Mechanotransduction-driven macrophage polarization via Integrin-SRC-STAT6 pathway in distraction osteogenesis

Abstract

Background

Mechanical stimuli are indispensable for bone regeneration. Distraction osteogenesis (DO) is a widely used clinical technique for limb lengthening and bone defect repair; however, its specific mechanobiological mechanisms remain unclear. Macrophages play crucial regulatory roles throughout bone fracture healing. Recent studies indicate that macrophages are mechanosensitive and can modulate the local immune microenvironment in response to mechanical cues. This study aims to investigate how macrophages respond to mechanical stimulation and regulate bone regeneration during DO.

Methods

Animal models of DO (with external fixation) and fracture healing (with internal fixation) were established to compare bone regeneration under different mechanical conditions. Immunohistochemistry (IHC) was used to quantify M1 and M2 macrophage infiltration. An in vitro model of cyclic mechanical stretch (10 %, 0.5 Hz, 12 h) was applied to RAW264.7 cells to study macrophage polarization. Flow cytometry, PCR, and western blot were used to assess macrophage phenotypes. An indirect co-culture system was employed to evaluate the effect of mechanically stimulated M2 macrophages on osteogenic differentiation. Single-cell RNA sequencing analysis of public data was performed to identify key biological processes in macrophage subpopulations during DO. Western blot and immunofluorescence were used to measure expression and phosphorylation levels of SRC and STAT-6. Pathway inhibitors were applied to elucidate regulatory mechanisms. In vivo, Saracatinib and TGF-β were administered locally in DO models. Bone regeneration was evaluated using micro-CT, mechanical testing, and histology.

Results

DO significantly enhanced M2 macrophage polarization at 1st, 2nd, and 4th week post-surgery compared to controls. Cyclic stretch promoted M2 polarization in vitro and increased secretion of TGF-β and IL-10. Mechanically induced macrophages enhanced osteoblast differentiation in co-culture. Mechanical activation of the Integrin-SRC-STAT6 pathway drove M2 polarization. Local SRC inhibition suppressed M2 polarization and impaired bone regeneration in DO, which was partially rescued by TGF-β supplementation.

Conclusion

Mechanical stimulation during DO promotes M2 macrophage polarization via the Integrin-SRC-STAT6 pathway. TGF-β appears to be a key cytokine secreted by mechanically induced M2 macrophages that facilitates osteogenesis. These findings reveal a novel mechano-immune regulatory axis that supports bone regeneration in DO.

The translational potential of this article

This research confirms the core concept of "mechano-immunoregulation" and identifies actionable therapeutic targets, enabling the development of targeted therapies for refractory bone defects by modulating the integrin-β1/SRC/STAT6 pathway and TGF-β1 to enhance bone regeneration.

Graphical abstract

A summary of mechanisms of mechanical induced M2 polarization and its osteogenic effect in DO. Mechanical stretch exerted by DO is sensed by Integrin leading to the phosphorylation of Src. p-Src phosphorylates Stat6 in the cytoplasm. p-Stat6 translocates to the nucleus to promote the expression of genes associated with M2 polarization. M2 polarized macrophages can secrete cytokines such as TGF-β, directly promoting osteoblast osteogenic differentiation and assisting in maintaining their own polarized state.

1. Introduction

The mechanical stimulation is an ancient therapeutic method for bone regeneration with a history spanning over a thousand years [1]. Based on those practices and theories, Gavriil Ilizarov, an orthopaedic surgeon from Russia, revolutionized orthopaedic surgery in the post-World War II era by formalizing the systematic technique of distraction osteogenesis (DO) during the 1950s [2]. This groundbreaking innovation involved the controlled, incremental application of mechanical tension to bone callus, enabling precise limb lengthening and regeneration. The DO technique amidst postwar surgical challenges, laid the scientific and practical groundwork for modern DO. This technology rapidly gained global recognition, becoming a cornerstone in orthopaedics [[3], [4], [5], [6], [7]] and craniofacial surgery [8] to repair large bone defects, correct deformities, and restore limb functional mobility.

As research on DO advances, researchers have found that its therapeutic effects extend beyond bone regeneration to surrounding blood vessels [9], lymphatic vessels [10], nerves [10] as well as damaged skin on the remote limb [11]. Notably, researchers discovered that DO stimulates angiogenesis synchronously with bone formation [12,13], and it with its efficacy closely tied to the rate of distraction [13]. Recently, BAI [10] and LU [14] reported lymphatic vessel regeneration during cranial bone transport, suggesting a potential application of DO in the treatment of brain disorders. Additionally, Cassio et al. demonstrated that late posterior vault DO can improve behavioral, learning and visual acuity outcomes without altering ventricle size [15], highlighting its multifaceted benefits. Collectively, the new emerging research opens new possibilities for leveraging DO in regenerative medicine and treating a wider range of diseases. These research findings not only deepen our understanding of tissue regeneration but also underscore the complex mechanobiological responses involving multiple cell types in DO.

At the bone lengthening site, cells sense mechanical stimuli and consequently response to it by altering their functions [[16], [17], [18]]. Those mechanoresponsive cells include not only osteoblasts, osteocytes, periosteal stem cells (PSCs) [19], and bone marrow mesenchymal stem cells(BMSCs) [20] that directly drive bone regeneration, but also H-type endothelial cells [21] and macrophages [22,23]. Emerging evidence highlights the role of macrophages in DO. Macrophages contribute to bone regeneration, a process tightly couples to their polarization into distinct functional phenotypes [22,24,25]. Recent studies have revealed that macrophages are mechanosensitive cells [18,23,[26], [27], [28], [29], [30], [31]] and capable of sensing various mechanical stimuli, including hydrostatic pressure [32], fluid shear stress [33], and mechanical stretch [17,34]. Upon mechanical stimulation, macrophages may polarize to different phenotypes and secret cytokines, then interact with a variety of neighboring cells [[35], [36], [37], [38], [39]].They may further facilitate bone [[40], [41], [42]]and blood vessels regeneration [40] by reshaping the local inflammatory environment [17,43]. In summary, macrophage may play an indispensable role in modulating the local immune environment as one of the mechanoresponsive cells, and the crosstalk between M2-polarized macrophages and osteoblasts as a key driver of osteogenesis.

Despite these advances, the precise mechanobiological mechanisms by which macrophages respond to mechanical signals and regulate surrounding immune environment remain poorly understood. This study aims to elucidate the pivotal role of macrophages in response to mechanical stretch during DO, and with a focus on identifying key molecules which mediated this mechanotransduction process. Building on prior findings, we propose the following hypotheses: (1) Mechanical stretch-induced macrophage polarization may play an essential role in DO. (2) The Integrin-SRC-STAT6 axis may orchestrates stretch-driven M2 polarization; and (3) Polarized macrophages may modulate the local immune microenvironment then promote osteogenesis through the secretion of various cytokines like TGF-β during DO.

2. Results

2.1. Macrophages persistently polarize to M2 phenotype during DO

The polarization of macrophages during bone repairment is a dynamic process involving the transition between its pro-inflammatory and anti-inflammatory phenotypes [44]. To investigate the impact of mechanical stimulation on this dynamic process, we designed the experiment as Fig. 1A showed. 2 groups were established: sham control (SC) group in which femur fracture and fixation were performed, and distraction osteogenesis group (DO) in which distraction osteogenesis was applied. Samples were harvested at postoperative day (POD) 7, 14, and 28 for both DO and SC groups.

Fig. 1.

The histological and immunohistochemical results of the DO group and SC group at different time points. (A) The timetable for animal models, and sacrifice. (B) H&E staining and Masson's staining figures (2X) showing the differences in fracture morphology as well as bone and cartilage formation at POD7, POD14, and POD28. (C) Immunohistochemical staining figures (2X & 10X) and statistical results of percentage of CD206⁺ CD86⁺ cells at bone callus in DO and SC group. n = 8, ∗P < 0.05, ∗∗∗P < 0.001 by two-way ANOVA and Tukey's post hoc tests.

H&E staining and Masson's staining were conducted to assess the healing progress of the bone fracture and to define the different stages of bone fracture healing (Fig. 1B). Besides, Immunohistochemistry (IHC) was performed for the purpose of visualizing the quantity and spatial distribution of M1 and M2 macrophages.

Immunohistochemistry results showed that a huge number of iNOS⁺ cells mainly showed around the fracture site at POD 7 in both DO and SC groups (Fig. 1C). CD206⁺CD86⁻ cells mainly showed in the bone callus and surrounding tissues indicating the healing process started shifting from the acute inflammation phase to the soft callus phase. At POD 14 when distraction was completed, a significant higher amount of CD206 + cells could be observed in the distraction area in the DO group than SC group (+5.07 folds, p < 0.001). The amount of iNOS⁺ cells in the SC group notably decreased indicating inflammation was fading. The number of iNOS⁺ cells in the DO group was higher than that of SC group. At POD 28, CD206 + cells in the DO group were significantly (+6.09 folds, p < 0.001) higher than in SC group. Besides, those cells were mainly concentrated in the distraction area. The SC group had fewer CD206⁺cells which mostly showed in the bone marrow rather than the callus area. iNOS + cells can hardly be seen in both SC and DO group at this time point.

These results demonstrated that macrophages underwent significant M2 polarization that persisted for at least 2 weeks after finishing distraction. Interestingly, the polarization of M1 macrophages is also enhanced to respond to mechanical stretch and/or balance M2 polarization.

2.2. Cyclic stretch promotes macrophage M2 polarization

After confirming the changes in the polarization status of macrophages in the DO rat model, we employed an in vitro stretching system to further investigate the impact of mechanical stimuli on function of macrophages. RAW 264.7 cells were seeded on flexible membranes as shown in the illustration (Fig. 2A).

Fig. 2.

Cyclical stretching promotes macrophage M2 polarization. (A)The illustration of in vitro stretch system and mechanical parameters used in the cell experiments. (B) Relative mRNA levels of IL-10, TGF-β, CD206, TNF-α and iNOS under different mechanical loading time (3h, 6h,12h and 24h). (C) Relative protein levels of iNOS and ARG-1 in the NC and ST group. (D) Flow cytometry data showing the percentage of CD206⁻CD86⁺ and CD86⁺CD206⁻cells in the NC and ST group. n = 4, ∗P < 0.05, ∗∗∗P < 0.001 by two-way ANOVA, Tukey's post hoc tests and unpaired Students’-t tests.

Given the complexity of localized force distribution, it is challenging to design an in vitro stretching system that fully replicates in vivo conditions. Therefore, we referenced established parameters from previous studies—10 % strain at 0.5 Hz [17]—and tested different durations of stretch. These parameters have been validated for macrophage responses in distraction osteogenesis models, providing justification for my experiments.

To minimize parameter-related bias, we conducted additional experiments testing variable conditions (frequency and strain magnitude). Results confirmed that while cytokine secretion exhibited dose-dependent trends under certain parameters such as frequency (Fig. S1A) and strain (Fig. S1B), adjustments within our tested range did not significantly alter macrophage polarization outcomes.

After cell attachment, we applied mechanical stretch in specific parameters (cyclic stretch, uniaxial, 10 %, 0.5Hz) to the macrophages. To determine the optimal duration of in vitro stimulation, we applied cyclic stretch (ST) for 3 h, 6 h, 12 h, and 24 h, and collected RNA of cells for q-PCR analysis (Fig. 2B). The results showed that compared to the negative control (NC) group, the expression of IL-10 significantly increased (+3.15 folds, p < 0.001) at 6 h, while TGF-β and CD206 expression significantly increased (+1.83 folds, p < 0.01 and + 1.25 folds, p < 0.001) at 12 h. Since the fold changes of mRNA level of il-10 and CD206 were most significant with 12-h stretching, we chose 12 h of stretching as the condition for the following experiments. Results of western blot showed that the expression of Arg-1 significantly increased (+1.79 folds, p < 0.001) in the 12-h stretch group (Fig. 2C). Data also showed that cyclic stretch significantly increased (+13.83 folds, p < 0.001) the proportion of CD206⁺/CD86^- macrophages (Fig. 2D).

To gain a more comprehensive understanding of the effects of mechanical stretch on macrophage polarization under different contexts, LPS was pre-administered to establish an inflammatory environment. Subsequently, RAW264.7 cells were subjected to mechanical stretch with specific parameters (cyclic stretch, uniaxial, 10 %, 0.5Hz). The mRNA expression levels of M1-related and M2-related cytokines were measured via PCR (Fig. S2A). The polarization status of macrophages under was verified using flow cytometry (Fig. S2B).

2.3. M2 macrophages facilitate osteogenic differentiation of MC3T3-E1 cells in an indirect co-culture system

The results from animal model and in vitro cell stretch experiments demonstrated that mechanical stretch could induce the M2 polarization of macrophages. To further investigate the effects of mechanically induced M2 macrophage on the functions of osteoblasts in the surrounding environment, we designed the following experiments: conditioned medium was prepared following the workflow shown in Fig. 3A. 3 groups were included: NC group in which conditioned medium was consist of 50 % (v/v) OIM and 50 % (v/v)α-MEM, ST group in which conditioned medium was consist of 50 % (v/v) OIM and 50 % (v/v) supernatant collected form stretched RAW 264.7, and Mφ group in which conditioned medium was consist of 50 % (v/v) OIM and 50 % (v/v) supernatant collected form static RAW 264.7.

Fig. 3.

Conditioned medium facilitates osteogenic differentiation of osteoblasts. (A)The workflow of preparing conditioned medium. (B) Representative figures of ALP staining as well as the statistical results in NC, Mφ, and ST group. (C)ARS and quantification results among different groups. (D) Relative protein levels of osteogenic markers, Col-1, Runx2, and OCN in NC, Mφ, and ST group. (E) Wound healing assay showing differences in migration of MC3T3-E1 cells between NC, Mφ, and ST group at 24 and 48 h n = 4, ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001 by one-way ANOVA and Tukey's post hoc tests.

ALP staining results indicated significantly higher ALP activity in the ST group compared to the NC group and Mφ group after cultured for 7 days (+4.29 folds, p < 0.001) and 14 days (+3.45 folds, p < 0.001) (Fig. 3B). Similarly, the results of ARS staining and quantification at day 14(+2.68 folds, p < 0.001) and day 21 (+4.09 folds, p < 0.001) suggested a significant increase in calcium nodule formation in the ST group compared to the other groups (Fig. 3C). The results of western blot showed a significant increase in the protein expression levels of Col-1(9.32 folds, p < 0.001), Runx-2 (5.23 folds, p < 0.001), and OCN (1.12 folds, p < 0.001) in MC3T3-E1 cells in the ST group compared to the NC group and Mφ group after 14 days of culture (Fig. 3D).

The wound healing assay was used to assess the effect of the conditioned medium on osteoblasts’ migration (Fig. 3E). The wound healing index of the ST group at 24 and 48 h was significantly higher than that of the NC group, but the increase was not significant compared to the Mφ group.

In summary, mechanically induced M2 macrophage significantly enhanced osteogenic differentiation of osteoblasts.

2.4. The sc-RNA sequencing demonstrates the impact of DO on macrophages

To investigate the key molecules and signaling pathways in macrophages polarization during DO, we analyzed the single cell sequencing data downloading from public databases (GSE169592). Based on the markers shown in Fig. S3A–3E, cells can be divided into 6 clusters.

Through evaluating the “mechanical response score” for each cluster, we found macrophages exhibit high mechanical responsiveness, suggesting that they may function as mechanosensitive cells in the Fig. S4A. Moreover, the abundant expression of ITGB1 and SRC in macrophages showed in the Fig. S4 B provides a foundation for further investigation into the Integrin-SRC signaling pathway. Besides, we expanded our investigation of intercellular signaling by specifically examining ligand–receptor interactions between macrophages and osteoblasts. The results demonstrated significant communication activity involving TGF-β pathways (Fig. S4 C-G). The results also showed that the percentage of p-SRC⁺ cells in the DO group was significantly higher than that in the Saracatinib and TGF-β groups. This indicates that SRC phosphorylation was effectively inhibited in the TGF-β group. Combined with the final osteogenic outcomes observed in the TGF-β group, it can be inferred that TGF-β may serve as the ultimate effector molecule secreted by M2 macrophages.

Single cell bioinformatic analysis of macrophages revealed significant differences in the expression levels of key genes in the DO group (Fig. 4A). Expression levels of Itga5 and Itgb1 were significantly increased in macrophages in the DO group. Besides, mRNA levels of tubulin, a critical component of the cell cytoskeleton, also increased significantly, suggesting that mechanical stimuli induce cytoskeletal remodeling. The expression levels of Il4r (+1.19 folds, p < 0.01), Lrp1 (+1.15 folds, p < 0.001) and Src (+1.14 folds, p < 0.01) significantly increased, indicating that the immunoregulatory functions of macrophages may be altered by mechanical stimuli. Subsequent GO analysis suggested that the expression levels of secretion function, response to external stimuli, and focal adhesion are significantly increased in the DO group (Fig. 4B). The osteogenesis related genes in the DO group were significantly reduced probably due to the unstable mechanical environment. Fig. 4C showed the differences in signaling pathways between the DO and SC groups, including BMP targets, myeloid cell development, TGF-β targets 1hr, inflammatory response LPS, TGF-β targets 10hr, tolerant macrophage.

Fig. 4.

Results of single cell sequencing analysis of DO animal model. (A) Differentially expressed mRNA levels of Tuba4a, Tubb4b, Itga5, Itgb1, Col1a1, Il4, Lrp1, and SRC. (B) Differences of biological processes, molecular functions, and cellular components between DO and SC group. (C) Signaling pathway that upregulated in DO group compared to SC group. (D) ELISA assay showing the supernatant protein levels of TNF-α, TGF-β, IL-10 and LRP1 under 12-h stretching. n = 4, ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001 by unpaired Students’-t tests.

ELISA was conducted to verify results of differentially expressed genes. ELISA results showed a significant increase in the levels of TGF-β (+84.88 %, p < 0.01), IL-10(+34.31 %, p < 0.05), and LRP1(+4.7 folds, p < 0.001) in the supernatant after 12 h of stretching, while the expression level of TNF-α remained unchanged (Fig. 4D).

In summary, sc-RNA sequencing analysis revealed that the DO group exhibited activation of signaling pathways and biological processes related to mechanical signal transduction and immune regulation compared to the SC group. Further analysis of key molecules indicated that macrophages in the DO group may respond to mechanical stimulation, secrete cytokines and regulate immune environments via the Integrin-SRC axis. ELISA results further confirmed our hypothesis.

2.5. Integrin-SRC-STAT6 signaling pathway mediates mechanical stimulation-induced macrophage polarization

Western blot was conducted to investigate the key molecules and pathways mediating mechanical stimulation-induced macrophage polarization. Based on sc-RNA sequencing results, we mainly focused on the differentially expressed levels of integrin-β, SRC, and macrophage polarization-related molecules, STAT6, in the mechanical stretch (ST) group and negative control (NC) group. Results of western blots showed a significant increase in integrin-β protein expression levels after cyclic stretch, while the protein levels of SRC and STAT6 showed no significant change (Fig. 5A).

Fig. 5.

Integrin-β/SRC/STAT6 signaling pathway mediates mechanical stimulation-induced macrophage polarization. (A) Relative protein levels of Integrin-β, total SRC, total STAT6 in the NC and ST group. (B) Levels of phosphorylation of SRC and STAT6 at different time points. (C) Relative expression levels of p-STAT6 and Tubulin, and their subcellular localization. (D) Semi-quantitative statistical analysis of fluorescence intensity of p-STAT6 and Tubulin. (E) Relative protein levels of Arg-1, Integrin-β, total SRC, p-SRC, total STAT6 and p-STAT6. (F) Flow cytometry data showing the percentage of CD206⁻CD86⁺ and CD86⁺CD206⁻cells in the NC and ST,Saracatinib and ST + Saracatinib group. n = 4, ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001 by unpaired Students’-t test and one-way ANOVA and Tukey's post hoc tests.

Due to Integrin activating signaling pathways through the phosphorylation of downstream molecules [45], it is essential to conduct western blotting of phosphorylated SRC and STAT6 at various time points. The results showed a significant increase (+2.9 folds, p < 0.001) in p-SRC levels at 1 h after stretching, with unchanged total SRC. p-STAT6 started to increase 6 h after stretching and significantly increased at 12 h (Fig. 5B). To investigate the role of SRC in the mechanically induced M2 polarization, Saracatinib were used 24 h before stretch to inhibit the expression of SRC in RAW 264.7 cells.We assessed the effect of Saracatinib on cell viability and its efficacy as an SRC inhibitor across a range of concentrations. The results showed that Saracatinib had no effect on cell viability at all tested concentrations (Fig. S5A), while SRC activity was significantly inhibited at a concentration of 10 Μm (Fig. S5B). Results of immunofluorescence showed a dramatic increase in the total protein level of p-STAT6 (Fig. 5C). The remodeling of cytoskeleton and nuclear translocation of p-STAT6 were observed in the ST group. After inhibition, we found the mechanically induced overexpression of p-STAT6 and its nuclear translocation in RAW 264.7 cells were impaired (−52.32 %, p < 0.001) (Fig. 5C and D) by immunofluorescence. A notable suppression (−68.97 %, p < 0.001) of Arg-1 expression in macrophages was observed after SRC inhibition.

Concurrently, there was a significant decrease (−80.67 %, p < 0.001) in p-STAT6 expression, while the expression levels of total STAT6 and Integrin-β remained unchanged (Fig. 5E). Flow cytometry results demonstrated the effect of SRC inhibition on mechanically induced macrophage polarization (Fig. 5F). There were no significant differences in cell proportions of CD206⁺/CD86^- cells between NC and sraracatinib group, which meant saracatinib had no impact on basic level of CD206+ cells under static environment. The proportion of CD206⁺/CD86^- cells in Saracatinib + ST group were significantly decreased (−72.00 %, p < 0.001) compared to the ST group.

GLPG0187 was used as a specific inhibitor of integrin-β1 at an in vitro concentration of 2 nM. By comparing the stretch(+)/GLPG0187(−) group with the stretch (+)/GLPG0187(+) group, we observed that inhibition of integrin-β1 significantly suppressed mechanical stress-induced phosphorylation of both SRC and STAT6 (Fig. S6 A). These results indicate that integrin acts as an upstream mechanosensor in the signaling pathway. To address the possible issue regarding potential off-target effects of Saracatinib, we have performed additional siRNA knockdown experiments targeting SRC. Before the formal experiments, the efficiency of siRNA (si-1503, si-1333, si-766) was validated, and the results were presented in Fig. S6 B-C. The Western blot analysis confirmed efficient SRC interference. Importantly, the phenotypic outcomes following SRC knockdown were consistent with those observed with Saracatinib treatment, further supporting the specificity of the inhibitor and the crucial role of SRC in our experimental context. These new data have been incorporated into the revised manuscript. We believe these results have significantly strengthened the reliability of our conclusions.

These findings suggest that SRC was a key molecule in the process of outside-in mechanotransduction and mediated the phosphorylation of STAT6 induced by Integrin-β activation under cyclic stretch.

2.6. The SRC inhibitor indirectly diminishes the osteogenic effects of conditioned medium

To further assess the influence of SRC inhibition of macrophages on the osteogenic environment, we designed the experiment with five groups. NC, ST and Mφ groups were established as same as grouping strategy described in the 3rd part of results. For Saracatinib group, conditioned medium consisted of 50 %(v/v) OIM and 50 %(v/v) supernatant collected from Saracatinib treated RAW 264.7. For Saracatinib + ST group, conditioned medium was consisted of 50 %(v/v) OIM and 50 %(v/v) supernatant collected from RAW 264.7 cell treated by both Saracatinib and mechanical stretch.

Results of the ALP staining (Fig. 6A) and ARS staining (Fig. 6B) of the Saracatinib group were not statistically higher than that of the Mφ group. Western blot results show that the levels of Col-1, Runx2, and OCN in the Saracatinib + ST group were not higher than those in the Mφ group (Fig. 6C). Results of ALP, ARS and western blot demonstrated that the use of the inhibitor dramatically abolished the osteogenic effect of the conditioned medium from mechanical stimulation-induced macrophages.

Fig. 6.

The SRC inhibitor indirectly diminishes the osteogenic effects of conditioned medium. (A) Representative figures of ALP staining as well as the statistical results in NC, Mφ, ST, Saracatinib and ST + Saracatinib group. (B) Representative figures of ARS staining and quantification results among different groups. (C) Relative protein levels of osteogenic markers, Col-1, Runx2, and OCN among different groups. n = 4, ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001 by two-way ANOVA and Tukey's post hoc tests.

2.7. SRC inhibitor impairs bone consolidation in DO, which can be rescued by TGF-β in situ injection

To study the role of SRC in the DO model, Saracatinib was locally injected during the distraction period following the method described in the method part. According to our results of single cell sequencing GO analysis, results of ELISA, and relevant literature reports, we hypothesized that TGF-β may be a secretory cytokine through which mechanically induced M2 macrophages exert their osteogenic effects. Based on our hypothesis, 3 groups were designed: DO group in which DO was performed in rats, Saracatinib group in which Saracatinib was injected into fracture site, and TGF-βgroup in which TGF-βwas injection to rescue SRC inhibition. Results of micro-CT analysis revealed that the total volume in the Saracatinib group significantly increased, while the bone volume showed no significant change compared to the DO group (Fig. 7A). The bone volume to total volume ratio (BV/TV) (−35.89 %, p < 0.001) and bone mineral density (BMD) (−33.19 %, p < 0.001) in Saracatinib group were significantly lower than that in the DO group (Fig. 7B). However, the use of TGF-β rescued the decrease in BV/TV and BMD. As shown in Fig. 7B, the Saracatinib group exhibited a higher presence of fibrous tissue rather than bone and cartilage at POD 42. Immunohistochemistry results indicate a significant decrease (−64.98 %, p < 0.001) in the number of CD206+/CD86-cells in the Saracatinib group, while the number of iNOS⁺ cells showed no striking decrease compared to the other two groups. In situ injection of TGF-β reversed these changes. Results of the mechanical testing revealed that compared to the DO group, the Saracatinib group exhibited a significant impair in stress@5 % (−44.25 %, p < 0.001), force@5 % (−57.15 %, p < 0.001), and the force at final fracture (−49.86 %, p < 0.001), while the TGF-β group showed no difference compared to the DO group (Fig. 7C). Curves showing the force-position relationship were provided in the Fig. S7.

Fig. 7.

SRC inhibitor impairs bone consolidation in DO, which can be rescued by TGF-β in situ injection. (A) Results of Micro-CT analysis showing bone volume (BV), total volume (TV), BV/TV, and bone mineral density (BMD) in different groups. (B) Histological and immunohistochemical results as well as statistical analysis results indicating percentage of CD206⁺, iNOS⁺ and OCN⁺ cells at bone callus in DO, Saracatinib and TGF-β group. (C) Results of mechanical tests demonstrating the mechanical data (stress at 5 %, force at 5 % and ultimate force) of the femur after different treatments. n = 8, ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001 by one-way ANOVA and Tukey's post hoc tests.

To validate the role of the Integrin-SRC signaling pathway in mechanically induced macrophage polarization and bone regeneration during DO, we performed IHC staining and quantitative analysis on samples collected 28 days after distraction. The results showed no statistically significant difference in Integrin-β expression among the DO group, Saracatinib group, and TGF-β group (Fig. S8). This may be attributed to the return of integrin expression to baseline levels following the cessation of mechanical stimulation. Subsequently, we quantified the expression of p-SRC (Fig. S8). To identify macrophages in vivo, immunofluorescence staining was performed. As shown in Fig. S9, we found a small number of cells co-expressing CD68⁺ and integrin-β⁺ surrounding the surface of newly formed bone in the DO group. These specific cells were exclusively presented in the DO group and were not observed in the saracatinib or TGF-β groups.

In summary, the results indicated that mechanical stretch exerted by DO is sensed by Integrin on the surface of the macrophages’ cell membrane, leading to the phosphorylation of downstream molecule, SRC. p-SRC phosphorylates STAT6 in the cytoplasm. Subsequently, p-STAT6 translocates to the nucleus to promote the expression of genes associated with M2 polarization. Mechanically induced M2 polarized macrophages can secrete cytokines such as TGF-β, directly promoting osteoblast osteogenic differentiation and assisting in maintaining their own polarized state.

3. Discussion

This study suggests a pivotal role of macrophages in transducing mechanical forces into regenerative outcomes during DO. Three key findings emerge in the study: first, Mechanical stimuli promote macrophage polarization during DO. Second, the Integrin-SRC-STAT6 signaling axis has been identified as a critical link between M2 macrophage polarization and mechanical stimulation. Third, polarized macrophages orchestrate the local immune microenvironment, promoting osteogenesis through the secretion of various cytokines like TGF-β during DO.

Based on the IHC results of the DO and fracture groups, we concluded that mechanical stretching may significantly enhance macrophage M2 polarization. This effect initiates from POD 14 and remains significant for up to 4 weeks after the end of the distraction phase. Interestingly, IHC results also suggest that DO concurrently promotes macrophage M1 polarization during the distraction phase. This paradox effect could potentially be attributed to the release of inflammatory factors and the hypoxic microenvironment due to microinjuries and insufficient blood supply caused by constantly mechanical stretching. And pro-inflammatory factors and hypoxia facilitate the polarization of macrophages towards the M1 phenotype. This finding is consistent with the pseudo-temporal analysis of single-cell sequencing results and other research [46].

In this study, we employed the same osteotomy procedure in the fracture model as that in the DO model, rather than a closed fracture procedure. This approach maximally preserved the surrounding soft tissue, thereby ensuring a consistent initial inflammatory milieu and allowing us to distinguish the effect of the mechanical environment from confounding factors. Compared to the more complicated polarization processes in vivo, results from in vitro experiments exhibit much clearer trends. We conducted the in vitro stretching experiment in an LPS-induced inflammation of macrophages, showing consistently M2 polarization.

Pro-inflammation or anti-inflammation? The impact of DO on the surrounding immune environment remains an ongoing debate. Two theories concerning this issue separately describe the influence of DO on the local immune microenvironment at different time points. One theory suggests that DO procedure induces bone callus tissue damage triggering inflammation and hypoxia [47], which initiate the tissue repair process. Inflammation and hypoxia are considered as two key mechanisms by which DO promotes tissue regeneration, particularly the coupling of "angiogenesis-osteogenesis" processes [21,48,49]. The other theory in recent years posits that during DO, mechanical stretch can be transduced from the tissue into the cell through mechanical structures, such as ECM-cell connections and mechanobiological signal transduction systems, which convert mechanical signals to biochemical signals, thereby influencing the functions of cells within this environment [16,50,51].

This study provides a new perspective on understanding the immune microenvironment in DO that these two biological processes may not occur alternately but could potentially coexist. Flow cytometry detected approximately 2 % of cells in the Q4 co-expressing CD86 (M1 marker) and CD206 (M2 marker). These flow cytometry results suggest that mechanical stretch in vitro may induce a mixed M1/M2 phenotypic state in macrophages. After mechanical distraction, both M1 and M2 macrophages increase simultaneously, indicating a delicate balance of pro-inflammatory and anti-inflammatory responses at the site of DO. At a macroscopic scale, mechanical stimuli cause micro-damage, leading to inflammation and hypoxia, recruiting macrophages, and promoting the polarization of some macrophages into the M1 phenotype. At the microscopic level, mechanical stimulation-induced M2 polarization can inhibit excessive inflammation and promote recruited macrophages differentiating towards the M2 phenotype. Newly generated M2 macrophages can secrete growth factors, further facilitating tissue regeneration.

Under physiological conditions, there is a natural conversion from M1 to M2, which is reinforced by the mechanical stimulation of DO. Through cycles of disruption and rebalance, tissue regenerates continually. Rapid distraction is known to result in poor bone formation [7]. According to our findings, rapid distraction leads to more severe tissue damage, increased release of inflammatory factors, and a higher percentage of M1 macrophages. The increase in M1 macrophages is a quick response to injury. In contrast, mechanical stimulation-induced M2 polarization in macrophages is a relatively slower process. This imbalance between M1 and M2 cells hinders fracture healing through extending the inflammatory phase and preventing covert to the soft callus phase. This M1-M2 dynamic equilibrium offers a fresh perspective on understanding the changes of surrounding immunoenvironment during DO.

Integrins play a vital role in cellular processes, such as cell adhesion, phagocytosis, and mechanosensing [45,52,53]. Integrins are heterodimeric transmembrane receptors composed of one α subunit and one β subunit with a function of phosphorylating downstream molecules [54,55]. Mechanotransduction mediated by integrin can be categorized into two types: outside-in, where signals from the extracellular matrix (ECM) are sensed and transmitted into the cell, and inside-out, where intracellular signals are received and alter cell mechanical properties through conformational changes [56]. Its downstream molecules include FAK, MAPK, SRC, and others [56]. Integrins activate downstream signaling pathways by phosphorylating downstream kinases. In this study, integrin altered the levels of p-SRC in macrophages through its mechanotransduction function. Moreover, mechanical stretching increased the protein levels of integrins in RAW 264.7, indicating that macrophages can respond to mechanical stimuli with an “inside-out” response.

The SRC family kinases act as the mechanotransmitter, receiving signals from integrins and activating downstream signaling pathways. The SRC protein comprises three domains: the kinase domain, SH-2 domain, and SH-3 domain [57]. There are two crucial phosphorylation sites on the SRC protein, Y416 and Y527. Phosphorylation at the Y527 site inhibits SRC activity, while phosphorylation at the Y416 site activates SRC [58]. In this study, phosphorylation of SRC occurred specifically at the Y416 site. Except for phosphorylation as the dominant regulatory mechanism, SRC can exert immunosuppressive effects by inhibiting histone acetylation [59]. SRC is closely linked to integrins, which phosphorylate SRC at the Y416 site through both FAK-dependent and FAK-independent pathways [57]. p-SRC not only regulates the activity of downstream signaling pathways such as NF-kappaB, AKT2, but can also provide feedback to integrins, adjusting their conformation to adapt to mechanical stimulation [54].

SRC plays a direct and indirect regulatory role in macrophage polarization. In this study, SRC, as a downstream molecule of integrins, mediates mechanical stretching, directly promoting M2 polarization of macrophages. Previous research by Hu et al. indicates that SRC mediates M2 polarization of bone marrow-derived macrophages and inhibits M1 polarization. The use of SRC inhibitors leads to more severe inflammatory responses [60]. Another research by Sirin et al. indicates that phosphorylation of SRC can promote M2 polarization via Stat3 [61]. A recent study by Mu et al. demonstrates that SRC induces M2 polarization through the STAT6 pathway rather than the STAT3 pathway [62]. Additionally, in Yong June Choi et al.'s study on the tumor microenvironment, researchers found that inhibiting SRC expression in fibroblasts can suppress M2 polarization of macrophages in the same environment, suggesting an indirect role of SRC in promoting M2 polarization of macrophages in vivo [63]. However, we also note some studies that propose different viewpoints from the above [64,65]. Jang et al. observed that the use of a SRC inhibitor produces anti-inflammatory effects, suggesting SRC promotes M1 polarization [64]. We believe that a potential reason for these discrepancies is attributed to the different mechanical environments where macrophages exist. Besides acting as mechanotransmitter, SRC can also act as a sensor directly sensing mechanical stimuli [57]. These findings indicate that SRC plays a crucial role in macrophage polarization induced by mechanical stimulation.

STAT6 is an essential transcription factor that plays a crucial regulatory role in DNA transcription and is widely known for mediating IL-4-induced immune responses [66]. The main parts of the STAT6 protein include the transactivation (TAD) domain, SH-2 domain and other domains. The TAD domain helps in DNA transcription and the SH-2 domain serves as a component that can be recognized by SFK and other kinases [67]. Activation of the STAT6 signaling pathway promotes macrophage M2 polarization. Recent research by Seong Mi Ji et al. reported that melatonin promotes M2 polarization in RAW 264.7 cells through the STAT6 pathway [68]. RUNX1 may be an upstream regulator of STAT6 because reduced RUNX1 levels lead to increased STAT6 phosphorylation, thus promoting M2 polarization [69]. Additionally, Deng et al.'s study suggests that the circATP8A1 induces macrophage M2 polarization through the STAT6 pathway rather than the Stat3 pathway [70]. Research by Sukka et al. on myocardial fibrosis indicates that STAT6 exhibits mechanosensitivity [71]. Our study provides the first evidence that nuclear-cytoplasmic translocation of p-STAT6 may be a key biological process underlying mechanical-induced M2 polarization in the DO procedure.

This article primarily focuses on the process of mechanical-induced M2 polarization and provides a primary exploration of the impact of M2 cells on osteoblast differentiation. Research suggest that TGF-β can promote osteogenesis both by direct effects [17,72] and by indirectly influencing the surrounding environment to promote bone formation [[73], [74], [75]].

Current results showed that the M2 polarization of macrophages is enhanced after the use of exogenous TGF-β, indicating that TGF-β activates signaling pathways related to M2 polarization through non-mechanical pathways. This process may additively strengthen mechanical-induced M2 polarization through autocrine mechanisms. Additionally, the results of cell–cell communications provide a contextual basis for the observed regulatory role of mechanical cues in macrophage-osteoblast crosstalk. However, it is noteworthy that many studies indicate that TGF-β also directly promotes osteoblast differentiation. Based on these findings, we believe that mechanically-induced M2 polarization favors osteogenesis and bone regeneration, although the specific mechanisms underlying its effects are yet to be further investigated.

There are several limitations in this study. First, concerning the animal models, the control settings in animal models are relatively simplistic, which is not conducive to a comprehensive investigation of the effects of mechanical stimulation on DO. In future studies, establishing different experimental groups may be a better strategy, as demonstrated in the research by Ransom et al. [76]. The spatiotemporally-specific knockout of Src in macrophages using the Cre-loxP system will help to further elucidate the underlying mechanism. Additionally, although for DO, the bone regeneration process appears to be more critical than bone resorption (compared to osteoporosis or remodeling phase in the process of fracture healing), studying the role of osteoclasts still contributes to a more comprehensive understanding of this research. Third, while we have observed that exogenous TGF-β can effectively rescue impaired bone regeneration resulting from SRC inhibition, because the possibility of TGF-β directly acting on osteoblasts was not excluded, the mechanism through which TGF-β promotes osteogenesis remains unclear. Finally, while TGF-β served as a major focus based on its established role in tissue regeneration and bioinformatic evidence, we acknowledge that other soluble mediators may also participate in the macrophage-osteoblast crosstalk.

In conclusion, this study elucidates a novel mechano-immunoregulatory role of macrophages in bone regeneration, highlighting the therapeutic potential of targeting mechanotransduction pathways of macrophage to optimize bone regenerative outcomes. This study innovatively reveals a novel mechanism by which mechanical force directly drives M2 macrophage polarization through the integrin-β1/SRC/STAT6 signaling axis, and identifies TGF-β1 as an effector molecule that executes the pro-osteogenic function of this pathway. This discovery not only confirms the core concept of "mechano-immunoregulation" but also provides clear therapeutic targets for clinical translation: targeting this pathway or combining with TGF-β application may lead to the development of targeted drugs or biologics for refractory nonunion and critical bone defects, thereby enhancing the efficacy of DO or accelerating bone healing.

4. Materials and methods

4.1. Animal experiments

All animal experimental protocols were approved by the Animal Experimental Ethical Committee of the Chinese University of Hong Kong (CUHK) (Approval No. 21/100/ITF). 6–8-week-old male SD rats were acquired from the animal center of the CUHK. The rats were held under around 20 °C cages with free access to water food and entertainment facilities. DO animal model was performed as follows: Under general anesthesia, the femur of SD rats was exposed, screws were placed into the femur and the external fixator was firmly assembled. For the distraction of the bone, we performed femoral osteotomy and marked the initial position. A period of 7 days was dedicated to waiting for callus formation. Subsequently, a continuous distraction was applied for 5 days at a rate of 0.25 mm per 12 h. The sham control group involved establishing a femur fracture model with internal fixation through a stainless-steel k-wire (Ø = 1.0 mm). To minimize confounding variables from soft tissue injury and subsequent inflammation, we created fractures in the control group using a standardized open osteotomy with a wire saw. Basically, the fracture and DO models shared the similar osteotomy procedures, while fracture remained fixation only without distraction procedure. Samples were collected at postoperative day (POD) 7, 14, 28, and 42. For inhibiting activity of SRC, Saracatinib (0.1 mg/kg) was injected in situ at the distraction site at POD 8, 10, 12. At POD 9, 11, and 13, TGF-beta administration (0.1 ug/kg) was conducted by in situ injection at the distraction site.

4.2. H & E staining, Masson's staining, and immunohistochemistry (IHC) staining

The samples were fixed in a 4 % paraformaldehyde solution overnight. Subsequently, the fixed samples were decalcified in a 0.5M EDTA solution at room temperature for 4 weeks. Following decalcification, the samples underwent dehydration and were embedded in paraffin. 5 mm sections were prepared for H & E staining, Masson's staining and IHC staining. H & E and Masson's staining procedures were carried out in accordance with the manufacturer's protocols. For immunohistochemistry staining, the sections were initially blocked with 1 % goat serum, followed by overnight incubation at 4 °C with primary antibodies. The antibodies employed in this study included: CD206 (1:500, GB113497, Servicebio, Wuhan, China), iNOS (1:500, GB11119, Servicebio, Wuhan, China), OCN (GB11233, Servicebio, Wuhan, China), Integrin β1(1:500, GB115173, Servicebio, Wuhan, China) and p-SRC (1:500, ab4816, Abcam, Cambridge, UK). After primary antibody incubation, the sections were treated with goat anti-rabbit secondary antibody (G1213, Servicebio, Wuhan, China) and visualized using a diaminobenzidine (DBA) staining kit (G1212, Servicebio, Wuhan, China). Quantitative analysis for IHC was conducted by ImageJ.

4.3. Cell cultrue and mechanical stimulation in vitro

The MC3T3-E1 cells were cultured in α-minimum essential medium (α-MEM, Gibco, Carlsbad, CA, USA), which contained 10 % fetal bovine serum (FBS, Gibco, Carlsbad, CA, USA) and 1 % penicillin-streptomycin (Gibco, Carlsbad, CA, USA), at 37 °C in a 5 % CO2 incubator. RAW 264.7 cells were cultured in Dulbecco's modified Eagle's medium (DMEM, Gibco, Carlsbad, CA, USA) with 10 % FBS and 1 % penicillin-streptomycin at 37 °C in a 5 % CO2 incubator. An in vitro culture system, the Cell Tank (SHINNING TECH, Hangzhou, China), was employed to apply cyclic mechanical stretch under specific parameters (10 % strain, 0.5Hz frequency, uniaxial direction [17]). Cells were seeded onto pre-coated flexible chambers (SHINNING TECH, Hangzhou, China) and allowed to adhere overnight before stretching. Saracatinib (MedChemExpress, Monmouth, USA) was used to inhibit the expression of SRC both in vitro and in vivo. GLPG0187 (MedChemExpress, Monmouth, USA) was used to inhibit the expression of Integrin-β in vitro. To find the optimized concentration in the cell experiments, we cultured RAW 264.7 cell on 96-well plate and treated cells with gradient concentration (1 mM, 5 μM, 10 μM, 50 μM). CCK-8 assay was conducted to evaluate the impact of Saracatinib on cell viability. Western blot analysis was employed to assess the efficacy of the inhibitor and siRNA, to confirm the optimal concentration of inhibitor or measure the efficiency of siRNA. Three siRNA sequences, si-1503, si-1333, and si-766 (sequences provided in Table S2), were synthesized to knock down SRC expression. Their interference efficiency was validated prior to formal experiments.

4.4. Indirect co-culture and conditioned medium

To explore the impact of macrophages on osteoblast function under mechanical stimulation, we utilized a co-culture system. The supernatant from RAW 264.7 cells which were subjected to 12 h of cyclic stretch as the parameters shown in Fig. 2A, was collected, filtered through a 0.22 μm filter, and mixed with an equal volume of osteogenic induction medium (OIM) to create conditioned medium for osteoblast culture (Fig. 3A).

4.5. RNA extraction and quantitative real-time PCR (qPCR)

Total RNA was extracted using Trizol (Takara, Japan). The plates or flexible chambers containing the cells were rinsed twice with ice-cold PBS. After removing the residual PBS, Trizol was applied to the surface and allowed to incubate for 5 min to lyse the cells. For RNA extraction from tissues, the brain and meninges were homogenized with liquid nitrogen and lysed using Trizol. Purified RNA was extracted following the Takara manufacturer's protocol. A reverse transcription kit (Takara, Japan) was utilized to generate cDNA, and the cDNA concentration was quantified using the Nanodrop 2000 system. Subsequently, 500 ng of cDNA in 1 μL DEPC-treated water was combined with well-designed primers (Thermo Fisher, USA) and SYBR Green (Thermo Fisher, USA) in a microtube and conducted the amplify on Thermo Fisher's PCR machine. The primers were specifically designed to amplify the target genes: IL-10, TGF-β, TNF-α, CD206, IL-6, Runx2, Osx, and Alp. Every target gene was amplified in 3 replicates to minimize errors and normalized to the expression of the housekeeping gene, glyceraldehyde-3-phosphate-dehydrogenase (GAPDH). Subsequently, the normalized expressions of the target genes were compared to the negative control group to obtain the relative mRNA expressions of the target genes. The specific primer sequences are described in Table S1.

4.6. Western blot

The RIPA Lysis Buffer (Servicebio, Wuhan, China) and 5x loading buffer (Servicebio, Wuhan, China) were mixed in a 4:1 vol ratio to prepare the working solution. Cocktail inhibitors (Servicebio, Wuhan, China) were added to prevent protein degradation. Cells cultured in flexible chambers or 12-well plates were harvested and lysed using the aforementioned working solution. The cell lysis solution was heated in a metal bath at 100 °C for 8 min. The proteins were separated through sodium dodecyl sulfate polyacrylamide (SDS-page) gel electrophoresis using a Bio-Rad electrophoresis system, and then the proteins were transferred onto a polyvinylidene fluoride (PVDF) membrane. The PVDF membrane with blots was blocked in the tris-buffered saline with 5 %(w/v) skim milk powder at room temperature for 1 h. And then, Incubated the membrane with the primary antibody against the target protein overnight at 4 °C, which included iNOS (1:1000,ab15323, Abcam, Cambridge, UK), Integrin-β (1:1000, #4706, Cell Signaling Technology, Danvers, USA), STAT6 (1:1000, # 5397, Cell Signaling Technology, Danvers, USA), p-STAT6 (1:1000, #56554, Cell Signaling Technology, Danvers, USA), SRC (1:1000, #2109, Cell Signaling Technology, Danvers, USA), p-SRC (1:1000, #6943, Cell Signaling Technology, Danvers, USA), Arg-1 (1:500, PTM-5730, PTMBio, Hangzhou, China), GAPDH (1:5000, GB11002, Servicebio, Wuhan, China), Col-1 (1:1000, ab88147, Abcam, Cambridge, UK), Runx-2 (1:1000, #12556, Cell Signaling Technology, Danvers, USA), OCN (1:1000, AF6300, Beyotime, Shanghai, China). After incubation with the primary antibody, the membrane was washed 3 times and then incubated with the secondary antibody for 2 h at room temperature. Subsequently, visualization was performed using an enhanced chemiluminescence reagent (Beyotime, Shanghai, China). The blots were visualized using Bio-Rad ChemiDoc Imaging System (Bio-Rad, California, USA) and quantified using ImageJ software (NIH, Bethesda, MD, USA).

4.7. Cell scratching and wound-healing assay

MC3T3-E1 cells were seeded onto flexible chambers and cultured in conditioned medium until reaching 80 % confluence. The sterile pipette tips were used to scratch the membrane, and then the detached cells were washed away with PBS. Subsequently, the medium was replaced with α-MEM without FBS. Wounds were photographed by microscopy at 0 h,24h and 48 h. The total scratch area and the area healed after 24 h were measured using ImageJ and the percentage of healed area was calculated. Wound healing index(24/48h) = (scratch area at 0h)- (scratch area at 24/48h)/scratch area at 0h.

4.8. Flow cytometry analysis

The cells seeded on flexible chambers were gently detached using pre-cooled PBS and collected into 1.5 ml Ep tubes. The cells were then centrifuged at 1000 rpm for 5 min at 4 °C to enrich them. The cells were re-suspended in staining buffer (Thermo Fisher, USA) and washed three times. Subsequently, the cells were fixed and permeabilized using 1X fixation/permeabilization solution (Thermo Fisher, USA). Afterward, the cells were washed and incubated together with antibodies on ice in the dark for 1 h. To investigate the polarization of macrophages under mechanical stimulation, the following antibodies were used: FITC Anti-Mouse F4/80 Antibody (Elabscience Biotechnology, Wuhan, China),APC Anti-Mouse CD86 Antibody (Elabscience Biotechnology, Wuhan, China),PE/Cyanine7 Anti-Mouse CD206/MMR Antibody (Elabscience Biotechnology, Wuhan, China). The stained cells were analyzed using flow cytometry (Beckman Coulter, USA) and the data were processed using FlowJo software (Tree Star, Ashland, USA) for visualization and subsequent analysis.

4.9. Alkaline phosphatase (ALP) staining and alizarin red S (ARS) staining

The cells on the 24-well plates were gently washed 3 times with PBS. Subsequently, the cells were fixed with 4 % PFA for 5 min at room temperature and were washed 3 times with PBS again. For ALP staining, a staining working solution was prepared by combining 3 ml of staining buffer, 20 μl of NBT, and 10 μl of BICP according to the specified proportions in the ALP staining assay (Beyotime, Shanghai, China). 200 μl of the ALP staining solution was added to each well. The plates were then left 2h at 37 °C for ALP staining. 200 μl of ARS staining working solution was added to each well. The ARS staining was completed in 5 min at room temperature. The stained wells were then washed three times with deionized water. The stained 24-well plates were scanned (EPSON, Japan), and ImageJ was employed for subsequent quantitative analysis.

4.10. ELISA

The supernatant of the treated RAW 264.7 cells was collected. The levels of TGF-β (Elabscience Biotechnology, Wuhan, China), IL-10 (Elabscience Biotechnology, Wuhan, China), LRP-1 (Biotechne, Minneapolis, USA), and TNF-α (Elabscience Biotechnology, Wuhan, China) in the supernatant were detected according to the manufacturer's instruction of ELISA kit.

4.11. Immunofluorescence

We employed immunofluorescence to observe changes of the cellular cytoskeleton under cyclic stretching in vitro, as well as the distribution of key molecules within the cells. Cells were cultured on circular cover grasses (Φ = 9 mm, Citotest, China) or Flexible chambers. Cells were incubated with primary antibodies against β tubulin (1:200, AF1216, Beyotime, Shanghai, China) and p-STAT6 (1:500, #56554, Cell Signaling Technology, Danvers, USA) overnight to visualize the cellular cytoskeleton and the intracellular distribution of p-STAT6. The cells were incubated with secondary antibodies against different species at room temperature for 2 h. The anti-fading reagent (thermo Fisher, USA) containing DAPI was utilized to stain the cytonuclear and prevent fluorescence from fading. For immunofluorescence staining, 7-μm sections embedded in paraffin were used. Following deparaffinization, the sections were subjected to antigen retrieval by incubation in citrate buffer (G1201, Servicebio, Wuhan, China) at 65 °C for 2 h. After blocking with 5 % BSA (11021029, Thermo Fisher, USA) at room temperature for 45 min, the sections were incubated with primary antibodies against CD68 (1:400, Ab31630, Abcam, Cambridge, UK) and integrin-β (1:400, Ab183666, Abcam, Cambridge, UK) overnight at 4 °C. This was followed by a 2-h incubation with corresponding species-specific secondary antibodies at room temperature. Finally, the sections were mounted with an anti-fade reagent containing DAPI (Thermo Fisher, USA) to stain the nuclei and minimize fluorescence quenching. Confocal microscope (Leica, Germany) was used for obtaining images and LASX (Leica, Germany) and ImageJ software was used for exporting images and semi-quantitative analysis.

4.12. Sc-RNA sequencing data analysis

Sc-RNA sequencing data of distraction osteogenesis sample and 11 sham operation sample were downloaded from the Gene Expression Omnibus (GEO) database (GSE169592). We filtered out the low-quality cells with the following criteria: (1) < 200 expressed genes, (2) > 20 % UMIs derived from the mitochondrial genome, (3) > 6000 expressed genes. The sc-RNA sequencing count data was normalized by NormalizeData function. The 2,000 highly variables genes (HVGs) in cells were identified using FindVariableFeatures function. Subsequently, batch effects were removed, and the object was integrated using Harmony (version 1.2.0). The FindNeighbors and FindClusters functions were performed at a clustering resolution of 0.6, and a total of 10 clusters were identified and visualized. The marker genes of each cluster were identified by FindAllMarkers function with the default parameter. Finally, the cell clusters were annotated based on the canonical cellular markers from CellMarker database (http://117.50.127.228/CellMarker/) The macrophage cell types were extracted from the integrated dataset to perform further analysis, and data were preprocessed in the same way as described above. Differentially expressed genes (DEGs) between two groups were determined by FindMarkers function with the default parameter. Gene Ontology (GO) functional enrichment analysis (Biological Process) of DEGs were carried out using clusterProfiler R package. The cut-off q < 0.05 was set to discover significantly enriched GO terms. Gene Set Enrichment Analysis (GSEA) was performed using GSEA R package. Mechanoresponsive genes include: cytoskeleton-related genes such as Actb, Actg1, Tubb, Tuba1b, Vim, Des, Krt18, Krt19, Krt8, Actn1, Actn4, Flna, Flnb, Flnc; extracellular matrix-related genes such as Col1a1, Col1a2, Col3a1, Col4a1, Col4a2, Col5a1, Col6a1, Fn1, Lama1, Lamb1, Lamc1, Sparc, Tnc, Tnr, Eln; integrin family genes such as Itga1, Itga2, Itga3, Itga4, Itga5, Itga6, Itga7, Itga8, Itga9, Itga10, Itga11, Itgav, Itgb1, Itgb2, Itgb3, Itgb4, Itgb5, Itgb6, Itgb7, Itgb8; adhesion molecule genes such as Cdh1, Cdh2, Cdh5, Pecam1, Vcam1, Icam1, Icam2, Ncam1; mechanosensitive ion channel genes such as Piezo1, Piezo2, Trpv4, Trpc1, Trpc6, Kcnk2, Kcnk4; mechanical signal transduction genes such as Yap1, Wwtr1, Src, Ctgf, Cyr61, Rhoa, Rock1, Rock2, Myl9, Myl12a, Myl12b, Myh9, Myh10, Myh11; matrix metalloproteinase genes such as Mmp2, Mmp9, Mmp14, Mmp1, Mmp3, Timp1, Timp2; mechanotransduction pathway genes such as Tgfb1, Tgfb2, Tgfb3, Smad2, Smad3, Smad4, Fak, Ptk2, Vcl, Tln1, Pxn, Zyx; and other mechanics-related genes such as Lox, Loxl1, Loxl2, Fbn1, Emilin1, Mgp. The mechanical scoring of clusters is based on this gene set and is used to identify mechanosensitive cells. The CellChat package was employed to analyze cell–cell communication and visualize receptor–ligand interactions.

4.13. Mechanical tests

Femurs were harvested at POD 42 by fully dislocating the hip and knee joints. After removal of external fixation devices, specimens were immersed in 0.9 % sodium chloride solution. The three-point bending test was conducted to assess mechanical properties of the femur at 42 days post-operation. The mechanical testing device (H25KS, Hounsfield Test Equipment Ltd., UK) consisted of a mechanical loading component, a sample holder, and a sensor. Samples were placed on the holder. A 250N sensor was connected to the mechanical loading component to record data. Data was presented and analyzed using Horizon software (Hounsfield Test Equipment Ltd., UK). Stress at 5 % strain(stress@5 %), force at 5 % strain (force@5 %), and ultimate load were measured, recorded and calculated as following formula: stress@5 % = force@5 %/original cross-section area.

4.14. Micro-CT scanning and analysis

The knee and hip joints of the samples were removed, leaving the femur shaft with only 4 nail holes above and below the distraction site as markers. The trimmed samples then packed into sample holders and underwent micro-CT scanning (Scanco Medical, Switzerland). The image acquisition process was conducted at 70 kV and 118 μA, with a scanning resolution of 10 μm/voxel. A total of 200 continuous slices surrounding the distraction site were selected as the volume of interest (VOI). 3-D reconstructions were conducted and data such as bone volume (BV), tissue volume (TV), BV/TV, and bone mineral density (BMD) of the bone callus were quantified and analyzed using Image Processing software (Scanco Medical, Switzerland).

4.15. Statistical analysis

All experiments were conducted with a minimum of four repeats. Unpaired Student's t-test was employed for statistical analysis between two datasets to determine significance of differences. One-way analysis of variance (ANOVA) and Tukey's post hoc tests were utilized to assess significance of differences among multiple groups. Two-way ANOVA along with multiple comparisons based on Tukey's post hoc tests were employed to determine significant differences among data sets in each group. When the p value < 0.05, we determined the differences to be statistically significant.

Author contributions

Conceptualization: Ling Qin, Gang Li, Sien Lin, Methodology: Haixing Wang, Xuan Lu, Linlong Li, Xu Yan, Investigation: Xu Yan, Su Han, Wu Rongjie, Yinuo Fan, Xie Lichun, Hui Chen, Visualization: Xu Yan, Zitong Li, Jiting Liu, Supervision: Sien Lin, Gang Li, Writing—original draft: Xu Yan, Writing—review & editing: Sien Lin, Haixing Wang, Xuan Lu, Shanshan Bai, Yinuo Fan

Declaration of generative artificial intelligence (AI) in scientific writing

In the research, all creative texts (including but not limited to the main body of the paper, discussions, conclusions, etc.) were independently completed by the researcher without the use of any AI tools for generation or assistance in creation. The data analysis in the study was based on the original data manually processed by the researcher, without the use of AI automation tools for data calculation, modeling, or result interpretation.

Funding

National Key R&D Program of China 2024YFA0919200

National Natural Science Foundation of China 82472454 & 82272505

National Natural Science Foundation of China 82172430

Shenzhen Science and Technology Program JCYJ20241202124908012

Natural Science Foundation of Guangdong Province 2023A1515011040

Research Grants Council of University Grants Committee Hong Kong 14119124, 14113723, 14121721, N_CUHK472/22, T13-402/17-N & AoE/M-402/20

Health and Medical Research Fund (HMRF), the Health Bureau of Hong Kong 09203436 & 08190416

Innovation and Technology Fund (ITF) of Innovation Technology Commission of Hong Kong (GHP/186/22GD)

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Data availability

All data are available in the main text or the supplementary materials.