Targeted muscle reinnervation surgery modulates apoptosis and synaptic plasticity to improve motor function following tibial nerve injury in rats

Abstract

Background

Targeted muscle reinnervation (TMR) is represents an advanced neural-machine interface that enhances prosthetic control and facilitates motor recovery in amputees. Although TMR is known to connect residual nerve fibers and supply neurotrophic factors, its impact on spinal cord motor neurons remains understudied. This study investigated the effects and possible mechanisms of TMR on spinal motor neurons in a rat model of tibial nerve transection (TNT).

Materials and methods

There were 30 Sprague Dawley rats grouped into control, TNT, and TMR groups. TMR was grafted proximal tibial nerve into the gastrocnemius muscle. Outcome measures included the sciatic functional index, the muscle wet weight ratio, muscle fibrosis via Masson’s trichrome staining, and immunohistochemical analysis of caspase-3 and Bcl-2 expression in spinal anterior horn. RT-PCR analysis of synaptic markers’ mRNA expression.

Results

The TNT group showed a marked SFI reduction, whereas the TMR group exhibited a significantly higher SFI (p < .01). Similarly, The operated muscle weight retention was preserved in the TMR relative to the TNT group (p < .01), indicating improved limb function and reduced atrophy. Masson trichrome staining demonstrated lower collagen deposition in the TMR group (p < .05). PCR analysis showed that TMR significantly downregulated spinal GAP43 mRNA (p < .05) while upregulating synapsin (SYN) and PSD-95 transcripts (both p < .05) versus TNT. Immunohistochemically, TMR decreased Bcl-2 (p < .05) and increased Caspase-3 (p < .01) expression relative to TNT.

Conclusions

These findings suggest that TMR promotes the spinal motor neuron recovery and synaptic remodelling, likely contributing to improve muscle morphology and overall post-injury functional outcomes.

Introduction

Limb loss remains a substantial public health and economic burden, with millions of individuals living with amputation worldwide and numbers projected to grow annually due to dysvascular disease and trauma [1]. Beyond the immediate loss of function, patients face lifelong risks of neuropathic pain and disability, contributing to healthcare costs often exceeding hundreds of thousands of dollars per person when factoring in repeated surgeries, rehabilitation, and lost productivity [2]. A major barrier to successful rehabilitation is the management of the injured nerve stump; traditional closure often leads to painful neuroma formation and irreversible muscle atrophy, severely limiting prosthetic fit and function. Targeted Muscle Reinnervation (TMR) offers a physiologic solution by redirecting transected nerves into nearby motor cohorts, effectively absorbing neural signals, preventing neuroma, and preserving muscle integrity for myoelectric control [3–5]. Given its potential to restore intuitive limb movement and reduce secondary complications, optimizing TMR techniques is critical for the growing amputee population [6]. Despite promising clinical outcomes, the precise molecular mechanisms by which TMR attenuates neuronal apoptosis and facilitates synaptic remodelling remain incompletely understood. While previous studies have focused on surgical technique optimization, the downstream effects on spinal cord plasticity and muscular homeostasis require further investigation.

Accumulating evidence indicates that during neural regeneration and synaptic remodelling, the efficiency of neural circuit reconstruction is determined by the coordinated expression of growth-associated protein-43 (GAP-43) – a key marker of axonal sprouting – and the formation of mature synaptic structures comprising synaptophysin (SYN) and postsynaptic density protein-95 (PSD-95) [7,8]. Furthermore, amputation-induced loss of spinal anterior horn motor neurons is frequently accompanied by apoptotic imbalance, wherein the dynamic equilibrium between Bcl-2 and Caspase-3 serves as a central hub regulating cell survival and death [6,9]. Notably, a close relationship exists between neuronal apoptosis and synaptic plasticity [10,11]. However, whether Targeted Muscle Reinnervation (TMR) can protect residual motor neurons and facilitate functional remodelling at the spinal level by upregulating neurotrophic support, coordinately modulating the expression of anti-apoptotic (Bcl-2) and pro-apoptotic (Caspase-3) proteins, and inducing the activation of synaptic plasticity-related molecules (GAP-43, SYN, and PSD-95) remains to be elucidated.

In this study, we established a rat model of nerve injury by surgically transecting the tibial nerve and the medial gastrocnemius muscle branch. Intervention was performed via TMR surgery, transplanting the tibial nerve into the medial gastrocnemius muscle belly. We evaluated the sciatic nerve function index (SFI) and the maintenance rate of wet weight in the operated medial gastrocnemius muscle, alongside the Masson collagen volume fraction. Additionally, RT-qPCR was employed to quantify the mRNA expression levels of Bcl-2, Caspase-3, GAP-43, SYN, and PSD-95in the spinal anterior horn. These assessments were designed to investigate the potential mechanism by which TMR improves motor function in the injured limb by modulating apoptosis and synaptic plasticity in residual spinal motor neurons.

Materials and methods

Experimental animals

Thirty adult Sprague Dawley (SD) rats (♂), weighed 220 ± 20 g each and classified as specific pathogen-free (SPF) grade, were sourced from Guangdong Medical Laboratory Animal Center (License number: SCXK (Yue) 2020-0002). All animals were housed in Shenzhen Institutes of Advanced Technology of the Chinese Academy of Sciences. Housing conditions maintained 22 ± 2 °C, 45% to 60% relative humidity, day/night cycle at 12h/12h. Rats underwent random assignment into the normal control, TNT, TMR groups.

In vivo experiments were approved by Laboratory Animal Ethics Committee of Zhuhai Campus of Zunyi Medical University (No. ZMU21-2203-164).

Establishment of rat TMR model

Following the procedure outlined by previous research [6], anesthetization of rats relied on inhalation of isoflurane, with an induction dose of 2%–3%. The maintenance dose of isoflurane was 1.5%–2%. Then, gastrocnemius muscle and corresponding muscular branches and tibial nerve got exposed (Figure 1(A)). Muscular branches to gastrocnemius muscle’s lateral and medial heads and tibial nerves were isolated. In normal control group, only tibial nerve, muscular branches aforementioned were exposed without further intervention (Figure 1(B,C)). In TNT group, the tibial nerve and the motor branch to the medial head of the gastrocnemius muscle were transected with ophthalmic scissors after ligation (Figure 1(B,C)). Within TMR group, following tibial nerve and medial gastrocnemius muscular branch transection, the nerve’s proximal segment was surgically implanted into medial gastrocnemius muscle belly. The tibial nerve was exposed in the lower popliteal fossa between the origins of the medial and lateral gastrocnemius heads. Transection was performed approximately 0.8 cm distal to the point where the branch to the medial gastrocnemius head arises from the tibial nerve. Nerve-muscle anastomosis was performed using epineural suture. Under a microscope, the proximal end of the tibial nerve was fixed to the muscle belly of the medial head of the gastrocnemius muscle with 10-0 suture, suturing the epineurium to the epimysium to ensure tension-free apposition between the nerve fascicles and muscle fibers (Figure 1(B,C)) [6]. After model establishment, researchers closed the skin of rats, placed them back to corresponding cages, providing all-time accessible food and water.

Figure 1.

Establishment of hind limb model in SD rat (posterior view). (A) SD rat (posterior view); (B) schematic diagram illustrating the establishment of the left hind limb model in SD rats, showing the Normal, TNT, and TMR group; (C) surgical view of the left hind limb (posterior view). Normal group: the tibial nerve and its branches were exposed; TNT group: the tibial nerve and the nerve branch innervating the medial head of the gastrocnemius muscle were transected; TMR group: the proximal end of the transected tibial nerve was sewn onto the medial belly of the gastrocnemius muscle.

At 6 weeks postoperatively, model evaluation was performed on the rats, and relevant data were recorded. The animals were then euthanized for tissue collection, specifically the gastrocnemius muscle and the lumbosacral enlargement of the spinal cord.

Perfusion and tissue collection

After behavioural testing, SD rats were deeply anesthetized with 10% chloral hydrate (5 mL/kg, i.p.). The abdominal cavity was opened, and transcardial perfusion was performed by first infusing 150 mL of normal saline through the left ventricle, followed by 4% paraformaldehyde (PF) fixative (100 mL rapidly, then 150 mL by dropwise infusion). The lumbosacral enlargement of the spinal cord was dissected, the meninges were removed, and the tissue was post-fixed in 4% PF for 6 h. The dura mater was carefully removed, and the tissue was dehydrated in 15% and 30% sucrose solutions until it sank. The medial head of the gastrocnemius muscle was similarly post-fixed in 4% PF for 24 h, followed by sucrose dehydration. Both tissues were embedded in paraffin, and 4-μm sections were cut using a microtome.

Tissue harvesting

At the designated time points, rats were overdosed with 10% chloral hydrate (5 mL/kg, i.p.) and euthanized by cervical dislocation. The ventral part of the lumbosacral spinal cord and the medial head of the gastrocnemius muscle were rapidly dissected on ice, documented, and stored at −80 °C until further analysis.

SFI

The sciatic functional index is a metric for assessing the degree of nerve injury through analysis of footprint parameters in rat gait (Figure 2(A–C)), with values ranging from 0 to −100. A value of 0 indicates normal nerve function, whereas −100 denotes complete nerve injury. The animal gait analysis system (BT60101, Shenzhen Zhongshi Co., Ltd., China) was used to record the footprints and gait of rats (Figure 2(D)). In each group, four to five footprints of both hind paws were collected. Then, the intermediate toe spread (IT), footprint length (PL) on experimental (E) and normal (N) sides, and toe width (TW) were recorded and analysed. SFI was then computed using formula below: SFI = − 38.3 × (EPL − NPL)/NPL + 109.5 × (ETW − NTW)/NTW + 13.3 × (EIT − NIT)/NIT − 8.8.

Figure 2.

Sciatic functional index (SFI) of each group. (A–C) Footprints and gait patterns of rats recorded using animal gait analysis system; (D) gait analysis and measurement; (E) bar graph showing SFI values for the three groups at 6 weeks post-surgery. *p < .05, vs. the normal control group; ^#p < .05, vs. the TNT group; TMR: targeted muscle reinnervation; TNT: tibial nerve transaction; SFI: sciatic functional index; IT: the intermediate toe spread; PL: footprint length; TW: toe width.

An SFI of 0 indicates function of normal nerve, the value of 100 indicates a complete loss of nerve function.

Gastrocnemius muscle’s wet weight ratio

Gastrocnemius muscle was isolated from both hindlimbs of rats, the fat and blood on the surface of the muscles was removed and absorbed. Then, the muscle was accurately weighed. Wet weight ratio was computed from formula below: wet weight on the side receiving operation (left side)/wet weight on normal side (right side) × 100%.

Masson’s trichrome staining

Paraffin sections from rat gastrocnemius muscles were dewaxed to water, oxidized in 1% potassium permanganate for 5 min, undergoing water rinse, bleached in oxalic acid for 1 min. Afterward, staining of sections relied on celestine blue solution for 5 min, undergoing water rinse. After the residual solution was removed, Mayer haematoxylin staining was conducted, lasting 3–5 min and then washed with running water keeping 5–10 min. Subsequently Ponceau-picric acid saturated solution was also conducted lasting 5 min, rinsed by 1% acetic acid aqueous solution, differentiated in 1% phosphomolybdic acid lasting 5 min, and washed with distilled water. Lastly, sections underwent staining with 1% toluidine blue for 30s, rinsed by 1% acetic acid aqueous solution, differentiated in 95% alcohol, dehydrated by absolute alcohol, with clearance by xylene, finally mounted by neutral balsam. All the agencies mentioned above were sourced from Sinopharm Chemical Reagent Co. Ltd., Shanghai, China.

Quantitative analysis of Masson’s trichrome-stained muscle sections

Muscle sections stained with Masson’s trichrome were imaged using a light microscope (Olympus BX53; ×200 magnification) under consistent lighting conditions. Images were exported in RGB format and analysed using ImageJ software (NIH, version 1.54f). Original RGB images were processed using the ‘Image > Color > Colour Deconvolution’ function in ImageJ to separate the blue channel (collagen fibers) and red channel (muscle fibers). For each channel, manual thresholding was performed using the Threshold tool: The pixel intensity distribution was evaluated via histogram analysis to identify the typical greyscale range for collagen (blue) and muscle fibers (red). A ‘Default’ algorithm was applied with fixed threshold levels (Lower: 0; Upper: 150), where the target structures (collagen or muscle fibers) were highlighted in red and the background set to black .For each section, ≥3 non-overlapping high-power fields (×200 magnification) were randomly selected, with priority given to central and peripheral regions of muscle bundles to minimize sampling bias.

After threshold segmentation, the following parameters were measured: Collagen area (CVF, collagen volume fraction): Calculated as the percentage of blue pixels (collagen-positive area) relative to the total field area (CVF = [collagen area/total area] × 100%). Muscle fibre area: Determined either as the proportion of red pixels or as the remaining area after subtracting collagen area from the total field area. Measurements were standardized by selecting ‘Area’ and ‘Limit to threshold’ in ‘Analyze > Set Measurements’ before quantification via ‘Analyze > Measure’. Final data for each sample were expressed as the mean ± standard deviation (SD) of the analysed fields (n = 30). All images were coded with random numbers before analysis, and the experimenter was blinded to group assignments (e.g. control vs. treatment groups) to prevent subjective bias. Group identities were revealed only after data collection and statistical analysis.

Immunohistochemical staining

The lumbosacral spinal cord segment was extracted, fixed in paraffin, and sectioned. Following deparaffinization and antigen retrieval, sections were processed by 3% hydrogen peroxide (Sinopharm Chemical Reagent Co. Ltd., Shanghai, China) for 30 min and rinsed with PBS. After membrane blockage in 3% BSA (room temperature, 30 min), sections were sent for subsequent incubation in solutions of primary antibodies (4 °C, overnight) rabbit anti-caspase-3 (1:200, Cat# ab32351, RRID: AB_725946, Abcam, Cambridge, UK), rabbit anti-Bcl2 (1:200, Cat#ab59348, RRID: AB_2064155, Abcam, Cambridge, UK). Followingly, another 50-min incubation was conducted with horseradish peroxidase–conjugated secondary antibody (1:1000, Cat# GK500705, RRID: 2895152, EnVisionTM Detection Kit, Dako, Glostrup, Denmark). Visualization relied on DAB with haematoxylin counterstaining and differentiation. The slides were then dehydrated, mounted, and microscopically examined. Image acquisition and analyses followed standardized protocols.

Fluorescence quantitative RT-PCR

Extraction of total RNA relied on lumbosacral enlargement of rat spinal cord utilizing the TRE-Trizol kit (Invitrogen, USA), the reversely-transcription relied on PrimeScript II 1st Strand CDNA Synthesis Kit (Cat#D6210A, Takara Biomedical Technology Co., Ltd., Beijing, China), all referring from professional instructions. Internal control gene was chosen as β-actin. The obtained cDNA was subjected to PCR reaction, prepared in ice-water bath. Conditions of PCR reaction conditions initiated at a denaturation step kept at 95 °C lasting 30s, then undergoing 40 denaturation cycles (95 °C, 5s), then annealing (60 °C, 45s). Amplification and quantification were conducted using ABI PRISM^® 7500 Sequence Detection System (SDSShell Software1.6, Thermo Fisher Scientific, USA). Relative expression of mRNA was computed from formula 2^−ΔΔCt. Table 1 lists primer sequences.

Table 1.

Primer sequences for RT-qPCR.

Target gene	Forward primer (5′->3′)	Reverse primer (5′->3′)	Product length
PSD95	ATCCTGTGTCCCTCATACGA	CCATAGAGGTGGCTGTTGTA	150bp
SYN	CATCTTCGCCTTTGCTACGT	CAGGAGGGTGCATCAAAGTA	150bp
GAP43	AGCTGTAGATGAAGCCAAACCT	ATCTGAGAAAGGGCAGGAGAGA	150bp
β-actin	AGGGAAATCGTGCGTGACAT	GAACCGCTCATTGCCGATAG	150bp

Statistical analysis

Statistical analyses relied on GraphPad Prism 7.0. Data expression presentation: mean ± standard deviation (SD). For data meeting assumption of homogeneity of variance, one-way analysis of variance (ANOVA) was completed to compare intra-groups, with the Bonferroni method applied for comparisons among multiple groups. Statistical significance: p value < .05.

Results

TMR surgery’s effects on motor function recovery in rats following TNT

SFI was utilized to evaluate motor functional recovery following injury in sciatic nerve relying on gait analysis. At 6 weeks post-surgery, SFI values were recorded and analysed using one-way ANOVA. The analysis revealed that rats in the normal control group exhibited clear and well-defined footprints. TNT group’s SFI results was markedly lower compared to normal control (p < .05, Figure 2(E)). Although TMR group’s SFI result remarkedly surpassed those of TNT group, it remained notably lower than that of normal control (p < .05, Figure 2).

TMR surgery’s effect on rat gastrocnemius muscle

To assess target muscle atrophy post nerve injury, gastrocnemius muscle wet weight ratio in each group was computed. Such data, meeting normal distribution assumptions and variance homogeneity, were assessed utilizing one-way ANOVA. The results indicated that wet weight ratio was significantly higher in the normal control group compared to both TNT and TMR groups (p < .05, Figure 3). Wet weight ratio on the left hind limb of the rat (operated side) was also notably higher within TMR group when compared to TNT group (p < .05, Figure 3).

Figure 3.

The wet weight ratio of the gastrocnemius muscle in rats of each group. *p < .05, vs. the normal control group; ^#p < .05, vs. the TNT group. TMR: targeted muscle reinnervation; TNT: tibial nerve transaction.

Results from gastrocnemius muscle’s Masson’s trichrome staining on the left hind limb (operative side) in each group showed that, there were only little collagen fibers in interstitial spaces between muscle fibers in normal control, the muscle cells were well-organized (Figure 4(A,D)). For both TNT (Figure 4(B,E)) and TMR groups (Figure 4(C,F)), the proliferation of collagen fibers was seen in muscle cells’ interstitial space, which was observed between the muscle bundles. The CVF evaluated by Masson’s trichrome staining was 0.126 ± 0.0137 in the normal control group, 0.537 ± 0.0452 in the TNT group, and 0.3125 ± 0.0317 in the TMR group, significant differences were noted among 3 groups (F(2,15)=235.8, p < .05, Figure 4(G)). The quantity of collagen fibre within muscle’s interstitial space in TMR group was notably elevated compared to normal control group (p < .05, Figure 4(G)), which was notably decreased than TNT group (p < .05, Figure 4(G)).

Figure 4.

Masson’s trichrome staining of the medial head of gastrocnemius muscle in each group. (A,D) Normal control group; B,E: TNT group; (C,F) TMR group; (G) bar graph showing the collagen volume fraction analysed by Masson’s trichome staining in each group. Scale bar = 100 μm, *p < .05, vs. the normal control group; #p < .05 vs. the TNT group. TMR: targeted muscle reinnervation; TNT: tibial nerve transaction.

Effects of TMR surgery on GAP43 and synaptic proteins in rat spinal cord motor neurons

To determine TMR surgery’s effects on GAP43 and synaptic proteins in rat spinal cord’s motor neurons, we evaluated mRNA expression of GAP43, SYN, and PSD-95 in each group by RT-PCR. The outcomes illustrated that mRNA expression of GAP43 within rat spinal cord’s anterior horn was notably up-regulated within TNT group than normal group (2.01 ± 0.04 vs. 1 ± 0.05, p < .05), significantly down-regulated within TMR group than TNT group (1.12 ± 0.02 vs. 2.01 ± 0.04, p < .05, Figure 5(A)).

Figure 5.

The mRNA expression levels of synaptic markers in rat spinal cord motor neurons in each group. (A) GAP43 mRNA expression in rat spinal cord motor neurons; (B) SYN mRNA expression in rat spinal cord motor neurons; (C) PSD95 mRNA expression in rat spinal cord motor neurons. *p < .05, vs. the normal control group; ^#p < .05 vs. the TNT group. TMR: targeted muscle reinnervation; TNT: tibial nerve transaction.

SYN mRNA expression in rat spinal cord’s anterior horn was notably down-regulated in TNT group than normal control (0.81 ± 0.02 vs. 1 ± 0.05, p < .05), while it was significantly up-regulated in TMR group in comparison to TNT group (0.9 ± 0.03 vs. 0.81 ± 0.02, p < .05, Figure 5(B)).

PSD-95 mRNA expression in rat spinal cord’s anterior horn notably dropped within TNT group than normal control (0.71 ± 0.03 vs, 1 ± 0.08, p < .05). Compared to TNT group, PSD-95 mRNA expression in rat spinal cord’s anterior horn was notably elevated in the TMR group (0.85 ± 0.02 vs. 0.71 ± 0.03, p < .05, Figure 5(C)).

Effect of TMR surgery on Bcl-2 expressions within rat spinal cord

Bcl-2 expression within rats’ spinal cord in each group was quantified by immunohistochemical staining. The integral optical density measurements revealed Bcl-2 levels of 0.05 ± 0.0053 in normal control, 0.1037 ± 0.01074 in TNT group, 0.078 ± 0.0097 in TMR group (Figure 6(A–G)). One-way ANOVA, after confirming the normal distribution of the data, indicated significant differences among groups (F(2,15) = 51.61, p < .05). Bcl-2 expression was significantly higher within TNT group compared to normal control group (p < .05, Figure 6(A–F)) and was notably reduced in TMR compared to TNT group (p < .05, Figure 6(G)).

Figure 6.

The expression level of Bcl-2 in rat spinal cord motor neurons in each group. (A–C) Representative images of immunohistochemical staining of Bcl-2 in spinal cord cross-sections from the normal control, TNT, and TMR groups. (D–F) High magnification images of the localized regions shown in A–C. (A,D) Normal control group; (B,E) TNT group; (C,F) TMR group. (G) Bar graph showing the Integrated Optical Density (IOD) of Bcl-2 expression in the normal control, TNT, and TMR groups. Scale bars: A–C = 300 µm; D– F= 75 µm. *p < .05, vs the normal control group; #p < .05 vs the TNT group. TMR: targeted muscle reinnervation; TNT: tibial nerve transaction; IOD: integral optical density.

Effect of TMR surgery on caspase-3 expressions within rat spinal cord

Expression level of caspase-3 was detected by immunohistochemical analysis. The data was conformed as with normal distribution (Figure 7(A–F))). The results showed statistical differences among 3 groups (F(2,15) = 83.68, p < .05). Caspase-3 expression was significantly higher in TNT group than normal group (0.0745 ± 0.0045 vs. 0.0321 ± 0.005, p < .05), notably lower in TMR group than in TNT group (0.0573 ± 0.007 vs. 0.0745 ± 0.0045, p < .05, Figure 7(G)).

Figure 7.

Caspase-3 expression in rat spinal cord motor neurons in each group. (A–C) Representative images of immunohistochemical staining of caspase-3 in the spinal anterior horn cross-sections from the normal control, TNT, and TMR groups. (D–F) High magnification images of the localized regions shown in A–C. (A,D) Normal control group; (B,E) TNT group; (C,F) TMR group. (G) Bar graph showing the IOD of caspase-3 expression in the normal control, TNT, and TMR groups. Scale bars: A–C = 300 µm; D– F= 75 µm. *p < .05, vs the normal control group; ^#p < .05 vs the TNT group. TMR: targeted muscle reinnervation; TNT: tibial nerve transaction; IOD: integral optical density.

Discussion

In our earlier research utilizing a rat TMR model, we found that TMR successfully reestablishes connections between remaining nerve fibers and targeted muscle, restoring motor function within injured hind limb, and enhances microenvironment to support neuronal cell bodies and residual nerve fibers survival. These effects contribute to the recovery of muscle nerve function within hind limb. Despite these findings, precise mechanisms through which TMR operates remain unclear. In this study, we replicated the rat TMR model using the methodology detailed in our earlier research [6]. We assessed the impact of TMR on hind-limb motor functional recovery following TNT, examined the degree of fibrosis in the target muscle post-TMR, and measured mRNA expression levels of GAP-43 and synaptic proteins (SYN and PSD-95). Additionally, we evaluated caspase-3 and Bcl-2 within the anterior horn of the rat spinal cord across TMR, TNT, normal control groups. The objective was to investigate the mechanisms by which TMR enhances motor function in the target muscle. Our results indicate that TMR surgery inhibits neuronal apoptosis by downregulating Caspase-3 and upregulating Bcl-2, while simultaneously regulating GAP-43/SYN/PSD-95 to improve synaptic plasticity, synergistically promoting the recovery of motor function.

This study used the SFI, gastrocnemius wet weight maintenance rate, and Masson’s trichrome staining to evaluate the effect of TMR surgery on motor function recovery following tibial nerve injury in rats. The SFI objectively reflects the motor function innervated by the target muscles [12,13]. Experimental results showed that after simple transection of the TNT group, the left hind limb of rats could not achieve normal plantar contact, and the SFI was significantly lower than that of the normal group, indicating severe motor dysfunction in the muscles innervated by the tibial nerve (e.g. the medial head of the gastrocnemius) due to denervation and neurotrophic disturbances. The SFI of the TMR group was significantly higher than that of the TNT group, indicating that TMR surgery helps improve hind limb motor function after injury, a finding consistent with previous studies and published reports. The gastrocnemius wet weight maintenance rate is an important indicator for assessing the degree of target muscle atrophy following nerve injury. In the TNT group, the wet weight maintenance rate of the gastrocnemius on the surgical side was significantly reduced, indicating marked atrophy of the target muscle due to denervation after tibial nerve transection; compared with the TNT group, the wet weight maintenance rate in the TMR group was significantly increased, suggesting that TMR surgery effectively alleviates denervation atrophy of the target muscle. This result further supports the positive effect of TMR surgery on motor function recovery, which is consistent with the trend observed in the SFI [14,15]. Masson’s trichrome staining showed that the collagen volume fraction of the target muscle in the TNT group was significantly higher than that in the normal group, indicating pronounced fibrosis of muscle tissue following denervation. The collagen volume fraction of the target muscle in the TMR group was significantly lower than that in the TNT group, indicating that TMR surgery reduces the degree of fibrosis in the target muscle and provides good protection of muscle tissue structure. In summary, the SFI, gastrocnemius wet weight maintenance rate, and Masson’s trichrome staining results all demonstrate that TMR surgery alleviates denervation atrophy and fibrosis of the target muscle after tibial nerve injury and promotes partial recovery of motor function in the injured hind limb of rats.

Peripheral nerve injuries often lead to damage in residual nerve fibers and neuronal cell bodies, making it crucial to understand how TMR surgery enhances the survival environment for these tissues. Neuronal apoptosis is a prominent feature of many neurodegenerative conditions, highlighting the importance of investigating the mechanisms through which TMR supports the survival of nerve fibers and neuronal cells. Studies have demonstrated that caspase-3 expression was significantly elevated [16,17] and Bcl2 expression was markedly reduced in rat spinal cord tissues after sciatic nerve injury [18,19]. Caspase-3 activation serves as an early biochemical marker of apoptosis, contributing to the morphological changes characteristic of apoptosis [20–22]. Bcl2 plays an important role in attenuating apoptosis in tissues of spinal cord following sciatic nerve injury in rats [23–25]. In this study, we detected caspase-3 and Bcl2 expressions within spinal cord’s anterior horn, and found a remarked elevation in both caspase-3 and Bcl2 within spinal cord on the side with injury post TNT. Our results were consistent with the previous studies [24–26]. Activation of caspase-3 induces occurrence of apoptosis and participates within initiation and execution of apoptosis during the early stages. Typically, caspase-3 activation occurs after 1 week of nerve injury and reaches a peak after 4–5 weeks. Persistent caspase‑3 activation mirrors the sustained oxidative stress and inflammatory response observed after spinal cord injury. Liu et al. [27] demonstrated that a hybrid conduit delivering catalase to scavenge reactive oxygen species significantly reduced cleaved caspase‑3 levels and increased neuronal survival at the lesion borders. Study outcomes illustrated that even after 6 weeks of TMR, caspase-3 expression in TNT group still surpassed normal control, suggesting that neuronal apoptosis still occurs in spinal cord’s anterior horn of rats after long periods of tibial nerve injury. The TMR group showed significantly lower caspase-3 expression compared to the TNT group, suggesting the potential inhibitory effect by TMR surgery on caspase-3 activation. Bcl2, an anti-apoptotic gene [28,29], is supported by a previous study to be weakened at the early stage of peripheral nerve injury, and then significantly elevated after 4 weeks of injury, persisting for 6 weeks [23]. Similar results were observed in this study. However, Bcl2 expression was notably weaker in the TMR group compared to the TNT group, potentially due to an improved neuronal survival microenvironment in the rat spinal cord after TMR. Since Bcl2 exerts its anti-apoptotic effect by inhibiting caspase activation, its expression tends to increase when caspase-3 levels rise. Our study revealed that Bcl-2 expression remained significantly higher in the TMR group compared to the normal control group. This finding suggests that Bcl-2 is crucial for inhibiting neuronal apoptosis and mitigating nerve tissue damage. It indicates that TMR surgery may effectively reduce neuronal apoptosis by modulating the expression levels of both caspase-3 and Bcl-2.

The apoptotic pathway and synaptic plasticity are not isolated entities; rather, they engage in profound molecular crosstalk. Emerging evidence suggests that Caspase-3 functions not only as an executioner of apoptosis but also modulates synaptic structure and function by cleaving synaptic-associated proteins (e.g. PSD-95), thereby directly influencing the efficiency of neural circuit reconstruction [30]. In this study, the TMR group exhibited concurrent suppression of Caspase-3 activity and upregulation of synaptic proteins (PSD-95 and SYN). This implies that TMR surgery may indirectly preserve synaptic proteins from proteolytic degradation by maintaining the homeostasis of the Bcl-2/Caspase-3 axis. Consequently, this provides a favourable cellular microenvironment for synaptic remodelling and supports the expression of neurotrophic factors essential for GAP-43-mediated axonal sprouting and synaptogenesis (PSD-95/SYN). This ‘anti-apoptotic–pro-plasticity’ synergistic mechanism likely represents a critical node in TMR-accelerated neurological functional recovery. Future investigations utilizing co-culture systems or specific gene knockout models are required to validate the causal interplay between these pathways.

Peripheral nerve injury and subsequent repair are known to influence motor function by modulating synaptic plasticity, the adaptive change in synaptic strength in response to experience and environmental shifts [31–35]. GAP43, a neural membrane protein implicated in regulating axonal growth and synaptic plasticity, is notably reduced in synaptic loss during cognitive disorders like Alzheimer’s disease [36–39]. Its expression surges during brain injury, neurodevelopment, and nerve regeneration, serving as a nerve regeneration biomarker [40,41] for nerve regeneration following nerve injury. In our study, RT-PCR analysis of the spinal cord’s ventral horns revealed a significant increase in GAP43 mRNA expression within TNT group, following a decrease in TMR group, although levels remained above those of normal control. It aligns with literatures suggesting that elevated GAP43 expression correlates with synaptic restructuring, with levels declining once the reconstruction process is complete [42,43]. Consistently, similar observations have been made in regenerating axons within DPSC conduits [44]. Notably, It reported that restoring synaptophysin expression was essential for re‑establishing neural circuits across a spinal cord lesion [45]. We further assessed mRNA expression levels of SYN and PSD95 proteins within spinal cord’s anterior horn. TNT group exhibited significantly reduced PSD95 and SYN mRNA expressions compared to the controls, while TMR group demonstrated a significant upregulation relative to TNT group. These findings suggest that tibial nerve injury inhibits synaptic protein expression in the anterior horn, and TMR surgery partially restores these levels, corroborating our prior findings [4]. TMR surgery significantly contributes to the protection of neuronal structures and the remodelling of synapses. Following TNT, residual neurons exhibited elevated levels of proteins such as GAP-43, which support neuron survival. Post-TMR surgery, these residual nerve fibres successfully formed connections with the target muscle, leading to a normalization of GAP-43 levels. Additionally, there was an observed increase in the expression of synaptic plasticity-related proteins (SYN and PSD95) in spinal cord neurons after surgery. This increase facilitates the formation of new neuromuscular junctions between the residual nerve fibres and the target muscle. These effects may be attributed to TMR surgery’s influence on the regulation of Bcl-2 and caspase-3 expression.

Nevertheless, this study has several limitations that warrant attention. First, although we assessed collagen deposition among muscle fibres using Masson staining, we did not quantify the degree of fat infiltration within the target muscle using specific stains such as Oil Red O. Given that fat infiltration is a critical factor affecting muscle atrophy and electromyographic signal quality, future studies are required to elucidate the specific impact of TMR on muscle composition remodelling. Second, this study focused primarily on the biological effects of the TMR procedure itself, without exploring the potential of combination therapies. In light of the cornerstone roles of rehabilitation training and electrical stimulation in neural recovery, future research should investigate the synergistic effects of combining TMR technology with specific regimens of motor training or functional electrical stimulation (FES) [46]. Such multimodal interventions are anticipated to maximize the restoration of motor function in amputees.

Conclusion

TMR surgery improves microenvironment for neuronal cell body survival, as well as enhances synaptic plasticity of motor neurons in the anterior horn of spinal cord, thereby partially restoring hind-limb motor function after injury. This restoration is hypothesized to be mediated through the regulation on Bcl2, caspase-3, and GAP-43 expression levels. Notably, the motor function in TMR-operated rats did not match that of normal control group, underscoring the urgent need for more efficacious rehabilitation strategies to augment targeted muscle recovery post-TMR. Additionally, further research is warranted to elucidate differential protein expression patterns, and the mechanisms of protein-protein interactions during nerve regeneration and repair.

Funding Statement

This work was supported by National Natural Science Foundation of China (Grants No.: 82260456), Special projects in key areas of ordinary universities in Guangdong Province (Grants No: 2023ZDZX2094), Science and Technology Planning Project of Shenzhen (Grants No.: JCYJ20230807140559047), Research Projects of Putian University (Grants No.: 2026020).

Ethical approval

In vivo experiments were approved by Laboratory Animal Ethics Committee of Zhuhai Campus of Zunyi Medical University.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.