Electrophysiological characterization of ion channels in osteoclasts isolated from human deciduous teeth

Bone Vol. 27, No. 1 July 2000:5–11

RAPID COMMUNICATION

Electrophysiological Characterization of Ion Channels in Osteoclasts Isolated From Human Deciduous Teeth A. F. WEIDEMA, S. J. DIXON, and S. M. SIMS Department of Physiology and Division of Oral Biology, Faculty of Medicine & Dentistry, The University of Western Ontario, London, ON, Canada

originate in the bone marrow.29 Resorption occurs in a compartment referred to as the resorption lacuna, where transport of H⫹ by an electrogenic H⫹-ATPase leads to dissolution of the mineral phase of the matrix, and secreted hydrolytic enzymes digest the organic phase.29 Because H⫹ transport is an electrogenic process, ion channels are required to provide pathways to dissipate charge that might otherwise accumulate, inhibiting resorption. A number of channel types have been demonstrated in osteoclasts, including several K⫹ channels, Cl⫺ channels and H⫹ channels.3,15,17,25,28,35,43 In addition, osteoclasts have been shown to express ligand-gated nonselective cation channels and Ca2⫹-dependent K⫹ channels activated by extracellular nucleotides.24,38 Whereas some types of ion channels are found in osteoclasts of all species studied, others are present only in certain species.39 For example, the large-conductance Ca2⫹dependent K⫹ channel has only been described to date in avian osteoclasts.40 To our knowledge, this is the first study describing ion channels in human osteoclasts. Several model systems have been used for studying human osteoclasts in vitro. These include giant cells obtained from osteoclastomas, which exhibit many characteristics of osteoclasts freshly isolated from long bones.9 However, some differences have been noted, such as the failure of giant cells to respond to nucleotides5 in a manner found for authentic osteoclasts.42 A second approach involves the generation of multinucleated, osteoclast-like cells during culture of mononuclear precursor cells21,26 or cells harvested from giant cell tumors.13 Such preparations are advantageous for examining factors regulating the generation of bone-resorbing multinucleated cells, but the cells may not exhibit all the characteristics of osteoclasts.13 In some cases, cells freshly isolated from surgical specimens, bone biopsies, or fetal long bones have been used to study human osteoclasts8,19,23; however, availability is often restricted. The osteoclasts that resorb the roots of deciduous teeth are often referred to as odontoclasts.7,30 The properties of these cells include: multinucleation; positive staining for tartrateresistant acid phosphatase (TRAP); expression of calcitonin receptors and ␣v␤3 integrins, the formation of a specialized ruffled border, and the ability to resorb mineralized substrates.7,30,33 We describe a novel method for isolation of viable authentic human osteoclasts from deciduous teeth, which enabled us to examine the ion channels in the cells using patch-clamp methods. We identified three types of K⫹ current and a Cl⫺ current, providing the first description of ion channels in human osteoclasts.

Ion channels contribute to several important processes in osteoclasts, including proton transport and volume regulation. Although ion channels have been described in osteoclasts from several species, little is known about their properties in human osteoclasts. We devised a method for isolation of authentic human osteoclasts from deciduous teeth undergoing root resorption, and characterized currents in these cells using patch-clamp techniques. Three types of Kⴙ current were identified. Hyperpolarization elicited an inwardly rectifying Kⴙ current in most osteoclasts, which was inhibited by Ba2ⴙ in a voltage- and time-dependent manner. Depolarization elicited an outwardly rectifying and tetraethylammonium-sensitive current, consistent with a largeconductance Ca2ⴙ-dependent Kⴙ channel. In addition to these basal currents, extracellular adenosine 5ⴕ-triphosphate (ATP) elicited a linear current that was identified as a Ca2ⴙ-dependent Kⴙ current, based on its reversal potential close to that predicted for Kⴙ, its blockade by quinine, and its activation by Ca2ⴙ ionophore. Last, an outwardly rectifying current was observed to activate spontaneously or in response to ATP, with properties of a swelling-activated Clⴚ current. This current reversed direction close to the Clⴚ equilibrium potential and was blocked by the anion channel blocker, niflumic acid, identifying it as a Clⴚ current. In summary, we have developed a novel method for isolation of authentic human osteoclasts and have characterized Kⴙ and Clⴚ currents. Clⴚ current mediates charge compensation during electrogenic Hⴙ transport, so activation of Clⴚ current may contribute to the stimulatory effects of extracellular ATP on bone resorption. (Bone 27:5–11; 2000) © 2000 by Elsevier Science Inc. All rights reserved. Key Words: K⫹ current; Cl⫺ current; Adenosine 5⬘-triphosphate; (ATP); Cibacron blue; Purinoceptor; Patch clamp. Introduction Osteoclasts are the cells responsible for the resorption of mineralized tissues during physiological processes such as bone remodeling and tooth eruption and in pathological states such as osteoporosis, rheumatoid arthritis, and periodontitis.41 These multinucleated cells arise from fusion of mononucleated precursors of the monocyte-macrophage lineage that Address for correspondence and reprints: Dr. Stephen M. Sims, Department of Physiology, The University of Western Ontario, London, ON N6A 5C1, Canada. E-mail: [email protected] © 2000 by Elsevier Science Inc. All rights reserved.

5

8756-3282/00/$20.00 PII S8756-3282(00)00287-8

6

A. F. Weidema et al. Ion channels in human osteoclasts

Bone Vol. 27, No. 1 July 2000:5–11

Figure 1. Human osteoclasts from deciduous teeth. (A) Photomicrograph shows cells that stain positive for tartrate-resistant acid phosphatase (TRAP) on the tooth root surface. Image shows a portion of a resorbed tooth root with the pulp chamber at left. Boxed region is enlarged at right, showing cells on the dentine surface (densely stained regions) as well as resorption pits (faint staining). (B) Multinucleated cells were isolated and allowed to settle onto glass coverslips and studied using standard patch-clamp recording methods. Image of a representative cell, with the live cell at left, viewed using phase-contrast optics. The same cell is shown at right, stained for TRAP after recording currents in whole-cell configuration.

Methods and Materials Isolation of Osteoclasts Deciduous teeth, extracted for a variety of reasons, were obtained from the pediatric dental clinic of the University of Western Ontario with permission of the patient and the parent or guardian, and in accordance with the guidelines of the university review board for research involving human subjects. Following extraction, teeth were placed in culture medium consisting of Medium 199 (Gibco Laboratories, Burlington, ON) with HCO⫺ 3 (26 mmol/L), HEPES (25 mmol/L), antibiotics (100 U/mL penicillin, 100 ␮g/mL streptomycin, 0.25 ␮g/mL amphotericin B), and heat-inactivated fetal bovine serum (15% vol/vol). Cells were isolated from the resorbing root surface within 1– 4 h. Teeth were placed in Ca2⫹- and Mg2⫹-free phosphate-buffered saline containing 1 mM ethylene glycol-bis(␤-aminoethyl ether) N,N,N⬘, N⬘-tetraacetic acid (EGTA) and 1 mM ethylene-diamine tetraacetic acid (EDTA) for 1 h, whereupon flushing of the root with saline, or in some cases, scraping the root surface with a scalpel blade, liberated the osteoclasts. To enhance cell adhesion, Ca2⫹ and Mg2⫹ (2 mM each) were added and droplets (50 –100 ␮L) of cell suspension were placed on glass coverslips in culture dishes. After incubation for 30 – 60 min at 37°C, supplemented medium was gently added to the dish and cells were maintained in culture (5% CO2 95% air, 37°C) for up to 2 days. Coverslips contained variable numbers (2–20) of large, multinucleated cells (ⱖ3 nuclei), which stained positively for TRAP (Sigma Kit 387-A). Some extracted teeth were fixed with formaldehyde

(3.7%), stained for TRAP, and examined for the presence of osteoclasts. Electrophysiology For patch-clamp experiments, we studied multinucleated cells that adhered to glass coverslips. Currents were recorded from a total of 27 cells from eight different preparations. For recording ionic currents, we used conventional whole-cell or cell-attached configurations. Electrode solution contained (in millimoles per liter): KCl, 140; HEPES, 20; MgCl2, 1; EGTA, 0.1, at pH 7.2 (adjusted with KOH); 280 –290 mOsm/L. Cells were superfused (1–2 mL/min) with Na⫹ solution consisting of (in millimoles per liter): NaCl, 130; KCl, 5; MgCl2, 1; CaCl2 1; glucose, 10; HEPES, 20, at pH 7.4 (adjusted with NaOH); 280 –290 mOsm/L. Currents were recorded with an Axopatch 1D amplifier (Axon Instruments, Foster City, CA), filtered (⫺3 dB at 1 kHz) and digitized at 2–5 kHz using pCLAMP 6.0 (Axon Instruments). Current voltage (I-V) relationships were obtained using voltage jumps or ramp protocols. Experiments were performed at room temperature (21–25°C). Where noted in the figure legends, currents were corrected for a linear leak between ⫺30 mV and 30 mV, or basal currents were subtracted. Cell capacitances ranged from 80 to 300 pF, as determined from the amplifier circuitry or integration of the capacitive current transients. Uncanceled capacitive current transients for osteoclasts of capacitance ⬎100 pF were blanked for display. Values are presented as means ⫾ SEM. Test solutions were applied to osteoclasts by local superfu-

Bone Vol. 27, No. 1 July 2000:5–11

Figure 2. Inwardly rectifying K⫹ current in freshly isolated human osteoclasts. (A) Cell was held at 0 mV and voltage steps from ⫺130 to ⫹10 mV in 20 mV increments were applied, as shown at top. Large inward currents were observed at membrane potentials more negative than ⫺80 mV. (B) I-V relationship of the peak currents from the cell shown in (A), with currents corrected for a linear leak as described in Methods and Materials. The current reversed between ⫺70 and ⫺90 mV, consistent with this being a K⫹ current. (C) I-V relationships were obtained in other cells using voltage ramps (⫺100 to ⫹50 mV over 340 msec). With standard bath solution containing 5 mmol/L K⫹ and 130 mmol/L Na⫹, the current reversed at ⫺80 mV, close to the calculated equilibrium potential for K⫹. Increasing [K⫹]0 to 135 mmol/L by equimolar replacement of Na⫹ shifted the reversal potential to 0 mV, as predicted for a K⫹-selective current (n ⫽ 7).

sion from micropipettes (5–10 ␮m tip diameter) positioned 30 –50 ␮m from the cell, and coupled to a Picospritzer II (General Valve Corp., Fairfield, NJ). Application of control solutions did not cause appreciable changes in membrane currents. Chemicals were obtained from Sigma Chemical Co. (St. Louis, MO) or BDH, Inc. (Toronto, ON). Results Appearance of Human Osteoclasts Upon extraction, some teeth were stained for TRAP, revealing regions of resorbed dentine with intense staining, representing multinucleated TRAP-positive cells (Figure 1A). Regions of fainter staining were also observed on the dentine, and likely represent resorption pits, as described previously.7,30 Cells isolated from the tooth roots adhered to glass coverslips and had many features characteristic of osteoclasts, including large size, multiple nuclei, and extensive lamellipodia (Figure 1B). These isolated cells also stained for TRAP (Figure 1B, at right, illustrating the same cell). Although some erythrocytes were also observed in these preparations, few other cell types were present, indicating that a relatively pure population of osteoclasts had been obtained. Inwardly Rectifying K⫹ Current Coverslips containing viable human osteoclasts were placed in a recording chamber and superfused continuously with Na⫹ solution. In the majority of cells studied (23 of 27 cells isolated from eight

A. F. Weidema et al. Ion channels in human osteoclasts

7

Figure 3. Inhibition of inwardly rectifying K⫹ current by extracellular divalent cations. (A) I-V relationships were obtained using a voltage ramp protocol (⫺100 to ⫹50 mV over 340 msec). Under control conditions, inwardly rectifying K⫹ current was evident at negative potentials. Increase of extracellular Ca2⫹ from 1 to 5 mmol/L partially inhibited the current, with no shift in the activation range and no effect at more positive potentials (n ⫽ 3). (B) Addition of Ba2⫹ (5 mmol/L, n ⫽ 5) to the bath resulted in complete block of the inward rectifier. I-V relationships were obtained using voltage step protocols. (C) At lower concentrations (0.2 mmol/L), Ba2⫹ showed time- and voltage-dependent blockade of the inward rectifier, studied here with elevated K⫹ (140 mmol/L) in the bath. Current traces show control currents and those recorded with Ba2⫹ during voltage steps, as indicated. (D) Steady-state I-V relationship for the cell shown in (C), with and without Ba2⫹. Data in (A) and (B) were corrected for leakage as described in Methods and Materials.

different individuals), an inwardly rectifying current was observed (Figure 2). Hyperpolarizing voltage commands elicited currents that were inward at potentials more negative than ⫺80 mV (Figure 2B), with mean reversal potential (Erev) determined in six cells of ⫺78 ⫾ 2 mV, close to the predicted equilibrium potential for K⫹. Increasing extracellular K⫹ from 5 to 135 mmol/L, by equimolar replacement of Na⫹, shifted Erev to ⫺1 ⫾ 3 mV (n ⫽ 7), confirming selectivity of the conductance for K⫹ (Figure 2C). Many inwardly rectifying K⫹ currents are blocked by extracellular divalent cations,2 including Ca2⫹ and Ba2⫹, so we examined the effects of these ions on human osteoclast K⫹ currents. Bath application of 5 mmol/L Ca2⫹ reduced the inwardly rectifying K⫹ current by ⬎50%, without affecting its voltage sensitivity (Figure 3A). No additional currents were detected following elevation of extracellular Ca2⫹. A more complete blockade of the inward rectifier was obtained with extracellular Ba2⫹ (Figure 3B, 200 ␮mol/L or 5 mmol/L, n ⫽ 5). The block by Ca2⫹ and Ba2⫹ was reversible after washout. Ba2⫹ caused time- and voltage-dependent inhibition of the inward current, which was studied with cells bathed in 135 mmol/L K⫹ to allow us to record inward currents over a wider range of potentials (Figure 3C). Under control conditions, voltage steps from 0 to ⫺100 mV resulted in a large and steady inward current. Addition of Ba2⫹ (200 ␮mol/L) reduced the initial inward current by approximately 50%, with further time-dependent blockade. The rapid decrease of inward current at ⫺100 mV is consistent with an open channel block. The I-V relationship illustrated in Figure 3D shows the voltage dependence of Ba2⫹ inhibition of the inward rectifier at the end of a voltage step.

8

A. F. Weidema et al. Ion channels in human osteoclasts

Bone Vol. 27, No. 1 July 2000:5–11

Figure 4. Outwardly rectifying K⫹ current in human osteoclasts. (A) Whole-cell currents were measured at various potentials (⫺120 to ⫹120 mV in 20 mV increments) from a holding potential of 0 mV. At potentials more positive than ⫹40 mV, large outward currents were observed. (B) I-V relationship of the peak currents from the cell shown in (A), illustrating activation of this current only at positive voltages. (C) Application of tetraethylammonium (TEA; 10 mmol/L in application pipette) reversibly inhibited the outward current (n ⫽ 3). (D) Single-channel currents of large amplitude were recorded in the cell-attached patch configuration. Channel activity was measured at membrane potentials 80, 100, or 120 mV positive to the resting potential. Single-channel conductance was estimated to be 150 to 200 pS, assuming a reversal potential of 0 mV (n ⫽ 3).

Depolarization-activated K⫹ Current ⫹

In addition to the inwardly rectifying K current, an outwardly rectifying current was observed in 12 of 27 cells studied. This current was activated only at potentials more positive than 40 mV (Figure 4A,B) and was not dependent on the holding potential (data not shown). The presence of outward currents at depolarized potentials is most consistent with either a K⫹ or nonselective-selective channel. To distinguish these possibilities, we examined the pharmacology of the current. As illustrated in Figure 4C, whole-cell currents were reversibly blocked by the K⫹ channel blocker tetraethylammonium (TEA; 10 mmol/L, n ⫽ 3), consistent with the outward current being due to K⫹ channels. At positive potentials, the outward currents were characteristically “noisy,” suggestive of the large conductance Ca2⫹-dependent K⫹ channels, referred to as BKCa channels.40 Consistent with this, large conductance channels where recorded in cellattached patches at positive potentials (Figure 4D). These channels had a unitary conductance estimated to be 150 –200 pS, assuming a reversal potential of 0 mV (n ⫽ 4). However, some error may arise in this estimate owing to the voltage dependence of the channel, as activity was not recorded at voltages close to the reversal potential. Ca2⫹-dependent K⫹ Current Elicited by Extracellular Adenosine 5⬘-Triphosphate Extracellular ATP acts through P2X and P2Y classes of purinoceptors in many cell types27 and stimulates osteoclast formation and resorptive activity in vitro.5,22 The P2Y class of receptors is coupled through G proteins to phospholipase C, with activation leading to release of Ca2⫹ from intracellular stores in many cell types, including osteoclasts.38,42 Application of ATP (100 ␮mol/L) by local superfusion caused transient activation of

outward current (Figure 5A, observed in two of five cells tested). Current values recorded at ⫺30 mV are plotted in Figure 5A, showing that periodic stimulation with ATP could repeatedly activate outward current which was blocked by quinine (Figure 5A, right). Using voltage ramp commands, the I-V relationship of the current activated by ATP was determined to be essentially linear, reversing direction at negative potentials (Figure 5B), shown after subtraction of the basal inward rectifier. The linear I-V relationship of the ATP-activated K⫹ current differed from that found for the inward and outward rectifiers just described, but resembles the intermediate conductance Ca2⫹-dependent K⫹ current (IKCa)36 previously identified in rat and rabbit osteoclasts.24,38 To examine the dependence of this current on cytosolic free Ca2⫹ concentration ([Ca2⫹]i) in human osteoclasts, we applied the Ca2⫹ ionophore 4-bromo-A23187 to elevate levels of cytosolic Ca2⫹. This resulted in activation of a current with features similar to that induced by ATP, including a linear I-V relationship and reversal at negative potentials (Figure 5C, mean reversal potential of ⫺67 ⫾ 4 mV, n ⫽ 5), consistent with K⫹ selectivity. The ionophoreinduced IKCa current was sensitive to blockade by quinine (200 ␮mol/L), a blocker of Ca2⫹-dependent and other K⫹ channels (not shown, n ⫽ 3). Activation of an Outwardly Rectifying Cl⫺ Current An additional outward current often developed after establishing whole-cell configuration (n ⫽ 12), reaching a maximum after several minutes (Figure 6A). In other cells, this current was elicited by extracellular ATP (Figure 6A, at right). This current activated rapidly during voltage commands (Figure 6B) and displayed outward rectification. However, this current was distinct from those just described, reversing direction at 5 ⫾ 3 mV (Figure 6C, n ⫽ 10), a value close to the predicted Cl⫺ equilibrium potential. These characteristics are similar to those de-

Bone Vol. 27, No. 1 July 2000:5–11

Figure 5. Ca2⫹-dependent K⫹ current elicited by extracellular ATP. (A) Cell was held steadily at ⫺30 mV and ATP (100 ␮mol/L) was applied focally from an application pipette while the bath was continuously perfused (1–2 mL/min) with standard Na⫹ solution. Periodic application of ATP (20 sec, recorded at 5 min intervals) caused transient activation of an outward current. This effect was repeatable, as shown by two sequential responses. Addition of quinine (200 ␮mol/L) to the bath blocked the ATP-evoked outward current in the same cell (at right). Effect of quinine was reversible (not shown). (B) I-V relationships were obtained from voltage ramps (⫺100 to ⫹50 mV over 340 msec), where control current was subtracted from peak current during application of ATP. The ATP-activated current exhibited a linear I-V relationship and reversed direction close to ⫺70 mV. (C) Increasing cytosolic free Ca2⫹ concentration by application of the calcium ionophore 4-bromo-A23187 (10 ␮mol/L) resulted in activation of a similar current, revealing it to be a Ca2⫹-dependent K⫹ current. All traces were recorded from the same cell.

scribed previously for swelling-activated Cl⫺ currents in osteoclasts.10,17,31 The ionic basis of this current was investigated further using Cl⫺ channel blockers. This outwardly rectifying current was blocked reversibly by the Cl⫺ channel blocker niflumic acid (500 ␮mol/L), providing further evidence for this being a Cl⫺ current (Figure 6A,D). P2 purinoceptors have been reported to activate Cl⫺ currents in other cell types.37 Cibacron blue 3GA, an anthraquinonesulfonic acid derivative, is a relatively nonselective P2 purinoceptor antagonist (Cibacron blue is synonymous with reactive blue 2). Cibacron blue inhibits the ATP-induced rise of [Ca2⫹]i in rat and rabbit osteoclasts, which is largely mediated by P2Y receptors (A. F. Weidema, S. J. Dixon, and S. M. Sims, unpublished observations), indicating that it is an effective antagonist for at least some osteoclast P2 purinoceptors. Therefore, we investigated the effect of Cibacron blue (200 ␮mol/L) on Cl⫺ current in human osteoclasts, where it caused prompt reduction of ATP-activated outwardly rectifying Cl⫺ current (Figure 6A, at right). Interestingly, we noted that Cibacron blue also blocked the spontaneously activated Cl⫺ current (Figure 7), suggesting that it may directly block the Cl⫺ channels. Consistent with this, inhibition of the Cl⫺ current by Cibacron blue showed marked dependence on membrane potential (Figure 7A). Subtraction of the current recorded in the presence of blocker from control current revealed that sensitivity to Cibacron blue was greater at more positive potentials. Fractional inhibition of the conductance was plotted as a function of voltage (Figure 7B), and was well described by a Boltzmann relation, with half blockade at 34 ⫾ 3 mV (n ⫽ 3) negative to the Cl⫺ reversal potential, and a maximum of 86% blockade at positive voltages. Thus, in addition to its effect as an inhibitor of P2 purinoceptors, these results indicate that Cibacron blue directly blocks outwardly rectifying Cl⫺ current in human osteoclasts.

A. F. Weidema et al. Ion channels in human osteoclasts

9

Figure 6. Cl⫺ current in human osteoclasts. (A) An outwardly rectifying conductance developed spontaneously following establishment of wholecell recording. Trace began ⬃10 sec after establishing whole-cell configuration, with current at ⫹85 mV plotted as a function of time. Cell was held at 0 mV and voltage ramps (⫺100 mV to ⫹100 mV in 340 msec) were applied every 5 sec. Bath application of niflumic acid (500 ␮mol/ L), a chloride channel blocker, inhibited the outward current. I-V relationships were recorded using step protocols during the 3 min break in the record. At right, a similar current was activated in another cell following stimulation with ATP (100 ␮mol/L, 10 sec), applied by puffer pipette. This current was completely blocked when Cibacron blue was added to the bath (200 ␮mol/L). (B) Currents activated by ATP were recorded during the 1.5 min break in the record for the cell illustrated at right in (A). Under control conditions, only an inwardly rectifying current was seen (at left). After application of ATP, outward current was activated at more positive potentials (at right). (C) I-V relationships for the cell shown in (A) at left, obtained using voltage ramp commands. Timing of the voltage ramps illustrated in (C) and (D) are indicated. The outwardly rectifying current reversed direction close to ⫹10 mV. (D) After maximal activation, the current was inhibited by niflumic acid (500 ␮mol/L). Basal current, including inwardly rectifying K⫹ current, was subtracted from traces in (D).

Discussion We have described a novel method for isolating authentic osteoclasts from human deciduous teeth. These cells are multinucleated, stain positively for TRAP, and are associated with regions of root resorption. Using patch-clamp electrophysiological techniques, we demonstrated the expression of multiple types of ion channels in human osteoclasts, including three different K⫹ currents and a Cl⫺ current, as well as their regulation by ATP. Human osteoclasts express a complement of ion channels similar to that found in several other species, although some notable differences are apparent. Most human osteoclasts display an inwardly rectifying K⫹ current with properties of Kir2.1 (IRK1), which has been previously identified in rat, rabbit, murine, and avian osteoclasts.15,16,28,34,43 This current was identified based on its characteristic voltage-activation properties and time- and voltage-dependent blockade by Ba2⫹. The inward rectifier plays an important role in setting the membrane poten-

10

A. F. Weidema et al. Ion channels in human osteoclasts

Figure 7. Voltage-dependent inhibition of Cl⫺ current by Cibacron blue. Outwardly rectifying Cl⫺ current activated spontaneously in this cell. (A) I-V relationships were obtained using voltage ramp protocols. The basal inwardly rectifying K⫹ currents were subtracted from the traces illustrated. Following development of the Cl⫺ current, bath application of Cibacron blue (200 ␮mol/L) reversibly reduced the current by ⬎85% at positive potentials, while blockade at negative potentials was less marked. (B) The fractional inhibition of the Cl⫺ conductance by Cibacron blue is plotted as a function of voltage, for the same cell shown in (A). The solid line represents a Boltzmann relation fitted to the data, with half-maximal inhibition at ⫺20 mV, slope factor at 26 mV, and maximal blockade at 0.86.

tial of mammalian osteoclasts. Ba2⫹, which blocks the inward rectifier, depolarizes rat osteoclasts from ⫺70 mV to around 0 mV.34 Another role for the inward rectifier K⫹ channel may be dissipation of charge arising from activity of the electrogenic H⫹-ATPase in osteoclasts. Transport of H⫹ across the ruffled border leads to hyperpolarization of the osteoclast membrane, which would ultimately prevent further H⫹ efflux. The inwardly rectifying K⫹ conductance allows inward movement of K⫹ at hyperpolarized potentials, making it ideally suited to counteract the change in membrane potential arising from activity of the H⫹ pump. This current can therefore prevent excessive hyperpolarization, while minimizing efflux of K⫹ at depolarized potentials. During bone resorption, the concentration of Ca2⫹ around osteoclasts is increased, a change that may affect channel properties. Indeed, high concentrations of extracellular divalent cations did reduce the inwardly rectifying K⫹ conductance, as reported for other mammalian osteoclasts.2,15 Elevation of extracellular Ca2⫹ blocks the channel, reducing the conductance at all voltages. However, it has also been suggested that inhibition of the inward rectifier K⫹ channel involves binding of Ca2⫹ to a cell-surface receptor that is coupled through G proteins to the channel.43 Human osteoclasts expressed two classes of Ca2⫹-dependent ⫹ K currents. At the whole-cell level, depolarization revealed an outwardly rectifying, TEA-sensitive current, with voltage-activation properties suggestive of the BKCa channel. Consistent with this, large conductance channels were recorded in the cell-attached configuration. This current is similar to that reported in chicken osteoclasts,40 and has not been described previously in mammalian osteoclasts.39 Elevation of cytosolic Ca2⫹ concentration by stimulation with nucleotides or Ca2⫹ ionophore activated a K⫹ current whose underlying conductance showed little voltage dependence. In these respects, this resembles the intermediate conductance KCa (IKCa) current identified previously in rat and rabbit osteoclasts.24,38 This category of KCa current has not, to our knowledge, been reported in avian osteoclasts. In addition to the K⫹ currents just considered, human osteoclasts exhibited an outwardly rectifying current that was blocked by niflumic acid, identifying it as a Cl⫺ current. A similar current has been described in osteoclasts of other species10,17,35 that also activates during recording.35 The spontaneous development of Cl⫺ current is likely due to subtle swelling of cells during recording, as this current is induced in rabbit

Bone Vol. 27, No. 1 July 2000:5–11

osteoclasts most reproducibly in response to osmotic swelling.17 It has not yet been established whether the Cl⫺ current in human osteoclasts is activated by osmotic swelling. We did not observe activation of the Cl⫺ current when human osteoclasts were exposed to an increased concentration of extracellular Ca2⫹, as described in murine osteoclasts.31 The Cl⫺ current in human osteoclasts was activated by extracellular nucleotides, a form of regulation not previously reported in osteoclasts. The ATP-activated current developed with a slower time-course than the intermediate conductance Ca2⫹-dependent K⫹ current just considered, but the mechanisms underlying its activation are not presently known. It has been suggested that ATP, released by hyposmotic shock, activates Cl⫺ channels in hepatoma cells via P2 purinoceptors,37 because P2 purinoceptor antagonists suramin, PPADS (pyridoxal-5-phosphate-6-azophenyl-2⬘,4⬘-disulphonic acid), and Cibacron blue inhibit the swelling-activated current. We also found that Cibacron blue inhibited Cl⫺ current in human osteoclasts, which could be taken as evidence for the involvement of P2 purinoceptors. However, the short latency and voltage dependence of the blockade observed in osteoclasts suggests that Cibacron blue acts directly to block Cl⫺ channels. Such voltage-dependent block of Cl⫺ current by Cibacron blue occurs in epithelial cells6,11 and is consistent with it acting as an open channel blocker at positive potentials. Blockade is partially relieved by the efflux of anions through the channel at negative potentials, contributing to the observed voltage dependence. Cl⫺ movement through channels is essential for acidifying ruffled border membrane vesicles,4 and a Cl⫺ channel has been identified in avian osteoclasts.32 Cl⫺ channels may provide a conductive pathway to dissipate charge arising from electrogenic H⫹ pumping.4 It is also likely that swelling-induced Cl⫺ channels play a role in regulatory volume decrease in osteoclasts, as proposed for other cell types.20 A number of findings provide evidence that Cl⫺ channels play an essential role in bone resorption. The Cl⫺ channel blockers DIDS (4,4⬘-diisothiocyanatostilbene-2,2⬘-disulfonic acid) and SITS (4-acetamido-4⬘-isothiocyanatostilbene,-2,2⬘-disulfonic acid) reduce bone resorption,14,18 an effect attributed to inhibition of the Cl⫺/HCO⫺ 3 exchanger. The possibility also exists that DIDS and SITS inhibit resorption through blockade of osteoclast Cl⫺ channels. Extracellular ATP and its analog, ATP␥S, stimulate osteoclastic resorption in vitro5,22; it is possible that activation of the Cl⫺ current reported here contributes to enhanced resorptive activity. There are several types of currents that we did not observe in human osteoclasts. For example, we did not observe transient outward currents, such as Kv1.3, which has been found to be expressed in a greater percentage of rat osteoclasts adhering to dentine or collagen than in osteoclasts adhering to glass.1 It remains to be determined whether the use of resorbable substrates will influence the expression of this channel in human osteoclasts. We also did not detect nonselective cation currents, due to P2X receptors, in human osteoclasts stimulated with ATP. This is not due to lack of responsiveness, as ATP did elicit IKCa current, presumably through a rise of [Ca2⫹]i mediated by P2Y receptors. In addition, we have not yet studied human osteoclasts under conditions where K⫹ currents are blocked, an experimental condition that is essential for detecting several other currents, including voltage-activated Na⫹ channels12 and proton channels.25 In conclusion, we have reported a novel method for isolating authentic human osteoclasts that has enabled us to characterize, for the first time, ion channels in these cells. Human osteoclasts express a variety of channels, with similarities to mammalian and avian osteoclasts. The finding that ATP regulates K⫹ and Cl⫺ currents may be relevant to the processes by which extracellular nucleotides stimulate osteoclast formation and resorptive activity.5,22

Bone Vol. 27, No. 1 July 2000:5–11

Acknowledgments: The authors thank N. L. Highfield, Dr. H. S. Bobier, and Dr. S. J. Weinberger for assistance with the collection of teeth and isolation of cells and Lin Naemsch and Mary Pilkington for helpful comments on the manuscript. This work was supported by The Arthritis Society and the Canadian Arthritis Network. S.M.S. was the recipient of a Scientist Award from the Medical Research Council of Canada.

A. F. Weidema et al. Ion channels in human osteoclasts

22.

23.

References 24. 1. Arkett, S. A., Dixon, S. J., and Sims, S. M. Substrate influences rat osteoclast morphology and expression of potassium conductances. J Physiol (Lond) 458:633– 653; 1992. 2. Arkett, S. A., Dixon, S. J., and Sims, S. M. Effects of extracellular calcium and protons on osteoclast potassium currents. J Membr Biol 140:163–171; 1994. 3. Arkett, S. A., Dixon, J., Yang, J. N., Sakai, D. D., Minkin, C., and Sims, S. M. Mammalian osteoclasts express a transient potassium channel with properties of Kv1.3. Receptors Channels 2:281–293; 1994. 4. Blair, H. C., Teitelbaum, S. L., Tan, H. L., Koziol, C. M., and Schlesinger, P. H. Passive chloride permeability charge coupled to H⫹-ATPase of avian osteoclast ruffled membrane. Am J Physiol 260:C1315–1324; 1991. 5. Bowler, W. B., Littlewood-Evans, A., Bilbe, G., Gallagher, J. A., and Dixon, C. J. P2Y2 receptors are expressed by human osteoclasts of giant cell tumor but do not mediate ATP-induced bone resorption. Bone 22:195–200; 1998. 6. Clarke, L. L., Harline, M. C., Otero, M. A., Glover, G. G., Garrad, R. C., Krugh, B., Walker, N. M., Gonzalez, F. A., Turner, J. T., and Weisman, G. A. Desensitization of P2Y2 receptor-activated transepithelial anion secretion. Am J Physiol 276:C777–C787; 1999. 7. Domon, T., Osanai, M., Yasuda, M., Seki, E., Takahashi, S., Yamamoto, T., and Wakita, M. Mononuclear odontoclast participation in tooth resorption: The distribution of nuclei in human odontoclasts. Anat Rec 249:449 – 457; 1997. 8. Edwards, M., Sarma, U., and Flanagan, A. M. Macrophage colony-stimulating factor increases bone resorption by osteoclasts disaggregated from human fetal long bones. Bone 22:325–329; 1998. 9. Flanagan, A. M., Nui, B., Tinkler, S. M., Horton, M. A., Williams, D. M., and Chambers, T. J. The multinucleate cells in giant cell granulomas of the jaw are osteoclasts. Cancer 62:1139 –1145; 1988. 10. Fujita, H., Matsumoto, T., Kawashima, H., Ogata, E., Fujita, T., and Yamashita, N. Activation of Cl⫺ channels by extracellular Ca2⫹ in freshly isolated rabbit osteoclasts. J Cell Physiol 169:217–225; 1996. 11. Galietta, L. J., Falzoni, S., Di Virgilio, F., Romeo, G., and Zegarra-Moran, O. Characterization of volume-sensitive taurine- and Cl⫺-permeable channels. Am J Physiol 273:C57–C66; 1997. 12. Gaspar, R., Jr., Weidema, A. F., Krasznai, Z., Nijweide, P. J., and Ypey, D. L. Tetrodotoxin-sensitive fast Na⫹ current in embryonic chicken osteoclasts. Pflu¨gers Arch 430:596 –598; 1995. 13. Grano, M., Colucci, S., De Bellis, M., Zigrino, P., Argentino, L., Zambonin, G., Serra, M., Scotlandi, K., Teti, A., and Zambonin Zallone, A. New model for bone resorption study in vitro: Human osteoclast-like cells from giant cell tumors of bone. J Bone Miner Res 9:1013–1020; 1994. 14. Hall, T. J. and Chambers, T. J. Optimal bone resorption by isolated rat osteoclasts requires chloride/bicarbonate exchange. Calcif Tissue Int 45:378 – 380; 1989. 15. Hammerland, L. G., Parihar, A. S., Nemeth, E. F., and Sanguinetti, M. C. Voltage-activated potassium currents of rabbit osteoclasts: Effects of extracellular calcium. Am J Physiol 267:C1103–C1111; 1994. 16. Kelly, M. E., Dixon, S. J., and Sims, S. M. Inwardly rectifying potassium current in rabbit osteoclasts: A whole-cell and single-channel study. J Membr Biol 126:171–181; 1992. 17. Kelly, M. E., Dixon, S. J., and Sims, S. M. Outwardly rectifying chloride current in rabbit osteoclasts is activated by hyposmotic stimulation. J Physiol (Lond) 475:377–389; 1994. 18. Klein-Nulend, J. and Raisz, L. G. Effects of two inhibitors of anion transport on bone resorption in organ culture. Endocrinology 125:1019 –1024; 1989. 19. Lambrecht, J. T. and Marks, S. C., Jr. Human osteoclast-like cells in primary culture. Clin Anat 9:41– 45; 1996. 20. Lang, F., Busch, G. L., Ritter, M., Volkl, H., Waldegger, S., Gulbins, E., and Haussinger, D. Functional significance of cell volume regulatory mechanisms. Physiol Rev 78:247–306; 1998. 21. Matsuzaki, K., Udagawa, N., Takahashi, N., Yamaguchi, K., Yasuda, H.,

25.

26.

27. 28.

29. 30.

31.

32.

33.

34. 35. 36. 37.

38.

39.

40.

41.

42. 43.

11

Shima, N., Morinaga, T., Toyama, Y., Yabe, Y., Higashio, K., and Suda, T. Osteoclast differentiation factor (ODF) induces osteoclast-like cell formation in human peripheral blood mononuclear cell cultures. Biochem Biophys Res Commun 246:199 –204; 1998. Morrison, M. S., Turin, L., King, B. F., Burnstock, G., and Arnett, T. R. ATP is a potent stimulator of the activation and formation of rodent osteoclasts. J Physiol (Lond) 511:495–500; 1998. Murrills, R. J., Shane, E., Lindsay, R., and Dempster, D. W. Bone resorption by isolated human osteoclasts in vitro: Effects of calcitonin. J Bone Miner Res 4:259 –268; 1989. Naemsch, L. N., Weidema, A. F., Sims, S. M., Underhill, T. M., and Dixon, S. J. P2X4 purinoceptors mediate an ATP-activated, non-selective cation current in rabbit osteoclasts. J Cell Sci 112:4425– 4435; 1999. Nordstrom, T., Rotstein, O. D., Romanek, R., Asotra, S., Heersche, J. N., Manolson, M. F., Brisseau, G. F., and Grinstein, S. Regulation of cytoplasmic pH in osteoclasts. Contribution of proton pumps and a proton-selective conductance. J Biol Chem 270:2203–2212; 1995. Quinn, J. M., Neale, S., Fujikawa, Y., McGee, J. O., and Athanasou, N. A. Human osteoclast formation from blood monocytes, peritoneal macrophages, and bone marrow cells. Calcif Tissue Int 62:527–531; 1998. Ralevic, V. and Burnstock, G. Receptors for purines and pyrimidines. Pharmacol Rev 50:413– 492; 1998. Ravesloot, J. H., Ypey, D. L., Vrijheid-Lammers, T., and Nijweide, P. J. Voltage-activated K⫹ conductances in freshly isolated embryonic chicken osteoclasts. Proc Natl Acad Sci USA 86:6821– 6825; 1989. Roodman, G. D. Cell biology of the osteoclast. Exp Hematol 27:1229 –1241; 1999. Sahara, N., Toyoki, A., Ashizawa, Y., Deguchi, T., and Suzuki, K. Cytodifferentiation of the odontoclast prior to the shedding of human deciduous teeth: An ultrastructural and cytochemical study. Anat Rec 244:33– 49; 1996. Sakai, H., Nakamura, F., and Kuno, M. Synergetic activation of outwardly rectifying Cl⫺ currents by hypotonic stress and external Ca2⫹ in murine osteoclasts. J Physiol (Lond) 515:157–168; 1999. Schlesinger, P. H., Blair, H. C., Teitelbaum, S. L., and Edwards, J. C. Characterization of the osteoclast ruffled border chloride channel and its role in bone resorption. J Biol Chem 272:18636 –18643; 1997. Shigeyama, Y., Grove, T. K., Strayhorn, C., and Somerman, M. J. Expression of adhesion molecules during tooth resorption in feline teeth: A model system for aggressive osteoclastic activity. J Dent Res 75:1650 –1657; 1996. Sims, S. M. and Dixon, S. J. Inwardly rectifying K⫹ current in osteoclasts. Am J Physiol 256:C1277–1282; 1989. Sims, S. M., Kelly, M. E., and Dixon, S. J. K⫹ and Cl⫺ currents in freshly isolated rat osteoclasts. Pflu¨gers Arch 419:358 –370; 1991. Vergara, C., Latorre, R., Marrion, N. V., and Adelman, J. P. Calcium-activated potassium channels. Curr Opin Neurobiol 8:321–329; 1998. Wang, Y., Roman, R., Lidofsky, S. D., and Fitz, J. G. Autocrine signaling through ATP release represents a novel mechanism for cell volume regulation. Proc Natl Acad Sci USA 93:12020 –12025; 1996. Weidema, A. F., Barbera, J., Dixon, S. J., and Sims, S. M. Extracellular nucleotides activate non-selective cation and Ca2⫹-dependent K⫹ channels in rat osteoclasts. J Physiol (Lond) 503:303–315; 1997. Weidema, A. F., Dixon, S. J., and Sims, S. M. Ion channels in osteoclasts. In: Zaidi, M., Ed. Advances in Organ Biology, Vol. 5B, Molecular and Cellular Biology of Bone. Stamford, CT: JAI; 1998; 423– 442. Weidema, A. F., Ravesloot, J. H., Panyi, G., Nijweide, P. J., and Ypey, D. L. A Ca2⫹-dependent K⫹-channel in freshly isolated and cultured chick osteoclasts. Biochim Biophys Acta 1149:63–72; 1993. Wiebe, S. H., Hafezi, M., Sandhu, H. S., Sims, S. M., and Dixon, S. J. Osteoclast activation in inflammatory periodontal diseases. Oral Dis 2:167– 180; 1996. Wiebe, S. H., Sims, S. M., and Dixon, S. J. Calcium signalling via multiple P2 purinceptor subtypes in rat osteoclasts. Cell Physiol Biochem 9:323–337; 1999. Yamashita, N., Ishii, T., Ogata, E., and Matsumoto, T. Inhibition of inwardly rectifying K⫹ current by external Ca2⫹ ions in freshly isolated rabbit osteoclasts. J Physiol (Lond) 480:217–224; 1994.

Date Received: November 29, 1999 Date Revised: March 3, 2000 Date Accepted: March 6, 2000