BMP-2 gene-modified mesenchymal stem cells enhance tendon–bone healing in ACL reconstruction: a rabbit micro-CT and biomechanical study

Abstract

Objective

Suboptimal tendon-bone interface healing remains a critical challenge affecting outcomes of anterior cruciate ligament (ACL) reconstruction. This study investigates whether BMP-2 gene-modified bone marrow mesenchymal stem cells (MSCs) enhances tendon-bone osseointegration.

Methods

An ACL reconstruction model was established in 18 New Zealand white rabbits, randomized into three groups: Control(fibrin glue), MSCs-Ad (adenovirus-vector transfected MSCs), MSCs-BMP2 (BMP-2-overexpressing MSCs). Micro-CT analysis and biomechanical testing assessed bone regeneration and interface properties at 4 weeks postoperatively.

Results

Micro-CT analysis revealed a significant enhancement in bone regeneration within the tunnel in the order of MSCs-BMP2 > MSCs-Ad > Control. The MSCs-BMP2 group demonstrated the highest values for bone volume fraction (BV/TV: 0.35 ± 0.05 vs. 0.27 ± 0.03 in MSCs-Ad vs. 0.22 ± 0.03 in Control), trabecular number (Tb.N: 0.73 ± 0.01 vs. 0.68 ± 0.01 vs. 0.64 ± 0.03), and trabecular thickness (Tb.Th: 0.46 ± 0.02 vs. 0.38 ± 0.01 vs. 0.36 ± 0.03). Biomechanical testing yielded consistent results: the maximum failure load for the MSCs-BMP2 group (32.80 ± 3.94 N) was significantly greater than that of both the MSCs-Ad (27.37 ± 2.15 N) and control (21.07 ± 1.94 N) groups, while the MSCs-Ad group also showed a significant improvement over the control.

Conclusion

BMP-2 gene-modified MSCs synergistically accelerate tendon-bone osseointegration, providing a theoretical foundation for biological augmentation in ACL reconstruction. This strategy may reduce postoperative graft failure risks. Future studies should validate clinical translation through long-term monitoring and mechanistic investigations.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12893-025-03277-x.

Introduction

Anterior cruciate ligament (ACL) injuries remain prevalent in orthopedics and sports medicine. Although recent epidemiological data from regions like the United States suggest a potential stabilization or even decline in overall incidence, surgical reconstruction demand continues to rise sharply, particularly among young, active populations seeking to return to high-demand activities [1, 2]. Regardless of incidence trends, ACL rupture profoundly compromises function. The ligament’s limited intrinsic healing capacity leads to persistent knee instability, substantially increasing risks of secondary meniscal tears, chondral damage, and accelerated progression to post-traumatic osteoarthritis [3, 4]. Consequently, surgical reconstruction remains the cornerstone treatment for restoring knee biomechanics.

Autologous hamstring tendon grafts have emerged as a preferred choice due to favorable biomechanics, lower donor-site morbidity than bone-patellar tendon-bone grafts, and high patient acceptance [5]. However, delayed or insufficient tendon-bone interface healing within osseous tunnels persistently undermines long-term outcomes. This weak link can cause graft laxity, tunnel widening, and suboptimal functional restoration [6]. Thus, developing biological augmentation strategies to accelerate and enhance tendon-bone integration represents a critical unmet need in ACL reconstruction.

Mesenchymal stem cells (MSCs) hold significant promise for addressing this challenge. Their therapeutic potential extends beyond multi-lineage differentiation (including osteogenic and tenogenic lineages) [7, 8]to encompass potent immunomodulation (e.g., via IL-10 and TGF-β1 secretion to suppress detrimental inflammation) and trophic support (e.g., promoting angiogenesis and fibroblast proliferation through VEGF) [9–11]. Collectively, these properties create a regenerative microenvironment conducive to interface healing. Nevertheless, translating unmodified MSCs faces hurdles, including unpredictable differentiation and limited site retention.

Bone morphogenetic protein-2 (BMP-2), a key osteoinductive cytokine in the TGF-β superfamily, potently stimulates osteoblast differentiation and bone matrix production via Smad signaling [12]. While recombinant human BMP-2 (rhBMP-2) promotes bone formation, its clinical application in tendon-bone healing is hampered by major limitations: supra-physiological doses—required to offset its short half-life and rapid clearance—often trigger serious adverse effects, including ectopic bone formation, significant inflammation, and osteolysis [13, 14]. These constraints necessitate alternative delivery strategies achieving sustained, localized, and physiologically relevant BMP-2 signaling.

To overcome challenges associated with both unmodified MSCs and direct rhBMP-2 delivery, this study employed genetic engineering. We hypothesized that BMP-2-overexpressing MSCs would enable continuous, low-dose paracrine release of the growth factor directly at the tendon-bone interface. This approach leverages synergistic benefits: BMP-2’s osteogenic drive combined with the inherent immunomodulatory and trophic support of MSCs. Using high-resolution micro-computed tomography (Micro-CT) for quantitative bone regeneration analysis [15, 16] and biomechanical testing of interface strength, we rigorously evaluated the efficacy of this BMP-2 gene-modified MSC therapy delivered locally within a fibrin-based scaffold. This combined strategy aims to provide a safer, more effective biological solution for enhancing tendon-bone healing in ACL reconstruction.

Method

Experimental design and animal model

Eighteen male adult New Zealand White rabbits (mean weight: 2.0–2.5 kg; Animal Center of Xinjiang Medical University) were randomly assigned to three groups (n = 6/group): Control Group: 100 µL fibrin glue (Sigma-Aldrich, St. Louis, MO, USA))injected at femoral and tibial tendon-bone interfaces.MSCs-Ad Group: 5 × 10⁶ adenovirus-transduced MSCs (Ad-GFP) suspended in 100 µL fibrin glue, injected at interfaces.MSCs-BMP2 Group: 5 × 10⁶ BMP-2-overexpressing MSCs (Ad-BMP2) suspended in 100 µL fibrin glue, injected at interfaces.

All animals underwent unilateral ACL reconstruction using an autologous semitendinosus tendon graft fixed to periosteum via suture suspension. The cell-scaffold mixture was injected into both bone tunnel entrances (∼5 mm depth) post-fixation. Animals were euthanized at 4 weeks for specimen collection (Section "Adenoviral transduction and BMP-2 expression").

Randomization: Rabbits were randomly assigned to groups using a computer-generated random number table.Allocation hidden: The group allocation was concealed in sealed opaque envelopes until the time of surgery.Blind method: The surgeons performing the surgery could not be blinded due to the nature of the intervention. However, the personnel performing the micro-CT analysis and biomechanical testing were blinded to the group allocation during data acquisition and analysis.

Isolation, culture, and characterization of bone marrow mesenchymal stem cells (BMSCs)

BMSCs were isolated from rabbit femurs/tibiae under isoflurane anesthesia (RWD Life Science, Shenzhen, China)via bone marrow flushing and heparinized density gradient centrifugation. Mononuclear cells were cultured in α-MEM (Procell Life Science, Wuhan, China) supplemented with 10% FBS. Non-adherent cells were removed after 48 h; adherent cells (P0) were passaged at 80% confluence using 0.25% trypsin.

Cell Proliferation: P3 BMSCs were seeded in 96-well plates (1 × 10³ cells/well). Growth kinetics were assessed over 6 days using a CCK-8 kit (Wuhan Three Eagles Biotechnology Co., Ltd., Wuhan, China). Absorbance (OD₄₅₀) was measured daily.

Adenoviral transduction and BMP-2 expression

P3 BMSCs were transduced with Ad-BMP2 or Ad-GFP (MOI = 100) in serum-free medium for 4 h, followed by complete medium addition. Transduction efficiency (> 90%) was confirmed by GFP fluorescence at 48 h.

BMP-2 Secretion: Supernatants from untransduced MSCs (Control), Ad-GFP-, and Ad-BMP2-transduced groups were collected 72 h post-transduction. BMP-2 concentration was quantified via ELISA (Lun Changshuo Biotech Co., Ltd., Beijing, China).

ACL reconstruction surgery

Under isoflurane/oxygen anesthesia, the right semitendinosus tendon (4–5 cm) was harvested. Anatomically positioned bone tunnels were drilled in the femur and tibia. The autologous graft (∼2 mm diameter) was passed through tunnels and fixed under mild tension at 30° knee flexion using periosteal suture suspension. The assigned treatment (fibrin glue + cells) was injected into femoral/tibial tunnel entrances using a dual-syringe system. The knee was held at 30° flexion for 3–5 min for scaffold gelation. Wounds were closed in layers. Postoperative care included intramuscular penicillin (800,000 U/day, 7 days) and daily monitoring without immobilization (Fig. 1).

Fig. 1.

Surgical workflow: (a) Knee exposure, (b) Tendon harvest, (c) Tunnel drilling, (d) Graft fixation

Euthanasia and specimen harvest

At 4 weeks, rabbits were deeply anesthetized (5% isoflurane/2 L/min O₂) until loss of corneal reflex, followed by intravenous pentobarbital sodium (Sigma-Aldrich, St. Louis, MO, USA) overdose (150 mg/kg) via the marginal ear vein. Death was confirmed by cessation of respiration and cardiac activity.

The surgical limb was disarticulated at the hip and ankle. Soft tissues were dissected to preserve the knee joint capsule, ligaments, and graft integrity. During dissection, the knee joint was carefully inspected grossly for any signs of ectopic mineralization or abnormal bone formation in the soft tissues surrounding the tunnels and the joint cavity. None were observed. The specimen (distal femur + proximal tibia, ∼5 cm) was rinsed in saline and fixed in 4% paraformaldehyde (4 °C).

Micro-CT analysis

Fixed specimens were scanned (NEMO^®; Beijing Medix Technology Co., Ltd., Beijing, China; 80 kV, 70 μm isotropic voxel size). Bone regeneration within tunnels was quantified by: Bone Volume/Total Volume (BV/TV, %),Trabecular Number (Tb.N, mm⁻¹),Trabecular Thickness (Tb.Th, mm).

Biomechanical testing

Specimens were thawed and hydrated. Femurs/tibiae were embedded in bone cement and mounted on an Instron 5848(Instron, USA)tester, aligning the graft parallel to the loading axis (simulating 30° flexion). After applying a 0.1 N preload, uniaxial tensile testing (20 mm/min) was performed until graft failure (load drop ≤ 30% peak). Hydration was maintained via saline spray.Recorded parameters: Maximum Failure Load (N), Tensile Stiffness (N/mm) (slope of the linear load-displacement region).

Statistical analysis

Primary and Secondary Outcomes: The primary outcomes were defined as bone volume fraction (BV/TV) and maximum failure load. Secondary outcomes included other microarchitectural parameters (Tb.N, Tb.Th) and tensile stiffness.Data normality was confirmed using the Shapiro-Wilk test, and homogeneity of variance was confirmed using Levene’s test. Data are expressed as mean ± standard deviation (SD).Inter-group comparisons used one-way ANOVA with Tukey’s post-hoc test (SPSS 26.0). P < 0.05 was considered significant.

Result

Morphological characterization of rabbit BMSCs

Primary rabbit bone marrow mononuclear cells displayed irregular, oval morphology with high refractivity after seeding. Pseudopodia extension was observed within 30 min of incubation. After 72 h (post-non-adherent cell removal), adherent cells exhibited spindle-shaped, fibroblast-like morphology (Day 4). By Day 8, cells formed confluent clusters with whorl-like growth patterns (Fig. 2).

Fig. 2.

Morphological observation of rabbit-derived bone marrow MSCs cells (×100)(scale bar = 10 μm)

BMSC proliferation kinetics

The P3 BMSC growth curve (CCK-8 assay) revealed: Days 1–3: Lag phase (slow proliferation), Days 3–5: Logarithmic growth phase (accelerated proliferation), Day 6 onward: Plateau phase (Fig. 3).

Fig. 3.

Growth curve of the 3rd generation MSCs

Adenoviral transduction efficiency

GFP fluorescence confirmed > 90% transduction efficiency in Ad-GFP-transduced BMSCs at 48 h post-infection (MOI = 100) (Fig. 4), validating robust adenoviral delivery.

Fig. 4.

Morphological observation of Ad-GFP infected MSCs 48 h post-infection under fluorescence microscopy (×100)(scale bar = 10 μm)

BMP-2 secretion analysis

ELISA quantification demonstrated significantly elevated BMP-2 secretion in the MSCs-BMP2 group (9.03 ± 0.20 ng/mL) versus both MSCs-Ad (2.25 ± 0.05 ng/mL) and untransduced controls (2.24 ± 0.05 ng/mL. No significant difference existed between MSCs-Ad and controls (Fig. 5).

Fig. 5.

Bar chart of BMP2 protein expression in the three groups(***: P < 0.01)

Micro-CT evaluation of bone regeneration

Qualitative analysis (Fig. 6): New bone formation (yellow areas) at tunnel entrances was markedly enhanced in the MSCs-BMP2 group vs. MSCs-Ad and controls (Fig. 6c vs. 6b/6a).

Fig. 6.

CT image of new bone in rabbit knee joints (a: Control group, b: MSCs-Ad group, c: MSCs-BMP-2 group; the yellow areas represent new bone. It can be observed that the new bone near the bone tunnel in the MSCs-BMP-2 group is more abundant than in the MSCs-Ad and control groups, and the MSCs-Ad group has more than the control group.)(scale bar = 10 mm)

Quantitative analysis (Fig. 7): The MSCs-BMP2 group demonstrated significantly higher values (0.35 ± 0.05) compared to both the MSCs Ad group (0.27 ± 0.03) and the control group (0.22 ± 0.03). Additionally, the MSCs Ad group exhibited a significant advantage over the control group.

Fig. 7.

Bar chart of three groups of Micro-CT analysis results (a: BV/TV, b: Tb. N, c: Tb. Th) (*: P < 0.05, **: P < 0.01, ***: P < 0.001)

Similarly, for Tb.N, the MSCs-BMP2 group (0.73 ± 0.004) showed significantly greater values than the MSCs Ad group (0.68 ± 0.01) and the control group (0.64 ± 0.03), with the MSCs Ad group also being significantly superior to the control group.

Regarding Tb.Th, the MSCs-BMP2 group (0.46 ± 0.02) again outperformed both the MSCs Ad group (0.38 ± 0.01) and the control group (0.36 ± 0.03), and the MSCs Ad group values were significantly higher than those of the control group.

Biomechanical testing

a:MSCs-BMP2 (32.80 ± 3.94 N) > MSCs-Ad (27.37 ± 2.15 N) > Control (21.07 ± 1.94 N). b:MSCs-BMP2 (13.37 ± 1.22 N/mm) > MSCs-Ad (11.70 ± 1.07 N/mm) and Control (11.37 ± 1.22 N/mm). No significant difference existed between MSCs-Ad and Control (P > 0.05) (Fig. 8).

Fig. 8.

Bar chart of biomechanical analysis for the three groups(a: Max Load; b: Stiffness) (*: P < 0.05, **: P < 0.01). Bar chart of biomechanical analysis for the three groups（a: Max Load;b: Stiffness） (*: P < 0.05, **: P < 0.01)

Discussion

Inadequate tendon-bone interface healing remains a critical limitation in ACL reconstruction, often leading to graft loosening and clinical failure. This study demonstrates that localized delivery of BMP-2 gene-modified MSCs within a fibrin scaffold improved osseointegration and biomechanical strength in a rabbit model at 4 weeks. Our principal findings reveal that the MSCs-BMP2 group yielded superior bone microarchitecture, evidenced by significant increases in BV/TV, Tb.N, and Tb.Th compared to both control and MSCs-Ad groups. This structural enhancement translated to functional superiority, with the MSCs-BMP2 group exhibiting a 56% increase in maximum failure load and an 18% increase in tensile stiffness versus controls. Micro-CT reconstructions further revealed strategic bone deposition at the tunnel entrances, a high-stress region, suggesting this optimized bone regeneration contributes to reduced graft micromotion and improved fixation strength.

Our strategy of using gene-modified MSCs offers distinct advantages over direct recombinant rhBMP-2 delivery. While rhBMP-2 is a potent osteoinductive agent, its clinical application is hampered by a short half-life, necessitating supraphysiological doses that carry risks of ectopic ossification, significant inflammation, and osteolysis [17, 18]. In contrast, BMP-2-overexpressing MSCs can provide sustained, low-dose paracrine signaling directly at the healing interface, more closely mimicking endogenous growth factor kinetics and potentially mitigating these adverse effects [19]. Furthermore, the therapeutic benefit of MSCs extends beyond single-factor delivery. The observed biomechanical improvement in the MSCs-Ad group over the control, albeit modest, underscores the inherent regenerative capacity of MSCs, likely mediated through their trophic support functions, such as secreting VEGF to promote angiogenesis. The synergistic effect seen in the MSCs-BMP2 group likely arises from the combination of potent BMP-2-driven osteogenesis and the conducive microenvironment created by MSC-secreted factors.

A key strength of our study is the integration of high-resolution micro-CT with biomechanical testing. This approach overcame the potential sampling bias of 2D histology by providing volumetric quantification of bone regeneration throughout the tunnel [15, 16]. It also established a strong structure-function correlation, strengthening the validity of our conclusions. Our findings align with recent work by Wang et al. (2024), who demonstrated that combined BMP-2 and sVEGFR1 treatment enhanced tendon-bone healing by regulating stem cell lineage [20], and Zhang et al. (2024), who utilized a dynamic hydrogel scaffold with BMP-2 to improve healing [21]. Our study contributes to this field by validating the efficacy of a cell-based gene therapy approach for sustained factor delivery.

Limitations and future work

We acknowledge several limitations. First, the lack of histological analysis precludes assessment of interface maturity and definitive exclusion of ectopic bone, though none was observed grossly. Second, we lack molecular data (e.g., Smad pathway activation) to mechanistically confirm BMP-2’s role in vivo. Third, the use of an adenoviral vector warrants investigation into potential immune responses. Future studies should include histology and molecular assays, employ longer timepoints to assess healing maturation, and validate findings in a large-animal model.

Conclusion

In conclusion, BMP-2 gene-modified MSCs synergistically improved tendon-bone healing by providing sustained osteoinduction alongside MSC-derived trophic support. This strategy significantly improves structural and functional outcomes in a preclinical model, offering a promising biological augmentation approach for ACL reconstruction. While further validation is needed, these findings provide a strong foundation for future research.

Abbreviations

BMP-2

Bone Morphogenetic Protein-2

MSCs

Marrow Mesenchymal Stem Cells

BMSCs

Bone Marrow Mesenchymal Stem Cells

ACL

Anterior cruciate ligament

BV

Bone volume

TV

Total Volume

Tb.N

Trabecular number

Tb.Th

Trabecular thickness

CCK-8

Cell Counting Kit-8

Ad

Adenovirus

OD

Optical Density

MOI

Multiplicity Of Infection

GFP

Green Fluorescence Protein

Authors’ contributions

ZHC: Established animal models, performed surgical procedures and postoperative management, and participated in data collection and analysis; ZYT: Isolated and cultured bone marrow mesenchymal stem cells, conducted adenovirus transfection, and performed ELISA to detect BMP-2 expression; Na : Responsible for Micro-CT scanning and three-dimensional image data processing, completed quantitative analysis of bone regeneration parameters; W: Led biomechanical testing (maximum load, tensile stiffness, etc.), optimized experimental equipment and parameter settings; T: Compiled literature, assisted in result interpretation and chart creation, and participated in the drafting and refinement of the manuscript. S: Proposed the research concept, designed the overall experimental plan, guided project implementation, and was responsible for manuscript writing and final revisions. All authors contributed to the discussion of experimental results and manuscript revisions and ultimately approved the final content.

Funding

This work has received funding from the Research and Innovation Team for Sports Medicine and Cartilage Regeneration Materials at the Sixth Affiliated Hospital of Xinjiang Medical University (LFYKYZXJJ2024001).

Data availability

Due to privacy concerns, the data is not publicly available, but can be obtained from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Ethics Review Committee of the Sixth Affiliated Hospital of Xinjiang Medical University, which is affiliated with the Sixth Affiliated Hospital of Xinjiang Medical University. The ethics approval number is LFYLLSC20201013-02.This research complies with international guidelines and adheres to the Basel Declaration.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Statement

Our study is reported in accordance with ARRIVE guidelines.