The exposed T8-9 spinal cord of 9-week-old female Sprague Dawley rats was completely transected for 2 mm with an ophthalmic blade. A 2-mm-long scaffold of HC400 or RPC was then implanted between the rostral and caudal spinal cord stumps. Four weeks after implantation, animals were perfused transcardially with 4 % paraformaldehyde in 0.1 M PB, and a 14-mm-long spinal cord section containing the implant was dissected under a microscope. Sagittal sections (30 μm) through the scaffolds were cut on a cryostat and immunostained with NF 200 antibody. In RPCs, NF-positive fibers had penetrated the implanted site, but few fibers had crossed the rostral border (Fig. 22.2a). In contrast, NF-positive fibers were observed beyond the borders between the HCs and the host spinal cords 4 weeks after implantation (Fig. 22.2b). At higher magnification, many NF-positive fibers were observed parallel to the longitudinal midline of the implants (Fig. 22.2d), while several fibers were seen crossing the border between the implanted area and the stumps (Fig. 22.2c, e). These results suggest that the serial tunnel structure of the HCs have a guidance ability which promotes the regeneration of axons across the implanted area.

Several rats were examined by transcranial electrostimulation 10 weeks after HC implantation. Muscle-evoked potentials were observed in the right quadrant muscle with HC implantation. In addition, the spinal cord was re-transected at the distal stump. Both MEPs and contraction of the quadriceps disappeared after the transection. These results indicate that some descending pathways across the HC-implanted area had reconnected.

BMSCs were prepared from adult Fischer 344 rats as previously described [15]. After one passage, BMSCs were transduced with lentiviral vectors expressing green fluorescent protein. HCs coated with 1 % fibronectin were prepared and mixed with dissociated BMSCs at a concentration of 2 × 10⁵ cells per scaffold. HCs with or without BMSCs were stabilized in the agar filled with DRG medium. Tuj-1-positive neurites and GFP-positive BMSCs were observed in the tubular HC sections after 10 days in vitro (Fig. 22.3a). Quantitatively, neurite length was increased over twofold in the HC combined with BMSCs compared with the control HC, indicating that BMSCs directly support the elongation of neurites into HC in vitro (unpublished data).

For in vivo study, the exposed spinal cord of adult female Fischer 344 rats was hemisected at T9 and a 3-mm hemicord section on the right side was resected. HC or HC combined with BMSCs was implanted between the rostral and caudal spinal cord stumps. Four weeks after implantation, animals were perfused, and a 9-mm-long spinal cord section containing the implant was dissected under a microscope. Horizontal sections of the scaffolds were cut on a cryostat and immunostained with mouse monoclonal anti-SMI31 (a marker for neurofilament M+H) and either rabbit polyclonal anti-GFP antibody or calcitonin gene-related peptide (CGRP) antibody. These antibodies were visualized with appropriate fluorescent secondary antibodies.

Four weeks after implantation, GFP-positive cells had survived in the implanted area of the HC combined with BMSCs, and SMI31-positive fibers were observed at the site (Fig. 22.3b). Quantitative analysis showed that more SMI31-positive fibers were observed in the HC combined with BMSCs than in the HC only (unpublished data). To analyze the growth of sensory fibers, we observed CGRP-positive fibers in the implanted spinal cord. These fibers were observed close to the GFP-positive area (Fig. 22.3c) as regenerated fibers possibly guided by the transplanted BMSCs.