Neuroimmune Consequences of eIF4E Phosphorylation on Chemotherapy-Induced Peripheral Neuropathy

Chemotherapy-induced peripheral neuropathy (CIPN) is a major dose-limiting side effect that occurs in up to 63% of patients and has no known effective treatment. A majority of studies do not effectively assess sex differences in the onset and persistence of CIPN. Here we investigated the onset of CIPN, a point of therapeutic intervention where we may limit, or even prevent the development of CIPN. We hypothesized that cap-dependent translation mechanisms are important in early CIPN development and the bi-directional crosstalk between immune cells and nociceptors plays a complementary role to CIPN establishment and sex differences observed. In this study, we used wild type and eIF4E-mutant mice of both sexes to investigate the role of cap-dependent translation and the contribution of immune cells and nociceptors in the periphery and glia in the spinal cord during paclitaxel-induced peripheral neuropathy. We found that systemically administered paclitaxel induces pain-like behaviors in both sexes, increases helper T-lymphocytes, downregulates cytotoxic T-lymphocytes, and increases mitochondrial dysfunction in dorsal root ganglia neurons; all of which is eIF4E-dependent in both sexes. We identified a robust paclitaxel-induced, eIF4E-dependent increase in spinal astrocyte immunoreactivity in males, but not females. Taken together, our data reveals that cap-dependent translation may be a key pathway that presents relevant therapeutic targets during the early phase of CIPN. By targeting the eIF4E complex, we may reduce or reverse the negative effects associated with chemotherapeutic treatments.


INTRODUCTION
Chemotherapy-induced peripheral neuropathy (CIPN) is one of the most prevalent and doselimiting complications in chemotherapy patients. It is characterized by pain and numbness or a "pins and needles" feeling in the extremities (1)(2)(3)(4). Approximately 70% of patients report the development of peripheral neuropathy during administration of chemotherapy, which leads to early treatment cessation and substandard regimens (5,6). Additionally, over 30% of patients experience persistent neuropathy for months or years after treatment cessation, making CIPN a long-term morbidity in patients that survive cancer and other diseases (7).
There is limited data on sex differences in neuroimmune effects during CIPN. Both clinical and preclinical investigations have reported conflicting observations ranging from marked sexual dimorphisms to no sex differences during CIPN (8)(9)(10)(11). There are few studies that investigate sex differences and the mechanisms involved in the development of CIPN. These findings necessitate a comprehensive and fully powered assessment of sex differences in the development of CIPN. Considering the recent interest in sex differences in chronic pain development and immune cell activation, it is purported that neuroimmune mechanisms vary between sexes. A benchmark paper in the pain field suggests sexual dimorphisms in the role of T-cells in pain initiation in females (12). Moreover, recent studies have shown that cytotoxic T-cells (CD8 + ) contribute to the resolution of CIPN (13,14). A more recent study showed that "primed" CD8 + T-cells, previously exposed to a chemotherapy agent, could prevent CIPN development when transferred to a new host (15). Based on these previous studies, we decided to assess the role of discrete T-cell subtypes: CD4 + (Th1, Th2), T regulatory (T regs ), Effector (T eff ), and activated cytotoxic T-lymphocytes (CD8 + ) in the draining lymph nodes of male and female mice after induction of CIPN.
Immune cells recruited to the dorsal root ganglion (DRG) in turn can influence the activity of nociceptors through the release of inflammatory mediators (16,17). This signaling could happen in tandem, in real-time, and this bi-directional crosstalk could be directly responsible for changes in nociceptors and immune cells, leading to changes in pain behaviors. These reciprocal interactions may contribute to maladaptive nociceptor plasticity if the homeostatic functioning of the cell-mediated immune response is dysregulated and has been shown to contribute to the development of CIPN (18)(19)(20)(21) Eukaryotic translation initiation factor 4E (eIF4E) is a cytosolic regulator that directs ribosomes to the cap structure of a subset of mRNAs to induce translation of proteins involved in inflammation and pain (22,23). Cellular signaling cascades responsive to external stimuli such as the mitogen-associated protein kinase (MAPK) pathway and the mammalian target of rapamycin (mTOR) pathway directly feed into central capdependent translation in all cell types (24,25). Regulation of eIF4E activation affects downstream signaling for production of a subset of pro-inflammatory cytokines involved in T-cell subsets activation, cellular metabolic pathways of nociceptors, and other components that regulate pain, inflammation, and behavior (26)(27)(28)(29)(30)(31)(32)(33). While we have begun to assess the role of eIF4E phosphorylation in sensory neuron populations (34), a complete breakdown of sex and sensory neuron metabolic phenotyping has yet to occur along with phenotyping of immune cell subtypes during the development of CIPN. It is known that sensory neuron metabolism is exceptionally responsive to nerve injury and inflammation (35)(36)(37). CIPN is accompanied with changes in mitochondrial bioenergetics and an energy deficit in DRG neurons (38,39). These changes can also alter ion signaling and other downstream metabolic processes to contribute to long-term development of nociceptor sensitivity and pain (40). The metabolic effect(s) paclitaxel has in late CIPN, has been assessed in sensory neurons from male rats only (38).
Therefore, we sought to elucidate the role of cap-dependent protein translation in CIPN-induced alterations that may be mediated by neuroimmune communication between activated immune cells and sensory neurons. We used eIF4E mice that harbor a point mutation at the serine 209 site on the eIF4E complex that codes alanine instead, designated as eIF4E S209A (41). This point mutation prevents phosphorylation of eIF4E, thus inhibiting its activation and consequently inhibiting capdependent protein translation (42,43). Paclitaxel is one of many chemotherapeutics used in rodent models to mimic clinical CIPN (44). Given that CIPN has an elusive etiology and the diverse array of cell types that could participate in CIPN development, we investigated whether there was a sexdependent role of eIF4E phosphorylation in DRG sensory neuron metabolism, local immune cell populations, peripheral T-cell subpopulations, and spinal microglia and astrocytes. We hypothesized that eIF4E phosphorylation regulates activation of specific macrophage and T-cell subpopulations, alters nociceptor metabolism, and glial activation during the early phase of paclitaxel-induced peripheral neuropathy.

Animals
Adult mice of both sexes (9-12 weeks old, 20-25g weight) were used for all experiments. eIF4E S209A mice were generated on a C57BL/6 background in the Sonenberg laboratory at McGill University as previously described (41). These were a gift and further bred to maintain genotypes at the University of Texas at (UT) Dallas vivarium to generate our experimental cohorts. Inhouse animals were weaned between 21 and 28 days of age and tail-clipped to verify genotypes. Both wild-type (WT) and eIF4E S209A gift mice were bred with breeder animals purchased from Jackson laboratory (WT). We used only homozygous eIF4E S 2 0 9 A mutant animals and WT littermates for experiments. All animals were housed at the UT Dallas vivarium with standard temperature (20-25°C) and a 12-hour light/dark cycle (lights on from 6 AM to 6 PM). Mice were group housed in polypropylene ventilated cages with 4-5 animals per cage and provided ad libitum access to food and water. All behavioral experiments and data analysis were carried out from 9am to 12pm during the light cycle. All experiments were performed in accordance with protocols and standard operating procedures (SOPs) approved by the University of Texas at Dallas Institutional Animal Care and Use Committee (IACUC) and the Institution Biosafety and Chemical Safety Committee. The experimenters were blinded to animal genotypes and treatment groups for all assays. A timeline of experiments and dosing regimen is shown in Figure 1.

Induction of Chemotherapy-Induced Peripheral Neuropathy (CIPN) With Administration of Paclitaxel
Paclitaxel (European pharmacopoeia (EP) Reference standard purchased from Sigma, Y0000698) was dissolved in a vehicle (1:1 ratio of Kolliphor oil (Sigma, C5135) and 100% ethanol), to prepare stock paclitaxel solution (5mg/mL stored at 4°C). Working stock was freshly prepared on injection days by diluting stock solution with sterile 1× PBS. Paclitaxel injections were administered intraperitoneally every other day for four (4mg/kg/day) injections to induce peripheral neuropathy (16mg/kg total dose per animal, Figure 1) as previously described (44,45), these studies show little to no differences of 2mg/kg versus 4mg/kg in the development of painful CIPN.

Behavioral Experiments
Baseline readings for all behavior assays were taken twice, a day apart (days -1 and -3) before administration of the 1 st paclitaxel injection on day 0. Behavioral assessment for spontaneous pain using the grimace scale, mechanical hypersensitivity, grip strength, and thermal hyperalgesia were performed on days 1, 3, 5, 7, and 9, in that order. This was important to maintain nonconfounding observations of evoked pain behavior using touch or heat stimuli after noting spontaneous pain. All behavior datasets are additionally represented as effect size. Effect size here is determined by calculating the cumulative difference between the value for each time point and the baseline value. All effect values were added to get effect size represented as an absolute number for each of the animal groups (46).

Assessment of Spontaneous Pain by Grimace Scale
Grimace is assessment of spontaneous pain using the observation of facial expression (47)(48)(49). Five parameters for facial grimace behavior were assessed: orbital tightening, ear position, cheek bulge, nose bulge, and whisker position. Each component was scored as follows: not present was given a score of "0", moderately present was given a score of "1", and obviously present was given a score of "2". The score of the five components was averaged to give a mean grimace score (MGS).

Assessment of Mechanical Hypersensitivity by von Frey Filaments
Mechanical hypersensitivity was assessed using the previously described von Frey assay (50). Von Frey acrylic chamber (11cm long, 10 cm wide, and 4.5 cm height) with mesh floor, elevated 4.25 feet off the ground, were used for this assay. Animals were acclimated to the testing room and habituated to the von Frey apparatus and test environment for at least 1 hour prior to baseline and testing measurements. After 2 baseline recordings (one day rest in between the recordings), animals were randomly FIGURE 1 | Experimental timeline with dosing regimen and terminal analysis. After habituation and baseline (BL) measures for behavior, all mice were subjected to intraperitoneal injection of paclitaxel (4 mg/kg per mouse) after BL i.e. on day 0, D2, D4, and D6 to induce peripheral neuropathy. Behavioral assays (mechanical hypersensitivity, grimace, grip strength, and thermal hypersensitivity) were measured on BL, D1, D3, D5, D7, and D9. On D10, animals were terminated and spinal cord, popliteal and inguinal lymph nodes, and DRG (L3 to T13) were collected. The lumbar portion of the spinal cord was used for immunohistochemical analyses for glia reactivity. The lymph nodes and DRG were used for flow cytometry analyses of immune cell populations. DRG neurons were plated and used for Seahorse analyses of bioenergetics. assigned to the different treatment groups. Mechanical hypersensitivity was determined by assessment of paw withdrawal threshold in response to the application of calibrated von Frey filaments (Stoelting, Illinois, USA) using the up-down method. A series of filaments with logarithmically incremental stiffness of 2.83, 3.22, 3.61, 3.84, 4.08, and 4.17 (converted to the 0.07, 0.16, 0.4, 0.6, 1, and 1.4 grams, respectively) were applied to the plantar surface of the hind paw and held for 2-3 seconds. A positive response was counted if there was a brisk withdrawal of the paw, paw shaking, and paw licking, or holding the paw in the air more than 2 seconds. The withdrawal threshold of both hind paws was measured and averaged.

Assessment of Muscle Dexterity by Grip Strength
Grip strength is a measure of functional pain in laboratory rodents and is measured with a grip strength test meter (IITC, California, USA). All animals were habituated to the apparatus for 5 minutes per animal, and were trained to grip the mesh wire surface prior to baseline or experiment. To measure, animals were gently placed on the rectangular wire meshed surface connected to the transducer in the grip strength meter and gently pulled away until they let go of the mesh. The maximal grip force was automatically measured by the transducer and recorded. Three trials were performed and averaged for both baseline measures and all experiments.

Assessment of Thermal Hyperalgesia by Hot Plate
Thermal sensitivity was assessed using the hot plate apparatus (IITC, California, USA). Animals were individually placed onto a surface maintained at 52°C, with their locomotion restricted by a Plexiglas chamber. Hind paw licking, shaking, and jumping was recorded as a positive response. Total time spent on the hot plate surface until a positive response was detected was measured and recorded as latency to response. The heating time cutoff was restricted to maximum 30 seconds to prevent tissue injury.

Immunohistochemistry
Animals were anesthetized with intraperitoneal injection of Ketamine/Xylazine (100 mg/kg) and intracardially perfused with 1x PBS solution followed by 4% paraformaldehyde (PFA). Lumbar spinal cord tissue was collected, post fixed in 4% PFA for 4 hours at 4°C and cryoprotected in 30% sucrose for 48 hours at 4°C. Tissues were embedded in optimal cutting temperature (OCT) (Fisher Scientific; 23-730-571) followed by cryosectioning into 20 mm sections that were mounted on positively charged glass slides (VWR, 48311-703). Mounted sections were treated with 1M HCL for 30 minutes followed by neutralization with 0.1 M sodium borate (pH 8.5) for 10 minutes. The slides were washed with 1x PBS and pre-incubated with 5% normal goat serum in 0.2% Triton X-100 in 1x PBS solution to block nonspecific binding. Subsequently, sections were incubated with the primary antibodies overnight at 4°C (see Table 1). The next day, sections were incubated in respective secondary antibodies and treated with DAPI solution (1:5000 dilution, Sigma-Aldrich D9542). Stained sections were covered with Gelvatol mounting medium and cover slips (51) and images were taken on a Zeiss Axio-observer 7 microscope (Jena, Germany) for quantitative analysis. Representative images were taken on an Olympus FV3000RS confocal laser scanning microscope (Shinjuku, Tokyo, Japan). Images acquired for quantitative analysis were taken using identical exposure times and representative images at identical laser power. Image analysis was done using FIJI (ImageJ) and Cell Sens version 3.1 (Olympus, Japan) software.

Flow Cytometry
Flow cytometric analysis of isolated immune cells from lymph node and dissociated DRG was performed based on previous protocols with few modifications (52). In brief, lymph nodes (popliteal and inguinal) and DRG (lumbar and thoracic, T13 to L5) were collected in ice cold sterile DPBS (Hyclone, Logan, UT). Lymph nodes were passed through a 70 micron nylon mesh with flow buffer (0.5% bovine serum albumin with 0.02% glucose) while DRG were first digested with enzymes (details described in section 2.6) and then passed through 70 micron nylon mesh. Resultant cell suspensions were centrifuged at 400 × g for 6 minutes at 4°C. Cell pellets were washed with cold 1x PBS and resuspended in pre-chilled flow buffer. Fc receptors were blocked by anti-CD16/CD32 purified antibody. Cells from lymph nodes were incubated with anti-CD3-Alexa Flour 700, anti-CD4-Fluorescein isothiocyanate, anti-CD8 Phycoerythrin, anti-CD25-eFluor 450, anti-CD44-eFluor780, and anti-CCR7-Allophycocyanin (see Table 1 for antibody details). Cells from DRG were incubated with anti-CD11b-Allophycocyanin-Cy7, anti-CD45-Brillient violet 421, anti-MHCII-Alexa Fluor 488, and anti-CD40 Phycoerythrin antibodies. Appropriate compensation controls and isotypes were used for determination and gating. Cells were washed twice in ice cold flow buffer and resuspended in ice cold flow buffer. T helper T-cells were initially identified with gating CD3 and CD4 cell surface markers, whereas cytotoxic T-cells were identified by gating for CD3 and CD8. These populations were further gated to identify activation and polarization (T h1 vs. T h2 ) with CD44 and CD25. CCR7 was also used to identify T-effector cells. Tissue activated macrophages were identified from CD45 separated cells with cell surface markers CD11b, MHC-II, and CD40. Stained samples were analyzed using a Becton-Dickinson Fortessa analyzer (Red Oaks, CA) and data were analyzed using FlowJo software (De Novo Software, Los Angeles, CA). For a complete list of antibodies used, please refer Table 1.

Primary Dorsal Root Ganglia (DRG) Culture and Cellular Energetics
On day 10, animals were deeply anesthetized with isoflurane and decapitated. DRG were dissected bilaterally starting from T13 to L5 and transferred to pre-chilled 1x PBS containing 1% penicillin/streptomycin (ThermoFisher Scientific, 15070063). Samples were centrifuged at 400 x g for 4 minutes. Supernatants were removed and DRG were treated with Collagenase A (Sigma, 10103586001) and incubated in water bath at 37°C for 20 minutes, followed with treatment of Collagenase D (Sigma, 1188866001) for another 20 minutes. Cells were centrifuged at 400 x g and pellet was resuspended in Enzyme T (soybean trypsin inhibitor made up in 1 part bovine serum albumin and 1 part DMEM/F12 media) to stop the enzymatic reaction. Digested tissues were triturated approximately 30 times using a 1 ml pipette tip and passed through the 70 micron nylon mesh, with a subsequent wash with DMEM/F12 media (supplemented with 10% Fetal bovine serum and 1% penicillin/streptomycin). Resultant suspension was centrifuged at 400 x g for 5 minutes and resuspended in DMEM/F12 medium. The number of cells was counted using a hemocytometer with trypan blue dye exclusion. For bioenergetic profile, 70,000 cells were seeded directly in respective XFp Cell Culture Miniplates (Agilent, 103025-100) coated with poly-D-Lysine (Millipore Sigma, P0899). Next day, XF Assay media (Seahorse Bioscience) supplemented with 2 mM GlutaMAX, 5 mM glucose (Sigma), 1 mM sodium pyruvate (Fisher Scientific, Loughborough, United Kingdom), and pH adjusted to pH 7.4, was warmed to 37°C. Cells were incubated with this Seahorse media for 1 hour at 37°C, 0% CO2 incubator. The media was replaced with the Seahorse XF calibrant solution for another hour before running the Mito Stress according to manufacturer's recommended protocol (53).

Statistical Analysis
Data were analyzed using GraphPad Prism software (version 8.4) and expressed as mean ± standard error of the mean (SEM). For behavior, flow cytometry, and Seahorse datasets, two-way ANOVAs were performed followed by Tukey's post-hoc for multiple comparisons. For Seahorse divided into phases of mitochondrial respiration, one-way ANOVA was performed with Sidak's multiple comparison test. For IHC datasets, two-way ANOVA followed by Bonferroni's post-hoc was performed for glia reactivity and Sidak's post-hoc was used for ATF3. A p value of ≤ 0.05 was considered significant.  Table 2). Interestingly, eIF4E S209A females treated with paclitaxel had a significantly reduced latency of response to heat stimuli compared to their WT counterparts ( Figure 2P and Table 2). Taken together, our data suggests that eIF4E is important for mechanisms involved in CIPN development early in both sexes.

Paclitaxel Dysregulates CD4 + and CD8 + T-Cell Subpopulations in an eIF4E-Dependent Manner
An obvious early contribution of adaptive immune cells in the development of paclitaxel-induced peripheral neuropathy has not been shown till date. Given that the paclitaxel was  (P) Thermal hypersensitivity in females shown as effect size, *p 0.0453. All data are presented as mean ± standard error of the mean, males: n = 6 mice per group; females: n = 12 mice per group. Two-way ANOVA was performed with Tukey's post-hoc for multiple comparisons. For all line graphs, asterisks indicate significant differences between vehicle VS paclitaxel-treated for each genotype at indicated time points. "a" (p < 0.0001), "b" (p < 0.001), "c" (p < 0.01), and "d" (p < 0.05) depict significant difference between paclitaxel-treated WT and eIF4E S209A mice at indicated time points.
administered peripherally, T-cells are activated in the lymph nodes, and neuropathy develops in the limbs, we chose to study subpopulations of T-cells from the draining (popliteal and inguinal) lymph nodes that are CD4 + (T helper ) CD8 + , CD25 +/and CD44 +/-(activated T-cells), CCR7 + (effector T-cells) ( Figures 3A-F). We found a significant increase in CD4 + Tcells in WT female mice injected with paclitaxel, but not males ( Figures 3A, B, E and Table 3). The CD4 + subset was further analyzed for CCR7, the marker for lymphocyte homing to lymph nodes as well as T-cell priming (55,56). The CD4+CD25-CD44+ subset was significantly reduced in eIF4E S209A females treated with paclitaxel compared to WT ( Figure 3G, Table 3). The CD4 + CCR7 + subpopulation was significantly higher in eIF4E S209A mice of both sexes independent of paclitaxel treatment ( Figures 3C, F and Table 3). This suggests that inhibiting cap-dependent translation upregulates T-cell priming and activation of T cells. We found significant downregulation of CD8 + T-cells in female eIF4E S209A mice compared to WT, irrespective of treatment ( Figures 3G-L, Table 3). This population was unchanged in male mice ( Figure 3H). CD8 + CCR7 + T-cells, i.e. activated memory Tcells were higher in eIF4E S209A mice ( Figures 3I, L and Table  3) regardless of sex or treatment. CD8 + CD44 + CD25 -T-cells were increased in paclitaxel-treated eIF4E S209A male mice compared to WT males treated with paclitaxel ( Figure 3J and Table 3). There were no differences in female mice for this subpopulation ( Figure 3M and Table 3). Overall, our data shows that paclitaxel induces activation and/or expansion of T-helper, memory T, and effector T-cells, which is modulated in an eIF4Edependent manner.

Paclitaxel Induces Increased Infiltration of Activated Myeloid Cells in DRG in Absence of eIF4E Phosphorylation
We further investigated if peripheral neuropathy leads to the recruitment of macrophages into the DRG. No studies to date have shown the recruitment of the macrophages and the activation state of macrophages in DRG after CIPN. We looked for the infiltration and activation of macrophages using the CD11b + -CD45 + and MHCII + -CD40 + cells respectively in male and female mice. We identified macrophages within DRG by gating for CD11b + and CD45 + cells. This population was further gated for MHCII + and CD40 + to assess activated macrophage subsets ( Figure 4A). We found that CD11b + -CD45 + -CD40 + -MHC-IIand CD11b + -CD45 + -CD40 --MHC-II + populations were unchanged in both sexes with paclitaxel or lack of eIF4E ( Figures 4B, C and Table 3), indicative of no effect on dendritic cells or tissue-resident macrophages. We found a significant increase in CD11b + -CD45 + -CD40 + -MHC-II + cells in male and female eIF4E S209A mice treated with paclitaxel compared to vehicle. Additionally, in female mice, this subpopulation was significantly higher in paclitaxel-treated Two-way ANOVA was performed followed by Tukey's post-hoc for multiple comparisons between genotypes and treatments for each sex. p ≤ 0.05 was considered significant (bolded).  <0.0001 Two-way ANOVA was performed followed by Tukey's post-hoc for multiple comparisons between genotypes and treatments for each sex. p ≤ 0.05 was considered significant (bolded). eIF4E S209A animals compared to paclitaxel-treated WT ( Figure 4C and Table 3). This indicates that antigenpresenting cells (APCs) are upregulated with paclitaxel treatment in absence of eIF4E. Interestingly, female eIF4E S209A vehicle-treated mice had a significantly lower proportion of activated APCs compared to WT ( Table 3). This shows that cap-dependent translation may be facilitating the recruitment of activated myeloid cells as a result of paclitaxel treatment, which in turn could be activating or expanding the Tcell populations.

Paclitaxel Alters Mitochondrial Function of DRG Neurons in an eIF4E-Dependent Manner
We investigated mitochondrial function in DRG neurons to examine if paclitaxel changed the bioenergetic profiles (Figures 5A-Q and Table 4) in an eIF4E-dependent manner. The overall oxygen consumption rate (OCR) was lower in paclitaxel-treated eIF4E S209A males and females compared to paclitaxel-treated WT ( Figures 5B, C, J, K and Table 4) whereas the overall extracellular acidification rate (ECAR) was higher with paclitaxel treatment in both sexes and lower in paclitaxel-treated eIF4E S209A males compared to paclitaxel-treated WT males ( Figures 5H, I, P, Q and Table 4). This suggests that lack of eIF4E causes a shift in mitochondrial respiration with paclitaxel treatment. We divided the different phases of the OCR bioenergetic profile into basal respiration (before oligomycin treatment), ATP production (after oligomycin treatment and be fo re FCC P -2-[2-[4-(trifl uo r om eth oxy ) p h eny l] hydrazinylidene]-propanedinitrile treatment), maximal respiration (after FCCP and before antimycin A and rotenone), and non-mitochondrial respiration or reserve capacity ( Figure  5A). For males, we found a significant increase in the basal respiration and non-mitochondrial respiration after paclitaxel treatment in WT, but not eIF4E S209A mice ( Figures 5D, G and Table 4), but no change in ATP turnover or maximal respiration ( Figures 5E, F and Table 4). For female mice, we found that basal respiration and non-mitochondrial respiration was increased after paclitaxel treatment for both genotypes ( Figures 5L, O and Table 4). ATP turnover and maximal respiration was also higher in WT females treated with paclitaxel compared to vehicle and in eIF4E S209A females compared to paclitaxel-treated WT ( Figures 5M, N and Table 4). Taken together, our data shows that sensory neurons in the DRG undergo an excited mitochondrial shift with paclitaxel treatment in an eIF4E-dependent manner in both sexes. Paclitaxel increased all aspects of mitochondrial respiration in females but did not affect ATP turnover and maximal respiration in males, which indicates divergent effects of paclitaxel in eIF4E absence in both sexes. All data are presented as mean ± standard error of the mean, n = 3-4 mice each for male/female WT vehicle-treated and WT paclitaxel groups, n = 4-7 mice each for male/female eIF4E S209A vehicle-treated and paclitaxel. For analyzing AUC data, two-way ANOVA was performed with Tukey's post-hoc for multiple comparisons. For analyzing each phase of mitochondrial respiration, an ordinary one-way ANOVA was used with Sidak's post-hoc for multiple comparisons.

Sensory Neuron Damage by Paclitaxel Is Similar for Both Sexes and Genotypes
Previous studies have shown no association between sickness behavior or change in the body weight of the animal with 16mg/ Kg dose of paclitaxel. Moreover, other researchers have used even higher doses of paclitaxel to induce peripheral neuropathy (57,58). We assessed number of DRG neurons positive for activating transcription factor 3 (ATF3). The number of ATF3 + neurons were significantly higher in the DRG of paclitaxel-treated animals indicating damage; however, there were no sex-based or genotype differences observed ( Figure 6 and Table 5). The increase is similar to previously published literature, albeit at a later time point (59,60), thus we conclude that this higher dose does not comparatively alter DRG neurons at day 10 after our regimen. Two-way ANOVA was performed on AUC values for OCR and ECAR followed by Tukey's post-hoc for multiple comparisons between genotypes and treatments for each sex. For breakdown data of mitochondrial respiration, ordinary one-way ANOVAs were used to analyze mean values of four groups (WT vehicle, WT paclitaxel, eIF4E S209A vehicle, and eIF4ES 209A paclitaxel), followed by Sidak's post-hoc for multiple comparisons between genotypes and treatments for each sex. p ≤ 0.05 was considered significant (bolded). OCR, oxygen consumption rate; ECAR, extracellular acidification rate.

Spinal Glia Reactivity Is Altered in a Sex-Dependent Manner With Paclitaxel Treatment
Previous data has shown activation of microglia and astrocytes at different phases of CIPN (61)(62)(63). Here, we investigated whether the eIF4E complex plays a role in the activation of microglia and astrocytes in the spinal cord. We found a significant increase in immunoreactivity of microglia and astrocytes in both male and female WT mice (Figure 7). Iba1 fluorescence was significantly higher in paclitaxel-treated WT mice of both sexes compared to   Figure 7C). For astrocytes, females lacking capdependent translation had increased GFAP fluorescence after paclitaxel compared to vehicle (Figures 7D, F, G, and H, and Table 5), which was different from the males where the GFAP fluorescence was at the same level as seen for the WT vehicle group (Figures 7D, E). This data shows that paclitaxel may mediate activation of astrocytes via an eIF4E-independent pathway in females but modulates microglia activation via eIF4E in both sexes and astrocyte activation in males.

DISCUSSION
The current study begins to uncover the importance of eIF4E phosphorylation in the development of CIPN and associated neuroimmune consequences in male and female mice. In our study, we found an eIF4E-dependent increase of mechanical and thermal hypersensitivity in both sexes, upregulated T-helper and reduced cytotoxic T-lymphocytes in draining lymph nodes in female but not male mice, whereas activated macrophages in DRG were significantly increased in eIF4E but not WT mice of both sexes. In addition to this, DRG neuronal mitochondrial function was significantly higher in WT mice but not in eIF4E mice of both sexes. Microglia were significantly activated in an eIF4E-dependent manner with paclitaxel treatment in WT mice of both sexes; however, astrocytes were activated in an eIF4Edependent manner with paclitaxel treatment only in male mice. Female eIF4E mice showed significant astrocyte activation.
Our study was designed and powered with the intent to appropriately assess sex differences in neuroimmune interactions during CIPN development-a question left unanswered in previous literature. We found an eIF4Edependent increase in spontaneous pain as well as mechanical and thermal hypersensitivity after paclitaxel. This suggests that the pain-like behavior changes due to paclitaxel are mediated via the central cap-dependent translation pathway that is common across diverse cell types, with more pronounced effects in female mice.
The cell bodies of primary afferent neurons reside in the dorsal root ganglia. The persistent signals from nerve terminals activate the machinery and signaling cascades in the nucleus and cytoplasm of the neuronal soma (64)(65)(66). Hyper-excited nociceptor neurons recruit immune cells by secreting chemokines, neurotransmitters, and/or other neuropeptides (16,67). It has been shown that paclitaxel treatment recruits macrophages into the DRG (20). Our findings corroborate this finding at an earlier time point and add to this knowledge by characterizing the macrophage population as CD45 + CD11b + CD40 + MHC-II + i.e. activated APCs. It has been shown that such infiltration may be beneficial to recruit Th2 antiinflammatory cells at later time points (62,68). We have presented evidence that paclitaxel injection leads to a significant increase in the activated APC population in eIF4E males and females, but not in WT mice. Thus, a lack of eIF4E phosphorylation or cap-dependent translation exacerbates the immune cell response and DRG infiltration during CIPN development.
A prior study performed only in male mice showed an increase in CD4 + T-cells at paclitaxel post-injection day 7 (14). We found a similar increase of CD4 + T-cells in female WT mice with paclitaxel treatment. Additionally, a novel finding was that CD8 + T-cells were decreased in female but not in male WT mice during early phase of CIPN development. This indicates sex differences in T-cell activation in the lymph nodes, which would affect the subsets of T-cells that would infiltrate other tissues such as DRG (69). A previously published study showed increased numbers of CD3 + , helper T-cells, and cytotoxic T-lymphocytes after paclitaxel on day 7 in mice DRG (70). Our dataset adds niche T-cell population characterization to this previously published literature. T-cell subset activation in the lymph nodes is highly dependent upon the antigen-presenting cells naïve T-cells encounter (71). Once activated, these subsets follow their divergent fates of redistribution to tissues in need of immune response. Thus, characterizing different subsets of T-cells in the lymph nodes allows us to determine what populations we can expect in interfaces where neuroimmune communication can occur, such as the meninges or DRG. T-cells are also important in the transition from acute to chronic pain (15). By showing alterations in specialized T-cell subsets, our study raises important questions about the implications for establishment of CIPN and its resolution. Animals lacking eIF4E had higher proportion of memory T-cells in both sexes. There were no changes in the effector T-cell population for the lymph nodes. These alterations in the activation and proportion of T-cell subsets and APCs suggest that although affective pain behavior is similar in male and female mice after development of CIPN, the mechanism driving the pain phenotypes in both sexes involves varied contribution from innate and adaptive immune cells.
Basal respiration and non-mitochondrial respiration were higher in male DRG neurons treated with paclitaxel compared to vehicle, which is contrary to a previous report for rats (38). We have further shown that the lack of eIF4E does not change this trend, even though oxygen consumption is lower, suggesting that paclitaxel may induce mitochondrial function independent of cap-dependent translation regulation. In females, while the effects on basal respiration follow the same trend as the males, paclitaxel seems to mediate an increase in non-mitochondrial respiration independent of eIF4E. Female DRG neurons also had significantly higher ATP turnover and maximal respiration with paclitaxel treatment in WT. Considering that eIF4E female mice show higher mechanical hypersensitivity with paclitaxel and that their DRG neuron ATP turnover and maximal respiration is higher, this may be indicative of the severity of the pain phenotype. Whereas for males, the increased basal respiration may reflect the pain phenotype in general. To our knowledge, this is the first study to show sex differences in mitochondrial bioenergetics in the premise of cap-dependent translation for paclitaxel-induced peripheral neuropathy. These differences seen early during CIPN development, may be important for identifying strategies to prevent its establishment in both sexes. Thus, it is necessary to understand the particular mitochondrial pathways downstream of eIF4E to identify potential targets to prevent CIPN. Alternatively, therapies that address restoration of normal mitochondrial function and morphology are also promising for either sex.
Previously published data has shown that paclitaxel leads to spinal activation of astrocytes but not microglia at different phases of paclitaxel-induced peripheral neuropathy in mouse and rat models (72,73). In contrast, we found that both microglia and astrocytes are significantly activated after paclitaxel injection in male mice, which is eIF4E dependent. However, the astrocyte reactivity increase in female mice was independent of eIF4E, suggesting that paclitaxel may have alternate mechanisms of astrocyte activation in females. Activated glia are known to secrete a number of proinflammatory mediators that act directly on CNS neurons, which can sensitize them if subjected to constant stimulation over extended periods (64). This has implications for the establishment of pain phenotypes post paclitaxel treatment. It has also been shown that intraperitoneal paclitaxel can activate astrocytes and mediate mechanical allodynia in the 1 st hour, possibly by miniscule amounts crossing the blood-brain barrier (62). Thus, the glia activation seen 10 days post our paclitaxel regimen may be important for onset as well as establishment of CIPN through cap-dependent translation in males but via other pathways in females.
Regulation of cap-dependent translation happens via multiple different pathways. This is partly related to the expression of cellspecific surface receptors such as TrkA/B, gp130, mGLuR1/5, and insulin receptors and this influences which pathway is activated in response to which stimuli (74)(75)(76). In behavioral experiments, we assessed the role of eIf4E for regulation of brain behaviors i.e the interaction of all cells in the body after insult, but in the molecular experiments, we target specific populations of cells to understand how they individually change as a result of eIF4E manipulation and paclitaxel treatment. Our study provides novel insight into how cap-dependent translation dysregulation of immune or neuronal cells may be important for sex-divergent mechanisms of early CIPN. But comprehensive future studies would still be required to further deduce the cellspecific responses after paclitaxel treatment in the absence of cap-dependent translation.
Overall, our data suggest that although pain-like behavior during the early development of the CIPN are similar, the mechanisms and cell types involved in engendering these behaviors are different in males and females. The differences in immune cell populations and activation of glia indicate that separate upstream pathways may regulate cap-dependent translation in both sexes and thus lead to changed downstream outcomes. The central player i.e. the eIF4E complex can thus be a valuable target for preventing establishment of CIPN and limiting associated pain.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The animal study was reviewed and approved by University of Texas at Dallas Institute for Animal Care and Use Committee (IACUC).

AUTHOR CONTRIBUTIONS
Methodology, NA. Data analysis and curation, NA, PM, TS-P, and HJ. Writing manuscript and drawing figures, NA, PM, TS-P, and HJ. MB conceptualized the study and participated and supervised in all aspects of the study and manuscript preparation. All authors contributed to the article and approved the submitted version.