Peripheral nerve injury results in a biased loss of sensory neuron subpopulations

Abstract

There is a rich literature describing the loss of dorsal root ganglion (DRG) neurons following peripheral axotomy, but the vulnerability of discrete subpopulations has not yet been characterised. Furthermore, the extent or even presence of neuron loss following injury has recently been challenged. In this study, we have used a range of transgenic recombinase driver mouse lines to genetically label molecularly defined subpopulations of DRG neurons and track their survival following traumatic nerve injury. We find that spared nerve injury leads to a marked loss of cells containing DRG volume and a concomitant loss of small-diameter DRG neurons. Neuron loss occurs unequally across subpopulations and is particularly prevalent in nonpeptidergic nociceptors, marked by expression of Mrgprd. We show that this subpopulation is almost entirely lost following spared nerve injury and severely depleted (by roughly 50%) following sciatic nerve crush. Finally, we used an in vitro model of DRG neuron survival to demonstrate that nonpeptidergic nociceptor loss is likely dependent on the absence of neurotrophic support. Together, these results profile the extent to which DRG neuron subpopulations can survive axotomy, with implications for our understanding of nerve injury-induced plasticity and pain.

Figure 1.

SNI_trans induces death of small primary afferent neurons, accompanied by a reduction in volume, not cell density, of the dorsal root ganglion. (A) Approach to differentially labelled intact afferents with tdTomato and damaged afferents with GFP after peripheral nerve injury using the Avil^FlpO;Atf3^CreERT2;RC::FLTG mouse line and schematic of experimental timeline. (B) Representative image of GFP, tdTomato, and NeuN expression in an L4 DRG, 2 weeks after SNI_trans. Scale bars = 100 µm. (C and D) Stereological quantification of the total number of DRG neurons (C) or number of axotomized and intact neurons (D) in the L4 DRG 1, 2, 4, and 8 weeks after SNI_trans or contralateral (contra) to injury. (C) One-way ANOVA with Tukey posttests; F_4,10 = 37.98, P < 0.001. (D) Two-way RM ANOVA; Timepoint × Color interaction F_4,10 = 39.04, P < 0.001, n = 3 mice; Tukey posttests (between injured groups): †P < 0.05 vs contra, ‡P < 0.05 vs 1-week. (E) Volume of DRG-containing cells (ie, excluding white matter tracts) following SNI_trans. One-way ANOVA with Tukey posttests; F_4,10 = 21.25, P < 0.001, n = 3. (F) Neuronal density within the DRG following SNI_trans. One-way ANOVA; F_4,10 = 2.77, P = 0.09, n = 3. (G) Population distribution of uninjured and injured afferents by cross-sectional area, 1 and 8 weeks post-SNI_trans. Kolmogorov–Smirnov tests of cumulative distributions; Uninjured: D = 0.08, P = 0.18; Injured: D = 0.32, P < 0.001; n = 310 to 427 neurons from 3 mice. *P < 0.05, **P < 0.01, ***P < 0.001 vs contra. ANOVA, analysis of variance; DRG, dorsal root ganglion; GFP, green fluorescent protein.

Figure 2.

Spared nerve crush and transection lead to a loss of small DRG neurons. (A) Approach to restrict analysis to damaged afferents: a subcutaneous injection of the tracer FB into both hindpaws labelled tibial afferents, before unilateral SNI_trans or SNI_crush surgery. (B) Representative image of FB labelling and NeuN immunostaining in the L4 DRG. The image is a projection of optical sections at 3-µm intervals through the entirety of a 30-µm-thick tissue section. Scale bar = 100 µm. (C and D) Quantification of the cross-sectional area of FastBlue labelled DRG neurons ipsilateral and contralateral to SNI_trans (C) or SNI_crush injury (D) reveals a loss of small afferents and subsequent shift in population distribution. Kolmogorov–Smirnov tests of cumulative distributions; SNI_trans: D = 0.25, P < 0.001; n = 183 or 191 neurons from 3 mice; SNI_crush: D = 0.22, P < 0.001, n = 319 or 325 neurons from 3 mice. (E) Experimental approach for whole DRG volumetric analyses after SNI_trans. (F) Representative 3D rendering of TDP-43 profiles and corresponding nuclear spot profiles following Imaris-based spot detection feature. Scale bar = 100 µm. (G) Quantification of DRG nuclear spot volume ipsilateral and contralateral to SNI_trans. Kolmogorov–Smirnov tests of cumulative distribution: D = 0.06, P < 0.001, n = 30,206 (contra) or 32,544 (ipsi) nuclei from 4 (contra) or 5 (ipsi) mice. (H) Total number of nuclear spots, by size, per DRG. Two-way RM ANOVA; size bin × injury interaction: F_2,14= 8.26, P = 0.004; n = 4 to 5 mice; Šídák multiple comparisons tests: **P < 0.01. ANOVA, analysis of variance; DRG, dorsal root ganglion; FB, FastBlue; RM, repeated measures.

Figure 3.

Spared nerve crush or transection results in death of nonpeptidergic neurons. (A) Schematic of experimental approach for (B and C). (B) MrgD^ChR2-YFP L4 DRGs 4 weeks after SNI, contralateral or ipsilateral to injury. Images are projections of optical sections at 3-µm intervals through the entirety of 30-µm-thick tissue sections. Scale bars = 100 µm. (C) Quantification of total number of MrgD-YFP+ cells per L4 DRG 4 weeks after SNI revealed a significant loss in ipsilateral DRG. Two-way RM ANOVA with Šídák multiple comparisons tests; Side x Treatment interaction: F_1,5 = 9.23, P = 0.029; n = 3 mice. (D) The experimental approach used to generate data presented in (E–G). (E and F) MrgD-YFP expression and FB labelling in the L4 DRG, 14 days after SNI or crush surgery or contralateral to injury. White boxes represent regions enlarged in (F). Scale bars = 100 µm (E) or 20 µm (F). (G) The proportion of FB-labelled DRG neurons decreased after spared nerve crush injury, and co-labelling is almost completely absent after SNI. Two-way RM ANOVA with Šídák multiple comparisons tests; side × injury interaction: F_1,4 = 7.80, P = 0.049; n = 3 mice. Posttests: *P < 0.05, **P < 0.01. ANOVA, analysis of variance; DRG, dorsal root ganglion; SNI, spared nerve injury; FB, FastBlue; RM, repeated measures.

Figure 4.

Spared nerve injury induces a loss of Trpm8+ and CGRP+ but not myelinated DRG neurons. (A) Schematic of experimental approach. (B–D) FastBlue labelling and Trpm8-tdTom (B), Calca-YFP (C), or Thy1-CFP expression (D) 28 days after SNI_trans in the L4 DRG, contralateral (top) or ipsilateral (bottom) to injury. Images are projections of optical sections at 3-µm intervals through the entirety of 30-µm-thick tissue sections. Scale bars = 100 µm. (E–G) Quantification of the proportion of FB-labelled neurons also expressing Trpm8-tdTom (E), Calca-YFP (F), or Thy1-CFP (G) in L4 DRG contralateral or ipsilateral to SNI_trans. Paired t tests; Trpm8-tdTom: t₂ = 5.31, P = 0.034, n = 3 mice; Calca-YFP: t₃ = 4.12, P = 0.026, n = 4 mice; Thy1-CFP: t₃ = 4.42, P = 0.022, n = 4 mice. *P < 0.05. CFP, cyan fluorescent protein; CGRP, calcitonin gene-related peptide; DRG, dorsal root ganglion; FB, FastBlue.

Figure 5.

Neurotrophic support ameliorates MrgD+ cell loss in vitro. (A–E) Representative fields of view of MrgD-YFP and β-tubulin III expression in neuronal cultures of isolated MrgD^CreERT2;Ai32 DRGs, 24 hours (A) or 4 weeks (B–E) after plating, with the addition of the growth factors (GFs) NGF, GDNF, both, or neither. Scale bars = 20 µm. (F) Quantification of the percentage of neurons that express MrgD-YFP under the conditions shown in (A–E). One-way ANOVA with Tukey posttests; F₄,₁₅ = 39.7, P < 0.001; n = 4 mice. Posttests: *P < 0.05, **P < 0.01, ***P < 0.001. (G–K) Representative fields of view of Calca-YFP and β-tubulin III expression in neuronal cultures of isolated Calca^CreERT2;Ai32 DRGs, with the addition of NGF, GDNF, both, or neither. Scale bars = 20 µm. (L) Quantification of the percentage of neurons that express Calca-YFP. One-way ANOVA; F₄,₁₄ = 0.46, P = 0.76; n = 3 to 4 mice. ANOVA, analysis of variance; DRG, dorsal root ganglion; GDNF, glial-derived neurotrophic factor; NGF, nerve growth factor; YFP, Yellow fluorescent protein.