Proton-Activated Chloride Channel Increases Endplate Porosity and Pain in a Mouse Spinal Degeneration Model

Chronic low back pain (LBP) can severely affect daily physical activity. Aberrant osteoclast-mediated resorption leads to porous endplates for the sensory innervation to cause LBP. Here, we report that the expression of proton-activated chloride (PAC) channel is induced during osteoclast differentiation in the porous endplates via a RANKL-NFATc1 signaling pathway. Extracellular acidosis evokes robust PAC currents in osteoclasts. An acidic environment of porous endplates and elevated PAC activation-enhanced osteoclast fusion provoke LBP. Further, we find that genetic knockout of PAC gene Pacc1 significantly reduces endplate porosity and spinal pain in a mouse LBP model, but it does not affect bone development or homeostasis of bone mass in adult mice. Moreover, both osteoclast bone resorptive compartment environment and PAC traffic from the plasma membrane to endosomes to form an intracellular organelle Cl channel have low pH around 5.0. The low pH environment activates PAC channel to increase sialyltransferase St3gal1 expression and sialylation of TLR2 in initiation of osteoclast fusion. Aberrant osteoclast-mediated resorption is also found in most skeletal disorders, including osteoarthritis, ankylosing spondylitis, rheumatoid arthritis, heterotopic ossification, enthesopathy. Thus, elevated Pacc1 expression and PAC activity could be a potential therapeutic target for LBP and osteoclast-associated pain.


Introduction
Skeletal disorders including osteoarthritis and spine degeneration are often associated with pain.
Pain is a major reason people seek medical attention.Chronic low back pain (LBP) profoundly affects quality of life and daily physical activity, especially in the elderly population, and thus it is a key risk factor for the development of future health decline (1)(2)(3).Most LBP is nonspecific with no apparent pathoanatomical cause (4).Therefore, to understand the source of LBP and its underlying mechanism is essential for its therapy.The cartilaginous endplate is composed of a thin layer of hyaline cartilage positioned between the vertebral endplate, the coronal surface of each vertebra, and the nucleus pulposus, which is the inner core of the vertebral disc that acts as the shock absorber for each spinal unit.Our previous studies have demonstrated that aberrant osteoclast-mediated resorption of calcified cartilaginous endplates during spine degeneration generates a porous structure, and osteoclasts in porous endplates secrete Netrin-1, that allows for sensory innervation of the spinal unit, thus leading to LBP (5)(6)(7).But it is still unclear how osteoclast activity becomes aberrant during spinal degeneration.
Pain in skeletal disorders is often associated with aberrant osteoclast resorption in a very low pH environments (8,9).Our recent study reveals that proton-activated chloride (PAC) channel is newly identified proton-activated Cl − channel to be encoded by Pacc1 (also known as TMEM206) opened by acidic pH.The channel is responsive to pathological acidic pH in ischemic brain injury and acid-induced neuronal cell death in mice (10).PAC represents a completely new ion channel family which has no obvious sequence homology to other membrane proteins, but it is highly conserved in vertebrates (11).There are two transmembrane (TM) helices in PAC, like the acid-sensing ion channel (ASIC) and the epithelial sodium channel (ENaC) (12).PAC structure study showed that the protein exists in two states: namely, a high-pH resting closed state and a low-pH proton-bound non-conducting state.PAC channel undergoes striking conformational changes when the pH drops from 8 to 4, leading to an opening of the channel and the conduction of anions across cellular membranes, thereby inducing diseases associated with tissue acidosis (acid-induced cell death).Interestingly, PAC also localizes in intracellular organelles (endosomes and macropinosomes) and regulates their pH and volume homeostasis (13).The aberrant osteoclast bone remodeling in some of major bone diseases such as osteoarthritis, rheumatoid arthritis and low back pain are likely -4 -associated with low pH, therefore we set out to investigate the function of the PAC channel in osteoclast differentiation and resorption for LBP.
Osteoclasts are formed from macrophages/monocytes that in response to NF-kappa B ligand (RANKL) signaling become committed to a tartrate-resistant acid phosphatase (TRAP)positive osteoclast lineage (14)(15)(16).TRAP + mononuclear cells first attach to the bone surface and then undergo fusion to form multinucleated osteoclasts (17).Osteoclasts are polarized to form ruffle membranes with abundant ATPase proton pump activity at the side attached to the bone (18).Chloride channels such as CIC-7 are important in the cell membrane and intracellular organelles (19).In osteoclasts, CIC-7 is predominantly localized to the ruffled border, a specialized membrane domain crucial for acidifying the resorption lacuna.This acidification process supports ATPase proton pump activity, thereby enabling bone resorption.It also expressed at the membrane of lysosome/endosome (20).At both lysosome/endosome membrane and the ruffled border, CIC-7 functions on the ionic homeostasis, and maintain the pH (21,22).The compartment between the ruffled membrane and the bone surface within the osteoclast sealing zone is acidified by secretion of protons, leading to the dissolvement of the bone matrix material (23,24).Aberrant osteoclast-mediated bone resorption and the secretion of Netrin-1 are associated with many major skeletal disorders, including low back pain (LBP), osteoarthritis, heterotopic ossification and ankylosing spondylitis, among others (5,(25)(26)(27)(28)(29).
PAC is activated at low pH environment with implication of bone diseases, and it is imperative to know its function in osteoclast differentiation and resorption in conjunction with CIC-7.
In this study, we investigated the potential role of PAC in osteoclast function of the porous endplates in spine degeneration-associated LBP.We found that PAC expressed on both cell

Knockout of Pacc1 significantly reduces spinal pain and endplate porosity in a mouse model of spine degeneration
During spinal degeneration, osteoclast activity is stimulated, leading to porosity of the endplates, and Netrin-1 secreted by osteoclasts induces CGRP + nerve innervation and thus causes LBP.To examine the potential role of PAC in osteoclast-mediated resorption of the endplate and the generation of its porosity, we utilized mice with genetic deletion of Pacc1 (10).
We then surgically manipulated the Pacc1 -/-mice and their wild-type (WT) littermates to generate a mouse model of lumbar spine instability (LSI) as a form of spinal degeneration.By pain behavior tests, we found that pressure tolerance was significantly lower in the WT LSI mice at 4-and 8-weeks post-LSI induction compared to sham-treated WT mice, whereas the degree of change in pressure tolerance in the Pacc1 -/-LSI mice was significantly less than in the WT LSI mice but still lower than the sham-operated mutant mice (Figure 1A).These results suggest that Pacc1-encoded activity is associated with endplate porosity-induced LBP.Furthermore, by measuring the paw withdrawal frequency (PWF) in von Frey tests to evaluate mechanical pain hypersensitivity, we found that mechanical hyperalgesia was significantly lower at 4-and 8-weeks post-LSI surgery in Pacc1 -/-LSI mice relative to their WT LSI littermates (Figure 1B).We also conducted a spontaneous activity behavior test.The distance traveled and active time per 24 hours were significantly greater in Pacc1 -/-LSI mice relative to their WT LSI littermates (Figure 1, C and D).
Next, we examined the sclerosis endplates by μCT and found that endplate porosity was significantly lower in Pacc1 -/-LSI mice relative to WT LSI mice 8 weeks post-LSI induction (Figure 1, E-G).By immunostaining for CGRP as a readout of innervation, we found that Pacc1 -/-LSI mice showed less sensory nerve innervation in the porous endplates compared to WT LSI mice (Figure 1H).We then stained L4-L5 caudal endplates with Safranin O and Fast Green staining and found greater cartilage area and less porosity proportion in the endplate of the Pacc1 -/-LSI mice relative to WT LSI mice at 8 weeks post-surgery (Figure 1, I and J).
Strikingly, by TRAP staining we found that the number of large multinuclear osteoclast cells were lower in the Pacc1 -/-LSI mice relative to WT LSI mice at 8 weeks post-surgery (Figure 1, K and L), suggesting that Pacc1 expression promotes aberrant osteoclast fusion and function.
Together, these results indicate that PAC plays a role in pathological osteoclast bone resorption during the generation of endplate porosity and LBP during LSI.
We also examined whether PAC expression plays a physiological role in normal osteoclast bone resorption, bone development and bone remodeling.First, we investigated the effect of PAC on skeletal development and adult bone homeostasis.We examined the body length and weight at 1, 3 and 6 months of age, and we found that Pacc1 -/-mice showed no difference with their WT littermates (Figure 2, A and B).Notably, we measured various bone parameters during development and adulthood by μCT at 1, 3 and 6 months of age.We found that important bone parameter bone volume in Pacc1 -/-mice was not different than WT Pacc1 +/+ mice at 3 months of age (Figure 2, C and D).Furthermore, by TRAP staining of femur sections, we found that osteoclast numbers did not change in Pacc1 -/-mice relative to WT Pacc1 +/+ mice at 3 months of age (Figure 2, E and F).Further, by osteocalcin staining, we found no difference in bone formation between Pacc1 -/-and WT mice at 3 months of age (Figure 2, G and H).Taken together, these results indicate that PAC expression does not play a critical role in osteoclast bone resorption during bone development or in bone homeostasis under physiological conditions.

PAC expression is induced during osteoclast differentiation
To examine whether PAC expression is induced during osteoclast differentiation and resorption, whole-bone marrow cells were isolated from C57BL/6 mouse hind limbs and cultured with M-CSF (50 ng/mL) for 2 days to induce growth of bone marrow macrophages (BMMs).BMMs were then treated with M-CSF (30 ng/mL) and RANKL (100 ng/mL) for 5 days to induce osteoclast differentiation, as well as their fusion into mature osteoclasts (mOCs) (Figure 3A).
Pacc1 mRNA expression was significantly induced by RANKL stimulation relative to the M-CSF group at 1, 3 and 5 days and peaked at day 3 as measured by RT-PCR (Figure 3B).To confirm expression of the Pacc1-encoded channel protein during osteoclast differentiation, BMMs were treated with RANKL and harvested at day 0, 1, 3 and 5 for Western blot analysis of PAC expression.Consistent with the pattern of mRNA expression, PAC expression was higher in the RANKL-treated cells compared to the M-CSF-only treated cells with peak expression at day 3 (Figure 3, C and D).We next examined the expression of NFATc1, a downstream transcription factor induced by RANKL signaling, and found that its expression was greater with treatment of RANKL in a time-dependent manner that matches that of PAC, suggesting that NFATc1 transcriptionally regulates Pacc1 expression.We found three potential NFATc1-binding sites in the Pacc1 promoter (Figure 3E), and by a chromatin immunoprecipitation (ChIP) assay we demonstrated that RANKL induced specific binding of NFATc1 to the most proximal NFATc1-binding site of the Pacc1 promoter to activate its expression in osteoclast cells (Figure 3, F and G).Thus, PAC expression is induced during osteoclast maturation and bone resorption.

Extracellular acidosis evokes the robust PAC currents in osteoclasts
To examine the functional activity of PAC, BMMs were isolated from WT and Pacc1 -/-mice and cultured with neutral (pH = 7.4) or acidic (pH = 6.8) medium containing M-CSF and RANKL.By RT-PCR and Western blot analysis, extracellular acidosis did not influence Pacc1 or PAC expression at 1, 3 and 5 days of osteoclast differentiation (Figure 4, A-C).Furthermore, co-immunostaining of PAC with TRAP on human bone sections demonstrated that PAC was expressed on the cellular membrane and intracellular organelles of osteoclasts and well colocalized with TRAP staining at bone surface (Figure 4D).Indeed, by whole-cell patch clamping we found that extracellular acidosis evoked the proton-activated chloride currents in Pacc1 wt preosteoclasts at day 3 after RANKL treatment, while they were absent in Pacc1 -/- cells (Figure 4, E and F).These results demonstrate that PAC is functionally expressed in osteoclast lineage cells and the channel is activated under acidic conditions.The results suggest that the function of PAC is not specific for osteoclast bone resorption like chloride transporter CIC-7, but likely in regulation of osteoclast fusion under low pH environment.

Knockout of Pacc1 in TRAP + cells reduce spinal pain and endplate porosity in the LSI model
To examine how PAC expression in osteoclast induces LBP, we crossed the floxed Pacc1 mice (Pacc1 flox/flox ) with TRAP-Cre to generate conditional Pacc1TRAP -/-mice.We then conducted a series of pain behavior tests with the Pacc1TRAP -/-mice and found that pressure tolerance was significantly greater in Pacc1TRAP -/-LSI mice relative to the Pacc1 wt LSI mice at both 4-and 8weeks post operation (Figure 5A).Moreover, mechanical hyperalgesia, as measured by von Frey tests, was lower in Pacc1TRAP -/-LSI mice at 4 and 8 weeks after operation compared to Pacc1 wt LSI mice (Figure 5B).In addition, distance traveled and active time per 24 hours spontaneous activity were significantly greater in Pacc1TRAP -/-LSI mice relative to Pacc1 wt LSI mice at both 4-and 8-weeks post-surgery (Figure 5, C and D).As expected, by μCT scanning we found that there was less endplate porosity and Tb.Sp in Pacc1TRAP -/-LSI mice relative to Pacc1 wt LSI mice at 8 weeks after surgery (Figure 5, E-G).And by immunostaining CGRP we found less sensory nerve innervation in the porous endplates in Pacc1TRAP -/-LSI mice compared to Pacc1 wt LSI mice (Figure 5H).Therefore, PAC expression enhanced osteoclast resorptive activity to generate porous endplates, leading to pain hypersensitivity in a mouse model of spine degeneration.

PAC-mediated I Cl,H current activity in response to extracellular acidosis enhances osteoclast fusion
Next, we investigated PAC functional activity during osteoclast-mediated bone resorption.
Osteoclasts prepared from either Pacc1 -/-or Pacc1 WT mice were cultured on bone slices to examine their bone resorptive activity, which is referred to as a pit assay.The bone resorptive areas were significantly greater underneath osteoclastic cells from Pacc1 WT mice in pH 6.8 medium relative to pH 7.4 (Figure 6, A and B), an effect that was blunted among osteoclastic cells from Pacc1 -/-mice in pH 6.8 medium, indicating that PAC expression enhances osteoclast reportion in a mildly low pH environment.We recorded a higher ICl,H current on the membrane in TRAP + mononuclear cells than in multinuclear cells at 3 days after RANKL treatment (Figure 6C).The PAC-mediated ICl, H current of multinuclear cells was significantly lower at day 5 relative to day 3 after RANKL treatment (Figure 6D), indicating that PAC primarily regulates TRAP + mononuclear cells prior to their osteoclast fusion.Moreover, by TRAP staining we found that extracellular acidosis accelerated osteoclast fusion in BMMs isolated from Pacc1 +/+ mice in pH 6.8 medium at day 3 and 5 after RANKL treatment (Figure 6, E and G, Supplemental Figure 1A).The results were further confirmed by phalloidin staining (Figure 6, F and H, Supplemental Figure 1A), which showed that osteoclast fusion in response to an acidic medium was blunted in BMMs from Pacc1 -/-mice.Taken together, our data indicate that PAC is required for osteoclast fusion at low pH.

PAC is essential for sialyltransferase St3gal1 expression for osteoclast fusion by inducing sialylation of TLR2
We have reported that sialyltransferase St3gal1-mediated sialylation of TLR2 on preosteoclast initiates osteoclast fusion (30,31).To investigate whether PAC regulates osteoclast fusion through the sialylation of TLR2, BMM isolated from Pacc1 +/+ and Pacc1 -/-mice were treated with M-CSF and RANKL for three days at acidic and physiological pH conditions as we described before.The total protein was harvested for Western Blot analysis.We found that the expression of St3gal1 was minimum expressed at pH 7.4 but significantly increased at pH 6.8.Importantly, St3gal1 expression was significantly decreased in Pacc1 -/-cells at both acidic and physiological pH medium (Figure 7, A and B ) suggesting that activation of PAC at low pH promotes expression of St3gal1 to initiate preosteoclast fusion through St3gal1-mediated TLR2 sialylation.To examine PAC essential for St3gal1 expression to initial osteoclast fusion, we knocked down St3gal1 in Pacc1 +/+ or Pacc1 -/-BMMs using siRNA against St3gal1 (150809, Thermo Fisher), and the result showed the number of fused cells significantly decreased in St3gal1 siRNA treated groups relative to control siRNA groups at different pH condition (Figure 7, C and D, Supplemental Figure 1, B-E).Osteoclast fusion marker, OC-STAMP, was also significantly decreased in the St3gal1 knocked down groups at different pH environments (Supplemental Figure 1, F-I).Thus, under low pH microenvironment, St3gal1 expression requires PAC activity to induce sialylation of TLR2 for osteoclast fusion.

Discussion
Endplates are cartilaginous structures connecting the vertebral body with the intervertebral disc in the spine.Endplates undergo porous sclerosis with partial ossification in patients with spine degeneration.Normally osteoclasts do not resorb cartilage, but they can target partially calcified cartilage in the porous endplates, in which the confined environment is marked by sensory innervation, angiogenic type H vessels and a low pH.Importantly, aberrant osteoclast resorption is found in many skeletal disorders with similar pathological environments including osteoarthritis (25,29), ankylosing spondylitis (26), rheumatoid arthritis (32,33), enthesopathy (34), spine degeneration(5), heterotopic ossification (28), and Paget disease (35).
The osteoclast-mediated resorptive activity in the pathological environment produces excessive Netrin-1, PDGF-BB, TGF-β and IGF-1 to promote the progression of skeletal disorders and pain (36)(37)(38).In our LSI animal model, the pH in the porous endplates we measured was 6.92 ± 0.08 (Supplementary Figure 2A).In addition, the Warburg effect is known as an important mechanism in generation of acidic microenvironment by elevated expression of lactic dehydrogenase A (LDHA) (39,40).LDHA significantly accumulated in endplates of LSI mice (Supplementary Figure 2, B and C).Moreover, PAC expression is induced by NFATc1 with RANKL stimulation on the membrane and intracellular organelles of osteoclasts as RANKL induces the commitment of macrophages to the TRAP + osteoclast lineage for osteoclast fusion and maturation (30,31).Interestingly, both the osteoclast bone resorptive compartment environment and PAC traffic from the plasma membrane to endosomes to form an intracellular organelle Cl channel have low pH around 5.0.The low pH environment activates PAC channel to increase St3gal1 expression for sialylation TLR2 in initiation of osteoclast fusion.
Chloride channels/transporters such as CIC-7, coded by Clcn7 gene, are important in the cell membrane and intracellular organelles (19).CIC-7 is primarily localized at the ruffled border of osteoclasts.Ruffled border is a special membrane area that important for the acidification and bone resorption.It also expressed at the membrane of lysosome/endosome (20).At both lysosome/endosome membrane and the ruffled border, CIC-7 contributes to the ionic homeostasis, and maintain the pH (21,22).Importantly, CIC-7 provides the chloride conductance in endosome/lysosome, along with the proton pumping in the osteoclast ruffled membrane (19,20,22).Comparing with CIC-7 expression is polarized in the osteoclast ruffled membrane to provide the chloride conductance in endosome/lysosome, along with the proton pumping, the PAC is evenly distributed on the membrane of TRAP + mononuclear cells and osteoclasts with primary function for osteoclast fusion.PAC is activated at low pH to induce expression of St3gal1 to sialylated TLR2 for fusion of TRAP + mononuclear cells (Figure 7, A     and B).Knockout of Pacc1 reduced osteoclast fusion in the endplate of LSI mice, whereas osteoclast fusion and bone resorption were not affected at normal bone.At low pH, Pacc1 -/- preosteoclasts are difficult to fuse for the formation of osteoclasts fusion.Thus, one of the important functions of PAC is to maintain osteoclast fusion at low pH environment.
We demonstrated that acidic environment promoted Pacc1 +/+ osteoclasts fusion but not Pacc1 - /-osteoclasts, while elevation of resorption was around 50% due to the "sudden stimulus" instead of continuous cultivation as Tim Arnett described.Tim Arnett et al. uncovered a significant impact of extracellular protons on the osteoclast bone resorption.This study represents the initial direct evidence that low pH enhances cell-mediated bone resorption (41).Furthermore, they found that rat osteoclasts may be more sensitive to stimulation by CO2 acidosis than by HCO3 -acidosis (42).His group investigated the effect of small shifts in extracellular pH on the resorptive activity of rat osteoclasts in vitro and found that very slight alterations in ambient hydrogen ion concentration can effectively "switch on" or "switch off' rat osteoclasts in vitro (43).His group also examined the effects of HCO3 -and CO2 acidosis on osteoclast-mediated Ca 2+ release from 3-day cultures of neonatal mouse calvaria and found that addition of H + reduced pH from 7.12 to 7.03 and increased Ca 2+ release 3.8-fold, and CO2 acidosis was a less effective stimulator of Ca 2+ release than HCO3 -acidosis over a similar pH range (44).
TRAP could be expressed in other cell types, such as leukocytes.Based on our results, bone homeostasis remained unchanged in PAC global knockout mice.Furthermore, TRAP conditional knockout Pacc1TRAP -/-mice displayed osteoclast functional outcomes similar to those in global knockout Pacc -/-mice.Thus, knocking out Pacc1 in leukocytes or other cell types is unlikely to exert significant indirect effects on the function of PAC in osteoclast fusion.
Attachment of TRAP + mononuclear preosteoclasts on the bone surface initiates fusion to form polarized multinucleated osteoclastic cells (45).The compartmentalized resorption environment is established by a circumferential attachment sealing zone (46).The plasma membrane within the sealing zone develops the ruffled border with abundant V-type H1-ATPase proton pump activity (47).Across this membrane the osteoclasts actively secrete HCl into the compartment to dissolve the bone matrix.First, the ATPase proton pump inserts H + into the resorption compartment and then chloride ions passively cross the membrane via chloride channel (22).These two steps of the proton pump and chloride channel forms HCl to acidify the resorption compartment and alkalinize the cytoplasm.At physiological condition of bone remodeling, the basolateral bicarbonate chloride exchanger corrects the cytoplasmic alkalinization by compensating for cytoplasmic chloride loss (48,49), while the PAC channel involved in chloride exchange on cell membrane as well.However, at low pH in the porous endplates, PAC is highly activated to promote chloride transport, which leads to aberrant osteoclast fusion and the development of LBP (Figure 7E).The aberrant osteoclast activity leads to the secretion of many factors, including Netrin-1 and PDGF-BB, to induce sensory innervation and angiogenesis in the porous endplates that ultimately leads to LBP and spine degeneration (5,38).Knocking out Pacc1 significantly reduced endplate porosity and LBP.
Therefore, PAC expression and activation in osteoclasts could be a potential therapeutic target for LBP or joint arthritis pain.

Sex as a biological variable
Our study exclusively examined male mice following our previous studies.It is unknown whether the findings are relevant for female mice.

Mice and in vivo treatment
Pacc1 flox/flox , Pacc1 -/-and Pacc1 +/+ mice were generated as previously described (10) We performed LSI surgery in 2-month-old male mice, that were either WT, Pacc1 -/-, Pacc1 wt and Pacc1 flox/flox .Briefly, the mice were anesthetized with ketamine (100 mg/kg, intraperitoneally) and xylazine (10 mg/kg, intraperitoneally).Then, the LSI mouse model was created by resecting the L3-L5 spinous processes and the supraspinous and interspinous ligaments to induce instability of the lumbar spine (10-12 mice per group).Sham operations involved detachment of the posterior paravertebral muscles from L3-L5 only in a separate group of mice (10-12 mice per group) (50).All mice were maintained at the animal facility of The Johns Hopkins University School of Medicine.

Human subjects
Following approval from the Institutional Review Board (IRB; Johns Hopkins Institutional Review Boards), we procured human joint tissue samples.Our institutional IRB granted a waiver of consent for the participants, as the specimens comprised de-identified tissue archived by the pathology department.This approach aligns with the US Food and Drug Administration's regulations on consent waiver (Organization Policy FDA 50.1), permitting the use of such samples for research without individual consent due to the anonymity and preexisting status of the tissue.

Behavioral testing
Behavioral testing was performed for Pacc1 -/-, Pacc1 wt mice with sham or LSI surgery.All behavioral tests were performed by the same blinded investigator in the study group Pressure thresholds were measure (SMALGO algometer, Bioseb) as pressure hyperalgesia (51).The L4-L5 spine was pressed by a 5-mm-diameter sensor tip while the mice were gently restrained.We gradually increased the pressure force at 50 g/s until an audible vocalization was heard.The pressure force was read by BIO-CIS software (Bioseb).A cutoff force was set as 500 g to prevent tissue trauma.The mice were allowed to rest for 15 min between tests, and the mean -14 -value was calculated as the pressure tolerance threshold.
For the Von Frey test, we used Von Frey filament of 0.4 g (Stoelting) to test the frequency of hind paw withdrawal.Mice were placed on a wire metal mesh grid covered with a black plastic cage.Mice were set to acclimatize to the environment for at least 30 min before testing.The mid-plantar surface was stimulated by the filament for 2 seconds.The withdrawal frequency was recorded as the result of mechanical nociceptive threshold of the mice in response to 10 applications.
Spontaneous wheel-running activity was recorded using activity wheels designed for mice (model BIO-ACTIVW-M, Bioseb).The software enabled recording the activities of a mouse in the wheel cage.The mice were acclimatized to the environment for overnight before testing.
And the test will be lasted for 48 hours for each mouse.The parameters of the spontaneous activities will be automatically recorded.

Histochemistry, immunofluorescence, and histomorphometry for histological section
At the time of killing, the L3-L5 lumbar spine and femur samples were collected and fixed in 10% buffered formalin for 24 h.Both human and mouse bone samples were decalcified at 4°C using 0.5M ethylenediamine tetra acetic acid for two months or 3 weeks with constant shaking and then embedded in paraffin or 8% gelatin in the presence of 20% sucrose and 2% -15 -polyvinylpyrrolidone.Four-μm-thick coronal-oriented sections of the L4-L5 lumbar spine were processed for Safranin O and fast green and TRAP (Sigma-Aldrich) staining using a standard protocol.Four-μm-thick sections of human tibia tissue were used for coimmunofluorescence staining of PAC and TRAP.Forty-μm-thick coronal-oriented sections were prepared for sensory nerve-related immunofluorescent staining, and 10-μm-thick coronaloriented sections were used for other immunofluorescent staining using a standard protocol.

Cell isolation and culture
The hind limbs of 8-week-old mice were harvested by carefully removing the attached soft tissue.We collected bone marrow cells by cutting both ends of the tibia and femur and then flushing the marrow with a syringe using α-minimum essential medium (α-MEM) (Sigma-Aldrich).Whole bone marrow cells were collected through centrifugation for 15 min at 1000 rpm and then cultured in α-MEM with 10% fetal bovine serum (FBS) (Sigma-Aldrich) at 37 °C in a 5% CO2-humidified incubator.After 24 h, the non-adherent cells floating in the culture media were collected and cultured in α-MEM with M-CSF (30 ng/ml).After 3 days, the macrophage-lineage cells were collected by digesting the adherent cells with Versene Solution (Thermo Fisher).The bone marrow macrophages (BMM) were reseeded in 6-well plates (5 × 10 5 cells per well), 24-well plates with cover slips (5 × 10 5 cells per well) or 96-well plates with bone slices (1 × 10 6 cells per well), and cultured in α-MEM containing 30 ng/ml M-CSF and 100 ng/ml RANKL (PeproTech).The pH level of cell culture medium was calibrated by blood gas analyzer at Johns Hopkins Medical Laboratory.

RT-PCR
before each measurement and the series resistance then was routinely compensated for at least 80%.
For PAC current recording, cells were held at 0 mV and voltage ramps (500 ms duration) were applied from -100 to +100 mV.Data was analyzed using Clampfit 10.6 and GraphPad Prism 6.

Bone resorption assay
To assess the effect of PAC-induced bone resorption in acidic condition, we performed an osteoclast bone resorption assay using a commercially available bone resorption assay kit (Cosmo Bio).Briefly, BMM isolated from Pacc1 -/-and Pacc1 +/+ mice were seeded on bone slices in 24-well plates and cultured in acidic or neutral osteoclastogenic medium for 7 days.
The resorption pits on the hydroxyapatite surface were imaged under a microscope.

Staining of the osteoclasts
Cells were cultured in osteoclastogenic medium for 1, 3 or 5 days, and fixed for 10 minutes by using 4% paraformaldehyde.A TRAP staining kit (Sigma-Aldrich) was used to detect TRAP + cells according to the manufacturer's instructions.We used fluorescence staining of F-actin with phalloidin to observe action ring formation.Sample images were captured by a fluorescence microscope (Olympus BX51, DP71).

Endplate pH measurement
To conclusively determine the presence of acidic conditions in the spinal endplate area, we employed a Micro combination pH electrode with a needle tip for accurate in vivo pH assessments (9863BN Micro pH Electrode, Thermo Fisher).Both LSI and sham-operated mice membrane and cytoplasm of osteoclasts and its open conformation only occurs under conditions of acidic microenvironment.Knockout of Pacc1 significantly reduced osteoclast fusion and endplate porosity, as well as relieved LBP in a mouse model of spine degeneration, but the genetic deletion did not have any effect on bone development or bone homeostasis in adult mice.Together, our results show that Pacc1-encoded Cl − channel activity is induced during acidosis, leading to abhorrent osteoclast-mediated resorption to generate endplate porosity that leads to LBP.As depletion of PAC activity prevented the development of LBP during spinal degeneration with no negative effect on bone modeling or remodeling, -5 -representing a potential therapeutic target for LBP.

Figures and Legends Figure 1 .
Figures and Legends

Figure 3 .
Figure 3. PAC expression is induced during osteoclast differentiation.(A) A schematic diagram illustrating the preparation of TRAP + osteoclast precursors and their maturation process.(B) Relative mRNA expression levels of Pacc1 in the cells cultured in the medium with or without RANKL at 0, 1, 3 and 5 days, n = 3. (C, D) Protein levels of NFATc1 and PAC at 0, 1, 3 and 5 days were analyzed by western blotting, n = 3. (E) Diagram of potential NFATc1 binding sites on the Pacc1 promoter in osteoclast precursors.(F) ChIP analysis of NFATc1 on the Pacc1 promoter.(G) Diagram of the Pacc1 promoter with an NFATc1 binding site.(B) *p < 0.05 compared with Day 0; # p < 0.05 compared with M-CSF group (without RANKL);

Figure 4 .
Figure 4. Extracellular acidosis evokes the I Cl, H current in the cell membrane of osteoclasts by activating PAC.(A) Relative mRNA expression levels of Pacc1 in the cells isolated from Pacc1 +/+ mice cultured in neutral or acidic medium at 1, 3 and 5 day, n = 3s.(B, C) Protein levels of PAC in the cells isolated from Pacc1 +/+ or Pacc1 -/-mice cultured in neural or acidic medium at 1, 3 and 5 days, n = 3. (D) The representative images of co-immunofluorescence staining of PAC (Green) and TRAP (Red) in human bone section.Scale bar: 200µm, 100µm.Blue: DAPI.(E) PAC currents monitored by voltage-ramp protocol at pH 5.5 in the cells isolated from Pacc1 +/+ or Pacc1 -/-mice cultured in osteoclastic medium for 3 days.(F) PAC-mediated current densities measured at +100 mV in the cells isolated from Pacc1 -/-(red) and Pacc1 +/+ (black) mice