Volume 17, Issue 6, Pages 992-1002 (June 2009) Combination of Microsurgery and Gene Therapy for Spinal Dorsal Root Injury Repair  Song Liu, Delphine Bohl, Stephane Blanchard, Josette Bacci, Gérard Saïd, Jean-Michel Heard  Molecular Therapy  Volume 17, Issue 6, Pages 992-1002 (June 2009) DOI: 10.1038/mt.2009.23 Copyright © 2009 The American Society of Gene Therapy Terms and Conditions

Figure 1 Surgical repair of C7 DRG neuron central axon connection to the spinal cord. (a) Schematic representation of the surgical procedure. Right C6 and C7 DRs were sectioned (rhizotomy) and the C6 DRG was eliminated. A predegenerated segment of the cutaneous peroneal nerve (PNG) was anastomosed end-to-end to the sectioned C7 DR and end-to-side to the C5 DR. (b) Photographs of GFP fluorescent signal in PNG cryosections examined at the time of implantation (upper row, before implantation) or 4 months after implantation in recipient rats (bottom row, 4 months after implantation). PNGs were exposed either to the HIV-GFP lentivirus vector alone (NF−) or to the combination of HIV-GFP, HIV-NT-3, and HIV-GDNF lentivirus vectors (NF+). The four pictures are at the same magnification (scale bar: 100 µm), showing that NF− PNG diameter did not increase during implantation in the host, whereas NF+ PNGs became larger. PNGs removed from recipient rats are surrounded by adventitial tissue in which some GFP-positive cells have migrated. (c) The amounts of NT-3 and GDNF mRNAs were measured by q-RT-PCR in tissue extracts prepared from NF− or NF+ PNGs that had been implanted in recipient rats for 4 months. Values are means ± SEM of three determinations. *P < 0.05 (Mann and Whitney test). DR, dorsal root; DRG, dorsal root ganglion; PNG, peripheral nerve graft; GDNF, glial cell line–derived neurotrophic factor; mRNA, messenger RNA; q-RT-PCR, quantitative reverse transcriptase polymerase chain reaction; NT-3, neurotrophin-3; VR, ventral root. Molecular Therapy 2009 17, 992-1002DOI: (10.1038/mt.2009.23) Copyright © 2009 The American Society of Gene Therapy Terms and Conditions

Figure 2 Recovery of nociception and proprioception. (a) Schematic representation of a rat forefoot showing footpads 2 (Fp-2) and 3 (Fp-3) in which sensory deficit was prominent after C6-C7 rhizotomy. Toes are numbered. (b) Thresholds of pain perception were measured at footpads 2 and 3 using a pinch coupled to a pressure sensor. Values were recorded before injury (pre), 1 week (1 w), or 4 months (4 m) after rhizotomy without repair (control, n = 6 rats), or followed by surgical repair with NF− (n = 8 rats) or NF+ (n = 8 rats) PNG. (c) Time before withdrawal from the hot plate was measured for the left forelimb (noninjured side) or the right forelimb (injured side in control or experimental rats) in the absence of rhizotomy (normal, n = 6 rats), or 4 months after rhizotomy without (control, n = 6 rats) or followed by surgical repair with NF− (n = 6 rats) or NF+ (n = 6 rats) PNGs. (d) Examples of footprints are shown on the left. Distances between the second and the fourth toe prints were measured for the left forelimb (ITS left, noninjured side) or the right forelimb (ITS right, injured side in control and experimental rats) in the absence of rhizotomy (normal, n = 120 prints from four rats), or 4 months after rhizotomy without repair (control, n = 180 prints from four rats) or followed by surgical repair with NF− (n = 230 prints from four rats) or NF+ (n = 300 prints from four rats) PNGs. Ratios of the values measured in the right and left sides are shown. (e) The pressure exerted by each forelimb was measured independently in the absence of rhizotomy (normal, n = 36 determinations in six rats), or 4 months after rhizotomy without repair (control, n = 35 determinations in six rats) or followed by surgical repair with NF− (n = 24 determinations in four rats) or NF+ (n = 24 determinations in four rats) PNGs. To account for weight differences between animals, compared values are ratios of the difference to the sum of pressures measured for the right and left forelimb, respectively. Values are means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 (upper row: Mann and Whitney test, bottom row: Student's t-test). Significant differences with normal rats values are not indicated. ITS, intermediate toe spread; PNG, predegenerated nerve graft. Molecular Therapy 2009 17, 992-1002DOI: (10.1038/mt.2009.23) Copyright © 2009 The American Society of Gene Therapy Terms and Conditions

Figure 3 Electrical connection of C7 DRG neuron peripheral axons with the spinal cord. (a) Electrical stimulation was delivered to the foot skin at ventral side and SSEPs were recorded at the C4 level of the spinal cord, as shown in the left part of the panel. Amplitude and surface of SSEPs recorded 1 week (1 w) or 4 months (4 m) after rhizotomy in sham controls rats (n = 6) or in experimental rats of the NF− (n = 7) or NF+ (n = 7) groups are shown. *P < 0.05, **P < 0.01 (Mann and Whitney test). (b) The C5 DR was sectioned distal to the anastomosis with the PNG to prevent direct conduction, and submaximal intensity electrical stimulation was delivered to the C7 spinal nerve. Late motor action potential interpreted as equivalent to H-reflexes was recorded in the biceps, as shown in the left part of the panel. Recording from two rats of the NF− group and two rats of the NF+ group are shown. Disappearance of this evoked potential after section of the C5 ventral root indicated electrical conduction through the C7 DR, the PNG, the C5 DR to the C5 motoneuron pool. DR, dorsal root; DRG, dorsal root ganglion; SSEP, somatosensory evoked potential; PNG, predegenerated nerve graft. Molecular Therapy 2009 17, 992-1002DOI: (10.1038/mt.2009.23) Copyright © 2009 The American Society of Gene Therapy Terms and Conditions

Figure 4 CTB transport from C7 DRG neuron peripheral axons to the spinal cord. (a) Four months after surgery, CTB-Alexa594 was injected in the radial, medial, and ulnar nerves in control rats (n = 3) or in experimental rats of the NF− (n = 3) or NF+ (n = 3) groups. Before dye injection, the C5 DR was sectioned distal to PNG anastomosis, the C4 DR was sectioned and the C4, C5, C6, and C7 ventral roots were sectioned, as shown in the upper part of the panel. CTB-Alexa594 was detected in the spinal cord after 7 days. The lower part of the panel shows various magnifications of CTB-Alexa594 signals detected in the C5 spinal cord in a NF+ rat. Laminae are indicated. With respect to anatomical constraints in these animals, the detection of CTB-Alexa594 signal in the spinal cord indicates transport of the dye through the C7 DRG neuron peripheral axons, the PNG and the C5 DR. (b) The intensity of CTB-Alexa594 signals was quantified as pixels per section surface in laminae IV, V, and VI, according to the area shown in panel a (total scored surface: 0.24 mm2). Three sections were scored separated by 1 mm in a region centered by the C5 DR. The mean number of pixels per section surface is shown. ***P = 0.001 (Mann and Whitney test). (c) A section stained with anti-NF200 antibodies was analyzed using confocal microscopy. The picture shows colocalization of the CTB and NF200 signals on a 0.3-µm optical section. Scale bars: 10 µm a and 5 µm c. DR, dorsal root; DRG, dorsal root ganglion; PNG, predegenerated nerve graft. Molecular Therapy 2009 17, 992-1002DOI: (10.1038/mt.2009.23) Copyright © 2009 The American Society of Gene Therapy Terms and Conditions

Figure 5 FG retrograde transport from the spinal cord to C7 DRG. (a) Four months after surgery, FG was injected in the spinal cord at the C4 level and CTB-Alexa594 was injected in the footpad skin, as shown in the upper part of the figure, in control rats (n = 6) or in experimental rats of the NF− (n = 8) or NF+ (n = 8) groups. (b) Fluorescent signals were detected in the C7 DRG after 7 days. The detection of FG-labeled C7 DRG neurons (upper row) indicated retrograde transport of FG through the C5 DR and the PNG. The detection of CTB indicates retrograde transport along the C7 DRG neuron peripheral axons. Costaining with FG and CTB-Alexa594 (medium and bottom rows) indicates anatomical connection between the footpad skin and the spinal cord. DRG sections stained with anti-CGRP or anti-NF200 antibodies show triply stained cells (FG + CTB + CGRP antibodies, medium row; FG + CTB + NF200 antibodies, bottom row). (c) The number of FG stained, or FG and CTB doubly stained cells were scored in C7 DRGs in control, NF− and NF+ rats. Mean total numbers ± SEM of cells scored in all sections covering the entire DRG are shown. ***P < 0.001 (Mann and Whitney test). Scale bars: 100 µm (upper row), 50 µm (medial and bottom rows). CGRP, calcitonin gene–related peptide; DR, dorsal root; DRG, dorsal root ganglion; FG, fluorogold; PNG, predegenerated nerve graft. Molecular Therapy 2009 17, 992-1002DOI: (10.1038/mt.2009.23) Copyright © 2009 The American Society of Gene Therapy Terms and Conditions

Figure 6 Axonal regeneration in PNGs. C6-7 rhizotomy was followed by surgical repair with NF− or NF+ PNGs. PNGs were removed 4 months after implantation in the host and processed for semithin or ultrathin sections or cryosections. (a) Semithin sections (upper row) show the aspect of a total degenerated superficial peroneal nerve (right panel) and degenerated/reinnervated PNGs removed from NF− (middle panel) or NF+ rats (left panel) with regenerating axons (arrows) and fibrosis (arrow heads). Ultrathin sections (bottom row) show regenerating myelinated (arrow) and nonmyelinated (arrowheads) fibers. Scale bars: upper row 20 µm, bottom row 2 µm. (b) CTB-Alex594 was injected in the medial, radial, and ulnar nerves and PNGs were examined after 7 days. Adjacent sections of the same NF− or NF+ PNGs are shown. PNG limits indicated by dashed white lines show that the NF+ PNG (middle column) has a much larger diameter than the NF− PNG (left column). Signal quantification (right column) shows that NF+ PNGs contain more numerous regenerating fibers and myelinating cells than NF− PNGs. Fluorescent signals from CTB-Alexa594 retrograde labeling show sensory fibers regenerating from the footpad skin. Immunostaining with anti-NF200 or anti-CGRP shows the abundance of large-size and small size axons regenerating in the PNGs, respectively. Immunostaining with anti-MBP or anti-S100β shows myelinating cells in PNGs. The typical morphology of MBP-stained myelinating cells is shown in the inserts. Scale bar: 50 µm, insert 5 µm. PNG, predegenerated nerve graft. Molecular Therapy 2009 17, 992-1002DOI: (10.1038/mt.2009.23) Copyright © 2009 The American Society of Gene Therapy Terms and Conditions