Proton-sensing G protein-coupled receptor mobilizes calcium in human synovial cells

Burgess N. Christensen, Mikhail Kochukov, Terry A. McNearney, Giulio Taglialatela, Karin N. Westlund

This article has a correction. Please see:


Lowered extracellular pH in a variety of tissues is associated with increased tissue destruction and initiation of inflammatory processes. Although the acid-sensing receptors described previously are ion channels, we describe a G protein-coupled proton-sensitive receptor that stimulates Ca2+ release from intracellular stores in a tumor-derived synoviocyte cell line (SW982) and in primary cultures of human synovial cells from patients with inflammatory arthropathies. We established a link between proton-dependent receptor activation and intracellular Ca2+ mobilization by demonstrating 1) dependence on the integrity of the intracellular Ca2+ store, 2) independence from extracellular Ca2+, and 3) proton-induced production of inositol phosphate and 4) by abolishing the effect with GTPase inhibitors. We propose that this G protein-coupled acid-sensing receptor linked to intracellular Ca2+ mobilization in synoviocytes can contribute to downstream inflammatory and cellular proliferative processes in synovial fibroblasts. The acid-sensing receptor has distinct characteristics as a metabotropic G protein-coupled receptor on human synoviocytes in this emerging new class of receptors.

  • calcium imaging
  • synoviocytes
  • arthritis
  • acid-sensing receptor

inflammatory processes are often accompanied by a decrease in extracellular pH. Protons can modulate cell signaling through acid-sensing ion channels, screening surface charge, or by protonation of residues on integral membrane proteins. Tissue acidosis is an important component of ischemic and inflammatory sequelae. Inflammatory arthropathies such as rheumatoid arthritis (RA) and septic arthritis are often characterized by acidic synovial fluids (SFs) and increased hydrogen ion levels, secondary to increased cellular proliferation, metabolic activation, and subsequent lactic acid production (6, 10, 20, 27, 31, 33). The effects of lowered extracellular pH on the tissue have been associated with increased destruction and worsening of the inflammatory parameters (8, 28, 29). In RA, synovial fibroblasts mediate the inflammatory process by releasing a number of factors, including proteases that destroy the surrounding supporting matrix. One of these, cathepsin, requires a lowered pH for activity in the degradation of collagen (2830). Inflammation or damage to synovial tissue releases protons necessary to activate cathepsin; thus protons represent a potential ligand for the activation of synoviocytes (21).

The potential role of integral membrane receptors in synovial cells that regulate cell function by Ca2+-mediated signaling has not been investigated. There are two distinct classes of receptors that may lead to a rise in intracellular Ca2+ concentration ([Ca2+]i). The ionotropic receptors that are activated by ligands or changes in membrane voltage allow for Ca2+ entry through Ca2+-permeable membrane channels. Metabotropic G protein-coupled receptors (GPCRs) allow, in many instances, for Ca2+ mobilization from intracellular stores after activation by an extracellular ligand (2, 34). A number of proton-sensitive GPCRs were characterized recently. Extracellular lipid receptors OGR1 (16) and G2A (19) respond to extracellular acidification by accumulation of inositol phosphates (IPs), whereas two related receptors, GPR4 and TDAG8, demonstrate pH-dependent increases in cAMP (16, 32). Herein we report a G protein-coupled proton-sensitive receptor that stimulates Ca2+ release from intracellular stores in a tumor-derived synoviocyte cell line (SW982) and in human synovial cells cultured from patients with inflammatory arthropathies.


Cell lines and cell culture.

The SW982 cells derived from a human synovial sarcoma are fibroblast-like synovial cells (7) obtained through the American Type Culture Collection (Bethesda, MD). The cells were maintained in Leibovitz's L-15 medium, 10% heated fetal bovine serum (FBS), 2 mM l-glutamine, 1,000 U/ml penicillin G, and 1,000 pg/ml streptomycin. Cells were incubated in a humidified cell incubator (37°C, ambient CO2) and used in passages 4–21. Primary lines of surface-adherent synoviocytes were derived from the SFs from therapeutic arthrocenteses of two patients (one patient had RA, one had psoriatic arthritis. Discarded fluid samples were collected with Institutional Review Board-approved informed consent in accordance with American College of Rheumatology criteria (13), the principles of the Declaration of Helsinki, as well as Title 45, U.S. Code of Federal Regulations, part 46, Protection of Human Subjects, revised November 13, 2001, effective December 13, 2001. Cells were maintained in DMEM, 10% heated FBS, 2 mM l-glutamine, and penicillin-streptomycin. Primary cultures were incubated in a humidified cell incubator (37°C with 5% CO2 atmosphere) for 2 wk before testing. Cells were split with 0.25% trypsin-0.02% EDTA disruption and plated on 25-mm circular glass coverslips at a density of 200,000 cells/mm3.

Intracellular calcium and pH imaging.

On the day of the experiment, the cells were loaded with the calcium-sensitive fluorescent dye fura-2 or the pH-sensitive fluorescent dye BCECF by incubating at 5 μM for 1 h (fura-2 AM) or 15 min (BCECF AM), after which they were washed in a mammalian physiological solution containing (in mM) 150 NaCl, 5.5 KCl, 1.2 MgCl2, 2 CaCl2, 10 glucose, and 10 HEPES, adjusted to pH 7.4 at room temperature with NaOH. For imaging, the coverslips were mounted in a Leiden chamber and placed in a Perspex holder on a Nikon 200E microscope. Cells were imaged with a ×40 oil immersion lens [1.3 numerical aperture (NA)], using a computer-controlled illumination system (Sutter Instruments, Novato, CA) equipped with a digital monochrome-cooled charge-coupled device Roper Coolsnap HQ camera (Roper Scientific, Tucson, AZ). Fluorescent emission at 510 nm for fura-2 and 535 nm for BCECF from the regions of interest (corresponding to a single cell) was acquired online with MetaFluor software (Universal Imaging, Downington, PA). Signal was obtained in dual excitation mode: 340 and 380 nm for fura-2 and 495 and 440 nm for BCECF. The 340- to 380-nm and 495- to 440-nm ratios reflect [Ca2+]i and intracellular pH (pHi), respectively. For electrophysiology, the chamber was placed on the stage of a Nikon Diaphot and cells were viewed with a ×100 oil immersion lens (1.3 NA). Illumination was supplied by a 75-W xenon lamp led via a beam splitter to two shutters (Uniblitz, Vincent Associates, Rochester, NY), each one containing a 340-nm or 380-nm filter. A bifurcated fused silica light guide then led the light source to the microscope. A computer with proprietary software controlled illumination by alternately opening the shutters and acquiring the emission at 510 nm after 340- and 380-nm excitation.

Electrophysiology and single-cell calcium measurement.

The chamber with coverslip was placed on the stage of a Nikon Diaphot, and the cells were viewed with a ×100 oil immersion lens (1.3 NA). Illumination was supplied by a 75-W xenon lamp led via a beam splitter to two shutters (Uniblitz), each one containing a 340-nm or 380-nm filter. A bifurcated fused silica light guide then led the light source to the microscope. A computer with proprietary software controlled illumination by alternately opening the shutters and acquiring the emission at 510 nm after 340- and 380-nm excitation.

Both ruptured-patch and perforated-patch techniques were used to record from dye-loaded cells. The perforated patch prevented dilution of dye during the recording period. The ruptured patch was used to apply guanosine 5′-O-(3-thiotriphosphate) (GTPγS) and guanosine 5′-O-(2-thiodiphosphate) (GDPβS) to the interior of the cell. Patch pipettes were made from 1.65-mm outer diameter borosilicate glass pulled on a Sutter Instruments puller (Sutter Instruments, Novato, CA). The pipette internal solution contained (in mM) 25 KCl, 65 K2SO4, 2 MgCl2, 10 HEPES, and 1 EGTA, with 100 μM fura-2, pH 7.2. In ruptured-patch mode 2 mM ATP and 0.5 mM GTP were included in the electrode solution just before the experiment. Amphotericin B stock solution (60 μg/μl) was made in DMSO and stored in 4-μl aliquots in Eppendorf tubes at −4°C. Working amphotericin B solution was made by diluting an aliquot in 500 μl of internal solution while vortexing. The solution was then sonicated for 1 min, and the solution was kept on ice. Pipette tips were filled by holding the tip in a drop of internal solution for 5 s. The pipette was backfilled with amphotericin B working solution, and the bubbles were removed mechanically. Gigaohm seals were made on cells by slight suction applied to the barrel of the pipette through the pipette holder (Axon Instruments, Union City, CA). Cell access was monitored by noting the change in the capacitance transient while applying a hyperpolarizing test pulse with an Axon Instruments 200A voltage-clamp amplifier connected to a Digidata 1200 data acquisition system and a computer running pCLAMP version 6.0 (Axon Instruments). Cells were voltage clamped at −65 mV, and membrane potential was stepped in 10-mV increments from the holding potential.

Solutions and recording conditions.

All measurements were performed at room temperature (21–23°C), most under ambient CO2. Mammalian physiological solutions at pH 7.4, 6.8, 6.4, or 5.5 were prepared as described above and adjusted to 330 mosM with sucrose. Some experiments were buffered with 25 mM NaHCO3 bubbled with 95% air-5% CO2 (pH 7.4). Solution changes were made during imaging by rapidly (5–10 s) adding a volume of 5–10 ml of new solution while removing the overflow by suction. There were no mechanical effects of rapid exchange, as repeating this procedure with solution of the same pH as in the chamber produced no change in [Ca2+]i. The recording chamber volume was 1–2 ml. Application of solution at pH 6.4 or 5.5 onto individual cell was achieved by pressure ejection from a pipette placed within 25 μm of the cell or by gravity flow from a large-bore pipette placed 300–400 μM from the cell. Experiments were repeated a minimum of four times each.

The same perfusion system configuration was used to apply ATP, thapsigargin, d-erythro-sphingosylphosphorylcholine (SPC), lysophosphatidylcholine (LPC), or CuCl2. 8,8′-[Carbonylbis(imino-3,1-phenylene)]bis-(1,3,5-naphthalenetrisulfonic acid) sodium salt (NF023) and pertussis toxin were added to the serum-free incubation medium at least 6 h before the experiment and were continuously present in the bath solution during measurements.

Generation of IPs.

A standard ion exchange chromatography protocol for isolation of IPs was used with the following modifications (4). The SW982 cells were plated at subconfluent density into six-well dishes in 2 ml of L-15 medium, 10% heated FBS, l-glutamine, and penicillin-streptomycin. myo-[3H]inositol (1 μCi/ml; NET114A, Perkin Elmer, Torrance, CA) was added to each well, and the cells were incubated for 48 h. At the time of cell harvest, the cells were washed three times with warmed PBS. After the third wash, 0.5 ml of the following pH buffers (36°C) was added to three wells: pH 7.4, 7.0, 6.8, 6.4, 6.0, and 5.5. These buffers also contained the following final concentrations (in mM): 118 NaCl, 4.7 KCl, 1.2 MgCl2, 2.5 CaCl2, 10 glucose, and 20 LiCl. LiCl was added to the buffers to block inositol monophosphatase activity. The solutions were buffered with either 10 mM HEPES (for pH 7.4, 7.0, and 6.8 buffers) or 10 mM MES (for pH 6.4, 6.0, and 5.5 buffers). Each condition was set up in triplicate, and the experiment was performed four times. One cell condition using PBS without Ca2+ or Mg2+ at pH 7.4 was also included as a negative control. One condition, in which the cells were incubated with 100 μM carbachol in pH 7.4 physiological solution, was also included as a positive control for these cells. Cells were incubated in the pH buffers for 5 min. The buffers were then aspirated, and the reaction was terminated by the addition of 0.5 ml of methanol-HCl (100:1 vol/vol) to each well. After extraction in chloroform (0.5 ml) and 250 ml of 10 mM EDTA, pH 7.4, the lysate was vortexed and centrifuged briefly to separate the phases. The aqueous phase was applied to washed Dowex columns (see below) for isolation of the IP. Standard 5-ml separation columns (Bio-Rad, Richmond, CA) were used by adding 0.5 ml of anion exchange resin (AG1-X8 resin, formate form; no. 140-1444, Bio-Rad) equilibrated with 10 mM myo-inositol and 0.1 M formic acid.

After the cell lysates were applied to the columns, the beads were washed three times with 5 ml of ice-cold 10 mM inositol-0.1 M formic acid. The IP was eluted with 1.0 ml of ice-cold 0.8 M/0.1 M formic acid. The eluates from each column were collected, diluted in 10 ml of aqueous scintillation cocktail (Aquasol Scintillation Fluid, Beckman, Fullerton, CA), and counted on a scintillation counter for 1 min. The eluate yields for each column were corrected by the activity (counts/min) of the third wash. The pH of the buffers and elution buffers had no independent effect on scintillation activity.


Results are reported as means ± SD. Statistical significance (P < 0.05) was evaluated with Student's t-test (SigmaStat 2.0 software, SPSS, Chicago, IL).


Leibovitz's L-15 cell culture medium and antibiotic solution were obtained from GIBCO (Grand Island, NY), and fetal calf serum was obtained from Gemini Bio-Products (Woodland, CA). ATP sodium salt, GTP sodium salt, GDPβS (lithium salt), GTPγS (lithium salt), NF023, pertussis toxin, thapsigargin, and fura-2 were obtained from Calbiochem (La Jolla, CA). Fura-2 AM and BCECF AM were obtained from Molecular Probes (Eugene, OR). SPC and 16:0 LPC were obtained from Avanti Polar Lipids (Alabaster, AL). myo-Inositol and carbachol were obtained from Sigma (St. Louis, MO).


Proton-induced Ca2+ mobilization.

Figure 1 illustrates the change in [Ca2+]i in clonal SW982 synoviocytes during repetitive applications of physiological solutions at different pH values. The change in [Ca2+]i is the average response measured from 8 of 22 cells in Ca2+-free physiological solution at pH 6.8, demonstrating that the increase in [Ca2+]i is not due to Ca2+ influx (Fig. 1A). When the physiological solution was changed to pH 6.4 in the presence of Ca2+, these eight cells responded again as shown, and an additional four cells previously unresponsive to pH 6.8 showed an increase in [Ca2+]i at pH 6.4. This suggests a dose-dependent response to proton concentration. Summarizing the data, about one-half of the cells (17 of 35) responded to pH 6.8, and 44 of 61 showed the response to pH 6.4 acidification. In Figs. 1, 2, and 4 showing changes in [Ca2+]i measured simultaneously from several cells, the large SD is due to the nonsynchronous rise in [Ca2+]i from the different imaged cells during the change in extracellular pH.

Fig. 1.

A: average and SD change of 340 nm-to-380 nm fluorescence ratio from imaged synoviocytes. Bars below trace indicate solution changes. Only the pH 6.8 solution was Ca2+ free. B: change in 340 nm-to-380 nm ratio of a single synoviocyte to application of pH 5.5 physiological solution by pressure ejection from a nearby pipette as indicated by the bars. With sufficient time for store refilling, cells responded to multiple applications. C: change in 340 nm-to-380 nm ratio of a single synoviocyte to application of pH 5.5 physiological solution by pressure ejection from a nearby pipette as indicated by the bar. The duration of application was 70 s, during which time the intracellular Ca2+ concentration ([Ca2+]i) increased slowly, as can be seen by comparison with the baseline indicated by the dashed line. On termination of the application, there was a spontaneous and rapid increase in [Ca2+]i. D: average and SD change of 340 nm-to-380 nm ratio from imaged synoviocytes. Bars below the trace indicate solution changes, including ATP and thapsigargin (Tpsg). Only the pH 6.8 solution was Ca2+ free.

Fig. 2.

Changes in the 340 nm-to-380 nm ratio recorded from single voltage-clamped cells loaded with fura-2 through the patch pipette. Bars above traces indicate application of pH 6.4 solution through a large-bore pipette. A: as a control, GTP (0.5 mM) was dialyzed into the cell, and LiCl (6 mM) was present in the intracellular pipette during activation with the pH 6.4 solution. B: under experimental conditions similar to those used in A, GTP and LiCl in the intracellular solution were replaced with 2 mM guanosine 5′-O-(2-thiodiphosphate) (GDPβS). Under these conditions, there was no response to applications of pH 6.4 solution.

In Fig. 1B, a pipette containing physiological solution at pH 5.5 was placed within 25 μm of a single cell, and solution was applied by pressure ejection (n = 25 separate single-cell experiments). The figure shows repetitive increases in [Ca2+]i in response to repeated applications of buffer at pH 5.5 from a single representative SW982 synovial cell under these conditions.

Ca2+ mobilization occurred after variable time delay, usually 30–60 s, and this was proven not to be due to perfusion or equilibrium artifacts. Once [Ca2+]i begins to increase, it develops in an all-or-none fashion. Figure 1C shows an example under conditions similar to those illustrated in Fig. 1B, where in this specific case Ca2+ mobilization was slow to develop. When the low-pH physiological solution was applied, there was a slow, small increase in [Ca2+]i. [Ca2+]i reaches what appears to be a threshold at about the time that the low-pH solution application is stopped, at which point the Ca2+ mobilization develops rapidly in the absence of continually applied protons. When extracellular solutions remained acidic for a long period of time, some cells responded only with one initial spike (on solution switching), whereas others demonstrated repeated Ca2+ spikes. This type of response is inconsistent with Ca2+ flux through a membrane pore and provides indirect evidence that proton-stimulated [Ca2+]i works with a receptor in a dose-dependent manner.

Role of extracellular Ca2+ and intracellular Ca2+ stores in acid-induced [Ca2+]i release.

Figure 1D shows the averaged response in a separate experiment from 2 of the 22 cells in which [Ca2+]i increased during treatment with Ca2+-free physiological solution at pH 6.8. After washout with a Ca2+-containing solution at pH 7.4, a decrease in pH to 6.4 produced a response in the 2 cells that reacted at pH 6.8 plus an additional 11 cells that had not responded at pH 6.8. The addition of ATP to the solution at pH 6.4 also produced a transient increase in [Ca2+]i in the 2 cells illustrated in Fig. 1D and an additional 16 cells. The four cells that did not respond to ATP also did not respond to pH 6.4. In every experiment, more cells responded to pH 6.4 than to pH 6.8. Addition of thapsigargin produced a small increase in [Ca2+]i in all cells. Finally, cells that were treated with thapsigargin for 1 h before imaging did not respond to physiological solution at pH 6.8 or pH 6.4 or to ATP, confirming that the proton-induced rise in [Ca2+]i was due to Ca2+ mobilization from intracellular stores.

We tested for the expression of ryanodine receptor by caffeine application to clonal SW982 synovial cells (not shown). Caffeine does not mobilize Ca2+ from intracellular stores, suggesting that these cells do not have ryanodine-sensitive Ca2+ stores.

Whole cell patch-clamp recordings did not detect any change in membrane conductance during the increase in [Ca2+]i when physiological solution at pH 6.4 or 5.5 was applied from a pipette placed close to the cell, further confirming that the increase in [Ca2+]i was not due to Ca2+ influx.

Nonhydrolyzable guanine nucleotide analogs abolish proton-activated calcium response.

In a few separate experiments, GTPγS and GDPβS were dialyzed into the cell via the patch pipette. A control series of recordings showed that about one-half (6 of 14) of the cells dialyzed with pipette solution containing ATP and GTP repetitively responded with [Ca2+]i increase to applying acid (pH 6.4 or 5.5) solution as long as the gigaohm seal was maintained. Because GTPγS and GDPβS are available as Li+ salts, control recordings were made to test the possible effect of this ion on the pH-sensing receptor. Addition of 6 mM LiCl in the pipette had no influence on acid-induced calcium response, which was observed in four of seven cells (Fig. 2A). After rupture of the patch membrane, GTPγS elicited a rapid increase in [Ca2+]i, after which application of buffer at pH 5.5 had minimal to no effect on [Ca2+]i (not shown). Cells dialyzed with GDPβS also did not show any Ca2+ spikes in response to acid (n = 12; Fig. 2B). Inhibition of proton-induced calcium response by the GTPase inhibitor GDPβS and the nonhydrolyzable GTP analog GTPγS strongly suggests the involvement of G protein in transducing the extracellular acidification signal to [Ca2+]i stores.

Effect of pHi on Ca2+ mobilization.

To test whether the increase in [Ca2+]i was due to a change in pHi, we repeated the measurement of [Ca2+]i during application of 20 mM NH4Cl in buffered saline at pH 7.4 or a bicarbonate-buffered solution (pH 7.4) bubbled with a mixture of 95% air-5% CO2. To demonstrate the changes in pHi, cells were loaded with the pH-sensitive fluorescent dye BCECF. Buffered saline at pH 6.4 did not change pHi over the time course of the Ca2+ mobilization experiments. A brief application of NH4Cl (20 mM) initially produced an alkalinization followed by a small acidification of the cell cytoplasm on washout with buffered solution at pH 7.2 (Fig. 3A). There was usually a small, slow increase in [Ca2+]i during the alkalinization phase but no change in [Ca2+]i during cytoplasmic acidification (Fig. 3B). However, cells responded with an increase in [Ca2+]i to a physiological solution at pH 6.4. Application of physiological saline buffered with HCO3 and bubbled with 5% CO2 showed a rapid acidification of the cell cytoplasm followed by a slow recovery after washout with solution at pH 7.4 (Fig. 3C). When [Ca2+]i was measured with this cell interior acidification protocol, some cells showed an increase in [Ca2+]i. When physiological solution at pH 6.4 was applied to the same preparation, many more cells responded (Fig. 3D). Although solution changes were applied rapidly and in large volumes compared with the volume of the cell chamber, it is possible that the CO2 diffused in the unstirred layer over the cell membrane faster than HCO3, thus transiently acidifying the solution near the cell surface. We tested this hypothesis by pretreating with a buffered bicarbonate solution that had not been bubbled with CO2, followed by the same solution with CO2. When synovial cells were pretreated with bicarbonate buffer, 1 of 30 cells tested measured increased [Ca2+]i when bubbled CO2 was added, whereas 23 of 30 cells responded when the extracellular pH was lowered to 6.4. With no bicarbonate buffer preincubation, 13 of 56 cells responded to bicarbonate buffer bubbled with CO2, compared with 46 of the 56 cells that responded when pH was lowered to 6.4. Thus a decrease in pHi did not release [Ca2+]i in synovial cells.

Fig. 3.

A: average and SD 495 nm-to-440 nm ratio from 16 of 16 cells loaded with BCECF to measure intracellular pH (pHi). There is little change in pHi for the first 100 s, during the application of the pH 6.4 physiological saline. The solution is then returned to pH 7.2, and an NH4Cl pulse is applied. There is a rapid alkalinization to the cell interior, followed by acidification on washout of the NH4Cl. B: average and SD 340 nm-to-380 nm ratio measured in 12 of 12 cells. During the NH4Cl pulse (alkalinization) there is a slow rise in [Ca2+]i but no change on washout of the NH4Cl. When the pH is decreased to 6.4, a rapid Ca2+ response is seen. C: average and SD 495 nm-to-440 nm ratio measured in 15 of 15 cells loaded with BCECF to measure pHi. Acidification of the cell interior during application of bicarbonate-buffered saline occurs with bubbled 95% air-5% CO2. The NH4Cl pulse shows the typical alkalinization and acidification on washout. D: average 340 nm-to-380 nm ratio measured in 6 of 13 cells that responded to the CO2 pulse followed by the increase in [Ca2+]i during a decrease in bath pH to 6.4. All 13 cells responded to the decrease in pH to 6.4.

Pharmacology of H+-induced Ca2+ release.

Two lipid compounds reported as possible ligands for known types of acid-sensing receptors (36) were tested for their effects on synovial cells. Application of 5 or 10 μM SPC produced a large Ca2+ spike with a short 10- to 20-s delay in all tested cells (n = 12). With a dose of 1 μM, the effect was less prominent or even absent in some cells. Applied on the declining slope of Ca2+ spike induced by the previous application of acid solution, SPC produced a new Ca2+ wave; acidification produced a new [Ca2+]i increase on the declining phase of the previous SPC-induced response (Fig. 4A). Furthermore, SPC was able to produce prominent [Ca2+]i spikes when cells were maintained at pH 6.4 or 5.5 and did not show any changes in Ca2+ in the absence of the lipid agonist (Fig. 4B). These results suggest that extracellular protons and SPC activate different subsets of receptors in synovial cells or have different mechanisms of activation of the receptor. Another lipid compound, LPC (10 μM), reported as the agonist for GPR4 (38) and an antagonist for G2A receptors (19), typically did not produce any changes in [Ca2+]i (n = 15 cells tested). Likewise, it had no effect on the SPC- or proton-induced Ca2+ responses.

Fig. 4.

A–C: average and SD change of 340 nm-to-380 nm ratio from imaged synoviocytes comparing responses to a lipid compound, d-erythro-sphingosylphosphorylcholine (SPC), and acid pulses. Changes in 340 nm-to-380 nm ratio were recorded from the cells, loaded with fura-2, under the same conditions as shown in Fig. 1. Each trace represents the signal ratio acquired from 3 or 4 adjacent cells. Bars above indicate application of pH 6.4 solution, 10 μM SPC, or 100 μM CuCl2. D: concentration-dependent inhibition of acid-induced synoviocyte calcium response by Cu2+. Fura-2 ratio was normalized to average changes in control group (without Cu2+, ΔRnorm). Each data point represents average ± SD for at least 4 separate experiments.

As was shown for OGR1 and TDAG8 receptors, CuCl2 effectively inhibits the proton-induced response, apparently by preventing the protonation of acid-sensing histidine residues. In the synovial cells, addition of 100 μM CuCl2 to the bath solution 1 min before acidification completely abolished any calcium response to pH 6.4 or 5.5, as well as changes elicited by SPC (Fig. 4, C and D). This inhibitory effect was reversible, i.e., the ability of cells to respond to acid pH on washout was restored.

Inhibitors of the Gi type of GTP-binding α-subunit, pertussis toxin (100 ng/ml) and NF023 (30 μM), were used to determine apparent association of the acid-sensing receptor to a specific G protein. However, treatment of synoviocytes with either compound did not have any effect on proton-induced [Ca2+]i changes. This experiment was repeated with increasing pretreatment time from 6 up to 18 h.

Calcium response to extracellular acidification in primary culture of human synoviocytes.

The effect of a decrease in extracellular pH on [Ca2+]i was tested in two separate studies with primary synoviocyte cultures obtained from two clinic patients, one with RA and one with psoriatic arthritis. Figure 5 illustrates the change in [Ca2+]i in adherent synoviocytes from one clinic patient with RA at pH 6.4 followed by ATP. The synoviocytes from primary cultures also demonstrated Ca2+ mobilization with lower-pH solutions. Similar results were obtained from primary cultures of synoviocytes from the patient with psoriatic arthritis.

Fig. 5.

The 340 nm-to-380 nm ratio from 1 of 4 primary synovial fibroblast cells derived from a patient with rheumatoid arthritis (ATP 100 μM). Bars below the trace indicate solution changes. This responsive cell represents many such imaged synoviocytes in several fields tested. Synovial cell cultures were derived from 1 patient with rheumatoid arthritis and 1 patient with psoriatic arthritis.

Effect of extracellular acidification on IP production.

To confirm a G protein-coupled mechanism for proton-induced Ca2+ mobilization, IP production was measured in cultured SW982 synoviocytes. The cells were incubated in buffers and conditions similar to those of the Ca2+ mobilization studies. Incubation in buffers at different pH values resulted in a significant increase in the amount of IP extracted from the solubilized cells (Fig. 6). The generated IP elution patterns were in agreement with the Ca2+ mobilization studies in that the maximum generation was seen at pH 6.4. Significant IP generation was also seen at pH 6.8, 6.0, and 5.5. A positive control of cells incubated with carbachol in physiological solution at pH 7.4 was used as a reference for IP generation in cultured synoviocytes. The IP generation in cultured synoviocytes at pH 6.4 paralleled the stimulation of IP seen with 100 μM carbachol. The value of IP generation at pH 7.4 was similar to that seen with the negative control using buffer without LiCl.

Fig. 6.

Measurement of inositol phosphate (IP) by ion exchange chromatography from cultured synoviocytes (SW982) incubated with physiological buffers at various pH values. *P < 0.05, statistical significance.

In summary, we measured membrane current and increases in [Ca2+]i from SW982 human synovial cells and from primary synoviocyte cultures obtained from patients with inflammatory arthropathies while applying buffered physiological solution at pH 7.4, 6.8, 6.4, or 5.5. Protons per se produced no change in membrane current. However, Ca2+ mobilization seen at decreased extracellular pH was abolished by intracellular Ca2+ store depletion with thapsigargin, but not by removal of extracellular Ca2+. pHi did not change during the time course of Ca2+ mobilization, and, conversely, an induced decrease in pHi did not mobilize Ca2+. Physiological solutions at decreased extracellular pH stimulated IP production in cultured synoviocytes. Our results describe a proton-activated GPCR linked to Ca2+ mobilization initiated by decreased extracellular pH in synoviocytes that may be involved in important signaling pathways during the development of arthritis and other systemic, ischemic, and inflammatory processes.


These results demonstrate the expression of a proton-sensing GPCR in SW982 cells obtained from a human synovial tumor and in primary synovial cells derived from patients with rheumatoid and psoriatic arthritis. We established a link between receptor activation and intracellular Ca2+ mobilization by demonstrating 1) dependence on the integrity of the intracellular Ca2+ store, 2) independence from extracellular Ca2+, and 3) proton-induced production of IP and 4) by abolishing the effect with GTPase inhibitors. Although future studies are necessary to determine what type of acid-sensing GPCRs are possessed by synovial cells, the GTPase inhibition clearly demonstrates the involvement of GPCRs in proton-mediated activation of this receptor. Four structurally similar acid-sensing GPCRs have already been described. Two of them, OGR1 and G2A, are proven to mediate phosphatidylinositol hydrolysis and intracellular Ca2+ increase (16, 19). These receptor-mediated effects can be variable, as well as ligand- and cell- specific. For example, Zhu et al. (38) demonstrated a [Ca2+]i increase elicited by SPC and LPC in GPR4-transfected cells. The data for LPC and SPC acting as agonists for GPRC pH-sensing receptors are very controversial. Ludwig et al. (16) reported that LPC and SPC do not appear to be ligands for OGR1 and GPR4 receptors. Although Murakami et al. (19) characterized LPC as an inhibitor of G2A receptor, this effect was not reproduced by Radu et al. (24). In our experiments, the effect of extracellular acidification on cytosolic Ca2+ in synoviocytes was neither reproduced nor inhibited by LPC. SPC was shown to be a very potent activator of Ca2+ release in synoviocytes, but this effect is not pH dependent and is apparently mediated by a different type of receptor.

The pH-sensing receptor observed in synoviocytes shows apparent similarity in its pH sensing ability to other known proton-sensitive receptors because it can be efficiently inhibited by Cu2+. The inhibitory effect of CuCl2 was demonstrated for OGR1 and GPR4 receptors by Ludwig et al. (16) and for TDAG8 receptor by Wang et al. (32). It has been proposed that Cu2+ stabilizes unprotonated histidine pairs, preventing proton-induced conformational changes in the receptor. This proton-sensing mechanism appears to be similar in all known acid-activated GPCRs, and our data suggest that the synoviocyte acid-sensing receptor also belongs to this group of receptors.

It is noteworthy that acid-activated Ca2+ responses in the SW982 synoviocytes, as well as IP production in these cells, were maximal at pH 6.4, whereas previous studies demonstrated maximum activation of receptors at pH 6.8 (16, 19, 32). This suggests that synoviocytes possess a distinct type of pH-sensing GPCRs compared with other known proton-activated receptors.

An important question is whether these acid-evoked responses are important in fibroblast-derived cells under normal physiological conditions or, more particularly, in ischemic and inflammatory conditions where destructive and reperfusion processes regularly occur, for example, in synovial and muscle tissues. The receptor on synoviocytes is sensitive to increased proton concentrations well below pH 7.0 and expands the pH range of proton-sensitive receptors. The increased proton sensitivity was noted in synoviocytes derived from two patients with inflammatory arthropathies. We propose that the G protein-coupled acid-sensing receptor is linked to intracellular Ca2+ mobilization in synoviocytes and contributes to downstream inflammatory and cellular proliferative processes in synovial fibroblasts.

There is good evidence that inflammatory disease leads to extracellular proton accumulation and subsequent tissue damage (17). A consequence of inflammatory arthropathies, such as RA, is the infiltration of synovial tissues by T cells, macrophages, and plasma cells (6, 10, 20, 27). Decreased SF pH correlates with leukocytosis (8, 33) and radiological joint damage (8). Increased metabolic activation seen in ischemic and inflammatory events can also cause increased acid production by synovial fibroblasts (21). Protons can also stimulate contractions in smooth muscle in bronchial explants (1). Acidic conditions reported in ischemic muscle and areas of infarction may reflect that there is an even greater potential for proton availability at the cellular level. For example, the release of NO analogs and cytokines with decreases in pH has been described in interactions of fibroblasts and myocytes derived from heart (3, 34). In articular joints, extracellular protons, local production of cytokines, and increases of neurotransmitters in SF derived from the nerve fibers may work in concert to contribute to the inflammatory response. Innervation of synovial tissue (5, 9, 12, 26) and the direct effects of the joint's neurotransmitter content have been reported (11, 12, 14, 15, 18, 23, 25, 37). Although acid-sensing ionic channels on nociceptors may contribute to sensory transduction (22) and inflammatory processes, our data are consistent with a role for activation of metabotropic acid-sensing receptors on synovial fibroblasts in peripheral inflammatory, ischemic, and cellular proliferative processes.

The characteristics of the receptor described here include an intracellular Ca2+ activation optimum at pH 6.4, independent from extracellular Ca2+, G protein-coupled linkage, dependence on the integrity of intracellular Ca2+ stores, and insensitivity to lipid agonists. Thus this acid-sensing receptor with characteristics of a metabotropic GPCR on human synoviocytes is distinctive in this emerging new class of receptors.


The authors thank Wen Ru Zhang for excellent technical assistance, Dr. Leoncio Veraga for excellent imaging laboratory support, and Pat Gazzoli for excellent secretarial assistance.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract