A distinct subgroup of small DRG cells express GDNF receptor components and GDNF is protective for these neurons after nerve injury

Abstract

Several lines of evidence suggest that neurotrophin administration may be of some therapeutic benefit in the treatment of peripheral neuropathy. However, a third of sensory neurons do not express receptors for the neurotrophins. These neurons are of small diameter and can be identified by the binding of the lectin IB4 and the expression of the enzyme thiamine monophosphatase (TMP). Here we show that these neurons express the receptor components for glial-derived neurotrophic factor (GDNF) signaling (RET, GFRalpha-1, and GFRalpha-2). In lumbar dorsal root ganglia, virtually all IB4-labeled cells express RET mRNA, and the majority of these cells (79%) also express GFRalpha-1, GFRalpha-2, or GFRalpha-1 plus GFRalpha-2. GDNF, but not nerve growth factor (NGF), can prevent several axotomy-induced changes in these neurons, including the downregulation of IB4 binding, TMP activity, and somatostatin expression. GDNF also prevents the slowing of conduction velocity that normally occurs after axotomy in a population of small diameter DRG cells and the A-fiber sprouting into lamina II of the dorsal horn. GDNF therefore may be useful in the treatment of peripheral neuropathies and may protect peripheral neurons that are refractory to neurotrophin treatment.

Fig. 1.

Cell size distribution of DRG cell profiles positively and negatively labeled for RET, GFRα-1, and GFRα-2 within L4/5 dorsal root ganglia. RET and GFRα-2 are present predominantly in small and intermediate diameter DRG cell profiles but are also present in some large diameter DRG cell profiles. GFRα-1 is more evenly distributed through the whole cell size spectrum.

Fig. 2.

Expression of RET, GFRα-1, and GFRα-2 in IB4-labeled DRG cells. In situ hybridization for RET (a), GFRα-1 (c), or GFRα-2 (e) was combined with IB4 labeling (b, d, f). a and bshow that many IB4 cells express RET (arrows indicate double-labeled cells). A similar pattern is seen in cand d and in e and f in relation to GFRα-1 and GFRα-2, except that the GFRα components are expressed in a smaller proportion of IB4 cells. Long arrows indicate IB4-labeled cells that express GFRα-1 or GFRα-2, whereas short open arrows indicate IB4 cells that do not express GFRα-1 or GFRα-2. Scale bar, 50 μm.

Fig. 3.

GFRα-1 and GFRα-2 expression in IB4 cells. GFRα-1 and GFRα-2 are coexpressed in one group of IB4 cells, expressed separately in other groups, and not expressed at all in a fourth group. Serial sections are shown triple-labeled for trkA (a, d), IB4 (b, e), and either GFRα-1 (c) or GFRα-2 (f) mRNAs. A IB4 cell expressing only GFRα-1 is identified by an arrow. Anarrowhead indicates a IB4 cell that expresses only GFRα-2. The double arrow indicates a IB4 cell that expresses both GFRα-1 and GFRα-2. Note that none of these cells are trkA immunoreactive. The star indicates a IB4 cell that expresses neither GFRα-1 nor GFRα-2. This cell is also trkA immunoreactive. Also shown is a large cell (asterisk) that expresses both GFRα-1 and GFRα-2. It is not IB4-labeled or trkA immunoreactive. Scale bar, 50 μm.

Fig. 4.

Colocalization of RET immunoreactivity with neurochemical markers in DRG cells and spinal cord.a–f, Dual labeling showing RET immunofluorescence (a, c, e) combined with IB4 labeling (b), trkA immunofluorescence (d), and CGRP immunofluorescence (f) in DRG cells. Arrowsindicate extensive colocalization of RET and IB4 in small diameter cells (a, b). Note that all IB4 cells show RET immunoreactivity. However, several RET positive cells do not bind IB4 (asterisks). RET labeling is not evident in many trkA cells (c, d). Asterisks denote trkA cells that are not co-labeled for RET. Similarly, few CGRP-expressing DRG cells are RET immunoreactive (e, f).Asterisks indicate cells that do not express RET but are labeled for CGRP. The arrow indicates a cell that is dual-labeled. g–j, Low-magnification (g, i) and high-magnification (h, j) micrographs showing RET immunofluorescence (g, h) and IB4 (i, j) double labeling in the dorsal horn of the spinal cord. Labeling is most intense in inner lamina II.Arrows in h and j indicate individual double-labeled axons. Scale bars (shown inf): a–f, 50 μm; (shown ini): g, i, 100 μm; (shown in j): h, j, 30 μm.

Fig. 5.

Histochemistry for TMP (a–d), IB4 labeling (e–h), and CGRP immunofluorescence (i–l) in dorsal root ganglia of control animals (a, e, i), animals with unilateral sciatic nerve section (b, f, j), animals with unilateral sciatic nerve section combined with intrathecal GDNF treatment (12 μg/d) (c, g, k), and animals with unilateral sciatic nerve section combined with intrathecal NGF treatment (12 μg/d) (d, h, l). Sciatic nerve section causes a loss of TMP (b) and IB4 (f) labeling, which is prevented by GDNF treatment (c, g) but not by NGF (d, h). In contrast, the loss of CGRP staining caused by sciatic nerve section (j) is prevented by NGF (l) but not by GDNF (k). Scale bar, 50 μm. CTRL, Control; AXOT, axotomized.

Fig. 6.

Histochemistry for TMP at the level of L4 in the dorsal horn of animals with unilateral sciatic nerve section (a), animals with unilateral sciatic nerve section combined with high (12 μg/d) (b) and low (1.2 μg/d) (c) dose intrathecal GDNF treatment, and animals with unilateral sciatic nerve section combined with high (12 μg/d) (d) and low (1.2 μg/d) (e) dose intrathecal NGF. Sciatic nerve section causes a loss of TMP in the sciatic termination territory within lamina IIi (demonstrated by arrows). GDNF treatment at either dose is effective at preventing this loss (b, c), whereas NGF is much less effective at either dose used (d, e). Scale bar, 100 μm. Axot., Axotomized.

Fig. 7.

a, The ratio of the area occupied by IB4, TMP, or CGRP stained terminals within lamina II of the dorsal horn of the spinal cord on the axotomized side versus the normal side in animals that have undergone axotomy (n = 4) or axotomy in combination with an intrathecal infusion of GDNF at a dose of either 1.2 μg/d (n = 3) or 12 μg/d (n = 4) or NGF at a dose of either 1.2 μg/d (n = 3) or 12 μg/d (n = 3). GDNF at a dose of 12 μg/d almost completely prevented the axotomy-induced reduction in staining intensity of IB4 and TMP (p < 0.001; unpaired t test; comparing GDNF with no treatment after axotomy). The lower dose of GDNF (1.2 μg/d) also had a significant effect in preventing the axotomy-induced reduction in staining intensity of these markers but was less effective than the higher dose. The high dose GDNF had a small but significant effect in preventing the axotomy-induced reduction in CGRP staining (p < 0.05; unpairedt test). NGF could almost completely prevent the axotomy-induced reduction in CGRP staining (p < 0.001; unpaired t test; comparing NGF with no treatment after axotomy). NGF at 12 μg/d had a small but significant effect on the axotomy-induced reduction in IB4 and TMP expression (p < 0.05; unpairedt test). b, The ratio of the area occupied by CTB-labeled terminals in lamina II compared with lamina III of the dorsal horn in normal (n = 5), axotomized (n = 4), and axotomy + GDNF (Axot. GDNF) 1.2 μg/d (n = 3) and 12 μg/d (n = 4) animals. Note that there is a significant increase in labeling in lamina II after axotomy (p < 0.01; unpaired t test), which is almost completely prevented by treatment with GDNF at the higher dose. GDNF treatment at the low dose also had a significant effect (p < 0.01 compared with no treatment; unpaired t test) but was less effective than the higher dose.

Fig. 8.

Transport of CTB to the dorsal horn of the spinal cord at the level of L3 (a–c), L4 (d–f), and L5 (g–i) after sciatic nerve label in control (CTRL) animals (a, d, g), animals that have undergone axotomy (AXOT) (b, e, h), and animals that have undergone axotomy combined with GDNF (AXOT+GDNF) treatment (12 μg/d) (c, f, i). In the normal animal, CTB-labeled terminals are present in lamina I and the deeper laminae of the dorsal horn (III–IV) but are excluded from lamina II (a, d, g). After axotomy, CTB-labeled terminals appear in lamina II (denoted byasterisks), and there is also more dense labeling of axon bundles within lamina I (arrows in b, e, h). These axotomy-related changes are prevented by treatment with GDNF, where the CTB labeling pattern appears the same as control, and this is seen consistently throughout L3–L5 (c, f, i). Scale bar, 100 μm.

Fig. 9.

The conduction velocity (CV) of C-fibers projecting through the tibial nerve was measured by stimulation of that nerve electrically and recording and averaging activity in fine strands of the L5 dorsal root. a shows a representative recording from an animal in which the tibial nerve had been cut and tied 2 weeks previously; the animal was treated continuously with intrathecal GDNF and NGF (each at 12 μg/d). Arrowsshow examples of individual C-fiber potentials occurring in response to the stimulation. b, Cumulative sum plots showing the average CV distributions constructed from groups of animals receiving different treatment (n = 3–5 animals per group). Error bars show SEM. Note that axotomy results in a significant slowing of C-fibers (seen as a leftward shift in the Qsum plots), and both NGF and GDNF partially prevent this slowing.