Neuroscience

Translational control in the spinal cord regulates gene expression and pain hypersensitivity in the chronic phase of neuropathic pain

Department of Anesthesia, McGill University, Montreal, Canada
Faculty of Dental Medicine and Oral Health Sciences, McGill University, Montreal, Canada
Alan Edwards Centre for Research on Pain, McGill University, Montreal, Canada
Department of Biochemistry and Goodman Cancer Research Centre, McGill University, Montreal, Canada
Département de biomédecine vétérinaire, Faculté de médecine vétérinaire, Université de Montréal, Saint-Hyacinthe, Canada
Centre de recherche sur le cerveau et l’apprentissage (CIRCA), Université de Montréal, Montreal, Canada
Patrick G. Johnston Centre for Cancer Research, Queen’s University Belfast, Belfast United Kingdom
Department of Neuroscience and Center for Advanced Pain Studies, University of Texas at Dallas, Dallas, United States
Ionis Pharmaceuticals, Inc., Carlsbad, United States
Biomedical Research Institute, Foundation for Research and Technology-Hellas, University Campus, Ioannina, Greece
Department of Psychology, Faculty of Science, McGill University, Montreal, Canada

https://doi.org/10.7554/eLife.100451.2

Open access
Copyright information

Figures and data

Transcriptional and translational analysis of gene expression using Ribo-seq.

(A) A schematic diagram showing the spared nerve injury (SNI) assay. S: Sural branch, T: Tibial branch and CP: Common peroneal branch. (B) An illustration of the ribosome profiling technique (Ribo-seq). Scatter plot shows ribosomal footprint (rFP) log₂ fold change (FC), reflecting translational changes, as a function of mRNA log₂ fold change for dorsal root ganglia (DRG, C) and spinal cord (SC, E), at day 4 and day 63 post-SNI in female mice. Each dot is a gene. Fold change evaluated between SNI and sham conditions. Color coding indicates modality of differential gene expression control, either at the transcriptional level (mRNA, magenta) or at the translational level (rFP, blue). Under each scatter plot, a list of the top 10 upregulated and downregulated genes (at the mRNA and rFP levels) is shown for each condition. Supplementary Figure 1 shows volcano plots for all conditions. Supplementary Table 1 includes complete datasets (worksheets 1–4), with yellow highlighting category of change in gene expression (translation only, transcription only, stable, opposite change, and homodirectional), grey indicating mRNA log₂FC, and blue indicating rFP log₂FC. (D) Number of genes showing changes at mRNA and rFP levels across independent biological replicates. The rFP/mRNA ratio for each condition is shown above the columns. (F) Pathway analyses of translationally regulated genes in the SC at day 63 post-SNI in the KEGG and Reactome databases.

Targeting spinal translation alleviates pain hypersensitivity at the late stage after peripheral nerve injury.

(A) A schematic showing the regulation of eIF4E via mTORC1/4E-BP1 and MAPKs/MNK pathways. eIF4E ASO (i.c.v) reduces eIF4E protein levels in the spinal cord (B) but not DRG (C) two weeks after administration (n = 4/group). (D) Time course of ASO (eIF4E and control) administration after SNI. The effect of ASO on von Frey (50% withdrawal threshold: E, n = 9/group) and Mouse Grimace Scale (MGS) (F, n = 9/10 mice per group). (G) Time course of ASO administration before SNI and its effect on the von Frey (50% withdrawal threshold: H, n = 11/12 mice per group) and MGS (I, n = 11/12 mice per group) tests. An unpaired two-tailed t-test was used in B, C, F, and I. Two-way ANOVA followed by Tukey’s post-hoc comparison was used in E and H. Each data point represents an individual animal. A comparable number of male and female mice was used in all experiments. Data are plotted as mean ± s.e.m. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, ns – not significant.

Assessment of protein synthesis using metabolic labeling.

(A) Illustration of protein synthesis assessment using FUNCAT. Anisomycin (100 mg/kg, i.p. injection 1 h before AHA injection) treatment blocked AHA incorporation (n = 3 female mice per group, normalized to the control group), demonstrating the validity of the approach. AHA signal in the superficial spinal cord (laminae I-III, defined based on NeuN staining) was quantified in inhibitory neurons (Pax2⁺, examples marked by white arrow) and excitatory neurons (Pax2^-/NeuN⁺) at day 4 (C, n = 6 female mice per group) and day 60 (D, n = 5 female mice per group) post-SNI. AHA signal intensity (integrated density on maximum-intensity projection images) in the soma of inhibitory and excitatory neurons were averaged across 25 cells/mouse to obtain a single value for each mouse (see Methods for details of the analysis). Scale bars: 50 μm for B and 30 μm for C, D. An unpaired two-tailed t-test was used. Each data point represents an individual animal. Data are plotted as mean ± s.e.m. *p < 0.05, **p < 0.01, ns – not significant.

Cell-type-specific profiling of spinal gene expression after peripheral nerve injury.

(A) A schematic illustrating the TRAP approach to assess gene expression in specific cell types. Confirmation of the specificity of IP fractions for inhibitory neurons in the L10a-eGFP; Gad2^Cre mouse line (B) and for excitatory neurons in the L10a-eGFP; Tac1^Cre mouse line (C). Experiments were performed in female mice. Dual flashlight plots (left) show the strictly standardized mean difference (SSMD) versus log₂ FC for genes in IP samples and panels on the right show the top 15 upregulated and downregulated genes for inhibitory neurons at day 4 (D) and 60 (E), and Tac1⁺ excitatory neurons at day 4 (F) and 60 (G) post-SNI. Positive Log₂ FC indicates increased expression in SNI compared to sham mice. Parameters for defining data as upregulated in SNI are indicated at the top. Supplementary Table 1 includes complete datasets (worksheets 5–8), with yellow highlighting log₂FC values in IP samples and orange highlighting log₂FC values in IN samples. (H) The number of altered genes in each condition (GAD2 D60: SNI versus sham day 60 in GAD2⁺ neurons; GAD2 D4: SNI versus sham day 4 in GAD2⁺ neurons; Tac1 D60: SNI versus sham day 60 in Tac1⁺ neurons; and Tac1 D4: SNI versus sham day 4 in Tac1⁺ neurons).

Activation of 4E-BP1-dependent translation in inhibitory neurons promotes mechanical hypersensitivity and contributes to reduced intrinsic excitability of PV neurons.

(A) A schematic of mTORC1 pathway. Deletion of 4E-BP1 in GAD2 (B, 4E-BP1 cKO: Eif4ebp1^fl/fl;Gad2^Cre, Control: Gad2^Cre, n = 9/10) and PV (C, 4E-BP1 cKO: Eif4ebp1^fl/fl;Pv^Cre, Control: Pv^Cre, n = 8/11) neurons induces mechanical (50% withdrawal threshold) but not heat hypersensitivity. A comparable number of male and female mice was used in B and C. Recording from PV neurons in spinal cord slices (identified by the expression of L10a-eGFP) shows that the deletion of 4E-BP1 in PV neurons (4E-BP1 cKO: Eif4ebp1^fl/fl: L10a-eGFP: Pv^Cre, Control: L10a-eGFP: Pv^Cre, n = 8/8 female mice) induces a decrease in firing frequency (D) and an increase in rheobase (E). No change in membrane capacitance (F), resting membrane potential (RMP, G), and input resistance (Rin, H) were found. AAVs (AAV-CAG-DIO-eGFP-eIF4E-shRNAmir or AAV-CAG-DIO-EGFP-scrambled-shRNAmir) were injected into the parenchyma of the dorsal horn of Pv^Cre female mice (illustration and time course are shown in I, n = 8/group), preventing the SNI-induced decrease in PV neuron firing frequency (J) and elevation of rheobase (K). No changes were found in capacitance (L), RMP (M), and Rin (N). An unpaired two-tailed t-test was used in B, C, E-H. Two-way ANOVA followed by Tukey’s post-hoc comparison was used in J-N. Each data point represents an individual animal. Data are plotted as mean ± s.e.m. *p < 0.05, **p < 0.01, ***p < 0.001, ns – not significant.

Effects of modulating eIF4E-dependent translation on pain hypersensitivity.

(A) eIF4E was downregulated in PV neurons by intraspinal injection of AAV-CAG-DIO-eGFP-eIF4E-shRNAmir (and control AAV-CAG-DIO-EGFP-scrambled-shRNAmir) into the parenchyma of the dorsal horn of Pv^Cre mice two weeks before SNI (n = 10/8). (B) Mice expressing a mutated non-phosphorylatable 4E-BP1 in PV neurons (4EBP1mt;Pv^Cre) and their controls (Tg-4EBP1mt) were subjected to SNI (n = 9/8). No reduction in SNI-induced mechanical hypersensitivity (von Frey, 50% withdrawal threshold) was observed in A or B. No changes were observed in mechanical (von Frey, 50% withdrawal threshold) and heat (radiant heat paw-withdrawal) thresholds in mice lacking 4E-BP1 in Vglut2 neurons (C, 4E-BP1 cKO: Eif4ebp1^fl/fl;Vglut2^Cre, Control: Vglut2^Cre) or Tac1 neurons (D, 4E-BP1 cKO: Eif4ebp1^fl/fl;Tac1^Cre, Control: Tac1^Cre, n = 7/8 mice). A comparable number of male and female mice was used in A-D. An unpaired two-tailed t-test. Data are plotted as mean ± s.e.m. ns – not significant.

Volcano plots showing changes in mRNA (top), ribosome footprint (rFP, middle), and translational efficiency (TE, bottom) levels in the DRG and SC tissues at day 4 post-SNI (A) and day 63 post-SNI (B).

π-values⁶⁰ calculated as log₂(FC) · -log₁₀(P), given an expression fold-change (FC; X axis) and its associated P-value (P; Y axis). Statistical significance at the alpha = 0.2 level; decreased (magenta) or increased fold-change (green) expression in SNI versus sham.

Confirmation of 4E-BP1 and eIF4E downregulation.

(A) Lumbar spinal cord tissue from 4E-BP1 cKO GAD2 (Eif4ebp1^fl/fl: L10a-eGFP: Gad2^Cre) and Control (L10a-eGFP: Gad2^Cre) mice was immunostained for 4E-BP1. eGFP expression indicates GAD2⁺ neurons (marked by white arrows). (B) Lumbar spinal cord tissue from 4E-BP1 cKO PV (Eif4ebp1^fl/fl: L10a-eGFP: Pv^Cre) and Control (L10a-eGFP: Pv^Cre) mice were immunostained for 4E-BP1. eGFP expression indicates PV⁺ neurons (marked by white arrows). (C) Lumbar spinal cord tissue from 4E-BP1 cKO Vglut2 (Eif4ebp1^fl/fl: Vglut2^Cre) and Control (Pv^Cre) mice was immunostained for 4E-BP1. Excitatory neurons were identified as NeuN⁺/Pax2^- (white arrows mark inhibitory neurons). (D) Lumbar spinal cord tissue from 4E-BP1 cKO Tac1 (Eif4ebp1^fl/fl: L10a-eGFP: Tac1^Cre) and Control (L10a-eGFP: Tac1^Cre) mice was immunostained for 4E-BP1. eGFP expression indicates Tac1⁺ neurons (marked by white arrows). (E) AAVs (AAV9-CAG-DIO-eGFP-eIF4E-shRNAmir and AAV9-CAG-DIO-eGFP-scrambled) were injected into the lumbar spinal cord of Pv^Cre mice and immunohistochemistry against eIF4E was performed 14 days later. Scale bar is 20 µm in all images. An unpaired two-tailed t-test was used. Each data point represents an individual animal. Data are plotted as mean ± s.e.m. *p < 0.05, **p < 0.01.

Sign up for email alerts