Cytohesin-2 mediates group I metabotropic glutamate receptor-dependent mechanical allodynia through the activation of ADP ribosylation factor 6 in the spinal cord

Group I metabotropic glutamate receptors (mGluRs), mGluR1 and mGluR5, in the spinal cord are implicated in nociceptive transmission and plasticity through G protein-mediated second messenger cascades leading to the activation of various protein kinases such as extracellular signal-regulated kinase (ERK). In this study, we demonstrated that cytohesin-2, a guanine nucleotide exchange factor for ADP ribosylation factors (Arfs), is abundantly expressed in subsets of excitatory interneurons and projection neurons in the superficial dorsal horn. Cytohesin-2 is enriched in the perisynapse on the postsynaptic membrane of dorsal horn neurons and forms a protein complex with mGluR5 in the spinal cord. Central nervous system-specific cytohesin-2 conditional knockout mice exhibited reduced mechanical allodynia in inflammatory and neuropathic pain models. Pharmacological blockade of cytohesin catalytic activity with SecinH3 similarly reduced mechanical allodynia and inhibited the spinal activation of Arf6, but not Arf1, in both pain models. Furthermore, cytohesin-2 conditional knockout mice exhibited reduced mechanical allodynia and ERK1/2 activation following the pharmacological activation of spinal mGluR1/5 with 3,5-dihydroxylphenylglycine (DHPG). The present study suggests that cytothesin-2 is functionally associated with mGluR5 during the development of mechanical allodynia through the activation of Arf6 in spinal dorsal horn neurons.


Introduction
The spinal cord dorsal horn is the first relay center where dorsal horn neurons form excitatory synapses with axon terminals of primary sensory neurons carrying nociceptive information from peripheral tissues. Additionally, it is also the principle integration center where nociceptive information is actively processed and modulated through elaborate neural networks, depending on the strength and duration of nociceptive stimuli (Peirs and Seal, 2016;Todd, 2010). Intense and persistent noxious stimuli caused by inflammation and nerve injury induce plastic changes in neurons and neural circuitry in the dorsal horn, leading to increased excitability of nociceptive transmission, a condition referred to as central sensitization (Basbaum et al., 2009;Latremoliere and Woolf, 2009;Woolf, 2011). It is clinically related to chronic pain with manifestations, such as spontaneous pain, augmented pain responses to noxious stimuli (hyperalgesia), and pain generated by low-threshold innocuous stimuli (allodynia).
In the spinal cord, glutamate mediates excitatory neurotransmission, synaptic plasticity, and central sensitization through two classes of receptors: ionotropic and metabotropic glutamate receptors (Bleakman et al., 2006;Liu and Salter, 2010). Ionotropic glutamate receptors mediate fast excitatory synaptic transmission, and are composed of the α-amino-3-hydroxy-5-methyl-4-isoxazolepropionate-type receptors (AMPARs), kainate-type receptors, and N-methyl-D-aspartate-type receptors (NMDARs). Metabotropic glutamate receptors (mGluRs) modulate neurotransmission through second messenger cascades via Gproteins, and are composed of group I (mGluR1 and mGluR5), group II (mGluR2 and mGluR3), and group III (mGluR4, mGluR6, mGluR7, and mGluR8). Among these, group I mGluRs are coupled to the Gq family heterotrimeric G protein, which stimulates phospholipase C-mediated hydrolysis of phosphatidylinositol 4,5-bisphosphate to diacylglycerol and inositol 1,4,5-trisphosphate, thereby leading to intracellular calcium mobilization and protein kinase C (PKC) activation. They are abundantly expressed in the spinal cord, with mGluR5 predominantly expressed in the superficial dorsal horn (laminae I and II) and mGluR1α in the deeper dorsal horn (laminae III -V) and the ventral horn (laminae VI -IX) (Alvarez et al., 2000). In dorsal horn neurons, mGluR5 is enriched preferentially in the marginal region of the postsynaptic density (PSD) of excitatory synapses, and are ideally positioned to respond only to persistent intense stimulation (Vidnyanszky et al., 1994). Since it was first reported that intrathecal administration of a selective group I mGluR agonist, 3,5-dihydroxylphenylglycine (DHPG), induces spontaneous nociceptive behaviors (Fisher and Coderre, 1996), evidence has accumulated that group I mGluRs play a pivotal role in central sensitization in various types of inflammatory and neuropathic pain models by modulating ion channels such as NMDARs and K + channels, and inducing gene transcription through multiple signaling pathways, including PKC, Src, and extracellular signal-regulated kinase 1/2 (ERK1/2) (Chiechio and Nicoletti, 2012;Guo et al., 2004;Hu et al., 2007;Karim et al., 2001;Kawasaki et al., 2004;Pereira and Goudet, 2019;Vincent et al., 2016;Vincent et al., 2017).
ADP ribosylation factors (Arfs) are small GTP-binding proteins that mediate membrane trafficking, actin cytoskeleton remodeling, and lipid metabolism/signaling (D'Souza-Schorey and Chavrier, 2006;Donaldson and Jackson, 2011;Gillingham and Munro, 2007). There are six mammalian Arf proteins, which are divided into three classes, based on their structural similarity: class I (Arf1, Arf2, and Arf3), class II (Arf4 and Arf5), and class III (Arf6). Both class I and II Arfs are localized primarily in the Golgi complex and have partially redundant functions in membrane trafficking within the Golgi complex. On the other hand, Arf6 is localized to the plasma membrane and endosomes, and regulates vesicular transport between the plasma membrane and endosomes, as well as the reorganization of the actin cytoskeleton beneath the plasma membrane. Like other small GTP-binding proteins, Arfs function as binary molecular switches by cycling between GTP-bound active and GDPbound inactive states. The proper timing, duration and location of the activation of Arfs are tightly controlled by two regulatory proteins: guanine nucleotide exchange factors (GEFs) that catalyze the exchange of GDP for GTP, and GTPase-activating proteins that stimulate GTP hydrolysis (Donaldson and Jackson, 2011;Gillingham and Munro, 2007;Sztul et al., 2019). All mammalian Arf-GEFs share a conserved catalytic domain of approximately 200 amino acids, known as the Sec7 domain, and are classified into seven subfamilies, i.e., the GBF1 (Golgi-specific brefeldin A-resistance factor 1), BIG (brefeldin A-inhibited ArfGEF), cytohesin, EFA6 (exchange factor for Arf6)/Psd (PH and Sec7 domaincontaining protein), BRAG (brefeldin A-resistant Arf-GEF)/IQSEC (IQ and Sec7 domain-containing), and FBXO8 (F-box only protein 8) subfamilies, based on their structural similarity and sensitivity to brefeldin A, a fungal metabolite secreted by Eupenicillium brefeldianum (Casanova, 2007;Cox et al., 2004;Gillingham and Munro, 2007).
Cytohesin-2, also known as ARNO (Arf nucleotide-binding site opener), is a member of the cytohesin Arf-GEF subfamily with guanine nucleotide exchange activity toward Arf1 and Arf6 in vitro (Chardin et al., 1996;Frank et al., 1998a;Kolanus, 2007). In the central nervous system (CNS), cytohesin-2 is abundantly expressed in developing and mature neurons (Ito et al., 2018), and is implicated in Arf6-dependent neuronal processes, including the formation of axons and dendrites of developing hippocampal neurons (Hernandez-Deviez et al., 2002;Hernandez-Deviez et al., 2004), and the pathfinding of commissural axons in the developing spinal cord during midline crossing (Kinoshita-Kawada et al., 2019). Cytohesin-2 has previously been shown to form a protein complex with group I mGluRs in the rat brain through the interaction with tamalin, also known as GRASP (GRP1-associated scaffold protein), a PDZ (postsynaptic density protein-95/discs large/zonula occludens-1) domain-containing scaffolding protein enriched in the PSD fraction (Kitano et al., 2002), suggesting the possible functional relationship between cytohesin-2 and mGluR1/5. Although a recent discovery of SecinH3 (Hafner et al., 2006), a cell-permeable selective inhibitor of the cytohesin family, has revealed the functional involvement of cytohesins in various pathological processes related to neurological disorders, such as amyotrophic lateral sclerosis (Hu et al., 2019;Zhai et al., 2015) and Alzheimer's disease (Yan et al., 2016), it remains unknown whether cytohesin-2 has a functional involvement in chronic pain.
In this study, we investigated the functional role of cytohesin-2 in the spinal cord, with special attention to its relationship with mGluR5. We first demonstrated that cytohesin-2 is expressed in subsets of excitatory interneurons and projection neurons in the dorsal horn, and forms a protein complex with mGluR5 in the spinal cord. We then demonstrated that CNS-specific cytohesin-2 conditional knockout (cKO) mice reduced mechanical allodynia in inflammatory and neuropathic pain models of mice. Pharmacological blockade of cytohesin catalytic activity with SecinH3 in the spinal cord reduced mechanical allodynia and activation of Arf6, but not Arf1, in both pain models. Furthermore, cytohesin-2 cKO mice exhibited marked reduction in mechanical allodynia and ERK1/2 activation after stimulation of spinal mGluR1/5 with DHPG. The present findings provide the first evidence for the involvement of the cytohesin-2-Arf6 pathway in inflammatory and neuropathic pain in spinal dorsal horn neurons.

Animals
All experimental protocols involving animals were approved by the Animal Experimentation and Ethics Committee of the Kitasato University School of Medicine. All efforts were made to reduce the number of animals used. To generate the CNS-specific cytohesin-2 cKO mice, we prepared the floxed cytohesin-2 mouse line carrying two loxP sites flanking the exons 6 and 7 (fCYTH2) (Torii et al., 2015) and the transgenic mouse line carrying Cre recombinase under the control of the nestin promoter (nestin-Cre, Stock No. 003771, Jackson Laboratory, Bar Harbor, ME). The founder mice (Cre/− , fCYTH2/fCYTH2) were crossed to homozygous fCYTH2 mice (− /− , fCYTH2/fCYTH2) to produce cKO (Cre/− , fCYTH2/fCYTH2) and control (− /− , fCYTH2/fCYTH2) littermates. For evaluation of the floxed cytohesin-2 allele, the PCR primers used were 5 ′ -TCAGGAAATGTCTCTCAAATAAGA-3 ′ (sense) and 5 ′ -AAATCTCTGCTCCAACTGTAGCTT-3 ′ (antisense). Floxed and wild-type cytohesin-2 alleles displayed ~390 bps and ~ 240 bps, respectively. The PCR primers used to identify the cre transgene were 5 ′ -TTTGCCTGCATTACCGGTCGATGCAAC-3 ′ (sense) and 5 ′ -GCGCGAGTT-GATAGCTGGCTGGTG-3 ′ (antisense). The PCR product was ~750 bps. Ninety male wild-type C57BL/6 mice (CLEA Japan) in the 8th to 12th postnatal weeks were used for pain models, eight for immunofluorescent staining, six for immunoelectron microscopic analysis, and six for immunoprecipitation assay. For analyses using cytohesin-2 cKO and control mice, 40 male mice of each genotype were used for pain models, 15 for DHPG models, 6 for immunoblotting, and 4 for immunoperoxidase analyses. Eight male nestin-Cre mice were included for pain models to exclude the possibility about differences in nociceptive behaviors between wild-type and nestin-Cre mice ( Supplementary Fig. S2). Mice (5 per cage) were housed and maintained in the air-conditioned room (about 20-24 • C) with a 12 h light-dark cycle with free access to food and water. Prior to the experiments, mice were anesthetized with sodium pentobarbital (50 mg/kg of body weight, i.p.) or inhalation of 2-3% isoflurane and assessed the anesthesia level by pedal reflex in order to minimize animals' suffering. Mice were killed by decapitation under deep anesthesia with sodium pentobarbital (100 mg/kg of body weight, i.p.) or the inhalation anesthesia with 4-5% isoflurane for the sampling of fresh brains and spinal cords.

Antibodies and antigens
The antibodies used in the present study are summarized in Table 1. To prepare antigens to produce antibodies against cytohesin-2 (N7) and Homer1b/c, the N-terminal region (amino acids 1-70) of mouse cytohesin-2 and C-terminal 215-amino acids (152-366) of mouse Homer1c, a region shared by Homer1b, were amplified by PCR using a mouse brain first-strand cDNA library and the following PCR primers, in which a SalI restriction enzyme site (underlined) and stop codon (small letters) were added to the 5 ′ end of forward and reverse primers, respectively: 5 ′ -GTCGACCATGGAGGACGGTGTCTACGAGCCC-3 ′ and 5'-ttaCTTCCTGCCCATTGCCATCTTCCG − 3 ′ for cytohesin-2 (1-70); 5 ′ -GTCGACTGCCAATGTGAAGCAGTGGAAGCAAC -3 ′ and 5'-ttaGCTG-CATTCCAGTAGCTTGGC -3 ′ for Homer1c (152-366). To prepare antigens for pre-absorption test in the post-embedding immunoelectron microscopic analysis, we also amplified cDNA fragments encoding Cterminal 28-amino acids of mouse mGluR5 and C-terminal 300-amino acids of human Gα11 from mouse brain and HeLa cell first-strand cDNA libraries, respectively, by PCR with the following primers: 5 ′ -GTCGACCTCCGAATCGGCCCTCTG − 3 ′ and 5 ′ -tcaCAACGATGAA-GAACTC -3 ′ for mGluR5 (1144-1171); 5 ′ -GTCGACGCGCATCATC-CACGGCGC -3 ′ and 5 ′ -ttaGACCAGGTTGTACTC -3 ′ for Gα11 (60-359). PCR products were subcloned into a pGEM-Teasy vector (Promega, Madison, WI). The inserts were then digested with SalI and NotI restriction enzymes and subcloned into the SalI and NotI sites of pGEX-4T-2 (Promega) and a modified pMAL-2c vector (New England BioLabs, Beverly, MA) that carries SalI and NotI restriction enzyme sites in the same reading frame as pGEX-4T-2. The antigens were bacterially expressed as fusion proteins of glutathione S-transferase (GST) or maltose-binding protein (MBP) in the presence of 0.1 mM isopropyl-β-Dthiogalactopyranoside at 20 • C overnight. GST-and MBP-fusion proteins were purified with glutathione-Sepharose 4B (GE Healthcare, Piscataway, NJ) and amylose-resin (New England Biolabs), respectively. The GST-cytohesin-2 (1-70) and GST-Homer1c (152-366) fusion proteins were emulsified with Freund's adjuvant and injected subcutaneously into guinea pigs or rabbits five times at 2-week intervals, and specific antibodies were purified from antisera by respective MBP-fusion proteins coupled to CNBr-activated Sepharose 4B beads (GE Healthcare). The specificity of anti-cytohesin-2 (N7) antibody was characterized by immunoblotting, in which a single band of 45-kDa detected by this antibody in lysates of the spinal cord and brain from control mice was markedly attenuated in those from cytohesin-2 cKO mice, consistent with the results obtained with the previously characterized anticytohesin-2 antibody (C2) raised against the C-terminal region (Ito et al., 2018) (Fig. 4A). The novel anti-Homer1b/c antibody detected a single 43-kDa immunoreactive band in the lysates of the mouse brain and HeLa cells transfected with pCMV-HA vector carrying Homer1c, but not in the naïve HeLa cell lysate ( Supplementary Fig. S1A). Furthermore, double immunofluorescence revealed that it produced almost completely overlapping immunolabeling in the spinal dorsal horn with a previously characterized anti-Homer1 antibody (Gutierrez-Mecinas et al., 2016) ( Supplementary Fig. S1B, C).
Post-embedding immunogold electron microscopy was performed as described previously (Fukaya et al., 2003) with minor modifications.
Microslicer sections of the spinal cord (400 μm in thickness) were cryoprotected with 30% sucrose in 0.1 M PB and frozen rapidly with liquid propane in a Leica EM CPC unit. Frozen sections were immersed in 0.5% uranyl acetate in methanol at − 90 • C in a Leica AFS freeze-substitution unit, infiltrated at − 45 • C with Lowicryl HM-20 resin (Lowi, Waldkraiburg, Germany), and polymerized with UV light. Ultrathin sections on nickel grids were treated successively with 1% human serum albumin (Wako, Osaka, Japan)/0.1% Tween 20 in Tris-buffered saline (HTBST, pH 7.5) for 1 h, primary antibodies (5 μg/mL for each) in HTBST overnight, and 10-nm colloidal gold-conjugated antibodies (1:100, British Bio Cell International, Cardiff, UK) in HTBST for 2 h. Finally, sections were stained with 2% uranyl acetate solution for 15 min and 1% lead citrate solution for 1 min, and examined using an electron microscope (H-7650, Hitachi). For quantitative analysis, the lateral distribution of cytohesin-2, mGluR5, and Gαq/11/14 at axo-dendritic asymmetric synapses in synaptic glomeruli of laminae I-II of the spinal dorsal horn were evaluated by measurement of the length from the center of gold particles to the edge of PSD on electron microscopic images using Image J software (NIH). The quantitative data of the postembedding immunogold analysis were obtained from more than 180 synapses from 3 mice for each genotype. For antigen pre-absorption test, ultrathin sections were treated with anti-mGluR5 and anti-Gαq/11/14 antibodies (5 μg/mL) in HTBST supplemented with the GST-fusion proteins of mGluR5 (1144-1171) and Gα (60-359) (20 μg/mL), respectively ( Supplementary Fig. S3).

Immunoprecipitation
Immunoprecipitation was performed as described previously (Fukaya et al., 2020). The dorsal half of the mouse lumbar spinal cord was homogenized in a buffer consisting of 10 mM Tris-HCl (pH 7.4), 320 mM sucrose, 10 mM EDTA, 10 mM EGTA, and a cocktail of protease inhibitors (11,697,498,001,Roche). After the nuclear and mitochondrial fractions were removed, the samples (P2 fraction, 200 μg) were lysed with a binding buffer consisting of 50 mM Tris-HCl (pH 7.4) and 1% Triton X-100 for 1 h at 4 • C, and immunoprecipitated with rabbit anti-cytohesin-2 (N7) IgG-or rabbit anti-mGluR5 IgG-conjugated magnetic beads (Dynabeads, DB10004, Thermo Fisher, Waltham, MA). The immunoprecipitates were washed extensively with the binding buffer. The immunoprecipitates and lysates were subjected to immunoblotting with antibodies against cytohesin-2 (C2, guinea pig) or mGluR5 (goat).

Arf activation assay
Arf activation assay was performed as described previously (Nakayama and Takatsu, 2005). Under deep anesthesia, the freshly dissected dorsal half of the lumbar spinal cord for each condition was homogenized with the lysis buffer containing 25 mM HEPES, 150 mM NaCl, 1% NP-40, 10 mM MgCl 2 , 1 mM EDTA, and 2% glycerol. After centrifugation for 5 min at 1000 xg, the supernatants were incubated with GST-Golgi-localized, γ-adaptin ear-containing, Arf-binding protein 1 (GGA1) fusion protein immobilized on glutathione-Sepharose 4B (GE Healthcare) for 1 h at 4 • C to bind active GTP-bound forms of Arf1 and Arf6. The beads were washed 4 times with the lysis buffer. The beads were lysed with SDS-PAGE sampling buffer containing 0.1 M DTT, and incubated for 15 min at 65 • C. The samples were electrophoretically separated on a 12.5% SDS-PAGE and transferred onto PVDF membranes. After blocking with 5% non-fat dry milk in PBS containing 0.1% Tween 20, membranes were incubated with anti-Arf1 IgG (1.0 μg/mL, rabbit, GeneTex) or anti-Arf6 IgG (1.0 μg/mL, rabbit) overnight and then with a horseradish peroxidase-linked secondary antibody (1: 10000, Thermo Fisher) for 2 h. Immunoreactive bands were visualized using a chemiluminescent detection kit (ECL-PLUS) and an image analyzer (Image-Quant LAS4000, GE Healthcare). ImageJ (NIH) was used to measure the densities of immunoreactive bands. Each immunoreactive intensity of GTP-Arf1 and GTP-Arf6 was normalized by that of total Arf1 and Arf6, respectively.

Pain models
For an inflammatory pain model, inflammation was induced by a subcutaneous injection of 10 μL of complete Freund's adjuvant (CFA; Sigma-Aldrich) into the right hind paw under anesthesia with isoflurane (2-3%). For the control, mice were injected with sterile saline. For a neuropathic pain model, chronic constriction nerve injury (CCI) model was performed as described previously (Bennett and Xie, 1988;Seltzer et al., 1990). Briefly, under anesthesia with isoflurane, mice had skin incision just below the right hip bone, parallel to the sciatic nerve. The right sciatic nerve was exposed by separating the fasciae of biceps femoris and gluteus maximus muscles. The injury was performed by the three loose ligations of one-half of sciatic nerve. The nerve was tied loosely with 6-0 silk at 1-mm spacing, until a brief twitch in the respective hind limb was observed.

Intrathecal administration
Mice were anesthetized with inhalation of 2-3% isoflurane and the skin of the lower back was incised to visualize the lumbar region. Intrathecal injection of the cytohesin inhibitor SecinH3 (2849, Toccris Bioscience, Minneapolis, MN) or group I mGluR agonist DHPG was performed by puncture between L5 and L6 vertebrae, which was confirmed to be successful by the observation that it evoked a tail-flick response. Control mice were injected with 10 μL of the artificial cerebrospinal fluid (ACSF). After the administration of SecinH3, mice were subjected to inflammatory or CCI pain models.

Behavioral analysis
Mechanical pain sensitivity was assessed by measuring the withdrawal threshold of the paw in response to mechanical stimuli using von Frey filaments (TACTILE TEST AESTHESIO, Muromachi Kikai, Tokyo). Mice were placed in a plastic cage with a mesh floor, and allowed to habituate over 20 min. The filaments were applied perpendicularly to the midplantar surface of the hind paw for 5 s. The response to the filament was considered to be positive when mice immediately withdrew or licked the hind paw. Each filament was applied 10 times at an interval of 30 s. Measurement was started from the smallest filament in ascending order, and the smallest filament eliciting response over six times was considered as the threshold stimulus. The strength of the von Frey stimuli ranged from 0.008 to 2.0 g. To evaluate the motor performance, a simple accelerating rotarod test was applied using a rotarod apparatus (O'hara, Tokyo, Japan) programmed to accelerate from 4 rpm to 40 rpm over 2 min and then remain constant speed at 40 rpm for 1 min. Three male mice of each genotype were placed on the rod rotating at 4 rpm and underwent three trials with a 30-min interval. The latency until mice fell off the rod was recorded.

Colocalization assay
To evaluate the colocalization between cytohesin-2 and synaptic or endosomal markers in double immunofluorescence, the colocalization coefficient, which represents the proportion of pixels colabeled with two fluorophores in all pixels labeled with a fluorophore, was measured using ZEN software (Manders split coefficient, Carl Zeiss). Colocalization coefficient data were obtained from more than 1000 μm 2 of neuropil region for each combination.

Statistical analysis
For quantitative comparison, Student's t-test was used for comparisons between the two evaluations within each group. Statistical significance was set at p < 0.05. For multiple comparisons, the one-factor analysis of variance (ANOVA) with Tukey-Kramer post hoc test and two-factor repeated measures ANOVA with post hoc test were used for Arf activation and von Frey threshold analyses, respectively. Data are presented as mean ± SD or SEM. Statistical analyses were performed using the StatView software 5.0 (SAS Institute Inc., Cary, NC).

Perisynaptic and endolysosomal localization of cytohesin-2 in dorsal horn neurons
To examine the expression of cytohesin-2 at the protein level in the spinal cord, we first performed immunohistochemical analyses using a specific antibody against cytohesin-2 (N7) raised in this study. In the lumbar spinal cord of wild-type adult male mice, the immunofluorescence for cytohesin-2 was distributed throughout the gray matter, especially enriched in the dorsal horn (Fig. 1A), which was completely abolished in cytohesin-2 cKO mice (Fig. 1B). Triple staining of the dorsal horn revealed that cytohesin-2 was concentrated particularly in laminae I and II, which were marked by substance P and calretinin, respectively (Fig. 1C).
We further examined the subcellular localization of cytohesin-2 in dorsal horn neurons at the ultrastructural level by immunoelectron microscopy. In pre-embedding immunoelectron microscopy, silverintensified immunoreactive metal particles for cytohesin-2 were distributed in various subcellular compartments, including cell bodies, axons, presynaptic terminals, and dendritic shafts. In soma and dendritic shafts, immunoreactive metal particles were frequently associated with vesicular or tubular membrane structures and with the plasma membrane (Fig. 2G). Quantitative analysis revealed that the density of immunoreactive particles was higher in somatodendritic compartments than in axonal compartments, with the highest density in the dendritic area adjacent to synapses (Fig. 2H). Post-embedding immunoelectron microscopy of axo-dendritic asymmetric synapses of glomeruli in the superficial dorsal horn revealed that immunogold particles for cytohesin-2 were distributed on the postsynaptic membrane with preferential accumulation around the edge of the PSD (Fig. 2I). In contrast, the immunolabeling was markedly attenuated in the same compartment of cytohesin-2 cKO mice (Fig. 2J, K; gold particles/μm 2 , mean ± SD; control: 6.268 ± 1.022; cytohesin-2-cKO: 0.520 ± 0.229, p = 0.00068, ttest, n = 3 mice for each), suggesting the specificity of the immunolabeling. A quantitative analysis of the lateral distribution of cytohesin-2 on the postsynaptic membrane revealed that cytohesin-2 was distributed most densely within 60 nm around the edge of the PSD, and gradually decreased with increasing distance from the PSD, suggesting the preferential perisynaptic localization of cytohesin-2 at excitatory postsynapses (Fig. 2L).

Cytohesin-2 forms a protein complex with mGluR5 in the spinal cord
Among group I mGluRs, mGluR5 is expressed predominantly in the superficial dorsal horn and localizes to the perisynapse of asymmetric synapses of dorsal horn neurons (Alvarez et al., 2000;Vidnyanszky et al., 1994). Furthermore, mGluR1/5 was shown to form a protein complex with cytohesin-2 through its interaction with tamalin, a synaptic scaffolding protein, in the rat brain (Kitano et al., 2002). Indeed, double immunofluorescence staining of the dorsal horn revealed a high degree of colocalization of immunoreactive puncta for cytohesin-2 with those for mGluR5 and Gαq/11/14, the cognate Gα protein for mGluR5 (Fig. 3A, B, Table 2). In the immunoprecipitation assay (Fig. 3C), the anti-mGluR5 antibody immunoprecipitated cytohesin-2 from the P2 fraction of the dorsal half of lumbar cord, but the reverse immunoprecipitation with the anti-cytohesin-2 antibody failed to detect mGluR5. These findings suggest that a subpopulation of cytohesin-2 forms a protein complex with mGluR5 in the spinal cord, consistent with the anatomical findings that cytohesin-2 exhibited wide subcellular distribution in dorsal horn neurons. Nuclear staining with DAPI (blue) is included in the images. Note that the immunolabeling for cytohesin-2 observed in the lumbar spinal cord of the wild-type mouse was completely attenuated in that of the cytohesin-2 conditional knockout mouse. (C) Triple immunofluorescence of the dorsal horn with antibodies against cytohesin-2 (green, C 1 ), calretinin (red, C 2 ), and substance P (SP) (blue, C 2 ), showing the prominent immunolabeling for cytohesin-2 in laminae I and II. (D-K) Double immunofluorescence staining of the dorsal horn with antibodies against cytohesin-2 (D 1 -K 1 ) and NeuN (D 2 , E 2 ), NK1R (F 2 , G 2 ), calbindin (H 2 ), calretinin (I 2 ), PKCγ (J 2 ), or Pax2 (K 2 ). Note the expression of cytohesin-2 in subsets of excitatory interneurons and projection neurons, but not in Pax2-positive inhibitory interneurons. Arrowheads point to neurons co-immunolabeled with cytohesin-2 and markers. Arrows in E indicate nuclear immunolabeling for cytohesin-2 in some

Cytohesin-2 cKO mice show reduced mechanical allodynia in inflammatory and neuropathic pain models
Group I mGluRs are implicated in nociceptive transmission and plasticity in the spinal cord dorsal horn (Chiechio and Nicoletti, 2012;Luo et al., 2014;Neugebauer, 2002). We thus examined the functional involvement of cytohesin-2 in inflammatory and neuropathic pain models using cytohesin-2-deficient mice. The conditional deletion of cytohesin-2 in Schwann cells has previously been shown to affect the myelination of the peripheral nervous system (PNS) (Torii et al., 2015). To avoid the effect of cytohesin-2 deletion on the PNS, we generated CNS-specific cytohesin-2 cKO mice by crossing floxed cytohesin-2 knock-in mice with nestin-Cre transgenic mice.
Immunoblotting using anti-cytohesin-2 antibodies raised against its divergent N-terminal or C-terminal region showed a marked reduction in a 45-kDa immunoreactive band for cytohesin-2 in total lysates of the spinal cord and brain from cytohesin-2 cKO mice, without any truncated products, compared with those from wild-type control mice (Fig. 4A). A faint residual immunoreactive band for cytohesin-2 in the spinal lysate from cytohesin-2 cKO mice was likely to reflect the expression of cytohesin-2 in primary afferents of dorsal root ganglion neurons entering the spinal cord, in spinal neurons and/or glial cells that escaped the targeted gene deletion due to inhomogeneous recombination or Cre recombinase expression, or in endothelial and/or blood cells that did not express Cre recombinase under the nestin promoter. In support of the first possibility, the primary sensory neurons in the dorsal root ganglia of both genotypes were similarly immunoreactive for cytohesin-2 ( Fig. 4G-J). Peroxidase immunostaining of the brain and spinal cord revealed that the immunolabeling for cytohesin-2 observed in control mice was almost completely attenuated in cytohesin-2 cKO mice (Fig. 4B, C, D). Nissl staining failed to detect any apparent differences in the gross anatomical structure of the spinal cord in terms of the  F 1 ) and Homer1b/c (A 2 ), VGluT2 (B 2 ), EEA1 (C 2 ), syntaxin13 (D 2 ), LAMP2 (E 2 ), or Arf6 (F 2 ). Merged images in C 3 -E 3 include nuclear staining with DAPI (blue). Note the partial overlapping of cytohesin-2-immunoreactive puncta with Arf6 and markers for excitatory postsynapses, endosomes, or lysosomes. (G) Pre-embedding immunoelectron microscopy showing the association of immunoreactive metal particles for cytohesin-2 with membrane vesicles and tubules (arrows) in the dendritic shaft (Dn). (H) Quantification of the density of immunoreactive metal particles for cytohesin-2 in neuronal subcellular compartments, showing the highest density of immunolabeling for cytohesin-2 in the dendritic area adjacent to excitatory synapses (Dn(Sy)). Ax, axonal shaft; CB, cell body; Dn(Sh), dendritic shaft region without synapses; NT, nerve terminal. (I, J) Post-embedding immunoelectron microscopy of the postsynaptic dendritic membrane of asymmetric synapses in glomeruli of control (I) and cytohesin-2 conditional knockout (cKO) (J) mice using anti-cytohesin-2 antibody. Arrowheads point to immunolabeling for cytohesin-2 along the postsynaptic membrane. (K) Quantification of the labeling density for cytohesin-2 in dorsal horn neurons of control and cytohesin-2 cKO mice (gold particles/μm 2 , mean ± SD; control: 6.268 ± 1.022; cytohesin-2-cKO: 0.520 ± 0.229, *p = 0.00068, t-test, n = 3 mice for each). Note significant attenuation of immunolabeling in cytohesin-2 cKO mice. (L) Quantification of the lateral distribution of immunogold particles for cytohesin-2 from the edge of the postsynaptic density (PSD) along the postsynaptic membrane of asymmetric synapses of glomeruli in the superficial dorsal horn, showing the highest density of immunolabeling for cytohesin-2 in the perisynaptic region. Mi, mitochondria. Scale bars: 10 μm in (A 3 -F 3 ); 500 nm in (G, I, J). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) cellularity and laminar structure between control and cKO mice (Fig. 4E, F). Furthermore, immunostaining using an anti-MBP antibody failed to reveal an apparent difference in myelination in the spinal cord and dorsal roots between the two genotypes ( Fig. 4K-N). Together, these findings confirmed the specific deletion of cytohesin-2 in the CNS, without apparent anatomical defects in the spinal cord.
We compared the nociceptive sensitivity to mechanical stimuli under the naïve condition between control and cytohesin-2 cKO male mice aged 8-12 weeks. In the von Frey test, the paw-withdrawal thresholds to mechanical stimuli were comparable between the two genotypes ( Fig. 5A; mean ± SEM; control: 0.76 ± 0.20 g; cytohesin-2 cKO: 0.82 ± 0.16 g, p = 0.58, t-test, n = 9 mice for each group), suggesting that deletion of cytohesin-2 in the CNS did not affect basal nociceptive sensitivity to mechanical stimuli. In addition, the rotarod test revealed that there were no significant differences in motor performance between the two genotypes (latency to fall, mean ± SEM; control: 79.66 ± 1.50 s; cytohesin-2 cKO: 68.00 ± 7.34 s, p = 0.097, t-test), excluding the possibility that nociceptive responses to mechanical stimuli were affected by differences in motor performance between the two genotypes.
We also examined the effect of cytohesin-2 deletion on neuropathic pain responses using the CCI model of the sciatic nerve. Following the loose ligation of one-third to one-half of the sciatic nerve, control mice developed mechanical allodynia in the ipsilateral hind paw at 2 days, which was sustained for at least 7 days (Fig. 5B). In contrast, cytohesin-2 cKO mice showed a significantly higher ipsilateral paw-withdrawal threshold to mechanical stimuli at 2 days post-operation compared with the control (Fig. 5B, Table 3; two-way repeated measures ANOVA, F(3, 117) = 2.953, p = 0.035, followed by Tukey-Kramer post hoc test, mean ± SEM, day 0, Control-Ipsi: 0.81 ± 0.07 g, cKO-Ipsi: 0.87 ± 0.13 g, p > 0.05; day 2, Control-Ipsi: 0.09 ± 0.03 g, cKO-Ipsi: 0.29 ± 0.07 g, p < 0.05; day 4, Control-Ipsi: 0.10 ± 0.03 g, cKO-Ipsi: 0.18 ± 0.05 g, p > 0.05; day 7, Control-Ipsi: 0.19 ± 0.05 g, cKO-Ipsi: 0.34 ± 0.08 g, p > 0.05, n = 19 mice for each group). The mechanical nociceptive thresholds tended to be higher in cKO mice than in control mice at 4 and 7 days post-operation, although they did not reach statistical significance. To exclude the potential effects of the nestin-Cre line on nociceptive behaviors, we confirmed that there were no significant differences in the nociceptive responses following CFA injection or CCI between nestin-Cre and wild-type mice ( Supplementary Fig. S2).

SecinH3 reduces mechanical allodynia and spinal Arf6 activation in pain models
To examine whether the catalytic activity of cytohesins in the spinal cord is required for mechanical allodynia in inflammatory and neuropathic pain models, SecinH3 was intrathecally administered to wildtype mice 2 h before CFA injection or CCI. Pretreatment with SecinH3 at two different doses (0.1 or 1 μg/g body weight) significantly increased the ipsilateral paw-withdrawal threshold to mechanical stimuli at 2 days after CFA injection, as compared to the control, and tended to be Data are presented as mean ± SD.

Cytohesin-2 cKO mice exhibit reduced mechanical allodynia induced by pharmacological activation of spinal mGluR1/5
Pharmacological activation of spinal mGluR1/5 by intrathecal administration of DHPG, a selective agonist of group I mGluRs, induces spontaneous nociceptive behaviors, such as licking of the flanks, hindpaws, and tail, as well as thermal and mechanical hypersensitivity (Adwanikar et al., 2004;Fisher and Coderre, 1996;Fisher and Coderre, 1998). To provide evidence for the functional linkage between cytohesin-2 and mGluR1/5 during chronic pain, we compared the effect of DHPG on nociceptive responses between control and cKO mice. Double immunofluorescence labeling of the superficial dorsal horn with antibodies against cytohesin-2 (A 1 , B 1 ), and mGluR5 (A 2 ) or Gαq/11/14 (B 2 ), showing the partial overlapping of cytohesin-2-immunoreactive puncta with mGluR5 and Gαq/11/14. (C) Immunoprecipitation analysis. The P2 fraction of the dorsal spinal cord of adult mice was subjected to immunoprecipitation with control, anti-cytohesin-2, or anti-mGluR5 IgG. The lysate (input) and immunoprecipitates were subjected to immunoblotting with anti-cytohesin-2 or anti-mGluR5 IgG. Scale bars: 10 μm. Fig. 4. Characterization of central nervous system-specific cytohesin-2 conditional knockout mice. (A) Immunoblotting. Total lysates of the spinal cord and brain from control (Cont) and conditional knockout (cKO) mice were subjected to immunoblotting with antibodies against N-and C-terminal regions of cytohesin-2, and α-tubulin.

Discussion
Accumulating evidence suggests that disturbances in Arf-mediated membrane trafficking and actin cytoskeleton remodeling in neurons are associated with various neurodevelopmental and neurological disorders, including periventricular heterotopia (Sheen et al., 2004), intellectual disability with seizure and autism spectrum disorder (Falace et (Sivadasan et al., 2016;Zhai et al., 2015), and Alzheimer's disease (Sannerud et al., 2011;Tang et al., 2015). In this study, we provided the first evidence for the functional involvement of the cytohesin-2-Arf6 pathway in inflammatory and neuropathic pain. One of the major findings is that conditional deletion of cytohesin-2 in the CNS reduced mechanical allodynia in CFA inflammation and CCI pain models. Torii et al. (2015) previously reported that conditional deletion of cytohesin-2 in Schwann cells reduced myelination in the sciatic nerve. However, the present immunohistochemical analyses revealed that there were no detectable differences in the immunoreactive intensity of MBP and number of myelinated fibers in the dorsal root of spinal nerves or white matter of the spinal cord between control and CNS-specific cytohesin-2 cKO mice, excluding the possibility that a disturbance in impulse transmission caused by impaired myelination affected the anti-allodynic phenotype. We also demonstrated that Total lysates of the dorsal spinal cords at 12 h following CFA injection (C) and CCI (D) were subjected to pull-down assay using GST-GGA1 fusion protein, and the total lysates (input) and precipitates (pull-down) were immunoblotted with antibodies against Arf6 and Arf1. (E-H) Quantification of the effects of SecinH3 on the activation of Arf6 (E, F) and Arf1 (G, H) in CFA inflammatory (E, G) and CCI (F, H) pain models (E, one-way ANOVA, F(2, 12) = 3.965, p = 0.047; F, one-way ANOVA, F(2, 6) = 5.956, p = 0.037; G, one-way ANOVA, F(2, 9) = 5.826, p = 0.024; H, one-way ANOVA, F(2, 6) = 8.514, p = 0.018, with Tukey-Kramer post hoc test). *p < 0.05 vs control in E-H. The activation level of Arf6 or Arf1 in each condition was normalized by that of control. Note that intrathecal SecinH3 administration blocked the activation of Arf6, but not Arf1, in both pain models.
cytohesin-2 was expressed in dorsal root ganglion neurons at a comparative level between the two genotypes. Thus, the anti-allodynic phenotype observed in the mutant mice was likely to be caused by the lack of cytohesin-2 in the CNS, but not in the PNS. Furthermore, intrathecal administration of SecinH3 similarly reduced mechanical allodynia, indicating that the cytohesin catalytic activity in the spinal cord is required for the development of mechanical allodynia. Although SecinH3 broadly inhibits the cytohesin family members (Hafner et al., 2006), the phenotypic similarity between mice lacking cytohesin-2 in the CNS and those administered intrathecally with SecinH3 suggests that cytohesin-2 in the spinal cord is responsible for the development of mechanical allodynia. Finally, the Arf pull-down activation assay demonstrated that, although both Arf6 and Arf1 were activated in both pain models, intrathecal SecinH3 administration blocked the activation of Arf6, but not Arf1, in the dorsal spinal cord. Interestingly, the activation of Arf6 in the dorsal spinal cord was peaked at 12 h after CFA injection or surgery, the genetic deletion of cytohesin-2 or pharmacological inhibition of cytohesins had a relatively long-lasting anti-allodynic effect. Together, it is reasonable to conclude that the cytohesin-2 mediates the initiation and development of mechanical allodynia through the activation of Arf6 in the dorsal horn. However, it should be mentioned that chronic pain can develop as a consequence of enhanced excitability in any neural circuits or nuclei that relay and process somatosensory information, including brain stem nuclei, periaqueductal gray, thalamus, somatosensory cortex, cingulate cortex, hippocampus and amygdala, as well as the spinal cord and PNS (Apkarian et al., 2005;Basbaum et al., 2009;Kuner, 2010). Since cytohesin-2 is widely expressed in the CNS (Ito et al., 2018), we cannot completely exclude the possibility that reduced mechanical allodynia observed in cKO mice may be attributed to the lack of cytohesin-2 in other brain regions in addition to the spinal cord.
Mechanical allodynia is considered to be a pathological condition in which the balance between excitatory and inhibitory neural circuits to projection neurons in the dorsal horn is disturbed, thereby allowing innocuous low-threshold mechanical inputs to gain access to and activate nociceptive projection neurons via polysynaptic pathways (Peirs et al., 2020;Peirs and Seal, 2016;Todd, 2010). Concerning cell types or neural circuits in which cytohesin-2 mediates mechanical allodynia in Fig. 7. Cytohesin-2 conditional knockout mice exhibit reduced mechanical allodynia and ERK1/2 phosphorylation following intrathecal DHPG administration. (A) Time-course changes in paw-withdrawal thresholds of control and conditional knockout (cKO) mice following intrathecal DHPG administration. Note that cytohesin-2 cKO mice exhibit significantly higher paw-withdrawal thresholds following intrathecal DHPG administration compared to the control (control DHPG vs cytohesin-2 cKO DHPG, two-way repeated measures ANOVA, F(3, 12) = 4.635, p = 0.022, with Tukey-Kramer post hoc test). *p < 0.05 vs control DHPG. (B) Representative immunoblots of the effects of DHPG on ERK1/2 activation and mGluR5 levels in the spinal cord between the control and cKO mice. Total spinal lysates from control and cKO mice treated with or without DHPG for 1 h were immunoblotted with antibodies against ERK1/2, phospho-ERK1/2, mGluR5, and α-tubulin. (C) Quantification of the effect of DHPG on ERK1/2 activation in the spinal cord. The phospho-ERK1/2 level was normalized to that of total ERK1/2 for each condition. Note the absence of the increase in ERK1/2 phosphorylation following intrathecal DHPG administration in cytohesin-2 cKO mice (control vehicle vs control DHPG, *p = 0.041, t-test; cytohesin-2 cKO vehicle vs cytohesin-2 cKO DHPG, p = 0.796, t-test). (D) Quantification of the expression level of mGluR5, showing no significant changes in the spinal mGluR5 level following intrathecal DHPG administration in both genotypes (control vehicle vs control DHPG, p = 0.837, t-test; cytohesin-2 cKO vehicle vs cytohesin-2 cKO DHPG, p = 0.481, t-test). the spinal cord, we demonstrated that cytohesin-2 was expressed abundantly in dorsal horn neurons, particularly in laminae I and II. By double immunofluorescence, we further found that cytohesin-2 was expressed primarily in a subset of excitatory interneurons labeled by PKCγ, calbindin, calretinin, or NK1R, and probably in NK1Rimmunoreactive projection neurons in the superficial dorsal horn. Recent evidence has indicated that excitatory interneurons in the dorsal horn, including neurons transiently expressing VGluT3, and neurons expressing PKCγ or calretinin, form an essential neural pathway for mechanical allodynia (Peirs et al., 2020;Peirs and Seal, 2016). Therefore, it is plausible that cytohesin-2 mediates mechanical allodynia in an excitatory neural circuit in the dorsal horn. Since dorsal horn neurons are highly diverse in their neurochemical, electrophysiological, and morphological properties (Peirs et al., 2020;Todd, 2017), it is difficult to define a unique functional neuronal population by any one of neurochemical markers such as neuropeptides, receptors, and intracellular signaling molecules. Thus, further studies using multiple lines of mice lacking cytohesin-2 in specific cell types in the dorsal horn are necessary to determine the cell type(s) and neural circuit(s) responsible for cytohesin-2-mediated mechanical allodynia in the dorsal horn.
Another important finding is that cytohesin-2 is functionally associated with mGluR1/5-mediated central sensitization. Cytohesin-2 was previously shown to be associated with mGluR1/5 in the rat brain through tamalin (Kitano et al., 2002). Consistently, several lines of evidence obtained by immunoprecipitation, immunofluorescence, and immunoelectron microscopy in the present study suggested that cytohesin-2 forms a protein complex with mGluR5 in the perisynaptic region, and is ideally positioned to mediate mGluR5-dependent nociceptive hypersensitivity in dorsal horn neurons. Indeed, mechanical allodynia induced by intrathecal DHPG administration was significantly reduced in cytohesin-2 cKO mice. Furthermore, cytohesin-2 cKO mice exhibited reduced DHPG-induced ERK1/2 activation in the spinal cord. The ERK1/2 pathway is the major signaling pathway downstream of mGluR1/5 in response to intense nociceptive stimuli (Adwanikar et al., 2004;Karim et al., 2001), and is implicated in chronic pain through the regulation of ion channels and gene transcription by phosphorylating various substrates, such as the A-type K + channel Kv4.2 and cyclic AMP response element-binding protein (CREB) (Hu et al., 2007;Hu et al., 2006;Ji et al., 1999;Ji et al., 2002;Kawasaki et al., 2004). Together, we propose that cytohesin-2-Arf6 mediates mGluR5-dependent central sensitization through the downstream ERK1/2 activation in dorsal horn neurons. However, it is noteworthy that cytohesin-2 is also associated with various membranous structures corresponding to early, recycling, and late endosomes, as well as lysosomes, as revealed by double immunofluorescence labeling and immunoelectron microscopy. The cytohesin-2-Arf6 pathway is implicated in intracellular trafficking of transmembrane proteins, such as the β2-adrenergic receptor (Claing et al., 2001) and integrin β1 (Oh and Santy, 2010). In addition, tamalin was recently shown to mediate activity-dependent trafficking of mGluR1 and AMPARs in primary hippocampal neurons (Pandey et al., 2020). Since endosomal trafficking is a critical determinant of the fate of the receptor and propagation of intracellular signaling from internalized receptors in endosomes (Di Fiore and von Zastrow, 2014), it is also possible that the cytohesin-2-Arf6 pathway may mediate mGluR5dependent plastic changes in the excitability of dorsal horn neurons, by regulating intracellular trafficking of mGluR5 and/or other receptors.
Using quantitative immunoelectron microscopy, we demonstrated that cytohesin-2 cKO mice exhibited a significant shift in the peak position of immunogold particles for mGluR5, from the perisynaptic region to the inside of the PSD, without apparent changes in the amount of membrane-associated mGluR5, as compared to that in control mice. In contrast, the preferential perisynaptic distribution of Gαq/11/14 did not differ between the two genotypes. Since spatial coupling of mGluRs with their downstream effectors is considered to be crucial for the efficacy and fidelity of signal transduction (Nakamura et al., 2004;Tanaka et al., 2000), it is possible that spatial dissociation of mGluR5 from Gαq/11/14 in the perisynapse is causatively related to reduced mGluR1/5dependent mechanical allodynia and spinal ERK1/2 activation in cytohesin-2 cKO mice. Although the precise mechanism for perisynaptic accumulation of mGluR1/5 remains largely unknown, specific interactions of mGluR1/5 with scaffolding proteins and with the actin cytoskeleton are believed to play a critical role in the positioning and lateral mobility of mGluR1/5 in the postsynaptic membrane (Scheefhals and MacGillavry, 2018). Previous studies using transgenic mouse rescue strategies, in which transgenes for mGluR1a, mGluR1b, or a mGluR1a mutant lacking the ability to bind to Homer proteins were introduced into mGluR1 knockout mice, revealed that the long C-terminal region of mGluR1a is essential for the perisynaptic accumulation of mGluR1 in Purkinje cells, although its interaction with Homer proteins is not sufficient for the perisynaptic targeting (Naito et al., 2018;Ohtani et al., 2014). This suggested that other mGluR1a-interacting proteins, such as tamalin (Kitano et al., 2002), Preso1 (Hu et al., 2012), and Norbin (Wang et al., 2009), participate in controlling the perisynaptic positioning of mGluR1a in a cooperative manner with Homer proteins. Among these, tamalin forms a multiprotein complex with various postsynaptic scaffold proteins, including S-SCAM, SAP95/PSD-95-associated proteins (SAPAPs), and PSD-95, in addition to mGluR1/5 and cytohesin-2 (Kitano et al., 2002;Kitano et al., 2003). Interestingly, cytohesin-2 was reported to regulate hepatocyte growth factor (HGF)-stimulated trafficking of tamalin from perinuclear recycling endosomes to the plasma membrane through the activation of Arf1 and Arf6 in the Madin-Darby canine kidney (MDCK) epithelial cell line (Koubek and Santy, 2016). Therefore, it can be speculated that the lack of cytohesin-2 may disarrange the protein-protein network in the PSD due to impaired trafficking of tamalin to the postsynapse, consequently reducing the molecular crowding of the PSD or increasing the number of slots in the PSD, and leading to the displacement of mGluR5 into the PSD. In addition, the actin cytoskeleton may also regulate the perisynaptic accumulation of mGluR5 by anchoring mGluR1/5 to the perisynaptic region and/or by limiting the entry of mGluR5 into the PSD. It was previously shown that polymerized actin is highly concentrated and dynamic in the perisynaptic region (Frost et al., 2010), and that mGluR5 is coupled to the actin cytoskeleton through its direct interaction with Factin-binding proteins, such as α-actinin-1 (Cabello et al., 2007) and Filamin-A (Enz, 2002). Since cytohesin-2 regulates various Arf6dependent actin-based cellular processes, such as cell morphology (Frank et al., 1998b), migration (Santy and Casanova, 2001), and neurite outgrowth (Hernandez-Deviez et al., 2004;Torii et al., 2012;Torii et al., 2014;Yamauchi et al., 2009), the disturbance in the Arf6dependent actin cytoskeleton dynamics caused by the lack of cytohesin-2 in the postsynapse could also account for the shift of mGluR5 into the PSD from the perisynaptic region.
The mechanism for the activation of cytohein-2 during chronic pain is also an important issue for future research. The catalytic activity of the cytohesin family is subject to complex regulation by intramolecular autoinhibition and phosphorylation. The catalytic activity of cytohesin-2 was reported to be autoinhibited by the intramolecular interaction of the linker region between the Sec7 and PH domain and a C-terminal amphipathic helix region containing a polybasic motif with the Sec7 domain in a pseudosubstrate mechanism, which is released by the phosphorylation of a serine residue at position 392 within the C-terminal polybasic region by PKC (DiNitto et al., 2007;Frank et al., 1998b). In addition, the Src family tyrosine kinase Fyn was shown to activate cytohesin-1 through phosphorylation of a tyrosine residue at position 382 in the C-terminal region of cytohesin-1, which is conserved in cytohesin-2, in Schwann cells during the myelin formation (Yamauchi et al., 2012). Interestingly, the PKC-Src pathway is a critical signaling cascade for central sensitization, which is activated by postsynaptic mGluR1/5 stimulation and leads to an increase in the channel activity of NMDARs through the phosphorylation of NMDARs, particularly the GluN2B subunit, in dorsal horn neurons (Guo et al., 2004). Therefore, it is attractive to speculate that cytohesin-2 may be activated by PKC and/ or Fyn downstream of mGluR5 in response to nociceptive stimuli, leading to the activation of the Arf6 pathway in dorsal horn neurons.
Finally, in actual clinical settings, it is frequently difficult to distinguish the primary cause of chronic pain clearly between neuropathy and inflammation. It is noteworthy that intrathecal SecinH3 administration had anti-nociceptive effects on both inflammatory and neuropathic mechanical allodynia. Therefore, manipulation of the cytohesin-2-Arf6 pathway could be an attractive target for chronic pain, with a broad therapeutic window. However, considering the fundamental roles of Arf-dependent cellular processes, general blockage of the cytohesin-2-Arf6 pathway with SecinH3 is unlikely to be clinically applicable. Indeed, mice fed with SecinH3 have been reported to develop hepatic insulin resistance associated with the preclinical stages of type 2 diabetes (Hafner et al., 2006). Further investigations of the cytohesin-2-Arf6 signaling pathway in nociception may reveal new opportunities for developing novel therapeutic strategies for chronic pain.

Author contributions
A.I., M.F., and H.S. designed the study, performed experiments, analyzed data, and drafted the paper. T.S. and Y.H. supervised part of the immunoprecipitation and immunohistochemical analyses; H.O. supervised the behavioral analyses; J.Y. generated and provided the cytohesin-2 cKO mice.

Declaration of competing interest
The authors declare no conflict of interest.