Objective To study the effects of embryonic spinal cord on reconstruction of nerve tract of injured spinal cord of adult rats.
Methods At days 30, 50, and 70 after E14 embryonic spinal cord transplantation into acutely injured spinal cord of adult rats, the fiber connection between the graft and the host tissue was traced with CB-HRP under transmission electron microscope.
Results CB-HRP labeling revealed ultrastructurally regeneration of axons of dorsal root into the graft. At day 30 after operation. At this time, motoneurons and interneurons also sent their sproutings to the graft. At day 50 after transplantation, dorsal afferent axons made synaptic connection with the embryonic neurons mainly in the form of axonal-dendritic synapse. In the meantime, rubrospinal fibers regenerated with relatively slow speed towards the graft and formed axonal-dendritic synaptic connection with the fetal neurons at day 70 after transplantation.
Conclusion Injured nerve tracts may be reconstructed in some degree by transplanting embryonic spinal cord into traumatized spinal cord.

Although nerve fibers of transplanted embryonic spinal cord (ESC) and host spinal cord(HSC) may interlace extensively, in the treatment of spinal cord injury (SCI), much more attention is paid to whether synaptic connection between ESC and HSC can be formed. This relates to whether interrupted nerve tract can be reconstructed by the transplanting procedure and whether transplanting procedure can provide key structural base for the recovery of HSC function. In order to answer these question, we use the Choleragen subunit B conjugated horseradish peroxides (CB-HRP) to trace nerve connection between the ESC and HSC under transmission electron microscope (TEM).

MATERIALS AND METHODS

CB-HRP Labeling
At days 30, 50 and 70 after transplantation, 5 recipients at each time point were anesthetized with urethane (1.25g/kg). The left sciatic nerve was exposed and the trunk of the nerve was clamped with a hemostat for 30 seconds so as to induce degeneration, then 10 μl CB-HRP was injected into the nerve with a micro-syringe. When the red nucleus was labeled, the skull of the rat was exposed and fixed in the brain stereotaxic instrument. The projection location of the red nucleus on the top of the skull was determined according to the Paxinos atlas of rat brain in stereotaxic coordinate.¹ With a dental drill, a small hole in diameter of 2 mm was created on the skull at the intersecting point of a line 5 mm posterior to the coronal suture with another line 1.2mm right lateral to the sagittal suture. The dura mater was ruptured with a meningeal forceps under surgical microscope and a glass micro-electrode with 10 μl CB-HRP was propelled for 7mm ventrally with the aid of a microelectrode propeller, and then at the determined level CB-HRP was injected. After the above procedures were finished, the wounds of hind limb and head were closed with interrupted sutures. Among the 15

Research Institute of Surgery, Daping Hospital, Third Military Medical University, Chongqing 400042
This study was supported by National Natural Scinece Foundation of China (No:39470715)
rats used for the CB-HRP labeling, 6 was for sciatic nerve labeling, 6 for red nucleus labeling, and 3 for both sciatic nerve and red nucleus labeling at the same time.

Infusion Fixation and Sampling
Forty-eight hours after CB-HRP labeling, the animals were reanesthetized. Thorax was opened to expose the heart. 150ml of 5% sucrose-PBS (0.1mol, pH 7.4) and 200ml of mixed fixative of 1% paraformaldehyde-2% glutaraldehyde were infused in succession through the left ventricle and aorta. After 2 hours of fixation, spinal cord at the transplantation site was removed and fixed in the same fixatives for 4 hours, and then passed overnight in a solution of 30% sucrose-PBS(0.1mol, pH 7.4) in a 4℃ refrigerator. Then the tissue blocks were taken out and cut into 50μm sections with a cryomicrotome.

HRP Histochemistry
The sections were collected and washed in redistilled water for 5min, then in histochemical solution for reaction of 20 min (contents of the solution: solution A: 5'-tetramethyl-benzidine 5mg, absolute alcohol 2.5ml; Solution B: sodium nitroprusside 100mg, redistilled water 92.5ml, 0.2mol acetic buffer 5ml. Before use, solution A and solution B were mixed to make reaction solution). After preincubation at room temperature for 20min, the sections were taken out of reaction solution, then soucked again at room temperature in the same reaction solution, into which 0.3% H₂ O₂ was added at a ratio of 100∶3. The sections was washed in acetic buffer (0.02 mol, pH 6.0) for 5min or the section might be stored at this step for a short time.

Embedding of the Sections for TEM Observation
The sections treated by the above procedures were transferred to 1% OsO₄ (0.1mol, pH 6.0, prepared with PBS) at 45℃ for 1h, washed in PBS (0.1 mol, pH 6.0) 2 to 3 times, dehydrated in series acetone from 50% to 100%, and transferred to epoxypropane for 15min and to a mixture of epoxypropane and Epon 618 at 37℃ for 1h. The sections were then put in embedding agent both with and without promoting agent at 37℃ for 2h. Subsequently, the sections were moved to the silicified glass slide with a toothpick, and 1 to 2 drops of embedding agent were added to the sections, then the sections were sandwiched between another silicified glass slide and a slight pressure was applied to smooth out the sections. The glass slides with the sections were allowed to polymerized at 45℃ for 12h, and at 60℃ for 48h.

Location of the Section and TEM Observation
Polymerized glass slides were carefully unclenched with a knife blade. The one with section was observed under light microscope with intense illumination. After the transplant was located, the glass slide was transferred to stage of anatomic microscope, then the transplant with the connected host tissue were cut out with razor-blade (Fig.1). Subsequently, the cut section, about 1mm² in size was sticked with 502 sticky agent on to the top of used and planished embedding capsule. Afterwards, 600nm ultrathin sections were cut with an ultramicrotome and counterstained with uranyl acetate and lead citrate and observed under JEM 100SX TEM.

RESULTS

Ultrastructural Characteristics of the Spinal Cord and Reaction Products
Because the cryosections were used for embedding, ultrastructure of the tissues was not as ideal as that of routinely processed samples under TEM. The cell matrix was loose, and some matrix was lost and mitochondria revealed swelling. However, various structures of the nerve tissue could still be clearly distinquished under TEM. Histochemical reaction products were shown as needle-like crystals with high electron density and often arranged in paralled and formed bundle. The other cell structures could be distinctively distinguished from the reaction products by these features. Density of the needle-like products was different, so that dense ones could completely mask other cell structure, and loose ones scattered among the cell matrix. The products were distributed in the perikaryon, axons, dendrites, myelineted fibers and nerve terminals. Besides, dense products could also be found in the cytoplasm of the oligocyte.

Difference in the Labeling Results of CB-HRP Between Sciatic and Red nucleus Injection
After CB-HRP was injected into sciatic nerve, the reaction products could be incidentally found in the myelinated fibers at day 30 after operation (Fig.2). At day 50 after operation, besides nerve fibers, the products could also be found in the nerve terminals (Fig.3), dendrites (Fig.4), and perikaryon. There was more chance to find the products in postsynaptic elements (Fig.5) at day 70 after operation than at day 50 after operation. Following CB-HRP injection into the red nucleus, it was not until day 50 after operation that the products could incidentally be found. At day 70 after operation, although the chance to observe the products in the same group of animas increased as compared with those of earlier time intervals, it was still not so easy to observe the products as that of injecting CB-HRP into the sciatic nerve. Besides the temporal difference, the products were often distributed in the myelinated and unmyelinated fibers as well as in the presynaptic elements (Fig.6) after CB-HRP was injected into the red nucleus. No reaction products could be found in the neuron body and dendrites.

FIg.1 50μm Cryosection used in embedding for TEM observation. H indicates host spinal cord, and G graft. Arrow(→) points to labeled nerve fibers. Thirty days after injection of HRP into the sciatic nerve. ×100. Fig.2 Ultrastructure of spinal cord at day 30 after transplantation. HRP labeling products (→) can be found in the myelinated fiber (MY) of the graft. HRP was injected into the sciatic nerve.×20000. Fig.3 Ultrastructure of spinal cord at day 50 after transplantation. HRP labeling products (→) appear in the nerve terminal (T) of the graft. HRP was injected into the sciatic nerve. ×21000.

Fig.4 Ultrastructure of spinal cord at day 50 after transplantation. HRP labeling products(→) distribute in dendrite (D) of the graft. Synapse (→), which is not very typical, is formed by junction of this dendrite with nerve terminal. HRP was injected into the sciatic nerve.×20000. Fig.5 Ultrastructure of spinal cord at day 70 after transplantation. HRP labeling products (→) can be found in dendrite (D) as a postsynaptic element of the graft. T indicates nerve terminal, and arrow (→) points to axonal-dendritic synapse. HRP was injected into the sciatic nerve.×28500.Fig.6 Ultrastructure of spinal cord at day 70 after transplantation. HRP labeling products (→) distribute in the nerve terminal (T) which forms a synapse with a dendrite (D). Arrow (→) points to axonal-dendritic synapse. HRP was injected into the red nucleus. ×23000.

DISCUSSION

Among the various HRP reagents, TMB was generally regarded as the most sensitive one.² In the present experiment the determined structures labeled by CB-HRP in the graft with this reagent could be identified under TEM. The results showed that at day 30 after transplantation the reaction products could be found in the myelinated and newly-grown axons if the CB-HRP was injected into the sciatic nerve. We deduce that these first labeled nerve fibers in the grafts may most probably originate from the central processes of neuron in the dorsal root ganglion of the same side. Under normal condition, although the motor neuron of the spinal cord may be labeled in a retrograde manner by HRP injected through the sciatic nerve, they connect with the surrounding neuron by the dendrites and neuronal body, and the axons, constituting nerve fibers, exit from the spinal cord through the anterior root to form somato-motor elements of the spinal nerve. The nerve fiber of the contralateral spinal cord may surely enter the grafts through the central gray matter, but it is almost impossible to label the neuron in contralateral side of the spinal cord by HRP which has been injected into the ipsilateral sciatic nerve. Local interneuron of the same side may be labeled by HRP transporting through the membrane. Going directly into formation of sciatic nerve, the pseudo-single polar neuron may most probably take up HRP injected into the sciatic nerve, then HRP was transported through central processes into the grafts. This suggestion is similar to that of the results reported by Tessler ^{3 ～5} and Pallini⁶ , where they injected HRP into the dorsal root directly. As sampling time was different, these authors observed the nerve terminals and synaptic structures originating from the dorsal roots in the grafts at about 2 months after operation.
It was further showed that at day 50 after the transplantation, HRP labeled structures expanded from nerve fibers to perikaryon, axonal terminals, and dendrites. According to the above deduction, we consider that labeled nerve terminals develop from the differentiation of regenerating dorsal root axon. Dendrites might be labeled through the following ways: one is that branches of dendrites of anterior horn motor neuron of near-by segment or interneuron grow into the graft; another is that HRP, transported by dorsal root axon, label transmembranely dendrites of the neuron in the graft. The products distributed in the perikaryon may most probably come from transmembrane transportation between the dorsal root axon and perikaryon of neuron located in grafts. These deduction can still fit to the observation under TEM at day 70 after operation in the present study. At this time, besides an increase in HRP-labeled perikaryon, axonal terminals and dendrites, more tracer labeled axonal-dendritic synapse can be easily observed. These results suggest that the embryonic spinal cord matures in structure after transplantation and makes more and more close connection with the regenerating dorsal root axon and dendrites of anterior horn motor neuron. Synthesizing the results mentioned above, we suggest that dorsal root axons of host spinal cord can regenerate into the transplanted embryonic spinal cord and can mainly form axonal-dendritic synapse with the neuron of the graft. At the same time, in order to make synaptic connection with the embryonic neuron, dendrites of anterior horn motor neuron located in near-by segment and/or interneuron also regenerate into the graft by branching.

Besides the sciatic nerve, HRP was also used to trace projection of the fibers of rubrospinal tract into the embryonic spinal cord. Although human rubrospinal tract is small and most of them terminate in the upper part of the spinal cord, well developed rat rubrospinal tract can be as long as the whole spinal cord. In lumbar region, the fibers of rubrospinal tract go into the lamina Ⅴ~Ⅶ of gray matter through the dorsal part of the funiculus laterallis and form axonal-dendritic synapse with dendrites of interneuron.⁷ After the embryonic spinal cord was transplanted into the injured spinal cord of rats, the regeneration of rubrospinal tract towards the grafts was observed under TEM. The results indicated that it was not until day 50 after operation that a few myelinated fibers, growing new axons and axonal terminals could be labeled by HRP. At day 70 after the transplantation, the above mentioned HRP-labeled structures, especially axonal-dendritic synapse, increased in number compared with those of earlier time intervals. Yet, no reaction products could be found in the perikaryon and dendrites throughout the experiment. It can be suggested from the results that descending fibers of rubrospinal tract can regenerate and enter into the graft, and form synapse mainly with dendrites of the neuron in the graft; but the time for nerve fibers to regenerate and enter into the graft is evidently later than that of dorsal root axon.

Synthesizing the above results of HRP tracing under TEM, our general impression about the embryonic spinal cord reconstructing the injured spinal cord is that ascending fibers from the dorsal root and descending fibers originating from rubrospinal tract can regenerate and enter into the transplanted embryonic spinal cord at different time intervals after the operation and form limited number of synaptic connection with its neuron. These, together with processes sent by near-by neuron, can repair the interrupted nerve tract to some degree.

REFERENCE