Spatially distributed bFGF and VEGFA gene delivery to tendon and peritendinous tissues enhances early tendon healing

Abstract

Tendon healing is limited by inadequate vascularity and low cellularity within the tendon, whereas surrounding peritendinous tissues include vascular, stromal, and extrinsic reparative cells. Building on prior work involving tendon-only co-delivery of basic fibroblast growth factor (bFGF) and vascular endothelial growth factor A (VEGFA), this study investigated whether spatially expanding gene delivery to both the tendon and adjacent peritendinous subcutaneous tissue would enhance early repair. PLGA/PEI nanoparticle complexes carrying bFGF and VEGFA plasmids were characterized for morphology, particle size, plasmid complexation, surface charge, and in vitro release, with local reporter expression assessed in vivo. Therapeutic effects were evaluated in chicken digital flexor tendon and rat Achilles tendon injury models through non-injection, tendon-only, peritendinous subcutaneous-only, and combined injection groups. The complexes demonstrated effective plasmid complexation, sustained release, and detectable local EGFP reporter expression within both tissue compartments. Combined injection significantly increased the ultimate strength of chicken tendons at 2, 4, and 6 weeks, and rat Achilles tendons at 2 weeks, compared to non-injection controls; tendon-only injection did not yield similar improvements in early strength. Higher total local bFGF and VEGFA protein levels, more visually prominent Ki-67-positive staining and CD31-positive vascular structures, together with qualitatively more organized collagen morphology. These results indicate that expanding gene expression from the tendon proper to the tendon-peritendinous microenvironment can mitigate an early limitation of tendon-only delivery, supporting the concept of enhanced dual-compartment delivery, though dose-matched studies are needed to establish synergy.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s11671-026-04885-z.

Introduction

Tendon injuries frequently result from exercise or daily activities [1–3]. However, healing of the injured tendon is hindered by inadequate growth factors and tendon cells, leading to prolonged recovery times that significantly impact patients’ quality of life [4, 5]. Consequently, maintaining a continuous supply of growth factors to facilitate tendon healing has become a primary research focus. Numerous studies have employed direct injection or collagen sponges to deliver growth factors for tendon repair [6–8]. Due to the susceptibility of growth factors to degradation by tissue proteases, direct delivery often yields little to no therapeutic benefit.

Gene therapy represents a promising strategy to translate target genes into proteins, effectively extending the availability of therapeutic proteins [9–11]. A variety of vectors, including viral vectors, liposomes, and other non-viral vectors, have been developed to enhance gene delivery [12–14]. Viral vectors are widely utilized due to their high delivery efficiency; however, their application is limited by challenges such as high production costs, cellular immune responses, insertional mutagenesis, and instability [10, 15]. In recent years, non-viral vectors have gained attention for their high biocompatibility and relative safety [16, 17]. Our team has developed polylactic-co-glycolic acid (PLGA) nanoparticles with sustained-release properties and favorable biocompatibility, which can effectively transfect genes into cultured primary tendon cells and tendon tissue cells [18–20].

Previous studies indicated that enhancing the expression of basic fibroblast growth factor (bFGF) and vascular endothelial growth factor A (VEGFA) in injured tendons promotes repair [12, 18, 21]. In a 2018 investigation, nanoparticle-mediated co-delivery of bFGF and VEGFA was limited to the tendon proper and primarily improved tendon strength during the middle and late phases of healing; it did not significantly enhance ultimate strength at 1 or 2 weeks post-injury [18]. This early limitation is clinically relevant, as sufficient repair strength is necessary for controlled mobilization, which is critical to preserving tendon gliding and supporting functional recovery [22, 23]. The restricted early response may arise not only from the low cellularity and dense extracellular matrix of the tendon proper [24, 25] but also from the limited number and uneven spatial distribution of cells available for plasmid uptake and sustained protein production following injury.

Tendon repair involves not only intrinsic cells within the tendon but alsoextrinsic cells originating from the tendon sheath and surrounding peritendinous tissues [26]. Both the tendon proper and peritenon house distinct progenitor cell populations [27], while perivascular-derived stem/progenitor cells can be recruited during the healing process [28]. Structural studies indicate that vascular and vascular-associated cell populations are more abundant in the paratenon and interfascicular matrix compared to densely collagenous fascicles [29]. These findings imply that the surrounding peritendinous subcutaneous tissue harbors a wider pool of transfectable stromal, vascular-associated, and progenitor cells, potentially serving as an additional local source of therapeutic factors. However, since extrinsic repair may also lead to fibrosis and adhesion, activation of this compartment should be considered in conjunction with tendon organization and adhesion outcomes [26, 30].

Thus, the hypothesis posits that intratendinous and peritendinous injections provide complementary rather than redundant functions. Intratendinous delivery aims to support resident tendon cells and intrinsic matrix repair, while delivery to the peritendinous subcutaneous tissue seeks to expand the number and spatial distribution of growth-factor-producing cells and generate paracrine signals across the tendon-peritendinous interface. This study addresses the early-stage limitations of the previous tendon-only strategy from 2018 rather than re-establishing the efficacy of bFGF/VEGFA co-delivery. The novelty of the current research lies in optimizing the spatial distribution of an established PLGA/PEI-based gene-delivery platform rather than developing a new nanoparticle formulation or gene-therapy vector. Treatments were compared across tendon-only, peritendinous subcutaneous-only, combined, and non-injection groups in chicken digital flexor tendon and rat Achilles tendon injury models (Fig. 1), with assessments of mechanical strength, local reporter expression, total local bFGF/VEGFA protein levels, Ki-67 and CD31 staining, collagen organization, and adhesion formation.

Fig. 1.

Schematic illustration of the nanoparticle preparation, spatial delivery strategy, and proposed mechanism for tendon repair. PLGA nanoparticles were fabricated using a double-emulsion method, surface-modified with polyethyleneimine (PEI), and complexed with plasmids encoding basic fibroblast growth factor (bFGF) and vascular endothelial growth factor A (VEGFA). The resulting nanoparticle/plasmid complexes were injected into the injured tendon and/or surrounding subcutaneous tissue in chicken digital flexor tendon and rat Achilles tendon injury models. Following local cellular uptake, released plasmids are proposed to support increased local bFGF and VEGFA availability, which may contribute to cell proliferation, vascular responses, collagen remodeling, and recovery of tendon mechanical strength

Methods

Preparation and architecture of plasmids

The open reading frames of the bFGF gene (NM_002006) and VEGFA gene (NM_001171626) were cloned into the pEGFP-N1 backbone (GeneChem, China) to create overexpression plasmids pEGFP-bFGF and pEGFP-VEGFA. The identities and orientations of the inserted sequences were confirmed through DNA sequencing. In these constructs, EGFP was expressed independently of the therapeutic coding sequences for bFGF and VEGFA. Simplified schematic maps of the pEGFP-bFGF and pEGFP-VEGFA constructs, which show the promoter, therapeutic coding sequence, EGFP expression cassette, polyadenylation signal, and relevant backbone elements, are illustrated in FigureS1. EGFP fluorescence served solely as a qualitative reporter of local plasmid expression and distribution, without serving as a quantitative surrogate for bFGF or VEGFA protein production.

Preparation of nanoparticle/plasmid complexes

The double emulsion (water/oil/water) technique was utilized to prepare PLGA nanoparticles, following a modified version of a previous method [31]. Briefly, 200 mg of PLGA (lactide: glycolide ratio, 65:35; molecular weight, 40 000–75 000; acid-terminated; catalog no. P2066; Sigma-Aldrich, St. Louis, MO, USA) was dissolved in 2 mL of dichloromethane and subsequently emulsified in 6 mL of a 7% (w/v) PVA (poly (vinyl alcohol), Sigma Aldrich) solution using an ultrasonic homogenizer (Sonoplus HD 2070, Bandelin electronic, Berlin, Germany) for 2 min to form the primary emulsion. This primary emulsion was then added to 100 mL of a 1% (w/v) PVA solution and further emulsified for 3 min to generate the double emulsion. Following 2 days of stirring, residual dichloromethane was completely removed from the double emulsion. Nanoparticles were collected via centrifugation (13,000 rpm, 5 min) and washed twice with deionized water. The nanoparticles were then dispersed in deionized water and surface-modified with branched polyethyleneimine (bPEI; molecular weight, ~ 25000 by LS; catalog no. 408727; Sigma Aldrich) at a PLGA-to-PEI mass ratio of 20:1 (w/w). Subsequently, PLGA nanoparticle/plasmid complexes were formulated with various nitrogen-to-phosphate (N/P) ratios. Specifically, PEI solution was combined with the PLGA nanoparticle solution, followed by the addition of plasmid solutions and incubation for 10 min at room temperature to form nanoparticle/plasmid complexes. Preparations used in this study were derived from 3 independent batches. The nanoparticle/plasmid-complex suspension had a nanoparticle-equivalent concentration of 5 µg/µL. Stock suspensions of nanoparticle-plasmid complexes were stored under sterile conditions at 4 °C and used within two weeks of preparation. Nanoparticle-plasmid complexes were prepared in three independent preparation batches.

The characterization of nanoparticle/plasmid complexes

The morphology of the nanoparticles or nanoparticle/plasmid complexes was examined using scanning electron microscopy (SEM, Hitachi, S-3400 N, Tokyo, Japan). The hydrodynamic diameter of unloaded PLGA nanoparticles and nanoparticle/plasmid complexes were assessed using dynamic light scattering (DLS) with a Brookhaven BI9000AT system (Brookhaven Instruments Corporation, Austin, TX, USA). Measurements were conducted on three independently prepared nanoparticle batches, with three technical measurements obtained for each batch. The mean values of the technical replicates were calculated for each batch, and results are presented as the mean ± SD of three independent batches (n = 3). A gel retardation assay assessed the interactions of the gene with the nanoparticles across different N/P ratios. Briefly, the nanoparticle/plasmid complexes underwent electrophoresis in a 1% Tris-acetate-EDTA agarose gel at 120 V for 20 min. Images were captured using a Gel Imaging System (SynGene, Cambridge, UK). The zeta potential of PLGA nanoparticles, PEI-modified nanoparticles, and nanoparticle/plasmid complexes was measured using a zeta-potential analyzer (90 Plus, Brookhaven Instruments, Holtsville, NY, USA).

In vitro release of gene from the nanoparticles

Nanoparticle complexes containing bFGF or VEGFA plasmids were suspended in phosphate-buffered saline (PBS, pH 7.4) and incubated at 37 °C with continuous agitation. At predetermined time points, samples were centrifuged, and 50 µL of the supernatant was collected to quantify plasmid DNA using a Biophotometer (Eppendorf, Germany). After each sampling, an equal volume (50 µL) of fresh PBS was added to maintain a constant volume of the release medium. Release experiments were conducted in triplicate, and cumulative plasmid release was calculated as a percentage of the total amount of plasmid initially incorporated into the formulation.

The injection of nanoparticle/plasmid complexes into tissues in vivo

White Leghorn chickens and Sprague-Dawley (SD) rats were purchased from the Animal Center of Nantong University. The digital flexor tendons of the chickens and Achilles tendons of the rats served as the experimental tissues. For the chicken tendon injury model, 104 completely transected digital flexor tendons from 52 White Leghorn chickens were randomized into four groups: non-injection, tendon injection, subcutaneous tissue injection, and combined tendon and subcutaneous tissue injection. Bilateral long toes of each chicken were used for the study. A total of 26 toes were used for each of the above groups. The chickens were anesthetized via intramuscular injection of 3% pentobarbital sodium, and elastic bandages were applied to the legs as tourniquets prior to the operation. The procedure was performed under sterile surgical conditions. The area between the proximal and distal interphalangeal joints, corresponding to zone II flexor tendons in humans, was identified as the ideal operation site. A zig-zag incision was made in the plantar skin, and the sheath was longitudinally opened for approximately 1.5 cm to expose the flexor tendon system. The flexor digitorum superficialis tendon was resected for about 1.5 cm, and the flexor digitorum profundus (FDP) tendon was completely transected. Nanoparticle/plasmid complexes were injected into the completely transected FDP tendons and/or surrounding peritendinous subcutaneous tissues. Each tendon received an injection of 20 µL of nanoparticle/plasmid complexes containing 5 µg of plasmids (2.5 µg bFGF plasmid + 2.5 µg VEGFA plasmid; bFGF: VEGFA plasmid mass ratio = 1:1). The surrounding peritendinous subcutaneous tissue of each tendon was injected with 40 µL of nanoparticle/plasmid complexes containing 10 µg of plasmids (5 µg bFGF plasmid + 5 µg VEGFA plasmid). For intratendinous delivery, the 20 μL injection was divided equally between two sites (approximately 10 μL per site), with one injection administered into the tendon proximal to the repair site and the other into the tendon distal to the repair site. At each site, the needle was inserted approximately 5 mm into the tendon end before injection. For peritendinous subcutaneous delivery, the 40 μL injection was divided equally among four subcutaneous sites surrounding the tendon repair site (approximately 10 μL per site). The four injection sites were located on both sides of the skin incision, with two sites positioned proximal to the repair site and two sites distal to the repair site, one on each side of the incision. Thus, the combined-injection group received a total of 15 µg plasmid DNA across the two compartments (7.5 µg bFGF plasmid + 7.5 µg VEGFA plasmid). After injection, the tendons were surgically repaired using the modified Kessler method with with the core suture knot placed on the external surface of the tendon (extra-tendinous knot), rather than buried between the two tendon ends, using 5 − 0 polypropylene sutures (HOLYCON, Nantong, Jiangsu, China), leaving the sheath unrepaired. The skin was then closed surgically, and the operated toes were immobilized in a semi-flexed position with sterile gauze and adhesive tape until the sample collection time points (with the exception of 6 weeks; at this time point, the operated toes had been fixed for 3 weeks). For the rat tendon injury model, 24 six-week-old male SD rats (n = 6 for each group) were anesthetized with 3% pentobarbital sodium. Following routine skin preparation and disinfection, the Achilles tendon was exposed and completely transected under sterile surgical conditions. The animals were randomly assigned to four groups consistent with the chicken tendon injury model: non-injection, tendon injection, subcutaneous tissue injection, and combined tendon and subcutaneous tissue injection. Following the respective treatments, the transected Achilles tendons were repaired using the modified Kessler method with 5 − 0 double-stranded polypropylene sutures (HOLYCON, Nantong, Jiangsu, China). The surgical skin incision was then closed, without postoperative immobilization of the operated limb. Rat Achilles tendon samples were collected two weeks post-surgery. At each designated endpoint, all animals were euthanized using an overdose of pentobarbital sodium prior to tissue collection. The number of animals, group assignments, and sampling time points for this study are illustrated in FigureS2. Two operated digital flexor tendons or Achilles tendons were harvested from each chicken or rat, with both tendons assigned to the same treatment group according to the randomization scheme. Each operated tendon was processed and analyzed separately, with the tendon defined as the experimental unit for biomechanical and histological evaluations. All animals, tissue samples, experimental datasets, and outcome measurements reported in this study were specifically generated for this investigation and did not overlap with those included in previous publications.

In vivo assessment of local EGFP expression

To assess local plasmid-associated reporter expression following the administration of nanoparticle/plasmid complexes, EGFP fluorescence was evaluated in the treated tendons and surrounding peritendinous tissues. Two weeks after injection, treated chicken and rat tendon tissues were harvested and fixed in 4% paraformaldehyde in PBS buffer for 24 h at 4 °C. The tissues were dehydrated using a gradient sucrose protocol, and 10 μm sections were cut from the samples using a cryostat and thaw-mounted onto polylysine-treated slides. After washing with PBS, sections were stained with DAPI (4’,6-diamidino-2-phenylindole) solution and incubated at room temperature for 10 min to visualize the total number of cells. The sections were subsequently washed with PBS and mounted with anti-fade fluorescence mounting medium (Life Technologies, Carlsbad, CA, USA). Finally, the sections were examined under a fluorescence microscope (Leica DMR 3000; Leica Microsystems, Bensheim, Germany).

The morphology of healing tendons and adhesion scores

In the chicken model, the FDP tendon was accessed through a volar longitudinal incision along the entire length of the operated toe. Adhesions were recorded using established grading criteria, focusing on the morphology of adhesions surrounding the tendon. Adhesion grade assessments were based on appearance, diameter, and healing status of the tendon. Scores for each sample were derived from a qualitative and quantitative evaluation of these key aspects.

Ultimate strength of the healing tendons

In the chicken model, the FDP tendon was carefully separated from adjacent tissues and adhesions. The distal phalanx, attached to the terminal FDP tendon, was secured in the lower clamp of a testing machine, while the proximal end of the tendon was fixed in the upper clamp. The tendon length measured 8 cm between the two clamps, with the repair site positioned centrally. The tendon was subjected to a constant pull speed of 25 mm/min. The load on the repaired tendons was continuously recorded until tendon rupture occurred, as indicated by a sharp decline in the load-displacement curve displayed on the monitor. For the SD rat Achilles tendon model, the Achilles tendon was also carefully separated from surrounding soft tissues and adhesions, ensuring that the repaired region remained undamaged. The distal calcaneus, connected to the terminal Achilles tendon, was fixed in the lower clamp of a mechanical testing machine using a custom fixture. The proximal end of the Achilles tendon, near the musculotendinous junction, was secured in the upper clamp. The distance between the clamps were adjusted according to the specimen size, with the repair site positioned in the middle of the test span. Tendons were pulled at a constant speed of 25 mm/min, and the load on the repaired Achilles tendon was continuously recorded until rupture, identified by a sharp decline in the load-displacement curve on the monitor. For both models, forces were measured to the nearest 0.1 N.

Measurement of total local bFGF and VEGFA protein levels

Western blot analysis was performed to compare total tissue levels of bFGF and VEGFA proteins between the non-injection and combined injection groups. Tissues were treated with lysis buffer containing protease inhibitors (Promega, Madison, WI, USA). The tissue lysates were then centrifuged at 12,000 rpm for 20 min at 4 °C, and the supernatants containing total proteins were collected. Protein concentration was determined using the Bio-Rad Protein Assay Reagent (Vancouver, Canada). Equivalent protein samples were separated on a 12% sodium dodecyl sulfate-polyacrylamide gel, followed by transfer to a polyvinylidene fluoride membrane. After transfer, membranes were washed with TBST buffer and incubated with a blocking solution for 1 h at room temperature. Subsequently, membranes were incubated overnight at 4 °C with anti-VEGFA antibody (1:1000, 26157-1-AP, Proteintech, China ) and anti-bFGF antibody (1:1000, 11234-1-AP, Proteintech, China ). Following additional washing with TBST, membranes were incubated with a secondary antibody conjugated to IRDye800 (1:5000, Rockland, Gilbertsville, CA, USA) for 2 h at 37 °C. After further washing with TBST, membranes were developed and visualized using an Odyssey infrared imaging system (LI-COR, Lincoln, NE, USA). The optical density of the bands was quantitatively measured, with β-actin used as a reference control.

HE and Masson’s staining

Repaired tendon tissues and corresponding surrounding peritendinous subcutaneous tissues were harvested from chicken and rat models at specified time points. Samples were fixed in 4% paraformaldehyde for 24 h at 4 °C, followed by dehydration through a gradient series of alcohol, clearing in xylene, and embedding in paraffin blocks. Sections measuring 4 μm in thickness were cut and mounted on positively charged slides. After drying, sections were deparaffinized in xylene and rehydrated through graded alcohol.

For hematoxylin and eosin (HE) staining, sections were initially stained with hematoxylin to visualize nuclei, followed by eosin staining to assess general tissue morphology. Masson’s staining was performed according to the manufacturer’s instructions to evaluate collagen deposition and organization within the repaired tendons and surrounding peritendinous subcutaneous tissues. After staining, all sections underwent dehydration, were cleared, and sealed with neutral gum before being examined under a light microscope (Leica DMR 3000; Leica Microsystems, Bensheim, Germany). Evaluation of histological morphology, inflammatory cell infiltration, tendon-peritendinous tissue interface, and collagen fiber deposition was conducted qualitatively. The images presented serve to illustrate representative morphological features, rather than providing quantitative comparisons between groups.

Immunohistochemical staining of the healing tendons and subcutaneous tissues

The repaired tendon tissues and their corresponding surrounding peritendinous subcutaneous tissues were harvested and fixed in 4% paraformaldehyde for 24 h at 4 °C. Following fixation, the tissues underwent dehydration through a gradient alcohol series before being embedded in paraffin blocks. Sections with a thickness of 4 μm were cut and mounted on positively charged slides. After drying, the slides were rinsed with PBS and blocked with 10% normal goat serum in PBS for 20 min. The slides were then incubated overnight at 4 °C with anti-VEGFA antibody (1:200) and anti-bFGF antibody (1:200). After washing with PBS, the slides were further incubated with a secondary antibody at 37 °C for 30 min. The slides were stained with 3,3-diaminobenzidine (DAB, Sigma, St. Louis, MO, USA) for 2–3 min and counterstained with hematoxylin to visualize nuclei for 1–2 min. Finally, the slides were sealed with neutral gum (Invitrogen, San Diego, CA, USA) and examined under a light microscope (Leica DMR 3000; Leica Microsystems, Bensheim, Germany).

Immunofluorescence staining of the healing tendons and subcutaneous tissues

Repaired tendon tissues and the corresponding surrounding peritendinous subcutaneous tissues were harvested and fixed in 4% paraformaldehyde for 24 h at 4 °C. The tissues were subsequently dehydrated through a graded alcohol series and embedded in paraffin blocks. Sections measuring 4 μm in thickness were cut and mounted on positively charged slides. After drying, the sections were deparaffinized in xylene, rehydrated through graded alcohol, and rinsed with PBS. Antigen retrieval was accomplished using citrate buffer, followed by cooling to room temperature. The sections were then permeabilized with 0.3% Triton X-100 for 10 min and blocked with 10% normal goat serum in PBS for 30 min at room temperature. Subsequently, sections were incubated overnight at 4 °C with primary antibodies against CD31 (1:200, 11265-1-AP, Proteintech, China), Ki67 (1:200, 84192-3-RR, Proteintech, China), VEGFA (1:200), and bFGF (1:200), respectively. After washing with PBS, sections were incubated with appropriate fluorophore-conjugated secondary antibodies at 37 °C for 30 min in the dark. Nuclei were counterstained with DAPI for 5 min, and the sections were mounted with antifade mounting medium before examination under a fluorescence microscope (Leica DMR 3000; Leica Microsystems, Bensheim, Germany). Immunohistochemical and immunofluorescence images were interpreted descriptively, and no quantitative analysis of image intensity or positive cell counts was performed.

Systemic biosafety evaluation

To evaluate the potential systemic toxicity of the nanoparticle/plasmid formulation, peripheral blood samples were collected from three animals per group (n = 3) in both the non-injection and combined-injection groups. Blood samples from rats were collected at 2 weeks post-surgery, while samples from chickens were obtained at 6 weeks post-surgery. EDTA-anticoagulated whole blood was utilized for routine hematological analysis. The counts of white blood cells (WBC), red blood cells (RBC), and platelets (PLT) were determined using an automated veterinary hematology analyzer (BC-30 Vet, Shenzhen Mindray Animal Medical Technology Co., Ltd., China). Serum biochemical parameters associated with hepatic and renal function, including alanine aminotransferase (ALT), aspartate aminotransferase (AST), creatinine (CREA), albumin (ALB), globulin (GLOB), and blood urea nitrogen (BUN), were measured using an automated veterinary chemistry analyzer (BS-240 Vet, Shenzhen Mindray Animal Medical Technology Co., Ltd., China).

Additionally, the heart, liver, spleen, lungs, and kidneys were harvested from three animals per group at the corresponding time points. These tissues were fixed in 4% paraformaldehyde, embedded in paraffin, sectioned, and subsequently stained with HE. The resulting tissue sections were examined for signs of inflammatory cell infiltration, necrosis, structural disruption, or other pathological abnormalities.

Statistical analysis

Given the heterogeneity of variances observed at several key time points and the unequal sample sizes in some groups, a consistent statistical approach robust to these conditions was applied across all time points. Statistical analysis of adhesion scores and ultimate strength data was performed using Welch’s one-way analysis of variance (Welch’s ANOVA) because of unequal variances and unequal sample sizes among groups. Pairwise comparisons among groups were performed using the Games–Howell post hoc test. Western blot protein-expression data and hematological and serum biochemical parameters were analyzed using an unpaired two-tailed Student’s t-test because these analyses involved comparisons between the non-injection and combined-injection groups. A p value < 0.05 was considered statistically significant. All statistical analyses were performed using SPSS version 17.0 software (SPSS Inc., Chicago, IL, USA). Histological, immunohistochemical, and immunofluorescence imaging endpoints were evaluated descriptively; no quantitative image-intensity or positive-cell analysis was performed, and no inferential statistical tests were applied to these representative imaging findings.

Results

Characterization of nanoparticle/plasmid complexes and local EGFP reporter expression in vivo

SEM images revealed that both PLGA nanoparticles and nanoparticle/plasmid complexes exhibited a spherical morphology with relatively uniform size and good dispersion (Fig. 2a). The mean hydrodynamic diameter was measured at 114.33 ± 11.59 nm for PLGA nanoparticles and 123.67 ± 12.90 nm for nanoparticle/plasmid complexes (Fig. 2b). To evaluate the complexation ability between the nanoparticles and plasmids, a gel retardation assay was conducted. The results indicated complete electrophoretic retardation of free plasmid DNA at an N/P ratio of 6:1 (Fig. 2c). Zeta-potential analysis demonstrated that PLGA nanoparticles, PEI-modified nanoparticles, and nanoparticle/plasmid complexes at an N/P ratio of 6:1 had zeta potentials of approximately − 12.3 mV, + 38 mV, and + 21 mV, respectively. The positive surface charge of the nanoparticle/plasmid complexes may enhance their interaction with negatively charged cell membranes, facilitating cellular uptake. In vitro release analysis demonstrated a sustained plasmid release profile without an obvious initial burst. Approximately 76% of the loaded genes were released within 28 days at 37 °C, with the bFGF and VEGFA plasmids displaying similar release patterns (Fig. 2d). To assess local reporter expression following delivery of the nanoparticle/plasmid complexes in vivo, EGFP fluorescence was examined in both chicken and rat tissues at 2 weeks post-injection. Detectable EGFP-positive signals were evident in the treated tendon and surrounding peritendinous tissues in both animal models at 2 weeks after injection (Fig. 2e and f). These representative images demonstrate local expression of the EGFP reporter following administration of the nanoparticle/plasmid complexes; however, they were not utilized for calculating quantitative transfection efficiency. These findings indicate successful nanoparticle-plasmid complex formation, sustained plasmid release, and detectable local reporter expression in vivo.

Fig. 2.

Characterization of PLGA nanoparticle/plasmid complexes and assessment of local EGFP reporter expression in vivo. a Representative scanning electron microscopy images of unloaded PLGA nanoparticles and nanoparticle/plasmid complexes. b Dynamic light-scattering size distributions of PLGA nanoparticles and nanoparticle/plasmid complexes. c Agarose gel-retardation assay of nanoparticle/plasmid complexes prepared at the indicated nitrogen-to-phosphate (N/P) ratios. d Cumulative in vitro release profiles of bFGF and VEGFA plasmids from the nanoparticle complexes in phosphate-buffered saline at 37 °C for 28 days. (e, f) Representative fluorescence micrographs showing EGFP expression in the surrounding peritendinous subcutaneous tissue and repaired tendon tissue of the chicken (e) and rat (f)models 2 weeks after injection. EGFP is shown in green, nuclei are counterstained with DAPI (blue), and merged images are shown in the bottom rows. Scale bar = 100 μm

Biomechanical analysis and adhesion assessments

To evaluate the therapeutic effect of nanoparticle/plasmid complexes across different injection groups, a biomechanical analysis was conducted. In the chicken model, tendons in the combined injection groups exhibited higher ultimate strength compared to the non-injection control group at 2, 4, and 6 weeks post-treatment (9.9 ± 2.1 N vs. 6.1 ± 1.7 N, p = 0.028 for 2 weeks; 32.0 ± 5.5 N vs. 21.0 ± 2.3 N, p = 0.012 for 4 weeks; 79.8 ± 22.6 N vs. 45.7 ± 5.4 N, p = 0.026 for 6 weeks) (Fig. 3b). At these time points, the mean ultimate failure loads of the combined-injection group were approximately 162%, 153%, and 174% of the corresponding values in the non-injection group, respectively. At 2 and 4 weeks, the tendon injection group did not show a statistically significant increase in ultimate strength compared with the non-injection control (p = 0.176 and 0.976, with statistical powers of 0.26 and 0.02, respectively). These findings indicate that a clear early biomechanical benefit of tendon-only delivery was not demonstrated under the present experimental conditions; however, the low statistical power at these time points limits definitive interpretation of the non-significant results. At 6 weeks, the tendon injection group did show a significantly higher ultimate strength compared to the non-injection control group (70.4 ± 9.0 N vs. 45.7 ± 5.4 N, p < 0.001 ) (Fig. 3b). Notably, the largest increases in ultimate strength for the combined injection group were observed at 4 weeks relative to the non-injection control group. In the rat model, tendons in the combined treatment group exhibited significantly higher ultimate strength compared to the non-injection group at 2 weeks post-treatment (35.7 ± 5.7 N vs. 20.1 ± 4.5 N, p = 0.002) (Fig. 3e). Specifically, the mean ultimate failure load in the combined-injection group was approximately 177% of that observed in the non-injection group at this time point. Additionally, the tendon injection group also showed a significantly higher ultimate strength compared to the non-injection control group ( 31.8 ± 7.7 N vs. 20.1 ± 4.5 N, p = 0.048).

Fig. 3.

Effects of nanoparticle-mediated bFGF and VEGFA gene delivery on tendon adhesion, biomechanical strength, and growth-factor expression. a Representative gross appearances of repaired chicken digital flexor tendons in the non-injection, subcutaneous injection, tendon injection, and combined injection groups. b Ultimate strength of repaired chicken tendons at 2, 4, and 6 weeks after surgery (Week 2: non-injection, n = 6; tendon injection, n = 6; subcutaneous injection, n = 6; combined injection, n = 6； Week 4: non-injection, n = 6; tendon injection, n = 6; subcutaneous injection, n = 6; combined injection, n = 6；Week 6: non-injection, n = 6; tendon injection, n = 8; subcutaneous injection, n = 7; combined injection, n = 7). c Peritendinous adhesion scores in the chicken model at 6 weeks (n = 8). d Representative gross appearances of repaired rat Achilles tendons in the four treatment groups at 2 weeks. e Ultimate strength of repaired rat Achilles tendons at 2 weeks (n = 6). f Representative Western blots and densitometric analyses of bFGF and VEGFA expression in chicken repaired tendon and surrounding peritendinous subcutaneous tissues from the non-injection and combined injection groups at 2, 4, and 6 weeks (n = 3). g Representative Western blots and densitometric analyses of bFGF and VEGFA expression in rat repaired tendon and surrounding peritendinous subcutaneous tissues from the non-injection and combined injection groups at 2 weeks (n = 3). β-actin was used as the loading control. Data are presented as mean ± SD. Biomechanical and adhesion data were analyzed by Welch’s ANOVA followed by Games–Howell post hoc test, whereas Western blot data were analyzed using an unpaired two-tailed Student’s t-test

The morphology of tendon adhesions in both the chicken and rat models was assessed and scored based on established grading criteria. Figure 3a and d depict typical adhesion images from the non-injection control and combined injection groups, respectively. However, in the chicken model, no significant differences in adhesion scores were observed across all four groups (Fig. 3c).

Effects of combined nanoparticle/plasmid delivery on total local bFGF and VEGFA protein levels

To evaluate the expression of bFGF and VEGFA proteins in repaired tissues and surrounding peritendinous subcutaneous tissues following the injection of nanoparticle/plasmid complexes, Western blot analysis was performed. In the chicken model, total proteins were isolated from repaired tissues and surrounding peritendinous subcutaneous tissues obtained from both the nanoparticle/plasmid complexes injection group and the non-injection group at 2, 4, and 6 weeks post-treatment. Figure 3f displays the bFGF and VEGFA protein expressions in the repaired tendon tissues and surrounding peritendinous subcutaneous tissues for both the combined injection and non-injection groups at these time points. Compared to the non-injection group, the nanoparticle/plasmid combined injection group significantly increased bFGF and VEGFA protein levels in both the repaired tendon tissues and the surrounding peritendinous subcutaneous tissues. For the rat model, total proteins were also isolated from repaired tissues and surrounding peritendinous subcutaneous tissues collected from the nanoparticle/plasmid complexes injection group and the non-injection group at 2 weeks post-treatment. Figure 3g illustrates the expression of bFGF and VEGFA proteins in the repaired tendon tissues and surrounding peritendinous subcutaneous tissues from both groups at this time point. Similarly, the nanoparticle/plasmid complexes combined injection group exhibited significantly elevated bFGF and VEGFA protein levels compared to the non-injection group in both tissue types. These findings indicate that the injection of nanoparticle/plasmid complexes can significantly enhance the expression of bFGF and VEGFA proteins in repaired tendon tissues as well as the surrounding peritendinous subcutaneous tissues. In the chicken model, total tissue levels of bFGF and VEGFA were higher in the combined injection group than in the non-injection group at the assessed time points. A comparable difference was noted in the rat model at 2 weeks. However, conventional Western blot analysis detects total tissue protein without distinguishing between plasmid-derived bFGF/VEGFA and endogenous proteins induced during tendon healing; thus, these results are interpreted as increased total local growth-factor levels rather than direct evidence of sustained transgene-derived protein expression.

To further verify the distribution and expression of bFGF and VEGFA proteins in tissues treated with nanoparticle/plasmid complexes, immunohistochemical analysis was performed. At 6 weeks for the chicken model and at 2 weeks for the rat model, tissue slices from the four groups were stained with anti-bFGF and anti-VEGFA antibodies and observed under a microscope. Figure 4c and d present typical sections of immunohistochemical staining for bFGF and VEGFA in repaired tendons and surrounding peritendinous subcutaneous tissues treated with nanoparticle/plasmid complexes. The results indicated that the repaired tendons and surrounding peritendinous subcutaneous tissues in the non-injection group displayed negative or weak positive signals for both bFGF and VEGFA expression. In contrast, the combined injection group exhibited clear positive signals for bFGF and VEGFA. These findings suggest a higher local level of bFGF and VEGFA protein expression in the combined injection group, in both repaired tendon tissues and surrounding peritendinous subcutaneous tissues. This further confirms that nanoparticle/plasmid complexes effectively facilitate gene transfer into tissues and enhance the corresponding target protein expressions in vivo.

Fig. 4.

Histological remodeling and local bFGF/VEGFA expression after nanoparticle/plasmid-complex treatment. a, b Representative hematoxylin and eosin (HE) and Masson’s staining of the surrounding subcutaneous tissue and repaired tendon tissue from the chicken model at 6 weeks (a) and the rat model at 2 weeks (b). c, d Representative immunohistochemical staining for bFGF and VEGFA in the corresponding subcutaneous and tendon tissues from the chicken model at 6 weeks (c) and the rat model at 2 weeks (d). Scale bar = 100 μm

Histological remodeling and peritendinous adhesion after nanoparticle/plasmid treatment

HE and Masson’s staining were utilized to assess tissue morphology, collagen deposition, and the tendon–peritendinous tissue interface in both the chicken and rat tendon injury models. In the non-injection group, HE staining revealed disorganized reparative tissue, inflammatory cell infiltration, and adhesion-like tissue formation surrounding the repaired tendons. The boundary between the tendon and surrounding peritendinous subcutaneous tissue appeared relatively indistinct, indicative of an immature and disordered tendon–peritendinous interface.

Qualitatively, sections from the nanoparticle/plasmid-treated groups, particularly the combined-injection group, exhibited more regular cellular organization, reduced inflammatory-cell infiltration, and a more distinct tendon–peritendinous tissue interface compared to representative sections from the non-injection group.

Masson’s staining illustrated differences in collagen deposition and organization among the experimental groups. In the non-injection group, collagen fibers were irregularly distributed within the repaired tendon and surrounding adhesion-like tissues, indicating disordered extracellular matrix deposition during the healing process. Conversely, treated groups displayed better-organized collagen fibers and less chaotic collagen accumulation in the peritendinous region, particularly in the combined injection group.

Similar qualitative histological patterns were noted in the rat model. Along with the quantitative adhesion-score data from the chicken model, these representative findings indicate that combined treatment did not lead to a significant increase in peritendinous adhesion and may be related to more organized tendon remodeling.

Descriptive immunofluorescence assessment of Ki-67 and CD31 staining

Tendon tissues and surrounding peritendinous subcutaneous tissues from chickens were harvested at weeks 2, 4, and 6, while samples from the rat model were collected at 2 weeks to assess proliferation ability. Ki-67, a nuclear protein associated with cellular proliferation, was shown to increase following nanoparticle-mediated co-delivery of bFGF and VEGFA genes in vitro. Representative immunofluorescence images indicated more pronounced Ki-67-positive staining in the combined-injection group compared to the non-injection group. Since the Ki-67-positive cells were not phenotypically identified and no quantitative positive-cell analysis was conducted, these results suggest a greater local abundance of proliferating cells rather than reflecting proliferation within a specific cell population (Fig. 5a and b). CD31, a vascular endothelial marker, sensitively highlighted blood vessels in tissues. Tendon tissue slices were immunofluorescently stained with a CD31 antibody to investigate angiogenesis following treatment with nanoparticle/plasmid complexes. The results revealed that CD31-positive vascular structures were more visually prominent in representative sections from the combined-injection group. CD31 staining provided a structural marker for vascular presence; however, direct assessment of functional tissue perfusion was not performed in this study (Fig. 5a and b).

Fig. 5.

Immunofluorescence assessment of angiogenesis, growth-factor expression, and cell proliferation in repaired tendon and surrounding peritendinous subcutaneous tissues. a Representative fluorescence images from the chicken model at 6 weeks. b Representative fluorescence images from the rat model at 2 weeks. For each animal model, the surrounding peritendinous subcutaneous tissue and repaired tendon tissue are shown separately, with the non-injection, subcutaneous injection, tendon injection, and combined injection groups arranged from left to right. Sections were stained for CD31, bFGF, Ki-67, and VEGFA (red), with DAPI nuclear counterstaining (blue); merged images and the corresponding single-channel images are presented. Scale bar = 100 μm

Furthermore, immunofluorescence staining in both chicken and rat models aimed to evaluate the expression of VEGFA and bFGF, thus confirming the ability of nanoparticle/plasmid complexes to promote the expression of angiogenesis- and repair-related factors in vivo. In non-injection groups of both models, only weak signals for VEGFA and bFGF were detected in the repaired tendons and surrounding peritendinous subcutaneous tissues. In contrast, following treatment with nanoparticle/plasmid complexes, particularly in the combined injection group, VEGFA and bFGF signals were significantly enhanced and widely distributed within both the repaired tendon tissues and surrounding peritendinous subcutaneous tissues in the chicken and rat models. These findings indicate that nanoparticle/plasmid complexes effectively mediate local expression of VEGFA and bFGF, contributing to a pro-angiogenic and pro-reparative microenvironment for tendon healing.

Systemic biosafety of nanoparticle/plasmid treatment

To assess the systemic biosafety of the nanoparticle/plasmid formulation, a histological examination of major organs was performed in both non-injection and combined-injection groups. HE staining of the heart, liver, spleen, lungs, and kidneys revealed preserved tissue architecture in both groups. No obvious treatment-related inflammatory-cell infiltration, tissue necrosis, structural disruption, or other pathological abnormalities were observed in rats at 2 weeks or in chickens at 6 weeks (Fig. 6a). Hematological and serum biochemical parameters were also examined to evaluate potential systemic toxicity. WBC, RBC, and PLT counts showed no significant differences between the combined-injection and non-injection groups in either animal model. Similarly, serum biochemical parameters related to hepatic and renal function, including ALT, AST, CREA, ALB, GLOB, and BUN, did not differ significantly between the two groups at the respective time points (Fig. 6b-g). These findings suggest that, under the tested conditions, local administration of the nanoparticle/plasmid complexes was not associated with detectable major organ, hematological, hepatic, or renal toxicity during the observation periods.

Fig. 6.

Systemic biosafety evaluation of local nanoparticle/plasmid treatment in chicken and rat tendon injury models. a) Representative hematoxylin and eosin (HE) staining of the lung, heart, liver, spleen, and kidney from the non-injection and combined-injection groups in chickens at 6 weeks and rats at 2 weeks after surgery. Scale bar = 100 μm. b–d Hematological and serum biochemical parameters in chickens at 6 weeks (n = 3). b White blood cell (WBC), red blood cell (RBC), and platelet (PLT) counts. c Alanine aminotransferase (ALT), aspartate aminotransferase (AST), albumin (ALB), and globulin (GLOB) levels. d Creatinine (CREA) and blood urea nitrogen (BUN) levels. e–g Corresponding hematological and serum biochemical parameters in rats at 2 weeks (n = 3). e WBC, RBC, and PLT counts. f ALT, AST, ALB, and GLOB levels. g CREA and BUN levels

Discussion

Tendon healing poses a significant clinical challenge due to the characteristics of tendon tissue, which include poor vascularity, low cellularity, and a dense extracellular matrix. These factors collectively hinder the local availability and biological utilization of therapeutic molecules. Numerous studies have sought to enhance tendon repair by increasing the local concentration of growth factors [32–34]. However, direct delivery of recombinant proteins often faces limitations such as rapid degradation, short half-life, unstable activity, and insufficient diffusion within the dense tendon matrix. Gene delivery offers an alternative approach by enabling local cells to continuously produce therapeutic proteins [35, 36]. Nevertheless, tendon-specific gene delivery alone may be inadequate during the early healing stage because the number of resident tenocytes is limited, potentially resulting in insufficient production of functional growth factors to support the high demand for cell proliferation, angiogenesis, and collagen synthesis following injury.

The principal finding of this study is that spatially distributed delivery of bFGF- and VEGFA-encoding nanoparticles to both the tendon proper and the surrounding peritendinous subcutaneous tissue significantly increased ultimate tendon strength compared with the non-injection at 2 and 4 weeks after repair, whereas tendon injection did not result in a statistically significant increase over the non-injection at either time point. However, the effect of tendon injection at 2 and 4 weeks remains inconclusive after statistical analysis because the statistical power for the comparisons of ultimate strength between the non-injection and tendon injection groups at 2 and 4 weeks was extremely low (0.26 and 0.02, respectively). At 2 weeks, the ultimate strength was 6.1 ± 1.7 N in the non-injection group and 10.6 ± 4.3 N in the tendon injection group. Although the mean value was numerically higher in the tendon-injection group, substantial within-group variability was observed, as reflected by its relatively large standard deviation. At 4 weeks, the mean ultimate strength was similar between the non-injection and tendon injection groups (21.0 ± 2.3 N and 19.2 ± 10.8 N, respectively), whereas the tendon-injection group exhibited markedly greater variability. The combination of substantial within-group variability and the limited sample size resulted in low statistical power and a substantial risk of type II error. Therefore, the absence of statistically significant differences at 2 and 4 weeks does not demonstrate the absence of an effect of tendon injection, and no definitive conclusion regarding its effect at these time points can be drawn from the present data. Combined injection increased local bFGF/VEGFA expression, Ki-67-positive cell proliferation, and CD31-positive vascular structures, and improved collagen organization without a detectable increase in adhesion. These results endorse a dual-compartment, microenvironment-oriented strategy as opposed to a treatment confined only to the relatively acellular tendon proper. The 2018 study established the feasibility of PLGA nanoparticle-mediated co-delivery of bFGF and VEGFA genes into the tendon proper, demonstrating benefits primarily in the mid to late stages of healing [18]. However, tendon injection did not yield significant improvements in ultimate strength at 1 or 2 weeks [18]. The results of the present study were generally consistent with those reported in our 2018 study. In both studies, tendon injection did not significantly increase ultimate strength compared with non-injection at 2 weeks, whereas a significant increase was observed at 6 weeks. A difference was noted at 4 weeks: tendon injection significantly increased ultimate strength in the 2018 study, whereas no significant difference from the non-injection control was detected in the present study. In the present study, at 4 weeks, the mean ultimate strength was 21.0 ± 2.3 N in the non-injection group and 19.2 ± 10.8 N in the tendon injection group. The two studies differed in both experimental conditions and tendon repair techniques. First, the repair techniques differed between the two studies. In the 2018 study, the suture knots were placed between the tendon ends, whereas in the present study, the core suture knots were placed on the surface of the tendon. The 2023 study from our group demonstrated that suture knot location can affect early tendon healing; at 4 weeks, Kessler repairs with knots placed on the tendon surface showed significantly greater ultimate strength than modified Kessler repairs with knots placed between the tendon ends. Second, the present study used 5-0 polypropylene sutures, which differed from the suture material used in the 2018 study. Thus, the different statistical outcomes observed at 4 weeks occurred under distinct tendon repair conditions and should be interpreted in the context of these methodological differences and inter-experimental variability. The current study addresses this limitation in four key ways: by expanding the therapeutic target from the tendon proper to the tendon-peritendinous microenvironment; directly comparing tendon-only, peritendinous-only, and combined delivery; demonstrating early mechanical benefits in two anatomically distinct animal models; and incorporating biological endpoints that correlate treatment location with proliferation, vascularization, collagen remodeling, and tissue-interface organization. Thus, the incremental value of this study lies in altering the spatial distribution of gene delivery and the local treatment field rather than developing a new nanoparticle platform or re-testing the co-delivery of bFGF and VEGFA.

From a biological perspective, combined injection may operate through spatial complementation between intrinsic and extrinsic healing compartments. Intratendinous delivery directly exposes resident tenocytes and cells within the repair site to the complexes; however, the dense collagen matrix and low cellularity of the tendon proper restrict the number of cells capable of producing therapeutic proteins early after injury. In contrast, the surrounding peritendinous tissue contains tendon-sheath and peritenon fibroblastic populations, vascular-associated cells, and distinct progenitor populations that can actively participate in extrinsic repair [26–29]. Transfected cells in this compartment may thus serve as an additional local source of bFGF and VEGFA, extending the effective treatment field through paracrine signaling across the tendon-peritendinous interface. This model elucidates why the two injection sites may be complementary rather than redundant.

Within this proposed mechanism, VEGFA is anticipated to enhance the vascular component of the repair microenvironment, while bFGF supports proliferation and matrix production. Representative staining showed more visually prominent Ki-67-positive signals and CD31-positive vascular structures, together with more organized collagen morphology in the combined-injection group. Because these imaging endpoints were not quantitatively analyzed, they are interpreted as descriptive associations rather than statistically established treatment effects. Collectively, the sequence of broader transgene expression, increased proliferation and vascular markers, enhanced matrix organization, and higher ultimate strength offers a plausible explanation for the mechanical benefits observed. However, since functional perfusion, cell lineage, and pathway-specific inhibition were not measured, these findings support a mechanistic association rather than definitive causality.

PLGA, a biodegradable synthetic polymer, is widely employed in drug delivery systems and tissue engineering materials due to its biocompatibility, biodegradability, and lack of significant toxicity in vivo [37–39]. In the present study, physicochemical characterization indicated that the PLGA-based nanoparticle/plasmid complexes were appropriate for local gene delivery. The complexes exhibited a spherical morphology, relatively uniform size distribution, qualitative plasmid complexation at an N/P ratio of 6:1, and a positive surface charge following PEI modification, which may enhance interaction with negatively charged cell membranes and promote cellular uptake. In vitro release analysis demonstrated sustained plasmid release without a notable initial burst, indicating that these complexes could provide prolonged plasmid availability during the early phase of tendon repair. Consistently, EGFP-positive signals were detected in both tendon tissues and surrounding peritendinous subcutaneous tissues at 2 weeks following injection in the chicken and rat models, providing qualitative evidence of local EGFP reporter expression following plasmid delivery in vivo.

Recruiting the peritendinous compartment raises safety concerns, as extrinsic fibroblastic repair can contribute to scar formation and adhesion [26], and tendon-associated progenitor niches can generate both regenerative and fibrotic cell populations [30]. In the present study, combined injection did not result in increased adhesion scores, while histological examination revealed a clearer tendon-peritendinous interface and more organized collagen deposition. These findings suggest that, under the tested conditions, the spatial expansion of bFGF/VEGFA expression enhanced repair without producing overt peritendinous fibrosis. However, adhesion was quantified only in the chicken model, and the histological assessments were primarily descriptive; thus, quantitative fibrosis markers and long-term gliding function analyses are necessary to more rigorously define the therapeutic window. Additionally, serum biochemical indices and histological examinations of major organs revealed no detectable systemic toxicity under the tested dosing conditions.

This study has several limitations. First, the combined group received the sum of the intratendinous and peritendinous doses; thus, the current design cannot fully differentiate the benefits of spatial distribution from those of a higher total plasmid dose. Future experiments should include dose-matched tendon-only, peritendinous-only, and dual-compartment delivery groups, and assess the interaction between injection sites through a factorial analysis before claiming a true synergistic effect. Second, lineage-tracing or cell-specific transfection analysis was not conducted, leaving the cellular sources of bFGF/VEGFA and the contributions of peritendinous progenitor or vascular-associated cells inferred. Third, while CD31 staining identifies vascular structures, it does not demonstrate functional perfusion; methods such as microangiography, lectin perfusion, or laser Doppler measurements would strengthen the investigation of the vascular mechanism. Fourth, the rat model was primarily evaluated at 2 weeks, necessitating longer time-course, dose-response, gliding function, and biosafety studies. Despite these limitations, the results suggest that dual-compartment gene delivery as a rational approach to overcome the early-stage constraints associated with tendon-only treatment. Finally, the biomechanical analyses were limited by specimen attrition and the relatively small effective sample size. Following exclusion of infected or failed specimens, several groups included only 6–8 samples, which have reduced statistical power in comparison of some groups. More specific studies at the early time points such as week 2 and 4 are needed to know exact effects of tendon-only injection as the present study has failed to obtain conclusion due to insufficient statistical power of analysis.

Conclusion

In conclusion, spatially distributing bFGF/VEGFA gene-loaded nanoparticles across the injured tendon and surrounding peritendinous subcutaneous tissue improved early mechanical repair in two animal models and was associated with higher total local growth-factor levels, more prominent Ki-67-positive staining and CD31-positive vascular structures, and qualitatively more organized collagen morphology, without a detectable increase in adhesion in the chicken model. Compared with our previous tendon-only strategy, this study broadens the therapeutic target to encompass the tendon-peritendinous microenvironment and supports potentially complementary roles of the two tissue compartments. Because the total plasmid dose was not matched and the cellular and vascular mechanisms were not directly traced, these findings should be interpreted as indicative of enhanced combined delivery rather than proven synergy.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

J.S. performed the experiments, collected and analyzed the data, prepared the figures, and drafted the manuscript. Q.Q.Y. contributed to the experiments and data analysis. Y.L.Z. conceived and designed the study, developed the methodology, supervised the research, acquired funding, and revised the manuscript. J.B.T. contributed to the study conception, acquired funding, and revised the manuscript. All authors reviewed and approved the final manuscript.

Funding

This study was funded by the National Natural Science Foundation of China (82372389 and 81871824) and the Natural Science Foundation of Jiangsu Province (BK20241841).

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval

The animal experiment protocol was approved by the Experimental Animal Care and Use Committee of Nantong University (approval number S20250213-001). All methods were performed in accordance with the relevant institutional guidelines and regulations.

Consent to participate

Not applicable. This study did not involve human participants.

Consent to publish

Not applicable. This study did not involve human participants or contain any identifiable personal data, images, or information.

Competing interests

The authors declare no competing interests.