Mechanically reinforced core-shell scaffold with integrated structure and function for accelerated tendon repair

PCL (Mn = 80 000), sodium dodecyl sulphate and fluorescein diacetate (FDA) were purchased from Sigma-Aldrich Co., Ltd (USA). Triton X-100, Hematoxylin-Eosin Staining Kit (G1120), Sirius Red Staining Kit (G1472) and Masson's Trichrome Staining Kit (G1340) were purchased from Beijing Solebo Technology Co., Ltd (China). Ethanol (EtOH), sodium hydroxide (NaOH) and dichloromethane (DCM) were purchased from Sinopharm Chemical Reagent Co., Ltd (China). Foetal bovine serum (FBS), trypsin, and Dulbecco's Modified Eagle Medium (DMEM) were purchased from Gibco Co., Ltd (USA). Propidium iodide (PI), phosphate buffer saline (PBS) and bovine serum albumin (BSA) were purchased from Dalian Meilun Biotechnology Co., Ltd (China). The CCK-8 kit (CCK-8) was purchased from Shanghai Tao Zhi Biotechnology Co., Ltd (China). TRIzolsolution, Alexa Fluor594 Goat Anti-Mouse IgG (H + L) and cell culture flasks were purchased from Thermo Fisher Scientific (USA). HiScriptIII RT SuperMix for qPCR (+gDNA wiper) and ChamQ Universal SYBR qPCR Master Mix were purchased from Vazyme Biotech Co., Ltd (China). Collagen type I primary antibody (COL-I, SC-59772) was purchased from Santa Cruz Biotechnology, Inc. (USA). Paraformaldehyde solution (PFA) was purchased from Biosharp (China). Antifade mounting medium with 4,6-diamidino-2-phenylindole (DAPI) was purchased from Solarbio Science & Technology Co., Ltd (China). Cell culture plates (24-well) were purchased from SuRui (China). Centrifuge tubes (5 mL, 10 mL) and pipettes (75 mL) were purchased from Sorfa Life Science Research Co., Ltd (China). Surgipro sutures of size 5-0 were obtained from COVIDIEN Co., Ltd (USA). TM TM ®

PCL solution (30 wt.%) was prepared as the ink using a mixed DCM/EtOH (7:3, v/v) solvent. The solution was fabricated into fibrous sheets at a pre-designed geometry using consistent parameters of needle size (21 G), applied voltage (3 kV), feed rate (20 µL min), tip-to-collector distance (2.5 mm) and stage speed (80 mm s). The as-fabricated sheet is denoted as PCLat different fibre angles of 15° and 90°. −1 −1 Fb

To fabricate the scaffold shell, PCL was first prepared as a film by two-roll milling followed by heat pressing. The obtained film is denoted as PCL. This film was drilled using a femtosecond laser (wavelength of 800 nm and frequency of 1000 Hz) to create the penetrated pores. The porous film was named P-PCLand was further rolled and heat-sealed into tubes with a diameter of ∼3 mm. By subjecting it to uniaxial stretching (HYC-2011, Hongjin, China) with a draw ratio of 5 and a stretching speed of 10 mm minat room temperature, the PCL tube was fabricated into the shell and named UXP-PCL. Fm Fm Fm −1

The PCLsheet was then rolled into a rod-like core. The core-shell tendon scaffold was obtained by inserting the core into the shell. The weight of the scaffold was labelled as W, and the weight calculated based on the volume of the scaffold and a density of PCL (1.145 g mLat 25°C) was labelled W. A porosity of the scaffold was calculated as follows: Fb s c −1

Scanning electron microscopy (SEM): Samples were gold-sputtered at 6 mA for 90 s before imaging using an SEM (MIRA3 LMN, TESCAN, Czech Republic). Images were acquired at 5 kV at random locations on the samples. To observe the cross section of the scaffold, samples were prepared with brittle fracture in liquid nitrogen.

Fourier transform infrared spectroscopy (FTIR): Samples were fixed on a flat diamond testing bench. Analysis was conducted utilizing an FTIR spectrometer (Nicolet iS50, Thermo Scientific, USA). The wavenumbers examined ranged from 400 to 4000 cm. Functional groups were identified by comparing the obtained spectra with the previously reported []. −1 ²⁹

X-ray diffraction (XRD): Samples were cut into 1 × 1 cmpieces and examined by XRD (MINIFLEX600, Rigaku, Japan) using Cu-Kα radiation under atmospheric conditions. Data were collected at a step width of 0.2°, duration of 1.2 s, wavelength of 1.5418 Å, and 2θ values of 5–90°. The phase structure was identified by comparing the diffraction patterns with the ICDD standards (JCPDS) []. The average grain size of PCL was estimated by the Scherrer Method []: 2 ^{30 31}

where κ, λ, β and θ represent the constant shape factor (0.9), X-ray wavelength, the width at half maximum and the diffraction angle, respectively.

Thermogravimetric analysis (TGA): Thermal stability (10 mg per sample) was performed using a thermogravimetric analyser (STA 2500, NETZSCH, Germany). Testing was performed under argon gas flow (100 mL min), at a heating speed of 10°C minfrom room temperature to 600°C. A derivative thermogravimetric curve was obtained by first-order derivative processing of the TGA curve. −1 −1

Differential scanning calorimetry (DSC): Material melting and crystallization were determined using DSC (DSC 300, NETZSCH, Germany). Testing was performed in an air environment with a temperature range of 20–90°C and heating rate of 10°C min. The crystallinity of the scaffold was determined by comparison with the melting enthalpy (139.5 J g) of 100% crystalline PCL []. −1 −1 ³²

Water contact angle (WCA): Surface wettability was examined using the built-in function of a contact angle analyser (SDC-200S, ShengDing, China). A 2-µL water droplet was used. Images were taken at 0, 2, 4, 6, 8 and 10 min after contact with the sample surfaces. For anisotropic samples, testing was performed along and across the surface structures. For each group, six samples were used, and at least three random locations of each sample were measured.

Mechanical testing: The rod-like testing samples were prepared at a length of 1 cm and a diameter of 3 mm. By pressuring the sample into a rectangular shape, the thickness was measured using a digital micrometre (0–25 mm, Meinaite, Germany). Testing was performed at a loading speed of 10 mm minand a force of 100 N using a tensile testing machine (HYC-2011, Hongjin, China). Data were analysed using the Origin software (Origin9.1, OriginLab, USA). For the sample exhibiting a yield phenomenon, a low yield point was employed to ascertain its yield stress and strain. For the left side, an offset strain of 0.5% was used to determine the yield stress and strain. Five samples from each group were analysed. −1

Degradation testing—Samples at 1 cm in length and 3 mm in diameter were prepared and immersed in 10 mL of NaOH solution (3 mol L) at room temperature. The initial weight of the sample was denoted as W. Optical images were captured at each time point before the solution was removed. The samples were washed three times with distilled water and dried for weighing (W). The degree of sample degradation was evaluated based on the mass loss determined as follows: −1 d

At each time point, five samples of each group were analysed.

Human foetal bone marrow MSCs (HUXMF-01001) were purchased from Cyagen Biotechnology Co., LTD (China). The cells were cultured in DMEM (1% PS + 10% FBS) in flasks (Gibco, USA) in a COincubator (BB50, Thermo Fisher Scientific, USA) at 37°C. The culture medium was replaced every three days. Cells were digested using 0.25% trypsin for passaging when the cell density reached 90%. For cell studies, all material samples (1 cm in length and 3 mm in diameter) were disinfected by soaking in a 75% EtOH solution for 12 h. For all studies, a 300-µL cell suspension was used for seeding on each sample (seeding density: 30 k/sample). After 6 h of incubation for the initial adhesion, 1 mL of D10 medium was added to each sample. The cells were used before passage 6. 2

The cellular adhesion and morphology were characterized using SEM. Briefly, the cells were cultured for 1 and 14 days after seeding. The cells were fixed with PFA (4 wt.% in PBS, 15 min), dehydrated by gradient EtOH solutions (30, 50, 70, 90, 95 and 100%, 15 min) and then surface gold coated for SEM characterization.

Cytotoxicity was evaluated using live and dead cell staining. After culturing for pre-determined periods, MSCs were incubated with FDA (4 µg mLin PBS) and PI (1 µg mLin PBS) solutions for 8 and 5 min, respectively. The cells were then washed thrice with PBS and examined using a confocal laser scanning microscope (CLSM, FV1000, Olympus, Japan). Live and dead cells were identified as FDA-labelled green and PI-labelled red, respectively. −1 −1

The metabolic activity was tested to evaluate cell proliferation. After culturing for 1, 3, 7, 14 and 21 days, the cell medium was removed, and the cells were incubated with CCK-8 testing medium (1:10 dilution in D10) for 2 h at 37°C and 5% CO. The supernatant was analysed with a microplate reader (2300, EnSpire, Singapore) at a wavelength of 450 nm. Four samples from each group were collected at each time point. 2

Tendon-associated gene expression was determined by a quantitative reverse transcription polymerase chain reaction (qRT-PCR, Q711-02, Vazyme, China). Briefly, MSCs were cultured on the scaffold for 5 and 10 days. PCLand PCL(90°) were used as the controls of the scaffold's shell and core, respectively. Extraction of total cell RNA, cDNA synthesis and real-time detection were performed using TRIzolSolution, HiScriptIII RT SuperMix and ChamQ Universal SYBR qPCR Master Mix, respectively. The primers specific to interested genes including scleraxis (Scx), decorin (DCN), tenascin-C (TNC) and COL-I were summarized inand provided by Sangon Biotech (China) []. Relative gene expression level was determined by comparing the quantified cDNA transcript level to that of GAPDH. The expression levels of all groups were then normalized to the group of cells before seeding on materials, which was set as 1. Triplicate detections were performed for each sample, and three samples were used for each group. Fm Fb TM ® Supplementary Table S1 ³³

The expression of major tendon matrix protein (COL-I) was examined by immunofluorescence staining. Briefly, MSCs were cultured on the scaffold for 5 and 10 days. PCLand PCL(90°) were used as the controls of the scaffold's shell and core, respectively. After fixation, permeabilization and blocking, the cells were incubated in sequence with mouse-derived anti-human primary monoclonal antibody (1:200 dilution of anti-COL-I in 1 wt% BSA solution) and 594 Goat Anti-Mouse IgG (H + L) antibody (1:200 dilution in 1 wt% BSA solution). Cells were finally incubated with DAPI for nucleus visualization. Images were taken using CLSM with identical parameters for all groups. Fm Fb

Tendon defect model: Animal procedures were approved by the Committee of Animal Experimental Ethical Inspection, Zhejiang University School of Medicine (Ethics Number: 2022-676). Briefly, 8-week-old male rats (320 ± 20 g) were randomly divided into three groups: the tendon defect group (TDG, six rats), the material treatment group (MTG, six rats) and the Sham group (Sham, two rats). Before surgery, all animals were fasted for 12 h, followed by an intraperitoneal injection of sodium pentobarbital (50 mg mL, 30 mg kg). To facilitate the operation, the rats were placed in a supine position. A 3-cm mid-lateral incision was made to expose the Achilles tendon. Using sharp scissors, a half-thickness transection of the Achilles tendon was performed, creating a defect measuring 0.5 cm in length. In the MTG procedure, the defect was semi-sutured using the fabricated tendon scaffold, after which the incision was closed using sutures. In the TDG procedure, the defect was left untreated before incision closure. Sham surgery served as a Sham procedure, with no Achilles tendon injury. The Kessler suture method was employed using absorbable suturing lines. For sample collection in each group, half of the animals were euthanized at 4 weeks, and the remaining animals were euthanized at 8 weeks. −1 −1

Gross observation: After euthanasia, the rats were placed in a supine orientation. A 5-mm incision was made along the right hind leg to fully expose the Achilles tendon. Gross observations of the animal models were conducted by the same surgeon, documenting any changes (including width, swelling, texture, and adhesion to the surrounding skin) in the tendon.

Histological analysis: The harvested Achilles tendon tissues, along with the materials, were fixed in 4% buffered formalin (pH 7.4) and subsequently embedded in paraffin blocks. Four-micrometre slices were prepared and stained using Hematoxylin-Eosin, Sirius Red and Masson's Trichrome Staining Kits, following the manufacturer's instructions. Pathological scoring was performed using a light microscope (BX51, Olympus, Tokyo, Japan) by two pathologists who were blinded to the experimental design. Tissue evaluation was based on the Tang score for tendon adhesion assessment and the Soslowsky, Svessen and Cook (SSC) score for tissue regeneration evaluation () []. Supplementary Table S2 ³⁴

Second harmonic generation (SHG) characterization: Pathological specimens from each group were imaged using two-photon microscopy (FVMPE-RS, Olympus, Tokyo, Japan) with an excitation wavelength of 850 nm. Images were captured at a resolution of 1024 × 1024 pixels and a pixel resolution of 4 µm through a 420–460 nm filter. SHG fluorescence of collagen fibres was observed in both injured and uninjured regions. Semi-qualitative analysis was conducted using ImageJ (NIH, MD, USA) to evaluate the density, organization and mean grey values (MGV) of the collagen fibres.

Data are expressed as the mean ± SD unless otherwise specified. Statistical comparisons between two groups were performed using Student's paired test with a two-tailed distribution in Excel (Microsoft Office 2021, USA). A value of < 0.05 is considered to be significantly different. P

We designed a core-shell scaffold for tendon rupture repair (). The scaffold showed a rod-like core composed of unique fibres for the guidance of cell and matrix alignment. The fibres printed by the EHD technique are tailorable in terms of structural geometry, which allows for controlled porosity and anisotropy. Around the core, a shell structure is designed to provide scaffold mechanical support as well as penetrated channels for nutrition transport and cell and vessel migration from peritendinous tissue. Optical images show a scaffold with an integrated diameter of 3 mm and a length of 70 mm assembled from the core and shell portions. Figure 1A

shows optical and SEM images of the scaffold. It is assembled by inserting the core into the shell and securing it firmly through structural relaxation of the core layers. The space formed by the shell was occupied by the printed fibres, which exhibited numerous interconnected pores that extended into the through holes of the shell. The porosity of this core-shell scaffold was calculated to be approximately 80%. This satisfies the requirement for adequate neotissue ingrowth and may facilitate the transport of essential nutrients and removal of toxic byproducts of cell metabolism. Figure 1B

Additional structural details of the core-shell scaffold are shown in. The fibres of the core are arranged periodically by EHD direct writing, forming rhomboids at intervals of 500 µm and at an angle of 15° as pre-determined. This leads to the well-known geometric anisotropy of scaffolds for contact cell guidance. PCL spheroids were observed on individual fibres, with material consolidation occurring at the interface between the upper and underlying fibres. Concurrently, the upper fibres exhibit morphological adaptability but not complete fusion to the underlying fibres. This permits interconnectivity among the rhomboidal units of the rolled core laminates. In contrast, owing to fibrous segment collapse as observed in existing 3D printed materials,continuous layer-by-layer deposition of the upper fibres would segregate the rhomboidal units completely, leading to restricted mass exchange in tissue engineering applications. Figure 1C in situ

SEM images of the shell exhibit multiple surface structures comprising staggered through-holes and non-hole regions with unique micro-ridges/grooves. The direct laser perforation of through-holes resulted in enlarged areas after uniaxial stretching (). This may have facilitated transverse mass transportation from the shell to the core. Meanwhile, micro-ridges/grooves were observed in the non-hole regions, forming a high geometric anisotropy along the scaffold length (). In our previous studies, it is suggested that mechanical stretching resulted in PCL recrystallization, and the structural deformation caused crystal exposure to the shell surface as ridges and the amorphous regions to depress into grooves []. Here, in regions near the through-hole edges, we observed a reduction in the formed inter-ridge-distance, probably because of the experience of a larger Poisson's contraction []. Although the shell exhibits surface micro-ridges/grooves with enlarged through-holes, the nature of the open pores negatively affects the mechanical strength of the scaffold []. However, we expect that the strain strengthening induced by uniaxial stretching can compensate for the weakening effect of through-holes. Supplementary Figure S1 Figure 1C ^{32 32 27}

shows the characteristic peaks of C–H, C=O, and C–O stretching vibrations at 2900, 1800 and 1400 cm, respectively []. This indicates that there were no chemical group changes in PCL during the electrohydrodynamic and laser processes. As shown in, the characteristic peaks of the PCL XRD spectrum were observed at 2θ = 21.4°, 22.0° and 23.8°. These peaks correspond to the PCL crystallographic planes of (110), (111) and (200), respectively []. Owing to the breakage of the PCL molecular chain by stretching, the characteristic peak at 2θ = 22.0° disappeared in the UX-PCL group, with significantly reduced grain size compared to the other groups () []. Figure 2A ²⁹ Figure 2B ³⁰ Supplementary Figure S2 ³² −1

The thermal stability of the scaffold is shown inand, which show a similar single-step decomposition for all the groups. Compared with raw PCL, the scaffold core and shell exhibited reduced initial decomposition temperatures, suggesting that the fabrication procedures affected the thermal stability of the composite. Consistent with this observation, the scaffold reached maximum decomposition at lower temperatures than raw PCL, further indicating alterations in thermal stability due to the scaffold fabrication processes. This leads to an evident early mass loss in the core-shell scaffold. Further comparisons indicated that the core had better thermal stability than the shell. This may be attributed to uniaxial stretching, which induces molecular chain breakage and thereby lowers the resistance of the material to thermal decomposition. Figure 2C Supplementary Figure S2

exhibits the thermo curves of the scaffold core and shell. Compared to raw PCL, the endothermic peak width of PCL became narrow, with reduced melting points observed for the core and shell. Furthermore, the core exhibited a higher degree of PCL crystallinity, indicating enhanced recrystallization of the PCL molecular chains during the EHD writing process of the PCL. This is consistent with the formation of PCL spheroids on the printed fibres (). Conversely, the shell exhibited a decrease in crystallinity (). Our previous study demonstrated that uniaxial stretching of PCL at a temperature close to its melting point, followed by cooling at room temperature, increases PCL crystallinity []. In this study, without any thermal treatment, our findings suggest that uniaxial cold stretching may lead to the degradation of PCL molecular chains, resulting in a reduction of the melting point and crystallinity. It is also valuable to note that the melting point of a polymetric material is not solely an intrinsic property, and it is influenced by extrinsic factors such as material structure, environmental conditions, processing history, impurity and experimental setup []. This explains why the change in melting point is inconsistent with that in crystallinity, due to the different processing for PCL materials. Figure 2D Figure 1C Figure 2D ^{32 35}

To assess the compliance of the fabricated core-shell scaffold for tendon applications,presents the tensile stress–strain curves. The scaffold incorporated the characteristics of both the core and shell, resulting in distinct toe and elastic regions, followed by plastic deformation. As the strain increases, the toe region is characterized by a slow, nonlinear rise in stress. This is primarily due to the EHD-printed fibre patterns of the core, where the angled fibres (15°) were able to align gradually, although the shell alone showed a minimal impact on the toe region. The observed toe region, ranging from 1-3% strain, is consistent with that of the natural tendons serving to buffer the abrupt tensile strain applied to the scaffold []. In previous studies, curved fibres have been investigated to mimic the mechanical behaviour in the toe region []. In comparison, the angled fibres in this study demonstrate a notable resilience, easily returning to their original state after experiencing tensile forces. Owing to their limited loading capability, after alignment, bioinspired fibres are prone to plastic deformation. However, with a reinforced shell, the core fibres can be protected to avoid overloading, thereby significantly enhancing the overall mechanical stability. Figure 3A ^{13 17}

In the elastic region, the scaffold demonstrated an intermediate behaviour between that of the core and shell. Due to the reinforcing effect of the shell, the scaffold achieved a yield stress 2.1 times and Young's modulus 2.4 times greater than those of the core (). This reinforcing effect is due to strain hardening, which offsets the weakening effect caused by hole enlargement in the shell. Compared to the reported scaffolds featuring core-shell (∼1–8 MPa) [], braided (∼5–121 MPa) [] and knitted (∼24–34 MPa) [] types, the resulting Young's modulus of the scaffold in this study was superior and approached more that of the human Achilles tendon (217–1200 MPa) []. It is also worth noting that the yield stress of the core is ∼5 MPa, above which the EHD-printed fibres alone will be subjected to plastic deformation and lose their elastic properties. In this study, with the core-shell structure, the fabricated scaffold can achieve a higher level of yield stress because of the shielding effect of the shell bearing the most excessive loading and protecting the core from excessive deformation. Table 1 ^{26 36 36 3}

In contrast to the core, the scaffold with a small strain increase displayed a fast stress increase in plastic deformation (), which is similar to that of the natural tendon []. This is attributed to the reinforcing effect of the shell due to its strain-hardening process by uniaxial stretching []. Consequently, the scaffold achieved a fracture stress 4.3 times greater than that of the core (). The ultimate fracture load of the scaffold (1 cm in length and 3 mm in diameter) is 78.5 N, which is larger than the sum of 13.4 N (the core) and 51.1 N (the shell,). This indicated a synergistic reinforcement effect between the core and shell, leading to a scaffold capable of withstanding high tensile loads. Finally, the fabricated scaffold exhibited a two-stage plastic deformation until fracture, corresponding to the preferential deformation in the shell and later deformation in the core. Figure 3A ^{37 17} Table 1 Supplementary Table S3

shows the WCA measured along and across the surface structures of the scaffold. The WCA along the surface geometry was much lower than that across the structures, demonstrating higher surface wettability. This is because of the stretching-induced micro-ridges/grooves. With prolonged contact, the anisotropic wettability was well maintained, indicating improved surface wettability with a decreased WCA. This is probably because the hydroxyl groups in PCL interact with water molecules and facilitate their adsorption on the surface. A comparative study indicated similar surface wettability between the scaffold and shell. In contrast, isotropic surface wettability was observed for the core, and the WCA was much higher than that of the scaffold and shell. As shown in theinsets, the water droplet demonstrates an approximately circular shape on the core surface, which may be because it covers multiple fibres and air gaps in all directions. This leads to the isotropic surface wettability observed for the core. Figure 3B Figure 3B

As shown in, the scaffold underwent sequential degradation from the core to the shell. After 21 days of degradation, the core preferentially lost its structural integrity and became entangled fibres. After 50 days of degradation, it completely broke down into debris. In contrast, the shell displayed significantly improved properties against degradation and maintained its structural integrity over the investigated hydrolysis period of 70 days. It is evident that PCL after uniaxial stretching follows surface-controlled degradation, because the improved polymer chain recrystallization and densification make it difficult for erosive liquids to penetrate the polymer surface []. In contrast, bulk degradation accounts for the loss of structural integrity, as observed for the spheroid structure in the core. In line with these results, the core exhibited a mass loss rate that was much higher than that of the shell, leading to approximately 100% degradation in the core on day 30 and only approximately 10% degradation in the shell until day 70. The scaffold achieved approximately 60% and 80% mass loss on days 30 and 80, respectively. This is because of the preferential degradation in the core, which provides space for the growth of new tissues. These results demonstrate the prolonged structural support of the scaffold, which compensates for the mechanical weakness of the core and ensures security before new tissue formation of the damaged tendon. Figure 3C ³⁸

shows fluorescence images of MSCs stained with FDA/PI. The predominance of FDA-labelled live cells was observed on the scaffold's shell. The porous structure of the shell facilitates cell migration into the core, enabling efficient nutrient exchange and supporting cellular viability. On day 1, some red-labelled dead cells were observed in the scaffold's core, which may be due to the entrapment of damaged cells during the seeding process. In line with this, after prolonged culture for 14 days, an increased cell population with no dead cells was observed. The details of the cellular proliferation are shown in. Compared to the core, intimal cell growth was enhanced on the shell, which accounted for higher levels of cell proliferation over the investigated period. This is probably because of the high porosity of the core, which causes cell loss during the seeding and subsequent adhesion stages. The scaffold exhibited a trend of cell proliferation similar to that of the core, which was elevated during the laterculturing period of day 7–21. Figure 4A Figure 4B in vitro

SEM images further indicated that the cells could elongate along the topographical structures of both the core and shell during the very early culture period (day 1,). Pseudopodia were mainly observed at the cell ends. The geometric curvature of individual fibres in the core and ridges in the shell provides contact guidance on each cell, whereas the overall anisotropy of these topographies accounts for the control of the cell alignment in a consistent direction over the scaffold. Prolonged culture results in cell proliferation. In the core, the cells tended to bridge different fibres and penetrate into the space among those fibres, as observed on day 21. On the shell, the cells exhibited nearly 100% coverage of the non-porous surface. In these observations, overall cell alignment was maintained along the surface structures. The cells in the core appeared to have less spreading morphology than those on the shell. Our results demonstrate the proper microenvironment of the tendon scaffold with no obvious cytotoxicity and good cell proliferation. Figure 4C

shows the positive expression of tendon-related genes in human MSCs, including Scx, DCN and TNC corresponding to the early- to middle-term tenogenic markers []. Compared to the Ctrl-PCLgroup, cells cultured for 5 days with the scaffold's shell showed higher expression levels of Scx and DCN and a lower level of TNC, although significant differences are not observed. In line with the above observations, the scaffold's core exhibited a lower expression level of Scx in MSCs and higher levels of DCN and TNC compared to those of the Ctrl-PCLgroup, respectively. Prolonged culturing promoted the gene expression in MSCs, shifting towards the more mature tenogenic markers. The higher expression levels of DCN and TNC were observed on the scaffold's shell than in its control group. A similar trend was obtained in the scaffold's core, where the expression levels of DCN and TNC remained higher than those of the control group with statistical differences. Figure 5A ¹³ Fm Fb

The expression of a major tendon matrix protein COL-I was also investigated, which at the genetic level showed comparable levels in groups of the scaffold's shell and its control. Differently, the scaffold's core obtained a higher level of COL-I expression than its control group. With the prolonged culturing of 10 days, the trend on COL-I expression remained unchanged in the scaffold's shell, while a declined expression level was observed in the scaffold's core.shows confocal fluorescence images of human MSCs labelled with COL-I and nuclear DNA. It was observed that at the protein level, the cells after 5 days of culturing showed obviously enhanced COL-I expression in both the scaffold's core and shell compared to their control groups, respectively. Cultured on the scaffold's components, the cells obtained directional growth toward the geometric anisotropy of the scaffold. Figure 5B

illustrates the animal modelling process. A 5-mm defect was created after exposing the Achilles tendon. In the MTG procedure, the defect was repaired using a scaffold, whereas in the TDG, the defect was repaired via direct suturing. The Sham group underwent no defect creation and served as the positive control group. As shown in, the scaffold exhibited suitable properties for suturing. Gross observation revealed scar tissue formation at the tendon gapping site in all injury groups, with varying degrees of adhesions, except in the Sham (). Similar tendon adhesion phenomena have been reported previously using ECM scaffolds []. However, the extent of scar tissue formation in the MTG was significantly reduced compared to that in the control of TDG, with longitudinal ingrowth of collagen fibres in the scaffold, indicating improved biomechanical properties resembling those of the normal tendon structure. Figure 6A Supplementary Figure S3 Figure 6B ³⁹

In terms of peritendinous tissue proliferation, the control group of TDG exhibited more prominent neo-angiogenesis; however, the newly formed vessels intersected with disorganized collagen. In the TDG, fibrosis of the injured tendon site was more pronounced, with the peritendinous tissue appearing rough and resilient, accompanied by significant tendon retraction. This indicates severe adhesion to the surrounding tissue. Peritendinous adhesion in the MTG was significantly reduced, and tendon regeneration was restricted within the implanted scaffold. The Tang score indicates that at week 4, both the TDG and MTG exhibited similar levels of peritendinous adhesion. However, by week 8, tendon adhesion improved in the MTG, whereas tendon adhesion worsened significantly in the TDG (). Figure 6C

Under the microscope, the control group of TDG also exhibited a high degree of disorganized fibre arrangement, with disrupted collagen orientation and larger interfibre gaps (), characteristic of unbenefited tendon remodelling. This was improved with scaffold implantation, where the regenerated collagen fibres were aligned in a pre-designed, linear pattern within the scaffold. Masson's and Sirius Red staining further confirmed enhanced collagen synthesis, improved fibre arrangement, and increased collagen content within the scaffold. The undegraded material of the scaffold caused porous gaps between the newly generated tendon tissues, suggesting the need for improved degradation capability. Figure 6B

From a cellular perspective, both TDG and MTG showed infiltration of adipose tissue around the newly formed collagen, along with an accumulation of fibroblasts and inflammatory cells. Neo-angiogenesis and cartilaginous formation were most pronounced in the control group and increased as the time post-injury proceeded (). In contrast, the MTG exhibited reduced cellularity along with significantly lower fibrous formation, vascularity and cartilaginous tissue, suggesting that it is a promising approach for Achilles tendon repair. The biomimetic structure of the tendon scaffold supports cellular activities that are essential for tissue regeneration. Figure 6D

Using a two-photon microscope, SHG analysis revealed a regenerated collagen structure at higher resolution and sensitivity (). In the TDG, the newly formed collagen structures appeared thin and disorganized, with several longitudinal defects in the middle section of the tendon at both 4 and 8 weeks. These findings suggest that the tissue has a diminished capacity to withstand mechanical stress. In contrast, the MTG demonstrated distinct longitudinal infiltration of collagen fibres into the tendon defect area within the scaffold at both 4 and 8 weeks. Signal intensity was significantly greater at week 8 compared to both TDG and Sham, suggesting enhanced collagen regeneration. Notably, the signal intensity in the TDG was significantly lower than in the MTG, and the newly formed collagen appeared thicker and more prominent. MGV analysis provided a comprehensive assessment of the collagen content and distribution (). In the TDG, the MGV was generally lower than in the MTG at both weeks 4 and 8, and became more dispersed by week 8. In contrast, the MTG exhibited lower but more organized MGV in a linear pattern at week 4, which became more pronounced around the scaffold by week 8, suggesting enhanced tendon regeneration and organization facilitated by the scaffold. Figure 6B Supplementary Figure S4

One limitation of our present study is the absence of comprehensive biomechanical testing on thescaffold constructs. Tendons are subjected to repetitive stress during movement, making fatigue resistance a critical factor in long-term function and surgical repair success. In addition, suture pull-through resistance is also a critical property of tendon graft, which resists tearing and pulling out of the tissue under tensile forces. Although material characterizations andassessments provided valuable insights of the scaffold's mechanical properties such as Young's modulus, elasticity and tensile strength, which are critical for tendon repair applications, future studies are needed to include systematic biomechanical evaluations for more accurately assessing the scaffold's functional performance and its ability to withstand physiological loads. It is also valuable to note that numerous regulations affect the process from DNA transcription to successful deposition of a tendon matrix protein. This results in the varied expression trends of a gene at the transcription and protein levels. Compared to tenogenic growth factors (e.g. bone morphogenetic protein and basic fibroblast growth factor), the remarkable performance of geometric cues as developed in our core-shell scaffold is more likely to induce cell morphogenesis into alignment, thereby resembling the structure and mechanical properties of natural tendon []. In our study of tendon regeneration, the tissue adhesion was evaluated by the Tang score. However, adhesions occur when scar tissue forms between a tendon and its adjacent tissues, restricting gliding motion and leading to stiffness, pain, and functional impairment. Therefore, balancing adhesion prevention vs. tendon healing strength remains a clinical challenge. in vivo in vitro in vivo ¹⁷