Unconventional auricular reconstruction using controlled scaffold buckling and chondrogenic cocktail containing muscle-derived cells Authors: Na Huang, Shui Kwong John Lee, Ka Ho Kenneth Lee, Chunyi Wen, Yun Chang, Ming Zhang, Cheng Dong, Chuanbin Mao, Dan Michelle Wang, Dai Fei Elmer Ker Published: 2026 ABSTRACT Tissue engineered auricles face challenges such as subpar graft mechanical robustness, insufficient stem/progenitor cells, and unidentified factors that specifically promote elastic cartilage. To tackle these, we developed resilient and bioactive 3D-printed scaffolds using two unconventional concepts, including controlled buckling compliant lattices for mechanical resilience and muscle-derived stem/progenitor cells (MDSCs) with defined biochemical factors for facile elastic cartilage-like regeneration. Our functionally graded design reduced stress by 63.1% relative to other graded designs, which was validated by finite element analysis and high-magnitude compressive testing. Notably, 91.7% of scaffolds remained undamaged in stark contrast to 100% failure for clinically-used MEDPOR® and 76.5% failure for other graded scaffolds. Further, we developed an elastic auricular regenerative cocktail (EARc) comprised of abundantly available MDSCs and elastic cartilage-specific biochemical factors. In vitro, mouse, and rabbit studies confirmed that EARc scaffolds facilitated mechanical resilience and elastic cartilage-like regeneration. In conclusion, EARc scaffolds demonstrate the unconventional application of buckling and muscle sourcing to enhance mechanical resilience and elastic-like cartilage-specific regeneration for auricular reconstruction. Introduction Numerous tissue engineered biomaterials such as auricular or external ear scaffolds must often satisfy multiple and potentially competing requirements for successful clinical translation. In auricular reconstruction, these requirements include ensuring that the graft is (i) mechanically resilient to maintain the ear's shape despite physical deformations from daily life activities, as well as (ii) using abundant and relatively atraumatic sources of elastic cartilage to regenerate pinna tissue. However, current clinical treatments do not adequately satisfy such requirements. For example, the current gold standard in clinical treatment is costal cartilage autograft, which exhibits high cartilage bioactivity but is highly traumatic due to significant donor site morbidity [1,2]. Also, costal cartilage autografts entail potential complications such as unintended rapid graft resorption, which can cause loss of graft mechanical resilience and ear shape [3,4]. In this context, alloplastic grafts such as high molecular weight polyethylene (PE)-based MEDPOR® are attractive clinical alternatives due to their elimination of donor site morbidity and well-defined degradation rates for long-term integrity. However, such synthetic grafts are brittle and prone to fracture (i.e., poor mechanical resilience) [5,6] and are limited by their lack of elastic cartilage bioactivity, which hinders graft-host integration. Therefore, improved ear reconstruction outcomes require innovative approaches that mechanically and biologically augment auricular scaffolds. Within this context, researchers have focused on tissue engineering strategies that enable ear scaffolds to closely mimic the biomechanical and biological aspects of native auricles. From a biomechanical perspective, prior efforts have sought to enhance the mechanical robustness of tissue engineered auricles using advanced fabrication techniques, notably 3D-printing and multi-material mold-casting with both natural and synthetic materials. These include 3D-printed polylactic acid (PLA) ear scaffolds combined with decellularized costal cartilage [7], 3D-printed polycaprolactone (PCL) scaffolds combined with collagen type I hydrogels [8], compression-molded PCL and polyglycolic acid (PGA) scaffolds [9], 3D-bio-printed and PCL-acellular cartilage matrix constructs [10], mold-casted PCL scaffolds combined with acellular cartilage matrix and gelatin [11], mold-casted PCL, mold-casted PGA, mold-casted polyhydroxybutyrate (PHB) scaffolds [12], and hydroxyapatite-epoxide acrylate malelic composites [13]. These approaches aim to aesthetically replicate the complex shape of the ear typically via 3D-printing while leveraging the intrinsic bioactivity and mechanical attributes of natural and synthetic material components, respectively. Despite considerable progress, the mechanical resilience of implanted tissue engineered auricles has not been explicitly addressed and may result in unintended graft fracture similar to current alloplastic grafts [5,6,14]. From a biological perspective, prior efforts have sought to address autologous chondrogenic cell insufficiency and graft-host tissue integration by utilizing alternative cell sources and pro-chondrogenic bioactive molecules. These alternative cell sources include autologous auricular chondrocytes [15], microtia chondrocytes [9,16], induced pluripotent stem cells (iPSCs) [17], and mesenchymal stem cells (MSCs) [18], while pro-chondrogenic bioactive molecules include transforming growth factor-βs (TGF-βs) [19], bone morphogenetic proteins (BMPs) [20], and insulin-like growth factor-1 (IGF-1) [21], and extracellular matrix (ECM) proteins [7]. Such studies have demonstrated proof-of-concept for their use in auricular tissue engineering, such as growth factor- [22] and cell-mediated [23] production of a cartilage-specific matrix. However, proposed chondrogenic cell sources still face issues of cell insufficiency while elastic cartilage-specific inductive cues have yet to be reported. Specifically, the estimated 200 million cells required for reconstituting a paediatric ear [24,25] are not readily achievable due to reasons such as rapid loss of chondrogenic ability during expansion [26] or safety concerns regarding genetically modified iPSC cells [27], while a majority of tissue engineering efforts have adopted biochemical factors typically associated with hyaline cartilage induction [[19], [20], [21]], which may result in production of an unwanted calcified cartilage matrix [[28], [29], [30]]. Therefore, further mechanical and biological augmentation of auricular scaffolds is needed, including increased mechanical resilience to withstand daily forces and avoid scaffold fracture, as well as an abundant source of pro-chondrogenic cells with elastic cartilage-specific cues to enhance auricular reconstruction. To address these issues, we have developed a 3D-printed, tissue engineered auricle that exhibits mechanical resilience and pro-chondrogenic bioactivity through unconventional means. Specifically, mechanical resilience was achieved through controlled buckling of 3D-fabricated structures containing functionally graded, compliant lattice structures while pro-chondrogenic bioactivity was attained via an elastic auricular regenerative cocktail (EARc) comprised of abundantly available muscle-derived stem/progenitor cells (MDSCs) and elastic cartilage-specific biochemical cues. The mechanical resilience of our scaffold was optimized in silico via finite element analysis (FEA) and subsequently validated ex vivo by large deformation, compressive testing. The elastic chondrogenic bioactivity of our scaffold was facilitated by ECM (plasminogen-depleted fibrin)-based delivery of growth factors and biochemical molecules, which was characterized ex vivo via Fourier Transform Infrared (FTIR) spectroscopy, release studies, and degradation experiments, while EARc composition was refined for elastic cartilage-like differentiation in vitro. Furthermore, the mechanical resilience and elastic cartilage bioactivity were demonstrated in vivo in both ectopic (mouse subcutaneous implantation) and orthotopic (rabbit ear defect) models. Altogether, this study investigated the unconventional use of buckling and stem/progenitor cells sourced from muscle to attain 3D-printed auricular scaffolds that exhibit mechanical resilience and pro-elastic chondrogenic bioactivity, respectively. Results Ex vivo physical and chemical characterization of 3D-printed PHT auricular scaffolds with functionally graded lattice structures To fabricate a graft suitable for ear reconstruction, auricular scaffolds with functionally graded lattice structures were 3D-printed and characterized chemically and physically ex vivo via FTIR spectroscopy (polymer composition and polymer-hydrogel integration), compressive testing (mechanical attributes), and micro computed tomography/microCT (porosity) (Fig. 1A). Using click-based chemistries, a 3D-printable resin based on a photocrosslinkable polythiourethane (PHT prepolymer) [31] was synthesized under ambient room temperature conditions in less than 30 min and 3D-printed into an auricular scaffold in under 10 s per layer with high fidelity (Fig. 1A and B and Supplementary Fig. S1). The auricular scaffold was comprised of functionally graded lattices with regions that corresponded to the helix, antihelix, tragus, antitragus, and scapha features of the native human ear. FTIR spectroscopy indicated formation of alkyl sulfide bonds and thiocarbamate from reduced peak signals at 2563 cm−1 (-SH) and loss of peak signal at 2265 cm−1 (-NCO), which were attributed to stretching vibrations of thiol and isocyanate groups, respectively, as well as photocrosslinking reactions from reduced peak signals at 1633 cm−1 (-CHCH2), which were attributed to stretching vibrations of acrylate alkene groups (Fig. 1C). FTIR spectroscopy also indicated integration of PHT polymer with fibrin hydrogel from the presence of peak signals at 1740-1706 cm−1 (-CO), and 670 cm−1 (-CS), which were attributed to vibrations of the ester groups [32,33] and stretching of alkyl sulfide bonds [34,35] in PHT polymer, respectively, as well as the presence of peak signals at 1650-1620 cm−1 (-CO) and 1540-1500 cm−1 (-NH), which were attributed to amide I and amide II groups in fibrin hydrogel, respectively (Fig. 1C). Uniaxial compressive testing showed that the intrinsic material properties of PHT polymer were similar to the reported values of auricular tissues [[36], [37], [38], [39]] and differed from clinically-used MEDPOR® by at least an order of magnitude (p = 0.003 for compressive modulus; p = 0.001 for compressive stiffness; Fig. 1D). MicroCT analysis demonstrated that PHT polymer could be 3D-printed as monophasic scaffolds with controlled lattice/pore sizes of 300 μm (Monophasic 1), 500 μm (Monophasic 2), and 700 μm (Monophasic 3) in length, whereas individual pore sizes for MEDPOR® ranged from 160 to 368 μm in diameter [40] (Fig. 1E). Increased lattice size resulted in increased porosity with 3D-printed Monophasic 1, 2, and 3 PHT scaffolds exhibiting 31.90 ± 0.38%, 55.91 ± 3.73%, and 64.97 ± 5.72% porosity, respectively, while MEDPOR® exhibited 71.63 ± 2.16% porosity (Fig. 1E). Together, these data demonstrated that a 3D-printable polyurethane-based scaffold could be rapidly fabricated with excellent ear shape fidelity and bonded with fibrin hydrogel with auricular-like mechanical properties and tunable porosity for ear tissue engineering.Fig. 13D fabrication of functionally graded auricular scaffold with associated chemical (component and fibrin hydrogel integration) as well as physical (mechanical attributes and porosity) characterization via mechanical testing, microCT assessment and FTIR analysis. (A) Experiment setup. The schematic illustrates a photocrosslinkable polythiourethane (PHT prepolymer) that can be easily prepared at room temperature for 3D-printing and was subjected to chemical and physical assessment via FTIR characterization, mechanical testing, and porosity testing. (B) A customizable auricular scaffold was fabricated via digital light processing 3D printing technology using PHT prepolymer with 10 s of visible light exposure per layer and subsequently integrated with fibrin hydrogel. Scale bar = 5 mm. (C) FTIR spectra of individual PHT monomer components (P, H and T) and photocrosslinked PHT polymer showed successful synthesis of PHT polymer and subsequent integration with fibrin hydrogel. (D) Mechanical testing of PHT polymer demonstrated compressive modulus and stiffness that approximated auricular cartilage and were at least an order of magnitude lower than MEDPOR®. n = 3 samples per group for MEDPOR®, n = 6 samples per group for 3D-printed PHT polymer. ∗ denotes statistical significance, which was established at p ≤ 0.05 using a 2-tailed independent sample T-test. (E) MicroCT reconstructions of 3D-printed PHT polymer showed tunable porosity relative to MEDPOR®. Scale bar = 0.5 mm. 3D fabrication of functionally graded auricular scaffold with associated chemical (component and fibrin hydrogel integration) as well as physical (mechanical attributes and porosity) characterization via mechanical testing, microCT assessment and FTIR analysis. (A) Experiment setup. The schematic illustrates a photocrosslinkable polythiourethane (PHT prepolymer) that can be easily prepared at room temperature for 3D-printing and was subjected to chemical and physical assessment via FTIR characterization, mechanical testing, and porosity testing. (B) A customizable auricular scaffold was fabricated via digital light processing 3D printing technology using PHT prepolymer with 10 s of visible light exposure per layer and subsequently integrated with fibrin hydrogel. Scale bar = 5 mm. (C) FTIR spectra of individual PHT monomer components (P, H and T) and photocrosslinked PHT polymer showed successful synthesis of PHT polymer and subsequent integration with fibrin hydrogel. (D) Mechanical testing of PHT polymer demonstrated compressive modulus and stiffness that approximated auricular cartilage and were at least an order of magnitude lower than MEDPOR®. n = 3 samples per group for MEDPOR®, n = 6 samples per group for 3D-printed PHT polymer. ∗ denotes statistical significance, which was established at p ≤ 0.05 using a 2-tailed independent sample T-test. (E) MicroCT reconstructions of 3D-printed PHT polymer showed tunable porosity relative to MEDPOR®. Scale bar = 0.5 mm. In silico optimization and ex vivo validation of lattice designs for mechanically resilient, 3D-printed PHT auricular scaffolds To optimize auricular scaffolds for mechanical resilience, 15 representative rectangular designs comprising monophasic, biphasic, and triphasic configurations with uniform or functionally graded lattice structures were evaluated in silico via linear FEA. These simulations modeled compressive forces from daily living activities [41], and selected designs were subsequently validated ex vivo through an unconfined compression test involving severe (80%) uniaxial displacement (Fig. 2A). For FEA simulations, two potential mechanically resilient designs were identified at a mesh size of 0.02 mm (Supplementary Fig. S2), with peak von Mises stress (σVM) that ranged from 1.117 × 10−2 to 3.652 × 10−2 MPa (Supplementary Table S1). Notably, the two designs with the lowest σVM were highly rigid Monophasic 1 (1.117 × 10−2 MPa; contains only rigid lattices) and semi-compliant Triphasic 6 (1.348 × 10−2 MPa; contains a compliant lattice in the middle flanked by rigid lattices) designs, while the design with the highest σVM was observed in Triphasic 9 (3.652 × 10−2 MPa; contains a semi-rigid lattice in the middle flanked by compliant lattices) design (Fig. 2B and C and Supplementary Table S1). Based on this, Monophasic 1 and Triphasic 6 designs were identified as optimal candidates as they achieved lower σVM by 69.4% and 63.1%, respectively, relative to suboptimal Triphasic 9 design (Fig. 2B and C). FEA qualitatively showed that suboptimal Triphasic 9 design exhibited stress concentrations at the corner of lattice structures, whereas both Monophasic 1 and Triphasic 6 designs exhibited a relatively homogenous σVM distribution (Fig. 2B). For mechanical testing, clinically-used MEDPOR® as well as 3D-printed monophasic (Monophasic 1, 2, 3), biphasic (Biphasic 1, 2, 3), and triphasic (Triphasic 1, 2, 4, 6) PHT scaffolds (Fig. 2D) were subjected to 80% uniaxial compression for ex vivo validation (Fig. 2E and Supplementary Movies S2-5). When summarized as an Ashby chart, Triphasic 6 PHT scaffold exhibited the highest proportion of undamaged scaffolds relative to other designs and clinically-used MEDPOR®. In addition, failure mode assessment generally predicted the behavior of scaffold designs with a direct relationship between low σVM-containing scaffolds and high proportions of 3D-printed PHT scaffolds with no visible damage. Notably, failure mode assessment confirmed the mechanical resilience of Triphasic 6 design, with a higher-than-expected proportion of undamaged 3D-printed Triphasic 6 PHT scaffolds (91.7%, p = 0.0012) (Fig. 2F and Supplementary Table S2). Unexpectedly, Monophasic 1 design was deemed suboptimal and reclassified as such, exhibiting a higher-than-expected proportion of damaged 3D-printed Monophasic 1 PHT scaffolds (76.5%, p = 0.0016) (Fig. 2F and Supplementary Table S2). This unexpected result was attributed to the rigid nature of Monophasic 1 design, which, similar to MEDPOR®, typically failed due to brittle fracture. In support of this, analysis of force-displacement graphs (Supplementary Fig. S3) showed that MEDPOR® and 3D-printed Monophasic 1 PHT scaffold exhibited the highest values of peak load and relatively later buckling time point compared to other PHT scaffold designs (Supplementary Fig. S4 and Supplementary Table S3). Although 3D-printed Triphasic 6 PHT scaffolds exhibited a lower peak load, lower average load for initial 50% deformation, and relatively earlier buckling time point compared to 3D-printed Monophasic 1 PHT scaffolds, they displayed a similar energy absorption density (Supplementary Fig. S4 and Supplementary Tables S3–7). This indicated superior stress dissipation capabilities compared to 3D-printed Monophasic 1 PHT scaffolds. Together, these data showed that Triphasic 6 design, which incorporated a compliant functionally graded lattice structure, exhibited compressive resilience with low σVM in silico and a higher-than-expected proportion of physically undamaged scaffolds ex vivo.Fig. 2In silico optimization of 3D-printed PHT auricular scaffold design with ex vivo validation via FEA and mechanical testing. (A) Experiment setup. The schematic illustrates the optimization of auricular scaffold design for mechanical resilience using 15 representative rectangular specimens comprised of uniform or functionally graded lattices via FEA optimization and ex vivo mechanical testing. (B) FEA results showed the distribution of σVM, with optimal auricular scaffold designs identified on the basis of low σVM (Monophasic 1 and Triphasic 6) relative to high stress designs (Triphasic 9). The orange dotted boxes show closeup views of peak σVM locations, indicated by the white arrows. σVM is shown using a colour scale with red and blue indicating regions of high and low effective stress, respectively. (C) Quantification of FEA results for peak σVM across monophasic, biphasic, and triphasic designs. Red arrows indicate the peak σVM value for low (Monophasic 1 and Triphasic 6) and high stress (Triphasic 9) designs. (D) Physical specimens for ex vivo validation were prepared by clinically-used MEDPOR® and 3D-printing of PHT scaffolds. Scale bar = 1 mm. (E) Compression test and Ashby chart analysis. Samples were secured to compressive plates using adhesives and subjected to 80% uniaxial displacement, which was greater than the displacement performed for FEA simulations. The Ashby chart compares peak load with the proportion of undamaged scaffolds, indicating that Triphasic 6 PHT scaffold has the highest undamaged rate compared to other 3D-printed PHT scaffolds and clinical grafts. (F) Failure mode analysis showed that Triphasic 6 PHT scaffolds exhibited a higher-than-expected proportion with no visible damage whereas Monophasic 1 PHT scaffolds exhibited a higher-than-expected proportion with major damage. Representative images showed various physical damage observed for MEDPOR® as well as 3D-printed PHT scaffolds following the 80% uniaxial compression test. Major damage (red) was defined as failed samples exhibiting complete fracture into two or more pieces, minor damage (blue) was defined as failed samples exhibiting partial fracture within a lattice unit, and no visible damage (green) was defined as samples exhibiting a return to their original shape without fracture. n = 7 samples per group for MEDPOR®, n = 8-17 samples per group for monophasic, biphasic, and triphasic PHT scaffolds. ∗ denotes statistical significance, which was established using a Bonferroni corrected p value =0.0083 using a Chi-squared test. In silico optimization of 3D-printed PHT auricular scaffold design with ex vivo validation via FEA and mechanical testing. (A) Experiment setup. The schematic illustrates the optimization of auricular scaffold design for mechanical resilience using 15 representative rectangular specimens comprised of uniform or functionally graded lattices via FEA optimization and ex vivo mechanical testing. (B) FEA results showed the distribution of σVM, with optimal auricular scaffold designs identified on the basis of low σVM (Monophasic 1 and Triphasic 6) relative to high stress designs (Triphasic 9). The orange dotted boxes show closeup views of peak σVM locations, indicated by the white arrows. σVM is shown using a colour scale with red and blue indicating regions of high and low effective stress, respectively. (C) Quantification of FEA results for peak σVM across monophasic, biphasic, and triphasic designs. Red arrows indicate the peak σVM value for low (Monophasic 1 and Triphasic 6) and high stress (Triphasic 9) designs. (D) Physical specimens for ex vivo validation were prepared by clinically-used MEDPOR® and 3D-printing of PHT scaffolds. Scale bar = 1 mm. (E) Compression test and Ashby chart analysis. Samples were secured to compressive plates using adhesives and subjected to 80% uniaxial displacement, which was greater than the displacement performed for FEA simulations. The Ashby chart compares peak load with the proportion of undamaged scaffolds, indicating that Triphasic 6 PHT scaffold has the highest undamaged rate compared to other 3D-printed PHT scaffolds and clinical grafts. (F) Failure mode analysis showed that Triphasic 6 PHT scaffolds exhibited a higher-than-expected proportion with no visible damage whereas Monophasic 1 PHT scaffolds exhibited a higher-than-expected proportion with major damage. Representative images showed various physical damage observed for MEDPOR® as well as 3D-printed PHT scaffolds following the 80% uniaxial compression test. Major damage (red) was defined as failed samples exhibiting complete fracture into two or more pieces, minor damage (blue) was defined as failed samples exhibiting partial fracture within a lattice unit, and no visible damage (green) was defined as samples exhibiting a return to their original shape without fracture. n = 7 samples per group for MEDPOR®, n = 8-17 samples per group for monophasic, biphasic, and triphasic PHT scaffolds. ∗ denotes statistical significance, which was established using a Bonferroni corrected p value =0.0083 using a Chi-squared test. Supplementary data related to this article can be found online at https://doi.org/10.1016/j.bioactmat.2026.05.049 The following are the Supplementary data related to this article:Multimedia component 2Multimedia component 3Multimedia component 4Multimedia component 5 Multimedia component 2 Multimedia component 3 Multimedia component 4 Multimedia component 5 In silico modeling, Euler buckling analysis, and ex vivo mechanical characterization for investigating the mechanism and robustness of 3D-printed PHT auricular scaffolds To understand the mechanism of action responsible for the compressive resilience of 3D-printed PHT scaffolds, Monophasic 1 and Triphasic 6 designs were investigated via non-linear FEA and Euler buckling analysis, while the ability to withstand multiple modes of physiological loading was assessed via manual localized compression, manual bending, drop weight, and cyclic tests (Fig. 3A). Non-linear FEA simulations were performed using a mesh size of 0.25 mm (Supplementary Fig. S5) and showed that Triphasic 6 design underwent global buckling of the entire scaffold and the local buckling within individual lattice struts, similar to ex vivo compressive testing (Fig. 3B). For global buckling, effective stress was initially homogenously distributed across the scaffold and concentrated in the middle compliant regions, as compression progressed. For local buckling, effective stress was redistributed from centrally located struts to peripherally located struts of the compliant region as compression progressed. Together, this redistribution phenomenon decreased the likelihood of stress concentration buildup. Euler buckling analysis showed that compliant regions had an increased propensity to undergo global buckling, with Triphasic 6 design exhibiting a smaller area moment of inertia (Ix = 3.33 × 10−14 m4, Iy = 3.20 × 10−13 m4) and lower Euler's critical load (Px = 0.55 N, Py = 5.32 N) compared to Monophasic 1 design (Ix = 5.83 × 10−14 m4, Iy = 4.49 × 10−13 m4; Px = 0.97 N, Py = 7.44 N), indicating higher susceptibility to buckling along the x-axis for global buckling (Fig. 3Ci). Both designs had identical slenderness ratios (50.57 along the x-axis and 12.64 along the y-axis), reinforcing the likelihood of global buckling along the x-axis. For local buckling, Triphasic 6 design containing a compliant lattice structure exhibited a lower Px relative to Monophasic 1 design, indicating a higher propensity to undergo both global and local buckling (Fig. 3Cii). In manual localized compression and bending tests, 3D-printed PHT auricular scaffolds demonstrated flexibility and robustness by returning to the original shape with little-to-no observable damage (Fig. 3Di and 3 Dii). In drop weight tests, the Triphasic 6 scaffold sustained an impact force 500× of its own weight with little-to-no damage (Fig. 3Diii, Supplementary Fig. S6, and Supplementary Movies S6-7. High-speed camera observations at 10,000 frames per second (fps) showed that, depending on the impact position, the scaffold underwent either buckling or bending motions. Notably, similarly sized auricular scaffolds made of commonly used PLA [7,42] exhibited brittle fracture when subjected to similar impact testing (Supplementary Fig. S7 and Supplementary Movie S8). In cyclic tests, 3D-printed PHT auricular scaffolds sustained at least 20 loading cycles, experiencing 30% compressive uniaxial displacement without failure (Fig. 3Div). Together, these data showed that 3D-printed PHT auricular scaffold's compressive resilience stemmed from a higher propensity to undergo global and local buckling to redistribute stress and minimize stress concentration buildup, and that this scaffold was able to withstand multiple modes of physiological loading including bending, impact, and cyclic regimes.Fig. 3Resilience evaluation of 3D-printed PHT auricular scaffold with functionally graded lattices via FEA, Euler buckling analysis and ex vivo mechanical testing. (A) Experiment setup. The schematic illustrates the mechanistic study of functionally graded lattices via non-linear FEA and Euler buckling analysis, and further mechanical characterization of 3D-printed PHT auricular scaffold via manual localized compression, manual bending, drop weight, and cyclic tests. Representative images show samples buckled and returned to their original dimensions after load removal. (B) Non-linear FEA results indicated that the Triphasic 6 design underwent global buckling of the entire scaffold and local buckling within individual lattice struts similar to ex vivo compressive testing. During global buckling, effective stress was initially evenly distributed across the scaffold and concentrated in the middle compliant regions as compression progressed. During local buckling, effective stress was redistributed from centrally located struts to peripherally located struts in the compliant region as compression progressed, decreasing the likelihood of stress concentration buildup. Effective stress is shown using a colour scale with red and blue indicating regions of high and low effective stress, respectively. Scale bar = 1 mm. (C) Euler buckling analysis showed that Triphasic 6 designs exhibited a smaller Euler's critical load relative to Monophasic 1 designs, with a higher propensity to buckle along the x-axis to undergo global buckling. Also, a compliant lattice design exhibited a higher propensity to undergo local buckling relative to a rigid lattice design, which redistributed stress and minimized stress concentration buildup. (D) Manual localized compression tests demonstrated that functionally graded lattice structures within 3D-printed auricular scaffold underwent controlled buckling and returned to their original shape, indicating the robust compressive resilience. Manual bending tests similarly showed full shape recovery after deformation. Drop weight tests showed that 3D-printed auricular scaffold remained intact with little-to-no damage after experiencing impact from a steel ball that was 500× of its own weight. Cyclic tests showed 3D-printed auricular scaffold withstood 20 loading cycles with 30% uniaxial compressive displacement and sustained approximately 200 N, equivalent to a child's weight. The 1st and 20th cycles are highlighted in green and red, respectively. n = 5 samples per test. Resilience evaluation of 3D-printed PHT auricular scaffold with functionally graded lattices via FEA, Euler buckling analysis and ex vivo mechanical testing. (A) Experiment setup. The schematic illustrates the mechanistic study of functionally graded lattices via non-linear FEA and Euler buckling analysis, and further mechanical characterization of 3D-printed PHT auricular scaffold via manual localized compression, manual bending, drop weight, and cyclic tests. Representative images show samples buckled and returned to their original dimensions after load removal. (B) Non-linear FEA results indicated that the Triphasic 6 design underwent global buckling of the entire scaffold and local buckling within individual lattice struts similar to ex vivo compressive testing. During global buckling, effective stress was initially evenly distributed across the scaffold and concentrated in the middle compliant regions as compression progressed. During local buckling, effective stress was redistributed from centrally located struts to peripherally located struts in the compliant region as compression progressed, decreasing the likelihood of stress concentration buildup. Effective stress is shown using a colour scale with red and blue indicating regions of high and low effective stress, respectively. Scale bar = 1 mm. (C) Euler buckling analysis showed that Triphasic 6 designs exhibited a smaller Euler's critical load relative to Monophasic 1 designs, with a higher propensity to buckle along the x-axis to undergo global buckling. Also, a compliant lattice design exhibited a higher propensity to undergo local buckling relative to a rigid lattice design, which redistributed stress and minimized stress concentration buildup. (D) Manual localized compression tests demonstrated that functionally graded lattice structures within 3D-printed auricular scaffold underwent controlled buckling and returned to their original shape, indicating the robust compressive resilience. Manual bending tests similarly showed full shape recovery after deformation. Drop weight tests showed that 3D-printed auricular scaffold remained intact with little-to-no damage after experiencing impact from a steel ball that was 500× of its own weight. Cyclic tests showed 3D-printed auricular scaffold withstood 20 loading cycles with 30% uniaxial compressive displacement and sustained approximately 200 N, equivalent to a child's weight. The 1st and 20th cycles are highlighted in green and red, respectively. n = 5 samples per test. Supplementary data related to this article can be found online at https://doi.org/10.1016/j.bioactmat.2026.05.049 The following are the Supplementary data related to this article:Multimedia component 6Multimedia component 7Multimedia component 9 Multimedia component 6 Multimedia component 7 Multimedia component 9 Ex vivo physicochemical assessment of fibrin physical integrity and fibrin-mediated bioactive factor delivery To identify a suitable delivery vehicle for the controlled release of pro-elastic cartilage bioactive factors, fibrin hydrogels with varying initial fibrinogen concentrations were assessed for physical integrity via degradation studies, and the release profiles of kartogenin (KGN), transforming growth factor beta 3 (TGF-β3), IGF-1, glycine, and proline were evaluated via UV-Vis spectrophotometry, enzyme-linked immunosorbent assay (ELISA), and fluorometric assays (Fig. 4A). For degradation studies, fibrin hydrogels fabricated from 10 mg/mL or higher fibrinogen concentrations maintained their integrity over 14 days with approximately 64% and 4% integrity in the absence and presence of plasmin-mediated degradation [43], respectively (Fig. 4B and Supplementary Fig. S8). The elastomer component (i.e., PHT polymer) was not investigated given prior characterizations showing long-term stability under accelerated degradation conditions [31]. Based on this, controlled release studies were performed using fibrin hydrogels formulated with 10 mg/mL fibrinogen. For controlled release studies, KGN, TGF-β3, IGF-1, glycine and proline showed cumulative releases of 8.07%, 6.65%, 9.73%, 7.77% and 8.96% within the first 24 h. By Day 14, sustained cumulative releases were achieved at 53.59%, 96.28%, 57.13%, 57.87% and 75.09%, respectively (Fig. 4C). Together, these data demonstrated that fibrin hydrogel formulated using 10 mg/mL fibrinogen or higher maintained adequate physical integrity for controlled, sustained release of pro-elastic cartilage bioactive factors.Fig. 4Ex vivo physicochemical evaluation of PHT-fibrin scaffold degradation and controlled release of elastic cartilage-promoting biochemical factors. (A) Experiment setup. The schematic illustrates the ex vivo physicochemical characterization of PHT-fibrin scaffold integrity and capability for therapeutic delivery via degradation studies and release kinetics assessment. (B) Degradation of fibrin hydrogels fabricated from 2 mg/mL, 5 mg/mL, 10 mg/mL, and 20 mg/mL of fibrinogen under plasmin-mediated degradation showed that fibrin hydrogels with an initial fibrinogen concentration of 10 mg/mL or higher maintained its integrity over a period of about 14 days. Fibrinolysis was monitored via changes in fibrin hydrogel wet weight over time. n = 3 samples per group per time point. (C) KGN, TGF-β3, IGF-1, glycine, and proline showed different controlled release profiles from fibrin component of PHT-fibrin scaffolds over 14 days. n = 5 samples per group per time point. Ex vivo physicochemical evaluation of PHT-fibrin scaffold degradation and controlled release of elastic cartilage-promoting biochemical factors. (A) Experiment setup. The schematic illustrates the ex vivo physicochemical characterization of PHT-fibrin scaffold integrity and capability for therapeutic delivery via degradation studies and release kinetics assessment. (B) Degradation of fibrin hydrogels fabricated from 2 mg/mL, 5 mg/mL, 10 mg/mL, and 20 mg/mL of fibrinogen under plasmin-mediated degradation showed that fibrin hydrogels with an initial fibrinogen concentration of 10 mg/mL or higher maintained its integrity over a period of about 14 days. Fibrinolysis was monitored via changes in fibrin hydrogel wet weight over time. n = 3 samples per group per time point. (C) KGN, TGF-β3, IGF-1, glycine, and proline showed different controlled release profiles from fibrin component of PHT-fibrin scaffolds over 14 days. n = 5 samples per group per time point. In vitro characterization and optimization of 3D-printed PHT auricular scaffolds with functionally graded lattice structures for cytocompatibility and elastic chondrogenic bioactivity To assess the cytocompatibility of 3D-printed PHT auricular scaffolds and define EARc, Live/Dead viability assays (Day 1 and Day 7) and chondrogenic differentiation studies via qPCR and immunofluorescence staining (Day 14 and Day 21) were performed using C2C12 cell spheroids in growth medium, along with six potential elastic chondrogenic media (Fig. 5A and Supplementary Table S8). In these studies, C2C12 cells were used to model MDSCs [44], which we propose as a novel cell source for elastic cartilage regeneration [[45], [46], [47], [48]] while candidate elastic chondrogenic media contained various combinations of pro-elastic cartilage bioactive factors from our controlled release studies. Live/Dead viability assays demonstrated high cytocompatibility, with increased viable cells from Day 1 to Day 7 and little-to-no cell death relative to methanol-treated negative controls (Fig. 5B). Pilot studies showed that fibrin hydrogels remained largely intact after 14-days with C2C12 spheroids and stained positive for glycosaminoglycans in various chondrogenic media (Supplementary Fig. S9). For gene expression studies, C2C12 cell spheroids cultured in Medium E exhibited high levels of elastic cartilage-associated markers, including SRY-Box Transcription Factor 9 (Sox9), collagen type II (Col2a1), and elastin (Eln) (Fig. 5C and Supplementary Tables S9–11). Notably, Medium E increased Sox9 mRNA expression compared to all other groups (p ≤ 0.001), including commercially available chondrogenic medium (CM), which showed higher Col2a1 mRNA expression relative to growth medium (p = 0.017), and exhibited increased Eln mRNA expression compared to both growth medium (p = 0.001) and CM (p = 0.002). Protein semi-quantification using immunofluorescence staining corroborated these findings, showing that C2C12 cell spheroids cultured in Medium E exhibited high levels of SOX9, COL2A1, and ELN (Fig. 5D and Supplementary Tables S12–14). For example, Medium E increased COL2A1 levels relative to growth medium (p < 0.001) and increased ELN levels relative to both growth medium (p < 0.001) and CM (p < 0.001). Together, these data indicated that PHT-fibrin scaffolds exhibited good cytocompatibility, and Medium E promoted elastic cartilage-like differentiation in C2C12 cell spheroids. Thereafter, compositions containing both Medium E and muscle cell-derived spheroids were referred to as EARc.Fig. 5In vitro characterization of cytocompatibility and elastic chondrogenic bioactivity of EARc scaffold. (A) Experiment setup. The schematic illustrates in vitro assessment of 3D-printed PHT-fibrin scaffold cytocompatibility and optimization of EARc via cell viability and elastic chondrogenic differentiation assays. (B) Live/Dead cell assays showed that C2C12 cell spheroids cultured under growth medium conditions within 3D-printed Triphasic 6 PHT-fibrin scaffolds at 1- and 7-day exhibited high viability. n = 6 samples per group from 2 independent experiments. Viable cells are shown in green while dead cells are shown in red. Scale bar = 0.8 mm. (C) Quantitative PCR (qPCR) showed that C2C12 cells cultured in Medium E generally showed high levels of elastic cartilage-associated gene expression for Sox9, Col2a1, and Eln. Gene expression was normalized to GAPDH expression. Growth: Growth medium containing 10% FBS (Negative control); CM: Commercial chondrogenic medium (Positive control); Medium A: Serum-free base chondrogenic medium/SFB; Medium B: SFB + 10 μM KGN; Medium C: SFB + 100 nM KGN; Medium D: SFB + 10 μM KGN + 1.5 mM glycine + 0.6 mM proline; Medium E: SFB + 100 nM KGN + 1.5 mM glycine + 0.6 mM proline. n = 6 samples per group from 2 independent experiments. Error bars indicate standard error of the mean (SEM). ∗ and + denote statistical significance, which was established at p ≤ 0.05 using ANOVA. ∗: Statistical difference between growth and Medium E groups; +: Statistical difference between CM and Medium E group. (D) Immunofluorescence (IF) staining showed that C2C12 cells cultured in Medium E showed high levels of elastic cartilage-associated protein levels for SOX9, COL2A1, and ELN. Scalebar = 50 μm. Protein levels were normalized to Hoechst 33342 staining intensity. Media component and statistical difference are as indicated in panel C legend. n = 6 samples per group from 2 independent experiments. Scale bar = 50 μm. In vitro characterization of cytocompatibility and elastic chondrogenic bioactivity of EARc scaffold. (A) Experiment setup. The schematic illustrates in vitro assessment of 3D-printed PHT-fibrin scaffold cytocompatibility and optimization of EARc via cell viability and elastic chondrogenic differentiation assays. (B) Live/Dead cell assays showed that C2C12 cell spheroids cultured under growth medium conditions within 3D-printed Triphasic 6 PHT-fibrin scaffolds at 1- and 7-day exhibited high viability. n = 6 samples per group from 2 independent experiments. Viable cells are shown in green while dead cells are shown in red. Scale bar = 0.8 mm. (C) Quantitative PCR (qPCR) showed that C2C12 cells cultured in Medium E generally showed high levels of elastic cartilage-associated gene expression for Sox9, Col2a1, and Eln. Gene expression was normalized to GAPDH expression. Growth: Growth medium containing 10% FBS (Negative control); CM: Commercial chondrogenic medium (Positive control); Medium A: Serum-free base chondrogenic medium/SFB; Medium B: SFB + 10 μM KGN; Medium C: SFB + 100 nM KGN; Medium D: SFB + 10 μM KGN + 1.5 mM glycine + 0.6 mM proline; Medium E: SFB + 100 nM KGN + 1.5 mM glycine + 0.6 mM proline. n = 6 samples per group from 2 independent experiments. Error bars indicate standard error of the mean (SEM). ∗ and + denote statistical significance, which was established at p ≤ 0.05 using ANOVA. ∗: Statistical difference between growth and Medium E groups; +: Statistical difference between CM and Medium E group. (D) Immunofluorescence (IF) staining showed that C2C12 cells cultured in Medium E showed high levels of elastic cartilage-associated protein levels for SOX9, COL2A1, and ELN. Scalebar = 50 μm. Protein levels were normalized to Hoechst 33342 staining intensity. Media component and statistical difference are as indicated in panel C legend. n = 6 samples per group from 2 independent experiments. Scale bar = 50 μm. In vivo characterization of 3D-printed, PHT, EARc scaffolds with functionally graded lattice structures Ectopic mouse dorsal subcutaneous implantation model To evaluate the biocompatibility, compressive resilience, and elastic chondrogenic bioactivity of EARc scaffolds, ectopic mouse subcutaneous studies were performed via histological and immunofluorescence staining, biomechanical testing, and microCT assessment at two- and four-week, and 16-weeks post-implantation (Fig. 6A). In these studies, 3D-printed Monophasic 1 PHT scaffold containing C2C12 cell spheroids with growth medium as well as 3D-printed Triphasic 6 PHT scaffold containing EARc were referred to as suboptimal and optimal EARc scaffold, respectively. For biocompatibility assessment, mice showed consistent weight gain and did not exhibit adverse clinical signs or mortality (data not shown), nor detrimental reactions such as necrosis, infection, or granuloma around the implants, as assessed by gross visual inspection. Also, Hematoxylin and Eosin (H&E) staining indicated that scaffolds did not induce an overt inflammatory response. In addition, little-to-no positive nitric oxide synthase (iNOS) and CD68 staining signals were observed in both groups, which further supported H&E findings (Fig. 6B). MicroCT-based morphometric analysis confirmed that EARc scaffolds retained >95% of their original dimensions post-implantation, with force-displacement curves exhibiting the characteristic plateau indicative of a buckling-induced stress dissipation mechanism (Supplementary Fig. S10). For biomechanical testing, optimal EARc scaffolds were more resilient to 80% uniaxial compressive testing compared to suboptimal EARc scaffolds at two-week post-implantation (Fig. 6C). Failure mode assessment revealed that optimal EARc scaffolds sustained less damage compared to suboptimal EARc scaffolds. In optimal EARc scaffold group, 20% showed no visible damage, 60% showed major damage, and 20% showed minor damage, whereas in suboptimal EARc scaffold group, 100% showed major damage. During the initial 50% uniaxial deformation of samples, optimal EARc scaffolds showed both lower values of peak load and average load, indicating that they experienced less compressive forces (p = 0.002 for peak load; p = 0.017 for average load; Fig. 6C and Supplementary Fig. S10). Also, optimal EARc scaffolds showed a relatively earlier buckling time point compared to suboptimal EARc scaffolds while still exhibiting similar energy absorption density (p < 0.001 for buckling time point; p = 0.287 for energy absorption density; Fig. 6C). This indicated that optimal EARc scaffolds displayed an increased propensity to buckle as a means of stress dissipation (Figs. 3 and 6C). Notably, 3D-printed PHT auricular scaffold with functionally graded lattice structures maintained its structural integrity under skin contraction after 16 weeks (Supplementary Fig. S11 and Supplementary Movie S8). For elastic chondrogenic bioactivity, optimal EARc scaffold group exhibited higher levels of elastic cartilage-associated markers compared to suboptimal EARc scaffold group after two- and four-week post-implantation (Fig. 6D and Supplementary Tables S15–19). Specifically, SOX9, COL2A1, and glycosaminoglycans (GAGs) were increased in optimal EARc scaffold group at four-week post-implantation compared to suboptimal EARc scaffold group at two-week post-implantation (p < 0.001, p = 0.001, p = 0.03). Furthermore, microCT analysis revealed that subcutaneous implantation of EARc scaffold in mice did not lead to ectopic mineralization at either two- or four-week post-implantation (Fig. 6E). Similarly, longer 16-week subcutaneous implantation of larger human ear-shaped auricular scaffolds (i.e., 3D-printed PHT auricular scaffolds) also showed no detectable calcium deposits (Supplementary Figs. S11–13). Together, these data showed that EARc scaffolds were biocompatible and that optimal EARc scaffolds exhibited ideal compressive resilience and elastic cartilage-like bioactivity.Fig. 6In vivo characterization of EARc scaffold with biomechanical, biocompatible properties, elastic cartilage-associated bioactivity and undesired mineralization in ectopic mouse subcutaneous model. (A) Experiment setup. The schematic illustrates the experimental setup and representative images of mouse dorsal subcutaneous implantation and EARc scaffolds collected 2-week post-implantation. The blue and purple rectangle indicate the implantation locations of suboptimal EARc scaffold and optimal EARc scaffold, respectively. (B) Representative images showed EARc scaffold biocompatibility. H&E staining indicated no obvious symptoms of severe inflammation and was similar for both groups. Similarly, few positive iNOS and CD68 staining signals were observed in both groups. n = 3 mice per group. Scale bar = 50 μm. (C) Biomechanical testing and analyses after 80% uniaxial compression test revealed the favourable mechanical properties of optimal EARc scaffold. Failure mode assessment showed that the major damage was reduced in optimal EARc scaffold compared to suboptimal EARc scaffold. Optimal EARc scaffold demonstrated lower peak load, average load, and earlier buckling time point while also maintaining similar energy absorption density, showcasing an increased propensity to buckle as an effective means of compressive stress dissipation. n = 5 mice per group per timepoint. ∗ denotes statistical significance, which was established at p ≤ 0.05 using a 2-tailed independent sample T-test. Scale bar = 1 mm. (D) Immunofluorescence and immunohistochemistry staining showed higher levels of SOX9, COL2A1 and GAGs in optimal EARc scaffold group compared to suboptimal EARc scaffold group, respectively. Protein expression was semi-quantified by the positive signal intensity normalized to nuclei intensity (Hoechst 33342 for IF, fast red for Alcian blue, and black for Verhoeff Van Gieson staining). Scale bar = 50 μm. n = 6 mice per group per timepoint. ∗ and + denote statistical significance, which was established at p ≤ 0.05 using ANOVA. ∗: Statistical difference from suboptimal EARc scaffold group at 2-week; +: Statistical difference from suboptimal EARc scaffold group at 4-week post-implantation. Scale bar = 50 μm. (E) MicroCT assessment demonstrated the absence of ectopic calcium deposition and mineralization after 2- and 4-week post-implantation. Dotted lines indicate position of EARc scaffold. n = 3 mice per group per timepoint. Scale bar = 1 mm. In vivo characterization of EARc scaffold with biomechanical, biocompatible properties, elastic cartilage-associated bioactivity and undesired mineralization in ectopic mouse subcutaneous model. (A) Experiment setup. The schematic illustrates the experimental setup and representative images of mouse dorsal subcutaneous implantation and EARc scaffolds collected 2-week post-implantation. The blue and purple rectangle indicate the implantation locations of suboptimal EARc scaffold and optimal EARc scaffold, respectively. (B) Representative images showed EARc scaffold biocompatibility. H&E staining indicated no obvious symptoms of severe inflammation and was similar for both groups. Similarly, few positive iNOS and CD68 staining signals were observed in both groups. n = 3 mice per group. Scale bar = 50 μm. (C) Biomechanical testing and analyses after 80% uniaxial compression test revealed the favourable mechanical properties of optimal EARc scaffold. Failure mode assessment showed that the major damage was reduced in optimal EARc scaffold compared to suboptimal EARc scaffold. Optimal EARc scaffold demonstrated lower peak load, average load, and earlier buckling time point while also maintaining similar energy absorption density, showcasing an increased propensity to buckle as an effective means of compressive stress dissipation. n = 5 mice per group per timepoint. ∗ denotes statistical significance, which was established at p ≤ 0.05 using a 2-tailed independent sample T-test. Scale bar = 1 mm. (D) Immunofluorescence and immunohistochemistry staining showed higher levels of SOX9, COL2A1 and GAGs in optimal EARc scaffold group compared to suboptimal EARc scaffold group, respectively. Protein expression was semi-quantified by the positive signal intensity normalized to nuclei intensity (Hoechst 33342 for IF, fast red for Alcian blue, and black for Verhoeff Van Gieson staining). Scale bar = 50 μm. n = 6 mice per group per timepoint. ∗ and + denote statistical significance, which was established at p ≤ 0.05 using ANOVA. ∗: Statistical difference from suboptimal EARc scaffold group at 2-week; +: Statistical difference from suboptimal EARc scaffold group at 4-week post-implantation. Scale bar = 50 μm. (E) MicroCT assessment demonstrated the absence of ectopic calcium deposition and mineralization after 2- and 4-week post-implantation. Dotted lines indicate position of EARc scaffold. n = 3 mice per group per timepoint. Scale bar = 1 mm. Supplementary data related to this article can be found online at https://doi.org/10.1016/j.bioactmat.2026.05.049 The following are the Supplementary data related to this article:Multimedia component 8 Multimedia component 8 Orthotopic rabbit auricular cartilage defect model To assess the orthotopic regenerative ability of EARc scaffold, rabbit ear injury studies were performed and assessed via cell viability experiments at one-day, two-week and four-week post-implantation, as well as a combination of histology, immunofluorescence staining, and microCT analysis at eight-week post-surgery (Fig. 7A). In these studies, rabbits either underwent sham surgery or received full thickness auricular cartilage defects (8 mm length × 3 mm width) that were left untreated (defect only), treated with 3D-printed Triphasic 6 PHT scaffold only (empty scaffold), or treated with EARc scaffold (3D-printed Triphasic 6 PHT scaffold containing human muscle-derived stem cell/hMDSC spheroids with Medium E treatment). Similar to mouse ectopic studies, rabbits showed consistent weight gain and did not exhibit unrelated adverse clinical signs or mortality (data not shown). H&E and Masson's Trichrome staining showed no obvious fibrous encapsulation or inflammatory infiltration (Supplementary Fig. S14). For assessment of cell viability post-implantation, technical challenges using human nuclear antigen staining were encountered (Supplementary Fig. S15) and fluorescence labeling with DiR dyes was used instead (Fig. 7B). Fluorescently-labeled hMDSC spheroids were detectable at two- and four-week post-implantation with about 42.60% and 13.79% DiR-fluorescence signal intensity compared to Day 1, respectively (Fig. 7B). For elastic chondrogenic bioactivity, EARc scaffold group demonstrated enhanced healing with higher levels of elastic cartilage-associated markers compared to defect only and empty scaffold groups (Fig. 7C). Specifically, tissues adjacent to EARc scaffold exhibited cartilage-like lacunae similar to the uninjured sham surgery group, which were absent in defect only and empty scaffold groups (Fig. 7C). ACAN and ELN were increased in EARc scaffold group compared to the empty scaffold group (p < 0.001, p = 0.04), while ACAN and elastic fibers were increased in EARc scaffold compared to defect only group (p = 0.001, p = 0.019) (Fig. 7C and Supplementary Tables S20–24). For microCT analysis, scaffold implantation did not result in undesired ectopic mineralization in either empty scaffold or EARc scaffold groups (Fig. 7D). Consistent with this, von Kossa and Alizarin Red staining revealed no obvious calcification or tissue structure indicative of ectopic bone at eight-week post implantation (Supplementary Fig. S14). Together, these data demonstrated that EARc scaffolds exhibited long-term viability and that EARc scaffolds orthotopically enhanced elastic cartilage regeneration.Fig. 7In vivo orthotopic characterization of EARc scaffold elastic cartilage-associated bioactivity in a rabbit auricular cartilage defect model. (A) Experiment setup. The schematic illustrates the experimental setup and representative images showing the creation of a rabbit auricular defect and implantation of EARc scaffolds. (B) Cell tracking studies using DiR-labeled hMDSC spheroids showed long-term retention and viability of hMDSC spheroids at 1-Day, 2- and 4-week post-implantation. n = 3 rabbits per group. DiR-fluorescence is shown using a colour scale with red and blue indicating fluorescence signals intensity of high and low of the transplanted cells, respectively. Scale bar = 3 mm. (C) Immunofluorescence and histochemical staining analysis showed enhanced auricular cartilage-like regeneration in EARc scaffold compared to defect only group and empty scaffold group. Specifically, elastic cartilage-associated markers including ACAN and ELN were increased in EARc scaffold group compared to empty scaffold group, while COL2A1, GAGs and elastin fibers were comparable to those in sham surgery group. Protein levels were semi-quantified by immunofluorescence signal intensity and normalized by nuclei signal intensity (Hoechst 33342 for IF, fast red for Alcian blue staining and hematoxylin for Verhoeff Van Gieson staining). Statistical significance was established at p ≤ 0.05; ∗, + and γ denote statistical significance, which was established at p ≤ 0.05 using ANOVA. ∗, significantly different from sham group; γ, significantly different from defect only group; +, significantly different from empty scaffold group. Scale bars as indicated in the images. n = 5 rabbits per group. Scale bar = 50 μm. (D) MicroCT assessment demonstrated that there were no ectopic calcium deposition and mineralization after 8-week post-surgery. Dotted lines indicate position of scaffolds. n = 3 rabbits per group. Scale bar = 1 mm. In vivo orthotopic characterization of EARc scaffold elastic cartilage-associated bioactivity in a rabbit auricular cartilage defect model. (A) Experiment setup. The schematic illustrates the experimental setup and representative images showing the creation of a rabbit auricular defect and implantation of EARc scaffolds. (B) Cell tracking studies using DiR-labeled hMDSC spheroids showed long-term retention and viability of hMDSC spheroids at 1-Day, 2- and 4-week post-implantation. n = 3 rabbits per group. DiR-fluorescence is shown using a colour scale with red and blue indicating fluorescence signals intensity of high and low of the transplanted cells, respectively. Scale bar = 3 mm. (C) Immunofluorescence and histochemical staining analysis showed enhanced auricular cartilage-like regeneration in EARc scaffold compared to defect only group and empty scaffold group. Specifically, elastic cartilage-associated markers including ACAN and ELN were increased in EARc scaffold group compared to empty scaffold group, while COL2A1, GAGs and elastin fibers were comparable to those in sham surgery group. Protein levels were semi-quantified by immunofluorescence signal intensity and normalized by nuclei signal intensity (Hoechst 33342 for IF, fast red for Alcian blue staining and hematoxylin for Verhoeff Van Gieson staining). Statistical significance was established at p ≤ 0.05; ∗, + and γ denote statistical significance, which was established at p ≤ 0.05 using ANOVA. ∗, significantly different from sham group; γ, significantly different from defect only group; +, significantly different from empty scaffold group. Scale bars as indicated in the images. n = 5 rabbits per group. Scale bar = 50 μm. (D) MicroCT assessment demonstrated that there were no ectopic calcium deposition and mineralization after 8-week post-surgery. Dotted lines indicate position of scaffolds. n = 3 rabbits per group. Scale bar = 1 mm. Discussion Auricular deformities represent a significant disability, with a high congenital and stochastic incidence due to microtia [49,50] and physical/medical trauma [51], respectively. Due to facial disfigurement and poorer hearing, such deformities can cause patients, especially children, to experience diminished self-esteem, social withdrawal [52], and learning difficulties [53]. This study has developed a tissue engineered, 3D-printed PHT auricular scaffold to address two longstanding challenges: insufficient mechanical resilience and inadequate elastic cartilage bioactivity. Native tissue-like mechanical resilience is achieved using controlled buckling of functionally graded, compliant lattices while elastic cartilage bioactivity is accomplished using stem/progenitor cells alternatively sourced from muscle together with defined biochemical cues. These approaches are highly unconventional given that buckling is generally considered inherently unstable from a civil engineering standpoint [54] while MDSCs are typically considered to be unipotent, producing skeletal myocytes under normal physiological conditions [55,56]. Together, this unconventional and integrated approach addresses mechanical and biological deficiencies that have constrained previous auricular reconstruction efforts. Prior tissue engineering advances for auricular reconstruction have sought to solve challenges posed by traditional surgical methods. Natural grafts such as autologous costal rib cartilage are the gold standard in clinic but require multiple, lengthy surgeries (lasting over 3 to 5 h each [57,58]) with significant donor site morbidity, tedious manual fashioning of an ear-shaped graft [[1], [2], [3], [4]], and potential complications such as undesired graft resorption. As a result, synthetic grafts such as MEDPOR®, based on non-degradable PE, have been pioneered to address these issues, but their alloplastic origin results in overly stiff and brittle biomaterials with little-to-no bioactivity. Such mechanical mismatch and poor host-biomaterial integration have resulted in undesired extrusion rates of 15% [59], where grafts dislocate from their intended position and become exposed beneath the skin. Recent tissue engineering advances have sought to address such auricular reconstruction challenges with (i) new fabrication techniques such as 3D-printing [9] to personalize patient ear shape, (ii) use of less stiff and/or more readily biodegradable biomaterials [10] for balancing auricle shape fidelity with eventual host-biomaterial integration including resorption, and (iii) application of bioactive factors such as different cell types [9,[15], [16], [17], [18], [19],24] and signaling molecules [[19], [20], [21], [22], [23]] to enhance cartilage bioactivity. While these innovations have addressed clinical needs such as avoidance of donor site morbidity, reduced surgery duration, etc., there are no approaches that allow an auricular graft to withstand severe mechanical deformations and promote elastic cartilage-specific bioactivity. Therefore, next-generation auricular grafts must actively possess stress-dissipating mechanisms and elastic cartilage tissue-specific inductive capabilities. Mechanically, EARc scaffold's ability to undergo controlled buckling strikes a balance between auricle shape fidelity and eventual host-biomaterial integration. As a clinical alternative to costal rib cartilage grafting, alloplastic MEDPOR® seeks to maintain auricle shape over long-term and avoid premature graft resorption due to its high molecular weight PE composition and rigidity. However, MEDPOR® is an excessively rigid biomaterial with a compressive modulus of 176 MPa and a stiffness of 237 N/mm (Fig. 1D) that exceeds native auricular cartilage attributes [[36], [37], [38], [39]] by more than 10-fold. Such a mismatch can impede proper tissue remodeling [60] and may increase the risk of graft failure owing to its brittle nature. Within this context, functional gradation is an effective stress reduction mechanism found in bone [61], bone-tendon interface [62], bamboo [63], and soft robots [64] and is utilized here as an energy absorption mechanism via controlled buckling. Buckling occurs when a structural component is subjected to compressive forces that result in a sudden lateral deflection relative to the axis of loading. In a traditional civil engineering context, buckling is almost always to be avoided due to the instability it entails. However, controlled buckling allows EARc scaffold to be simultaneously flexible for withstanding sudden severe deformations without mechanical failure while exhibiting adequate rigidity and low degradation for maintaining ear shape. EARc scaffold's functionally graded design allows controlled buckling by gradually varying mechanical properties throughout the scaffold. This gradation enables specific (i.e., compliant) regions of the scaffold to absorb energy and dissipate forces by buckling-mediated deformations. This allows the usage of a high molecular weight polymer that exhibits sufficient rigidity and associated slow degradation to maintain long-term structural integrity while providing flexibility to withstand unexpected dynamic impacts to preserve overall shape. Indeed, in silico, ex vivo, and in vivo studies demonstrated that optimized EARc scaffolds reduced stress concentration (Fig. 2, Fig. 3 and Fig. 6), exhibited little-to-no damage after 80% uniaxial compression (Fig. 2), and demonstrated lower peak load and average load, reducing structural damage compared to suboptimal EARc scaffolds after animal implantation (Fig. 6). While infiltration of neo-tissue creates a bio-hybrid construct with distinct mechanical properties compared to pristine scaffolds, the fundamental deformation mode remains consistent. MicroCT-based morphometric analysis confirmed that EARc scaffolds retained >95% of their original lattice geometry after 4 weeks, preserving the structural foundation required for buckling. The persistence of the characteristic plateau in the in vivo force-displacement curves [23] (Supplementary Fig. S10) demonstrated that buckling-induced stress dissipation is a geometry-dependent property that remains functional in a physiological environment. These results validate that the EARc design effectively governs the mechanical response even after biological conditioning. Besides, the 3D-printed PHT auricular scaffold remained intact after withstanding an impact force about 500× of its own weight, withstood 180° bending, and endured 20 loading cycles with 30% displacement deformation (Fig. 3D). This demonstrated EARc scaffold's mechanical resilience. Notably, structural integrity was maintained following 16-weeks of in vivo skin contraction forces (Supplementary Fig. S11 and Supplementary Movie S8). Furthermore, the highly porous nature of functionally graded lattices facilitates loading of regulatory signals (Fig. 4) that support tissue formation and eventual graft-host integration (Fig. 5, Fig. 6, Fig. 7 and Supplementary Fig. S14). Therefore, controlled buckling can effectively dissipate stress, and this concept has potential applications in avoiding failure in other tissue engineering, soft robotics, and structural design applications. Biologically, EARc facilitated elastic cartilage-like regeneration through the synergistic use of an alternative cell source with tissue-specific differentiation protocols. In microtia patients, auricular chondrocytes are compromised and exhibit inferior in vivo cartilage regeneration [65], while the remaining auricular cartilage in the vestigial pinna (if present) is often insufficient for complete auricular reconstruction [66]. This limitation poses a significant challenge, as it has been estimated that high cell concentrations of approximately 25-90 million cells per mL are required for effective tissue engineering [9,25]. Also, current protocols using TGF-βs or BMPs promote general chondrogenesis but often yield hyaline-like cartilage instead of the desired elastin-rich matrix. Our solution leverages MDSCs, which are abundant and easily accessible, comprising about 40% of body mass [67] and obtainable via minimally invasive biopsy to yield approximately 5.3 × 104 MDSCs per gram of muscle tissue [68]. Although MDSCs are typically regarded as unipotent and only capable of generating myocytes under normal physiological conditions, these cells have documented chondrogenic capacity [69,70] and show potential for addressing cell source insufficiency. In addition, plasminogen-depleted fibrinogen was employed to facilitate slow degradation and sustained release of a biochemical cocktail in fibrin hydrogel [31]. The fibrin hydrogels remained largely intact under plasmin-mediated degradation with an initial fibrinogen concentration of 10 mg/mL or higher (Fig. 4B and Supplementary Fig. S8), as well as in various elastic chondrogenic media in vitro (Supplementary Fig. S9). This stability supported localized cell retention for at least 4-weeks in vivo (Fig. 7B). Regarding the fate of transplanted cells, the observed decrease (∼86%) in DiR fluorescence at 4-weeks post-implantation likely reflects signal dilution due to a combination of proliferation-induced signal dilution, natural cell turnover, and metabolic clearance of the lipophilic dye. Although we sought to provide direct histological evidence of hMDSC engraftment using multiple antibodies to stain for human-specific nuclear antigen staining, technical challenges such as non-specific cross-reactivity with the rabbit host tissue were encountered despite extensive optimization of antigen retrieval protocols (Supplementary Fig. S15). While these antibodies successfully labeled hMDSCs in vitro, the high background in vivo precluded definitive quantification of human cell survival. Future studies utilizing constitutive genetic labeling (e.g., GFP) or qPCR for human-specific genes will be necessary to precisely delineate the long-term contribution of transplanted cells to cartilage regeneration. Also, a biochemical cocktail comprised of KGN [71], TGF-β3 [19], IGF-1 [21], dexamethasone [72], ascorbic acid [72], glycine [73,74], and proline [73] was optimized to specifically induce elastin-rich matrix. This cocktail demonstrated strong upregulation of elastic cartilage-associated markers such as COL2A1 and ELN in vitro (Fig. 5) as well as SOX9, ACAN, ELN, COL2A1, GAGs, and elastin fibers in vivo for mouse subcutaneous implantation (Fig. 6) and rabbit auricular defect models (Fig. 7). Notably, high ELN levels (Fig. 5D, 6D and 7C) play a key role in elastic cartilage's mechanical resilience [38,75] and flexibility, and potentially explain its reduced risk for mineralization [[76], [77], [78], [79]]. Furthermore, the pronounced levels of ELN are striking given that elastin production primarily occurs during fetal development, with minimal synthesis observed after birth [80]. In contrast with other studies that require long periods of maturation (up to 17-weeks in vitro culture) [9,81], high levels of ELN were observed after 2-weeks in vitro culture (Fig. 5). Therefore, combining an abundant cell source with pro-elastic chondrogenic induction generates an optimal cocktail that overcomes longstanding biological barriers in auricular reconstruction. While EARc scaffold exhibits promising performance in the auricular defect, several limitations to this study exist. First, although functional gradation allows for compliant lattice designs that withstand 80% deformation, incorporating multi-axial controlled buckling would better protect the auricle against unpredictable loading directions in daily life. Second, the auricle's natural composition of cartilage, muscle, and connective tissues [82] suggests that multi-tissue patterning strategies could further enhance graft-host integration. Third, our histological and microCT data confirm the absence of chronic inflammation, fibrosis, or ectopic calcification in mice up to 4-weeks post-implantation and in rabbits up to 8-weeks post-implantation (Figs. 6D and 7E). Since ectopic calcification can manifest as early as 3 weeks in mice [83] and 4 weeks in rabbits [84], our scaffolds show in vivo evidence of biological safety and efficacy. Nevertheless, the potential for off-target signaling such as SMAD2/3-mediated fibrosis or osteochondral transition, warrants further investigation to fully elucidate the mechanism-of-action. Fourth, while nutrient diffusion was sufficient to support our 1 mm-thick scaffolds a dimension matching native auricular cartilage thickness [36,85,86], scaling to larger 3D volumes remains a challenge due to diffusion limits. As such, future iterations should incorporate vascularization strategies to prevent central necrosis in full-scale grafts [81]. Despite these limitations, EARc scaffold offers an unconventional paradigm that lowers graft fracture risk and promotes elastic cartilage-specific regeneration. Conclusion In summary, we have developed a mechanically resilient and bioactive 3D-printed scaffold for auricular reconstruction using controlled buckling to dissipate mechanical stress and EARc to biologically augment elastic cartilage-like regeneration. Mechanically, this study identified a 3D-printable and functionally graded, compliant lattice design that withstood severe deformation via controlled buckling. Biologically, the use of MDSCs as an abundant, alternative cell source with elastic cartilage-specific biochemical cues enhanced chondroinductive bioactivity. Together, these concepts represent significant advancements in auricular reconstruction and can be broadly applied in tissue engineering and regenerative medicine applications. Materials and methods PHT design and biomaterial naming convention To distinguish between different types of designs, materials, and scaffolds used in this study, a naming convention was established. For scaffold design, simple rectangular-shaped scaffolds with square lattices, as opposed to complex irregularly shaped auricles with tailored lattices, were used to facilitate ease-of-optimization. These designs included uniform/non-graded (monophasic) and non-uniform/functionally graded (biphasic and triphasic) scaffolds. For biomaterials, five different types of PHT biomaterials were used to facilitate various studies. These include: (i) PHT polymer, which refers to mold-casted biomaterial. These polymers were utilized for ex vivo mechanical and ex vivo chemical characterization. The rationale for using PHT polymer was to accurately determine intrinsic material attributes independent of scaffold geometry; (ii) 3D-printed PHT scaffolds and 3D-printed PHT-fibrin scaffolds, which comprised of monophasic, biphasic, and triphasic designs that had been physically fabricated using digital light processing (DLP)-based printing alone or coated with fibrin, respectively. These scaffolds were used for ex vivo mechanical, ex vivo chemical, and in vitro characterization. The rationale for using 3D-printed PHT scaffold or 3D-printed PHT-fibrin scaffold was to accurately determine mechanical, chemical, and biological attributes that were dependent on scaffold geometry; (iii) 3D-printed PHT auricular scaffolds, which refer to Triphasic 6-based auricular scaffold with a human ear shape. These scaffolds feature rigid lattices for the helix and antihelix, and compliant lattices for the concha. They were used to evaluate the buckling efficiency within the complex auricle geometry for mechanical robustness in ex vivo manual localized compression, bending, drop weight test, cyclic compression test, and long-term in vivo study; (iv) EARc scaffold includes suboptimal and optimal, which comprised of 3D-printed Monophasic 1 PHT-fibrin scaffold without elastic cartilage-promoting factors (i.e., growth medium and C2C12 cell spheroids; negative control) and 3D-printed Triphasic 6 PHT-fibrin scaffold with elastic cartilage-promoting factors (i.e., Medium E and C2C12 cell/hMDSC spheroids; experimental group), respectively. These scaffolds were used for in vivo mouse subcutaneous implantation and rabbit auricular defect studies. The rationale for using Monophasic 1 PHT scaffold and Triphasic 6 PHT scaffold was based on in silico optimization and ex vivo mechanical validation; (v) Empty scaffold, which comprised of 3D-printed Triphasic 6 PHT scaffold without elastic cartilage-promoting factors (i.e., no EARc). Empty scaffold and EARc scaffold (i.e., containing Medium E and hMDSC spheroids) were used for in vivo rabbit auricular defect studies. The rationale for using an empty scaffold as opposed to suboptimal EARc scaffold was to allow for the bioactivity of EARc to be investigated. Fabrication of PHT polymer and 3D-printed PHT scaffold PHT polymer and PHT scaffold were fabricated using mold-based visible light fabrication and DLP-based 3D-printing based on a two-step synthesis reaction as previously described, with minor modifications [31] (Supplementary Fig. S1). For mold fabrication, PHT pre-polymer was poured into a custom polypropylene mold and subjected to visible light photocrosslinking at 405 nm using the same light projector as the 3D-printer (Optoma, Taiwan, model no. VDHDNT). For 3D-printing, a DLP-based 3D-printer (Kudo3D, Taiwan, model no. Titan 2 HR) was used. Prior to 3D-printing, computer-aided design (CAD) 3D-models were digitally prepared and processed using Creation Workshop software (Envision Labs, USA, version 1.0.0.75). This included digital slicing of 3D-CAD models with a thickness layer of 0.1 mm, followed by uploading into a 3D-printer. Thereafter, 3D-printing was performed according to the manufacturer's instructions. Following fabrication, PHT polymer or 3D-printed PHT scaffold samples were subjected to post-processing via heat, washing, and sterilization treatments. Heat curing was performed in a vacuum oven (Shel Lab, USA, model no. SVAC1-2 Compact Vacuum Drying Oven) at 150 °C for 0.5 h. Sample washing was performed once in 5 N sodium hydroxide (Sigma-Aldrich, USA, catalog no. 795429) for 1 h, followed by three washes in 1 × PBS. Sterilization of in vitro and in vivo samples was performed under UV light for 30 min in a cell culture hood. Ex vivo FTIR characterization Synthesis of PHT polymer and PHT-fibrin scaffold were monitored using a FTIR spectrometer (Thermo Fisher Scientific, USA, model no. Nicolet iS10) as previously described [31]. For the synthesis of PHT polymer, spectra were acquired for individual components: P, H, T, and PHT polymer. Specifically, 200 μL drops of P, H, T, and PH prepolymer were placed between two sodium chloride plates and then positioned into the plate holder. PH prepolymer sample was obtained in the first step reaction, while a thin film of PHT polymer was obtained by UV irradiation. For PHT polymer, film samples (less than 50 μm thick) were prepared by casting the pre-crosslinked solution between two clamped glass slides, followed by UV irradiation using a 12.5 W lamp (Shenzhen He Sheng Bang Technology Co. Ltd, China) for 2 s. Fibrin film was prepared by mixing 4 U/mL thrombin (Enzyme Research Laboratories, USA, catalog no. HT 1002a) with 10 mg/mL of fibrinogen (Enzyme Research Laboratories, USA, catalog no. FIB 1) and sandwiching the freshly prepared mixture between two glass slides at 37 °C for 30 min. For PHT-fibrin scaffold, film samples were prepared by casting fibrin-thrombin mixtures onto the top surface of a photocured PHT polymer film. Subsequently, fibrin and PHT-fibrin scaffold films were lyophilized using a freeze dryer (Labconco, USA, model no. FreeZone 2.5 L Benchtop Freeze Dry System). Transmission spectra were acquired at ambient temperature, using eight scans per spectrum over the range of 4000 – 400 cm−1 with a resolution of 8 cm−1. Ex vivo mechanical characterization Uniaxial compression tests of ex vivo samples were performed using an ADMET's MTESTQuattro material testing system (Admet, USA) equipped with a 10 lb load cell in accordance with American Society for Testing Materials standard D695-10 for assessing the compressive properties of rigid plastics and as previously described [[36], [37], [38], [39]]. Protective adhesive tapes (Shenzhen Benyida Technology Co., Ltd, China, adhesive roll size: 1.5 cm × 50 m) were placed on compression plates, and samples were secured to the tape using an epoxy glue (Shenzhen Sinwe electronic material Co., Ltd, China, catalog no. Sinwe 605) to prevent sample flipping during testing. Compression testing was conducted at a constant displacement rate of 10 mm per min [39,87], which was equivalent to a compressive displacement of 80% based on sample length. For intrinsic material attributes independent of scaffold geometry, PHT polymer and MEDPOR® (Stryker, USA, catalog no. SYK8312) were fabricated as dog bone-shaped specimens with an overall width of 9.5 mm and a length of 39.7 mm, with the narrow region measuring 1.5 mm (thickness) × 6.35 mm (width) × 19.05 mm (length), similar to ASTM D695-10. For mechanical attributes dependent on scaffold geometry (i.e., ex vivo testing), ten representative 3D-printed scaffolds, including (i) Monophasic 1, Monophasic 2 and Monophasic 3 PHT scaffolds, (ii) Biphasic 1, Biphasic 2 and Biphasic 3 PHT scaffolds, (iii) Triphasic 1, Triphasic 2, Triphasic 4 and Triphasic 6 PHT scaffolds, and MEDPOR® were fabricated as rectangular-shaped specimens in accordance with FEA studies. During ex vivo analysis of PHT polymer, the compressive modulus and stiffness of materials were calculated within the 1% linear region of the slope of the stress-strain curve and the load-displacement curve, respectively. During ex vivo analysis of 3D-printed PHT scaffolds, samples were quantified for peak load, average load for the initial 50% deformation, buckling time point, and energy absorption density, as well as classified into three categories: major damage, minor damage, and no visible damage. Major damage was defined as a fracture that resulted in a complete dislocation of the sample into two or more pieces, minor damage was defined as a partial fracture that resulted in an incomplete dislocation of the sample, and no visible damage was defined as a sample that returned to its original shape and form without any observable fracture. Failure modes were recorded using a 12-megapixel phone camera and assessed via a Chi-squared test. Peak load was defined as the maximum force observed in the force-displacement plots within 10% deformation. The average load during the initial 50% deformation was derived from the output data. The buckling time point was defined as the time when a characteristic load drop occurred. The load drop exceeded 10% of the peak load and occurred between the linear elastic region and a plateau, with no observable fracture or damage to the scaffold [88]. Energy absorption density was calculated as the total area under the force-displacement curve normalized by the scaffold volume. Ex vivo microCT imaging The porosity of 3D-printed PHT scaffolds was assessed via microCT scans (Scanco Holding AG, Switzerland, model no. Scanco Medical μCT 35) as similarly described [31]. The X-ray source voltage and current were set to 45 kV and 117 μA, respectively, with the detector camera configured to acquire 500 projections over 180° of rotation, yielding a voxel resolution of 12 μm. The acquired DICOM data were reconstructed using the open-source software 3D Slicer (https://www.slicer.org/, USA, version 5.6.1). Reconstructed images were visually inspected to facilitate proper segmentation of the scaffold and scaffold pores. This included applying a suitable threshold range of grayscale intensity values for 3D-printed PHT scaffolds and MEDPOR®, which were 80-2000 and -650-2000, respectively. For porosity measurements, the total scaffold porosity was calculated using Equation (1).(Equation 1)Totalporosity(%)=1‐ScaffoldvolumeTotalROIvolume×100% In silico scaffold optimization The mechanical resilience of scaffold designs was computationally optimized using linear static FEA in SolidWorks (Dassault Systemes, USA, version 2022 SP1.0) as similarly described [89]. As previously stated in ‘PHT design and biomaterial naming convention’, 3D-printed PHT scaffolds were comprised of simple, rectangular-shaped scaffolds with square lattices to facilitate ease-of-optimization. These designs were modeled as uniform/non-graded (monophasic) and non-uniform/functionally graded (biphasic and triphasic) scaffolds. Square lattice sizes of 0.3 mm, 0.5 mm, and 0.7 mm were used to endow designs with low, medium, and high compliance, respectively. As a result, the dimensions of monophasic and triphasic scaffolds were 7.3 mm (length) × 2.5 mm (width) × 1 mm (thickness), while the biphasic scaffolds were 6.8 mm (length) × 2.5 mm (width) × 1 mm (thickness). A mesh size of 0.02 mm with 6,498,448 elements was chosen as the model achieved convergence upon mesh refinement (Supplementary Fig. S2). For the material attributes, a linear elastic isotropic material with a Poisson's ratio of 0.3 [[90], [91], [92]] and an elastic modulus of 22.24 MPa (Fig. 1D) was used. For the boundary conditions, the model was affixed on one face, which was denoted by its width and thickness, with zero displacement constraint (x = 0, y = 0 and z = 0) to simulate the attachment of the auricle to the skull while a compressive stress mimicking daily activities was applied on the opposing end, which was also denoted by its width and thickness. The stress was applied along the x-axis, with a magnitude of 1.06 kPa to mimic a phone pressed against the auricle [41,93]. This magnitude was determined based on prior reports indicating that the mobile phone exerts a force of 5 N [41] on an auricle with an estimated surface area of 47.3 cm2 [93]. Linear FEA simulation results were reported as σVM and optimal scaffold designs were identified based on the lowest stress values. Ex vivo mechanical characterization Manual localized compression test Manual localized compression tests were conducted to simulate common daily physical activities, such as ear twisting during rough play or the ear being caught on clothing during dressing, and to assess the mechanical response of 3D-printed PHT auricular scaffold with functionally graded lattice structures, as previously described [9]. Samples were manually compressed with forceps until visible deformation occurred, and the process was documented using a 48-megapixel smartphone camera. Manual bending test Manual bending tests were performed to simulate common daily physical activities, such as side-sleeping and contact sports, and to assess the bending performance of 3D-printed PHT auricular scaffold with functionally graded lattice structures, as previously described [94]. Samples were manually bent with forceps to 180°, and the process was documented using a 48-megapixel smartphone camera. Drop weight test Drop impact tests were performed to simulate high-impact scenarios, such as a child falling or accidently striking his/her ear, and to assess the durability of 3D-printed PHT auricular scaffold with functionally graded lattice structures. 3D-printed PLA auricular scaffolds, fabricated via fused deposition modeling (Bambu Lab X1 Carbon, China), served as the control group [42]. Samples were assessed using a custom-made setup and recorded via a high-speed camera (TMX 7510, AMETEK, USA) at a resolution of 1280 × 800 pixels and a sampling rate of 10,000 fps (Supplementary Fig. S6), as previously described [95]. Briefly, a 100 g spherical steel impactor (29 mm in diameter) was released from a height of 12 cm. This impactor mass represented approximately 500× and 600× the mass of the PHT and PLA scaffolds, respectively. Each sample featured a complex auricle shape, measuring 11.0 mm width × 17.7 mm length × 3.4 mm thickness. A 500:1 to 600:1 mass ratio was selected to approximate the effective inertial mass relevant to localized auricular trauma [96] based on clinical studies for human head mass of 2.9-5.7 kg [97] and auricle mass of 5.4-7.5 g [39,93]. Following drop weight tests, samples were inspected for visible damage. Cyclic compression test Cyclic compression testing was conducted to simulate repetitive loading that may occur in a dynamic implantation environment, such as wearing a tight headgear (e.g., helmet during cycling), and to assess the performance of 3D-printed PHT auricular scaffold with functionally graded lattice structures, as previously described [98]. In brief, samples were evaluated using a mechanical tester (Admet, USA) equipped with a 500 lb load cell in accordance with American Society for Testing Materials standard D7791-12 for assessing the compressive properties of rigid plastics, as previously described [31]. Samples were fabricated as human ear-shaped specimens with an overall width of 10 mm (width) × 23 mm (length) × 1 mm (thickness). Compression testing was conducted at a constant displacement rate of 10 mm per min [39,98,99] with a compressive displacement of 30% based on sample length. The ultimate load was defined as the maximum load recorded within 20 cycles. In silico buckling analysis The global and local buckling behaviors of 3D-printed PHT scaffolds were analyzed using nonlinear FEA (University of Utah, USA, FEBio software version 1.7.1) [100]. Nonlinear FEA facilitated accurate modeling of systems involving large deformations, including buckling [101]. For global buckling, Triphasic 6 design, identified as optimal based on linear FE analysis and mechanical validation, was selected for comparison with Monophasic 1 design control. For local buckling, analysis specifically focused on the compliant region of Triphasic 6 design, which contained nine 0.7 mm square lattices. A mesh of 0.25 mm with 8123 elements was chosen as the model achieved convergence upon mesh refinement (Supplementary Fig. S5). For scaffold material, a Neo-Hookean material with a Poisson's ratio of 0.3 [[90], [91], [92]] and an elastic modulus of 22.24 MPa (Fig. 1D) was used. For the boundary condition, two rigid bodies were used to fix the scaffold's two ends to simulate the 3D-printed PHT scaffold encountered during the uniaxial compression testing. A prescribed displacement was along the x-axis and was 41% scaffold length (1.5 mm for global buckling, 0.5 mm for local buckling) from two end surfaces to simulate compressive plate ends. Symmetry constraints at y = 0 and z = 0 to prevent lateral movement were applied. Global buckling was identified by a sudden loss of overall structural stability. Local buckling was defined as excessive deformation of individual strut units, and this was based on experimental observations. Nonlinear FEA simulation results were reported as effective stress. Euler buckling analysis Euler buckling analysis was applied to investigate the buckling propensity of the Monophasic 1 and Triphasic 6 designs. The Euler equation has the form [102]:(Equation 2)Pcr=ℼ2ELe2Iwhere Pcr is the compressive load, E is Young's modulus, I is the moment of inertia obtained from Equation (3), and Le is the effective length obtained from Equation (4).(Equation 3)I=bh312(Equation 4)Le=kLwhere b is the width of the column, h is the height of the column, and k is dependent on the geometry of the ends (k = 0.5 for fixed ends). The slenderness ratio, S, was obtained by dividing the effective column length by its radius of gyration from Equation (5):(Equation 5)S=Le12h The propensity to buckle for the Monophasic 1 design and Triphasic 6 design was determined by calculating the moment of inertia, Euler's critical load, and the slenderness ratio for each design. Ex vivo physicochemical assessment of fibrin physical integrity and fibrin-mediated bioactive factor delivery Degradation and bioactive factor release studies were performed on fibrin hydrogel samples to evaluate physical integrity under physiological conditions and capacity for the controlled delivery of bioactive factors, as similarly described [31,43]. For degradation studies, the physical integrity of fibrin hydrogels was assessed ex vivo by measuring the wet weight of hydrogels over 14 days. Fibrin hydrogels, approximately 0.1 cm3 in volume, were prepared by mixing varying concentrations of fibrinogen (2, 5, 10 and 20 mg/mL) with 4 U/mL thrombin and PBS, followed by incubation at 37 °C for 1 h. Subsequently, fibrin hydrogels were immersed in 3 mL PBS containing 2.5 nM plasmin (Enzyme Research Laboratories, USA, catalog no. HPlasmin) to mimic the physiological concentration of plasmin in blood [43] and maintained at 37 °C, 5% CO2 with daily medium change for 14 days (Supplementary Fig. S8). Fibrin hydrogel physical integrity was quantified by measuring the wet weight of hydrogels at days 1, 3, 5, 7, 10, and 14 and normalized to the initial weight at day 0. The degradation (% of wet mass remaining) was calculated from Equation (6):(Equation 6)Degradation(%ofWetmassremaining)=WetweightatendpointWetweightonday0×100% For bioactive factor release studies, EARc components, including KGN (Sigma-Aldrich, USA, catalog no. SML0370), TGF-β3 (PeproTech, Hong Kong, catalog no. 100-36E/50μg), IGF-1 (ProSpec-Tany TechnoGene, Israel, catalog no. Cyt-216), glycine (Sigma-Aldrich, USA, catalog no. G8790), and proline (Sigma-Aldrich, USA, catalog no. P5607) were assessed over 14 days ex vivo using either UV−vis spectrophotometry or ELISA kits according to the manufacturer's instructions. Fibrin hydrogels were prepared by mixing 10 mg/mL fibrinogen, 4 U/mL thrombin, and PBS with one of the following: 1 mM KGN, 100 ng/mL TGF-β3, 100 ng/mL IGF-1, 30 mM glycine, or 0.6 mM proline, followed by incubation at 37 °C for 1 h. Similar to previous studies [103], the concentrations of KGN, TGF-β3, and glycine were increased to facilitate ease-of-detection. Subsequently, fibrin hydrogels were immersed in 3 mL PBS and maintained at 37 °C, 5% CO2 for 14 days. Samples (0.05 mL medium supernatant) were collected with PBS replenishment at the following time points (Day 1, 3, 5, 7, 10, and 14) and stored at −80 °C immediately. The released amounts of KGN were quantified by UV−vis spectrophotometry (Thermo Fisher Scientific, USA, catalog no. NanoDropTM ND-2000C) following the construction of a standard curve with known KGN quantities at 279 nm [104]. The released amounts of TGF-β3 and IGF-1 were determined using colorimetric ELISA kits (Bio-Techne, USA, catalog no. DY243 and DG100B) on a microplate reader (Thermo Fisher Scientific, USA, model no. Multiskan GO) according to the manufacturer's instructions. The released amounts of glycine and proline were determined by assay kits (Glycine assay kit: Abcam, UK, catalog no. ab211100 and ab303731) using a fluorescence microplate reader (Thermo Fisher Scientific, USA, model no. Varioskan™ LUX Multimode Microplate Reader) according to the manufacturers' instructions. The concentrations, cumulative mass, and cumulative percentage of these released factors at different time points were calculated. In vitro 2D cell culture Two different types of cells were used to model stem/progenitor cells for in vitro and in vivo studies. Multipotent mouse C2C12 myoblasts (ATCC®, USA, catalog no. CRL-1772™) were cultured in Dulbecco's modified Eagle's medium (DMEM; Thermo Fisher Scientific, USA, catalog no. 11995073) with 10% fetal bovine serum (FBS; Thermo Fisher Scientific, USA, catalog no. 26140079) and 1% penicillin-streptomycin (PS; Thermo Fisher Scientific, USA, catalog no.15140122). This medium is referred to as C2C12 cell growth medium [31]. Two hMDSCs from 14-year-old male and 13-year-old female (Cook® MyoSite, Inc., USA, catalog no. P01059-14M and P10530-13F) were cultured in DMEM supplemented with 10% heat-inactivated FBS (heat-inactivated for 30 min), 10% heat-inactivated horse serum (Thermo Fisher Scientific, USA, catalog no. 16050122; heat-inactivated for 30 min), 0.5% chicken embryo extract (Life Science Group Ltd, UK, catalog no. MDL-004E-UK), and 1% PS. This medium is referred to as hMDSC growth medium [105,106]. All cells were kept at 37 °C, 5% CO2 in a humidified incubator. No mycoplasma contamination was found in cell cultures, which was monitored using a mycoplasma detection kit (Nanjing Vazyme Biotech Co., Ltd., China, MycoBlue, catalog no. D101-02) and Hoechst 33342 (Thermo Fisher Scientific, USA, catalog no. PI62249) staining. In vitro 3D cell spheroid culture C2C12 cells or hMDSCs were cultured as 3D spheroids using a commercially available microwell platform (AggreWell™ 400 microwell plates; STEMCELL Technologies Inc., Canada, catalog no. 34460) according to the manufacturer's instructions. The microwell platform was coated with an anti-adherence rinsing solution (STEMCELL Technologies Inc., Canada, catalog no. 07010) to avoid cell attachment, centrifuged to ensure a homogeneous coating, and washed with medium prior to cell seeding. Cells were seeded at a density of 1.2 × 106 cells per well (200 cells/microwell) in candidate chondrogenic medium (Supplementary Table S8) and kept at 37 °C, 5% CO2 in a humidified incubator. After 16 h, cell spheroids were formed and used for in vitro and in vivo studies. In vitro cytocompatibility of PHT-fibrin scaffolds Cell viability was evaluated using Live/Dead staining kit (Thermo Fisher Scientific, USA, catalog no. L3224) as previously described [31]. C2C12 cell spheroids were seeded on Triphasic 6 PHT-fibrin scaffolds and cultured in C2C12 cell growth medium, with medium changes every 48 h. Cytocompatibility was assessed on days 1 and 7 post-seeding. Unstained samples served as autofluorescence control, while negative controls were prepared by treating samples with ice-cold methanol prior to staining. Fluorescence images were acquired using an inverted confocal microscope (Leica microsystems, Germany, model no. Leica TCS SP8) equipped with a pulse laser source (WLL) and an acousto-optical tunable filter. Data acquisition was performed using sequential scan via photomultiplier tube with detection ranges set to 488-525 nm for viable fluorescence signals (i.e., live cells), and a hybrid detector with detection ranges set to 580-630 nm for non-viable fluorescence signals (i.e., dead cells). In vitro elastic chondrogenesis induction The elastic chondrogenesis induction effect of candidate elastic chondrogenic media in C2C12 cell spheroids was assessed by measuring expression and levels of elastic cartilage-associated genes at 21 days via quantitative RT-PCR and proteins at 14 days via semi-quantitative immunofluorescence staining, respectively. C2C12 cell spheroids were seeded on 3D-printed PHT-fibrin scaffolds and cultured with different candidate media for inducing elastic chondrogenesis (Supplementary Table S8). Optimal EARc composition was selected based on assessment of elastic cartilage-associated gene (Sox9, Col2a1 and Eln) expression and protein (SOX9, COL2A1 and ELN) levels, which are described in the following. In vitro assessment of elastic cartilage-associated gene expression Expression levels of elastic cartilage-associated genes under various candidate elastic chondrogenic media conditions were assessed using TaqMan®-based PCR on the samples harvested after 21 days of culture, as similar described [107]. Total RNA was extracted using RNeasy Mini Kit (Qiagen, USA, catalog no. 74134) and converted into cDNA using the Lunascript RT Supermix Kit (New England Biolabs, USA, catalog no. E3010). Real-time qPCR was performed using a PCR mix containing TaqMan® Universal PCR Master Mix and Assays-on-Demand Gene expression probe (Thermo Fisher Scientific, USA, catalog no. 4331182) for Sox9 (assay ID. Mm00448840_m1), Col2a1 (assay ID. Mm01309565_m1), and Eln (assay ID. Mm00514670_m1). After 40 rounds of amplification, gene expression levels were estimated using the comparative threshold cycle method (2−ΔΔCt) and normalized to that of glyceraldehyde-3-phosphate dehydrogenase (Gapdh; Thermo Fisher Scientific, USA, assay ID. Mm99999915_g1, catalog no. 4354370) as an endogenous reference gene. Quantitative gene expression levels were expressed relative to serum-containing growth medium, which served as a negative control. In vitro assessment of elastic cartilage-associated protein levels Protein levels of elastic cartilage-associated markers under various candidate elastic chondrogenic media conditions were assessed by immunofluorescence staining of SOX9, COL2A1, and ELN on samples harvested after 14 days of culture. Freshly harvested samples were fixed in 4% paraformaldehyde. For paraffin embedding, samples were dehydrated through a graded ethanol series, followed by three washes in xylene, and infiltrated with paraffin at 60 °C under vacuum. Paraffin-embedded samples were sectioned at 5 μm using a microtome (Leica Biosystems, Germany, model no. Leica RM2235), then deparaffinized, rehydrated, and processed for immunohistochemical staining. SOX9 immunofluorescence staining was performed by incubating sections in citrate buffer (pH 6.0; Abcam, UK, catalog no. ab64214) at 98 °C for 20 min [108,109]. COL2A1 immunofluorescence staining was performed by pre-digesting sections in 0.1% pepsin (Worthington Biochemical Corporation, USA, catalog no. LS003317) in 0.01 N hydrochloric acid (pH 2.0; Duksan, South Korea, product no. 4606) at room temperature for 10 min [110,111]. ELN immunofluorescence staining was performed by pre-digesting sections with 10 μg/mL proteinase K (Sigma-Aldrich, USA, catalog no. 3115801001) in Tris and EDTA buffer (pH 9.0; Abcam, UK, catalog no. ab93684) at room temperature for 10 min [112]. Additional permeabilization with 0.1% Triton™ X-100 (Sigma-Aldrich, USA, catalog no. X100-500 mL) in PBS was performed for SOX9 staining, followed by three washes with PBS. All sections were blocked with 10% donkey serum (Sigma-Aldrich, USA, catalog no. D9663-10 mL) in PBS for 30 min before incubation with primary antibody: rabbit monoclonal anti-SOX9 (5 μg/mL; Abcam, UK, catalog no. ab185966), rabbit polyclonal anti-Collagen II (5 μg/mL; Abcam, UK, catalog no. ab34712), rabbit polyclonal anti-Elastin (5 μg/mL; Abcam, UK, catalog no. ab21610) at 4 °C for 16 h. Antibodies were diluted in a wash buffer, which was comprised of 0.3% donkey serum in PBS. After three washes with PBS, sections were incubated with Alexa Fluor 647-labeled donkey anti-rabbit antibody (10 μg/mL; Thermo Fisher Scientific, USA, catalog no. A31571) for 1 h in the dark at room temperature. Nuclei counterstaining was performed by rinsing and staining sections with 0.3 μg/mL Hoechst 33342. Images were acquired with an inverted fluorescence microscope (Nikon, Japan, model no. Ti2-E) equipped with a fluorescence light source (Lumencor, USA, model no. SPECTRA III Light Engine), Hoechst- and Cy5-equivalent fluorescent filters, and a DP80 Double-CCD monochrome/colour camera. Semi-quantitative analyses were performed using Adobe Photoshop (Adobe Inc., USA, version 22.0.0) by measuring the average pixel intensity via the histogram tool. Elastic cartilage-associated markers were semi-quantified by the fluorescence signal intensity of SOX9, COL2A1, and ELN normalized to the nuclei fluorescence signal intensity of Hoechst 33342. Fabrication of empty and EARc scaffold EARc scaffolds were fabricated by embedding C2C12 cell- or hMDSC spheroids into 3D-printed PHT-fibrin scaffolds, with or without elastic cartilage-promoting biochemical factors identified from our in vitro study. As previously stated in ‘PHT design and biomaterial naming convention’, suboptimal EARc scaffolds referred to 3D-printed Monophasic 1 PHT scaffold containing C2C12 cell or hMDSC spheroids with growth medium, while optimal EARc scaffolds referred to 3D-printed Triphasic 6 PHT scaffold containing C2C12 cell or hMDSC spheroids and Medium E (i.e., optimal elastic cartilage-promoting biochemical factors/biomolecules). Empty scaffolds comprised of 3D-printed Triphasic 6 PHT scaffold. Fibrin hydrogels, approximately 0.04 cm3 in volume, were prepared by mixing 10 mg/mL fibrinogen with 4 U/mL thrombin and corresponding medium, which contained 6,000 C2C12 cell or hMDSC spheroids. These fibrin hydrogels were immediately cast into the 3D-printed PHT scaffolds. The assembly was incubated at 37 °C for 1 h before adding the corresponding medium. After 24 h, EARc scaffolds were formed and used for in vivo studies. In vivo evaluation using mouse dorsal subcutaneous implantation model Mouse subcutaneous implantation was performed as previously described [31]. This study was performed according to the guidelines established by the University's Animal Experimentation Ethics Committee (approved protocol: 21-199-MIS). Wild-type C57BL/6 mice (both male and female) of eight weeks old with a body weight of 20-25 g were used and maintained in a 12:12 h light-dark cycle with free access to standard laboratory food as well as water, and housed for one week before surgery. In brief, mice were anesthetized by inhalation with isoflurane (4% for induction, 2% for maintenance; Piramal critical care Inc., UK, catalog no. NDC 0044-5260-05) and oxygen (0.6 L/min) followed by subcutaneous injection of 0.05 mg/kg buprenorphine (VetViva Richter GmbH, Austria, catalog no. Bupaq) and 25 mg/kg cephalexin (Santa Cruz Biotechnology, USA, catalog no. sc-487556) as an analgesic and anti-infective, respectively. The surgical area was shaved and disinfected with alternating applications of betadine and 70% ethanol in circular motions. Using aseptic technique, a 2 cm incision was made on the dorsal skin with a surgical scalpel. For the long-term evaluation of auricular graft stability and calcification risk, a human ear-shaped scaffold measuring 15 × 26 × 5 mm [113], was designed with optimized functionally graded lattice structures and was subcutaneously transplanted into the back of mice. For the biomechanical, biocompatible properties, elastic cartilage-associated bioactivity and mineralization of EARc scaffold, two subcutaneous pockets were created on the left and right sides. Suboptimal EARc scaffolds (3D-printed Monophasic 1 PHT-fibrin scaffold containing C2C12 cell spheroids with growth medium) were implanted on one side/flank of the body while optimal EARc scaffold (3D-printed Triphasic 6 PHT-fibrin scaffold containing EARc, which comprised of C2C12 cell spheroids and Medium E) on the other side/flank. 5-0 Silk suture was used to avoid migration and anchor the biomaterial to the initial implantation site. 4-0 PGA was used to close the wound. After surgery, mice were allowed to recover on a heating pad and were allowed free-cage activity. 0.05 mg/kg buprenorphine was administered subcutaneously twice a day for three days for post-operative analgesia. At two- and four-week post-implantation, mice were euthanized via carbon dioxide overdose, and samples were collected for biomechanical assessment, biocompatibility, and histological analyses. For quantification of EARc scaffold morphology, EARc scaffolds harvested at 2 and 4 weeks, were carefully debrided of peripheral connective tissue and analyzed via microCT to assess macroscopic dimensions and internal lattice integrity. The volume and morphology of the internal lattice pores were assessed to identify any signs of structural collapse or significant degradation. Dimensional retention was calculated relative to pre-implantation values, and only samples maintaining >95% of their initial height and architecture were used for subsequent biomechanical testing. For biomechanical assessment of EARc scaffolds, uniaxial compression testing was performed on freshly harvested suboptimal and optimal EARc scaffolds to quantify peak load, average load for initial 50% deformation, buckling time point, and energy absorption density, as previously described in ‘Ex vivo mechanical characterization’. Samples were also similarly classified into three categories as previously described for failure mode analysis. Mechanical testing of samples was recorded using a 12-megapixel phone camera. For biocompatibility assessment of EARc scaffolds, assessment was performed on paraffin-embedded and sectioned mouse subcutaneous implantation samples, as previously described [31]. Briefly, freshly harvested samples were fixed in 4% paraformaldehyde, processed for paraffin embedding and tissue sectioning, followed by H&E, CD68 and iNOS staining. H&E staining was performed on deparaffinized and rehydrated sections according to the manufacturer's instructions (Electron Microscopy Science, USA, catalog no. 26042 and 14851). To facilitate immunocytochemical staining of inflammation-associated markers, CD68 and iNOS, sections were deparaffinized, rehydrated, and incubated in citrate buffer (pH 6.0) at 98 °C for 20 min for antigen retrieval. Additional permeabilization with 0.1% Triton™ X-100 in PBS was performed for iNOS staining, followed by three washes with PBS. All sections were blocked with 10% goat serum for 30 min, followed by a second blocking of donkey polyclonal anti-mouse Fab fragment antibody (40 μg/mL; Jackson ImmunoResearch Inc., USA, catalog no. 715-007-003) for mouse-on-mouse CD68 staining. Thereafter, incubation with primary antibodies including mouse anti-CD68 and rabbit anti-iNOS (5 μg/mL; Arigo Biolaboratories Corp., Taiwan, catalog no. ARG10514 and ARG56509) at 4 °C overnight, followed by goat anti-mouse IgG H&L horseradish peroxidase (4 μg/mL; Abcam, UK, catalog no. ab6789) for 30 min before staining with 3, 3′-diaminobenzidine (DAB; Thermo Fisher Scientific, USA, catalog no. 34065). For the elastic chondrogenic bioactivity of EARc scaffolds, the assessment was performed on paraffin-embedded and sectioned mouse subcutaneous implantation samples collected at two- and four-week post-implantation, as previously described [31]. Freshly harvested samples were fixed in 4% paraformaldehyde for 4 h, processed for paraffin embedding and tissue sectioning before staining. SOX9, COL2A1, and ELN immunofluorescence staining were performed as previously described in ‘In vitro assessment of elastic cartilage-associated protein levels’. Alcian blue staining was performed on deparaffinized and rehydrated sections using Alcian blue dye (1% Alcian blue in 3% acetic acid, pH 2.5; Electron Microscopy Sciences, USA, catalog no. 26026-13) for 30 min, then washed with tap water. Subsequently, sections were counterstained in nuclear fast red solution (Electron Microscopy Sciences, USA, catalog no. 26078-05) for 3 min, followed by a 1-min tap water wash before dehydration and mounting. Verhoeff van Gieson staining was performed on deparaffinized and rehydrated sections using freshly prepared Verhoeff working solution (5% alcoholic hematoxylin, 10% ferric chloride, Weiger's iodine solution; Electron Microscopy Sciences, USA, catalog no. 26374-01, 26374-02 and 26374-03) for 1 h. After washing with tap water for 30 s, sections were differentiated in 2% ferric chloride (Electron Microscopy Sciences, USA, catalog no. 26374-02). Subsequently, they were counterstained in Van Gieson's solution (Electron Microscopy Sciences, USA, catalog no. 26374-06) for 3 min, followed by a 1-min tap water wash before dehydrating and mounting. Images were captured and semi-quantitative analyses were performed as previously described in ‘In vitro assessment of elastic cartilage-associated protein levels’. Elastic cartilage-associated markers were semi-quantified by the fluorescence signal intensity of SOX9, COL2A1, and ELN, along with the signal intensity of GAGs and elastic fibers. All measurements were normalized to the nuclei intensity (Hoechst 33342 for immunofluorescence staining, fast red for Alcian blue, and black for Verhoeff Van Gieson staining). For ectopic mineralization assessment, freshly harvested samples were scanned as previously described [114]. A HA (Ca10(PO4)6(OH)2) microCT phantom was used as a positive control for all images (Supplementary Fig. S12). To ensure sensitivity and reproducibility in calcification detection, microCT thresholds were calibrated against HA phantoms ranging from 0 to 1000 mg HA/cm3 (Supplementary Fig. S13) [114]. For each phantom insert, the mean attenuation was measured to construct a calibration curve (grayscale vs. HA density), enabling the conversion of voxel grayscale values into HA-equivalent density. Calcified regions were segmented using a fixed lower threshold of 100 mg HA/cm3, a value chosen to ensure sensitive detection of mineralized tissue while excluding non-mineralized matrix, consistent established standards in calcification research [115,116]. In vivo evaluation using rabbit auricular defect model In vivo orthotopic characterization of EARc scaffold elastic cartilage-associated bioactivity was performed using a rabbit auricular defect model, as similarly described [18,117]. This study was performed according to the guidelines established by the University's Animal Experimentation Ethics Committee (approved protocol: 22-216-GRF). New Zealand White rabbits (both male and female) from 12 to 16 weeks old with a body weight of 3.0-4.5 kg were used in the current study and maintained in a 12:12 h light-dark cycle with free access to standard laboratory food and water. Rabbits were housed for one week for acclimation before surgery. The rabbits were randomly divided into four groups: sham, untreated (defect only), treated with 3D-printed Triphasic 6 PHT scaffold only (empty scaffold), or treated with optimal EARc scaffold (3D-printed Triphasic 6 PHT scaffold containing EARc, which comprised hMDSC spheroids and Medium E). In brief, rabbits were initially anesthetized by intramuscular injection of 35 mg/kg ketamine (AlfaMedic Ltd, Holland, catalog no. 013004) and 5 mg/kg xylazine (AlfaMedic Ltd, Holland, catalog no. 013006). Subcutaneous injection of 0.05 mg/kg buprenorphine and 20 mg/kg cephalexin was administered as an analgesic and anti-infective, respectively. Rabbits were placed on a covered heating pad during surgery to maintain body temperature. The surgical area was shaved and disinfected with alternating applications of betadine and 70% alcohol. Using aseptic technique, a 1 cm × 2 cm skin flap was made, and rabbits either underwent sham surgery or received full-thickness auricular cartilage defects (8 mm length × 3 mm width). Thereafter, the defect was either left empty (i.e., defect only) or immediately repaired by placing a 3D-printed Triphasic 6 PHT scaffold only (empty scaffold) or optimal EARc scaffold (3D-printed Triphasic 6 PHT scaffold containing EARc, which comprised of hMDSC spheroids and Medium E) into the defect and followed by skin flap closure. After surgery, rabbits were allowed to recover on a heating pad and were allowed free-cage activity. 0.05 mg/kg buprenorphine was administered subcutaneously twice a day for three days for post-operative analgesia. At one-day, two-, four-, and eight-week post-surgery, rabbits were euthanized via intracardiac overdose of pentobarbital (AlfaMedic Ltd, Holland, catalog no. 013003). One treatment was performed on one rabbit ear. Both genders and the surgical side were randomized. Rabbit ears were collected for cell viability and histological analyses. hMDSC spheroids retention and survival post-implantation were evaluated using near-infrared fluorescent labeling, as similarly described [118]. Prior to implantation, hMDSC spheroids were labeled by incubating cells with 1 μg/mL DiR dye (1, 1′- dioctadecyl-3, 3, 3′, 3′-tetramethylindotricarbocyanine iodide, Thermo Fisher Scientific, catalog no. D12731) for 30 min at 37 °C. For imaging, rabbits were euthanized at predetermined time points, and ear specimens were immediately analyzed using an IVIS Spectrum CT imaging system (PerkinElmer, USA, model no. 128214/F). Three experimental groups were compared: uninjured controls, unlabeled implants as unstained controls, and DiR-labeled EARc scaffolds. The acquired data were analyzed using Living Image Software (PerkinElmer, USA, version 4.7.4). Fluorescence was quantified under standardized conditions: the excitation wavelength of 745 nm and the emission wavelength of 800 nm. The fluorescence signal was shown as radiant efficiency (photons/sec/cm2/sr) and quantified by selecting the region of interest (ROI) and calculating the total (photons/sec). For assessment of graft-host tissue integration and potential tissue structures associated with undesired ectopic calcification, H&E and Masson's Trichrome staining were performed. Briefly, sections were deparaffinized, rehydrated, and stained according to the manufacturer's instructions (H&E: Electron Microscopy Sciences, USA, catalog no. 26042 and 14851; Masson's Trichrome: BioGnost Ltd., Croatia, catalog no. MST-100T). For the elastic chondrogenic bioactivity of EARc scaffolds, ACAN, COL2A1, and ELN immunofluorescence staining were performed as previously described in ‘In vitro assessment of elastic cartilage-associated protein levels’. SOX9 staining was not performed due to the unavailability of antibodies for non-rabbit sources. Primary antibodies included mouse anti-aggrecan (5 μg/mL; ACAN, Thermo Fisher Scientific, USA, catalog no. MA3-16888), goat polyclonal anti-collagen type II (5 μg/mL; COL2A1, Southern Biotech, USA, catalog no. 1320-01), and guinea pig anti-elastin (5 μg/mL; ELN, Abcam, UK, catalog no. ab52115). After primary antibody staining, sections were rinsed and incubated with secondary antibodies for 1 h at room temperature. Secondary antibodies included Alexa Fluor 488-labeled donkey anti-mouse (10 μg/mL; Thermo Fisher Scientific, USA, catalog no. A21202), Alexa Fluor 555-labeled donkey anti-goat (10 μg/mL; Thermo Fisher Scientific, USA, catalog no. A21432), and Alexa Fluor 647-labeled donkey anti-guinea pig (10 μg/mL; Southern Biotech, USA, catalog no. 706-605-148) antibodies. Nuclei counterstaining was performed by rinsing and staining sections with 0.3 μg/mL Hoechst 33342. Alcian blue staining and Verhoeff van Gieson staining were performed as previously described in ‘In vivo evaluation using mouse dorsal subcutaneous implantation model’. Images were captured using an automatic slide scanner (Zeiss, Germany, model no. Axioscan 7) equipped with four LED lamps (630 nm; 555 nm; 475 nm; 385 nm; Zeiss, Germany, model no. Colibri 7) for fluorescence light source and illuminator VIS-LED (Zeiss, Germany, model no. WL-LED) for transmitted light, Hoechst-, GFP-, Cy3-, and Cy5-equivalent fluorescent filters, and Axiocam 705 colour/Orca Flash 4.0 V3 mono camera. Semi-quantitative analyses were performed using Adobe Photoshop (Adobe Inc., USA, version 22.0.0) by measuring the average pixel intensity via the histogram tool. Elastic cartilage-associated markers were semi-quantified by the fluorescence signal intensity of ACAN, COL2A1, and ELN, along with the signal intensity of GAGs and elastic fibers. All measurements were normalized to the nuclei intensity (Hoechst 33342 for immunofluorescence staining, fast red for Alcian blue, and black for Verhoeff Van Gieson staining). For ectopic mineralization assessment, von Kossa and Alizarin Red staining were performed on deparaffinized and rehydrated sections, and freshly harvested samples were scanned as previously described in ‘In vivo evaluation using mouse dorsal subcutaneous implantation model’. For von Kossa staining, sections were incubated in 1% silver nitrate solution (Sigma-Aldrich, USA, catalog no. 1.93200) for 30 min under ultraviolet light. After incubation, sections were washed in distilled water, treated with 5% sodium thiosulfate (Sigma-Aldrich, USA, catalog no. 217263) for 5 min. The sections were counterstained with nuclear fast red (Sigma-Aldrich, USA, catalog no. N-8002), followed by a final rinse in distilled water. For Alizarin Red staining, sections were incubated in Alizarin Red S solution (Sigma-Aldrich, USA, catalog no. A-3757) for 5 min. Excess stain was removed, and sections were counterstained with fast green (BDH Reagents & Chemicals, Hong Kong, catalog 340304F) for 30 s. All sections were dehydrated in increasing ethanol concentrations and mounted. Images were captured as previously described. Statistical analysis Statistical analysis was performed using SPSS® Statistics 27.0 (IBM, USA) software. A p value of <0.05 was used to establish statistical significance. Sample sizes ranged from 3 to 17 replicates per group from 2 to 5 independent experiments, with the sample number indicated in the figure legends. Quantitative data were presented as means ± standard error of mean (mean ± SEM). Prior to p value computation, parametric testing assumptions were assessed using the Shapiro–Wilk and the Levene tests for normal data distribution and equal variances among groups, respectively. Unpaired t-test was performed to determine statistical significance between two groups. For multiple comparisons, one-way analysis of variance (ANOVA) was performed, followed by Tukey's honestly significant difference for data that satisfied assumptions of normality and equal variance, while Brown-Forsythe ANOVA with Dunnett's T3 post hoc test was performed for data that did not meet assumptions of normality and equal variance [119]. For failure mode analysis, a Chi-squared test was performed using a Bonferroni-corrected p value of 0.0083. Data availability statement The data that support the findings of this study are available from the corresponding author upon reasonable request. Ethics approval and consent to participate All the animal experiments were approved and conducted according to the guidelines established by the University's Animal Experimentation Ethics Committee (approved protocol 21-199-MIS for mice and 22-216-GRF for rabbits). CRediT authorship contribution statement Na Huang: Formal analysis, Methodology, Writing – original draft. Shui Kwong John Lee: Writing – review & editing. Ka Ho Kenneth Lee: Writing – review & editing. Chunyi Wen: Writing – review & editing. Yun Chang: Writing – review & editing. Ming Zhang: Writing – review & editing. Cheng Dong: Writing – review & editing. Chuanbin Mao: Writing – review & editing. Dan Michelle Wang: Funding acquisition, Writing – review & editing. Dai Fei Elmer Ker: Funding acquisition, Supervision, Writing – review & editing. Declaration of competing interest The authors declare the following personal relationships which may be considered as potential competing interests: Ka Ho Kenneth Lee is currently employed by Occam Biosciences. Patent(s) concerning the stress-dissipating and elastic cartilage-inducing biomaterial have been filed.