Coaxial electrohydrodynamic printing of core–shell microfibrous scaffolds with layer-specific growth factors release for enthesis regeneration

The rotator cuff tear has emerged as a significant global health concern. However, existing therapies fail to fully restore the intricate bone-to-tendon gradients, resulting in compromised biomechanical functionalities of the reconstructed enthesis tissues. Herein, a tri-layered core–shell microfibrous scaffold with layer-specific growth factors (GFs) release is developed using coaxial electrohydrodynamic (EHD) printing for in situ cell recruitment and differentiation to facilitate gradient enthesis tissue repair. Stromal cell-derived factor-1 (SDF-1) is loaded in the shell, while basic fibroblast GF, transforming GF-beta, and bone morphogenetic protein-2 are loaded in the core of the EHD-printed microfibrous scaffolds in a layer-specific manner. Correspondingly, the tri-layered microfibrous scaffolds have a core–shell fiber size of (25.7 ± 5.1) μm, with a pore size sequentially increasing from (81.5 ± 4.6) μm to (173.3 ± 6.9) μm, and to (388.9 ± 6.9 μm) for the tenogenic, chondrogenic, and osteogenic instructive layers. A rapid release of embedded GFs is observed within the first 2 d, followed by a faster release of SDF-1 and a slightly slower release of differentiation GFs for approximately four weeks. The coaxial EHD-printed microfibrous scaffolds significantly promote stem cell recruitment and direct their differentiation toward tenocyte, chondrocyte, and osteocyte phenotypes in vitro. When implanted in vivo, the tri-layered core–shell microfibrous scaffolds rapidly restored the biomechanical functions and promoted enthesis tissue regeneration with native-like bone-to-tendon gradients. Our findings suggest that the microfibrous scaffolds with layer-specific GFs release may offer a promising clinical solution for enthesis regeneration.

Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence.Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.

Introduction
The rotator cuff of the shoulder, which is frequently subject to injury from excessive loading and degenerative tendinopathy, has become a major global health concern, with millions of new cases reported annually [1,2].Conventional surgical suturing techniques for repairing rotator cuff tears are often ineffective due to a lack of innate healing ability, resulting in the formation of scar tissues instead of the desired enthesis tissues with tendon-to-bone gradient interfaces [3].The inferior regenerated enthesis tissues result in a high retear ratio of the repaired rotator cuffs, leading to ongoing suffering for patients [4].
Emerging tissue engineering strategies may provide promising ways for the functional reconstruction of the enthesis tissues [5,6].For example, adipose-derived stem cells (ADSCs) were cultured and harvested as cell sheets, which were then delivered to rats with chronic rotator cuff tears as a therapeutic intervention.The engineered ADSCs sheets were able to improve fibrocartilage remodeling and enhance bone formation for the healing of enthesis tissues [7].However, the main drawbacks associated with cell sheet engineering are the limited thickness (<80 µm) [8] and mismatched cellular architectures between the engineered homogeneous cell sheets and native gradient enthesis.Jiang et al employed 3D bioprinting and melt electrospinning to fabricate zonal-specific constructs containing ADMSCs, markedly improving the reconstruction of gradient fibrocartilaginous interfaces in enthesis tissues [9].Chae et al presented an innovative study on bioprinted biomimetic living tissue constructs with a compositional gradient bioink, which exhibited the potential in zone-specific inducibility and multi-tissue formation mimicking the tendon-tobone interfaces [10].Similarly, we previously proposed to bioprint stem cells-laden constructs with layer-specific growth factors (GFs) [11].The GFs were able to facilitate the regionspecific differentiation of loaded stem cells in vitro and significantly enhanced the functional enthesis tissue regeneration when implanted in vivo.Nevertheless, the clinic usage of exogenous stem cells is restricted by invasive and painful harvesting procedures, short self-renewal in vitro, and potential morbidity of donor sites [12][13][14].
Unlike directly incorporating exogenous cells, cellinstructive tissue engineering scaffolds with specifically designed structural and/or compositional configurations have gained extensive attention owing to their unique capability to guide endogenous stem cell growth for improving the reconstruction of enthesis interfaces [15,16].For instance, tape-casted bilayered polycaprolactone (PCL)/calcium phosphate silicate spacers [17], and electrospun bilayered poly (L-lactide) thin films with/without nano-hydroxyapatite [18] were employed for the treatment of rotator cuff tears, and these scaffolds with layer-specific compositions of polymers and osteoinductive materials were able to guide collagen alignment and gradient mineralized cartilage formation.Recent developments in 3D printing further enabled the construction of highly porous scaffolds with designer structural and compositional distributions [19].In one interesting study, Tarafder et al [15] encapsulated tenogenic, chondrogenic, and osteogenic GFs inside poly (lactic-co-glycolic acid) microspheres and then implemented the GFs-laden microspheres for the extrusion-based printing of porous PCL scaffolds with layerspecific GFs distributions.The spatially-released GFs can act as local regulators for modulating cellular activities, which successfully guided regional differentiation of mesenchymal progenitor cells in vitro and the integrative healing of tendonto-bone interfaces in vivo.
Yet, extrusion-based printing possesses relatively low resolution, which commonly results in microstrands with a diameter of 200-400 µm and a spacing larger than 200 µm, and thus limits the cell guiding capabilities of the resultant scaffolds.Recently, electrohydrodynamic (EHD)-printed microfibrous scaffolds were developed for tissue engineering applications due to their unprecedented advantages for generating cell-favorable microenvironments [20][21][22].It is reported that EHD-printed PCL scaffolds with specifically designed microscale fibers and pores can enhance cellular activities, including alignment [23], adhesion, proliferation and migration [24].However, the absence of stem cells at the wound locations remains one of the main causes of the failure of rotator cuff healing [19,25].Therefore, microfibrous scaffolds printed by EHD printing with the ability to recruit and direct regional differentiation of endogenous stem cells have great potential for healing tendon-to-bone interfaces and warrant further investigation.
Here we proposed to use coaxial EHD printing for the fabrication of cell-instructive microfibrous scaffolds with layerspecific structural and compositional organizations, and the potential of the scaffolds for tendon-to-bone enthesis tissue reconstruction was investigated (figure 1).To facilitate stem cell recruitment and regional differentiation, the EHD-printed microfibers were designed with core-shell structures, where chemokines (stromal cell-derived factor-1, SDF-1) laden in the sheath and differentiation-inductive GFs laden in the core region of the printed fibers.Because natural enthesis tissues display a gradient interface transitioning from tendon to fibrocartilage and bone, scaffolds with three distinct layers were designed and EHD-printed with the corresponding compositions of SDF-1/basic fibroblast GF (bFGF) and a pore size of 100 µm, SDF-1/transforming GF-beta (TGFβ) and a pore size of 200 µm, and SDF-1/bone morphogenetic protein-2 (BMP-2) and a pore size of 400 µm as tenogenic, chondrogenic, and osteogenic instructive layers, respectively.The structural organizations, mechanical properties, and GFs release properties of the microfibrous scaffolds were tested.The effect of the scaffolds on cellar viability, proliferation, and recruitment was investigated using Live/Dead staining and transwell experiments.Following this, the ability of the instructive properties of the scaffold to direct cell differentiation was explored using immunofluorescence staining and real-time quantitative polymerase chain reaction (RT-qPCR).In vivo experiments were further conducted to confirm the effect of the layer-specific microfibrous scaffolds on collagen fiber remodeling, fibrocartilage reconstruction, and osteointegration.

Preparation and characterization of GFs-loaded chitosan nanoparticles
Chitosan nanoparticles were synthesized as carriers for GFs using a previously established method [26,27], which involves the ionotropic gelation of chitosan with sodium tripolyphosphate (TPP).Briefly, 75 mg of chitosan was dissolved in 25 ml of 1% acetic acid.10 mg TPP was mixed with 10 µl of diluted GFs in 1 ml of 1% acetic acid solution.The TPP solution was then added to the chitosan solution dropwisely, stirred magnetically for 5 min, and subjected to ultrasonication (60 W, BILON96-II, Bilon Biotechnology Co., China) for 10 min to form a uniform emulsion.The resulting GFs-loaded chitosan nanoparticles were obtained by filtration, washing, and freezedrying, and the size and shape were analyzed using a particle analyzer (Delsa™ Nano, Beckman Coulter, USA) and a scanning electron microscope (SEM, SU8010, Hitachi, Japan).

Coaxial EHD printing of microfibers with tunable core-shell structures
As depicted in figure 2(a), we employed a coaxial EHD printing system (Shaanxi Baipusheng Medical Technologies Co., Ltd, China) to fabricate microfibrous scaffolds with layerspecific structural and compositional configurations [28].The corresponding nanoparticles containing chemokine and differentiation GFs were mixed with the sheath and core inks, each at a fixed concentration of 100 mg•ml −1 , and loaded into separate 1 ml syringes, which were independently controlled by a syringe pump.A coaxial printing nozzle was mounted on the z-axis and connected to a high-voltage supplier.The core inlet of the nozzle was linked to the syringe loaded with differentiation GFs, while the sheath inlet was connected to the syringe loaded with chemokine.The process of coaxial EHD printing was initiated by placing an indium tin oxide-coated glass substrate as the collecting surface onto the moving stage with a nozzle-to-collector distance of 2.5 mm.The core and sheath PCL solutions were simultaneously fed into the coaxial nozzle by opening the syringe pump and applying a high voltage of 2 kV to the nozzle.This resulted in the formation of coreshell PCL microfibers through the evaporation of the HFIP solvent.
The effect of ink feeding rate on the structure of the coreshell microfibers produced by the coaxial EHD printing process was investigated.Rhodamine was included in the core ink to enable the visualization of the core-shell structure.Two sets of experiments were performed.In the first experiment, the total ink feeding rate was set at 50 µl•h −1 , and the ratios of core ink to sheath ink were varied from 1:9 to 9:1.In the second experiment, the ratio of the core ink to sheath ink was fixed at 1:1, and the total ink feeding rate was increased from 10 µl•h −1 to 50 µl•h −1 and 100 µl•h −1 .Bright-field images of the printed fibers were captured using an inverted fluorescence microscope (ECLIPSE Ti, Nikon, Japan) for the determination of the size of the core and sheath components.Table S1 demonstrates the detailed variables for the optimization of the coaxial EHD printing.

Coaxial EHD printing of microfibrous scaffolds with layer-specific structural and compositional configurations
Coaxial EHD printing was employed to fabricate microfibrous scaffolds with layer-specific structural and compositional configurations by changing the printing inks and precisely directing the movement of the x-y moving stage according to a userspecific designed trajectory in a layer-by-layer manner.As abovementioned, a tri-layered microfibrous scaffold consisting of tenogenic, chondrogenic, and osteogenic instructive layers was designed for enthesis tissue reconstruction (figure 1).For the construction of the tenogenic layer, SDF-1 and bFGF were added in the sheath and core inks, respectively, and the core-shell microfibers were stacked in a cross-hatched configuration with a fiber spacing of 100 µm and a printing layer number of 60.Similarly, the chondrogenic/osteogenic layers were fabricated by replacing the bFGF with TGF-β/BMP-2, with a fiber spacing of 200/400 µm, respectively.The scaffolds designed with structural and compositional gradients were thus termed the 'S + R + D group'.By contrast, the microfibrous scaffolds without any GFs (henceforth 'S group') and microfibrous scaffolds with differentiation GFs only (henceforth 'S + D group') were fabricated as the control groups.
The structural organizations of the coaxial EHD-printed layer-specific microfibrous scaffolds and the distribution of encapsulated chitosan nanoparticles were captured with a confocal laser scanning microscope (Olympus, Japan) and SEM.The mechanical properties of the coaxial EHD-printed layerspecific microfibrous scaffolds (20 mm × 20 mm × 0.3 mm) were evaluated at a 10 mm gauge length with a cross-head speed of 5 mm•min −1 [29,30].Tensile modulus, strength, ultimate strain, and toughness were determined from stress-strain curves following established methods [29].

Release of the GFs from the coaxial EHD-printed microfibrous scaffolds
The core-shell structure is engineered with an expectation to achieve the sequent release of chemotactic and differentiation GFs, which consequently allows the recruitment and differentiation of autologous stem cells in a sequential process.The release behaviors of four GFs, BMP-2, TGF-β, bFGF, and SDF-1, were analyzed utilizing enzyme-linked immunosorbent assay kits (SAB, USA) following a protocol described in previous literature with modifications [31].Briefly, the coaxial EHD-printed layer-specific microfibrous scaffolds (n = 3) were soaked in 2 ml of PBS solution.At various time points, including 0.5 h, 1 h, 2 h, 3 h, 6 h, 12 h, 24 h, 2 d, 3 d, 7 d, 14 d and 28 d, the entire supernatants were collected, and the released GFs were quantified.

Decipher the effect of microfibrous scaffolds on cellular viability, recruitment, and proliferation
2.6.1.Cell viability and proliferation on each layer of microfibrous scaffolds in the S + R + D group.The cytocompatibility of the osteogenic, chondrogenic, and tenogenic layers of the microfibrous scaffolds in the S + R + D group was investigated by directly seeding BMSCs on them.The scaffolds were first immersed in 75% alcohol for 15 min and exposed to UV light for disinfection.After carefully washing the scaffolds three times in PBS, 500 µl culture medium with rabbit BMSCs was seeded on each scaffold with a density of 1 × 10 6 cells•ml −1 .Adequate culture medium was added to each dish after 4 h of cell attachment.After being cultured for 1 d, 2 d and 3 d, the viability of the seeded BMSCs on the scaffolds was evaluated by Live/Dead assay (Thermo Fisher, USA) and then observed with a fluorescent microscope.Cytotoxicity of different layer scaffolds was evaluated via a CCK-8 assay kit (Biosharp Life Science, China).

Cell recruitment and proliferation on each layer of microfibrous scaffolds in S + D and S + R + D groups.
Transwell cell culture inserts (Corning, USA) were utilized to study the cell recruitment capacity of the S + D and S + R + D groups.Briefly, different layer scaffolds of the S + D and S + R + D groups were placed at the bottom chamber in 24 well plates under the transwell inserts.5 × 10 4 BMSCs labeled with Cell Tracker™ Green CMFDA (Invitrogen, USA) were seeded into the top chamber of the transwells, and the plates were incubated at 37 • C for 24 h and 48 h before measurement.At specific time points, the cells above the membranes were eliminated, while those across the membranes were visualized using a confocal laser scanning microscope (A1+, Nikon, Japan).Images of migrated cells were captured and analyzed using ImageJ software.
The osteogenic, chondrogenic, and tenogenic layer scaffolds from the two groups were then utilized to analyze the seeding efficiency, spreading, and proliferation of rabbit BMSCs.Briefly, the microfibrous scaffolds were immersed in a culture medium with a BMSCs density of 1 × 10 6 cells•ml −1 and incubated at 37 • C for 24 h.The scaffolds were then moved to new culture plates and incubated for another 6 d before analyzing.The cellular morphology on the scaffolds was visualized by phalloidin and DAPI staining at 24 h and 7 d.Cell numbers were then quantified using ImageJ software's 'Watershed' function [32,33].

Directed cell differentiation on each layer of the microfibrous scaffolds in S, S + D, and S + R + D groups
The osteogenic, chondrogenic, and tenogenic layer scaffolds of the S, S + D, and S + R + D groups were EHDprinted individually to investigate their directional differentiation capabilities [34].After being cultured for 14 d, phenotypic differences at each layer scaffolds in the S group, the S + D group, and the S + R + D group were analyzed using immunofluorescence staining of tendon fibroblast marker Tenomodulin (TNMD), chondrocyte markers SRY-Box Transcription Factor 9 (SOX-9) and osteoblast marker Runt-related transcription factor 2 (RUNX-2), respectively.Detailed procedures can be found in supplementary information.Finally, the cellular constructs were rinsed with PBS and observed via laser confocal microscopy.The expression of TNMD, SOX-9, and RUNX-2 was semi-quantified based on the fluorescence intensity using ImageJ software and normalized by the expression in the corresponding layer scaffolds in the S group.
In addition, RT-qPCR was employed to quantitatively assess the expression of specific genes, including tendon fibroblast markers TNMD and scleraxis (SCX), chondrocyte markers SOX-9 and aggrecan (ACAN), as well as osteogenic markers BMP-2 and osteocalcin (OCN), in the various layer scaffolds of the S + R + D group.Following a 14 day culture, BMSCs on each layer scaffold were harvested.mRNA was extracted and converted into cDNA, and RT-qPCR was performed on an ABI 7500 RT-PCR system (Applied Biosystems, USA) according to the manufacturer's protocol [35].The primer sequences are provided in table S1.Target gene expression was quantified using the ∆∆CT method and normalized to the corresponding layer scaffolds in the S group.

In vivo validation of the coaxial EHD-printed tri-layered microfibrous scaffolds
48 adult male New Zealand white rabbits (2.5-3.0 kg) were randomly divided into four groups.A rotator cuff tear model was created in both extremities based on the supraspinatus tendon, following established procedures [35][36][37].Coaxial EHDprinted tri-layered microfibrous scaffolds were implanted and secured to form experimental groups (S, S + D, S + R + D), while sutures directly connected bone channels in the control group (Ctl) without scaffolds [38,39].At 6 weeks and 12 weeks post-operation, 6 rabbits (totaling 12 rotator cuff samples) in each group were sacrificed for biomechanical and histological analysis.Animal experiments were approved by the animal research committee of Xi'an Jiaotong University.The details of biomechanical tests, histological analysis, and immunofluorescence analysis of the retrieved samples are available in Methods in the supplementary information.

Statistical analysis
All experimental data was analyzed with SPSS 25.0 (SPSS Inc., Chicago, USA), presenting as means ± standard deviations.One-way ANOVA and Tukey's test were used to compare the differences among different groups.A significant difference was noted at P < 0.05.

Coaxial EHD-printing of layer-specific core-shell microfibrous scaffolds
The chitosan nanoparticles exhibited spherical shapes and smooth surfaces as revealed by SEM (figure S1(a)).A unimodal distribution in particle size was observed, with an average diameter of (515.1 ± 146.5) nm (figures S1(b) and (c)).The printability of the chitosan nanoparticles was evaluated by incorporating them into PCL/PEO inks for EHD printing.Figure 2(a) depicts the coaxial EHD-printing platform and an enlarged image of the coaxial printing nozzle.Upon application of high voltage, a continuous, tiny fiber much smaller than the orifice size was ejected and deposited on the collector (figure 2(b)).Optical microscopy revealed that the chitosan nanoparticles were uniformly distributed and maintained their spherical shape, suggesting that the printing process had little effect on their morphology (figure S1(b)).
As depicted in figures 2(c) and (e), by tuning the flow ratio of the core ink and sheath ink to 1:9 with a fixed total feeding rate of 50 µl•h −1 , the sizes of the core line and sheath line were determined to be (1.6 ± 0.4) µm and (27.1 ± 6.1 µm), respectively.When the ratio was adjusted to 1:4, the core line and sheath line width reached (6.3 ± 2.3) µm and (28.7 ± 5.2) µm, respectively.With the ratio further increasing to 1:1, the core and sheath line widths changed to (10.0 ± 3.5) µm and (25.7 ± 5.1) µm, respectively.However, no evident coaxial structure was observed when the ratio further increased to 4:1.Therefore, we adopted a 1:1 ratio for the subsequent experiments to ensure the stability of the printing and maximize the loading capacity of GFs in both core and sheath lines.
The size of the coaxial fibers produced by EHD printing is dependent on the total feeding rate of the inks (figures 2(d) and (e)).A reduction in the feeding rate to 10 µl•h −1 resulted in a decrease in the size of the core and sheath lines to

Characterization of the coaxial EHD-printed layer-specific microfibrous scaffolds
The tri-layered scaffolds, designed to consist of osteogenic, chondrogenic, and tenogenic instructive layers, were successfully printed using the developed coaxial EHD printing strategy, as illustrated in figure 3(a).The microfibrous scaffolds exhibited a gross fiber width of (24.3 ± 6.3) µm and a thickness of (314.5 ± 16.0) µm.As determined by laser confocal microscopy, the tenogenic, chondrogenic, and osteogenic regions were clearly distinguished and displayed respectively accumulated thicknesses of (96.2 ± 17.9) µm, (187.3 ± 12.4) µm and (314.5 ± 16.0) µm from the bottom to the top (figure 3(b)).There was no noticeable separation between the different layers, indicating robust adhesion and interconnections among the various layers (figure 3(c)).
Additionally, the structural integrity and stability of the coaxial EHD-printed microfibrous scaffolds were wellpreserved even after the dissolution of PEO in aqueous environments, which is critical for their applications both in vitro and in vivo [40].After 24 h of incubation in PBS, the coaxial EHD-printed microfibrous scaffolds exhibited an increased surface roughness, characterized by a hierarchical nanofibrous morphology on the printed microfibers (figure 3(d)).This hierarchical nanofibrous morphology might be mainly caused by the leaching of the water-soluble PEO from the PCL matrix, which forms nanopores and rough nano-fibrillar morphology along with the axial direction of the EHD-printed microfibers.Similar phenomena have also been reported in previous literature [41,42].Furthermore, the emerging complex nanomorphology significantly increased the surface area as compared with that of the original EHD-printed microfibers, which may enhance cellular adhesion and proliferation, thus providing a more favorable environment for tissue regeneration [43,44].
The designed tri-layered scaffolds aim to promote enthesis healing when placed between the tendon end and attachment point [45].Their function is not to replicate the exact mechanical properties of tendon, cartilage, or bone, but to ensure sufficient durability to endure the implantation process and subsequent shoulder movements post-surgery.As shown in figure 3(e), the tensile modulus of the microfibrous scaffolds decreased from (8.56 ± 4.83) MPa to (4.82 ± 2.26) MPa after incubation in PBS solution.The tensile strength also decreased from (667.3 ± 99.8) kPa to (451.4 ± 60.16) kPa (figure 3(f)).However, the ultimate tensile strain (figure 3(g)) and toughness (figure 3(h)) did not exhibit any significant changes, indicating their potential to withstand surgical implantation procedures and maintain local stability post-operatively.
The surface morphology of the scaffolds changed from dense to porous after incubation in PBS, which might lead to the rapid release of the chitosan nanoparticles and thus the loaded GFs.To evaluate the cumulative release profiles of GFs from the coaxial EHD-printed microfibrous scaffolds, we measured the release amounts over time.As shown in figure 3(i), the release speed of the GFs loaded in both the core and the sheath of the microfibers showed a similar fast-release trend within the first 48 h.At the 48th hour, the cumulative release amounts of SDF-1 in the sheath, and BMP-2, TGF-β, and bFGF in the core were (2389.0± 59.5) pg, (2170.7 ± 97.9) pg, (2158.7 ± 102.7) pg and (2130.6 ± 107.2) pg, respectively.However, a significantly rapid release of SDF-1 was detected in the following 4 weeks, with the cumulative release of SDF-1 on day 28 ((3791.58± 118.70) pg) being significantly higher than that of the other three factors (BMP-2 of (3354.59± 196.00) pg, TGF-β of (3 426.66 ± 97.38) pg, and bFGF of (3 234.41 ± 97.08) pg).The results suggest that the release of differentiation GFs loaded in the core fibers lags behind that of chemotactic GFs in the sheath, with both exhibiting effective release times exceeding 28 d.

Viability of BMSCs seeded in each microfibrous layer scaffold in the S + R + D group
Rabbit BMSCs were characterized using a combination of surface morphology, growth curve analysis, three lines of differentiation assays, and surface marker expression analysis (figure S3).Results indicated that the viability of BMSCs in the osteogenic, chondrogenic, and tenogenic layer scaffolds remained higher than 95% at 24 h, 2 d and 3 d, respectively (figure S4(a)).Additionally, each layer of the microfibrous scaffold in the S + R + D group was found to effectively support cell proliferation, as confirmed by the CCK-8 assay (figure S4(b)).

SDF-1 promoted the migration of BMSCs toward microfibrous scaffolds
As illustrated in figures 4(a) and (b), a significantly higher number of green fluorescently labeled BMSCs migrated across the microporous membrane in the S + R + D group compared to that observed in the S + D group after a 12 h incubation period.This trend was further validated after 24 h of incubation.These findings were quantitatively confirmed by analyzing the number of migrated cells using ImageJ software (figures 4(c) and (d)).After 12 h of incubation, the cell densities across the membrane in tenogenic, chondrogenic, and osteogenic layer scaffolds of the S + R + D group reached (893.for tenogenic, chondrogenic, and osteogenic layer scaffolds of the S + R + D group, respectively.The results derived from the transwell assay showed that the scaffolds in the S + R + D group exhibited an enhanced capacity to recruit BMSCs as compared with those in the S + D group, which could potentially be harnessed for recruiting stem cells for tissue regeneration in vivo [46,47].

Coaxial EHD-printed microfibrous scaffolds supported cell adhesion and proliferation
As depicted in figures 5(a) and (b), after seeding with equal cell densities and culturing for 24 h, the number of cells attached to each layer scaffold in the S + R + D group was significantly higher compared to the S + D group.The number of cells continued to increase significantly from 24 h to 7 d in both groups.As shown in figure 5(c), the number of cells in the tenogenic ) in the S + D group at 24 h was relatively similar.By contrast, the number of cells adhering to each layer scaffold in the S + R + D group was consistently higher compared to the corresponding S + D groups, as SDF-1 is prone to recruit more stem cells [48,49].Moreover, different cell numbers were observed on the tenogenic layer scaffolds ((527.3± 13.5) cells•mm −2 ), chondrogenic layer scaffolds ((363.7 ± 16.0) cells•mm −2 ), and osteogenic layer scaffolds ((261.3± 9.1) cells•mm −2 ) in the S + R + D group, which may be due to the distinct structural configurations.When the surface area of the layer scaffolds becomes insufficient for promoting cell adhesion and spreading, it may impose a constraint on the maximum number of cells that can be attached.After 7 d of incubation, the rabbit BMSCs in each layer scaffolds of the S + R + D group proliferated substantially and spread over the microfibers' surfaces of the scaffolds.In contrast, the cell numbers in each layer scaffold of the S + D group were significantly lower than that of the S + R + D group, as confirmed by DAPI staining and cell count analysis (figure 5(d)).The abovementioned results demonstrated once again that the local release of SDF-1 can effectively recruit a substantial number of stem cells to the scaffolds.

Directed stem cell differentiation on each microfibrous layer scaffold in S, S + D, and S + R + D groups
The effect of the loaded differentiation GFs on the regional differentiation of the recruited stem cells was further investigated.Figure 6 presents immunofluorescence images of each layer scaffold in S, S + D, and S + R + D groups, which were stained for the tendon fibroblast marker (TNMD), chondrocyte marker (SOX-9), and osteoblast marker (RUNX2), respectively.Nuclear and cytoskeletal components were visualized using specific fluorochromes: DAPI bound to DNA in nuclei in Blue fluorescence, while phalloidin targeted actin filaments in Green fluorescence.Figure 6(a) depicts the majority of cells in the tenogenic cellular construct of the S group spreading along the fibers.The fiber spacing was only about 100 µm, and thus the stem cells could span the gaps obliquely.In the S + D group, a significant portion of cells grew along the microfibers.Red fluorescence indicated TNMD expression in certain cells.Meanwhile, the tenogenic constructs in the S + R + D group exhibited significant cellular proliferation filling the inter-fiber spaces, resulting in a complete cell sheet that covered the entire surface of the scaffold, with only a few circular pore-like vacuities present.Additionally, there was a high expression of red fluorescent markers.
Figure 6(b) illustrates the development of chondrogenic cellular constructs in the S, S + D, and S + R + D groups.The cells in the chondrogenic cellular construct of the S group developed into a homogeneous network structure with little red fluorescence.In contrast, the chondrogenic cellular construct of the S + D group displayed an increased number of cells on the surface of the fibers, with many cells spanning the microfibers and completely covering the lattice structure.Additionally, red fluorescence was detected in several cells.Similar to the S + D group, the cells in the chondrogenic cellular construct of the S + R + D group proliferated extensively and bridged the spacing between fibers, resulting in complete coverage of the microfibrous scaffolds.A high expression of red fluorescent marker was also observed.
The amount and shape of cells in the osteogenic cellular construct in the S group were similar to those in the S + D group (figure 6(c)).There was no obvious red fluorescence in the osteogenic cellular construct in the S group, but clear signals can be observed surrounding the nucleus in the S + D group.For the osteogenic cellular construct in the S + R + D group, spreading cells could be observed on the surface of the scaffold, and the cells proliferated vigorously even through the relatively large spacing of the microfibers, exhibiting a homogeneous sheet-like construct.Fluorescence intensities of TNMD, SOX-9, and RUNX2 were semi-quantitative analyzed as shown in figures 6(d)-(f), with the results showing a significantly higher expression of specific proteins in the S + R + D groups, followed by the S + D group and the S group.

Coaxial EHD-printed microfibrous scaffolds enhanced enthesis tissue regrowth in vivo
Microfibrous scaffolds were implanted between the rabbit supraspinatus tendon stump and the humeral footprint (figure 8(a)).As shown in figure 8(b), the gross morphology at 6 weeks postoperative indicates that the tendon-to-bone interface of the rotator cuff was poorly healed in the Ctl group.In contrast, the S, S + D, and S + R + D groups showed a more continuous tendon-to-bone interface with significant fibrous tissue growth.
At 12 weeks postoperatively, despite improved continuity of the tendon-to-bone interface in the Ctl group, the locally generated new tissue lacked native-like gradient features.In the S group, a substantial amount of tendon-like tissue was generated in the repaired area.However, a discontinuous area was observed, indicating mechanical weakness of the newly-formed tissues.Notably, the tendon-tobone junction site exhibited better healing in the S + D and S + R + D groups, with significantly more tendon-like tissues generated at the tendon end in the S + R + D group.The biomechanical performances of the tissue-scaffold complexes were evaluated at 6 weeks and 12 weeks postoperatively.No significant differences in the cross-sectional area of the rotator cuff insertion area were observed between these experimental groups (figure 8(c)).The maximum failure loads of the S + R + D group, however, reached (131.47 ± 8.95) N and (150.97 ± 8.25) N at 6 weeks and 12 weeks respectively, which were significantly greater than those of the other three groups (P < 0.05, figures 8(d) and S5).
H&E staining was performed at 6 and 12 weeks postoperatively to evaluate the rotator cuff healing (figure 8(e)).At 6 weeks postoperatively, host tissues and cells infiltrated into the tendon-to-bone interface regions in all groups, but the arrangement of cells varied among the groups.The Ctl group showed several fissures at the tendon-to-bone interface and a random distribution of tightly arranged cells around these fissures.In the scaffold groups (S, S + D, and S + R + D), microfibrous structures of the scaffolds were visible at the tendon-to-bone interfaces.Many cellular nuclei were observed to be randomly distributed in the vicinity of the scaffolds in the S and the S + D groups.By contrast, the S + R + D group possessed a reduced number of cells with the appearance of thick connective tissue fiber at the tendon side, suggesting their potential to reduce inflammation and facilitate enthesis regeneration.
At 12 weeks postoperatively, no clear scaffold remnants were found at the tendon-to-bone interfaces in the S, S + D, and S + R + D groups (figure 8(e)).The Ctl group still showed a high density of cells with disorganized tissue growth at the tendon-to-bone interfaces.The cell density in the S group was comparable to that in the Ctl group.Some cracks were observed on the fibrous connective tissue near the tendon side in the S group, which might be caused by the residual structures of scaffolds.The cell number in the S + D group was further reduced compared to the prior two groups, and localized cells showed an aligned trend along with the collagen fibers.The S + R + D group showed favorable tissue repair, with the tendon fibers aligned perpendicularly to the tendon-to-bone interface.The only shortfall compared to the native tendon-to-bone interface was the absence of the intact tidemark (figure 8(f)).Figure 8(g) also confirmed that the reconstructed tendons in the S + R + D groups exhibited significantly superior maturation compared to the other groups, with statistically significant differences (P < 0.05).
Fibrocartilage formation is crucial to the regeneration of functional enthesis tissues [53,54].Here, the cartilage was stained red when combined with Safranine O, whereas the bone tissue was stained green or blue when combined with Fast Green [55,56].The intensity and size of the red metachromasia area were used as indicators to evaluate the content of the fibrocartilage matrix [57].As shown in figure 9(a), 6 weeks after surgery, a large area of metachromasia was observed at the tendon-to-bone interface in the S + R + D group.There were no obvious metachromasia areas in the Ctl, S, and S + D groups.The results at 12 weeks postoperatively showed that the Ctl group had sporadic areas of redstained chondrogenic tissue at the interface, while the bone side of the S group showed lighter red staining.The S + D group exhibited a larger area of metachromasia, however, the disordered cellular arrangement on the tendon side suggested incomplete tendon reconstruction.In contrast, the S + R + D group exhibited the largest area of metachromasia that was well-aligned in structure and similar to the natural interface (figure 9(b)).The metachromasia area was then measured as shown in figure 9(c), and the quantitative results also confirmed that the S + R + D group exhibited a significantly elevated regeneration of cartilage-like tissues.These phenomena are potentially attributed to the recruitment of a substantial number of endogenous stem cells and efficient induction of regional differentiation through the release of various GFs from the coaxial EHD-printed microfibrous scaffolds.
The functional recovery of rotator cuffs also relates to the remodeling of collagen fibers [58,59].Picrosirius red staining was used to evaluate collagen orientation at given time points (figure 9(d)).The results showed that 6 weeks postoperatively, the collagen fiber arrangement in all groups was disorganized and remnants of scaffold material were observable.At 12 weeks postoperatively, the collagen fibers in the Ctl group and S group were still disorganized and the tendon-to-bone interface in the S group was porous and sparsely packed.The S + D group displayed some partially oriented collagen fibers, but the overall structure remained disorganized.In contrast, a large proportion of collagen fibers in the S + R + D group were highly oriented, although the ratio was still different from that of the native tendon-to-bone interface (figures 9(f) and (g)).
We further investigated the of key proteins for the enthesis interface through immunofluorescence staining [60,61].Notably, remodeling of the tendon-to-bone interface of the rotator cuff in the S + R + D group can be confirmed by the gradual convergence of the protein markers SOX9 and SCX towards their respective native distribution positions over time.At 6 weeks postoperatively, both red and green fluorescence exhibited extensive, homogeneous expression in zones 1 and 2 of the S + R + D group, respectively (figure 10(a)).By 12 weeks postoperatively, green fluorescence was reduced in Zone 3 and elevated in Zone 4, while conversely, the red fluorescence displayed increased expression in Zone 5 and decreased in Zone 6 (figure 10(d)).
We found that the integrated optical density (IOD) and mean optical density (MOD) of SCX in the S + R + D group were significantly higher than those in the other groups at both 6 weeks and 12 weeks postoperatively (figures 10(c) and (f)).The IOD of SOX-9 in the S + R + D group exhibited  and functionally integrated tendon-to-bone interface.The collective findings suggest that the S + R + D group initially promotes a wide generation of both cartilage and tendon tissues within the interface region.Subsequently, a process of tissue remodeling occurs, gradually repositioning the cartilage and tendon tissues towards their native-like locations.
Although the results are promising, it should be noted that there are several limitations in this study.Firstly, the current utilization of a rabbit rotator cuff tear model may not fully capture the intricacies of human enthesis tissue repair processes.Therefore, it is crucial to employ large animal models that bear greater resemblance to humans to further validate the effectiveness of the coaxial EHD-printed tri-layered microfibrous scaffolds.Secondly, the 12-week observation period conducted in this study falls short in providing comprehensive insights into the long-term stability and integration of regenerated enthesis tissues.Longer-term observations may still need to evaluate both the durability of the repair process and any potential adverse effects of the present layer-specific core-shell microfibrous scaffolds.

Conclusion
In this study, we developed a tri-layered core-shell microfibrous scaffold with layer-specific GFs release using coaxial EHD printing for enhanced enthesis tissue repair.The shell of the scaffold was loaded with SDF-1, while bFGF, TGF-β, and BMP-2 were loaded in the core in a layer-specific manner.The tri-layered scaffolds were fabricated to have a coaxial fiber width of (24.3 ± 6.3) µm and a thickness of (314.5 ± 16.0) µm, which closely resembled the native enthesis thickness.A more rapid release of SDF-1 and a slightly delayed release of differentiation GFs over approximately four weeks were achieved, which enabled stem cell recruitment and guided their differentiation towards tenocyte, chondrocyte, and osteocyte phenotypes in vitro.The efficacy of the core-shell microfibrous scaffold with layer-specific GFs release was further evaluated using a rabbit rotator cuff injury model.It was found that the regenerated enthesis tissues in S + R + D groups exhibited significantly enhanced biomechanical properties, aligned collagen fibers, and higher expression and ordered distribution of key proteins, surpassing the performance of the other groups.These findings highlight the promising applications of coaxial EHD-printed core-shell microfibrous scaffolds in facilitating enthesis tissue healing and demonstrate significant potential for future clinical usage.

Figure 1 .
Figure 1.Schematic illustration of coaxial EHD printing of layer-specific core-shell microfibrous scaffolds with GFs release for enthesis regeneration.

Figure 2 .
Figure 2. Fabrication and characterization of the coaxial EHD-printed microfibers.(a) The coaxial EHD printing platform.(b) Enlarged image of the ejected microfiber during coaxial EHD printing.(c) Coaxial EHD-printed microfibers with the total feeding rate fixed at 50 µl•h −1 while changing the ratio of core ink and sheath ink from 1:9 to 4:1.(d) Coaxial structure with the ratio of core ink and sheath ink set to 1:1 while changing the total feeding rate from 10 µl•h −1 to 100 µl•h −1 .(e) The line width of the coaxial fibers with various core-sheath ink feeding ratios (n = 3).(f) The line width of the coaxial structure with different total ink feeding rates (n = 3).Blue dashed line: Sheath line.Green dashed line: Core line.

( 4 . 3 ± 1 . 1 )
µm and (6.2 ± 1.8) µm, respectively.Conversely, an increase in the feeding rate to 100 µl•h −1 led to an increase in the size of the core and sheath lines to (29.8 ± 5.7) µm and (53.7 ± 11.5) µm, respectively.Coaxial EHD printing also enables the fabrication of diverse microfibrous patterns by stacking the microfibers in a layer-by-layer manner.Microfibrous lattice scaffolds with fiber spacing of 100 µm, L Bai et al 200 µm and 400 µm were successfully printed as shown in figure S2.

Figure 4 .
Figure 4. Recruitment of BMSCs by coaxial EHD-printed microfibrous scaffolds.(a) Schematic diagram of the chemotaxis experiment.(b) Transwell test of different layer scaffolds in S + D and S + R + D groups for 12 h and 24 h.Quantified results of cell numbers across the membrane after (c) 12 h and (d) 24 h of culture (n = 3).* * * * P < 0.001.

Figure 5 .
Figure 5.The morphology and proliferation of the recruited BMSCs in each layer scaffold in the S + D group and S + R + D group.(a) Schematic diagram of the experiment.(b) Phalloidin and DAPI staining of the scaffolds from S + D and S + R + D groups after co-culturing with BMSCs for 24 h and 7 d.Quantified results of cell numbers on each layer scaffold in the S + D group and S + R + D group after (c) 24 h and (d) 7 d of incubation (n = 3).* P < 0.05.

Figure 8 .
Figure 8. Scaffold implantation and histological assessments of the reconstructed enthesis tissues (n = 6).(a) The scaffold was implanted at the rabbit's rotator cuff injury site.Yellow arrow: implanted scaffolds.(b) The gross morphology of the repaired supraspinatus tendon at 6 weeks and 12 weeks after surgery.Yellow dashed circle: tendon-to-bone interface.(c) The cross-sectional area of the rotator cuff insertion area.(d) The ultimate failure loads of the repaired rotator cuffs.(e) Exemplary images of H&E-stained reconstituted enthesis.(f) H&E stained native enthesis.Green stars: tendon side.Blue star: bone side.Black asterisk: implanted scaffolds.Black arrowhead: degraded scaffolds.Yellow dashed line: tidemark.(g) Tendon maturing scores for the repaired complex.* P < 0.05.

Figure 10 .
Figure 10.Delineating key protein expression at the reconstructed enthesis tissues via immunofluorescence staining (n = 6).Representative immunofluorescence staining images of SOX-9 and SCX at (a) 6 weeks and (d) 12 weeks postoperatively.Yellow asterisk: tendon side.White asterisk: bone side.White arrowhead: implanted scaffolds.Red and green dashed area: region of interest.Semi-quantitative evaluation of the IOD and MOD of (b) and (e) SOX-9 and (c) and (f) SCX.Significant differences (P < 0.05) were marked when compared to the Ctl group ( ∧ ), the S group ( * ), the S + D group (#), and the S + R + D group (&).