of viscosity (η) versus shear rate ($\dot{\mathrm{\gamma }}$) with a power law model (η = K $\dot{\mathrm{\gamma }}$ⁿ⁻¹), where K is a consistency factor. A Newtonian fluid has a power law index (n) of 1, while shear-thinning fluids have 0 < n < 1. All three bioink formulations showed strong shear-thinning behavior, with n values of 0.16 ± 0.0 (soft), 0.05 ± 0.03 (medium), and 0.03 ± 0.04 (stiff), all below 0.3 (Fig. 2D). The soft bioink group showed weaker shear-thinning behavior, likely due to the larger width of the soft μRB and decreased porosity that limit μRB rearrangement. Yield strains of all groups showed no significant difference, except that the stiff μRB bioink showed a slightly lower yield strain, indicating that they required less deformation to start flowing (Fig. 2E). However, the stiff and medium groups exhibited greater yield stresses due to their higher storage moduli, which is consistent with the results of injection force tests (Supplementary Fig. S8B). Self-healing tests using alternating 1 % and 500 % strain showed the storage modulus recovered after two cycles in all bioink groups, confirming excellent mechanical stability post-printing (Fig. 2F). Together, rheological analyses of all μRB bioinks showed no significant difference except in yield stress.

To assess the injectability of μRB bioinks with different stiffnesses, the injection forces required to extrude the bioinks loaded in syringes through 18G or 20G needles were measured (Fig. 2G). Injection force is a method for quantifying the stress applied to the μRBs and cells directly during the injection, which is a critical factor impacting cell viability after bioprinting. The injection force was recorded continuously for 120 s at 2 ml h⁻¹ injection speed, and the maximum injection force was identified based on the highest reading during the injection (Fig. 2H). The maximum injection forces were proportional to the stiffness of μRBs, as more stress was required to align and disentangle the stiffer μRBs (Fig. 2I). As expected, the needle diameter also affected the injection force, where the 20G needle required significantly greater force than 18G needle for all μRB stiffnesses. To ensure high cell viability and reduce injection force, the 18G needle was chosen for the following cell studies. In summary, μRB stiffness played a key role in injectability, with stiffer μRBs requiring higher stress to realign and flow during extrusion.

To visualize the printability of microgel-based bioinks, several widely-used tests, including printing grid patterns to assess filament uniformity, filament drop tests, and filament collapse tests, can be applied as indicators of bioink shape fidelity and self-supporting capability [46,52,61]. Next, we performed these tests to assess the printability of μRB bioinks. All μRB bioinks groups supported printing grid pattern, and varying μRB stiffness impacted filament drop and filament collapse (Fig. 3A). Furthermore, μRB bioinks exhibited self-supporting properties, as demonstrated by stable structures with 6, 9, and 12 printed layers (Fig. 3B). The printability index (Pr) quantifies bioink printability by measuring the circularity of pores in a printed grid, with a Pr value close to 1 indicating ideal square-shaped pores and good structural fidelity. All μRB bioinks displayed Pr values close to 1 and similar widths of the printed filament, indicating excellent printability (Fig. 3C and D). While all groups exhibited self-supporting ability in the collapse tests, the stiff bioink supported more continuous extrusion with a longer filament (>100 mm) in the filament drop test (Fig. 3E). This indicates that the stiff μRBs could endure greater pulling forces that typically disrupt adhesion or entanglement among granular microgels. In addition, the stiff bioink retained clearer pore geometry after printing 9 layers, compared to soft bioink. The printed height accuracy, calculated as printed height divided by targeted height, remained close to one for all three bioinks (Fig. 3F). Together, these results confirm that μRB bioinks across the tested stiffness range all display excellent printability and self-supporting capability, indicating their suitability for a broad range of bioprinting applications.

Fig. 3.

Printability assessment of μRBs bioinks with varying microgel stiffness. (A) Visualization of printability through a printed grid pattern, filament drop test, and filament collapse test (Scale bar = 1 mm). (B) μRB bioinks exhibit self-supporting properties, demonstrated by stable structures with 6, 9, and 12 printed layers (Scale bar = 5 mm). (C) Pore geometry (PR) measured from the printed grid patterns (n = 10 per group). (D) Width of printed filaments on grid patterns (n = 20 per group)). (E) Maximum unsupported hanging length of filament from drop tests (n = 3 per group). (F) Height accuracy of the printed multilayer structures compared to digital models (n = 3 per group). Data are presented as mean ± S.D. and p-values were determined by one-way analysis of variance (ANOVA) with Tukey's multiple comparisons test; ∗∗∗p < 0.005.

3.3. Anisotropic μRB bioinks promote cell spreading and alignment

Given the inherent anisotropic nature of ribbon-shaped microgels, we next explored their potential for bioprinting applications aiming to mimic extracellular matrix (ECM) anisotropy, a feature commonly observed in tissues such as vasculature, tendon, and muscle (Fig. 4A) [37,38,[56], [57], [58]]. Individual μRBs were previously shown to support the spreading, alignment, and proliferation of human adipose-derived stromal cells [33]. To assess if μRB bioink support cell alignment, human mesenchymal stromal cells (MSCs) were seeded onto the surface of soft, medium, and stiff μRBs, allowed to attach overnight, and imaged with Calcein AM live cell stain (Fig. 4B). Quantification of MSC length and MSC alignment (relative μRB long axis) demonstrated greater cell spreading and cell alignment as substrate stiffness increased (Fig. 4C and D). Overall, MSCs exhibited excellent attachment and alignment along the μRBs.

Fig. 4.

Effect of extrusion process on μRB bioink and cell alignment. (A) Schematic of MSCs encapsulated in μRB bioink with HUVECs seeded on top of the printed scaffolds. (B) Live cell staining of MSC alignment on individual μRBs (Scale bar = 100 μm). (C) Distribution of MSC cell length (n = 250 per group); p-values were determined by one-way analysis of variance (ANOVA) with Tukey's multiple comparisons test; ∗∗∗∗p < 0.0001. (D) Cell orientation relative to μRB orientation, where 0° is parallel to the axis of the μRB (n = 250 per group). (E) Confocal images of F-Actin staining for cell morphology and VE-cadherin staining for endothelial cell junctions. Color survey visualization of directional analysis conducted using OrientationJ (Scale bar = 200 μm). (F, G) Quantification of F-actin and VE-Cadherin alignment. For alignment quantification (n = 10 per group) data reported as mean ± S.D., statistical analysis by Watson–Wheeler test, ∗p ≤ 0.05.

Given the ability of μRB bioinks to align during extrusion in a stiffness-dependent manner, we next explored the potential of μRB bioinks to promote cell alignment in multicellular bioprinted constructs. μRB bioinks were mixed with MSCs before extrusion and then photo-crosslinked into macroporous scaffolds after bioprinting. Human umbilical cord-derived endothelial cells (HUVECs) were subsequently seeded on top of the MSC-containing, bioprinted μRB scaffolds. After 7 days of culture in growth medium, immunostaining for F-actin and VE-cadherin was performed to assess cell alignment and formation of endothelial cell junctions. Visualization of F-actin and VE-cadherin alignment was shown by the color survey representation from the ImageJ OrientationJ plugin (Fig. 4E). Analysis of both F-actin and VE-cadherin showed cell alignment across all three μRB bioinks, with greater alignment observed as μRB stiffness increased (Fig. 4F and G). The difference in F-actin alignment among the three stiffness groups (p-value = 0.0065) was more statistically significant than that observed for VE-cadherin alignment (p-value = 0.036). Since VE-cadherin is exclusively found in endothelial cells and the HUVECs were seeded onto the surface of the crosslinked scaffolds, this difference may be explained by the stronger shear stress experienced at the surface of the extruded filaments, which induced alignment of the exterior μRBs regardless of their stiffness.

Together, these results validate that all μRB bioinks supported robust cell alignment, which is further enhanced as the stiffness of μRB bioinks increase. These results demonstrate the promise of μRB bioinks for mimicking ECM anisotropy and promoting cell alignment, which addresses a key unmet need for printing hierarchically organized tissues. While the present study focuses on MSCs and endothelial cells to demonstrate proof-of-principle, future studies may explore the potential of μRB bioinks for inducing alignment of cell types in other tissues with aligned structures such as muscle, tendon, and nerves.

3.4. μRB bioinks modulate MSC osteogenesis and bone formation in a stiffness-dependent manner

Substrate stiffness plays a key role in directing stem cell differentiation [43,44], and granular microgel-based bioinks have been widely used to support stem cell in tissue regeneration applications [[62], [63], [64]]. μRB bioinks with tunable stiffness are well suited for engineering lineage-specific tissues, including adipose, cartilage, and bone. As a proof of concept, to demonstrate the potential of μRB bioinks in directing lineage-specific differentiation, we chose the osteogenic lineage of MSCs. Bone tissue is unique due to its anisotropic structure and high mineral content [65,66], with aligned collagen fibers providing load-bearing functions [67]. Gelatin μRBs are especially suitable for bone tissue engineering as gelatin is derived from Type I Collagen, the primary organic component of bone ECM [68].

To evaluate how varying stiffness of μRB bioinks modulates MSC osteogenesis and bone formation, human MSCs were mixed with gelatin μRBs of varying stiffness and cultured in osteogenic differentiation medium for 28 days. The soft μRB bioinks supported higher cell viability post-extrusion compared to the medium and stiff bioinks (Fig. 5A). The trend in cell viability after extrusion was consistent with the trend in injection force, as stiffer μRB bioink required higher injection forces, which could compromise cell viability. Consistent with our previous findings, quantification of cell angle revealed cell alignment in the direction of printing in all groups (Supplementary Fig. S9). By day 28, cells in the soft μRBs exhibited round morphology with higher density, whereas cells in stiffer μRBs exhibited an elongated morphology with lower cell density. These differences were driven in part by greater scaffold contraction by cell contractile force. For acellular scaffolds, after 14 days of incubation in PBS at 37 °C, brightfield imaging confirmed all three stiffness groups retained their shape. Furthermore, all groups retained a stable wet weight over 14 days (Supplementary Fig. S10A and B). However, for cell-encapsulated μRB scaffolds in osteogenic media, both medium and stiff groups retained its shape and dimension over time, whereas the soft group showed greater contraction over time (Supplementary Fig. S11). This is consistent with the result from previous bioprinting studies, in which cell-generated forces induce predictable shape changes in compliant scaffolds [69,70]. Overall, the scaffold weight remained stable over 28 days in culture with MSCs with no noticeable degradation.

Fig. 5.

μRB bioinks modulate MSC morphology, osteogenesis and bone formation in a stiffness-dependent manner. (A) Live/dead staining of MSCs after extrusion (Day 0) or after 28 days of culture in osteogenic medium. Green: live cells, Red: dead cells. (B) Metabolic activity of MSCs measured by PrestoBlue assay at day 0 and day 14 after bioprinting (n = 6 per group). (C) DNA content per scaffold measured by PicoGreen assay at day 28 (n = 3 per group). (D) Alizarin red S (ARS) staining for mineralized bone matrix, (E) Aniline Blue staining for total collagen, and (F) immunostaining of Osteocalcin (OCN), a mature bone marker, at day 14 and day 21 of osteogenesis (n = 4 per group). (G–I) Quantification of ARS and Aniline Blue percent positive area and OCN mean fluorescence intensity (MFI). Scale bar = 100 μm. Values are presented as mean ± S.D. and p-values were determined by one-way analysis of variance (ANOVA) with Tukey's multiple comparisons test; ∗p ≤ 0.05, ∗∗p ≤ 0.01, ∗∗∗p ≤ 0.005, ∗∗∗∗p ≤ 0.001.

Similar to the trend observed with Live/dead staining, a PrestoBlue assay also showed higher cell metabolic activity in soft μRBs immediately after extrusion, though all groups recovered by day 14 (Fig. 5B). DNA content per scaffold was measured by PicoGreen assay at day 28, which showed higher cell number in soft μRBs compared to stiff μRBs (Fig. 5C). This trend is consistent with previous reports, where increasing substrate stiffness reduced cell proliferation while accelerating MSC osteogenesis [71].

To further evaluate MSC osteogenesis and new bone formation, samples were harvested on day 28 for histology. All scaffolds became increasingly opaque over time, indicating new matrix deposition (Supplementary Fig. S12). Alizarin red S staining revealed significantly greater and faster mineral deposition in the medium and stiff μRB bioink groups by day 14, with average positive areas of 58 % and 49 %, respectively, compared to only 1 % in the soft group (n = 3). The same trend continued to day 21, where the soft μRB group showed minimal mineralization (Fig. 5D–G). Aniline Blue staining of collagen showed a similar trend, where the stiff μRB bioink led to faster collagen deposition than the medium stiffness group on day 14, and both groups resulted in comparable high collagen deposition by day 21 (Fig. 5E–H). Immunostaining of osteocalcin, a late-stage osteogenesis marker, further supported stiffness-dependent osteogenesis, with the stiff μRB bioink showing the highest osteocalcin signal at day 21 (Fig. 5F–I). Evidence of reduced porosity in the soft μRBs is also visible from histology, and macroporosity has been demonstrated to be beneficial for osteo-induction [72]. Higher porosity can enhance cell viability and differentiation by facilitating greater nutrient exchange and cell infiltration, but may reduce the mechanical strength of the scaffold [73,74]. Stiffer μRBs promote cell spreading (Fig. 5A), thereby enhancing activation of mechanotransduction signaling to promote osteogenesis [43,75].

Together, these findings demonstrate that μRB bioinks are mechanically tunable and support MSC osteogenesis after bioprinting in a stiffness-dependent manner. Increasing μRB stiffness enhanced MSC spreading and accelerated mineralized bone matrix deposition while retaining printed scaffold shape over time. Featuring tunable stiffness and composition, μRB bioinks can match the biochemical and mechanical requirements of different tissues. These results establish ribbon-shaped microgels as a new class of anisotropic bioink for directing lineage-specific stem cell differentiation via mechanotransduction, with broad potentials in stem cell-based bioprinting applications.

3.5. μRB bioinks enable hierarchical tissue patterning to model cancer invasion

μRB-based bioinks are self-supporting and mechanically stable, enabling the fabrication of complex, hierarchically patterned tissue constructs. Previous research has shown that cell-cell and cell-ECM interactions are critical for building 3D bone co-culture systems [39], including bioprinted multiple-compartment models [76,77]. Unlike spherical microgels, high-aspect-ratio μRBs provide inherently high pore inter-connectivity, which enhances cell migration in 3D [28]. To demonstrate the potential of μRB bioinks for hierarchical, multicellular patterning, we chose to bioprint a 3D breast cancer-bone metastasis model mimicking breast cancer invasion into long bone through the bone marrow [78,79]. In previous studies using non-printed μRB hydrogels, the biological relevance of the model was demonstrated, including bone-specific metastasis, differential invasion and cancer-induced bone remodeling [39]. In the present work, we used bioprinting to recreate the model containing multiple cell types in a spatially organized manner, and bioprinting would enable greater scalability of fabricating such complex models.

We first bioprinted MSCs with stiff μRBs into a 4-by-4 grid with three layers and cultured them in osteogenic medium for 28 days to generate bone tissue, building upon prior results demonstrating robust MSC-based bone formation using the stiff μRB bioink (Fig. 5). The construct contained 16 open compartments, which modeled the bone marrow compartment in 16 replicates. Breast cancer cells were mixed with the soft μRB bioink and extruded into the open compartments within the grid to mimic the bone marrow/bone interface (Fig. 6A). To visualize the bone construct, the scaffold was imaged with fluorescence microscopy (Fig. 6B). To monitor cancer cell invasion into MSC-derived bone, MSCs were pre-labeled with CellTracker Red, and GFP-expressing breast cancer cell lines (MDA-MB-231 and MCF-7) were used (Fig. 6C). After extruding the breast cancer cells into the open compartments, the scaffold was UV-crosslinked and co-cultured with bone in growth medium for 14 days. Confocal imaging at day 14 revealed substantial breast cancer invasion into the surrounding bone compartment (Fig. 6D). Quantification of breast cancer cell invasion, based on cell count and fluorescence area (p = 0.0026 and p = 0.0012), demonstrated significantly higher invasion by MDA-MB-231 cells compared to MCF-7 cells. This behavior is consistent with known cell line characteristics, as MDA-MB-231 is more aggressive in bone metastasis [39]. Each compartment mimicked the bone marrow/bone interface, and the 4-by-4 grid allows high-throughput screening with 16 replicates in a single bioprinted construct. These results highlight the potential of μRB bioinks to support hierarchical patterning of multiple cell and tissue types, and future studies could leverage μRB bioinks for creating scalable multi-tissue patterns that mimic physiological structures for modeling diseases, drug screening, or regenerative applications.

Fig. 6.

μRB bioinks support multicellular patterning to model breast cancer-bone invasion at tissue interface. (A) Schematic of experimental design: bioprinting of MSCs and osteogenic differentiation to derive bone grid, followed by injection of breast cancer bioink, and monitoring invasion over time using confocal microscopy. (B) Confocal images of MSCs (red) after printing (Scale bar = 1 mm). (C) Confocal images of scaffold sections containing CellTracker-labeled MSCs (red) after 28 days of osteogenic differentiation and GFP + MDA-MB-231 cells extruded into the open pores of the grids (green) (Scale bar = 200 μm). (D) Confocal images of patterned MSC-derived bone (red) with MDA-MB-231 and MCF-7 breast cancer cells (green) after 14 days of co-culture (Scale bar = 1 mm). (E) Quantification of breast cancer cell invasion: percentage that remain in open pores vs. invading into the MSC-bone compartment (n = 5 per group). Values are reported as mean ± S.D. and p-values were determined by two-way analysis of variance (ANOVA) with Tukey's multiple comparisons test; ∗∗p ≤ 0.01, ∗∗∗p ≤ 0.005, ∗∗∗∗p ≤ 0.001.

4. Conclusions and future directions

In summary, here we report μRBs as a new class of anisotropic microgel-based bioinks generated by jamming anisotropic ribbon-shaped microgels made by wet-spinning. By controlling the intra-crosslinking density of gelatin, we can tune the stiffness of individual μRBs, and the resulting porosity and compressive moduli of the crosslinked μRB scaffolds. We demonstrated excellent printability of μRB bioinks with three different stiffnesses, which support shape fidelity and structural integrity. All three μRB bioinks supported high cell viability and metabolic activity after bioprinting. The shear stress generated during the extrusion process aligned anisotropic μRBs, which further guided human MSC and endothelial cell alignment in 3D. The stiff μRB group experienced the highest shear stress during extrusion, resulting in the greatest alignment. These results support that μRB bioinks hold great potential in bioprinting tissues with aligned structures, such as tendon, ligand and muscle. Additionally, tuning μRB bioink stiffness can enhance osteogenesis, demonstrating the ease of tuning μRB bioink properties to promote desirable cell differentiation. Finally, we demonstrated the versatility of μRB bioinks for hierarchical patterning of multiple cell types by bioprinting of model of breast cancer-bone metastasis. μRB bioinks present great advantages in mimicking tissue anisotropy, while maintaining macroporosity, modularity, and excellent printability. The results from the present study have established μRB-based bioinks as anisotropic microgels for bioprinting that enables the fabrication of various anisotropic tissues for regenerative medicine and disease modeling applications.

While this study establishes the feasibility of μRBs for bioprinting, a remaining limitation is printing resolution. The elongated geometry of the μRBs induces higher shear forces during extrusion, which restricts the minimum size of the printing needle. To expand the printing resolution, future work can incorporate rheological modifiers to reduce μRB entanglement while maintaining the anisotropic architecture. Furthermore, the current study focuses on gelatin, and future studies may further expand the μRB-based bioinks by optimizing the printability of μRBs fabricated from various natural or synthetic polymers [34,35]. In the present study, we demonstrate the versatility of μRB bioinks in supporting diverse bioprinting applications using well established biological models. Future studies can build on this foundation and apply μRB bioinks to answer novel biological questions and further assess their translational potential using relevant disease models in vivo [[80], [81], [82]].

CRediT authorship contribution statement

Hung Pang Lee: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Michelle Tai: Writing – review & editing, Writing – original draft, Visualization, Validation, Software, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Sarah J. Jones: Investigation, Formal analysis, Data curation. Xinming Tong: Conceptualization. Sungwon Kim: Data curation. Michelle M.T. Jansman: Data curation, Formal analysis, Methodology. Tony Tam: Software, Resources. Jianyi Du: Data curation. Mark A. Skylar-Scott: Writing – review & editing, Writing – original draft, Resources. Fan Yang: Writing – review & editing, Writing – original draft, Supervision, Funding acquisition, Conceptualization, Formal analysis, Investigation, Methodology, Project administration, Resources.

Ethics approval and consent to participate

This study does not include any clinical investigations, animal experimentation, or research involving human subjects, including the use of donated organs or tissues.

Declaration of competing interest

The authors declare the following personal relationship which may be considered as potential competing interests: Mark A. Skylar-Scott is currently employed by Chan Zuckerberg Biohub.

Appendix A. Supplementary data

The following are the Supplementary data to this article: