DLP bioprinting of cartilage organoid-laden bioinks yields high-fidelity auricular constructs with enhanced chondrogenesis

Abstract

Purpose

To develop and validate a cartilage organoid (CO)-laden digital light processing (DLP) bioprinting strategy for auricular reconstruction and to compare its performance with conventional chondrocyte-laden prints.

Materials and methods

Rat bone-marrow stromal cells (BMSCs) were aggregated into spheroid and chondrogenically induced to form CO. Organoid construction and characterization included EdU proliferation assay, CD73/CD90 immunofluorescence (IF), and real-time quantitative polymerase chain reaction (qPCR) of chondrogenic genes. O-nitrobenzyl functionalized gelatin (GelNB)/ methacrylated hyaluronic acid (HAMA) bioinks were screened by gross morphology, tensile/ compressive mechanics, enzymatic degradability (0.1% collagenase II), and swelling, identifying 10% (w/v) GelNB + 1% (w/v) HAMA as the working bioink. Full-scale ears were DLP-printed from acellular, chondrocyte-laden, or CO-laden inks. In vitro, whole-mount Live/Dead imaging was performed at 7/14/21 days and qPCR quantified COL2A1, aggrecan (ACAN), COL10A1. In vivo, constructs were implanted subcutaneously in nude mice and analyzed at 4/8 weeks by Hematoxylin–eosin (H&E), Safranin O/Fast Green, ACAN and COL II IHC staining, and compressive Young’s modulus measurement on implants.

Results

The 10% GelNB + 1% HAMA ink balanced print fidelity, mechanical robustness, and controlled degradability. Loading CO did not impair printability. During 21-day culture, viability remained high; at day 21, CO-laden ears showed higher COL2A1/ACAN and lower COL10A1 than chondrocyte-laden controls. After implantation, CO-laden explants exhibited more abundant lacuna-like cartilage morphology, stronger ACAN and COL II staining, increased EVG-positive elastic fiber-associated matrix, and a higher compressive Young’s modulus at 8 weeks.

Conclusions

CO-laden DLP bioprinting enhances chondrogenesis and mitigates hypertrophy versus chondrocyte-laden printing, offering an exploratory high-fidelity strategy for auricular cartilage-like tissue engineering rather than definitive mature elastic cartilage regeneration.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s13287-026-05156-5.

Introduction

Microtia is a common craniofacial malformation that imposes functional and psychosocial burdens on children and families. Current clinical reconstructions rely on autologous costal cartilage frameworks or alloplastic/3D-printed polymer implants, yet both strategies have well-recognized limitations: autologous approaches demand multi-stage procedures, entail donor-site morbidity and postoperative pain, and often yield auricular elasticity inferior to native tissue, whereas implants risk exposure/extrusion and infection and do not generate autologous neotissue [1, 2]. These trade-offs motivate strategies that couple anatomical precision with true biological regeneration of elastic cartilage [3]. Recent auricular biofabrication studies have further highlighted the rapid progress of scaffold-assisted ear reconstruction. For example, full-scale auricular constructs have been generated using 3D-printed external scaffolds combined with decellularized cartilage xenograft [4], while decellularized human cartilage-derived bioink reinforced with polycaprolactone has also been explored for ear reconstruction [5]. These studies underscore the importance of maintaining patient-relevant ear shape and mechanical support, but they also mainly emphasize scaffold or template guidance and matrix-derived bioactivity rather than directly using living multicellular cartilage-forming units as the active building blocks of the printed construct.

In parallel, the convergence of 3D bioprinting and tissue engineering has accelerated and is now being investigated across bone, blood vessels, cardiac, and skin applications [6–10]. Achieving robust post-print viability and maturation, however, requires a delicate balance: the bioink must present a densely cellular architecture that supports cell–cell communication while maintaining a crosslinked hydrogel network that prints with high fidelity and withstands handling [11, 12]. The most common practice—directly mixing single cells into a hydrogel—simplifies preparation but disperses cells as isolated units, which can blunt intercellular signaling, limit cooperative matrix secretion, and hinder chondrogenic maturation [13]. Because effective cartilage regeneration depends on sustained extracellular matrix (ECM) production and deposition, the micro-organization of living building blocks in the ink is as important as the polymer chemistry itself [14–16].

Cartilage organoid (CO) addresses this architectural gap [17]. By aggregating progenitors/chondrocytes into 3D microtissues, CO recapitulate aspects of mesenchymal condensation and early chondrogenesis in vitro and outperform dispersed single-cell suspensions in maintaining lineage stability and driving cartilage matrix deposition during regeneration [18]. In parallel, recent studies on organoid- and spheroid-based cartilage biofabrication have reinforced the value of multicellular microtissues as regenerative building blocks. Cartilaginous organoids have been shown to support adaptive cell-matrix interactions and enhanced cartilage regeneration [18], while spheroid-based bioink systems have demonstrated that maintaining multicellular architecture can improve print support, fusion, and post-print culture performance [19]. In addition, recent reviews of bone/cartilage organoid bioprinting have emphasized both the promise of these living building blocks and the remaining challenges in standardization, scale-up, and translation [17, 20]. Against this backdrop, the present study differs in that it uses CO as the active biological building blocks within a DLP-printable GelNB/HAMA system and directly evaluates whether this strategy can improve chondrogenic maturation in high-fidelity full-scale auricular constructs. Guided by these principles, we reasoned that loading CO into a digital light processing (DLP)-printable bioink could keep cells in supportive multicellular neighborhoods during printing and early culture, while DLP would reproduce ear-scale macro-geometry and fine surface features with minimal mechanical insult to cells compared with shear-based extrusion [21–23]. In combination, this design should favor stable chondrogenesis [Upregulate type II collagen alpha 1 chain (COL2A1)/SRY-related high mobility group-box gene 9 (SOX9)], limit hypertrophic drift (Downregulate type X collagen alpha 1 chain (COL10A1), and ultimately improve in-vivo maturation relative to chondrocyte-laden prints.

To implement this strategy, we selected a hydrogel system based on o-nitrobenzyl–functionalized gelatin (GelNB) blended with methacrylated hyaluronic acid (HAMA). GelNB/HAMA offers a set of properties aligned with auricular biofabrication: cytocompatibility for sensitive cell populations; high-resolution DLP printability with controllable cure depth; tunable mechanics spanning soft cartilage-like modulus; and programmable degradability that permits matrix deposition and remodeling over weeks [24]. After a formulation screen, 10% (w/v) GelNB + 1% (w/v) HAMA provided a balanced window for ear-scale printing, gross handling, and subsequent biological maturation.

Here, leveraging our prior DLP experience, we formulate and test a CO-laden digital light processing approach for auricular bioprinting alongside chondrocyte-laden and acellular controls [6, 25, 26]. We show that CO loading does not compromise printability or shape fidelity; that in vitro CO-laden ears sustain viability over 21 days while upregulating COL2A1/ aggrecan (ACAN) and downregulating COL10A1; and that in vivo (subcutaneous nude mouse model) they develop more abundant lacuna-like cartilage morphology, stronger ACAN/COL II staining, increased EVG-positive elastic fiber-associated matrix, and higher compressive modulus at 8 weeks, supporting enhanced cartilage-like maturation. Figure 1 summarizes the process workflow, including CO construction and characterization, the bioink formulation comprising GelNB/HAMA and CO, and the high-fidelity DLP printing of the auricle followed by its implantation subcutaneously in nude mice.

Fig. 1.

Technical Roadmap of This Study. Rat bone marrow mesenchymal stem cells (rBMSCs) were aggregated into spheroid and induced to form chondrogenic organoids (CO). These CO were mixed with 10% GelNB + 1% HAMA to create a CO-laden bioink, which was then printed into full-scale auricular constructs using digital light processing (DLP) technology. The printed constructs were cultured in vitro and subcutaneously implanted into nude mice for histological, IHC, and compressive mechanical testing and analysis

Materials and methods

BMSCs and chondrocyte isolation and in vitro culture

The isolation and culture of rat BMSCs were performed according to the previously established protocol. In brief, neonatal Sprague–Dawley (SD) rats (3–5 days old) were euthanized by cervical dislocation and soaked in 75% ethanol for 5 min. Femurs and tibias were dissected, attached soft tissues were removed, and both epiphyses were cut off. The marrow cavities were flushed three times with complete low-glucose DMEM (L-DMEM, Hyclone, USA) containing 10% fetal bovine serum (FBS, Hyclone, USA) and 1% penicillin–streptomycin (P/S, Beyotime, China) until the shafts appeared white. The flushed bone-marrow suspension was seeded into 10-cm culture dishes and maintained at 37 °C, 5% CO₂; medium was changed every 2 days. Upon reaching ~ 80–90% confluence, r-BMSCs were passaged with trypsin–EDTA and used at P2–P4.

Cartilage was harvested from both femoral epiphyses of the same-age SD pups under aseptic conditions. Tissues were minced to ~ 1 mm fragments, incubated with trypsin–EDTA (Beyotime, China) for 30 min at 37 °C, then centrifuged and the supernatant discarded; high-glucose DMEM (H-DMEM, Hyclone, USA) with 10% FBS and 1% P/S was added to quench trypsin. Pellets were digested with 0.2% (w/v) collagenase II for 3 h at 37 °C, passed through a 70-µm strainer, and cells were collected by centrifugation. Chondrocytes were seeded in culture dishes with HG-DMEM + 10% FBS + 1% P/S (Beyotime, China) at 37 °C, 5% CO₂; medium was changed every 2–3 days.

Construction of CO from BMSC spheroid

BMSCs (P2–P4) were seeded into commercial cell hive plates (EXCellentBiosciences, China) at ~ 2000 cells per well to form BMSCs spheroid. The spheroid formation medium was L-DMEM, supplemented with 10 ng/mL basic fibroblast growth factor (bFGF, MCE, USA), 10% FBS, and 1% P/S. Medium was changed every other day. After 3 days of aggregation (D3), spheroid were switched to chondrogenic induction medium to generate cartilage organoid (CO).

The induction medium contained 20 µM Y-27632 (Yeasen, China), 1× ITS (Beyotime, China), transforming growth factor β1 (TGF-β1, MCE, USA) 10 ng/mL, transforming growth factor β3 (TGF-β3, MCE, USA) 10 ng/mL, bFGF 10 ng/mL, and 1% FBS in DMEM/F12 (or L-DMEM when indicated). Unless otherwise specified, the same induction cocktail was also used to maintain chondrocyte phenotype (chondrocytes expanded in H-DMEM + 10% FBS + 1% P/S during proliferation, then switched to DMEM/F12 + induction cocktail for phenotype maintenance). CO were used at D3 + 3 (3 days spheroid + 3 days induction) for early assays and D3 + 7 for gene expression.

Characterization of BMSCs spheroid and CO

IF of spheroid/CO: Spheroid/CO were fixed with 4% paraformaldehyde for 10–15 min at room temperature, permeabilized with 0.1% Triton X-100 for 10 min, and blocked in QuickBlock™ blocking buffer (Beyotime, China) for 10 min. For the first round, specimens were incubated with rabbit anti-rat CD73 (1:200, overnight at 4 °C, Servicebio, China), washed three times with PBS, and then incubated with goat anti-rabbit IgG–Alexa Fluor 488 (1:500, 1 h, room temperature, protected from light, Servicebio, China). After three PBS washes, samples were incubated with QuickBlock™ Blocking Buffer for 10 min and washed again with PBS. For the second round, rabbit anti-rat CD90 (1:100, overnight at 4 °C, Servicebio, China) was applied under light protection, followed by three PBS washes and incubation with goat anti-rabbit IgG–Alexa Fluor 594 (1:500, 1 h, Servicebio, China). Nuclei were counterstained with DAPI (Beyotime, China). Imaging was performed on a confocal laser microscope (Nikon, Japan) with identical settings across groups.

EdU assay: Proliferation was assessed by EdU incorporation (Beyotime, China) with 3 h incubation at 37 °C, followed by the manufacturer’s click-chemistry labeling protocol. EdU⁺ nuclei (%) was quantified as EdU⁺/DAPI×100%.

Quantitative real-time PCR (qRT-PCR)

For spheroid/CO (Fig. 4) and DLP-printed ears (Fig. 5), total RNA was extracted with TRIzol (ThermoFisher, USA) following the manufacturer’s instructions. Briefly, spheroid/CO were collected by gentle trypsinization and PBS rinse; printed constructs (chondrocyte-laden or CO-laden) were rinsed in PBS and manually minced and triturated on ice to fully release cells before adding TRIzol. RNA was reverse-transcribed to cDNA and qRT-PCR was performed on a StepOnePlus Real-Time PCR System (Applied Biosystems). Primers (COL2A1, ACAN, COL10A1, SOX9) are listed in Table S1. Gene expression was normalized to β-actin using the 2^–ΔΔCt method. Each group was tested in three independent experiments, with technical triplicates.

Fig. 4.

CO Construction and characterization. A Schematic of BMSC spheroid formation and chondrogenic induction to generate CO. B EdU staining of Day 3 spheroid (D3 spheroid), Day 6 spheroid (D6 spheroid), and Day 3 induced CO (CO D3 + 3), with DAPI labeling all cell nuclei (scale bar = 100 μm). C Percentage of EdU⁺ nuclei in D3 spheroid, D6 spheroid, and CO D3 + 3 (*p < 0.05, **p < 0.01). D Immunofluorescence staining for CD73 and CD90 in D3 spheroid, D6 spheroid, and CO D3 + 3 (Rabbit anti-rat primary antibody; goat anti-rabbit AF488/AF594 secondary antibody; sequential staining), with DAPI staining for nuclei (scale bar = 100 μm). E qPCR results for COL2A1 and ACAN in uninduced BMSC spheroid and fully induced CO, normalized to β-actin and expressed as 2^−ΔΔCt (***p < 0.001)

Fig. 5.

Long-term in vitro culture of DLP bioprinted constructs using GelNB/HAMA-based CO-laden bioink. A Fluorescent images of live/dead staining and 3D reconstructed confocal images of ear cartilage printed with chondrocyte-laden and CO-laden bioinks at days 7, 14, and 21 (scale bar = 1000 μm). B qPCR analysis of COL2A1, ACAN, and COL10A1 gene expression in constructs printed with chondrocyte-laden and CO-laden bioinks at days 7, 14, and 21 (*p < 0.05, **p < 0.01, ****p < 0.0001, ns = not significant)

Bioink preparation

The bioink was freshly prepared under sterile conditions with a composition of 10% (w/v) GelNB + 1% (w/v) HAMA, supplemented with 0.05% (w/v) LAP (EFL, China). The mixture is gently homogenized to prevent bubble formation and serves three purposes: (i) as an acellular ink; (ii) as an ink loaded with chondrocytes (1 × 10⁷ cells/mL); (iii) as an ink loaded with CO (5000 organoids/mL). The chondrocyte density of 1 × 10⁷ cells/mL was selected based on commonly used cellular densities in cartilage bioprinting studies [27–31]. The CO loading density was designed to match the total cellular input of the chondrocyte-laden group. Because each CO was generated from approximately 2000 BMSCs, 5000 COs/mL corresponded to approximately 1 × 10⁷ cells/mL. Therefore, the CO-laden group was established as a cell-equivalent comparison to the chondrocyte-laden group, allowing us to evaluate the biological effect of organoid organization rather than differences in initial cell number.

DLP bioprinting

The auricle and other intricate models were designed using UG 10.0 software with a slice thickness of 30 micrometers. Printing was performed on a DLP printer (EFL-BP-8600, China) at a light intensity of 14 mW cm⁻² and a layer thickness of 30 μm. Following printing of cartilage cell/CO-laden constructs, the constructs were transferred using sterile forceps, rinsed three times with PBS, and subsequently placed in the corresponding culture medium for further cultivation.

Live/dead staining of in-vitro printed constructs

Printed auricles (chondrocyte-laden and CO-laden) were rinsed with PBS 3 times. Dilute the Calcein-AM stock solution with PBS to 1 µg mL⁻¹ to prepare the working solution for live cells, propidium iodide (PI) was prepared at 1 µg mL⁻¹ in PBS per the supplier’s instructions. Samples were incubated with calcein-AM at 37 °C for 30 min, then PI was added and incubated 5 min at room temperature in the dark. After a PBS rinse, imaging was performed on confocal laser microscope. We collected z-stacks over ~ 300 μm and generated 3D reconstructions; laser/PMT settings were kept identical across groups and time points.

SEM

Hydrogel samples (10% GelNB composite with 0.5%, 1.0%, or 2.0% HAMA) were fixed in 2.5% (v/v) glutaraldehyde for 4 h, dehydrated using a gradient ethanol solution replaced with tert-butanol, and freeze-dried. The dried samples were sputter-coated with gold under vacuum (20 mA, 150 s) and subsequently imaged using a scanning electron microscope (Quanta 200, FEI, USA).

Mechanical property testing

Cylindrical compression specimens (9 mm diameter × 5 mm height) and rectangular tensile specimens (40 mm × 10 mm × 1 mm) were prepared. Testing employed a 20 g load cell (ZHIQU ZQ770, China), loading at 0.01 mm s⁻¹ to 60% strain. Young’s modulus for compression and tension were determined from the linear strain range of 2%–10%. Each test series comprised three independent replicates, with three specimens per replicate.

Swelling

Weigh the initial mass of the dry sample (W_d). Immerse it in PBS at 37 °C, remove at predetermined time points, gently blot surface moisture, and reweigh (W_t). Calculate the swelling rate using the formula Q = (W_t − W_d)/W_d and plot the curve over time.

Enzymatic degradation of acellular hydrogels

The hydrogel samples were placed in a 0.1% (w/v) collagenase II solution and incubated at 37 °C. Samples were retrieved at predetermined time points, rinsed, freeze-dried, and weighed. Degradation extent was calculated as the percentage of residual mass relative to the initial dry mass, and degradation curves were plotted.

In vitro degradation of acellular, chondrocyte-laden and CO-laden constructs

Printed acellular, chondrocyte-laden and CO-laden constructs were maintained in culture medium under standard cell culture conditions. At predetermined time points, constructs were collected, excess surface liquid was gently removed, and the samples were weighed. The remaining mass was normalized to the initial mass at day 0 and expressed as a percentage.

DNA quantification of printed constructs

To compare the total cellular content of the printed constructs, chondrocyte-laden and CO-laden constructs were fully digested with collagenase II to release the encapsulated cells and organoids. The collected samples were then lysed in a buffer containing 50 mM Tris-HCl (pH 8.0), 10 mM EDTA, 0.5% SDS, and 0.2 mg/mL proteinase K. DNA was purified by sodium acetate/ethanol precipitation, redissolved in nuclease-free water, and quantified using a NanoDrop spectrophotometer.

Subcutaneous nude mice implantation

Eighteen male BALB/c-nu immunodeficient mice (7 weeks old, weighing 17–20 g) were randomly divided into three groups: the acellular group, the chondrocyte-laden group, and the CO-laden auricular construct group. For each group at each time point, three mice were used, and each mouse received two constructs in bilateral subcutaneous pockets, yielding six explanted constructs per group per time point. The constructs were printed under sterile conditions using digital light processing technology and pre-soaked in complete medium prior to implantation. Mice were anesthetized via intraperitoneal injection of 2.5% (w/v) sodium pentobarbital. A 10-mm skin incision was made along the dorsal midline. Subcutaneous pockets were formed bilaterally by blunt dissection, with one auricular construct implanted into each pocket. The skin was closed with 5-0 interrupted sutures to secure the construct position and prevent deformation. Mice were euthanized at the 4-week and 8-week time points, and the grafts were harvested for subsequent analysis.

Histological analysis

Explants were fixed overnight in 4% paraformaldehyde and embedded in paraffin. Sections were stained with hematoxylin and eosin (H&E) as well as Safranin O/Fast green staining. For immunohistochemical staining (IHC) of ACAN and COL II, sections underwent dewaxing, antigen retrieval (Beyotime, China), and blocking. They were then incubated overnight at 4 °C with rabbit anti-ACAN and anti-COL II primary antibody (Servicebio, China). HRP-labeled secondary antibody was subsequently applied, followed by DAB development, counterstaining with hematoxylin, dehydration, and mounting. To evaluate elastic fiber-associated matrix deposition, Elastica–van Gieson (EVG) staining was performed on paraffin sections and incorporated into the main histological analysis. Briefly, EVG-stained sections were stained with EVG working solution, differentiated in acid alcohol, counterstained with van Gieson solution, and then dehydrated, cleared, and mounted. All histological and immunohistochemical images were acquired under identical microscopic conditions (Nikon, Japan). ImageJ software was used to quantify the percentage areas of ACAN-positive, COL II-positive, and EVG-positive staining within tissue-containing regions. EVG-positive area was interpreted as elastic fiber-associated matrix deposition rather than elastin-specific protein expression.

In vivo and in vitro graft compressive modulus

After graft removal (at 4 and 8 weeks), compression testing was performed using a 20 g load sensor (ZHIQU ZQ770, China). Samples were loaded at a rate of 0.01 mm s⁻¹ to 50% strain. Young’s modulus was calculated from the linear segment of the stress-strain curve (2–10% strain range). For in vivo explants, six constructs obtained from three mice were analyzed per group per time point. Each mouse contributed two constructs from bilateral subcutaneous pockets.

Statistical analyses

Data normality was assessed using GraphPad Prism 9.5 software. Results conforming to normal distribution were expressed as mean ± standard deviation (SD). Intergroup comparisons were performed using Student’s t-test or analysis of variance (ANOVA), followed by Tukey’s post hoc test for multiple comparisons where appropriate. For qPCR experiments, each group included three independent biological experiments, and each biological experiment was analyzed with technical triplicates. For in vivo experiments, each group at each time point included three mice and six explanted constructs, with two constructs implanted in bilateral subcutaneous pockets of each mouse. For histological, immunohistochemical, and EVG quantification, multiple non-overlapping fields from each explanted construct were analyzed. Differences were considered statistically significant at p < 0.05 and non-significant at p > 0.05.

Results

The gross appearance of hydrogels prepared with 10% GelNB and 0.5%, 1.0%, or 2.0% HAMA is shown in Fig. 2A. As the HAMA content in the composite system increased, SEM images revealed that the cross-section of the hydrogel gradually shifted from a loose, macroporous structure to a more compact network structure (Fig. 2B). Tensile quantitative analysis revealed a stepwise increase in tensile modulus with rising HAMA concentration (0.5% < 1% < 2%): p = 0.0023 between 0.5% and 1%, p < 0.0001 between 0.5% and 2%, and p < 0.0001 between 1% and 2% (Fig. 2C). Regarding compressive properties, the 2% HAMA group was significantly stiffer, while no significant difference existed between the 0.5% and 1% groups (p = 0.7333). Significant differences were observed between the 0.5% and 2% groups and between the 1% and 2% groups (p < 0.0001, p < 0.0001) (Fig. 2D). Degradation assays demonstrated that high-concentration HAMA slowed the enzymatic degradation rate of hydrogels in 0.1% collagenase II, indicating a stable network structure capable of prolonged in vivo persistence (Fig. 2E). Swelling tests revealed that higher HAMA concentrations imparted denser structures to the hydrogel system, reducing water absorption effects. All formulations reached maximum swelling after approximately 12 h of immersion (Fig. 2F). Considering printability, workability, swelling properties, and degradation, the 10% GelNB + 1% HAMA formulation was selected for all subsequent printing experiments.

Fig. 2.

Characterization of GelNB/HAMA hydrogels at different concentrations. A Macroscopic images of hydrogels containing 10% GelNB with 0.5%, 1.0%, or 2.0% HAMA (scale bar = 5 mm). B Scanning electron microscopy (SEM) images of the network structures corresponding to hydrogels at different concentrations (scale bar = 100 μm). C Tensile stress-strain curves and corresponding tensile Young’s modulus (**p < 0.01, ****p < 0.0001). D Compression stress-strain curves and corresponding compressive modulus (****p < 0.0001, ns = not significant). E Enzymatic degradation profiles (remaining mass percentage) of hydrogels with different concentrations in 0.1% collagenase II. F Swelling curves of hydrogels with different concentrations

Three distinct ink formulations (acellular, chondrocyte-laden and CO-laden) enabled the preparation of structurally distinct auricles via DLP printing. Macroanatomical features such as the helix, antihelix, and concha cavity were clearly discernible in the printed specimens, with localized magnification revealing sharp model edges. When lifting the 3D-printed auricles from each group with tweezers, no significant collapse or structural imbalance occurred, indicating sufficient post-printing strength to support subsequent operations like transfer (Fig. 3A, B). Notably, loading of CO did not cause structural distortion or feature blurring. Beyond auricle models for subsequent experiments, we successfully reproduced other complex structures, demonstrating the ink’s printability and the reliability of the printing process (Fig. 3C).

Fig. 3.

DLP Bioprinting Using GelNB/HAMA Bioink. A DLP-printed auricular structures prepared using acellular, chondrocyte-laden and CO-laden inks, with magnified details (scale bar = 1 mm). B Tweezers lifting the printed specimen demonstrate its shape retention capability. C Display of other complex geometries (scale bar = 1 cm)

For the construction of CO (Fig. 4A), we characterized it based on proliferation activity, MSC-associated marker expression, and chondrogenic-related gene expression. Because adult BMSC-derived cartilage organoids do not yet have a universally accepted marker framework, CD73 and CD90 were not used as definitive cartilage organoid markers. Instead, they were used as exploratory MSC-associated phenotypic markers to monitor the transition from a stem-like BMSC spheroid state toward a chondrogenically induced CO state [32]. EdU staining provided complementary information on proliferative activity, whereas qPCR analysis of SOX9 and COL2A1 was used to assess chondrogenic commitment. For EdU analysis, starting with BMSC spheroid cultured for 3 days (D3 spheroid), both continued maintenance culture for 3 days (D6 spheroid) and chondrogenic induction for 3 days (CO D3 + 3) resulted in a significant decrease in positive rates (CO D3 + 3 compared with D3 spheroid, p = 0.0015; induced CO D3 + 3 vs. D3 spheroid (p = 0.0125), and D6 spheroid vs. D3 spheroid (p = 0.0354; Fig. 4B, C). IF results for stem cell-related markers aligned with this trend—CD73/CD90 exhibited stronger expression on day 3, with fluorescence intensity decreasing after continuing without induction until day 6, and significantly diminishing after 3 days of chondrogenic induction (Fig. 4D). These findings were interpreted as descriptive evidence of phenotypic transition rather than as proof that CD73/CD90 downregulation directly determines functional chondrogenic capacity. qPCR analysis (Fig. 4E) revealed elevated expression of chondrogenic genes in BMSCs after 3 days of spheroid formation followed by 7 days of chondrogenic induction, compared to BMSC spheroid maintained for 10 days. Notably, SOX9 (p = 0.0009) and COL2A1 (p = 0.0002) showed significant upregulation.

Live/dead staining performed on days 7, 14, and 21 revealed strong calcein green fluorescence signals throughout the entire ear thickness in both groups, with only scattered propidium iodide signals and no obvious necrotic cores observed (Fig. 5A). qPCR analysis of the entire construct revealed divergent trends over time between the two groups (Fig. 5B). Compared with chondrocyte-laden controls, CO-laden constructs showed higher ACAN expression at days 7, 14, and 21, higher COL2A1 expression by day 21, and reduced COL10A1 expression at day 21, indicating enhanced chondrogenic matrix formation and reduced hypertrophic drift. In summary, both cell groups exhibited robust viability throughout prolonged in vitro culture. Over time, the CO-laden group demonstrated greater potential to efficiently guide tissue differentiation toward a more stable chondrocyte phenotype. To compare the total cellular content of the printed constructs, chondrocyte-laden and CO-laden constructs were fully digested after printing and subjected to DNA quantification. No significant difference in total DNA content was observed between the two groups (Fig. S2), suggesting that the subsequent differences in chondrogenic performance were unlikely to be explained simply by differences in initial cell dose.

To validate the biocompatibility and ectopic chondrogenic potential of the printed constructs in vivo, acellular, chondrocyte-laden, and CO-laden constructs were implanted subcutaneously in nude mice, and tissue samples were harvested at 4 and 8 weeks after implantation. The explants were well tolerated in vivo and remained intact at retrieval (Fig. 6A). Histological and IHC analyses revealed progressive cartilage-like tissue formation in the cell-laden groups, particularly in the CO-laden group (Fig. 6B–D). Because hydrogel matrices may retain Safranin O/Fast Green staining nonspecifically, as reported in previous cartilage hydrogel studies, Safranin O staining was not interpreted as a standalone indicator of GAG deposition [33–36]. Instead, cartilage-like maturation was assessed by integrating tissue-containing histological fields, lacuna-like chondrocyte morphology, ACAN/COL II immunostaining, EVG staining, and mechanical testing.

Fig. 6.

CO-laden printed constructs promote in vivo cartilage-like matrix maturation. A Representative photographs showing subcutaneous implantation in nude mice and retrieved explants after harvesting (scale bar = 5 mm). B Histological evaluation of explanted constructs at 4 and 8 weeks, including H&E staining, Safranin O/Fast Green staining, and EVG staining. Dashed boxes indicate the magnified regions shown below each low-magnification image (scale bar = 500 μm). C Immunohistochemical evaluation of ACAN and COL2A1 expression in explanted constructs at 4 and 8 weeks. Dashed boxes indicate the magnified regions shown below each low-magnification image (scale bar = 500 μm). D Quantitative analysis of COL2A1-positive area, ACAN-positive area, and EVG-positive area fraction in explanted constructs at 4 and 8 weeks. EVG-positive area fraction was quantified within tissue-containing regions and interpreted as elastic fiber-associated matrix deposition rather than elastin-specific protein expression. E Unconfined compression testing of explanted constructs, including representative stress-strain curves and compressive Young’s modulus values. Data are presented as mean ± SD from six explanted constructs obtained from three mice per group at each time point. For each group at each time point, three mice were used, and two constructs were implanted in bilateral subcutaneous pockets of each mouse. *p < 0.05, **p < 0.01, ****p < 0.0001, ns = not significant

At week 4, the advantage of the CO-laden group was not yet fully evident in routine histology. H&E staining revealed more obvious cartilage lacuna-like structures in the chondrocyte-laden group, whereas the CO-laden group showed stronger matrix staining in Safranin O/Fast Green sections. ACAN IHC signals remained relatively weak at this stage, and quantitative analysis of ACAN-positive area showed no significant difference between the CO-laden and chondrocyte-laden groups (p = 0.3756), although the CO-laden group was significantly higher than the acellular group (p = 0.0386). In contrast, COL II IHC already revealed more abundant positive staining in the CO-laden group at week 4, with quantitative analysis showing a markedly higher COL II-positive area in the CO-laden group (20.207 ± 2.060%) than in the chondrocyte-laden group (8.413 ± 1.511%) and the acellular group (1.867 ± 2.043%). EVG staining showed limited elastic fiber-associated matrix deposition at week 4, and EVG-positive area fraction did not differ significantly among groups.

By week 8, the intergroup differences became more pronounced. H&E staining showed that the CO-laden group contained more numerous and mature cartilage lacuna-like structures. Safranin O/Fast Green staining exhibited broader distribution and stronger matrix staining in the CO-laden group. In parallel, both ACAN and COL II IHC signals were markedly enhanced in the CO-laden group. Quantitative analysis confirmed that ACAN-positive area in the CO-laden group was significantly higher than that in both the acellular and chondrocyte-laden groups (p < 0.0001 for all comparisons). Likewise, the COL II-positive area remained highest in the CO-laden group at week 8 (21.377 ± 6.740%), compared with the chondrocyte-laden group (2.837 ± 0.858%) and the acellular group (1.067 ± 0.846%), indicating superior cartilage-specific extracellular matrix deposition in the CO-laden constructs. Importantly, EVG staining showed increased EVG-positive elastic fiber-associated matrix in the CO-laden group at week 8. Quantitative analysis confirmed that EVG-positive area fraction was significantly higher in the CO-laden group (16.753 ± 4.953%) than in both the acellular (1.110 ± 0.900%) and chondrocyte-laden groups (3.083 ± 0.887%). These results support enhanced cartilage-like maturation and preliminary elastic fiber-associated matrix deposition in CO-laden constructs, while not proving complete native-like elastic cartilage regeneration.

The mechanical properties of the explanted constructs were generally consistent with the histological findings (Fig. 6E). At week 4, the CO-laden group exhibited a lower compressive modulus than both the chondrocyte-laden group (p = 0.0167) and the acellular group (p < 0.0001). This early mechanical weakness may reflect a transient imbalance between cell/organoid-associated hydrogel degradation and the accumulation of newly deposited cartilage-like ECM. By week 8, with further maturation of the cartilage organoids and sustained matrix secretion, the CO-laden group achieved the highest compressive Young’s modulus, significantly exceeding both the chondrocyte-laden and acellular groups (all comparisons p < 0.0001). Together with the increased ACAN/COL II-positive areas, more abundant lacuna-like cartilage morphology, and higher EVG-positive area fraction, these data suggest a mechanical evolution from early construct loosening to later matrix-associated mechanical recovery. To further interpret these early mechanical differences, the in vitro degradation profiles of acellular, chondrocyte-laden, and CO-laden constructs were additionally evaluated over the first 14 days. The acellular group showed the slowest mass loss, whereas the cell-laden groups, particularly the CO-laden group, exhibited faster decreases in construct remaining mass (Fig. S1). Therefore, the faster mass loss of CO-laden constructs is consistent with, but does not directly prove, a remodeling-like phase. Direct evidence from MMP activity, pericellular matrix accumulation, or other remodeling-associated markers will be required in future studies to clarify this mechanism.

Discussion

Ear reconstruction stands at the crossroads of form and function: patients seek ears that appear lifelike, feel elastic, and endure over time. Current clinical options primarily include autologous rib cartilage grafts or synthetic prosthetic implants, each with inherent limitations [37, 38]. Rib cartilage grafts entail donor site complications, require staged procedures, and offer limited elasticity while implants may exhibit exposure or extrusion and cannot generate autologous cartilage [39]. This study aims to bridge this gap through a strategy that maintains anatomical fidelity of ear dimensions while promoting cartilage-like tissue maturation. It should be noted, however, that the present work represents an early ectopic proof-of-concept study based on a fully rodent system rather than a clinically equivalent auricular reconstruction model. Rat BMSCs and chondrocytes were selected primarily because they are readily available, reproducible, and suitable for repeated formulation screening, organoid induction, and full-construct validation. The nude mouse dorsal subcutaneous model is useful for preliminary evaluation of in vivo chondrogenic potential [40, 41], but it does not reproduce the mechanical loading, vascular environment, immune context, or soft-tissue coverage conditions of clinical auricular reconstruction. Future studies should further evaluate this strategy using human MSCs, human auricular chondrocytes, and clinically relevant large-animal auricular defect models.

We combined CO with DLP bioprinting technology and an optimized GelNB/HAMA bioink to fabricate full-scale auricles. This study (Fig. 1) conducted parallel in vitro and in vivo testing on acellular, chondrocyte-laden, and CO-laden auricles. Material testing indicated that 10% GelNB + 1% HAMA provided a practical balance, including printability, ease of handling and allowing tissue remodeling. Structural mechanical properties improved with increasing HAMA content, though this was accompanied by reduced hydrogel toughness. Swelling/degradation behavior also correlated with HAMA concentration (Fig. 2). Using this bioink, DLP printing clearly reproduced the ear’s landmark structures across all groups; loading with CO did not impair printing performance or structural integrity (Fig. 3).

We selected the GelNB/HAMA network system based on recent research confirming that, through a light-triggered radical coupling mechanism, GelNB reinforced with HAMA can form dispersed hard phases within a relatively soft gelatin matrix [24, 42]. The resulting hydrogel undergoes rapid photopolymerization and exhibits exceptional strength and toughness, which are rare in gelatin-based systems, while maintaining its biodegradability and cell adhesion [43]. This unique combination directly facilitates bioprinting at the auricular scale. In practice, this manifests as: (i) compatibility with digital light processing and self-supporting capability during printing of intricate auricular structures; (ii) precise tuning of stiffness and swelling properties by adjusting HAMA content to achieve an “optimal balance point” for shape fidelity while avoiding excessive restriction of cellular remodeling (consistent with our formulation choice of 10% GelNB + 1% HAMA in Fig. 2); (iii) excellent cell compatibility and tissue compliance, supporting enhanced cell adhesion/migration compared to common polymer scaffolds while inducing milder acute inflammatory responses in vivo; and (iv) complete enzymatic degradation, creating space for new cartilage matrix deposition during organoid maturation. These properties align with our findings: printed structures maintained geometric integrity (Fig. 3), supported cell viability in vitro while promoting ACAN/COL2A1 upregulation and COL10A1 downregulation (Fig. 5), and enabled robust GAG/ACAN and COL II in vivo with higher compressive modulus (Fig. 6). In summary, GelNB/HAMA provides the required balance of processability, mechanical properties, and biological functionality for ear scaffolds loaded with CO: rapid, precise digital light processing curing; sufficient initial strength to maintain the clarity of the ear’s landmark structures; and a degradation, cell-friendly microenvironment ensuring organoids can effectively perform their functions. Our findings should also be interpreted in the context of our group’s previous work on DLP cartilage bioprinting. In that earlier study, microtissue-based DLP bioinks were shown to function as chondrocyte “microshelters,” supporting both printability and cartilage regeneration. The present study extends this concept by using stem cell-derived cartilage organoids rather than chondrocyte microtissues, adopting a GelNB/HAMA-based photocurable system, and focusing specifically on auricle-scale constructs with anatomically recognizable external ear features [25].The principle advancement of this study lies on the use of CO as active building blocks for bioink, demonstrating that DLP printing loaded with CO can be achieved at the auricular scale [44, 45]. Unlike individual chondrocytes dispersed in gel, CO retain tight cell-cell junctions and pericellular-like matrix—features that mimic early mesenchymal aggregates and support lineage stability [18, 44]. In our assays, the transition from BMSC spheroid to CO reduced proliferative activity and downregulated mesenchymal stem cell markers while upregulating chondrogenic transcription factors (Fig. 4). Because there is still no widely accepted characterization framework for adult stem cell-derived cartilage organoids, we used the downregulation of canonical MSC markers together with the upregulation of chondrogenic markers as practical evidence that the spheroids had transitioned toward a chondrogenically committed organoid state. We have clarified that CD73/CD90 immunofluorescence was included as exploratory source-cell phenotypic characterization, not as a definitive organoid maturity marker or a standalone predictor of functional chondrogenesis. Functional chondrogenic maturation was instead supported by the combined increase in chondrogenic gene expression, cartilage-like morphology, ACAN/COL II deposition, EVG-positive matrix, and in vivo mechanical evolution [32]. This “first-mover advantage” proves critical post-printing. During 21 days of in vitro culture, qPCR analysis of entire constructs revealed superiority in CO-laden ears: ACAN expression was higher at days 7/14/21, COL2A1 was significantly elevated at day 21 compared to chondrocyte-loaded groups, and hypertrophy marker COL10A1 was markedly reduced at day 21 (Fig. 5). In short, maintaining cells within an organoid microenvironment helps optimize the balance between matrix construction (ACAN, COL2A1) and hypertrophy control (COL10A1). A further limitation is that qPCR analysis was performed on whole printed constructs and therefore reflects an average transcriptional profile across the entire auricle. The observed differences may partly reflect spatial heterogeneity, differences in local organoid persistence, or regional matrix deposition. We therefore interpreted qPCR together with spatial histological and immunohistochemical analyses, including ACAN, COL II, and EVG staining. Future work using region-defined sampling, spatial transcriptomics, or single-cell RNA sequencing will be needed to determine whether chondrogenic commitment is uniformly distributed throughout the construct or concentrated in focal organoid-rich regions.

To further support this conclusion, we conducted in vivo studies. At week 4, grafts loaded with CO exhibited more cartilage-like island structures, although ACAN IHC staining signals were not intense, quantitative analysis showed they were significantly higher than in the acellular group and comparable to the chondrocyte-laden group. At this early stage, type II collagen immunohistochemistry already showed stronger positive staining in the CO-laden group, suggesting that cartilage-specific matrix deposition in the organoid-laden constructs may precede the full emergence of histological differences in routine staining. By week 8, differences became pronounced: H&E staining revealed more numerous and mature cartilage lacunae, while Safranin O/Fast green staining showed a more extensive GAG-rich matrix, with ACAN IHC positive areas exceeding both controls and with COL II immunohistochemical staining also remaining highest in the CO-laden group (Fig. 6B–D). Following in vivo retrieval, specimens underwent mechanical testing. At week 4, the acellular group exhibited maximal mechanical property preservation, potentially due to concurrent cellular degradation of the hydrogel without timely formation of a mature extracellular matrix in the other two groups. This interpretation was further supported by the early in vitro degradation profiles of the constructs, in which the acellular group showed the slowest mass loss, whereas the cell-laden groups—particularly the CO-laden group—degraded more rapidly during the first 14 days (Figure S1). By week 8, the CO-laden group achieved the highest compressive Young’s modulus, showing significant differences in pairwise comparisons with both the chondrocyte-laden group and the acellular group (Fig. 6E). However, the absolute modulus of the CO-laden constructs remained below reported values for mature native auricular cartilage [46]. To place this result in the context of engineered auricular cartilage, the 8-week modulus of the CO-laden constructs in this study (~ 25.6 kPa) was higher than that reported for some hydrogel-dominant auricular cartilage models, such as chondrocyte-laden alginate constructs with moduli of approximately 3.5–6 kPa [47], but remained lower than more mechanically reinforced or longer-matured bioprinted auricular cartilage systems, which have been reported to reach tens to hundreds of kPa depending on construct design, culture duration, and testing mode [48, 49]. Therefore, the current mechanical data should be interpreted as evidence of improved functional maturation relative to internal controls and some soft hydrogel-based systems, rather than complete restoration of native auricular cartilage mechanics. Additional maturation time, mechanical conditioning, and structural reinforcement may be required to achieve clinically sufficient early mechanical stability and long-term shape retention. For the composition of the bioink, we did not merely select the hardest gel or one which degrades the slowest. The auricle requires fine surface details to achieve realistic appearance and sufficient remodeling space to form cartilage. The 10% GelNB + 1% HAMA formulation strikes this optimal balance, exhibiting appropriate modulus, moderate swelling properties, and controllable collagenase response. These characteristics allow the construct to maintain its shape while enabling cells and CO to deposit their matrix. For bioprinting, DLP offers high resolution with minimal shear forces, which helps preserve CO integrity compared to high-shear extrusion printing [50–52]. In practice, we only slightly extend exposure time for printed constructs loaded with CO to compensate for light scattering, while maintaining sharp feature definition. However, the spatial distribution and tissue-filling efficiency of multicellular COs within DLP-printed constructs remain technically challenging. Compared with single cells, COs are larger multicellular aggregates and may sediment more rapidly during bioink handling and layer-by-layer printing. Simply increasing the total CO concentration may improve tissue filling, but it may also increase optical scattering, reduce curing uniformity, and compromise DLP printing fidelity. Therefore, future optimization should distinguish total cell-equivalent loading from local organoid packing density per unit print area. Increasing CO packing density through spatial patterning or controlled organoid placement, without necessarily increasing the overall cell concentration beyond the printability window, may improve construct colonization and tissue occupancy.

In addition, a hybrid seeding strategy that combines COs with interstitial single chondrocytes or progenitor cells may represent a conceptually distinct approach to improve construct filling. In such a design, COs could serve as chondrogenic organizing centers, while dispersed interstitial cells could occupy the spaces between organoids, bridge organoid-rich regions, and promote more homogeneous matrix deposition throughout the printed auricle. Higher-density CO bioinks, anti-sedimentation formulations, dynamic mixing strategies, spatially patterned organoid deposition, and hybrid CO/single-cell printing will therefore be key directions for future optimization [53, 54].

Furthermore, while single-cell bioink handling is straightforward, dispersed chondrocytes may weaken collaborative signaling and promote chondrocyte hypertrophic drift during culture [18, 55, 56]. Comparative data in this study clearly demonstrate the distinction: CO-laden auricles exhibit a chondrogenic transcriptional profile in vitro (upregulation of ACAN/COL2A1, downregulation of COL10A1) and translate these advantages in vivo to richer GAG content, stronger ACAN expression, more abundant COL II deposition, and higher stiffness. Crucially, it is the intact microarchitecture of the CO—not merely the accumulation of cells—that drives maturation. Supporting this interpretation, post-print DNA quantification showed no significant difference in total DNA content between the chondrocyte-laden and CO-laden constructs (Figure S2), suggesting that the superior performance of the CO-laden group is unlikely to be explained simply by differences in effective cell dose. Compared with decellularized-matrix or scaffold-reinforced auricular strategies, the CO-based approach places greater emphasis on living self-organized biological units that can actively contribute cell-cell signaling and matrix secretion from the outset. By contrast, decellularized cartilage-derived bioinks and external scaffold-based systems offer strong structural guidance and tissue-specific matrix cues, but rely less directly on the intrinsic self-organization capacity of living multicellular building blocks. These approaches should be viewed as complementary rather than competing: scaffold-heavy systems may be advantageous for immediate shape support, whereas CO-based systems may offer stronger biologically driven remodeling potential [4, 5]. This principle is believed to extend beyond auricular reconstruction to other cartilage morphologies [57–61].

From a clinician’s perspective, the study’s implications are twofold: (i) DLP printing technology delivers shape fidelity for patient-specific anatomical structures; (ii) the superior biological properties of the GelNB-HAMA-based bioink used to construct CO facilitate stable cartilage tissue formation with reduced hypertrophic signaling. A feasible translational pathway may involve personalized model construction based on imaging data, printing of CO-laden implants, brief precultivation to establish ECM and implantation combined with soft tissue coverage [62]. At the same time, our data should be interpreted against the broader goal of true elastic cartilage engineering. Prior studies on auricular or nasal/alar cartilage fabrication have highlighted the continuing importance of elastin deposition, elastic fiber organization, and tissue-specific mechanical behavior, often requiring additional structural support, specialized maturation conditions, or longer-term remodeling to approach native tissue properties [63–65]. More broadly, recent organoid-based tissue engineering reviews have also emphasized that long-term translation will depend not only on biological performance, but also on reproducibility, batch standardization, and scalable manufacturing [17, 20].

Nevertheless, several limitations of the present study should be acknowledged. First, although the auricle is composed of elastic cartilage, our evaluation mainly focused on chondrogenic matrix markers such as GAG, ACAN, and type II collagen. EVG staining quantified as EVG-positive area fraction to provide additional evidence of elastic fiber-associated matrix deposition. However, EVG staining is histochemical rather than elastin-specific, and the current data remain insufficient to conclude that the engineered tissue fully recapitulates mature native elastic cartilage, particularly with respect to elastin organization, fibrillin-rich elastic fiber networks, and mechanical anisotropy. Future studies should include elastin-specific immunostaining, fibrillin-1 analysis, ultrastructural imaging, and proteomic profiling to further define the elastic cartilage phenotype. Second, auricular chondrocytes are an important target cell source for ear reconstruction; however, in this study we prioritized BMSC-derived CO as a scalable and practical biological building block for proof-of-concept validation. Direct comparison with auricular chondrocyte-based constructs, especially using human cells, will be an important direction for future work. Third, the current in vivo study used only 4- and 8-week ectopic subcutaneous timepoints, which are appropriate for early feasibility evaluation but insufficient to address long-term maturation, durability, and clinically relevant implantation scenarios. Although each group at each time point included six explanted constructs, these constructs were obtained from three host mice, with two bilateral implants per mouse. Therefore, the host animal number remains relatively limited. The current in vivo results should be interpreted as preliminary proof-of-concept evidence. Larger animal cohorts, longer follow-up periods, and more clinically relevant implantation models will be required to confirm reproducibility, long-term stability, and functional performance.

In summary, this study demonstrated that DLP auricle printing loaded with CO can be achieved at a scalable level by utilizing CO as the active unit in the bioink. Within the scope of this proof-of-concept rodent study, we demonstrate a viable pathway that preserves anatomical integrity while guiding maturation toward the desired trajectory—achieving higher ACAN/COL2A1 expression, lower COL10A1 expression, stronger type II collagen deposition, stronger ACAN expression, and stiffer grafts over time. This approach addresses genuine clinical challenges in auricular reconstruction and provides a template for organoid-based bioprinting applications in other cartilage scenarios.

Conclusions

Our study demonstrates an exploratory approach for auricular cartilage-like tissue engineering that combines anatomical fidelity with enhanced cartilage-like maturation: loading cartilage organoid (CO) into O-nitrobenzyl functionalized gelatin (GelNB)/ methacrylated hyaluronic acid (HAMA) bioink and printing full-scale auricles via digital light processing (DLP) technology. This approach not only preserves the ear’s intricate architecture but also creates a microenvironment conducive to chondrogenesis—printed constructs loaded with CO maintain robust viability in vitro and transition to a stable cartilage phenotype with reduced hypertrophic drift. In vivo, compared to cartilage cell-loaded and acellular control groups, it formed richer glycosaminoglycan (GAG)-positive matrix, more mature lacuna-like structures, stronger ACAN staining intensity, and higher compressive modulus. The selected formulation (10% GelNB + 1% HAMA) balances printability, mechanical properties, swelling capacity, and degradation, accommodating both operational feasibility and tissue remodeling potential. In vivo, compared with chondrocyte-laden and acellular controls, CO-laden constructs showed more abundant lacuna-like cartilage morphology, stronger ACAN and COL II staining, increased EVG-positive elastic fiber-associated matrix, and higher compressive modulus at 8 weeks. Consistency across molecular, histological, and mechanical outcomes in the heterotopic transplantation model supports CO-laden DLP bioprinting as a promising proof-of-concept strategy for future personalized auricular cartilage-like tissue engineering.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Yuxuan Chen: Conceptualization, methodology, investigation, data curation, formal analysis, visualization, writing—original draft. Yifan Zhang: methodology, investigation, formal analysis, data curation, visualization, writing—original draft. Pengxiang Luo: methodology, investigation, data curation, visualization, writing—review & editing. Shanshan Su: resources, investigation, validation, formal analysis, visualization, writing—review & editing. Wenbin Jiang: resources, investigation, validation, visualization, writing—review & editing. Chao Luo: resources, investigation, methodology, validation, writing—review & editing. Xia Cai (Corresponding author): conceptualization, supervision, project administration, writing—review & editing, funding acquisition. Jiaming Sun (Corresponding author): conceptualization, supervision, resources, writing—review & editing, funding acquisition. Zhenxing Wang (Corresponding author): conceptualization, supervision, project administration, resources, writing—review & editing, funding acquisition.

Funding

This work was supported by the National Key Research and Development Program of China (No. 2024YFC3407600), the National Natural Science Foundation of China (No. 82322046, 82020108020, 82072198), the Open Foundation of Hubei Key Laboratory of Regenerative Medicine and Multi-disciplinary Translational Research (2024zsyx08, 2024KQHM02). Interdisciplinary Research Program of HUST 2023JCYJ026.

Data availability

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

Declarations

Ethics approval and consent to participate

[2025] IACUC Number 4813 was approved by Experimental Animal Ethics Committee of Huazhong University of Science and Technology for the studies involving animal, the name of approved project “Smart Responsive Hydrogels for the Culturing of Bone/Cartilage Organoids and Application in In Vivo Regeneration” and the approval date is September 26, 2025.

Competing interests

The authors declare no competing interests.