Protein Kinase C Lambda Mediates Acid-Sensing Ion Channel 1a-Dependent Cortical Synaptic Plasticity and Pain Hypersensitivity

Chronic pain is a serious debilitating disease for which effective treatment is still lacking. Acid-sensing ion channel 1a (ASIC1a) has been implicated in nociceptive processing at both peripheral and spinal neurons. However, whether ASIC1a also contributes to pain perception at the supraspinal level remains elusive. Here, we report that ASIC1a in ACC is required for thermal and mechanical hypersensitivity associated with chronic pain. ACC-specific genetic deletion or pharmacological blockade of ASIC1a reduced the probability of cortical LTP induction and attenuated inflammatory thermal hyperalgesia and mechanical allodynia in male mice. Using cell type-specific manipulations, we demonstrate that ASIC1a in excitatory neurons of ACC is a major player in cortical LTP and pain behavior. Mechanistically, we show that ASIC1a tuned pain-related cortical plasticity through protein kinase C λ-mediated increase of membrane trafficking of AMPAR subunit GluA1 in ACC. Importantly, postapplication of ASIC1a inhibitors in ACC reversed previously established nociceptive hypersensitivity in both chronic inflammatory pain and neuropathic pain models. These results suggest that ASIC1a critically contributes to a higher level of pain processing through synaptic potentiation in ACC, which may serve as a promising analgesic target for treatment of chronic pain. SIGNIFICANCE STATEMENT Chronic pain is a debilitating disease that still lacks effective therapy. Ion channels are good candidates for developing new analgesics. Here, we provide several lines of evidence to support an important role of cortically located ASIC1a channel in pain hypersensitivity through promoting long-term synaptic potentiation in the ACC. Our results indicate a promising translational potential of targeting ASIC1a to treat chronic pain.


Introduction
Chronic pain is a challenging clinical problem seriously affecting millions of people everyday (Holmes, 2016). Unfortunately, the current treatment for chronic pain is not very effective (Woolf, 2010;Yekkirala et al., 2017). The most widely used analgesics for chronic pain are opioids and aspirin-like nonsteroidal anti-inflammatory drugs, for which the clinical efficacy is limited by severe adverse effects and abuse liability (Roques et al., 2012;Latremoliere and Costigan, 2018). On the other hand, pain perception eventually resides in the brain (Chen, 2018). ACC is an important forebrain structure involved in a variety of physiological and pathological functions, including chronic pain (Chen et al., 2018;Sellmeijer et al., 2018). Synaptic plasticity in ACC is considered one of the most critical mechanisms underlying the transition of pain from acute to chronic (Luo et al., 2014;Bliss et al., 2016). Peripheral inflammatory or neuropathic pain elicits dramatic plastic changes in synaptic transmission and morphology of ACC neurons (Zhao et al., 2006;Ko et al., 2018), whereas blocking or erasing aberrant cingulate plasticity produces significant analgesic effects in animals (Li et al., 2010;Wang et al., 2011).
Accordingly, molecules involved in the induction and expression of plastic changes in ACC are potential drug targets for treating chronic pain (Bliss et al., 2016). Among several candidates, NMDAR, a major glutamate receptor important for cortical LTP (Paoletti et al., 2013), has been intensively investigated as an analgesic target by academic laboratories and pharmaceutical companies. However, the use of a general NMDAR antagonist is limited by side effects, such as ataxia and sedation (Chizh et al., 2001;Wu and Zhuo, 2009). Another important glutamate receptor, AMPAR, contributes to the expression of LTP through regulated trafficking into and out of synapses (Huganir and Nicoll, 2013). However, perampanel, a noncompetitive AMPAR antagonist used to treat partial seizures for people older than 12 years, can induce serious adverse psychiatric and behavioral reactions (Ettinger et al., 2015). Because of the limitations in targeting classical glutamate receptors, it is necessary and important to explore other molecule targets for chronic pain treatment.
Ion channels are emerging as promising targets for devising and developing novel analgesics with promising therapeutic outcomes (Waxman and Zamponi, 2014;Zamponi, 2016). Acidsensing ion channel (ASIC) family consists of six isoforms (ASIC1a, 1b, 2a, 2b, 3 and 4) encoded by four genes (Kellenberger and Schild, 2015), among which ASIC1a has been most frequently studied because of its ubiquitous expression in the brain (Wemmie et al., 2003). Previous work has unequivocally demonstrated that ASIC1a plays crucial roles in modulating pain sensitivity (Wemmie et al., 2013;Deval and Lingueglia, 2015). Peripheral administration of a Texas coral snake toxin elicits strong pain in mice through activation of ASIC1-expressing nociceptors (Bohlen et al., 2011). In contrast, intrathecal injection of mambalgin-1, a toxin isolated from black mamba venom, or psalmotoxin 1 (PcTx1), a peptide extracted from South American tarantula, produces potent analgesic effect in an ASIC1adependent manner (Mazzuca et al., 2007;Diochot et al., 2012). However, despite these exciting results, fewer reports have illustrated the role of supraspinally localized ASIC1a in chronic pain.
Here, we investigated the role of ASIC1a in ACC neurons in chronic pain and cortical plasticity. We found that genetic deletion or pharmacological blockade of the ASIC1a channel in ACC substantially attenuated thermal and mechanical hypersensitivities in inflammatory pain. Moreover, ASIC1a modulates pain sensitivity by promoting AMPAR GluA1 subunit trafficking and induction of LTP in ACC through protein kinase C (PKC). Importantly, post-treatments with ASIC1a inhibitors can reverse preestablished pain hypersensitivity in both inflammatory and neuropathic pain models, demonstrating a promising translational potential of targeting ASIC1a-mediated aberrant cortical plasticity to treat chronic pain.

Materials and Methods
Animals. The experiments were performed on C57BL/6J, ASIC1a WT/ KO, and ASIC1a flox/flox mice (male, 8 -10 weeks old). The global ASIC1a KO mice (RRID:MGI:2654038) were the generous gifts of Prof. Michael J. Welsh (Howard Hughes Medical Institute, University of Iowa, Iowa City, IA) (Wemmie et al., 2002). The conditional ASIC1a flox/flox mice (RRID:IMSR_RMRC13158) were graciously provided by Prof. Cheng-Chang Lien (National Yang-Ming University, Taiwan) (Wu et al., 2013). All mice were housed in groups of 4 or 5 per cage under standard environment (12 h light/dark cycles at 21°C and 50%-60% humidity) with mouse chow and water ad libitum. Animal care and the experimental procedures were approved by the Animal Ethics Committee of Shanghai Jiao Tong University School of Medicine, Shanghai, China (Policy Number DLAS-MP-ANIM.01-05).
Animal model of pain. To induce inflammatory pain, complete Freund's adjuvant (CFA, 50% in saline, 20 l; Sigma-Aldrich) was injected into the plantar surface of the left hindpaw (Berta et al., 2014). Sham control animals received 20 l of saline. Unless otherwise stated, pain-related behavioral tests were performed at day 1, 3, 5, and 7 after the CFA injection. For neuropathic pain, we adopted a well-established spared nerve injury (SNI) model as previously described (Bourquin et al., 2006;Chen et al., 2015;Tan et al., 2017). Briefly, the mice were anesthetized with isoflurane, and an incision was made to the lateral skin surface of the thigh to expose the sciatic nerve and its branches. Then a 5.0 silk tight ligation of the tibial and common peroneal nerves was performed, followed by transection and removal of a 3 to 5 mm portion of the nerve. However, the third peripheral branch of the sciatic nerve, the sural nerve, was left intact, and any contact with or stretching of this nerve was carefully avoided. The muscle and skin were subsequently sutured, and animals were left to recover in a heated cage for 24 h. For the sham group, the surgical procedures were almost the same, except that the nerves were only exposed but not ligated/cut. Behavioral testing was performed at 7 and 14 d after the operation.
Cannula implantation and drug microinjection. Cannulation and microinjection were performed as described previously (Wang et al., 2011;Ko et al., 2018). Animals were anesthetized and placed in a stereotaxic frame (RWD Life Science). Guide cannulas were implanted bilaterally into ACC with the following coordinates: AP 1.00 mm, lateral Ϯ1.47 mm with a 30°angle, DV Ϫ1.65 mm from the skull. The cannulas were fixed in the position with acrylic dental cement and secured with skull screws. To prevent clogging, a stylus was placed in the guide cannula. Animals were allowed to recover from surgery for at least 1 week before experimental manipulations. For drug microinfusion, the stylus was removed from the guide cannula, and a 30-gauge injector (1.00 mm lower than the guide) cannula was used. The infusion cannula was conducted via PE20 tubing to a motorized microsyringe infusion pump (KDS 310, KD Scientific). PcTx1 (10 M in saline, 0.5 l per side, Peptide Institute), compound 5b (100 M in saline, 0.5 l per side), and their respective vehicles were microinfused into ACC at a rate of 0.1 l/min. These drug doses were chosen based on the previous literature using the similar pharmacological strategy to study the function of ASIC1a (Buta et al., 2015;Wang et al., 2018), as well as our preliminary experiments and calculations on the diffusion range of the drug in the ACC according to the estimated average volume of this brain region in an adult mouse. Also, we did not observe any obvious neurological deficits in the druginfused animals, indicating the lack of significant side effects at the selected doses. After the microfusion, the injector was left in place for 10 min for the drug to diffuse and act on its targets before experimental testing. The injection sites were mapped postmortem by sectioning the brain and performing Nissl staining. Animals with incorrect diffusion scope were excluded from the data analysis.
Virus constructs. Adeno-associated virus (AAV) vectors (pAAV-MCS) carrying the full-length of WT mouse ASIC1a (NM_009597.2) or PKC (NM_008857.3) cDNA were constructed with the coding sequences of EYFP and mCherry fused to that of the N termini of ASIC1a and PKC, respectively, via the coding sequence for the "self-cleaving" 2A-peptide. To ensure specific expression of ASIC1a (or PKC) in neurons, human synapsin I (hSyn) promoter was used. The construct was then packaged into AAV2/8 chimeric virus with AAV8 capsids and AAV2 ITR (inverted terminal repeat) element. The pAAV-EGFP (or pAAV-mCherry) vector driven by the hSyn promoter without ASIC1a (or PKC) was used as a control AAV vector (AAV-hSyn-Ctrl). The AAV vector for expression of Cre recombinase or EGFP driven by the hSyn promoter was constructed using similar methods.
To make the short hairpin RNA (shRNA) constructs, oligonucleotides that contain 21 base sense and antisense sequences targeting mouse ASIC1a (NM_009597.2, sense sequence, 5Ј-GGACATTCAGCAAGAT GAATA-3Ј) were connected with a hairpin loop followed by a poly(T) termination signal . These sequences were then cloned into the vector pGLV1-U6-GFP (Genepharm). The negative control sequence was 5Ј-GTTCTCCGAACGTGTCACGT-3Ј. The shRNA vector targeting mouse PKC, 5Ј-CTGGATTTGCTGACATCCAAG-3Ј (Wang et al., 2016), was prepared by inserting oligonucleotides containing these sequences in pAAV-U6-MCS-EF1␣-EGFP. Each plasmid was then packaged into rAAV2/8 particles (AAV2 ITRs; AAV8 capsid). These procedures were performed by Sunbio Medical Biotechnology. To knock out ASIC1a specifically in excitatory neurons in ASIC1a flox/flox mice, AAV vector for expression of Cre recombinase driven by CaMKII promoter was constructed. AAV vector for expression of EGFP driven by CaMKII promoter was used as a control in the same set of experiments.
Viral injection. For viral injection, mice at the age of 4 -6 weeks were anesthetized and placed in a stereotaxic apparatus (RWD Life Science). The coordinates of ACC according to the mouse brain atlas were as follows: AP 1.00 mm, lateral Ϯ0.30 mm, DV Ϫ1.80 mm. Viruses (titers of all vectors Ͼ1.0 ϫ 10 12 viral genome-containing particles per milliliter) were injected bilaterally (0.5 l/side) into ACC using glass micropipettes connected with a microinfusion pump (KDS 310, KD Scientific) at a rate of 0.1 l/min. After the microfusion, the microelectrode was left in place for an additional 10 min to allow the viruses to diffuse. Mice were allowed to recover for at least 4 weeks before behavioral or electrophysiological experiments. Targeting was confirmed postmortem with fluorescence above background in ACC bilaterally.
Open field test. After intra-ACC administration of saline or PcTx1 as described above, each mouse was placed in the center of a square Plexiglas open field apparatus (40 ϫ 40 ϫ 35 cm) and allowed to freely explore for 30 min. Activity was recorded and analyzed using the Ethovision video-tracking system (Noldus Information Technology). Time spent and distances traveled in the center zone (covering 20 cm ϫ 20 cm) were used as the measure of anxiety. Total distance traveled in the open field was also measured and analyzed to assess the motor ability of the animal.
Hargreaves' test. Thermal hyperalgesia was assessed by measuring paw withdrawal latencies in response to a radiant heat source using Hargreaves' test. Briefly, mice were placed in a transparent plastic box (10 ϫ 10 ϫ 5 cm) on an elevated glass plate. Thermal stimulation was delivered to the hindpaw from an infrared source (Ugo Basile) placed under the glass plate. Paw withdrawal latency was recorded by the length of time between the start of the infrared light and the foot withdrawal. Paw withdrawal latencies were averaged from at least three consecutive tests, with a minimum of 10 min between individual measurements.
von Frey test. Mechanical allodynia was examined by measuring paw withdrawal thresholds in response to mechanical stimuli (calibrated von Frey filaments). Briefly, the animal was placed in a transparent plastic box (10 ϫ 10 ϫ 5 cm) on an elevated mesh screen. Calibrated von Frey filaments with ascending order (0.07, 0.16, 0.40, 0.60, 1.00, 1.40, and 2.00 g) were used to stimulate the plantar surface of mouse hindpaw. Paw withdrawal threshold was determined by a modified up-down method as previously described (Chaplan et al., 1994).
Place escape/avoidance paradigm (PEAP). PEAP testing was conducted as described previously (LaBuda and Fuchs, 2000;LaGraize et al., 2004;Han et al., 2014). Individual mice were placed in a 50 ϫ 30 ϫ 30 cm chamber on top of a raised mesh floor with one-half painted white (light area) and the other half painted black (dark area). Animals were allowed unrestricted movement throughout the test chamber. A suprathreshold mechanical stimulus (4.0 g, von Frey filament) was applied to the plantar surface of the hindpaws at 15 s intervals throughout the 30 min test period. The mechanical stimulus was applied to the affected paw (ipsilateral to the CFA/SNI) while the animal was located in the dark area, and the naive paw (contralateral to the CFA/SNI) was stimulated when the animal was in the light area. Saline-injected and sham-ligated animals were mechanically stimulated in an identical manner as the experimental group. The percentage of time spent on the nonpreferred light side was used to quantify the level of pain affect according to previous methods (LaBuda and Fuchs, 2000;Han et al., 2014). Subcellular fractionation. The ACC membrane fractions were prepared 3 d after CFA treatment as described previously (Xu et al., 2008;Duan et al., 2012). Briefly, ACC samples were homogenized in a lysis buffer (10 mM Tris, pH 7.4, 300 mM sucrose, and 1 mM EDTA) containing protease inhibitor mixture and then centrifuged at 8000 ϫ g for 5 min at 4°C. The pellet (P1) containing nuclei and cell debris was discarded. The supernatant (S1) was then centrifuged at 40,000 ϫ g for 30 min at 4°C to obtain crude membranes in pellet (P2). This pellet (P2), which included cellular membranes, was resuspended in the lysis buffer with protease inhibitor mixture. Protein contents in P2 were quantified using Pierce BCA Protein Assay Kit (Thermo Fisher Scientific). The expression of GluA1 and ASIC1a in P2 (membrane fractions) was characterized by Western blotting. Cadherin was used as the loading control of membrane extracts.
Primary culture of mouse cortical neurons. Cultured cortical neurons were prepared as described previously (Zeng et al., 2013;Wang et al., 2015). Postnatal day 1 C57BL/6 WT or ASIC1a KO mice (with a congenic C57BL/6 background) were anesthetized with halothane. Brains were removed rapidly and placed in ice-cold Ca 2ϩ -and Mg 2ϩ -free PBS. Tissues were dissected and incubated with 0.05% trypsin-EDTA for 15 min at 37°C, followed by trituration with fire-polished glass pipettes, and plated in poly-D-lysine-coated 35 mm culture dishes (5 ϫ 10 5 cells per dish). Neurons were cultured with Neurobasal medium containing 2% B27 and 1% Glutamax supplements and maintained in a humidified 5% CO 2 atmosphere incubator. Cultures were fed twice a week and used for all the assays 14 -16 d after plating. Glial growth was suppressed by the addition of 5-fluoro-2-deoxyuridine (20 g/ml; Sigma-Aldrich) and uridine (20 g/ml; Sigma-Aldrich). For Western blot analysis, cultured cortical neurons were treated with an acidic solution, pH 6.0, for 10 min before being homogenized for protein sample preparation. In some experiments, we applied a Ca 2ϩ -free (with no Ca 2ϩ added and the addition of 0.1 mM EGTA) pH 6.0 solution or a pH 6.0 solution pretreated with BAPTA-AM (30 M) for 30 min to test the Ca 2ϩ dependence of acidosisinduced PKC phosphorylation. The pH 7.4 extracellular solution (Ca 2ϩ -free or not) was used as a control.
Neuron transfection and calcium imaging. Calcium phosphate transfection was performed in cultured cortical neurons grown on glass coverslips on 7-10 DIV. Before transfection, the culture medium was changed for fresh Neurobasal medium, and the original medium was readded. For the 35 mm dish, 1-4 g of GCaMP6 plasmid was added into 60 l CaCl 2 solution (0.3 M) with pipetting; then 60 l HBSS (280 mM NaCl, 1.5 mM Na 2 HPO 4 , 50 mM HEPES, pH 6.9) was added. After fully mixing, the transfection solution was immediately transferred into the dish. After incubation at 37°C for 1-1.5 h, the medium was replaced with CO 2saturated Neurobasal medium (wash medium) to remove excess calcium phosphate particles. After that, the wash medium was replaced with 1 ml of the original medium and 1 ml of fresh Neurobasal medium plus B27 supplement and the dish returned to the culture incubator.
At 14 -16 DIV, calcium imaging experiments were performed in neurons transfected with GCaMP6. The neurons on coverslip were mounted in standard extracellular solution and consecutively imaged using a Nikon A1R laser scanning confocal microscope with a 20ϫ lens. GCaMP6 was excited with a 488 nm laser. Baseline recording under pH 7.4 (60 s) was rapidly changed to pH 6.0 (20 s), followed by pH 7.4 perfusion for another 50 s. GCaMP6 fluorescence change (⌬F/F 0 ) representing Ca 2ϩ alterations was calculated by (F Ϫ F 0 )/F 0 , where F 0 refers to the median of the fluorescence values during the baseline period (from 0 to 60 s).
Brain slice preparation and patch-clamp recordings. Experiments were performed on ACC slices as described previously (Koga et al., 2015;Song et al., 2017). Briefly, after decapitation, the mouse brain was quickly removed and placed in well-oxygenated (95% O 2 /5% CO 2 , v/v) ice-cold aCSF containing the following (in mM): 125 NaCl, 2.5 KCl, 12.5 D-glucose, 1 MgCl 2 , 2 CaCl 2 , 1.25 NaH 2 PO 4 , and 25 NaHCO 3 , pH 7.35-7.45. Three coronal brain slices (300 m) were obtained at the level of ACC with a vibratome (Leica Microsystems VT 1000S) and incubated at 31°C in oxygenated aCSF for at least 1 h. Then the brain slice was transferred to a recording chamber placed on the stage of a microscope (Olympus, BX51WI) equipped with an infrared differential interference contrast video monitor. EPSCs were recorded from layer II/III pyramidal neurons with an Axon 200B amplifier (Molecular Devices), and the stimulations were delivered by a bipolar tungsten stimulating electrode placed in layer V/VI of the ACC (Song et al., 2017). The recording pipettes (3-5 M⍀) were filled with a solution containing the following (in mM): 132.5 Cs-gluconate, 17.5 CsCl, 2 MgCl 2 , 0.5 EGTA, 10 HEPES, 4 Mg-ATP, and 5 QX-314 chloride (280 -300 mOsm, pH 7.2 with CsOH).
AMPAR-mediated EPSCs were induced by repetitive stimulations at 0.03 Hz, and neurons were voltage-clamped at Ϫ70 mV, except where indicated otherwise. The input-output relation of EPSCs was obtained by giving different stimulation intensities onto layer V/VI of the ACC slice. Paired-pulse facilitation was induced by delivering two consecutive pulses with a 25, 50, 75, or 100 ms interpulse interval and calculated as the ratios of the second response peak values over the first response peak values. To isolate ASIC1a-dependent synaptic currents, we treated the ACC slice from either WT or ASIC1a KO mice with CNQX (20 M) and D-APV (50 or 100 M) after obtaining stable AMPAR-mediated EPSCs with different stimulation intensities under the voltage-clamp mode.
Then amiloride (200 M) or PcTx1 (100 nM) was applied to ensure the involvement of ASIC1 in the relatively small remaining current after blocking AMPAR-and NMDAR-mediated synaptic transmission Kreple et al., 2014). To determine AMPAR/NMDAR ratio, peak amplitude of EPSCs at Ϫ70 mV in the presence of picrotoxin (100 M) was measured as AMPAR-mediated currents, and peak amplitude of EPSCs at 40 mV in the presence of picrotoxin (100 M) and CNQX (20 M) was measured as NMDAR-mediated currents. The mEPSCs of superficial layer (layer II and III) pyramidal neurons in ACC were obtained at Ϫ70 mV in the presence of TTX (0.5 M) and picrotoxin (100 M) without stimulation and analyzed using MiniAnalysis (Synaptosoft, RRID:SCR_002184). Spiking activity of ACC neurons was measured under current-clamp mode with an internal solution containing the following (in mM): 145 K-gluconate, 5 NaCl, 10 HEPES, 2 MgATP, 0.1 Na 3 GTP, 0.2 EGTA, and 1 MgCl 2 (280 -300 mOsm, pH 7.2 with KOH). The initial access resistance was 15-30 M⍀, and it was monitored throughout the experiment. Data were discarded if the access resistance changed Ͼ20% during experiment.
Multichannel field potential recordings. A commercial 64-channel multisite recording system (MED64; Alpha-Med Sciences) was used for extracellular field potential recordings. The procedure for preparation of the MED64 probe (MED64 probe, P515A) followed standard methods as previously described . Slice preparation for the MED64 recording was based on the procedures mentioned above. Acutely prepared ACC slices were allowed to recover for at least 2 h at 31°C in oxygenated aCSF. Then one slice was positioned on the MED64 probe in such a way that the area of ACC was entirely covered by the recording dish mounted on the stage of an inverted microscope (IX51, Olympus). The slice was continuously perfused with oxygenated and fresh aCSF at a rate of 2 ml/min with the aid of a peristaltic pump (Minipuls 3; Gilson) throughout the experimental period.
After the slice had been allowed to recover for 20 min on the probe, one of the microelectrodes covering the deep layer V-VI of the ACC slice was selected for stimulation by visual observation through a charge-coupled device camera (DP70; Olympus) connected to the inverted microscope. For test stimulation, monopolar, biphasic constant-current pulses (0.2 ms in duration) were delivered by the data acquisition software (Mobius; Panasonic, Alpha-Med Sciences) at 0.017 Hz through the selected stimulating microelectrode. The fEPSPs recorded at the remaining microelectrodes were amplified by a 64-channel amplifier, displayed on the monitor screen, and stored on the hard disk of a microcomputer for offline analysis. For LTP induction, after the baseline synaptic responses were stabilized for at least 30 min, a theta burst stimulation (TBS) protocol (5 bursts at 5 Hz, repeated 5 times at 10 s intervals, 4 pulses at 100 Hz for each burst) was delivered with the stimulation intensity eliciting 40%-60% of the maximum response (Li et al., 2010;Song et al., 2017;Ko et al., 2018). For LTD induction, stable baseline responses were recorded for at least 20 min before delivery of the low-frequency stimulation (LFS) protocol (1 Hz, 900 pulses) (Liu and Zhuo, 2014). To test the effects of channel inhibitors on LTP induction, the ASIC1a inhibitor PcTx1 (100 nM) was applied from 15 min before until 15 min after the TBS delivery. For LFS-LTD, PcTx1 or another ASIC1a channel inhibitor compound 5b (100 nM) was applied from 15 min before LFS and maintained continuously during the LFS. All multichannel electrophysiological data were analyzed offline by the MED64 Mobius software. For quantification of LTP and LTD data, the slope of fEPSPs was measured, normalized, and expressed as percentage change from the baseline level. The numbers of activated channels versus LTP-showing or LTD-showing channels were counted, and the induction ratios of cingulate LTP/LTD were calculated as follows: (number of LTP-or LTD-showing channels/number of all activated channels) ϫ 100% Liu et al., 2016).
Experimental design and statistical analysis. Before the start of experiments, all animals were acclimatized to their environment for at least 1 week. Mice were assigned to different groups randomly in pharmacological and genetic studies. All behavioral tests were conducted in a temperature-controlled room by an experimenter blinded to the identity of drug treatment and/or mouse genotype. Group sample size was chosen on the basis of previous studies using the same methodologies. Animals with incorrect viral injections or cannula locations were excluded from final analyses. All immunoblotting experiments were repeated at least three times with similar results. Prism software (GraphPad 6.0; RRID: SCR_002798) was used to conduct the statistical analysis. Data distribution was assumed to be normal, but this was not formally tested. The variance between the groups that are being statistically compared is similar. For every figure, statistical tests are justified as appropriate. Tests for statistical differences between groups were performed using Student's t test or a one-way ANOVA or two-way repeated-measures ANOVA followed by Tukey's or Sidak's post hoc test as described in the text. All data are presented as the mean Ϯ SEM. In all cases, p Ͻ 0.05 was considered statistically significant.
In addition to genetic manipulation of ASIC1a expression, we also pharmacologically blocked ASIC1a activation by injecting PcTx1 bilaterally into ACC (10 M, 0.5 l per side), on the day before, during, and 1 d after the CFA injection ( Fig. 2 A, B). The treatment with PcTx1 prevented development of inflammatory thermal hyperalgesia compared with the saline control (F (1,9) ϭ 24.9, p ϭ 0.0007, two-way repeated-measures ANOVA; Tukey's multiple-comparison test: day 1: p ϭ 0.0001; day 3: p ϭ 0.00005; day 5: p ϭ 0.0056; day 7: p ϭ 0.1450; Fig. 2C). Control experiments showed that PcTx1 had no effect on basal nociception, locomotor ability, and anxiety-related emotional state of naive mice ( Fig. 2D-I ). Thus, pharmacological blockade of ASIC1a yielded similar results as genetic deletion of ASIC1a gene, pointing to the importance of cingulate ASIC1a function in inflammatory pain.

ASIC1a is required for LTP induction in ACC
ASIC1a-mediated synaptic plasticity plays pivotal physiological roles in specific brain regions, such as spatial memory in hippocampus and conditioned taste aversion in insular cortex (Wemmie et al., 2002;. Meanwhile, LTP in ACC is considered as a critical cellular mechanism of chronic pain (Bliss et al., 2016). Thus, we hypothesized that ASIC1a may contribute to inflammatory pain hypersensitivity through regulation of ACC LTP. To test this hypothesis, we evaluated the role of ASIC1a in cingulate LTP inducibility. First, we showed that global deletion of ASIC1a did not significantly alter basal synaptic transmission and intrinsic membrane excitability of ACC neurons (Fig. 3). Neither input-output curve of AMPAR-mediated EPSC (Fig. 3 F, G) nor the mEPSC frequency/amplitude (Fig.  3L-N ) was altered by ASIC1a KO. There was equally no significant change in the presynaptic neurotransmitter release probability as demonstrated by the lack of any effect on paired-pulse ratio (Fig. 3 H, I ). In particular, we did not detect any significant change in AMPAR/NMDAR ratio in the ASIC1a KO mice (Fig.  3 J, K ), suggesting no apparent alteration of NMDAR currents after genetic deletion of ASIC1a.
We next examined the possible involvement of ASIC1a in LTP induction, using a previously established 64-channel multielectrode array (MED64) recording system . The relative position of the MED64 probe with an ACC slice is exemplified in Figure 4A. An electrode channel (Fig. 4A, right, red dot) was chosen to deliver electrical stimulation in the deep layer of the ACC slice, whereas those other channels of the 64-channel array that could reliably detect clear fEPSPs were assumed to be recording sites. These are referred to as activated channels. With the application of TBS from the stimulation channel, the percent-age of activated channels that developed LTP, defined as LTP induction ratio (see Materials and Methods), reached 75.2 Ϯ 5.3% in slices from control ASIC1a flox/flox mice injected with AAV-hSyn-Ctrl (Fig. 4 B, C). Remarkably, with conditional deletion of ASIC1a in ACC (AAV-hSyn-Cre-injected), the LTP induction ratio greatly reduced to 29.4 Ϯ 8.0% (t (10) ϭ 4.784, p ϭ 0.0007 vs AAV-hSyn-Ctrl, unpaired Student's t test; Fig. 4C).
To confirm that the impaired ACC LTP indeed resulted from the loss of ASIC1a in ACC, we injected AAV driven by the human synapsin I promoter for neuron-specific ASIC1a overexpression (AAV-hSyn-ASIC1a) into ACC of adult ASIC1a KO mice. Intra-ACC injection of AAV-hSyn-ASIC1a, but not AAV-hSyn-Ctrl, successfully rescued LTP (Fig. 4D), with the induction ratio increased from 32.4 Ϯ 3.6% for AAV-hSyn-Ctrl to 81.1 Ϯ 5.6% for AAV-hSyn-ASIC1a (t (13) ϭ 7.451, p Ͻ 0.0001 vs AAV-hSyn-Ctrl, unpaired Student's t test; Fig. 4E). Aside from genetic manipulations, we also performed pharmacological experiments and found that bath application of either PcTx1 (100 nM) or another ASIC1a inhibitor, compound 5b (100 nM) (Buta et al., 2015), blocked cingulate LTP induction (Fig. 4 F, G). These results further substantiate that ASIC1a function is crucial for cingulate LTP induction. The animals were pretreated with the selective ASIC1a blocker, PcTx1, injected in ACC. CFA-induced inflammatory pain was then behaviorally assessed. B, Example Nissl-stained image depicting the location of the injection cannula (indicated by the asterisk). Scale bar, 500 m. C, PcTx1 pretreatment prevented CFA-evoked thermal hyperalgesia. n ϭ 10 mice for each group. D, Schematics of the experimental protocol for E. E, PcTx1 did not affect normal heat nociception (n ϭ 8 mice per group). F, Schematics of the experimental protocol for G-I. G-I, Open field test showing no difference between saline-and PcTx1-treated groups in the time spent in the center zone (G), distance traveled in the center arena (H ), or total distance traveled (I ) (saline group, n ϭ 8 mice; PcTx1 group, n ϭ 9 mice). Data are mean Ϯ SEM. **p Ͻ 0.01, ***p Ͻ 0.001, N.S., not significant.

ASIC1a is not required for LTD induction in ACC
We next investigated the effect of ASIC1a deletion/inhibition on cingulate LTD, another form of synaptic plasticity induced by LFS (Liu and Zhuo, 2014). With the conditional KO of ASIC1a in ACC of ASIC1a flox/flox mice using AAV-hSyn-Cre, LTD was still induced by LFS with an induction ratio of 79.0 Ϯ 8.8%, which did not differ significantly from the AAV-hSyn-Ctrl-injected ASIC1a flox/flox mice (77.1 Ϯ 0.5%, t (8) ϭ 0.2215, p ϭ 0.8303 vs AAV-hSyn-Ctrl, unpaired Student's t test; Fig.  4 H, I ). Consistently, inhibiting ASIC1a function with PcTx1 or 5b also failed to prevent LFS-induced cingulate LTD (Fig.  4 J, K ). Together, these findings indicate that ASIC1a selectively regulates the inducibility of LTP in ACC, without affecting LTD.

ASIC1a in excitatory neurons of ACC is the major player in LTP and pain hypersensitivity
The aforementioned results show an important role of ASIC1a in the cingulate LTP and behavioral sensitization to peripheral inflammation. We next asked whether there is any cell type-related difference for the involvement of ASIC1a in pain hypersensitivity. Previous optogenetic studies have revealed that excitatory neurons are main contributors in ACC that modulate pain perception (Kang et al., 2015;Sellmeijer et al., 2018). Therefore, we first specifically knocked out ASIC1a in ACC excitatory neurons, by injecting AAV driven by the CaMKII promoter for excitatory neuron-specific Cre recombinase expression (AAV-CaMKII-Cre) into the ACC of ASIC1a flox/flox mice (Fig. 5 A, B). The AAV-CaMKII-EGFP served as a control (AAV-CaMKII-Ctrl). showing no effect of ASIC1a deletion on action potential firing properties in ACC pyramidal neurons. B, Intensity-dependent increase in number of spikes elicited in ACC. C, Action potential amplitude. D, Action potential half-width. E, Resting membrane potential of ACC neurons in ASIC1a WT and KO mice (n ϭ 8 neurons/3 or 4 mice). F, G, Representative traces (F ) and quantification of evoked EPSCs (G) showing no effect of ASIC1a deletion on input-output relationship of excitatory synaptic transmission in ACC (n ϭ 8 neurons/4 mice). H, I, Representative plots (H ) and quantification (I ) of paired-pulse ratio recordings showing no effect of ASIC1a deletion on the probability of presynaptic neurotransmitter release (n ϭ 9 -11 neurons/4 or 5 mice). J, K, Representative traces (J ) and quantification (K ) of AMPAR/NMDAR ratios in ACC demonstrating no effect of ASIC1a deletion (n ϭ 10 -12 neurons/3 mice). L-N, Representative traces (L) and quantification (M,N ) revealing no obvious difference in the amplitude (M ) or frequency (N ) of mEPSCs recorded in ACC pyramidal neurons (n ϭ 14 -16 neurons/4 or 5 mice). Data are mean Ϯ SEM. N.S., not significant. Confocal imaging analysis showed that most GFP-positive cells colocalized with the marker of excitatory neurons, Vglut2, confirming the specificity of virus infection (Fig. 5C). Western blotting data demonstrated the significant reduction of ASIC1a protein expression in the conditionally deleted group compared with the control (t (6) ϭ 2.98, p ϭ 0.0246, unpaired Student's t test; Fig. 5D). Behaviorally, the selective deletion of ASIC1a in ACC excitatory neurons significantly reduced CFA-evoked thermal and mechanical hypersensitivity (thermal: F (1,9) ϭ 6.75, p ϭ 0.0288, two-way repeated-measures ANOVA; Tukey's multiplecomparison test: day 1: p ϭ 0.0003; day 3: p ϭ 0.0097; day 5: p ϭ 0.0483; day 7: p ϭ 0.0489; Fig. 5E; mechanical: F (1,9) ϭ 12.26, p ϭ 0.0067, two-way repeated-measures ANOVA; Tukey's multiplecomparison test: day 1: p ϭ 0.0134; day 3: p ϭ 0.0205; day 5: p ϭ Figure 5. ASIC1a in excitatory neurons of ACC is the major contributor to cingulate LTP and behavioral hypersensitivity. A, Schematic of the experimental procedure for excitatory neuron-specific ASIC1a deletion. B, Schematic diagram of the viral vectors for AAV-CaMKII-Ctrl and AAV-CaMKII-Cre. C, Confocal images showing the colocalization of GFP (i.e., the AAV-CaMKII-EGFP virus) and Vglut2 (excitatory neuronal marker) immunofluorescence. Filled arrowheads indicate the colabeled neurons. Scale bar, 50 m. D, Representative images and quantification of Western blots demonstrating that AAV-CaMKII-Cre injection reduces ASIC1a expression in ACC (n ϭ 4 mice per group). E, F, Conditional deletion of ASIC1a in excitatory neurons significantly attenuated CFA-evoked thermal hyperalgesia (E) and mechanical allodynia (F ). n ϭ 10 mice for each group. G, H, Conditional deletion of ASIC1a in excitatory neurons reduced induction probability of cingulate LTP (G, summary plot of fEPSP slopes; H, induction ratios of LTP). n ϭ 6 or 7 slices/6 or 7 mice. I, J, LTD was not affected by ASIC1a deletion in excitatory neurons of ACC (I, summary plot of fEPSP slopes; J, induction ratios of LTD). n ϭ 6 slices/6 mice. G, I, Insets, Representative fEPSP traces at the time points indicated by numbers in the graph. Calibration: 100 V, 10 ms. G, Arrow indicates the TBS application. I, Horizontal bar represents the LFS delivery to induce LTD. Data are mean Ϯ SEM. *p Ͻ 0.05, **p Ͻ 0.01, ***p Ͻ 0.001, N.S., not significant. 0.0266; day 7: p ϭ 0.0234; Fig. 5F ). Electrophysiologically, LTP induction ratio reduced from 71.5 Ϯ 5.9% for AAV-CaMKII-Ctrl-injected to 21.0 Ϯ 9.3% for AAV-CaMKII-Cre-injected ASIC1a flox/flox mice (t (11) ϭ 4.389, p ϭ 0.0011, unpaired Student's t test; Fig. 5G,H ). However, LTD was induced successfully with an induction ratio of 82.4 Ϯ 7.3% in the AAV-CaMKII-Cre group, comparable with the control (76.2 Ϯ 2.8%, t (10) ϭ 0.7967, p ϭ 0.4441, unpaired Student's t test; Fig. 5 I, J ).

ASIC1a is required for chronic pain-induced membrane translocation of GluA1 in ACC
So far, the above results suggest that ASIC1a in excitatory neurons of the ACC contributes to inflammatory pain sensitization through promoting LTP of synaptic transmission. To explore the molecular mechanisms underpinning the involvement of ASIC1a in cingulate LTP and inflammatory pain, we first examined whether the expression of ASIC1a was upregulated in ACC after CFA injection. However, in contrast to our previous finding that CFA-induced inflammation resulted in increased expression of ASIC1a in dorsal horn of spinal cord (Wu et al., 2004), the total protein levels of ASIC1a remained unaltered at day 1, day 3, and day 7 after CFA injection (Fig. 6 A, B). Yet, the amount of ASIC1a protein on the plasma membrane appeared to increase, suggesting an enhanced membrane trafficking of ASIC1a in response to CFA-induced peripheral inflammation (t (8) ϭ 3.723, p ϭ 0.0058, WT-saline vs WT-CFA, unpaired Student's t test; Fig. 6 E, F ). Second, GluN2B, a critical subunit of NMDAR, has been shown to play critical roles in ACC LTP and chronic pain (Wu and Zhuo, 2009). We observed a significant increase in total protein levels of GluN2B, but not that of GluA1, in ACC of CFA-injected mice (GluN2B: F (3,11) ϭ 4.47, p ϭ 0.0276, one-way ANOVA; Tukey's multiple-comparison test: WT-saline vs WT-CFA, p ϭ 0.0421; Fig. 6C,D). However, this change was unaffected by ASIC1a deletion, suggesting that the upregulation of GluN2B was independent of ASIC1a (one-way ANOVA followed by Tukey's multiple-comparison test: KO-saline vs KO-CFA, p ϭ 0.0476; Fig. 6D).

ASIC1a works through PKC to promote LTP and pain hypersensitivity
Next, we examined the downstream signaling molecules required for ASIC1a-dependent GluA1 membrane trafficking and LTP induction. The atypical PKC isoforms, including PKC and PKM, have been reported to play critical roles in the initiation and maintenance of hippocampal LTP, respectively (Wang et al., 2016). Interestingly, atypical PKC isoforms are also involved in the pathogenesis of chronic pain (Li et al., 2010;Marchand et al., 2011). Because the electrophysiological data showed an induction deficit in the early phase of cingulate LTP following ASIC1a deletion (Fig. 4B), we explored the possibility that PKC is the missing link between ASIC1a and AMPAR tracking during cingulate LTP. We first checked whether PKC activation was affected in the ACC after peripheral inflammation and whether such an effect required ASIC1a. Three days after saline or CFA injection, ACC tissues were collected and subjected to Western blotting analysis, which showed an increase in the phosphorylation levels of PKC, without any detectable change in total expression of PKC. Importantly, ASIC1a deletion blocked the CFA-induced increase in PKC phosphorylation (F (3,20) ϭ 2.523, p ϭ 0.0368, one-way ANOVA; Tukey's multiple-comparison test: WT-saline vs WT-CFA, p ϭ 0.0427; KO-saline vs KO-CFA, p ϭ 0.7440; Fig. 7 A, B), indicating that the phosphorylation of PKC in ACC during inflammatory pain proceeds in an ASIC1adependent manner.

ASIC1a induces PKC phosphorylation through intracellular Ca 2؉ increase
Next, we asked what might be the molecular mechanism linking ASIC1a activation to PKC phosphorylation in ACC. It has been shown that PKC activation relies on elevation of intracellular Ca 2ϩ , which further triggers phosphoinositide 3-kinase activation (Suzuki et al., 2003;Ren et al., 2013). Here, we used primary cultures of cortical neurons to determine whether intracellular Ca 2ϩ elevation caused by ASIC1a is involved in increasing PKC phosphorylation. First, we performed Ca 2ϩ imaging experiments to show that application of the acidic, pH 6.0, solution elicited a rise in intracellular Ca 2ϩ concentration in cultured mouse cortical neurons, and this effect was completely lacking in ASIC1a KO neurons (Fig. 9A). Second, we performed Western blot to illustrate that the acidic solution treatment induced a significant in-crease in the phosphorylation of PKC in the cultured cortical neurons; however, this increase was blocked by the removal of Ca 2ϩ from the treatment solution (F (5,18) ϭ 1.27, p ϭ 0.0448, one-way ANOVA; Tukey's multiple-comparison test: WT-pH 7.4 vs WT-pH 6.0, p ϭ 0.0040; WT-pH 7.4 vs WT-pH 6.0, Ca 2ϩ free, p ϭ 0.8800; Fig. 9 B, C) or the incubation with BAPTA-AM to prevent the elevation of the intracellular Ca 2ϩ (F (5,18) ϭ 2.94, p ϭ 0.0410, one-way ANOVA; Tukey's multiple-comparison test: WT-pH 7.4 vs WT-pH 6.0, p ϭ 0.0072; WT-pH 7.4 vs WT-pH 6.0, BAPTA-AM, p ϭ 0.9459; Fig. 9 D, E). Moreover, the acid-induced increase in PKC phosphorylation was not seen in ASIC1a KO neurons (Fig. 9B-E). These findings suggest that, in cortical neurons, acid activation of ASIC1a triggers PKC phosphorylation in a Ca 2ϩ -dependent manner.
It has been suggested that ACC mainly plays a role in the affective (or aversive) component of pain. Several research groups, using multiple pain conditions and interventions, have shown that ACC excitation correlates with increased aversive responses to pain (LaGraize et al., 2004;LaGraize and Fuchs, 2007;Qu et al., 2011;Navratilova et al., 2015;Sellmeijer et al., 2018). Therefore, we tested whether ASIC1a in the ACC contributes to pain-related negative affect. We used pain-evoked PEAP to evaluate the affective responses to chronic pain as previously described (LaBuda and Fuchs, 2000;Han et al., 2014). Surprisingly, we found that postinjection of compound 5b, at the same dose as that to produce the aforementioned antiallodynic effects, could not relieve the aversiveness of either CFA-induced inflammatory pain or SNI-induced neuropathic pain (Fig. 11). Compared with the control group (intraplantar saline plus intra-ACC saline, or sham injury plus intra-ACC saline), animals with intra-ACC administration of either saline or 5b spent significantly more time in the light half of the chamber to avoid stimulation of the inflamed hindpaw (F (2,18) ϭ 268.3, p Ͻ 0.0001, two-way repeatedmeasures ANOVA; Tukey's multiple-comparison test: CFAsaline vs CFA-5b, p ϭ 0.3155, p ϭ 0.1334, p ϭ 0.6613, p ϭ 0.9904, p ϭ 0.8928, and p ϭ 0.7176 for 5, 10, 15, 20, 25, and 30 min, respectively; Fig. 11 A, B) Fig. 11 E, F ). These findings indicate that ASIC1a in the ACC contributes pain hypersensitivity through sensory, rather than affective, aspect of pain processing.

Discussion
ASIC1a channel has been well documented to contribute to pathophysiology of several neurological diseases, such as chronic pain (Duan et al., 2012), stroke (Wang et al., 2015), and epilepsy (Ziemann et al., 2008). The present study reveals a novel function of ASIC1a in ACC, where it is abundantly expressed and criti- Figure 10. Inhibiting ASIC1a in ACC reverses the preestablished inflammatory and neuropathic pain hypersensitivity. A, Schematic diagram of the experimental protocol for B. PcTx1 was microinfused into ACC at day 3 after CFA injection. B, Post-treatment with PcTx1 at ACC reversed CFA-induced thermal hyperalgesia. n ϭ 10 mice for each group. C, Schematic diagram of the experimental protocol for D. Similar to A, but another ASIC1a inhibitor, compound 5b, was injected in ACC to determine the analgesic effects. D, Compound 5b significantly reversed inflammatory pain in WT mice but had no effect on ASIC1a KO mice, excluding any nonspecific action of compound 5b. n ϭ 10 mice for each group. E, Schematic diagram of the experimental protocol for F. PcTx1 or saline was microinfused into ACC at day 7 after SNI. F, Post-treatment with PcTx1 reversed SNI-induced mechanical allodynia. n ϭ 10 mice for each group. G, H, Compound 5b produced a similar antihyperalgesia effect in the SNI model of neuropathic pain. n ϭ 10 mice for each group. I, Schematic diagram of the experimental protocol for J. PcTx1 or saline was microinfused into ACC at day 14 after SNI. J, Post-treatment with PcTx1 still reversed SNI-induced mechanical allodynia at the later stage of neuropathic pain. n ϭ 10 mice for each group. K, L, Intra-ACC administration of compound 5b produced a similar antihyperalgesia effect at day 14 after SNI. n ϭ 10 mice for each group. Data are mean Ϯ SEM. *p Ͻ 0.05, **p Ͻ 0.01, ***p Ͻ 0.001. cally involved in nociceptive hypersensitivity through modulating the strength of excitatory synaptic transmission. Specifically, we show that disrupting ASIC1a expression/or function in ACC by either genetic deletion or pharmacological blockade significantly relieves CFA-induced heat hyperalgesia and mechanical allodynia, as well as the enhancement of cortical synaptic adaptations caused by chronic inflammatory pain. Moreover, we reveal that ASIC1a participates in cingulate LTP induction through a PKC-dependent increase in GluA1 receptor membrane trafficking, eventually leading to enhanced synaptic efficacy and behavioral sensitization. This represents the first demonstration of the synaptic role of ASIC1a in ACC. Another novelty of this study lies in the demonstration of PKC in cortical processing of pain, while most previous studies have focused on the role of spinal PKC in nociceptive processing. More importantly, we show that the preestablished pain hypersensitivity can be reversed by drug inhibition of cingulate ASIC1a. These results not only extend our understanding of the supraspinal role of ASIC1a but also provide proof of principle for the therapeutic potential of targeting ASIC1a to treat chronic pain.
ASICs are excitatory cation channels directly gated by extracellular protons (Kellenberger and Schild, 2015). Growing studies have reported the importance of ASIC1a in multiple aspects of nociception as well as several animal models of chronic pain. However, the majority of these studies focused on the peripheral and spinal ASIC1a, with little attention paid to the function of cortical ASIC1a in pain processing (Wemmie et al., 2013;Deval and Lingueglia, 2015). Here, we demonstrate the role of ASIC1a in ACC by first showing its contribution to the development of inflammatory pain hypersensitivity. To selectively manipulate ASIC1a in ACC, we used a number of genetic and pharmacological strategies. First, we injected AAV-Cre virus into ACC of ASIC1a flox/flox mice to specifically delete the channel in this brain area. This resulted in attenuated heat and mechanical hypersen-sitivity in the CFA model of inflammatory pain. Second, we knocked down ASIC1a expression in ACC by shRNA and obtained similar analgesic effects. Third, preemptive treatment with PcTx1 reduced the elevated inflammatory pain sensitivity. Last, we selectively disrupted ASIC1a in excitatory neurons of ACC and showed them to be critical for inflammatory sensitization. Together, these findings strongly argue for the importance of cingulate ASIC1a in gating sensory hypersensitivity under inflammatory conditions.
It is well documented that ASIC1a, apart from being involved in multiple pathological processes, exerts its physiological functions by regulating synaptic transmission and plasticity in the brain . However, ASIC1a appears to play differential roles in synaptic plasticity in different brain regions. For instance, ASIC1a is important for LTP induction in hippocampus and amygdala Chiang et al., 2015;Liu et al., 2016), but it mainly regulates LTD inducibility in insular cortex . These differences may underlie the variable effects of the channel on behavioral responses of various tasks. Here, we found ASIC1a to selectively regulate LTP induction in ACC. The exact reason(s) for the differential effects of ASIC1a on synaptic plasticity is not fully understood. The different locations of ASIC1a in distinct cell types from different brain regions, as well as the selective alterations in specific signaling pathways related to LTP or LTD resulting from ASIC1a inhibition, could all contribute to the diverse outcomes . More interestingly, ASIC1a appears to use different downstream signaling pathways to regulate the same type of synaptic plasticity, such as LTP. In hippocampus, ASIC1a modulates LTP through both NMDAR-dependent and -independent pathways . In ACC, however, ASIC1a works through PKC to facilitate GluA1 synaptic delivery and in turn synaptic potentiation (see below). Therefore, the mechanisms by which ASIC1a regulates synaptic plasticity can be quite complex and diverse, Figure 11. Inhibiting cingulate ASIC1a fails to affect chronic pain-related negative aversion. A, Schematic diagram of the experimental protocol for B. The PEAP test was performed at 10 min after 5b infusion at day 3 after CFA injection. B, Intra-ACC administration of 5b did not reverse inflammatory pain-induced negative aversion in the PEAP. Compared with the saline-injected control group, CFA-inflamed animals with intra-ACC injection of either saline or 5b spent similarly more time in the light side of the chamber to avoid the noxious stimulation in the dark side. n ϭ 10 mice for each group. C, Schematic diagram of the experimental protocol for D. The PEAP test was performed at 10 min after 5b infusion at day 7 after SNI induction. D, Intra-ACC administration of 5b did not reverse neuropathic pain-induced negative aversion in the PEAP test. n ϭ 10 mice for each group. E, Schematic diagram of the experimental protocol for F. The PEAP test was performed at 10 min after 5b infusion on day 14 after SNI induction. F, Intra-ACC administration of 5b did not reverse SNI-induced negative aversion at the later stage of neuropathic pain. n ϭ 10 mice for each group. ***p Ͻ 0.001, N.S., not significant.
varying greatly depending on brain regions, cell types, and downstream signaling pathways.
LTP in sensory pathways has been proposed to be one of the most important cellular substrates for pain chronicity (Luo et al., 2014). The present results clearly show that ASIC1a is involved in both cingulate LTP and inflammatory pain hypersensitivity. To define the underlying mechanism, we focused on molecular processes that have been reported to contribute to LTP. A large body of literature supports the importance of AMPAR trafficking in LTP expression in several brain regions, including hippocampus and ACC (Toyoda et al., 2007;Penn et al., 2017;Zhou et al., 2018). Accordingly, we reveal that ASIC1a contributes to CFA-induced upregulation of membrane content of GluA1containing AMPARs in ACC. This finding is in line with the previous report showing that ASIC1a promotes calciumpermeable AMPAR activity under conditions of ischemia or acidosis (Quintana et al., 2015). Interestingly, previous studies have stressed the NMDAR dependence of multiple ASIC1a-regulated functions, which occurs indirectly through ASIC1a-induced membrane depolarization due to Na ϩ influx mediated by the channel (Gao et al., 2005;Wemmie et al., 2013;Yu et al., 2018). Although we do not exclude the possibility that ASIC1a contributes to cingulate LTP also through an NMDAR-dependent mechanism, our previous work indicates that ASIC1a mediates hippocampal LTP induction through both NMDAR-dependent and NMDAR-independent mechanisms. Thus, it would be important for future studies to determine whether and to what extent ASIC1a exerts its effect on cingulate LTP induction through NMDARs. The evidence collected here supports PKC as a critical mediator that links ASIC1a to GluA1 trafficking. Whereas the knockdown of PKC in ACC resulted in LTP impairment and behavioral analgesia, overexpression of exogenous PKC in ACC restored inflammatory hyperalgesia in ASIC1a KO mice. More critically, the knockdown of PKC did not further increase the pain threshold in conditional ASIC1a KO mice, making it unlikely that ASIC1a and PKC play parallel or additive roles in cingulate LTP and inflammatory pain. Our findings strongly suggest that ASIC1a promotes cingulate LTP and pain sensitization through, at least partially, PKC-regulated membrane trafficking of GluA1. In line with this assessment, previous work has shown PKC to be crucial for synaptic incorporation of AMPARs during hippocampal LTP and trace fear conditioning (Ren et al., 2013;Wang et al., 2016). Together, our present study adds a new dimension for cortical function of PKC, revealing its regulation by ASIC1a and contribution to cortical synaptic potentiation and behavioral sensitization to inflammatory pain.
Modulation of ion channel signaling has the potential to effectively treat pain. Ion channel modulators, such as lidocaine and pregabalin, act on voltage-gated sodium channels and calcium channels to reduce pain in the clinical setting (Skerratt and West, 2015). As a type of nonclassical ion channels, ASIC1a plays an important role in chronic pain processing at both peripheral and central levels of the sensory axis (Deval and Lingueglia, 2015). Moreover, deletion or inhibition of ASIC1a does not obviously affect normal sensation or other important physiological functions, which may make ASIC1a a better druggable target with fewer adverse effects. The present results clearly show that not only can a peptide inhibitor of ASIC1a, PcTx1, reverse the established pain hypersensitivity with intra-ACC administration, but the novel small-molecule inhibitor of ASIC1a, compound 5b, also exerts a similar analgesic effect. Importantly, the analgesic effect was demonstrated in both inflammatory and neuropathic pain models. Because of great difficulties in delivering large pep-tides for clinical use, the small-molecule-based 5b provides an excellent base for developing more potent and orally available ASIC1a drugs to alleviate chronic pain in the future.
In conclusion, we report the supraspinal role of cortically located ASIC1a in pain processing using multiple neurobiological approaches. ASIC1a critically gates pain sensitivity through regulating the probability of cingulate LTP induction, in which PKC serves as a crucial mediator. Together with previous reports on the pronociceptive function of ASIC1a in peripheral and spinal levels, our findings highlight the promising potential of targeting ASIC1a for future development of new molecular medications for chronic pain. This is of great clinical significance considering the huge unmet need in clinical pain management.