TNFR1/p38αMAPK signaling in Nex+ supraspinal neurons regulates sex-specific chronic neuropathic pain

Upregulation of soluble tumor necrosis factor (sTNF) cytokine signaling through TNF receptor 1 (TNFR1) and subsequent neuronal hyperexcitability are observed in both animal models and human chronic neuropathic pain (CNP) [1–4]. To test the hypothesis that supraspinal circuitry is critical to pain chronification, we studied the intersect between supraspinal TNFR1 mediated neuronal signaling and sex specificity by selectively removing TNFR1 in Nex + neurons in adult mice (NexCreERT2::TNFR1f/f). We determined that following chronic constriction injury (CCI), pain resolves in males; however, female acute pain transitions to chronic. Subsequently, we investigated two downstream pathways, p38MAPK and NF-κB, important in TNFR1 signaling and injury response. We detected p38αMAPK and NF-κB activation in male cortical tissue; however, p38αMAPK phosphorylation was reduced in NexCreERT2::TNFR1f/f males. We observed similar behavioral results following CCI in NexCreERT2::p38αMAPKf/f mice. Previously, we established estrogen’s ability to modulate sTNF/TNFR1 signaling in CNP, which may contribute to female prevalence of CNP [5–9]. To explore the intersection between estrogen and inflammation in CNP we used a combination therapy of an estrogen receptor β (ER β) inhibitor with a sTNF/TNFR1 or general p38MAPK inhibitor. We determined both combination therapies lend “male-like” therapeutic relief to females following CCI. These data suggest that TNFR1/p38αMAPK signaling in Nex + neurons in CNP is male-specific and lack of therapeutic efficacy following sTNF inhibition in females is due to ER β interference. These studies highlight sex-specific differences in pathways important to pain chronification and elucidate potential therapeutic strategies that would be effective in both sexes.


INTRODUCTION
Chronic neuropathic pain (CNP) a icts 7-20% of adults and can result from disease or injury to the somatosensory nervous system [1,10,11].Critical mediators of persistent pain include neuroin ammation, such as increased levels of the in ammatory cytokine soluble tumor necrosis factor (sTNF), and maladaptive plasticity within the central nervous system (CNS) which extend beyond resolution of the initial insult [12].While 7-20% of adults suffer from CNP, there is greater prevalence seen in females Sorge and Totsch, 2017).Previous research has established the in uence of biochemical sex differences that contribute to this health disparity and established that sex hormones are linked to pain tolerance [7,8].Speci cally, estradiol-17β has been linked to enhanced immune reactivity, but it remains unclear whether it serves in a protective or injurious capacity for pain development [14][15][16].
In several animal models of CNP, increased production of sTNF is observed at both acute and chronic timepoints following injury [9,[17][18][19].sTNF preferentially signals through its receptor, TNFR1, which is ubiquitously expressed.Under physiological conditions, sTNF/TNFR1 signaling regulates glutamatergic receptor transmission and mediates glial production of pro-in ammatory cytokines [20][21][22].However, under pathological conditions TNFR1 signaling is associated with both initiation and maintenance of neuroin ammation in part through activation of mitogen-activated protein kinase (MAPK) transduction pathways [23,24].Interestingly, increased production of sTNF is not limited to the site of injury.After chronic constriction injury (CCI) of the sciatic nerve increased expression of sTNF is observed in both the peripheral and central nervous systems (Andrade et al., 2016).Studies in human patients re ect similar dispersion of sTNF levels where pain following a peripheral surgery corresponded with increased central sTNF levels [25].This may be a result of downstream effects resulting from increased sTNF/TNFR1 activity.
Speci cally, increased sTNF/TNFR1 signaling is associated with increased phosphorylation of p38MAPK which is linked to chronic pain states [26][27][28].Increased MAPK signaling then further stimulates sTNF production by the target cell creating a positive feedback loop exacerbating cytokine expression levels [24].p38MAPK phosphorylation is a known indicator of proin ammatory microglia and may promote prolonged in ammation in the spinal cord.Increased sTNF levels have been directly associated with increased activation of proin ammatory microglia both in the spinal cord and dorsal root ganglia of several models, including CCI [19].Together, these data would suggest that increased production of sTNF may trigger proin ammatory microglial activity via p38MAPK phosphorylation.
A strong correlation exists between the presence of CNS maladaptive plasticity, highlighted by increased excitatory signaling, and CNP.Speci cally, increased production of sTNF by proin ammatory glia can cause nearby neurons to overexpress membranous glutamate receptors Haroon et al., 2017;Parsons and Raymond, 2014).Overexcitation promotes rapid synaptic remodeling characterized by increased expression of excitatory receptors, such as AMPAR and NMDAR.Supraspinal structures undergoing such adverse changes have been linked to aversive learning and memory outcomes [31][32][33].While both microglia and neurons respond and contribute to maladaptive CNS changes, their role in CNP remains unclear.
In this study, we nd that chronic pain in males is dependent on TNFR1 signaling and p38αMAPK activation in a subset of supraspinal Nex + excitatory neurons, sometimes referred to as neuronal differentiation 6 (NeuroD6) neurons.However, when estrogen receptor β (ER β) is inhibited, females become more 'male-like' with respect to the role of sTNF in the development and treatment of chronic pain.

Tamoxifen injections
To induce Cre recombination, tamoxifen (Sigma Aldrich, T5684; 75mg/kg) was administered via intraperitoneal injection for 5 consecutive days to both mutants and oxed controls at 10-weeks-old.
Tamoxifen administration was completed 2 weeks prior to surgery to ensure recombination e cacy.

Chronic Constriction Injury (CCI)
As previously described by del Rivero et al, we performed chronic constriction injury (CCI) of the right sciatic nerve in 12-week-old mice.Prior to surgery, anesthesia was provided via intraperitoneal injection of ketamine (100 mg/kg) and xylazine (10 mg/kg).The sciatic nerve was exposed with an incision made in the mid-thigh.Once exposed, the sciatic nerve was tied with 3 silk ligatures (Oasis Silk 6 − 0, MV-711-V), spaced 1mm-1.5mmapart starting 3mm proximal to the sciatic bifurcation, and each subsequent ligature placed closer to the hip.The overlying skin was then closed with nylon sutures (Oasis Nylon 5 − 0, MV-661).

Behavioral Testing
The von Frey test was performed 3 days prior to CCI, establishing baseline mechanical sensitivity, and repeated weekly post-CCI.To perform von Frey testing, mice were placed in clear individual boxes on a wire mesh and habituated for 40 minutes prior to testing.Using an up-down method, an incremental series of 8 von Frey laments (force range: .02g-2g)were applied to the plantar surface of the hind paw, as previously described [34].For all mice, we tested both ipsilateral and contralateral hind paws.Each test consisted of 6 trials, with 5 minutes between each trial.Positive responses to a lament were noted as both hind paw withdrawal and a cognitive component, such as guarding behavior or hind paw grooming.This dual attention/response criteria allows the experimenter to discount re exive movement as a falsepositive response.We quanti ed responses according to the standard up-down analysis method (26).For behavioral testing, the contralateral hind paw served as the uninjured side and the ipsilateral hind paw served as the injured side.The experimenter remained blinded to experimental conditions/treatments until the termination of the experiment.

Drugs
Faslodex (20 mg/kg; Abcam ab120131), a clinically used general estrogen receptor inhibitor, and ERβ speci c inhibitor PHTPP (1.7mg/kg; Sigma Aldrich SML1355) were diluted in corn oil and administered systemically.Drug or vehicle (corn oil) was administered via intraperitoneal injection rst at CCI. Treatment with Xpro1595 (10 mg/kg; intraperitoneal; INmuneBio), a dominant negative sTNF inhibitor, or vehicle (0.1M PBS) began 7 days post-CCI.Previously mentioned drugs were given every 48 hours until experiment termination.SB203580, a p38 αβ MAPK inhibitor (Invivogen tlrl-sb20), was diluted to 4.54mg/mL in PBS and delivered to the CNS at 0.11µL/hr via osmotic pump (Azlet 1004) administering the drug at 0.5µg/hr with intrathecal cannula (Azlet Brain Infusion Kit3).Pumps containing SB203580 or vehicle were inserted during CCI via cannulation at the T12 vertebrae level.

2.6b Synaptosome protein extract
For synaptic preparations, mice were perfused with HEPES buffer (2mM HEPES, 0.32 M sucrose, 10 mM sodium pyrophosphate, 10 mM sodium uoride, 2 mM EGTA, 1 mM sodium orthovanadate, and 0.1 mM PMSF).Cortical tissue was isolated, frozen on dry ice and homogenized in HEPES buffer with 1 µg/mL leupeptin and 1 µg/mL aprotinin.First, the homogenate was centrifuged (3500 rpm, 4°C, 10 minutes) and the resulting supernatant was removed, and the pellet was kept as a purity control.The supernatant was centrifuged (13500 rpm, 4°C, 15 minutes) and the remaining pellet, which is the synaptosomal fraction, was resuspended in HEPES buffer while the supernatant was kept also for purity control.

Statistics
Data was analyzed and graphed on GraphPad Prism software (version 9.4.1).Western blots comparing 3 or more groups were analyzed by one-way ANOVA (α = 0.05), with each group compared to each other using Tukey's multiple comparisons test.Western blots comparing 2 groups were analyzed using a nonparametric, two-tailed Mann-Whitney t-test (α = 0.05).Changes in mechanical thresholds were evaluated by two-way ANOVA (α = 0.05), with time post-CCI and experimental group (genotype or treatment) as factors.The Geisser-Greenhouse correction was used to control for within group variability following by Tukey's multiple comparisons test for all timepoints.Values for all data are expressed as mean ± SEM.No data outliers were identi ed (ROUT test, Q = 1%).

RESULTS
3.1 TNFR1 signaling in supraspinal excitatory projection neurons is required for CNP in males, not females.
Previous work from our group demonstrated sex differences from systemic disruption of sTNF/TNFR1 signaling; speci cally global TNFR1 knockout prevented post-CCI allodynia in males while females had only partial alleviation [9].Based on these data, we hypothesized that supraspinal TNFR1 signaling is critical for processing somatosensory inputs that then drive CNP.To test this, we generated NexCre ERT2 ::TNFR1 f/f mice to interrogate the development of CNP in males and females.Nex, a neuronal basic helix-loop-helix protein, is expressed among supraspinal glutamatergic pyramidal projection neurons which project processes between the neocortex and hippocampus [35].NexCre ERT2 mice are an inducible Cre line that speci cally targets cortical and hippocampal pyramidal neurons that project between both regions [35].By selectively ablating TNFR1, which is ubiquitously expressed, among the Nex + neuronal subset, we are able determine the contribution of selective supraspinal TNFR1 signaling in excitatory neurons.TNFR1 selective deletion in these neurons prevents the development of chronic pain solely in males and no effect is observed on acute pain for either sex (Fig. 1A, B).To ensure recombination e ciency, we measured TNFR1 expression levels in the cortex (Cx), hippocampus (Hc), and spinal cord (SC).We observed a signi cant reduction in cortical and hippocampal TNFR1 expression levels for both males and females to the extent anticipated for TNFR1 loss in the Nex + neuronal subset; however, no changes were observed in the spinal cord (Fig. 1C-H; supplemental Fig. 3A-F).
3.2 p38MAPK signaling in supraspinal excitatory projection neurons is required for CNP in males, not females TNFR1 signaling is known to elicit downstream effects among both NF-κB and p38MAPK signaling pathways.Pharmacological inhibition of p38MAPK has been demonstrated to be therapeutic for CNP in a sexually dimorphic manner [36][37][38].Mechanistically, this has largely been attributed to attenuating p38MAPK-dependent microglial activation in the spinal cord of male, but not female mice [39].However, following injury we observed a signi cant reduction in cortical p38αMAPK activation selectively in male NexCre ERT2 ::TNFR1 f/f mice (Fig. 2B, E; supplemental Fig. 4C, D).Cortical NF-κB activation in males and females remained elevated in NexCre ERT2 ::TNFR1 f/f mice following injury compared to naïve controls (Fig. 2A, D; supplemental Fig. 4A, C).To determine whether p38αMAPK activation in these neurons is critical for development of chronic pain we generated NexCre ERT2 ::p38αMAPK f/f mice.As shown in Fig. 2C, F, selective deletion of p38αMAPK in supraspinal Nex + neurons prevented the development of chronic but not acute pain in male, but not female mice.To our knowledge, this is the rst genetic evidence that p38αMAPK signaling in supraspinal excitatory projection neurons is crucial for the sexspeci c effects of CNP.Whole tissue western blots of total p38αMAPK con rmed successful recombination in males and females with decreased cortical (supplemental Fig. 2A,B), but not spinal (supplemental Fig. 2C,D), p38αMAPK activation.

sTNF/TNFR1, maladaptive plasticity and chronic pain
Maladaptive plasticity within cortical areas, such as the primary somatosensory cortex (S1), has been demonstrated to play an important role in the development of CNP.In humans, the extent of pathological plasticity in S1 appears to be directly related to the sensorimotor intensity of chronic pain following peripheral or central injuries [40][41][42].Importantly, attenuation of maladaptive neuronal plasticity is associated with reductions in CNP [43].Unfortunately, the intersection of maladaptive plasticity, neuroin ammation and chronic pain has not been adequately investigated.To address this, we investigated changes in glutamate receptor expression in cortex of naïve, TNFR1 f/f and NexCre ERT2 ::TNFR1 f/f mice.We observed a robust increase in NMDAR1 expression in males following CCI in TNFR1 f/f controls, but an attenuation in the NexCre ERT2 ::TNFR1 f/f group.(Fig. 3A; supplemental Fig. 5A); however, no changes were observed in females (Fig. 3B; supplemental Fig. 5B).Previous reports have demonstrated that extra-synaptic upregulation of NMDAR1/NMDAR2B containing subunits occurs in chronic pain [44][45][46].To investigate this further, we isolated synaptic fractions and reexamined NMDAR1 expression in male TNFR1 f/f and NexCre ERT2 ::TNFR1 f/f mice and determined there was no change in synaptic NMDAR1 receptor subunits (supplemental Fig. 1).However, NMDA receptors containing NMDAR2B subunits remained elevated following CCI in both male and female TNFR1 f/f and NexCre ERT2 ::TNFR1 f/f mice (Fig. 3C, D; supplemental Fig. 5C, D).To determine if NMDAR1 expression was elevated following CCI in males due to Nex + neuronal TNFR1 signaling speci cally via p38MAPK activation, we quanti ed expression in NexCre ERT2 ::p38MAPK f/f mice.Following CCI, in males we observed a similar increase in NMDAR1 expression in p38MAPKf/f controls and decreased expression in NexCre ERT2 ::p38MAPK f/f males (Fig. 3E; supplemental Fig. 5E).Again, no changes were observed in females (Fig. 3F; supplemental Fig. 5F).These data suggest expression of NMDAR1 containing receptors are regulated through a sTNF/TNFR1 signaling nexus in a sex-dependent manner.

The role of ER β in female-typical neuropathic pain
The data presented above (Figs.1-3) and our previously published work demonstrate sex-speci c effects of inhibiting sTNF/TNFR1 and p38MAPK on maladaptive plasticity and CNP [9].Therefore, we sought to interrogate which estrogen receptors (e.g., ERα and/or ERβ) contribute to the sex-speci c effects of sTNF/TNFR1-p38MAPK signaling in CNP.In these studies, we used Xpro1595, a sTNF inhibitor, with either Faslodex (aka Fulvestrant) or PHTPP.Faslodex is an inhibitor of both ERα and ERβ used clinically to treat breast cancer and other disorders [47,48].PHTPP, on the other hand, is a highly selective ERβ antagonist (e.g., silent antagonist), with ~ 36-fold selectivity for ERβ over ERα [49].With these inhibitors, we sought to determine the relationship between estrogen and therapeutic e cacy of sTNF inhibition.Following injury, female mice were treated with vehicle, Faslodex, Xpro1595, or combinations of Faslodex and Xpro1595 as previously described [9].As shown in Fig. 4A, combination therapy of Faslodex at the time of CCI and Xpro1595 one-week post-CCI signi cantly reduced CNP in females whereas either drug alone did not.Next, we investigated the therapeutic potential of combining PHTPP with Xpro1595 or SB203580, a potent p38MAPK inhibitor.Combination therapies of PHTPP with either Xpro1595 (Fig. 4B) or SB203580 (Fig. 5A) were highly therapeutic for CNP in females, whereas individual drugs were not (Figs.4B,5A).Interestingly, combination therapies of SB203580 and PHTPP in females were equally as effective as SB203580 alone in males (Fig. 5A).These, and other data, suggest that inhibition of ERβ signaling renders females "male-like" with respect to CNP.Finally, we demonstrated that female cortical ERβ expression after CCI was similar to that of naïve levels when treated with PHTPP (Fig. 4C; supplemental Fig. 6A).However, spinal ERβ expression remained unchanged in all groups (Fig. 4D; supplemental Fig. 6B).Following PHTPP administration, cortical p38αMAPK activation appears to be reduced compared to vehicle-treated controls; although, there was no statistical signi cance (Fig. 5B; supplemental Fig. 7A).No changes were observed in the lumbar spinal cord (Fig. 5C; supplemental Fig. 7B).Mechanisms through which PHTPP may modulate ERβ expression have yet to be determined.

DISCUSSION
The contribution of biochemical sex differences to the increased prevalence of CNP amongst females and males and the role of sex hormones in relation to pain tolerance has been well established [7,8].Crucial mediators of CNP include neuroin ammation and maladaptive plasticity; however, we have yet to elucidate the intersect between these critical mediators and sex differences, such as estrogen.We demonstrate that CNP in males, but not females, is dependent on TNFR1 signaling and p38αMAPK activation in Nex + supraspinal excitatory neurons.Further, inhibition of ER β renders females "male-like" with respect to inhibiting sTNF and p38MAPK as viable therapeutic targets.Altogether, these data indicate the importance of supraspinal circuitry in chroni cation of pain and elucidate potential mechanisms through which male and female pain may be treated.

TNFR1 and p38αMAPK signaling in supraspinal excitatory projection neurons mediate CNP in a sex-speci c manner
The link between neuroin ammation and persistent neuronal hyperexcitation and their contribution to both the induction and maintenance of CNP is well established.Previously, we have demonstrated that neuroin ammatory mechanisms of male CNP are heavily mediated by TNFR1 signaling [9].TNFR1 signaling is known to activate apoptotic and pro-in ammatory pathways that include NF-κB and p38MAPK [50,51].Both pathways exacerbate in ammation and chronic glial activation that then furthers the production of cytokines and promotes increased neuronal activity [3].
Previous research has demonstrated that supraspinal signaling in cortical and hippocampal excitatory projection neurons is critical for the chroni cation of pain in males [52,53].In the brain, pyramidal neurons comprise most excitatory neurons in the cortex, hippocampus and amygdala and heavily employ NMDAR pathways [54,55].Nex + neurons are a subset of these pyramidal neurons located supraspinally.
Excitatory pathways associated with male CNP may, in part, be mediated via the TNFR1-p38αMAPK pathway where sTNF leads to increased presence of extrasynaptic NMDARs as a consequence Woods et al., 2021).In several disease and injury paradigms, extrasynaptic NMDARs are linked to cell death through pro-apoptotic pathways and are activated by glutamate via synaptic spillover or ectopic release [45,46].When TNFR1 and p38αMAPK are knocked out on Nex + neurons, cortical NMDAR1 extrasynaptic expression is reduced solely in males.However, extrasynaptic NMDAR1 expression is found in both neuronal and glial populations aiding in neuron-to-glia communication [46,56,57].Future studies are needed to determine whether the aforementioned changes in NMDAR1 are glial or neuronal in their extrasynaptic presence.
Microglial activity via p38 signaling was previously demonstrated by others to preferentially contribute to neuropathic pain in male mice [58].However, p38MAPK signaling is not restricted to spinal microglia and occurs in neuronal/non-neuronal populations located supraspinally and within dorsal root ganglion [26,59].These studies employed the use of intrathecal skepinone, a p38MAPK selective inhibitor, which may also inhibit supraspinal, spinal, and DRG neuronal/non-neuronal p38MAPK signaling (Paige et al., 2018) (56) (Paige et al., 2018).Our studies used a powerful genetic approach to determine that neuronal p38αMAPK signaling in neuropathic pain plays a critical role in pain chroni cation.While previous studies demonstrate phosphorylation of p38MAPK is more frequently associated with sTNF-associated microglial in ammation, our data support that supraspinal neuronal p38αMAPK is also a critical regulator of CNP following in ammation via peripheral injury.

ERβ regulates the behavioral phenotype of female CNP
Chronic pain is more prevalent among females, and this is observed across a spectrum of conditions [60][61][62].Mounting evidence exists that shows differential immune system response between females and males during pain, which may account for differences in immune targeted therapeutic e cacy between sexes [63,64].Estrogen receptors (ERs) are widely distributed among nociceptive regions of the CNS, including the thalamus and anterior cingulate cortex, and are known to modulate sensory processing and transduction [8, 65-68].Speci cally, estrogen and ERs are known to contribute to supraspinal neuronal maladaptive plasticity associated with chronic pain and modulate immune cell response [67,69,70].We have previously demonstrated that following CCI, pharmacological and genetic inhibition of TNFR1 signaling is selectively therapeutic in males [9].However, in ovariectomized females, inhibition of TNFR1 signaling has "male-like" therapeutic e cacy.Our ndings suggest that ER β interferes with therapeutic e cacy via TNFR1 signaling inhibition.More speci cally, combination therapy in females with ER β, sTNF, and p38MAPK inhibitors rendered them more "male-like" with regards to recovery from allodynia.Previous work has shown that ER β activation in cortical neurons increases the density of dendritic spines as well as clustering of postsynaptic density-95 at the membrane [71].Together, these data suggest that cortical ER β may modulate post-synaptic excitatory signaling and, in turn, enhance glutamate-mediated excitotoxicity.Inhibition of sTNF/TNFR1 signaling pathways reduces excitatory signaling; however, ER β may enhance cortical excitotoxicity.By blocking ER β in females, we reduce this enhanced excitotoxicity and allow females to be more amenable to anti-TNF therapies.Future studies are needed to determine e cacy for other pain categories such as thermal and spontaneous pain.

CONCLUSIONS
In summary, our data demonstrate that supraspinal neuronal TNFR1/p38αMAPK signaling is critical in chroni cation of pain with male speci city.When TNFR1/p38αMAPK signaling is selectively knocked out on supraspinal excitatory neurons, the transition from acute to chronic pain is avoided in males.Furthermore, when ER β is inhibited in females and sTNF and p38MAPK inhibitors are administered, pain does not transition from an acute to chronic state.With these data we propose that supraspinal neuronal TNFR1/p38αMAPK signaling is critical to pain chroni cation in a sex-speci c manner but poses a novel therapeutic route for treatment of CNP.Declarations 6.1 Ethics approval and consent to participate

Figure 4 Inhibition
Figure 4 Mice were acclimated to the colony housing room for 2 weeks prior to experiments.All animal-use experimental protocols were carried out with approval of Drexel University's Institutional Animal Care and Use Committee (#20295, #20895) and in accordance with the United States Public Health Service's Policy on Humane Care and Use of Laboratory Animals.
Animals were housed in a 12/12 h light/dark cycle room and given food and water ad libitum.