To study changes in gene expression at both transcriptional (transcriptome) and translational (translatome) levels, we employed Ribo-seq on DRG and lumbar spinal cord tissue obtained from mice subjected to an experimental assay of peripheral nerve damage-induced (i.e., neuropathic) pain, spared nerve injury (SNI, Figure 1A; Decosterd and Woolf, 2000), or sham surgery. SNI prominently features mechanical hypersensitivity, which develops within 2–4 days of the nerve injury and persists for many months (Millecamps et al., 2023). L3–L5 DRGs and the ipsilateral dorsal half of the corresponding segment of the lumbar spinal cord (illustrated in a schematic diagram in Figure 1A) were collected on day 4 (early phase) and day 63 (late phase) post-SNI and processed for Ribo-seq. Ribo-seq allows the identification of mRNA fragments (ribosome footprints [rFPs]) where translating ribosomes are bound. These mRNA fragments thus remain protected from nuclease-mediated RNase degradation, and thereby reveal the number and location of ribosomes on specific transcripts (Figure 1B; Ingolia et al., 2012). Normalization to the corresponding transcript abundance from the parallel mRNA-seq analysis provides a measure of mRNA translation efficiency on a genome-wide scale. Using this approach, we identified transcriptionally and translationally regulated genes in each tissue and time point (n = 3 biological replicates/condition, 15 mice pooled per replicate). In the DRG, a significant number of transcripts were altered at both transcriptional and translational levels at day 4 post-SNI (Figure 1C, D; Figure 1—figure supplement 1A provides volcano plots for all conditions; full datasets are provided in Supplementary file 1). At day 63 post-SNI, the number of transcriptionally altered mRNAs in the DRG decreased compared to day 4 (Figure 1C, D; Figure 1—figure supplement 1B). In the spinal cord, changes in gene expression were less pronounced and on day 4 post-SNI, a comparable number of differentially transcribed genes and differentially translated genes were identified (Figure 1D, E; Figure 1—figure supplement 1A). Surprisingly, on day 63 post-SNI, changes in gene expression in the spinal cord occurred predominantly at the translational but not transcriptional level (Figure 1D, E; Figure 1—figure supplement 1B). Gene Ontology pathway analysis of translationally regulated genes in the spinal cord on day 63 post-SNI revealed alterations in processes related to extracellular matrix organization and its interaction with cell surface receptors, cell adhesion, and protein turnover (Figure 1F). Collectively, these results indicate that both transcriptional and translational mechanisms mediate changes in gene expression in the DRG in both the early and late phases of neuropathic pain and in the spinal cord in the early phase. In the late chronic phase, however, changes in gene expression in the spinal cord are controlled more prominently at the translational level.

The important role of spinal translation in regulating changes in gene expression in the chronic phase of neuropathic pain prompted us to test whether targeting translation in the spinal cord can alleviate established pain hypersensitivity. Translation initiation and the activity of cap-binding protein eIF4E, which facilitates the recruitment of ribosomes to the mRNA, are rate-limiting steps in protein synthesis (Gingras et al., 2001; Tahmasebi et al., 2018). eIF4E is a key translation initiation factor regulating pain-related plasticity as it integrates information from two signaling pathways that are activated in neuropathic pain to stimulate mRNA translation and promote pain hypersensitivity: The mechanistic target of rapamycin complex 1 (mTORC1), and mitogen-activated protein kinases (MAPKs), such as ERK and p38 (Khoutorsky and Price, 2018; Ji et al., 2009; Melemedjian and Khoutorsky, 2015; Figure 2A). Moreover, previous studies have revealed that a partial reduction in eIF4E expression (e.g., in Eif4e^+/− mice and in mice treated with eIF4E shRNA) is both well-tolerated and alleviates adverse phenotypes in cancer (Truitt et al., 2015) and autism spectrum disorder (Gkogkas et al., 2013; Santini et al., 2013) mouse models. To study the role of mRNA translation in chronic pain hypersensitivity, we used an antisense oligonucleotide (ASO) against mouse Eif4e to modulate eIF4E expression. To target eIF4E in the central nervous system but not the DRG, we injected eIF4E-ASO via the intracerebroventricular (i.c.v., 100 mg/kg) route (Mohan et al., 2018), resulting in a ~38% reduction in eIF4E protein expression in the lumbar spinal cord 2 weeks post-injection (Figure 2B), without changing eIF4E levels in the DRG (Figure 2C). We first assessed the effect of eIF4E downregulation on established pain hypersensitivity. Injection of eIF4E-ASO at week 6 post-SNI alleviated mechanical pain hypersensitivity in the von Frey test 2 weeks later (at week 8 post-SNI; the experimental time course is provided in Figure 2D, von Frey data in Figure 2E). Reduced hypersensitivity persisted for 4 additional weeks (up to week 12 post-SNI), demonstrating a long-lasting effect following a single eIF4E-ASO administration. eIF4E-ASO also attenuated spontaneous pain, as assessed using the Mouse Grimace Scale (MGS) on week 8 post-SNI (Figure 2F).

We then assessed the effect of downregulating eIF4E during the early development phase of neuropathic pain by administering eIF4E-ASO and control-ASO 10 days before the peripheral nerve injury. Surprisingly, we found no alleviation of mechanical hypersensitivity at day 4 post-SNI in eIF4E-ASO-injected mice (the experimental time course is provided in Figure 2G, von Frey data in Figure 2H). However, testing at later time points showed that mice injected with eIF4E-ASO exhibited reduced mechanical pain hypersensitivity at week 2 after the nerve injury (Figure 2H), and the effect became more pronounced at weeks 4 and 8. The MGS was also reduced in eIF4E-ASO-injected mice at week 4 and 8 post-SNI (Figure 2I). To control for potential non-specific effects of eIF4E-ASO following i.c.v. administration, we performed rotarod and open field tests, which revealed no differences in locomotor function between mice injected with eIF4E-ASO and control-ASO (Figure 2J, K). Together, these results indicate that downregulation of eIF4E in the spinal cord using ASO alleviates pain hypersensitivity in the late but not the acute stages of neuropathic pain.

Ribo-seq provided a comprehensive characterization of translational landscape in DRG and spinal cord tissues during the early and late phase of neuropathic pain. However, this approach does not allow the measurement of gene expression in distinct neuronal subtypes and cannot distinguish between neuronal and non-neuronal cells. To assess protein synthesis in specific neuronal subpopulations, we used fluorescence noncanonical amino acid tagging (FUNCAT) (Hooshmandi et al., 2024; Dieterich et al., 2010), focusing on two major neuronal subtypes in the spinal cord, excitatory and inhibitory neurons. In FUNCAT, mice are injected with a noncanonical amino acid, azidohomoalanine (AHA), which is charged onto methionine tRNA and incorporated into newly synthesized proteins (Figure 3A). Visualization of AHA incorporation using click chemistry and fluorescent labeling provides a measure of de novo general protein synthesis in spinal cord sections. The specificity of this approach was validated using the protein synthesis inhibitor anisomycin, which blocked AHA incorporation in the spinal cord (Figure 3B). FUNCAT analysis showed that AHA incorporation increased at day 4 and 60 post-SNI in Pax2⁺ inhibitory neurons (day 4: Figure 3C; day 60: Figure 3D), whereas no statistically significant changes were found in excitatory neurons (NeuN⁺ and Pax2⁻).

Next, we employed the TRAP approach to identify specific mRNAs that are actively translated in excitatory and inhibitory neurons. In TRAP, the eGFP-tagged ribosomal protein, L10a, is expressed in a genetically defined cellular population (via a specific gene promoter), followed by IP of tagged ribosomes with an anti-eGFP antibody and the sequencing of ribosome-bound mRNAs (Figure 4A). We performed TRAP analysis on two major subpopulations of neurons: a subset of excitatory neurons, defined by Tac1 (using L10a-eGFP; Tac1^Cre mice); and inhibitory neurons, defined by GAD2 (using L10a-eGFP; Gad2^Cre mice). Tac1 is expressed in a subset of excitatory interneurons and projection neurons in the spinal cord that play important roles in processing nociceptive information, as well as driving spinal plasticity and chronic pain-related behaviors (Barik et al., 2021; Huang et al., 2019). GAD2⁺ neurons encompass numerous subpopulations of spinal cord inhibitory neurons that are critical for the development and maintenance of neuropathic pain (Todd, 2010; Peirs et al., 2020). To this end, L10a-eGFP; Tac1^Cre and L10a-eGFP; Gad2^Cre mice were subjected to SNI or sham surgery (bilaterally), and lumbar dorsal spinal cord tissue was collected at days 4 and 60 after the nerve injury. mRNAs isolated from the immunoprecipitated (IP) and input (IN) samples were sequenced. Expression levels (IP/IN) of excitatory (Slc17a7, Tac1, Cck, Nts) and inhibitory (Slc32a1, Gad2, Pax2, Pvalb) neuronal markers, as well as markers of non-neuronal cells (e.g., Aldh1l1, Gfap, Tmem119, Cx3cr1), are shown in Figure 4B for L10a-eGFP; Gad2^Cre mice and in Figure 4C for L10a-eGFP; Tac1^Cre mice, demonstrating the specificity of the approach. Changes in ribosome occupancy were found in 126 mRNAs in the early phase and 223 mRNAs in the late phase in GAD2⁺ neurons, and 118 mRNAs in the early phase and 161 in the late phase in Tac1⁺ neurons (IP: Figure 4D–H, IP/IN datasets are provided in Supplementary file 1), suggesting higher translational changes at day 60 post-SNI in GAD2⁺ inhibitory neurons compared to Tac1⁺ excitatory neurons. Together, these results establish translational changes in GAD2⁺ and Tac1⁺ neurons in the early and late phases of neuropathic pain.

The pronounced upregulation of mRNA translation in inhibitory and excitatory neurons after peripheral nerve injury prompted us to study its functional role in mediating pain hypersensitivity. mTORC1, a master regulator of mRNA translation, stimulates protein synthesis via phosphorylation of the translational repressor eIF4E-binding proteins (4E-BPs), triggering their dissociation from eIF4E to allow cap-dependent translation initiation (Figure 5A). Accordingly, deletion of 4E-BPs, which mimics the activation of the mTORC1–eIF4E axis, stimulates translation. There are three isoforms of 4E-BP (4E-BP1, 4E-BP2, and 4E-BP3), which exhibit similar functions but have different tissue distribution (Khoutorsky et al., 2015). 4E-BP1 is the main isoform in the pain pathway, as 4E-BP1, but not 4E-BP2 whole-body knockout mice show mechanical pain hypersensitivity (Khoutorsky et al., 2015), while 4E-BP3 expression is very low in the nervous system. To increase translation selectively in inhibitory or excitatory neurons, we generated mice lacking 4E-BP1 in each cell type (confirmation of 4E-BP1 downregulation is shown in Figure 5—figure supplement 1A–D). Deletion of 4E-BP1 in inhibitory neurons (Eif4ebp1^fl/fl;Gad2^Cre) induced mechanical hypersensitivity without affecting heat sensitivity (Figure 5B). A subpopulation of inhibitory neurons, PV⁺ interneurons, specifically gate mechanical allodynia (Petitjean et al., 2015; Boyle et al., 2019; Cao et al., 2022). Peripheral nerve injury induces substantial plasticity in PV neurons, resulting in a decrease in their intrinsic excitability and spiking activity, and the consequent disinhibition of postsynaptic PKCɣ interneurons and engagement of myelinated primary afferents in spinal nociceptive circuits (Petitjean et al., 2015; Boyle et al., 2019; Cao et al., 2022). Deletion of 4E-BP1 in PV neurons (Eif4ebp1^fl/fl;Pvalb^Cre) induced robust mechanical hypersensitivity, similar to that observed in Eif4ebp1^fl/fl;Gad2^Cre mice (Figure 5C; no change was found in heat sensitivity). Recording from lumbar spinal cord slices showed that PV neurons lacking 4E-BP1 exhibit reduced excitability, as evident by a decreased firing rate in response to a depolarization pulse (Figure 5D) and elevated rheobase compared to PV neurons from control mice (Figure 5E). No change was observed in membrane capacitance (Figure 5F), resting membrane potential (Figure 5G), and input resistance (Figure 5H). To study the role of translation in peripheral nerve injury-induced plasticity in spinal PV neurons, we selectively downregulated eIF4E in PV neurons in the lumbar spinal cord before SNI. To this end, an adeno-associated virus (AAV)-expressing shRNAmir against eIF4E (AAV9-CAG-DIO-eGFP-eIF4E-shRNAmir) was injected into the lumbar dorsal horn parenchyma of Pvalb^Cre mice 14 days before the SNI (Figure 5I shows experimental design, Figure 5—figure supplement 1E shows confirmation of reduced eIF4E levels). Recordings from PV neurons in spinal cord slices revealed that downregulation of eIF4E in PV neurons prevented the SNI-induced decrease in intrinsic excitability (Figure 5J). Whereas PV neurons from control mice (Pvalb^Cre mice injected with AAV9-CAG-DIO-eGFP-eIF4E-scrambled) exhibited reduced spiking activity and increased rheobase 4 weeks post-SNI compared to sham animals, PV neurons with reduced eIF4E showed no change in their excitability after nerve injury (Figure 5J: firing frequency; Figure 5K: rheobase). No change was found in membrane capacitance (Figure 5L), resting membrane potential (Figure 5M), and input resistance (Figure 5N).

We next investigated the effect of downregulating eIF4E-dependent translation in PV neurons on SNI-induced pain hypersensitivity using two complementary approaches. We first downregulated eIF4E by injecting AAV9-CAG-DIO-eGFP–eIF4E-shRNAmir into the lumbar dorsal horn parenchyma of Pvalb^Cre mice 14 days prior to SNI. Unexpectedly, this manipulation did not attenuate mechanical hypersensitivity compared with mice injected with AAV9-CAG-DIO-eGFP–scrambled shRNA (Figure 6A). To confirm this finding, we used an alternative genetic approach. Specifically, we generated mice expressing a mutated, non-phosphorylatable form of 4E-BP1 (harboring threonine-to-alanine mutations at two key mTORC1 phosphorylation sites, amino acids 37 and 46; referred to as Tg-4EBP1mt) (Tsai et al., 2015) selectively in PV neurons (Tg-4EBP1mt; Pvalb^Cre). In these mice, the non-phosphorylatable 4E-BP1 binds to and inhibits eIF4E activity specifically in PV neurons. Consistent with the results of the AAV-eIF4E-shRNAmir-mediated knockdown experiment, we observed no alleviation of mechanical hypersensitivity in Tg-4EBP1mt; Pvalb^Cre mice compared to control Tg-4EBP1mt animals (Figure 6B). Together, these results indicate that inhibiting eIF4E-dependent translation in PV neurons is sufficient to prevent the SNI-induced decrease in their intrinsic excitability, but is not sufficient to alleviate SNI-induced mechanical hypersensitivity.

Finally, we generated mice lacking 4E-BP1 in a broad population of excitatory Vglut2⁺ neurons (Eif4ebp1^fl/fl; Slc17a6^Cre) (Figure 6C), as well as in a subpopulation of excitatory neurons defined by Tac1 (Eif4ebp1^fl/fl; Tac1^Cre) (Figure 6D). These two mouse lines exhibited no changes in mechanical or heat withdrawal thresholds, suggesting that increasing eIF4E-dependent translation in excitatory neurons does not confer pain hypersensitivity.

C57BL/6 mice were purchased from Charles River Laboratories, Inc (St. Constant, Quebec, Canada) at 6–7 weeks of age. On arrival at the in-house animal facility, the mice were placed in groups of 5 animals per cage with food and water ad libitum under a 12:12-hr light/dark cycle (light period from 07:00 to 19:00 hr) with ambient temperature (22°C) and humidity maintained at 40%. Eif4ebp1^fl/fl mice (Aguilar-Valles et al., 2021) were crossed with Gad2^Cre (The Jackson Laboratory, stock #010802) to generate Eif4ebp1^fl/fl;Gad2^Cre animals, Pvalb^Cre (Jackson Laboratory, stock #008069) to generate Eif4ebp1^fl/fl;Pvalb^Cre animals, Tac1^Cre (The Jackson Laboratory, stock #021877) to generate Eif4ebp1^fl/fl;Tac1^Cre animals, and Slc17a6^Cre (The Jackson Laboratory, stock #028863) to generate Eif4ebp1^fl/fl; Slc17a6^Cre animals (corresponding Cre lines were used as controls). Tg-4EBP1mt mice (Jackson Laboratory, stock #029735) Tsai et al., 2015 were crossed with Pvalb^Cre (Jackson Laboratory, stock #008069) to generate Tg-4EBP1mt;Pvalb^Cre animals. L10a-eGFP; Tac1^Cre and L10a-eGFP; Gad2^Cre mice were generated by crossing L10a-eGFP mice (Sanz et al., 2009) with the corresponding Cre mouse lines. We also generated Eif4ebp1^fl/fl: L10a-eGFP: Pvalb^Cre, Eif4ebp1^fl/fl: L10a-eGFP: Gad2^Cre, and Eif4ebp1^fl/fl: L10a-eGFP: Tac1^Cre mice and their controls (no Eif4ebp1^fl/fl) for confirmation of 4E-BP1 deletion experiments and recording from spinal PV neurons (Figure 5D–H). Sample sizes were determined based on similar previous studies in the field. Female mice were used in Ribo-seq (Figure 1), TRAP (Figure 4), FUNCAT (Figure 3), and electrophysiology experiments. All experiments were performed and analyzed by an experimenter blind to genotypes and treatments. Both sexes (with comparable numbers of males and females) were used in all behavioral experiments. No main effects of sex were observed; therefore, data were pooled for all reported analyses. All procedures complied with the Canadian Council on Animal Care guidelines and were approved by the McGill University’s Downtown Animal Care Committee (protocol #7869).

Mice were anesthetized under 4% isoflurane for induction and 2% for maintenance. The mice were placed on a heated (36–37°C) surgical bed during the surgical procedure. The sciatic nerve was exposed by making an incision in the upper thigh and cutting through the femoris muscle. The tibial and common peroneal branches were ligated with 7.0 silk (Covidien, S-1768K), and a 2- to 4-mm section of the nerve below the ligation was removed using micro self-opening scissors and forceps, leaving the sural nerve fully intact. The muscle and skin were sutured using 6.0 Vicryl (Ethicon, J489G). All sham surgeries featured incisions to the thigh and cutting of the femoris muscle; however, the sciatic nerve was left untouched and intact. Mice returned to their cage placed on a heated (36–37°C) surface for recovery.

Mice were decapitated 2 weeks post-ASO injection. The lumbar section of the spinal cord and the DRGs were extracted and immediately homogenized in a homogenization buffer 200 mM HEPES, 50 mM NaCl, 10% glycerol, 1% Triton X-100, 1 mM EDTA, 50 mM NaF, 2 mM Na3VO4, 25 mM β-glycerophosphate, and EDTA-free complete ULTRA tablets (Roche, Indianapolis, IN) before being centrifuged at 14,000 rpm for 15 min at 4°C to obtain a supernatant. Bradford protein assay was used to measure the protein concentration of the lysates, followed by loading 30 μg of the lysates on a 12% SDS–PAGE gel, and ran at a constant current (0.03 A/gel). The gel was transferred to a nitrocellulose membrane overnight on ice at 20 mV. The membrane was blocked (5% milk or BSA in TBS-T) for 1 hr and then incubated in primary antibodies overnight at 4°C. Afterwards, the membrane was washed three times and incubated in HRP-conjugated secondary antibody at room temperature. The membrane was then washed three times, and an Enhanced Chemiluminescent (ECL) reagent was used to enhance the signal before visualizing it using a ChemiDoc Imaging System (Bio-Rad). Primary antibodies were eIF4E BD (Biosciences, Cat# 610270, 1:1000), and beta-tubulin (Cell Signaling, Cat# 2146S, 1:1000). Secondary antibodies were anti-Rabbit IgG – Horseradish Peroxidase antibody (GE Healthcare, Cat# NA9340 RRID:AB_772191), and sheep anti-Mouse IgG – Horseradish Peroxidase antibody (GE Healthcare, Cat # NA931; RRID:AB_772210).

Female mice, aged 8–10 weeks, were fed a methionine-free diet (Envigo RMS Inc, TD.110208) for 1 week, followed by an intraperitoneal injection of AHA (100 μg/g body weight, i.p., Click-IT AHA [L-azidohomoalanine], Cat No. C10102, Thermo Fisher Scientific). After 3 hr, mice were anesthetized and perfused transcardially with 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS), pH 7.4. The L4 and L5 lumbar sections of the spinal cord were extracted and kept at 4°C in PFA overnight. Tissue was then cut into 30 μm sections and washed three times (10-min increments) in PBST (0.2% Triton X-100 in PBS) before being blocked overnight in a solution composed of 10% normal goat serum, 0.5% Triton-X100, and 5% sucrose in PBS. Afterward, click chemistry was performed on the sections overnight in a click buffer, consisting of 200 μM triazole ligand, 400 μM TCEP, 2 μM fluorescent Alexa Fluor 555 alkyne (Alexa Fluor 555 Alkyne, Cat No. A20013, Thermo Fisher Scientific), and 200 μM CuSO₄ in PBS. The sections were washed followed by immunohistochemistry (described below). Quantification (25 neurons were quantified per mouse) was performed as described below for immunohistochemistry. Anisomycin (100 mg/kg, A9789, Sigma-Aldrich) was injected intraperitoneally 1 hr before AHA administration.

Mice were anesthetized and perfused with 4% PFA in PBS, pH 7.4, and spinal cords were extracted and left overnight in PFA at 4°C. Spinal cords were then transversely cut into 30 μm sections followed by three washes (10 min each) using 0.2% Triton X-100 in PBS. After washing, sections were blocked using a solution consisting of 5% normal donkey and 5% normal goat serum and 0.2% Triton X-100 in PBS for 1 hr. After blocking, tissue was incubated overnight in primary antibodies diluted in PBS. Sections were washed three times in PBS and incubated in the corresponding secondary antibody diluted in PBS for 2 hr.

Primary antibodies for immunohistochemistry were: NeuN (1:1000, mouse, Millipore, MAB377), 4E-BP1 (1:200, rabbit, #2855S, Cell Signaling and Technology Laboratories), Pax2 (1:500, goat, Novusbio, AF3364), and GFP (1:1000, chicken, Abcam, ab13970). Secondary antibodies were Goat anti-rabbit Alexa Fluor 568 (1:500, Thermo Fisher Scientific, A-11011), Donkey anti-goat Alexa Fluor 647 (Thermo Fisher Scientific, A-21447), Goat anti-mouse Alexa Fluor 488 (Thermo Fisher Scientific, A-21042), and Alexa Fluor 555 alkyne (Alexa Fluor 555 Alkyne, Cat No. A20013, Thermo Fisher Scientific). Antibodies against eIF4E (1:500, mouse, Santa Cruz, SC-271480) and GFP (1:1000, chicken, Abcam, ab13970) were used to confirm knockdown of eIF4E in PV⁺ cells. Secondary antibodies were Goat anti-mouse Alexa Fluor 488 (Thermo Fisher Scientific, A-21042) and Goat anti-chicken Alexa Fluor 647 (Thermo Fisher Scientific, A-21449).

After three washes in PBS, the sections were mounted and imaged using a Zeiss confocal microscope (LSM 880, 63X/1.40 Oil DIC f/ELYRA objective) equipped with Argon, DPSS, and HeNe lasers. Alexa Fluor 488 was excited using the Argon laser and detected with an emission window of 507–551 nm. Alexa Fluor 555/568 was excited using the DPSS 561 nm laser and detected with an emission window of 578–631 nm. Alexa Fluor 647 was excited using the HeNe 633 nm laser and detected with an emission window of 654–690 nm. Images were acquired using identical settings across experimental groups. Integrated density of AHA signal was measured within Pax2⁺ cells for inhibitory neurons, or Pax2^-/NeuN⁺ cells for excitatory neurons in the superficial dorsal horn (laminae I–III, defined using a lamina overlay based on NeuN staining) and quantified using ImageJ on maximum intensity projection images. To confirm the knockout of 4E-BP1 signal, integrated density of 4E-BP1 signal was measured in eGFP⁺ neurons using Eif4ebp1^fl/fl: L10a-eGFP: Pv^Cre, Eif4ebp1^fl/fl: L10a-eGFP: Gad2^Cre, and Eif4ebp1^fl/fl: L10a-eGFP; Tac1^Cre transgenic mice. Corresponding mouse lines without Eif4ebp1^fl/fl were used as controls. Background noise was subtracted from the final calculations in all experiments. For all image quantification experiments, three sections were analyzed per mouse, with three images acquired from each section (nine images per mouse). The integrated density of AHA, eIF4E, or 4E-BP1 signals in the somata of 25 neurons per mouse was quantified using ImageJ on maximum-intensity projection images (generated using the same parameters and number of optical sections), and the values were averaged per mouse and presented.

SNI and Sham surgeries were performed on 8- to 10-week-old female (L10a-eGFP; Tac1^Cre and L10a-eGFP; Gad2^Cre) mice 4 or 60 days prior to tissue extraction. TRAP-sequencing was performed as previously described (Megat et al., 2019; Sanz et al., 2009). Mice were decapitated and the lumbar section of the spinal cord was extracted under cold and RNase-free conditions and transferred to ice-cold dissection buffer (1X HBSS, 2.5 mM HEPES-NaOH [pH 7.4], 35 mM Glucose, 5 mM MgCl₂; 100 µg/ml cycloheximide and 0.2 mg/ml emetine were added just before use). Next, the extracted spinal cord tissues were homogenized in lysis buffer (20 mM HEPES-NaOH [pH 7.4], 12 mM MgCl₂ and 150 mM KCl in RNAse free water; 0.5 mM DTT, 1 µl/ml Protease Inhibitors Cocktail [Roche], 100 µg/ml cycloheximide, 20 µg/ml Emetine, 40 U/ml RNasin [Promega] and 2 U/ml TURBO DNase [Invitrogen] were added immediately before use) using Minilys Personal High Power Tissue Homogenizer (Bertin Technologies) on medium speed for 10 s for a total of eight times with 10-s intervals (incubation on ice) in a cold room at 4°C. A post-nuclear fraction was generated from spinal cord homogenates by centrifuging at 2000 × g for 5 min at 4°C. The supernatant was collected and NP-40 (AG Scientific) and DHPC (Avanti Polar lipids) were added at a final concentration of 1% and incubated on ice for 5 min. Afterward, post-mitochondrial fractions were generated by centrifuging samples at 15,000 × g for 10 min. A 200-µl aliquot was taken from the post-mitochondrial fraction and used as the input, and the remaining fraction was incubated overnight with protein washed G-coated Dynabeads (Invitrogen) bound to 50 µg of anti-GFP antibodies (HtzGFP-19F7 and HtzGFP-19C8 antibodies were acquired from Sloan Memorial Kettering Centre) on an end-over-end mixer. On the following day, the beads were washed four times with a high salt buffer (20 mM HEPES-NaOH [pH 7.4], 12 mM MgCl₂ and 0.35 M KCl, 1% NP-40 [AG Scientific] in RNAse free water; 100 µg/ml cycloheximide and 0.5 mM DTT were added just before use) to collect the IP fraction. After the removal of the final wash, RNA was extracted by incubating 300 µl of Trizol to the IP fraction and 600 µl to the input fraction for 10 min at room temperature. Equal amounts of ethanol were added to each of the samples before eluting the RNA using a Direct-zol RNA kit (Zymo Research) using manufacturer’s protocol. RNA yields were quantified using a NanoDrop Spectrophotometer ND-1000 (NanoDrop Technologies, Inc) and RNA quality was determined by a 2100 Bioanalyzer (Agilent Technologies).

Mice were deeply anesthetized (induced with 5% isoflurane and maintained on 2% isoflurane) and their head was secured using ear bars in a stereotaxic frame (Kopf). The head was shaved, and an incision was made to expose the skull. Aiming occurred via coordinates relative to bregma (anterior/posterior (AP): –0.5 mm, medial/lateral (ML): 1 mm, and dorsal/ventral (DV): –2.2 mm), and the skull was drilled for injection of the ASO into the lateral ventricles (i.c.v.). Five µl of eIF4E-ASO or scrambled ASO was injected using a 10-µl Hamilton microsyringe mounted on a perfusion pump. The perfusion rate was set to 0.5 µl/min, and the needle stayed in place for an additional 5 min to prevent leakage. The timeline of ASO administration, SNI, and behavioral testing is shown in Figure 2.

eIF4E-ASO targeting mouse EIf4e mRNA and Control scrambled ASO were developed and synthesized by Ionis Pharmaceuticals, see Supplementary file 2 for sequences. Each ASO consists of 5 nucleotides on the 5′ and 3′ ends of the ASO with a 2′-O-methoxyethyl modification, and a central 10-base DNA ‘gap’. ASO was diluted in PBS to a concentration of 100 mg/ml and was delivered i.c.v. at a single dose of 100 mg/kg.

Eight- to ten-week-old Pv^Cre mice were injected with AAV-CAG-DIO-eGFP-eIF4E-shRNAmir or AAV-CAG-DIO-EGFP-scrambled-shRNAmir using a minimally invasive (non-laminectomy) technique. Mice were deeply anesthetized (induced with 5% isoflurane and maintained on 2% isoflurane) and steel clamps were attached to the vertebral column. An incision was made in the skin and muscle at T12–L3, removing the muscle from the space between the T13 and L1 vertebrae. A glass electrode was inserted 250 μm into the spinal dorsal horn. A total of 500 nl of AAV-shRNA was injected over a 10-min period (David Kopf instruments, 99236B) followed by suturing of the skin using 6-0 Vicryl silk (Ethicon, J489G).

The microRNA-adapted short hairpin RNAs (shRNAmir) packaged in adeno-associated virus (AAV9-CAG-DIO-eGFP-eIF4E-shRNAmir and AAV9-CAG-DIO-eGFP-scrambled-shRNAmir) were prepared by Vector Biolabs. The validated sequence targeting mouse eIF4E was TCCAGTTGTCTTAATTTAAGTCAGTCAGTGGCCAAAACTTAAATTACTAGACAACTGGACAG, and the scrambled sequence used as a control was GAAATGTACTGCGCGTGGAGACGTTTTGGCCACTGACTGACGTCTCCACGCAGTACATTTCAG.

GraphPad Prism v.9 software was used to analyze data. All data are presented as means ± SEM. An α = 0.05 was used to determine statistical significance. Data analysis included unpaired Student’s t-test (two-tailed), one- and two-way ANOVA, and repeated measures ANOVA, followed by between-group comparisons using Tukey’s post hoc test, as appropriate.

This study did not generate new unique reagents.