Sensory Neuron TLR4 mediates the development of nerve-injury induced mechanical hypersensitivity in female mice

Abstract

Recent studies have brought to light the necessity to discern sex-specific differences in various pain states and different cell-types that mediate these differences. These studies have uncovered the role of neuroimmune interactions to mediate pain states in a sex-specific fashion. While investigating immune function in pain development, we discovered that females utilize immune components of sensory neurons to mediate neuropathic pain development. We utilized two novel transgenic mouse models that either restore expression of toll-like receptor (TLR) 4 in Na_v1.8 nociceptors on a TLR4-null background (TLR4^LoxTB) or remove TLR4 specifically from Na_v1.8 nociceptors (TLR4^fl/fl). After spared nerve injury (SNI), a model of neuropathic injury, we observed a robust female-specific onset of mechanical hypersensitivity in our transgenic animals. Female Na_v1.8-TLR4^fl/fl knockout animals were less mechanically sensitive than cre-negative TLR4^fl/fl littermates. Conversely, female Na_v1.8-TLR4^LoxTB reactivated animals were as mechanically sensitive as their wild-type counterparts. These sex and cell-specific effects were not recapitulated in male animals of either strain. Additionally, we find the danger associated molecular pattern, high mobility group box-1 (HGMB1), a potent TLR4 agonist, localization and ATF3 expression in females is dependent on TLR4 expression in dorsal root ganglia (DRG) populations following SNI. These experiments provide novel evidence toward sensory neuron specific modulation of pain in a sex-dependent manner.

1. Introduction

The CDC lists chronic pain as a leading cause of long-term disability, with current treatments remaining ineffective (Dahlhamer et al., 2018). Grim side effects have tallied over 600,000 opioid-related deaths, with 64% of these cases linked to chronic pain patients since 2010 (Seth et al., 2018). These issues become more complex with females being more susceptible to certain forms of chronic pain (Joseph et al., 2003). Neuropathic pain results from disease or trauma to the nervous system, with injury-induced neuropathic pain estimated to occur in 20–50% of patients following routine operations (Kehlet et al., 2006). Sufferers often report a significantly reduced quality of life and the prevalence of chronic pain continues to rise as the number of effective therapeutics remain limited (McDermott et al., 2006). Recent literature highlighting sexual dimorphisms in the onset and chronicity of pain states have begun to alter the way researchers and clinicians approach mechanisms in pain and potential therapeutics. Therefore, a dire need exists to bolster research efforts to identify crucial therapeutic time windows as well as novel and effective alternatives for the treatment and abatement of chronic pain. Here we hope to Identify early time-points in the development of neuropathic pain that could represent a viable therapeutic window.

Toll-like receptors (TLRs) are pattern recognition receptors (PRRs) that initiate responses to infection and tissue injury that lead to inflammation and nocifensive behaviors (Hanamsagar et al., 2012; Nicotra et al., 2012; Xiang et al., 2015). TLR4 plays a pivotal role in chronic pain through the activation of pro-inflammatory cytokine production in non-neuronal cells. Pro-inflammatory mediators have been shown to directly sensitize neurons in the dorsal root ganglia (DRG) and exacerbate maladaptive plasticity (Ma et al., 2006). The dynamics of TLR4-mediated activation via microglia is well established where pharmacological inhibition of spinal TLR4 attenuates mechanical hypersensitivity in male, but not female mice (Sorge et al., 2011; Woller et al., 2016). Moreover, the role for T-cells regulating development of neuropathic pain in females has been alluded to, however, recent evidence leads us to believe that nociceptive sensory neurons are the primary facilitators of chronic pain etiology (Lopes et al., 2017). Differential expression of genes involved in nociceptive pathways have been identified with females expressing upregulated genetic markers that correlate with increased inflammation, synaptic transmission, and extracellular matrix (ECM) reorganization (Mecklenburg et al., 2020; Wangzhou et al., 2021). In the periphery, direct TLR4 stimulation is involved in dental pain via sensitization of trigeminal ganglion neurons (Diogenes et al., 2011), while TLR4 in dorsal root ganglia (DRG) cultures are implicated in chemotherapy-induced neuropathy (Li et al., 2015a). Recent studies suggest that TLR4 on peripheral monocytes may play a role in female-specific nociceptive states (Huck et al., 2021), while others refute this claim (Peng et al., 2016; Yu et al., 2020a). Due to its expression pattern in immune, mesenchymal, and neuronal cell populations, peripheral TLR4 signaling remains an elusive topic and important in both sexes. Importantly, the sex-specific role of TLR4 signaling in DRG neurons is unclear (Allette et al., 2014; Feldman et al., 2012; Hutchinson et al., 2008; Klawitter et al., 2014; Rudjito et al., 2021). In this study, we ask how endogenous TLR4 activation on peripheral nociceptors mediates the development of neuropathic pain.

Tissue injury upregulates danger-associated molecular patterns (DAMPs) (Man et al., 2015; Wan et al., 2016). High-mobility group box 1 (HMGB1) is a DAMP capable of initiating inflammatory cascades via TLR4 (Yamasoba et al., 2016). HMGB1 is a nuclear-bound redox sensitive cytokine, that is released from the cell upon activation, during tissue damage and changes to oxidative states; disulfide HMGB1 (dsHMGB1), is the predominant species that has affinity for TLR4 (Agalave and Svensson, 2015; Yamasoba et al., 2016). Intrathecal administration of HMGB1 has been shown to produce robust mechanical allodynia (Agalave et al., 2021b; O’Connor et al., 2003). Elevated levels of circulating HMGB1 promote activation and priming states of immune cells; however, its action on neurons that express TLR4 to elicit pain behavior has not been assessed (Bestall et al., 2018). The notion of neuronal TLR signaling remains contentious as emerging evidence suggests DRG macrophages are involved in communication with sensory neurons following nerve injury in both sexes (Yu et al., 2020b). Interestingly, this does not relegate the fact that sensory neurons are activated and directly interact with immune cells (reverse communication). While the magnitude of neuropathic pain may not differ between sexes, the mechanisms which drive its development are clearly different. The timing and mode of communication between immune cells and neurons may help elucidate the mechanisms by which males and females differ.

Utilizing recently developed and validated transgenic models, we were able to specifically remove or reactivate TLR4 on Na_v1.8⁺ nociceptors in a cre dependent fashion (Jia et al., 2021; Jia et al., 2014). We identified a distinct neuronal-mediated pathway in female mice which contributes to the early onset of peripheral nerve injury-induced mechanical hypersensitivity in addition to uncovering a direct mechanism of neuronal sensitization. We demonstrated that TLR4, expressed specifically on Na_v1.8⁺ peripheral nociceptors, is responsible for mechanical hypersensitivity during the onset of peripheral nerve injury in female, but not male mice. Additionally, HMGB1 localization and autocrine signaling via these nociceptors, as well as paracrine action on large diameter sensory neurons in the DRG, contribute to pain sensitization in female mice. We believe TLR4 activation on Na_v1.8⁺ nociceptors is one mechanism of pain sensitization unique to females.

2. Material and Methods

2.1. Animals

All animal experiments were carried out in accordance with protocols approved by the Institutional Animal Care and Use Committee of the University of Texas at Dallas. Mice were housed (4–5 per cage) in a temperature-controlled facility and maintained on a 12-hour light/dark cycle (lights on: 7am/ lights off 7pm). Mice had ad-libitum access to food and water and were eight to twelve-weeks-old during the experiments (male, 25–30 g: female 20–25 g). Transgenic mice expressing Cre recombinase under control of the Scn10a (Na_v1.8) promoter were obtained initially from Professor John Wood (University College London), but are commercially available from Ifrafrontier (EMMA ID: 04582). Characterization of these mice showed that heterozygous cre animals have no pain phenotype and normal electrophysiological properties (Stirling et al., 2005). Genetically modified TLR4 floxed animals (has loxP sites flanking exon 2 and 3; TLR4^fl/fl) (Jia et al., 2014) and TLR4 null-reactivatable animals (has a transcriptional blocker inserted into the TLR4 gene in between exon 2 and 3; TLR4^LoxTB) (Jia et al., 2021)were a gift from Joel K. Elmquist, (UT Southwestern Medical Center) were bred with Na_v1.8^cre animals in-house (Na_v1.8cre × TLR4^fl/fl, TLR4^fl/fl, Na_v1.8cre × TLR4^LoxTB, TLR4^LoxTB) and used for all behavioral and biochemical assays (Jia et al., 2021; Jia et al., 2014). All cell-specific knockout (Na_v1.8^TLR4fl/fl) or cell-specific reactivated (Na_v1.8^TLR4LoxTB) mice are heterozygous for the Na_v1.8-cre (Na_v1.8^cre) and are homozygous for the TLR4^fl/fl or TLR4^LoxTB gene, respectively. Phenotypically normal littermates lack Na_v1.8^cre and are homozygous for the TLR4^fl/fl gene. Whole-body null or knockout mice (TLR4^TB/TB) lack Na_v1.8^cre and are homozygous for the TLR4^LoxTB gene. We purchased ROSA26^LSLtdtomato (tdtomato) animals from Jackson lab (stock no. 007909) and crossed them with Nav1.8^cre animals (Nav1.8^tdT+) for reporter, IHC, and flow cytometry experiments. All strains were backcrossed at least eight generations to maintain C57BL/6J genetic background with animals from Jackson Lab (stock no. 000664). We used in-house bred C57BL/6J animals as wild-type (WT) controls.

2.2. Surgical Procedures

The spared nerve injury (SNI) model of neuropathic pain was used. Mice were anesthetized under isoflurane anesthesia (1.0–2.5 %). The ipsilateral thigh was shaved and cleaned with betadine (Dynarex, 1425) and 70% ethanol (Decon Labs, 2701). The skin and muscle of the left thigh were incised with a #11 scalpel blade (Thermo, 22–079-691) and the sciatic nerve along with its three branches (common peroneal, tibial, and sural) was exposed. A tight ligature using a 5–0 silk suture (VWR, MV-682) was placed around the proximal tibial and common peroneal branches, after which the nerve distal to the ligature was transected, taking care to not stretch or damage the sural nerve (Decosterd and Woolf, 2000). Sham surgeries were done identically to the SNI surgery; however, no portion of the sciatic nerve was ligated or transected. The skin was closed using an auto clip (Fine Science Tools, 12022–09) and mice were then given a subcutaneous injection of 5 mg/mL Gentamicin (Sigma, G1272) as a preventative antibiotic and returned to their home cages to recover (Decosterd and Woolf, 2000). Mice were monitored daily for the duration of the experiment. Mechanical hypersensitivity and cold allodynia were then assessed on postoperative days 1, 3, 5 and 7. Baseline values were taken 24 hours prior to surgery.

2.3. Behavioral Testing

To measure mechanical hypersensitivity, mice were individually placed on an elevated wire grid inside acrylic behavior racks and allowed to habituate for approximately 2 hours. The ipsilateral hind paw was then stimulated with von Frey filaments (Stoelting, 58011) using the up-down experimental paradigm (Chaplan et al., 1994). To assess cold allodynia in our SNI model, mice were individually placed on the same elevated wire grid and behavior racks and allowed to habituate for approximately 2 hours before testing. Approximately 100 μL of biology grade acetone (Fisher, AI6P-4) was then applied to the ipsilateral hind paw using a 1 mL syringe (VWR, 309659) attached to a 25 G needle (VWR, 305125). Latency of behavioral response over a 60 second period was measured (Choi et al., 1994). Behavior racks were cleaned with a 1:3 ratio of a plant-based deodorant-free cleaner (Seventh Generation™, 22719BK-5) to eliminate odor cues between each reading, baseline, and experiment. Baseline values were taken 24 hours prior to performing surgery. Mechanical hypersensitivity and cold allodynia were then measured on postoperative days 1, 3, 5, and 7. Mechanical measures were always taken before thermal. All behavioral testing was done between 10 a.m. and 2:00 p.m. All behavior data sets are additionally represented as effect sizes. It is determined by calculating the cumulative difference between the value for each time point and the baseline value (Agalave et al., 2021a; Hassler et al., 2019). Behavioral experiments were performed by T.A.S., L.R.B., and M.D.B. Experimenters were blinded to genotype, surgical condition, or both.

2.4. Immunohistochemistry

Three days post SNI, mice were anesthetized using 100% isoflurane and were subsequently euthanized via decapitation. Dorsal root ganglia (L3–5) were extracted and post-fixed in 4% paraformaldehyde made in 1 × phosphate buffered saline (PBS) (Sigma, 158127) for 4 hours and then cryoprotected for 48 hours in a 30% sucrose (Sigma, S0389) solution made in 1 × (PBS). Frozen tissues (DRGs: 16 μm) were then transversally sectioned on a cryostat and mounted onto SuperFrost Plus (Fisher, 12–544-7) charged microscope slides. Slides were then allowed to dry for 1 hour before being placed in the −80° C freezer overnight. Slides were then rehydrated in 1 × PBS for 5 minutes before being incubated in a blocking/permeabilization solution consisting of: 1 × PBS, 2% heat-inactivated normal goat serum (Sigma, G9023), 1% bovine serum albumin (Sigma, A9576–50ML) 0.1% Triton (Sigma, X100), 0.05% Tween-20 (Sigma, P1379) and 0.05% sodium azide (Ricca Chemical, R7144800). Slides were then incubated overnight at 4° C in a primary antibody cocktail diluted in blocking/permeabilization solution. The following day slides were washed three times for 5 minutes each in 1 × PBS + 0.05% Tween-20 and incubated for 2 hours at room temperature in their respective secondary antibodies diluted in blocking/permeabilization solution. Slides were then washed three times for 5 minutes each in 1 × PBS + 0.05% Tween-20 and mounted using Prolong gold (Thermo, P36974) mounting media and a 1.5 mm glass coverslip (Thermo, 08–774-384). Slides were stored at 4° C until imaging. Images used for analysis and quantification were taken on a Zeiss Axiobserver 7 epifluorescent Microscope at 20x. Representative images were taken at 20x (ATF3 and tdTomato images) or 40x (HMGB1 images) using an Olympus FluoView 3000 RS confocal microscope. Analysis of images was done using ImageJ Version 1.48 (National Institutes of Health, Bethesda, MD) for Mac OS X (Apple). Immunohistochemistry (IHC) experiments and analyses were performed by T.A.S., and L.R.B. Experimenters were blinded to sex, genotype, and surgical condition. For a complete list of antibodies used refer to Table 1.

Table 1.

Antibodies used in IHC and flow experiments.

Antibody	Company	Catalog number	Working dilution
Antibodies used for IHC
Anti-Nav1.8 sodium channel	Neuromab	75-166	1:500
Anti-Neurofilament 200	Millipore Sigma	MAB5266	1:1000
Anti-HMGB1	Abcam	AB79823	1:500
Anti-ATF3	Abcam	AB207434	1:500
Goat anti-mouse Alexa Fluor 488	Invitrogen	A21131	1:1000
Goat anti-mouse Alexa Fluor 647	Invitrogen	A21240	1:1000
Goat anti-rabbit Alexa Fluor 568	Invitrogen	A11011	1:1000
Antibodies used for flow
Anti-CD45 Super Bright 645 conjugate	eBioscience	64-0451-82	1:200
Anti-TLR4/MD-2 Complex APC conjugate	eBioscience	17-9924-82	1:200
Anti-CD16/32	eBioscience	16016185	1:1000
Anti-NeuN Alexa Fluor 488 conjugate	Millipore Sigma	MAB377X	1:200

2.5. Flow Cytometry

Flow cytometric analysis of isolated DRGs was performed based on previous protocols with some modifications (Chiu et al., 2014). In brief, ipsilateral and contralateral DRGs (L3–4) were collected in ice cold sterile 1 × DPBS (Hyclone, SH30028). Samples were centrifuged at 400 × g for 3 minutes. Supernatants were removed and samples were treated with Collagenase A (Sigma, 10103586001) and incubated in a 37°C water bath for 20 minutes. Samples were then centrifuged at 400 × g for 3 minutes. Supernatants were removed and the samples were treated with Collagenase D (Sigma, 1188866001) and a 10% v/v of papain (Roche, 10108014001) in HBSS for another 20 minutes. Cells were centrifuged at 400 × g and the pellet were resuspended in Enzyme T (Sigma, 10109886001) (soybean trypsin inhibitor made in 1:1 bovine serum albumin and DMEM/F12 media (Thermo, 10565161) supplemented with 10% fetal bovine serum (FBS) (Hyclone, SH30088.03) and 1% pen/strep (Sigma, P4333)) to stop the enzymatic reaction. Digested tissues were triturated using a 1 mL fire-polished Pasteur pipette tip and passed through a 70-micron nylon mesh cell strainer (Sigma, CLS431751–50EA), with a subsequent wash using flow buffer (0.5% bovine serum albumin with 0.02% glucose (Sigma, G7528) made in 1 × DPBS). Resultant suspension was centrifuged at 400 × g for 3 minutes and resuspended in flow blocking buffer (anti-CD16/32 purified antibody diluted in flow buffer) for 20 minutes to block fc receptors. Samples were incubated with pre-conjugated extracellular flow antibodies (CD45 & TLR4) for 45 minutes. Samples were then washed with flow buffer, then centrifuged at 400 × g for 3 minutes and resuspended in a fixation/permeabilization buffer (BD Biosciences, 555028) for 45 minutes. Samples were then centrifuged at 400 × g for 3 minutes and were washed twice using the manufacturer provided wash buffer (1×). Samples were incubated with pre-conjugated intracellular flow antibodies (NeuN) for 60 minutes and washed with DAPI for 5 minutes. Samples were then centrifuged at 400 × g for 3 minutes and were washed twice using wash buffer and were then resuspended in flow buffer. Appropriate compensation controls and isotypes were used for determination and gating. After gating DAPI positive cells (to determine debris), immune cells were identified with gating for CD45 and were further gated to identify expression of TLR4. Neurons were initially identified with gating for NeuN and were further gated to identify expression of TLR4. Stained samples were analyzed using a Special Order (4-laser) Becton-Dickinson Fortessa analyzer (Red Oaks, CA) and data were analyzed using FlowJo and FCS Express software (De Novo Software, Los Angeles, CA). Flow cytometry experiments and analyses were performed by T.A.S., M.E.L., and M.D.B. Experimenters were blinded to genotype and surgical condition. For a complete list of antibodies used refer to Table 1.

2.6. Statistical Analysis

Prism 8.01 software (GraphPad, San Diego, CA, USA) was utilized to generate all graphs and statistical analysis. Behavioral data was analyzed using repeated measures Two-Way ANOVA with Tukey’s post hoc tests. Effect size, immunohistochemical, and flow cytometric data were analyzed using Ordinary Two-Way ANOVA with Tukey’s post hoc tests. All data are represented as the standard error of the mean (SEM). A p-value of <0.05 was used to determine statistical significance. All statistical values and corresponding tests can be found in Tables 2-7. All behavioral, immunohistochemical, and flow cytometric experiments and analyses were performed by blinded experimenters.

Table 2.

Statistical values for analyses performed within Figure 1. Paw withdrawal threshold and cold allodynia behavioral data were analyzed using repeated measures Two-Way ANOVA with Tukey’s post hoc. Effect size data were analyzed using Ordinary Two-Way ANOVA with Tukey’s post hoc. ATF3 and HMGB1 localization datasets were analyzed using Ordinary Two-Way ANOVA with Tukey’s post hoc. Significance was set at p<0.05 for all datasets. Bolded values are statistically significant.

Dataset	Main Effect			Interactions		Multiple Comparisons
		F (DFn, DFd)	p-value	F (DFn, DFd)	p-value	Effect	Groups	POD	p-value
Mechanical Sensitivity	Paw Withdrawal	Surgery: F (3, 24) = 140.6 Time: F (2.218, 53.24) = 66.65	p<0.0001 p<0.0001	F (12, 96) = 19.05	p<0.0001	Surgery	Male Female	BL 1 3 5 7 BL 1 3 5 7	p=0.7619 p=0.0093 p<0.0001 p<0.0001 p<0.0001 p=0.4393 p=0.0030 p<0.0001 p<0.0001 p<0.0001
	Effect Size	Surgery: F (1, 24) = 81.93 Sex: F (1, 24) = 1.397	p<0.0001 p=0.2487	F (1, 24) = 0.0411	p=0.8411	Surgery	Male Female		p<0.0001 p<0.0001
Cold Allodynia	Response	Surgery: F (3, 24) = 58.11 Time: F (2.891, 69.39) = 31.56	p<0.0001 p<0.0001	F (12, 96) = 9.429	p<0.0001	Surgery	Male Female	BL 1 3 5 7 BL 1 3 5 7	p=0.4227 p=0.0293 p=0.0027 p=0.0001 p=0.0028 p=0.9964 p=0.2647 p=0.0020 p=0.0056 p=0.0009
	Effect Size	Surgery: F (1, 24) = 129.9 Sex: F (1, 24) = 1.013	p<0.0001 p=0.3242	F (1, 24) = 0.0726	p=0.7899	Surgery	Male Female		p<0.0001 p<0.0001
ATF3 Localization	NF200+ Neurons	Surgery: F (1, 12) = 120.4 Sex: F (1, 12) = 4.019	p<0.0001 p=0.0681	F (1, 12) = 1.347	p=0.2684	Surgery	Male Female		p<0.0001 p<0.0001
	Nav1.8+ Neurons	Surgery: F (1, 12) = 646.2 Sex: F (1, 12) = 48.07	p<0.0001 p<0.0001	F (1, 12) = 34.88	p<0.0001	Surgery Sex	Male Female Sham SNI		p<0.0001 p<0.0001 p>0.9999 p<0.0001
HMGB1 Localization	NF200+ Neurons	Surgery: F (1, 12) = 11.47 Sex: F (1, 12) = 0.2560	p=0.0054 p=0.0622	F (1, 12) = 0.2705	p=0.6124	Surgery	Male Female		p=0.1267 p=0.0341
	Nav1.8+ Neurons	Surgery: F (1, 12) = 16.92 Sex: F (1, 12) = 1.880	p=0.0014 p=0.1954	F (1, 12) = 5.222	p=0.0413	Surgery Sex	Male Female Sham SNI		p=0.9148 p=0.0034 p=0.5843 p=0.0959

3. Results

3.1. The severity of mechanical hypersensitivity and cold allodynia in a neuropathic pain model is similar between males and females, but protein expression levels in the lumbar DRG are different.

To establish if any sexual dimorphisms exist in the magnitude of pain following neuropathic injury, SNI was performed on both male and female WT mice. No significant differences were found between male and female SNI groups in either mechanical hypersensitivity or cold allodynia (Fig. 1A–D, Table 2). Emerging evidence suggests that the mechanisms that mediate neuronal sensitization during the early onset of neuropathic pain differ between sexes (Ross et al., 2018; Tajerian et al., 2015). Myeloid-derived immune cells are strongly implicated in driving the early onset of pain following injury, however; differences between sexes in direct neuronal sensitization in the lumbar DRG are poorly understood (Peng et al., 2016). To test the hypothesis that neuronal injury after SNI is different between males and females we first utilized our Na_v1.8^tdT reporter line to verify colocalization between our Na_v1.8 antibody and cre-mediated tdTomato expression in Na_v1.8⁺ nociceptors. Naïve male and female Na_v1.8^tdT lumbar DRGs (L3–5) were immunoassayed with Na_v1.8 and NF200 antibody (Fig. 1E, Table 2). They were then assessed for colocalization with tdTomato. We found there to be a high degree of colocalization (75%) between Na_v1.8 antibody and tdTomato expression (Fig. 1F, Table 2). This indicates robust specificity of the Na_v1.8 antibody with endogenous expression of tdTomato. Next, we used histochemical markers to identify small (Na_v1.8) and large (NF200) diameter nociceptors in the DRG (L3–5) of male and female WT mice. We then examined HMGB1 cytosolic localization and ATF3 expression in these distinct neuronal populations 3 days post SNI (Fig. 1G–J, Table 2). We discovered a sexual dimorphism where both male and female small diameter nociceptors (Na_v1.8⁺) exhibit significantly elevated expression of ATF3 3 days post SNI, but females have significantly more ATF3 expressed as compared to males. (Fig. 1K, Table 2). Both males and females have significantly increased ATF3 expression in their large diameter sensory neurons (NF200⁺) 3 days post SNI as compared to sham controls (Fig. 1L, Table 2). It is evident that these neuronal subpopulations exhibit signs of cellular injury in both males and females after SNI. We then measured cellular localization of HMGB1 in these neuronal subpopulations. Only female WT mice exhibit significantly elevated cytosolic localization of HMGB1 3 days post SNI in their Na_v1.8⁺ nociceptors as compared to sham controls (Fig. 1M, Table 2). Additionally, only female WT mice exhibit significantly elevated cytosolic localization of HMGB1 3 days post SNI in their NF200⁺ neurons as compared to sham controls (Fig. 1N, Table 2). Taken together, our data reveals an inherent sexual dimorphism in both inflammatory signaling and neuronal injury in the context of neuropathic pain. This simple, yet enlightening finding pushed us to further investigate the dichotomy of sex and cell-specific sensitization in the peripheral nervous system.

Figure 1.

The magnitude of pain after SNI does not differ between sexes following SNI, however; the mechanisms during its onset at the level of the lumbar (L3–5) DRGs are. Spared-nerve injury was performed on the left hindlimb of male and female mice. A, Hind paw mechanical withdrawal thresholds were measured prior to surgery and on days: 1, 3, 5, and 7 post-surgery in both male and female sham (n=7) and SNI (n=7) wild type mice. B, Data for males and females combined shown as mechanical effect size. C, Hind paw response to application of acetone were measured prior to surgery and on days: 1, 3, 5, and 7 post-surgery in both male and female sham (n=7) and SNI (n=7) mice. D, Data for males and females combined shown as acetone effect size. E, Naïve Na_v1.8^tdT female and male lumbar (L3–5) DRGs were immunostained with Na_v1.8 (green), tdTomato (red) and NF200 (Blue); representative images from n=4 mice). Scale bar: 50 μm. Magnification: 20x. F, Quantification of colocalization between Na_v1.8, tdTomato and NF200 (n=4). G-H Female and male WT lumbar DRGs (L3–5) were immunostained 3D post SNI with DAPI (teal), Na_v1.8 (green), ATF3 (red) and NF200 (blue; representative images from n=4 mice). Scale bar: 50 μm. Magnification: 20x. K, Quantification of ATF3 colocalization with Na_v1.8⁺ neurons of both sexes 3 days after SNI (n=4 per group). L, Quantification of ATF3 colocalization with NF200⁺ neurons of both sexes 3 days after SNI (n=4 per group). I-J, Female, and male WT lumbar DRGs (L3–5) were immunostained 3D post SNI with DAPI (teal), Na_v1.8 (green), HMGB1 (red) and NF200 (blue; representative images from n=4 mice). White arrows point to an example of nuclear or cytosolic localization of HMGB1. Scale bar: 20 μm. Magnification: 40x. M, Quantification of HMGB1 cytosolic localization in Na_v1.8⁺ neurons of both sexes 3 days after SNI (n=4 per group). N, Quantification of HMGB1 cytosolic localization in NF200⁺ neurons of both sexes 3 days after SNI (n=4 per group). *p < 0.05; **p < 0.01; ****p < 0.0001. BL = Baseline.

3.2. Sensory neuron and immune cell expression of TLR4

To investigate baseline and surgery-induced levels of TLR4 on DRG cell populations, ipsilateral and contralateral DRGs (L3-L5) were isolated from males and females 3 days post SNI (Chiu et al., 2014). Moreover, we verified protein expression of our null-re-expression animals when we added whole-body null (TLR4^LoxTB) and cell-specific re-expression (Na_v1.8^TLR4LoxTB) groups to the study. Neuronal (NeuN⁺) and immune cell (CD45⁺) populations were identified and assessed for TLR4 at various stages and genotypes (Fig. 2A). We determined that there were no significant sex or surgery-induced differences in TLR4 protein expression patterns in DRG neurons in ipsilateral DRGs, compared to contralateral DRGs, with roughly 30% of the neurons expressing TLR4 in males and females (Fig. 2B & 2C). We saw that our whole-body null animals expressed minimum amounts of TLR4 across the neuronal populations, as expected (Fig. 2B & 2C). Opposingly, we observed a male-driven sex-dependent upregulation of ipsilateral TLR4 expression in CD45⁺ immune cells 3 days post SNI (Fig. 2E), but no female differences. This data juxtaposes recent findings of no sex-differences in DRG immune cells after neuropathic injury (Yu et al., 2020a). Apparently, it is imperative to understand not only the cell population amount, but the expression pattern of those cells in the DRG.

Figure 2.

There are no basal or surgery-induced sex differences in TLR4 expression in sensory neurons and a surgery-induced upregulation of TLR4 in DRG immune cells 3D post SNI. A, Lumbar DRGs (L3–5) were enzymatically dissociated, stained with TLR4, NeuN, and DAPI and subjected to flow cytometry. After gating on both neurons and immune cells based on forward and side scatter, NeuN⁺ and CD45⁺ cells were further gated to TLR4⁺ populations. B, NeuN⁺/TLR4⁺ cells in control (contra) and surgerized DRGs (ipsi) 3D post SNI in males. C, NeuN⁺/TLR4⁺ cells in control (contra) and surgerized DRGs (ipsi) 3D post SNI in females. D, Combined male and female data; ratio of ipsilateral-to-contralateral NeuN⁺/TLR4⁺ DRG cells 3D post SNI. E, CD45⁺/TLR4⁺ cells in control (contra) and surgerized DRGs (ipsi) 3D post SNI in males. F, CD45⁺/TLR4⁺ cells in control (contra) and surgerized DRGs (ipsi) 3D post SNI in females. G, Combined male and female data; ratio of ipsilateral-to-contralateral CD45⁺/TLR4⁺ DRG cells 3D post SNI. Asterisks on the bar graphs indicate significant differences between the ^groups. *p < 0.05; **p < 0.01; ***p < 0.001, (n= 3–4).

3.3. Sensory Neuron TLR4 is necessary during the onset of neuropathic mechanical hypersensitivity in female mice.

Investigating the cell-specific implications of TLR4 activation in behavioral pain phenotypes is necessary to understand the biological relevance of cell-molecule interactions. Danger signals are released by both neuronal and non-neuronal cells in response to tissue injury, where neuronally expressed TLR4 allows for rapid detection of these proteins (Liu et al., 2014). Here, we asked if there are sex differences in the response to neuropathic tissue injury by neuronally expressed TLR4. To test this, we used transgenic male and female mice that lack TLR4 on their Na_v1.8⁺ nociceptors combined with a spared nerve injury (SNI) model of peripheral neuropathic injury (Fig. 3A). Genotypes for Na_v1.8^TLR4fl/fl mice are confirmed using polymerase chain reaction (PCR) and gel electrophoresis (Fig. 3B, C). We assessed the onset of neuropathic pain by measuring mechanical hypersensitivity and cold allodynia and our data reveals robust sex and genotype-dependent behavioral effects. (Mecklenburg et al., 2020)Our data demonstrates the role of TLR4 in nociceptors and their female-specific role in regulating neuropathic pain development. Male Na_v1.8^TLR4fl/fl mice exhibit similar mechanical withdrawal thresholds to their TLR4^fl/fl littermates after SNI (Fig. 3D, Table 3). Female Na_v1.8^TLR4fl/fl mice exhibit reduced mechanical withdrawal thresholds days 1, 3 and 5 after SNI as compared to their TLR4^fl/fl littermates (Fig. 3E, Table 3). Sexes are directly compared using effect sizes (Fig. 3F, Table 3). Interestingly, male but not female mice show reduced cold allodynia on day 3 post-surgery as compared to their TLR4^fl/fl littermates (Fig. 3G–H, Table 3). Sexes are directly compared using effect sizes (Fig. 3I, Table 3). This demonstrates a distinction in the role of TLR4 in mechanical vs. thermal during neuropathic pain development. It is possible that the mechanisms which mediate cold and mechanical hypersensitivity are sexually dimorphic in nature, and a distinguished subset of transient receptor potential (TRP)⁺ sensory neurons in the DRG differ mechanistically between males and females in a neuropathic state, however; this was not the focus of the current study.

Figure 3.

Development of neuropathic pain is attenuated in Na_v1.8^TLR4fl/fl female, but not male mice. Spared-nerve injury was performed on the left hindlimb of male and female mice. A, Schematic representing the murine genetic model, Na_v1.8^TLR4fl/fl. B, Example PCR, and gel electrophoresis for Na_v1.8cre. A – indicates homozygous WT mice, a −/+ indicates heterozygous Na_v1.8cre mice, a +/+ indicates homozygous Na_v1.8cre. C, Example PCR, and gel electrophoresis for TLR4^fl/fl. A - indicates homozygous WT mice, a −/+ indicates heterozygous TLR4^fl/fl mice, a +/+ indicates homozygous TLR4^fl/fl mice. D-E, Hind paw mechanical withdrawal thresholds were measured prior to surgery and on days: 1, 3, 5, and 7 post-surgery in both male and female sham (n=8), TLR4^fl/fl (n=9), Na_v1.8^TLR4fl/fl (n=8) mice. F, Data for males and females combined shown as mechanical effect size. G-H, Hind paw response to application of acetone were measured prior to surgery and on days: 1, 3, 5, and 7 post-surgery in both male and female sham (n=8), TLR4^fl/fl (n=9) and Na_v1.8^TLR4fl/fl (n=8) mice. I, Data for males and females combined shown as acetone effect size. Asterisks on the line graphs indicate significant differences between the Na_v1.8^TLR4fl/fl and TLR4^fl/fl groups. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. BL = Baseline.

Table 3.

Statistical values for analyses performed within Figure 3. Mechanical sensitivity and cold allodynia behavioral data were analyzed with a repeated-measures Two-Way ANOVA with Tukey’s post hoc. Effect size datasets were analyzed using Ordinary Two-Way ANOVA with Tukey’s post hoc. Statistical significance was set at p<0.05. All values deemed statistically significant are bolded.

Dataset	Main Effect			Interactions		Multiple Comparisons
		F (DFn, DFd)	p-value	F (DFn, DFd)	p-value	Effect	Groups	POD	p-value
Mechanical Sensitivity	Males	Genotype: F (2, 24) = 89.06 Time: F (3.068, 73.64) = 38.10	p<0.0001 p<0.0001	F (8, 96) = 9.794	p<0.0001	Surgery Genotype	Mixed Sham vs TLR4fl/fl Mixed Sham vs Nav1.8TLR4LoxTB TLR4LoxTB vs Nav1.8TLR4LoxTB	BL 1 3 5 7 BL 1 3 5 7 BL 1 3 5 7	p=0.3123 p=0.0013 p<0.0001 p<0.0001 p<0.0001 p=0.2072 p=0.0036 p<0.0001 p<0.0001 p<0.0001 p=0.7957 p=0.7581 p=0.2611 p=0.4406 p=0.2021
	Females	Genotype: F (2, 24) = 69.27 Time: F (3.278, 78.67) = 19.64	p<0.0001	F (8, 96) = 9.249	p<0.0001	Surgery Genotype	Mixed Sham vs TLR4fl/fl Mixed Sham vs Nav1.8TLR4LoxTB TLR4LoxTB vs Nav1.8TLR4LoxTB	BL 1 3 5 7 BL 1 3 5 7 BL 1 3 5 7	p=0.9777 p=0.0012 p<0.0001 p<0.0001 p<0.0001 p=0.9158 p=0.0266 p=0.0783 p=0.0054 p<0.0001 p=0.7957 p=0.0419 p=0.0074 p=0.0010 p=0.1120
	Effect Size	Genotype: F (2, 49) = 44.27 Sex: F (1, 49) = 0.8985	p<0.0001 p=0.3478	F (2, 49) = 3.384	p=0.0420	Genotype	Male SNI Female SNI		p=0.5528 p=0.0007
Cold Allodynia	Males	Genotype: F (2, 20) = 24.29 Time: F (3.155, 63.10) = 57.32	p<0.0001 p<0.0001	F (8, 80) = 16.68	p<0.0001	Surgery Genotype	Mixed Sham vs TLR4fl/fl Mixed Sham vs Nav1.8TLR4LoxTB TLR4LoxTB vs Nav1.8TLR4LoxTB	BL 1 3 5 7 BL 1 3 5 7 BL 1 3 5 7	p=0.9239 p=0.8930 p=0.0100 p=0.0013 p<0.0001 p=0.7419 p=0.2450 p=0.0010 p=0.0340 p=0.0005 p=0.9249 p=0.3318 p=0.4565 p=0.0850 p=0.0260
	Females	Genotype: F (2, 20) = 21.04 Time: F (2.406, 48.12) = 10.32	p<0.0001 p<0.0001	F (8, 80) = 4.242	p=0.0003	Surgery Genotype	Mixed Sham vs TLR4fl/fl Mixed Sham vs Nav1.8TLR4LoxTB TLR4LoxTB vs Nav1.8TLR4LoxTB	BL 1 3 5 7 BL 1 3 5 7 BL 1 3 5 7	p>0.9999 p=0.7482 p=0.0029 p=0.0009 p=0.0001 p=0.9749 p=0.1593 p=0.0010 p=0.0325 p=0.0004 p=0.9718 p=0.4756 p=0.1864 p=0.9781 p=0.5727
	Effect Size	Genotype: F (2, 40) = 24.96 Sex: F (1, 40) = 2.021	p<0.0001 p=0.1629	F (2, 40) = 0.1819	p=0.1819	Genotype	Male SNI Female SNI		p=0.0368 p=0.9843

3.4. Sensory Neuron TLR4 is sufficient during the onset of neuropathic mechanical hypersensitivity in female mice.

Whole body TLR4 has been shown to influence recovery from a non-terminal neuropathic injury (Stokes et al., 2013). Moreover, TLR4 knockout mice have reduced mechanical and thermal hypersensitivity in response to neuropathic injury (Piao et al., 2018). It is still unclear what specific populations of cells that express TLR4 contribute to these behavioral effects. To investigate the direct action of TLR4 on peripheral sensory neurons, we utilized Na_v1.8^TLR4LoxTB mice combined with an SNI model of neuropathic pain (Fig. 4A). Genotypes for Na_v1.8^TLR4LoxTB mice are confirmed using PCR and gel electrophoresis (Fig. 4B, C). Here, we tested the sufficiency of TLR4 expressed only on Na_v1.8⁺ DRG neurons to produce a neuropathic behavioral pain phenotype similar to wild type mice. Analysis reveals significant genotype-dependent behavioral phenotypes and robust sex differences. Male TLR4^LoxTB mice show significantly diminished mechanical withdrawal thresholds during days 1, 3 and 5 post-surgery as compared to WT littermates. Interestingly, male Na_v1.8^TLR4LoxTB mice do not recapitulate the severe mechanical hypersensitivity seen in WT littermates on days 1, 3 and 5 post-surgery (Fig. 4D, Table 4). Female TLR4^LoxTB mice show significantly diminished mechanical withdrawal thresholds during days 1, 3 and 5 post-surgery as compared to WT littermates, however; only female Na_v1.8^TLR4LoxTB mice recapitulate the severe mechanical hypersensitivity seen in WT littermates on days 1, 3 and 5 post-surgery (Fig. 4E, Table 4). Sexes are directly compared using effect sizes (Fig. 4F, Table 4). Surprisingly, no sex or genotype-dependent effects are seen across time in cold allodynia development (Fig. 4G–H, Table 4). Sexes are directly compared using effect sizes (Fig. 4I, Table 4). These data suggest direct activation of neuronally expressed TLR4 by endogenous danger signals released during injury is enough to cause a neuropathic behavioral pain phenotype only in female mice.

Figure 4.

Development of neuropathic pain is phenotypically normal in Na_v1.8^TLR4LoxTB female, but not male mice. Spared-nerve injury was performed on the left hindlimb of male and female mice. A, Schematic representing the murine genetic model, Na_v1.8^TLR4LoxTB. B, Example PCR, and gel electrophoresis for Na_v1.8cre. A - indicates homozygous WT mice, a −/+ indicates heterozygous Na_v1.8cre mice, a +/+ indicates homozygous Na_v1.8cre. C, Example PCR, and gel electrophoresis for TLR4^LoxTB. A – indicates homozygous WT mice, a −/+ indicates heterozygous TLR4^LoxTB mice, a +/+ indicates homozygous TLR4^LoxTB mice. D-E, Hind paw mechanical withdrawal thresholds were measured prior to surgery and on days: 1, 3, 5, and 7 post-surgery in both male and female sham (n=12), wild type (n=4), TLR4^LoxTB (n=9), F, Data for males and females combined shown as mechanical effect size. G-H, Hind paw response to application of acetone were measured prior to surgery and on days: 1, 3, 5, and 7 post-surgery in both male and female sham (n=12), wild type (n=4), TLR4^LoxTB (n=9), Na_v1.8^TLR4LoxTB (n=7) mice. I, Data for males and females combined shown as acetone effect size. Asterisks on the line graphs indicate significant differences between the Na_v1.8^TLR4LoxTB and TLR4^LoxTB groups. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. BL = Baseline.

Table 4.

Statistical values for analyses performed within Figure 4. Mechanical sensitivity and cold allodynia behavioral data were analyzed with a repeated-measures Two-Way ANOVA with Tukey’s post hoc. Effect size datasets were analyzed using Ordinary Two-Way ANOVA with Tukey’s post hoc. Statistical significance was set at p<0.05. Comparisons between wild-type and sham groups can be found in Table 2. All values deemed statistically significant are bolded.

Dataset	Main Effect			Interactions		Multiple Comparisons
		F (DFn, DFd)	p-value	F (DFn, DFd)	p-value	Effect	Groups	POD	p-value
Mechanical Sensitivity	Females	Genotype: F (3.28) = 55.08 Time: F (3.39, 94.92) = 67.17	p<0.0001 p<0.0001	F (12. 112) = 20.37	p<0.0001	Surgery Genotype	Sham vs TLR4LoxTB Sham vs Nav1.8TLR4LoxTB TLR4LoxTB vs Nav1.8TLR4LoxTB WT vs TLR4LoxTB WT vs Nav1.8TLR4LoxTB	BL 1 3 5 7 BL 1 3 5 7 BL 1 3 5 7 BL 1 3 5 7 BL 1 3 5 7	p=0.2686 p=0.1743 p=0.9994 p=0.1102 p<0.0001 p=0.0181 p=0.0026 p=0.0003 p=0.0265 p<0.0001 p=0.2939 p=0.0003 p=0.0004 p=0.1353 p=0.0299 p=0.8767 p=0.0019 p<0.0001 p<0.0001 p=0.0147 p=0.9267 p=0.2831 p=0.2463 p=0.2698 p=0.8959
	Males	Genotype: F (3, 28) = 34.18 Time: F (3.612, 101.1) = 78.45	p<0.0001 p<0.0001	F (12, 112) = 9.249	p<0.0001	Surgery Genotype	Sham vs TLR4LoxTB Sham vs Nav1.8TLR4LoxTB TLR4LoxTB vs Nav1.8TLR4LoxTB WT vs TLR4LoxTB WT vs Nav1.8TLR4LoxTB	BL 1 3 5 7 BL 1 3 5 7 BL 1 3 5 7 BL 1 3 5 7 BL 1 3 5 7	p=0.3918 p=0.0540 p>0.9999 p=0.9731 p<0.0001 p=0.0441 p=0.7980 p=0.9997 p=0.9497 p=0.0171 p=0.7958 p=0.1310 p=0.9994 p>0.9999 p=0.5950 p=0.8051 p=0.0060 p<0.0001 p<0.0001 p=0.0110 p=0.9994 p=0.0266 p<0.0001 p<0.0001 p=0.1004
	Effect Size	Genotype: F (3, 56) = 42.25 Sex: F (1, 56) = 0.1764	p<0.0001 p=0.6761	F (3, 56) = 5.613	p=0.0019	Genotype	Males: Sham vs TLR4LoxTB		p=0.5843
							Males: Sham vs Nav1.8TLR4LoxTB		p=0.2216
							Males: WT vs TLR4LoxTB		p<0.0001
							Males: WT vs Nav1.8TLR4LoxTB		p<0.0001
							Males: Nav1.8TLR4LoxTB vs TLR4LoxTB		p=0.9959
							Females: Sham vs TLR4LoxTB		p=0.3275
							Females: Sham vs Nav1.8TLR4LoxTB		p<0.0001
							Females: WT vs TLR4LoxTB		p=0.0002
							Females: WT vs Nav1.8TLR4LoxTB		p=0.9990
							Females: Nav1.8TLR4LoxTB vs TLR4LoxTB		p<0.0001
						Sex	Male TLR4LoxTB vs Female TLR4LoxTB		p=0.9990
							Male TLR4LoxTB vs Female Nav1.8TLR4LoxTB		p=0.0004
							Male Nav1.8TLR4LoxTB vs Female TLR4LoxTB		p=0.9079
							Male Nav1.8TLR4LoxTB vs Female Nav1.8TLR4LoxTB		p=0.0094
Cold Allodynia	Females	Genotype: F (3, 28) = 31.54 Time: F (2.890, 80.93) = 41.56	p<0.0001 p<0.0001	F (2, 112) = 6.852	p<0.0001	Surgery Genotype	Sham vs TLR4LoxTB Sham vs Nav1.8TLR4LoxTB TLR4LoxTB vs Nav1.8TLR4LoxTB WT vs TLR4LoxTB WT vs Nav1.8TLR4LoxTB	BL 1 3 5 7 BL 1 3 5 7 BL 1 3 5 7 BL 1 3 5 7 BL 1 3 5 7	p=0.6547 p=0.9513 p=0.0951 p=0.0022 p=0.0016 p=0.2301 p=0.7320 p=0.0155 p=0.0169 p=0.0042 p=0.8570 p=0.4368 p=0.9923 p=0.5887 p=0.5554 p=0.6584 p=0.4968 p=0.0233 p=0.8123 p=0.5127 p=0.4585 p=0.7330 p=0.0030 p=0.9948 p=0.9593
	Males	Genotype: F (3, 28) = 38.43 Time: F (2.694, 75.42) = 44.19	p<0.0001 p<0.0001	F (12, 112) = 9.260	p<0.0001	Surgery Genotype	Sham vs TLR4LoxTB Sham vs Nav1.8TLR4LoxTB TLR4LoxTB vs Nav1.8TLR4LoxTB WT vs TLR4LoxTB WT vs Nav1.8TLR4LoxTB	BL 1 3 5 7 BL 1 3 5 7 BL 1 3 5 7 BL 1 3 5 7 BL 1 3 5 7	p>0.9999 p=0.3785 p=0.0033 p=0.0172 p<0.0001 p=0.9660 p=0.6250 p=0.1112 p=0.0095 p=0.0033 p=0.9576 p>0.9999 p=0.5759 p=0.7327 p=0.8198 p=0.5393 p=0.6010 p=0.0320 p=0.0625 p=0.4515 p=0.8149 p=0.6146 p=0.9946 p=0.2385 p=0.8953
	Effect Size	Genotype: F (3, 57) = 61.34 Sex: F (1, 57) = 3.348	p<0.0001 p=0.0725	F (3, 57) = 1.065	p=0.3712	Genotype	Males: Sham vs TLR4LoxTB		p<0.0001
							Males: Sham vs Nav1.8TLR4LoxTB		p<0.0001
							Males: WT vs TLR4LoxTB		p=0.7256
							Males: WT vs Nav1.8TLR4LoxTB		p=0.9843
							Males: Nav1.8TLR4LoxTB vs TLR4LoxTB		p=0.8547
							Females: Sham vs TLR4LoxTB		p<0.0001
							Females: Sham vs Nav1.8TLR4LoxTB		p<0.0001
							Females: WT vs TLR4LoxTB		p=0.0177
							Females: WT vs Nav1.8TLR4LoxTB		p=0.9364
							Females: Nav1.8TLR4LoxTB vs TLR4LoxTB		p=0.0183

3.5. Injury marker, ATF3, in small diameter neurons is upregulated in a sex specific and TLR4 fashion

To investigate the sexual dimorphisms driving the development of neuropathic pain we used a neuronal marker of injury, ATF3, and assessed colocalization with small diameter nociceptors (Na_v1.8⁺) and large diameter neurons (NF200⁺) (Fig. 5A–B, Table 6). In female Na_v1.8^TLR4LoxTB mice, we show significantly upregulated ATF3 expression in their Na_v1.8⁺ DRG neurons 3 days post-SNI as compared to TLR4^LoxTB mice. Conversely, in male Na_v1.8^TLR4LoxTB mice there is no significant upregulation of ATF3 3 days post-SNI in their Na_v1.8⁺ DRG neurons as compared to TLR4^LoxTB counterparts. There is, however, a significant upregulation in both male and female WT mice 3 days after SNI as compared to shams (Fig. 5C, Table 6). This sets a precedent that endogenous TLR4 activation on Na_v1.8⁺ nociceptors after nerve injury regulates transcription factors (ATF3) that are important in neuronal injury in female, but not male mice. Both male and female Na_v1.8^TLR4LoxTB mice exhibit no significant differences in ATF3 expression after SNI as compared to TLR4^LoxTB counterparts in NF200⁺ neurons. There is, however, a significant upregulation in both male and female WT mice 3 days after SNI as compared to shams (Fig. 5D, Table 6). This data demonstrates that TLR4 has an important role in regulating neuronal injury after neuropathic injury in both sexes, but the cell types that express TLR4 which are responsible for these effects differ between males and females. This interesting finding led us to investigate the downstream implications of TLR4 activation on these small diameter nociceptors via HMGB1 localization.

Figure 5.

The injury marker, ATF3, is upregulated in small diameter nociceptors in a sex and genotype dependent manner via TLR4 expression. A, Female sham (mixed genotypes), WT, Na_v1.8^TLR4LoxTB, and TLR4^LoxTB lumbar (L3–5) DRGs were immunostained 3D post SNI with DAPI (teal), Na_v1.8 (green), ATF3 (red) and NF200 (blue; representative images from n=4 mice). B, Male sham (mixed genotypes), WT, Na_v1.8^TLR4LoxTB, and TLR4^LoxTB lumbar (L3–5) DRGs were immunostained 3D post SNI with DAPI (teal), Na_v1.8 (green), ATF3 (red) and NF200 (blue; representative images from n=4 mice). C, Quantification of ATF3 colocalization with Na_v1.8⁺ neurons of both sexes (n=4 per group). D, Quantification of ATF3 colocalization with NF200⁺ neurons of both sexes (n=4 per group). Scale bar: 50 μm. Magnification: 20x. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. ns = not significant.

Table 6.

Statistical values for analyses performed within Figure 6. HMGB1 localization datasets were analyzed using Ordinary Two-Way ANOVA with Tukey’s post hoc. Significance was set at p<0.05 for all datasets. Bolded values are statistically significant.

Dataset		Main Effect		Interactions		Multiple Comparisons
		F (DFn, DFd)	p-value	F (DFn, DFd)	p-value	Effect	Groups	p-value
HMGB1 Localization	Nav1.8+ Neurons	Genotype: F (3, 24) = 8.720 Sex: F (1, 24) = 3.194	p=0.0004 p=0.0866	F (3, 24) = 2.697	p=0.0684	Genotype	Males: Sham vs TLR4LoxTB	p=0.9996
							Males: Sham vs Nav1.8TLR4LoxTB	p>0.9999
							Males: WT vs TLR4LoxTB	p=0.4787
							Males: WT vs Nav1.8TLR4LoxTB	p=0.7364
							Males: Nav1.8TLR4LoxTB vs TLR4LoxTB	p=0.9994
							Females: Sham vs TLR4LoxTB	p>0.9999
							Females: Sham vs Nav1.8TLR4LoxTB	p=0.0958
							Females: WT vs TLR4LoxTB	p=0.0005
							Females: WT vs Nav1.8TLR4LoxTB	p=0.2405
							Females: Nav1.8TLR4LoxTB vs TLR4LoxTB	p=0.0948
	NF200+ Neurons	Genotype: F (3, 24) = 4.233 Sex: F (1, 24) = 0.01456	p=0.0155 p=0.9050	F (3,24) = 0.252	p=0.8593	Genotype	Males: Sham vs TLR4LoxTB	p=0.9202
							Males: Sham vs Nav1.8TLR4LoxTB	p=0.1687
							Males: WT vs TLR4LoxTB	p=0.6859
							Males: WT vs Nav1.8TLR4LoxTB	p=0.9784
							Males: Nav1.8TLR4LoxTB vs TLR4LoxTB	p=0.4483
							Females: Sham vs TLR4LoxTB	p=0.9914
							Females: Sham vs Nav1.8TLR4LoxTB	p=0.3730
							Females: WT vs TLR4LoxTB	p=0.1849
							Females: WT vs Nav1.8TLR4LoxTB	p=0.8830
							Females: Nav1.8TLR4LoxTB vs TLR4LoxTB	p=0.5369

3.6. TLR4 expression in small diameter neurons mediates HMGB1 translocation from the nucleus to the cytosol of the cell

We demonstrated the sufficiency of TLR4 to upregulate the injury marker, ATF3, in Na_v1.8⁺ DRG neurons in female, but not male mice. Now that we have established a significant role for TLR4 in regulating neuronal injury, we further investigated the downstream implication of TLR4 signaling in these populations of small and large diameter DRG neurons. Here, we immunoassayed lumbar DRGs (L3–5) from female and male Na_v1.8^TLR4LoxTB, TLR4^LoxTB, and WT mice 3 days post-SNI for cellular localization of HMGB1, a DAMP, in both small (Na_v1.8⁺) and large (NF200⁺) diameter neurons (Fig. 6A, B). We found there to be a trend of increased cytosolic HMGB1 in female Na_v1.8⁺ DRG neurons of Na_v1.8^TLR4LoxTB mice as compared to TLR4^LoxTB mice. Interestingly, male Na_v1.8^TLR4LoxTB mice do not show significantly increased cytosolic HMGB1 in their Na_v1.8⁺ DRG neurons 3 days after SNI as compared to TLR4^LoxTB mice. These results are recapitulated only in female WT mice, indicating direct activation of TLR4 on Na_v1.8⁺ DRG neurons through endogenous DAMP signaling is sufficient to induce mobilization of HMGB1 to the cytosol of the cell (Fig. 6C, Table 6). We find no significant differences between both male and female Na_v1.8^TLR4LoxTB and TLR4^LoxTB mice regarding cytosolic HMGB1 localization in NF200⁺ large diameter DRG neurons 3 days after SNI (Fig. 6D, Table 6). These highly sex and genotype dependent results indicate that endogenous activation of TLR4 on sensory neurons in the DRG of female mice after neuropathic injury initiates cytosolic mobilization of HMGB1. This, taken together with the upregulation of ATF3 expression in nociceptors indicates females utilize a TLR4-dependent pathway to facilitate pain behaviors during a neuropathic injury.

Figure 6.

HMGB1 translocation following neuropathic injury is upregulated in a sex and genotype dependent manner in small diameter nociceptors and is mediated by TLR4. A, Female sham (mixed genotypes), WT, Na_v1.8^TLR4LoxTB, and TLR4^LoxTB lumbar (L3–5) DRGs were immunostained 3D post SNI with DAPI (teal), Na_v1.8 (green), HGMB1 (red) and NF200 (blue; representative images from n=4 mice). White arrows point to an example of nuclear or cytosolic localization of HMGB1. B, Male sham (mixed genotypes), WT, Na_v1.8^TLR4LoxTB, and TLR4^LoxTB lumbar (L3–5) DRGs were immunostained 3D post SNI with DAPI (teal), Na_v1.8 (green), HGMB1 (red) and NF200 (blue; representative images from n=4 mice). White arrows point to an example of nuclear or cytosolic localization of HMGB1. C, Quantification of cytosolic localization of HMGB1 in Na_v1.8⁺ neurons of both sexes (n=4 per group). D, Quantification of cytosolic localization of HMGB1 in NF200⁺ neurons of both sexes (n=4 per group). Scale bar: 20 μm. Magnification: 40x. *p < 0.05; **p < 0.01; ****p < 0.0001. ns = not significant.

4. Discussion

In the nociceptive system, bi-directional communication between neuronal and non-neuronal cells is a core mechanism in mediating the response to injury (Iwata and Shinoda, 2019; Szabo-Pardi et al., 2021). Emerging evidence suggests that both sexes utilize myeloid-derived immune cells in the central and peripheral nervous system to drive cytokine production and subsequent neuronal sensitization (Peng et al., 2016; Yu et al., 2020b). Currently, little is known about direct neuronal response to activation by endogenous DAMP signaling during injury. Here, we report that female mice utilize a specific pathway where direct activation of TLR4 by endogenous danger signals on small diameter nociceptors in the DRG leads to upregulation of stress-induced ATF3 and DAMPs (HMGB1) in the lumbar DRG. Additionally, these molecular changes lead to a delay in the onset of neuropathic pain associated with peripheral nerve injury in a TLR4 dependent manner (Fig. 7). Our genetic and histological experiments demonstrate a robust sexual dimorphism wherein peripheral nociceptors expressing TLR4 mediate the early onset of neuropathic mechanical hypersensitivity in female mice, in addition to regulating molecular changes in peripheral nervous tissue. Moreover, TLR4 expression on Na_v1.8⁺ DRG neurons is both necessary and sufficient for the behavioral response to neuropathic injury. Use of this genetic model has allowed us to elucidate a novel pathway to neuronal injury in female mice. Our results support the idea that TLR4 signaling on peripheral nociceptors plays an important role in direct sensitization through the up-regulation of stress-induced transcription factors and DAMPs in these cells. In our view, this identifies a novel mechanism of direct neuronal sensitization in female mice crucial for the early onset of neuropathic pain. This provides novel evidence for the dissection of the sexually dimorphic response to tissue injury (Anwar et al., 2019).

Figure 7.

Schematic model indicated that TLR4 expressed on Na_v1.8⁺ nociceptors in the lumbar DRGs (L3–5) plays an important role in regulating the localization of HMGB1 in response to neuropathic injury in females, but not males. This is both behaviorally and molecularly significant, as female mice lacking TLR4 expression on their Na_v1.8+ nociceptors exhibit reduced mechanical hypersensitivity following SNI. Moreover, re-expression of TLR4 only on these Na_v1.8 nociceptors recapitulates a behavioral phenotype similar to that of wild type mice. This effect is not present in male mice of either genotype.

Women are found to have a higher incidence of neuropathic pain as compared to males, in addition to lowered efficacy of common therapeutics (DiBonaventura et al., 2017). As such, there exists a need to understand the core mechanisms that differentiate male and female nociceptive circuitry. Compared with the well-documented behavioral phenotypes seen after SNI, we show that male and female wild type mice exhibit no differences in the magnitude of both mechanical hypersensitivity and cold allodynia (Bravo-Caparros et al., 2019; Sorge et al., 2015). Interestingly, evidence suggests a fundamental disconnect between the mechanisms involved in the onset of neuropathic pain (Mogil, 2020). For example, male mice lacking chemokine receptor 1 (CX3CR1) expressing immune cells exhibit delayed development of mechanical hypersensitivity; however, female mice remain unaffected (Peng et al., 2016). It has also been shown that neuronally expressed NOD-like receptor protein 3 (NLRP3) inflammasome has sex specific effects in pain, where females are dependent on NLRP3 in sensory neurons to facilitate mechanical hypersensitivity (Cowie et al., 2019). NLRP3 inflammasome activation is enhanced via intracellular cascades initiated through TLR4 activation (Bauernfeind et al., 2009). Importantly, inflammasome activation in tissue injury increases production of IL-1β, which plays a prominent role in pain, specifically mechanical hypersensitivity (Chen et al., 2015). Sensory neurons express large amounts of interleukin (IL) −1β receptor, which leads us to believe that autocrine signaling through the TLR4 and NLRP3 pathways in sensory neurons presents a point of sexually divergent mechanisms of pain sensitization.

In response to this, we sought to characterize endogenous DAMP signaling and chose to focus on HMGB1, specifically, as it has been implicated by numerous studies in chronic pain development (Campana et al., 2009; O’Connor et al., 2003; Zhang et al., 2015). HMGB1 is localized in the nucleus of cells in homeostasis, however; with adequate stimulation, hyperacetylation of the protein causes translocation to the cytosolic and subsequent release in both neuronal and non-neuronal cells (Lotze and Tracey, 2005). Secreted HMGB1 acts on a variety of receptors, namely the receptor for advanced glycation and end products (RAGE), TLR2, TLR4 and TLR5 all of which are involved in immunomodulation and neuroinflammation (Das et al., 2016; Frasnelli et al., 2015). Interestingly, HMGB1 is capable of upregulating canonical NLRP3 activation, both of which are dependent on the nuclear factor (NF)-κB pathway which is directly linked to TLR4 signaling (Chi et al., 2015). Downstream consequences of neuronally-expressed TLR4 activation are poorly understood, however; evidence does suggest that neurons are able to produce cytokines through its activation (Leow-Dyke et al., 2012). Moreover, it has been demonstrated that activation of TLRs on sensory neurons does promote the production of neuropeptides and chemokines which may facilitate neuronal sensitization (Diogenes et al., 2011). To our surprise, we have identified a robust sexual dimorphism in the lumbar DRG 3 days after SNI. Wild type female mice exhibit elevated cytosolic localization of HMGB1 in Na_v1.8 expressing nociceptors. Cytosolic localization of HMGB1 after injury is decreased in both male and female TLR4 knockouts, but only females with TLR4 reactivated in their Na_v1.8 neurons see an increase that trends towards a wild type phenotype. Moreover, this same population of peripheral nociceptors in females express higher levels of neuronal injury measured by upregulated ATF3 expression. We demonstrate that this increase in ATF3 expression is linked to TLR4 expression, as both male and female TLR4 knockouts have lower expression of ATF3 after injury. ATF3 has been shown to be a negative regulator of TLR4 activity by dampening the activity of NF-κB. Although co-expression of ATF3 and HMGB1 were not performed, it is a potential that the increase in ATF3 localization within Na_v1.8⁺ neurons is due to endogenous TLR4 activation from tissue injury (Kwon et al., 2015; Rao et al., 2015). These data provide evidence that neuronally-expressed HMGB1 is involved in the onset of neuropathic pain in females, however; in future experiments it may be necessary to distinguish specific subsets of nociceptor populations (isolectin b4 (IB4) vs. calcitonin gene related peptide (CGRP)) that express HMGB1 to enhance granularity as to the specific circuitry involved. Nonetheless, this novel finding leads us to believe that females utilize more neuronally-driven mechanisms to facilitate chronic pain development.

To better understand the implications of direct neuronal activation through endogenous TLR4 signaling, we use two recently developed transgenic lines wherein TLR4 is either deleted (Na_v1.8^TLR4fl/fl) or reactivated (Na_v1.8^TLR4LoxTB) constitutively in a cre dependent manner (Jia et. Al., 2020). Evidence suggests TLR4 plays a major role in regulating the onset of neuropathic mechanical hypersensitivity, but not it’s persistence (Hu et al., 2018). Here, we investigate the nocifensive response to SNI in both male and female mice using mechanical and thermal measures. Interestingly, removal of TLR4 only on Na_v1.8⁺ nociceptors delay the onset of mechanical hypersensitivity but not cold allodynia in female mice. Additionally, re-expression of TLR4 on these same nociceptors confers a phenotype that recapitulates a wild type behavioral response to neuropathic injury only in females. Lastly, whole body knockouts of TLR4 in both sexes behave similarly, conferring necessity during the onset of neuropathic pain in both sexes. Together, these data indicate that neuronally-expressed TLR4 is both necessary and sufficient during the onset of neuropathic mechanical hypersensitivity. The involvement of TLR4 signaling in thermal hyperalgesia is unclear. TLR4 deficiency has been shown to reduce mechanical allodynia, but not thermal hyperalgesia in a model of trigeminal neuropathic pain (Hu et al., 2018). Conversely, administration of a TLR4 antagonist, lipopolysaccharide (LPS)-RS, after chronic constriction injury (CCI) reduced both mechanical allodynia and thermal hyperalgesia (Jurga et al., 2016). We report no significant differences in cold allodynia, measured by behavioral response to acetone application, throughout the entirety of our behavioral experiments, independent of both sex and genotype. Interestingly, a recent study highlights how over 80% of cold-sensitive neurons in the DRG do not express Na_v1.8 (Luiz et al., 2019). As we study the role of TLR4 in the Na_v1.8 expressing neuronal population, it is a potential that co-expression between TLR4 and cold sensitive neurons (TRPM8) is not high enough to confer biological relevant behavioral changes in a murine model of neuropathic pain.

A major unanswered question from this work is how the spontaneous ectopic activity of Nav1.8⁺ DRG neurons is affected by direct TLR4 activity following neuropathic injury. Our histological data provides some insights as to what molecular changes these neurons undergo soon after injury, however; changes in their excitability are unknown. Following SNI, there is a large increase in the spontaneous activity of DRG neurons in addition to afferent fibers in the injured sciatic nerve (Seltzer et al., 1990; Wall and Gutnick, 1974). As seen in our model and consistent with others, whole body knockouts of TLR4 exhibit attenuated pain responses following injury (Piao et al., 2018; Tanga et al., 2005). While we currently cannot pinpoint the effects of direct TLR4 activation on DRG neuron excitability, there is evidence in the brain where TLR4 activation leads to increased neuronal excitability in the dentate gyrus after injury (Li et al., 2015b). Although not directly translatable, this does provide compelling evidence that TLR4 can modulate neuronal excitability. Therefore, understanding the role of TLR4 in regulating the spontaneous activity and subsequent excitability of DRG nociceptors will yield significant insights as to how the transient changes in behavior we see evolves into long-term nociceptor plasticity. This is a research endeavor we plan on pursuing as a follow-up to the present study.

In conclusion, we have identified a novel and sexually dimorphic mechanism of neuronal injury in a murine model of neuropathic pain. Female, but not male mice utilize endogenous TLR4 activation on their Na_v1.8 expressing nociceptors to facilitate the onset of mechanical hypersensitivity following SNI. Additionally, TLR4 expression on Na_v1.8 expressing nociceptors plays an important role in regulating neuronal localization of HMGB1 in female mice. Taken together, these data indicate a unique mechanism of neuronal injury and behavioral output following peripheral injury in female mice. This work is poised as an initial endeavor into the sexually dimorphic mechanisms that regulate pain plasticity.

Table 5.

Statistical values for analyses performed within Figure 5. ATF3 localization datasets were analyzed using Ordinary Two-Way ANOVA with Tukey’s post hoc. Significance was set at p<0.05 for all datasets. Bolded values are statistically significant.

Dataset	Main Effect			Interactions		Multiple Comparisons
		F (DFn, DFd)	p-value	F (DFn, DFd)	p-value	Effect	Groups	p-value
ATF3 Localization	Nav1.8+ Neurons	Genotype: F (3, 24) = 55.51 Sex: F (1, 24) = 43.95	p<0.0001 p<0.0001	F (3, 24) = 10.03	p=0.0002	Genotype	Males: Sham vs TLR4LoxTB	p=0.4919
							Males: Sham vs Nav1.8TLR4LoxTB	p=0.3250
							Males: WT vs TLR4LoxTB	p=0.0031
							Males: WT vs Nav1.8TLR4LoxTB	p=0.0064
							Males: Nav1.8TLR4LoxTB vs TLR4LoxTB	p>0.9999
							Females: Sham vs TLR4LoxTB	p=0.1037
							Females: Sham vs Nav1.8TLR4LoxTB	p<0.0001
							Females: WT vs TLR4LoxTB	p<0.0001
							Females: WT vs Nav1.8TLR4LoxTB	p=0.9360
							Females: Nav1.8TLR4LoxTB vs TLR4LoxTB	p<0.0001
						Sex	Male TLR4LoxTB vs Female TLR4LoxTB	p=0.8948
							Male TLR4LoxTB vs Female Nav1.8TLR4LoxTB	p<0.0001
							Male Nav1.8TLR4LoxTB vs Female TLR4LoxTB	p=0.9720
							Male Nav1.8TLR4LoxTB vs Female Nav1.8TLR4LoxTB	p<0.0001
	NF200+ Neurons	Genotype: F (3, 23) = 17.49 Sex: F (1, 23) = 2.067	p<0.0001 p=0.1639	F (3,23) = 1.088	p=0.3741	Genotype	Males: Sham vs TLR4LoxTB	p=0.0035
							Males: Sham vs Nav1.8TLR4LoxTB	p=0.0402
							Males: WT vs TLR4LoxTB	p>0.9999
							Males: WT vs Nav1.8TLR4LoxTB	p=0.9166
							Males: Nav1.8TLR4LoxTB vs TLR4LoxTB	p=0.9797
							Females: Sham vs TLR4LoxTB	p=0.0355
							Females: Sham vs Nav1.8TLR4LoxTB	p=0.0007
							Females: WT vs TLR4LoxTB	p=0.1901
							Females: WT vs Nav1.8TLR4LoxTB	p=0.9888
							Females: Nav1.8TLR4LoxTB vs TLR4LoxTB	p=0.5423

Highlights:

• TLR4 on sensory neurons drive the early development of neuropathic pain in females

• ATF3 in Na_v1.8 neurons is higher in females early after neuropathic injury

• Neuronal HMGB1 activation is mediated by TLR4 expression in females after injury

• TLR4 does not contribute to the development of cold allodynia after injury

Abbreviations:

TLR

Toll-like receptor

WT

Wild-type

HMGB1

High mobility group box-1

ATF3

Activating transcription factor-3

DRG

Dorsal root ganglia

PRRs

Pattern recognition receptors

DAMPs

Danger-associated molecular patterns

SNI

Spared nerve injury

FBS

Fetal bovine serum

SEM

Standard error of the mean

ECM

Extracellular matrix

TRP

Transient receptor potential

NLRP3

NOD-like receptor protein 3

RAGE

Receptor for advanced glycation and end products

IB4

Isolectin B4

CGRP

Calcitonin gene related peptide

LPS

Lipopolysaccharide

CCI

Chronic constriction injury