Transplantation of a vascularized pedicle of hemisected spinal cord to establish spinal cord continuity after removal of a segment of the thoracic spinal cord: A proof‐of‐principle study in dogs

Abstract Introduction Glial scar formation impedes nerve regeneration/reinnervation after spinal cord injury (SCI); therefore, removal of scar tissue is essential for SCI treatment. Aims To investigate whether removing a spinal cord and transplanting a vascularized pedicle of hemisected spinal cord from the spinal cord caudal to the transection can restore motor function, to aid in the treatment of future clinical spinal cord injuries. We developed a canine model. After removal of a 1‐cm segment of the thoracic (T10–T11) spinal cord in eight beagles, a vascularized pedicle of hemisected spinal cord from the first 1.5 cm of the spinal cord caudal to the transection (cut along the posterior median sulcus of the spinal cord) was transplanted to bridge the transected spinal cord in the presence of a fusogen (polyethylene glycol, PEG) in four of the eight dogs. We used various forms of imaging, electron microscopy, and histologic data to determine that after our transplantation of a vascular pedicled hemisection to bridge the transected spinal cord, electrical continuity across the spinal bridge was restored. Results Motor function was restored following our transplantation, as confirmed by the re‐establishment of anatomic continuity along with interfacial axonal sprouting. Conclusion Taken together, our findings suggest that SCI patients—who have previously been thought to have irreversible damage and/or paralysis—may be treated effectively with similar operative techniques to re‐establish electrical and functional continuity following SCI.


| INTRODUC TI ON
Spinal cord injury (SCI) results in major issues for both individuals suffering from SCI and for society. [1][2][3] Indeed, tens of thousands of people suffer from severe disabilities as a result of SCI, as well as multiple complications related to SCI. Apart from the direct impact of SCI on patients and their families, SCI also places a heavy burden on society, in terms of social and financial costs. 4 Studies have shown that extensive damage occurs within 30 min of trauma, during which axons die back hundreds of micrometers.
In addition, some in-vivo imaging studies have demonstrated that axonal regeneration can occur within 6-24 h of the lesion. 5,6 Recent research has shown that there are two parallel cellular pathways from the brain to the spinal cord. 7,8 In addition to the pyramidal tract, there is a gray matter-based network of interneurons extending from the brainstem to the spinal cord that is involved in the conduction of command signals from cortical motor areas to peripheral motor neurons. This short-fiber pathway is known as the corticotruncoreticulo-propriospinal pathway, which embeds and links the central pattern generators located in both the cervical and lumbar spinal cord. 7,8 After spinal cord injury, this short fiber regenerates much faster than the long fibers in the pyramidal tract. Thus, if two fresh and healthy transected spinal cord sections are created after spinal cord injury, and they can be fused quickly to rebuild the electrophysiological connection, further damage can be minimized, thereby increasing the patient's chances of healing.
Polyethylene glycol (PEG) is a water-soluble polymer that is synthesized from ethylene oxide and has a molecular weight ranging from 0.4 to 100 kDa. Non-toxic and non-irritating, it is widely used in various pharmaceutical preparations. PEG has been approved by the U.S. Food and Drug Administration (FDA) and the China National Food and Drug Administration (CFDA) for clinical use. Extensive experimental evidence has shown that polyethylene glycol (PEG) can act as a neuroprotective agent to not only prevent apoptosis but also to seal damaged membranes and thereby promote membrane fusion, allowing for acute restoration of the integrity of sharply severed nerve fibers. [9][10][11][12][13] Hence, when PEG is utilized after resection of a limited segment of spinal cord and re-establishes apposition of the severed ends of the spinal cord, through the membrane fusion function of PEG, the spontaneous regrowth of severed axons/dendrites of the apposed spinal cord occurs at the contact point, reestablishing the gray matter neurophil. 11 As such, the bridged gray matter is then connected to the white matter on both sides, thereby restoring electrophysiological conduction.
We and others have shown this dramatic re-establishment of electrical continuity in several animal models of complete spinal cord transection (mice, rats, dogs, pigs, and monkeys) using PEG as a neuroprotective agent at the site of transection. In these studies, all animals eventually recovered a large part of their motor functions. 10,[14][15][16][17][18][19] However, this experimental model does not recapitulate traumatic SCI in humans; in contrast, patients with paraplegia secondary to traumatic SCI have glial scars. Therefore, a surgical method that meets the following conditions needed to be designed: (1) The glial scar needs to be resected acutely, leaving the cranial and caudal ends of the otherwise normal spinal cord; (2) These two freshly transected sections need to be connected immediately without any tension or a gap. Given the above two conditions, there are three potential surgical methods, as follows: (1) transplantation of a stem cell-based tissue engineering scaffold; (2) operative shortening of the vertebral column (by vertebrectomy or multiple discectomies), and approximation of the two freshly severed spinal cord stumps; and (3) autotransplantation of a vascularized segment of the caudal spinal cord to act as a neural bridge that will allow re-establishment of electrical and functional continuity across the gap left after resection of the glial scar. In the present study, we employed autotransplantation since it circumvents immune-related issues and allows for a vascularized spinal segment that can be mobilized to fit the SCI gap.

| Spinal cord excision and autotransplantation
During the entire operation, each dog was kept in a prone position.
Ketamine (0.1 mg/kg given intramuscularly) was administered to induce general anesthesia before intubation for artificial ventilation.
Anesthesia was maintained through intravenous administration of remifentanil (0.2 μg/kg/min), vecuronium bromide (0.1 mg/kg), and propofol (10 mg/kg/h). An incision overlying the thoracic spinal column was made at T10-T11, and a laminectomy was conducted using standard neurosurgical techniques to expose the spinal cord. Then, 1 cm of the spinal cord at T10 was removed via sharp transection using a sapphire blade (Shanghai Jingming Fine Technology Co.) both cranially and caudally to create a gap, as described previously 7 (Figure 1). Thereafter, the distal 1.5 cm of the caudal spinal cord was hemisectioned transversely, and the blood supply was preserved to the now mobilized hemisection to be used to bridge the gap from the excised spinal cord. Thereafter, random numbers were used to randomize dogs to either the experimental condition, where the autotransplantation site was bathed in PEG 600 (2 ml) for 5 min, or the control condition, where the autotransplantation site was bathed in 0.9% NaCl (2 ml) for 5 min (0.9% NaCl is an isotonic solution that will not damage cells).
Standard closure by layers was subsequently performed.
For three consecutive days following the above procedures, all dogs were given an intravenous solution containing sulbactam sodium and cefoperazone (25 mg/kg; Harbin General Pharmaceutical Factory's Sales Company) Urine was removed from the bladder via abdominopelvic compressions twice a day throughout the period of observation of the study, or until the voiding reflex was restored. In the absence of recovery, most dogs resumed oral intake of a normal diet (Bridge PetCare Co., Ltd) by postoperative day 3. Dogs had free access to water or were hydrated intravenously with electrolyte solutions when they were unable to eat or drink during the first postoper- wheelchair-like support that allowed mobilization until useful voluntary motor function was regained; this support also helped to alleviate or prevent pressure sores and skin irritations. All eight dogs survived the operation and completed the study without obvious discomfort.

| Motor assessments
Motor function was assessed 12 times after the operation at 3, 10, F I G U R E 1 Surgery. STEP ONE: a dorsal view of the complete spinal cord is shown, with the left side being the cranial side and the right side being the caudal side. STEP TWO: complete removal of a 1-cm-long segment of the spinal cord via sharp dissection at the T10 segment of the spinal cord is shown. STEP THREE: a hemisegment of the distal spinal cord was cut along a cross-section at a distance of 1.5 cm from the gap, after which we continued to cut along the posterior median sulcus of the spinal cord to the gap (the side close to the gap was not completely cut, so the blood supply to this half of the spinal cord was maintained). STEP FOUR: the hemisection was then mobilized to fill the gap left after removing the T10 section of the spinal cord. STEP FIVE: we then injected PEG at the transplant site. The half of the spinal cord that was recently flipped was tightly connected to the cranial face of the spinal cord and the remaining half of the caudal spinal cord of 0 signifies total paraplegia, while a cBBB score of 19 signifies normal functioning. 22 Statistical significance at each time point was calculated using two-way analysis of variance and Mann-Whitney U-test.

| Tissue preparation
At 6 months after operations, all dogs were fully anesthetized with ketamine (0.1 mg/kg given intramuscularly) and perfused with 0.9% NaCl solution and 4% paraformaldehyde. Immediately thereafter, the thoracic spinal cord was extracted en bloc by sectioning 2 cm above and below the point of fusion ( Figure 2).
Subsequently, a small section (approximately 2 mm × 3 mm) was cut from the transplant area for electron microscopic observations. The remaining samples were used for histopathology and were immersed in 4% paraformaldehyde for 48 h, followed by paraffin embedding.

| Transmission electron microscopy
At the deep site of the transplant area, we cut small blocks of approximately 1 mm 3 and soaked them in 2.5% glutaraldehyde at 4°C for 24 h. The tissue blocks were washed three times in phosphatebuffered saline (PBS) for 15 min. After fixation in 1% osmic acid for 1 h, 1% uranyl acetate was used to stain the tissues for 2 h. The tissues were then embedded for coronal sections after using gradient acetone solution for dehydration. Following localization by toluidine blue staining and semi-thin sectioning, the tissues were observed using a transmission electron microscope (H-7700, Hitachi) after being cut into ultrathin sections.

| Scanning electron microscopy
At the deep site of the transplant area, we cut small blocks of approximately 1 mm 3 . These blocks were fixed in a solution containing 2% glutaraldehyde in 0.1 M of sodium cacodylate buffer at 4°C for 12 h. Thereafter, the tissue blocks were washed three times with 0.2 M of sodium cacodylate buffer and were then post-fixed in a 1% osmium tetroxide solution for 1 h. Subsequently, they were dehydrated and mounted on aluminum stubs. A MED020 sputter coater (Balzers) was used to sputter coat the samples with copper, and they were then examined using a Hitachi 10-kV scanning electron microscope (S-3400N, Hitachi).

| Standard histology
Sagittal sections (involving the entire cross-section of the transplanted hemisegment of the spinal cord) at a thickness of 7 μm each F I G U R E 2 Spinal cord sample. A1 (dorsal) and A2 (ventral) show images from the TRANSPLANT group. It can be seen that there is a glial scar in A1, as indicated by the red arrow. B1 (dorsal) and B2 (ventral) show images of the TRANSPLANT+PEG group. The shape of the transplanted spinal cord was basically the same as that of the normal spinal cord. However, the other half of the gap consisted of a glial scar (red arrow in B1) were prepared from the paraffin-embedded spinal cords and were stained with either hematoxylin-eosin (HE), luxol fast blue (LFB), or 1% cresyl violet (Nissl staining). For HE staining, the 7μm-thick sagittal sections were placed directly into xylene twice for 10 min for dewaxing, before being placed into alcohol at decreasing concentrations (100%, 95%, 85%, and 70%) for 5 min each for hydration. PBS was used to rinse the sections three times (5 min each) before they were stained with hematoxylin for 10 min and subsequently rinsed with distilled water. Finally, a 1% HCl solution was used to separate the cytoplasm from pigment; then, the sections were dipped in distilled water. Sections were stained with eosin for 3 min before dehydrating them with alcohol at increasing concentrations (70%, 85%, 95%, and 100%) for 5 min each. Thereafter, the sections were washed twice with xylene (5 min each) for vitrification.
For Nissl and LFB staining, the 7μm-thick sections were dewaxed and hydrated, after which they were submerged in xylene for 20 min, 100% alcohol for 1 min, and 95% alcohol for 5 min. Nissl and LFB were then used to stain these sections for 10 min, followed by a triple rinse in 0.2% glacial acetic acid and counterstaining for 10 min in 0.5% brilliant-green glacial acetic acid solution. Next, distilled water was used to rinse the sections for 2 min, prior to them being dehydrated in gradient ethanol and incubated twice in xylene (5 min each).
An Olympus IX73 optical microscope was used to visualize all sections. For image analysis, Image-Pro-Plus (Media Cybernetics) software was utilized.

| Immunohistochemistry
The spinal cord tissues of the TRANSPLANT+PEG group and the

| Statistical analysis
Motor assessments (cBBB scores) were analyzed by two-way analysis of variance (α = 0.05) and Mann-Whitney U-test.
Histologic indices (the AIOD of neurofilament-positive and MBP-positive areas) were analyzed via Image-Pro-Plus 6.0 (Media Cybernetics).

| Motor function is partially restored proved by motor assessment
Normal cBBB scores were observed initially for all eight dogs (Table 1) Table 1. The Mann-Whitney U-test indicated a statistically significant difference between the two groups after the 29th day (p < 0.05), and the mean value of the TRANSPLANT+PEG group was higher than that of the TRANSPLANT group.  TA B L E 2 (a) Two-way analysis of variance shows that there was a statistically significant difference between the TRANSPLANT+PEG group and the TRANSPLANT group (F = 329.8, p < 0.0001), and there were also statistically significant differences across different time points (F = 109.5, p < 0.0001). (b) Through the mixed-effects analysis of the TRANSECTION+PEG group and the TRANSPLANT+PEG group, the p value and F value of the factor 'time' are <0.0001 and 114.6, respectively; thus, time has an impact on the recovery of spinal cord injury. The p value and F value of the factor 'treatment (TRANSPLANT+PEG group or TRANSECTION+PEG group)' were 0.6312 and 0.2469, respectively, so there was no significant difference between the TRANSPLANT+PEG group and the TRANSECTION+PEG group

| A large amount of new myelin and axons found by transmission electron microscopy
Prior research has shown that axons undergo demyelination and ne-

| The myelin of the TRANSPLANT group and the TRANSPLANT+PEG group were significantly different observed by scanning electron microscopy
Changes in myelin and axons can be observed in more detail via scanning electron microscopy. In the scars of the TRANSPLANT group at the site of transection, most of the observed myelin was damaged ( Figure 6, arrow + star in B2 and B3) and incomplete ( Figure 6, arrow + square in B1 and C2). In comparison, the myelin of the TRANSPLANT+PEG group had a smooth surface and a complete structure ( Figure 6, arrow + heart in A1, A2 and A3). In addition, the myelin from the TRANSPLANT+PEG group was generally thinner, suggesting that there was a smaller volume of new myelin. At the edge of the sample block (the position of the knife cut when sampling) in the TRANSPLANT group, degenerated myelin ( Figure 6, arrow + square in C1) was seen to envelope a degenerated axon ( Figure 6, arrow + triangle in C1), similar to that seen in Figure 5 (arrow + heart in A2). In the glial scars of the TRANSPLANT group, axons with spherical enlarged stumps after degeneration were also found ( Figure 6, arrow + triangle in C3), which represent another sign of axonal degeneration.

| Evidence of nerve recovery showed by histologic assessment
In HE-stained sections, the PEG-treated and untreated spinal cords differed tremendously. In contrast to that of the TRANSPLANT group, vacuolization due to tissue injury was minimal in the TRANSPLANT+PEG group (Figure 7, red arrows in A and B), which likely demonstrates the neuroprotective effects of PEG. In terms of Nissl staining, neurons labeled with Nissl bodies could be seen only in the bridging of the TRANSPLANT+PEG group (Figure 8, red arrow in A1) but not the TRANSPLANT group (Figure 8, A2). After LFB staining, irregularly arranged myelin was found in the bridging tissue of the TRANSPLANT+PEG group (Figure 8, red arrow in B1), but only stained myelin debris was seen in the TRANSPLANT group (Figure 8, B2).

| A large amount of MBP and neurofilament protein exist in the bridging tissue found by Immunohistochemistry
CNS myelin contains a large proportion of MBP; thus, MBP is a well-known marker for CNS myelin. 27 A large quantity of MBP was observed in the bridging tissue of the TRANSPLANT+PEG group ( Figure 9, B1, C1 and red arrows in B2, C2), but was seen only rarely in the TRANSPLANT group (Figure 9, A1 and A2). Surprisingly, a highly concentrated area of MBP was observed in the transverse section of the TRANSPLANT+PEG group (from the bridging tissue), which may have represented the transplanted spinal cord (Figure 9, black square in B1). However, in the TRANSPLANT group, no highly concentrated area of MBP was found. Semi-quantitative analysis of F I G U R E 6 Scanning electron microscopy. In the bridging tissue of the TRANSPLANT group, most of the myelin is damaged (arrow + star in B2 and B3) and incomplete (arrow + square in B1 and C2). By comparison, the myelin of the TRANSPLANT+PEG group has a smooth surface and a complete structure (arrow + heart in A1, A2 and A3). At the edge of the sample blocks in the TRANSPLANT group, we find degenerated myelin (arrow + square in C1) enveloping a degenerated axon (arrow + triangle in C1, similar to the results in Figure 5, namely arrow + heart in A2). In the scars of the TRANSPLANT group, we also find axons with a spherically enlarged stump after degeneration (arrow + triangle in C3) AIODs revealed that there was a significant difference between the

| Axons and myelin pass through the bridging tissue detected by Immunofluorescence
When the immunofluorescent slides were assessed, we found that myelin was present in the bridging tissue of the TRANSPLANT+PEG group ( Figure 11, B1, B2, B3, C1, and C2), but not in the TRANSPLANT group (Figure 11, A1). On one representative section, one can see a complete axon appearing to pass through the bridging tissue (Figure 11, B2). In the transverse section (from the bridging tissue), dense myelin was present ( Figure 11 The immunofluorescence identified that GFAP protein was present in the bridging tissue of the TRANSPLANT+PEG group and the TRANSPLANT group. This indicates that PEG may not have a significant inhibitory effect on the growth of glial scars, but it may have an inhibitory effect on fibrotic scars, which ultimately inhibits the growth of astroglial-fibrotic scars.

| DISCUSS ION
There are at least four barriers to the effective treatment of SCI that must be overcome, as follows. 29 First, amelioration/prevention of primary and secondary tissue damage at the SCI site is needed.
CNS injury triggers microglial activation within minutes, constitute the innate immunity responsible for debris clearing and providing a source of trophic and antiinflammatory factors to promote tissue repair, but they also release inflammatory cytokines to fuel secondary injury, 30  re-approximation in our present study. The GEMINI SCF protocol is predicated on propriospinal neurons of the gray matter sprouting fibers after the apposition of spinal cord surfaces under PEG protection, rather than on the regrowth of long pyramidal axons. 7,55 After the short fibers in the gray matter re-connect, they bridge long-range fibers (i.e., the pyramidal tract) on both sides, thereby restoring electrical continuity. This finding was confirmed in the present study. Notably, sphincter control did not recover, which is controlled by long-range fibers, which apparently did not regrow or fuse at the site of the autograft.
Before the translation of this present research into clinical practice, several points should be noted. First, the hemicord to potentially be used as the bridge in patients with a chronic SCI will have undergone years of some form of plastic adaptive changes, some of which may be detrimental. 56 These changes might interfere with the process of reinnervation in some situations or different types of injury; however, it is well known that so-called entrenched plasticity is reversible. 57  However, the chest and abdominal wall muscles in this limited area would still be affected by motor deficits. Third, motor recovery and evidence of fiber regrowth indirectly confirmed that the graft retained its vascular inflow in the present study. However, spinal angiography should be a consideration in clinical trials. Finally, the F I G U R E 1 2 Immunofluorescence for neurofilament protein and GFAP: in the bridging tissue of the TRANSPLANT+PEG group, we observe a large amount of NF immunofluorescence indicative of neuronal fibers (B1 and C1), but we did not find similar immunofluorescence in the bridging tissue of the TRANSPLANT group (A1). In the magnified view, many apparent axons penetrate through the bridging tissue (B2, B3 and C2). In the coronal view, there is dense myelin (white square in C1) use of PEG alone is associated with an impressive degree of recovery following spinal transection, but it is likely boosted by stimulation from the spinal cord and motor cortex, as per the GEMINI SCF protocol. 7,55

| CON CLUS ION
In summary, our present study provides the first evidence of reversal of motor paralysis after removal of a complete segment of the spinal cord and immediate bridging with an autologous vascular pedicle of the ventral hemisected spinal cord and PEG treatment (per the GEMINI SCF protocol). If these findings are confirmed, this phenomenon may lead to a novel and viable strategy for the treatment of selected patients with traumatic spinal paralysis.

ACK N OWLED G M ENTS
This study received financial support from the National Natural

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.