Sections of C5 and C6 DRG (10 mm) were double-immunostained for bIII tubulin (Promega, 1:1000) to identify all neurons, and for the growth-associated protein GAP-43 (gift from G. Wilkin, 1:2000), using 7-amino-4-methylcoumarin-3-acetic-acid-and tetramethylrhodamine isothiocyanate (TRITC)-conjugated secondary antibodies. The percentage of GAP-43-positive cells was determined in four sections per animal.
Axon tract tracing in the spinal cord
Ascending dorsal column axons were labelled using the CTB tracer 29 , injected into the left median nerve of anaesthetized rats. Longitudinal sections (20 mm) were immunostained for glial fibrillary acidic protein, to identify the lesion site, and for CTB, to identify the labelled dorsal column axons 29 . Descending CST axons were labelled with 1 ml BDA (Molecular Probes; 10% in saline) injected under anaesthesia at four sites over the left primary motor cortex, 1 mm below the dorsal surface of the brain. BDA-labelled fibres were visualized in parasaggittal spinal cord sections (20 mm) with extra-avidin conjugated to fluorescein isothiocyanate. All BDA-labelled fibres observed within a 1-mm square grid were counted at measured intervals from 4 mm above to 5 mm below the lesion site by an experimenter, blinded to treatment. Due to variability in labelling, axon numbers were calculated as a percentage of the fibres seen 4 mm above the lesion, where the CST was intact. To confirm a complete CST lesion, transverse lumbar spinal cord sections (20 mm) were immunostained with an antibody against the g-subunit of protein kinase C (PKC-g, Santa Cruz, 1:1000), visualized with a TRITC-conjugated secondary antibody.
Electrophysiology
In terminal electrophysiological experiments, the sensory motor cortex and cervical spinal cord were exposed in urethane-anaesthetized (1.5 g kg 21 ) rats. Cortical evoked potentials were elicited by electrical stimulation (five square-wave pulses at 400 Hz, 100 mA, 200 ms, delivered every 2 s) of the left sensory motor cortex using a 0.5-mm concentric needle electrode lowered 1 mm into the cortex. For each experiment we mapped the optimal stimulation site, located 1-2 mm lateral and 1 mm rostrocaudal from Bregma. Postsynaptic potentials evoked by the cortical stimuli were recorded with a silver ball electrode placed medially on the contralateral cord surface. At each recording site (1 segment above and 1-7 mm below the lesion at C4), 64 responses were averaged and stored for off-line analysis of response magnitude (area under curve) and latency.
Behavioural assessment
After baseline testing, we tested animals once a week for 6 weeks after lesion. We averaged right and left forepaw scores, as no differences were observed between them. Experimenters were blind to the treatment. A tape removal test (adapted from ref. 24) produced separate scores for sensory and motor behaviour. Adhesive tape (0.3 inch ยฃ 1 inch) was placed on the forepaw, and the time taken to sense the presence of the tape (indicated by paw shake) was determined. For animals that sensed the tape, the removal time was also scored. Rats were tested on two locomotor tasks requiring sensorimotor integration (adapted from ref. 30). Rats were trained to cross a narrow metal beam (1.25 inch ยฃ 36 inches) and a wire grid (12 inches ยฃ 36 inches with 1 inch ยฃ 1 inch grid squares). Forepaw foot slips were recorded (determined by a paw slipping off the beam or below the plane of the grid). In a footprint analysis (adapted from ref. 30), rat forepaws were covered with ink to record walking patterns during continuous locomotion across a wooden runway (4 inches ยฃ 36 inches), and stride length and width were calculated.