Figure 2. Axonal retraction and regeneration after SCI. a, After spinal cord lesion, injured axons initially retract as evident at 2 d post-SCI. b, Quantitation of the number of neuronal process at different distances from the lesion site reveals a significant increases in regeneration and sprouting toward the lesion site at 10 d (white bars) compared with 2 d (black bars) post-SCI at all distances analyzed (n = 7 per group, **p < 0.001, n.s., not significant). c, d, Examples of regenerating axons at 10 d post-SCI from neurons at the lesion edge. At this time, when glial bridges have not yet formed, new neurites present at the lesion are not yet aligned in the anterior–posterior axis, but instead growing perpendicular (arrowheads, c′, d). New small-bodied neurons near the lesion edge show extensive sprouting with multiple branches (arrowheads, d) arising from individual neurites (arrow). f, In nestin:GFP fish, low and high-power images of traced regenerated neurons, showing that regenerating axons at 10 d post-SCI have not yet entered the lesion site (f2). At the lesion, traced axons (arrowhead, f1) do occasionally align with glial processes (nestin positive—arrow, f1) (n = 5), whenever present. e, Quantitation shows that at all times post-SCI the majority of axons regenerate along glia processes. g–k, By 2 or 3 weeks post-SCI, the majority of TMRD-labeled neurites extend across the lesion site along GFAP-positive glial processes (2 weeks, n = 7; 3 weeks, n = 10) (g, h). New islet1:GFP-labeled neurons also regenerate along GFAP-positive glia processes (k, k′). Processes of islet1:GFP-labeled neurons, however, do not colabel with the tracer, suggesting that although they have grown across the lesion, they have not reached far enough downstream to contact the tracer injection site (n = 11) (i, j). l–l″, A proportion of new neurons with regenerating processes (arrows) at the lesion site incorporate BrdU (arrowhead). Retrograde label accumulates both in mature neuron cell bodies upstream of the lesion site (m, m′, n, p, arrowheads) and also in labeled processes terminating around unlabeled neuronal cell bodies (n, arrow; o). Scale bars: a, 200 μm; c, i, 100 μm; c′, d, f1–f1′, f2–f2′, i′, j–l′, m′, 50 μm; g, h, 25 μm; n–p, 10 μm.