Author's email addess A. F. Weidema: firstname.lastname@example.org
Extracellular Nucleotides Activate non-Selective Cation and Ca2+-Dependent K+ Channels in Rat Osteoclasts
Article first published online: 30 SEP 2004
The Journal of Physiology
Volume 503, Issue 2, pages 303–315, September 1997
How to Cite
Weidema, A. F., Barbera, J., Dixon, S. J. and Sims, S. M. (1997), Extracellular Nucleotides Activate non-Selective Cation and Ca2+-Dependent K+ Channels in Rat Osteoclasts. The Journal of Physiology, 503: 303–315. doi: 10.1111/j.1469-7793.1997.303bh.x
- Issue published online: 30 SEP 2004
- Article first published online: 30 SEP 2004
- Received 06 January 1997; accepted 16 May 1997.
- 1Extracellular ATP elevates cytosolic free Ca2+ concentration ([Ca2+]i) in osteoclasts, but its effects on ion channels have not been reported previously. Membrane currents and [Ca2+]i were recorded in isolated rat osteoclasts using patch clamp and fluorescence techniques.
- 2At negative membrane potentials, ATP (1–100 μm) activated an inward current that peaked rapidly and then declined. A later current was outward at potentials positive to the equilibrium potential for K+ (EK) and showed oscillations.
- 3The initial inward current, studied in isolation using Cs+ in the electrode solution, showed rapid activation, inward rectification and reversal at +3 ± 4 mV. Reduction of [Na+]o to 10 mM shifted the reversal potential to –21 ± 3 mV, indicating that ATP activates a non-selective cation current, consistent with involvement of P2X receptors.
- 4The later current activated by ATP, studied with K+ in the electrode solution, exhibited a linear I–V relationship, and reversed at –71 ± 4 mV. The reversal potential shifted 51 mV per 10-fold change of [K+]o, indicating that ATP activates a K+ current (IK).
- 5In fura-2-loaded cells, ATP caused elevation of [Ca2+]i that persisted in Ca2+-free solution, indicating that ATP induced release of Ca2+ from intracellular stores, consistent with involvement of P2Y receptors. Simultaneous patch clamp and fluorescence recordings revealed that IK was associated with the elevation of [Ca2+]i. Using a Ca2+ ionophore (4Br-A23l87) to elevate [Ca2+]iIK activated when [Ca2+]i exceeded ∼400 nm, with half-maximal activation at 580 ± 50 nM.
- 6In cell-attached patches, ATP activated a channel with a conductance of 48 ± 6 pS, that reversed direction near EK Channel open probability increased with elevation of [Ca2+]i, indicating the Ca2+ dependence of this channel.
- 7These results demonstrate that rat osteoclasts express two types of purinoceptors. P2X receptors give rise to non-selective cation current. P2Y receptors mediate Ca2+ release from stores, causing activation of a Ca2+-dependent K+ channel.
Osteoclasts are the cells responsible for resorption of bone and other mineralized tissues, and participate in physiological processes such as bone remodelling and tooth eruption. These multinucleated cells arise from fusion of mononucleated precursors of the monocyte-macrophage lineage that originate in the bone marrow (Roodman, 1996). Osteoclasts exhibit different phases of activity, with cells alternating between motile and resorbing phases. Motile osteoclasts typically exhibit extensive pseudopods and are flattened in appearance (spread morphology). In contrast, resorbing osteoclasts lack extensive pseudopodia and are more dome shaped (rounded morphology). Resorption is accomplished by acidification of a compartment referred to as the resorption lacuna. Transport of H+ by an electrogenic H+-ATPase causes dissolution of the mineral phase of bone, while secreted hydrolytic enzymes digest the organic matrix (Roodman, 1996). Since H+ transport is electrogenic, ion channels are required to provide pathways to dissipate charge.
Previous studies have revealed that mammalian osteoclasts express a number of channel types, including an inwardly rectifying K+ channel (IRK1), a transient outwardly rectifying K+ channel (Kv1.3), H+ channels and Cl− channels (Arkett, Dixon & Sims, 1992; Kelly, Dixon & Sims, 1992; Arkett, Dixon, Yang, Sakai, Minkin & Sims, 1994; Yamashita, Iahii, Ogata & Matsumoto, 1994; Kelly, Dixon & Sims, 1994; Nördstrom et al. 1995). The expression of channels is related to the morphology of the osteoclast, with spread osteoclasts exhibiting IRK1 and Cl− current, whereas rounded osteoclasts exhibit Kv1.3 and Cl currents (Arkett et al. 1992, 1994; Hammerland, Parihar, Nemeth & Sanguinetti, 1994).
Extracellular nucleotides are important signalling molecules mediating a number of processes, including neurotrans-mission, mechanosensation and regulation of proliferation (Dubyak & El-Moatassim, 1993; Burnstock, 1996; Nakamura & Strittmatter, 1996). Nucleotides, released by damaged cells and activated platelets and leukocytes (Born & Kratzer, 1984), may act as paracrine factors in early responses to trauma and inflammation. Nucleotides act through two classes of purinoceptors, P2Xand P2Y (Burnstock, 1996; North, 1996). P2X receptors are a family of ligand-gated channels that are non-selective for cations, and in many cases permit Ca2+ influx (Collo et al. 1996). In contrast, P2Y purinoceptors are members of the seven transmembrane spanning family of receptors that couple via G proteins to phospholipase C, causing the generation of inositol 1,4,5-trisphosphate and release of Ca2+ from intra-cellular stores (Burnstock, 1996).
Recent studies have revealed the presence of nucleotide receptors in osteoclasts. Rabbit osteoclasts respond to extracellular nucleotides with elevation of [Ca2+]i (Yu & Perrier, 1993). Exposure of murine osteoclasts to high concentrations of ATP leads to their permeabilization (Modderman, Weidema, Vrijheid-Lammers, Wassenaar & Nijweide, 1994). Furthermore, a P2Y2 (P2U) purinoceptor has been identified in human osteoclastoma, suggesting that at least this subtype is present in osteoclasts (Bowler, Birch, Gallagher & Bilbe, 1995). The regulation of osteoclast ion channels by nucleotides has not been described previously. In this report, we demonstrate that ATP activates two ionic conductances in mammalian osteoclasts via distinct signalling mechanisms. Portions of this work have appeared in abstract form (Weidema, Barbera, Dixon & Sims, 1996).
Osteoclasts were isolated from femora and tibiae of Wistar rat pups (up to 1 week old) that were killed by decapitation. Cells were plated onto plain or type I collagen-coated glass coverslips as described previously (Arkett et al. 1992). Coverslips with adherent osteoclasts were washed after 20 min to remove non-adherent cells and placed in culture medium consisting of Medium 199 with HCO3− (26 mm), Hepes (25 mm) (Gibco), antibiotics (100 u ml−1penicillin, 100 μg ml−1 streptomycin and 0.25 μg ml−1 amphotericin B) and heat-inactivated fetal bovine serum (15% v/v). Cells were studied within 14 h of isolation. Osteoclasts were identified as cells having multiple nuclei (≥3), and the identity of rounded osteoclasts was confirmed by counting nuclei upon disruption of the cell with a pipette at the end of recording.
For recording macroscopic currents, we used nystatin-perforated patch or conventional whole-cell configuration. The K+ electrode solution contained (mM): KCI, 140; Hepes, 20; MgCl2, 1; CaCl2 0.4; EGTA, 1 (∼100 nM free Ca2+); pH 7.2 (adjusted with KOH); 290 ± 5 mosmol l−1. In some experiments, the electrode solution contained CsCl to block K+ currents. In others, CI− in the electrode solution was reduced to 30 mw (isosmotic substitution with aspartate) to shift the chloride equilibrium potential (ECl to negative potentials. Cells were superfused (1–2 ml min−1) with Na+solution consisting of (mM): NaCl, 130; KC1, 5; glucose, 10; MgCl2, 1; CaCl2, 1; and Hepes, 20; pH 7.4 (adjusted with NaOH); 290 ± 5 mosmol l−1. To examine the selectivity of the ATP-evoked current, [Na+]o was reduced by replacement with N-methyl-D-glucamine+. Currents were recorded with Axopatch-1D or 200A amplifiers (Axon Instruments), filtered (–3 dB at 1 kHz) and digitized at 2–5 kHz using pCLAMP 6.0 (Axon Instruments). Currents were also stored on videotape using a pulse code modulator. Current–voltage (I–V) relationships were obtained using voltage ramp protocols, where voltage was shifted from –100 to +100 mV over 340 ms. Experiments were performed at room temperature (21–25 °C).
Fluorescence recording of [Ca2+1]i
[Ca2+]i was measured using fura-2 fluorescence from single osteoclasts. Cells were loaded by incubation with 2 μM fura-2 acetoxymethylester (Molecular Probes) for 30–60 min at room temperature. After washing, the cells were incubated at 37 °C for 30–60 min to allow for ester hydrolysis. Coverslips containing fura-2-loaded cells were placed in a chamber (0.75 ml) mounted on a Nikon Diaphot inverted microscope and continuously superfused with Na+ solution. Cells were illuminated by epifluorescence with alternating 340 and 380 nm light from a Xenon lamp and a Nikon Fluor × 40 objective lens. The emission signal was filtered using a 510 nm bandpass filter, detected by a photomultiplier (Photon Technology International (PTI), South Brunswick, NJ, USA) and sampled at 5–20 ratios s−1 (Felix software, PTI). [Ca2+]i, was calculated from the ratio of the fluorescence intensities at 340 and 380 nm following correction for background. System calibration constants were obtained using fura-2 solutions as described by Grynkiewicz, Poenie & Tsien (1985). In experiments involving simultaneous patch clamp and fluorescence recordings, cells were loaded with fura-2 prior to establishing perforated patch or cell-attached configurations. In cases where the whole-cell configuration was used, 30 μM fura-2 salt was included in the electrode solution and EGTA was reduced to 0.01 mM. Current and Ca2+ traces were recorded simultaneously using Felix software to ensure proper temporal alignment, while currents were also recorded at high bandwidth on digital video tape and/or digitized using pCLAMP software.
Test solutions were applied to cells by local superfusion from micropipettes (5–10 μM diameter) positioned 30–50 μM from the cell (Picospritzer II, General Valve Corporation, Fairfield, NJ, USA). The delay in the application system was 50–150 ms, as determined by the shift of the reversal potential of IRK1 upon application of high K+ solution. Application of control solutions did not cause appreciable changes in membrane currents or [Ca2+]i. Agonists tested included ATP, adenosine 5’-O-3-thiotriphosphate (ATPγS), 2-methylthioadenosine 5’-triphosphate (2-MeSATP), AMP and adenosine. 2-MeSATP, charybdotoxin and apamin were from Research Biochemicals International. Kaliotoxin was from Alomone Labs (Jerusalem, Israel). Unless otherwise indicated, chemicals were from Sigma or BDH. Values are presented as means ± standard deviation.
Effects of extracellular ATP on whole-cell currents of osteoclasts
To study the effects of ATP on osteoclast conductances, we first used a K+-containing electrode solution, and held cells steadily at various potentials, to minimize the contribution of voltage-activated currents, such as Kv1.3 (Arkett et al. 1992, 1994). ATP was applied periodically to cells, allowing at least 4 min between applications for recovery. When held at –60 or –90 mV, ATP evoked inward currents that were rapid in onset and that declined with continued application (Fig. 1A). At –30 mV, ATP evoked an initial current that declined and was followed by oscillations of outward current. At 0 mV, ATP only elicited an outward current with small oscillations (Fig. 1A). This biphasic pattern suggested the sequential activation of two currents, an initial inward current and a later outward current. Support for this was obtained using voltage ramp commands to investigate the I–V relationship of the ATP-induced currents. The initial inward current was inwardly rectifying and reversed direction close to 0 mV, whereas the later current was linear and reversed close to –55 mV (Fig. 1B). Experiments described below were designed to separate and characterize each of these ATP-induced currents.
We have previously shown that osteoclasts express two phenotypes (Arkett et al. 1992). Spread osteoclasts are flattened, have lamellipodia, and express an inwardly rectifying K+ conductance (IRK1). In contrast, rounded osteoclasts are dome shaped, lack lamellipodia and express an outwardly rectifying K+ conductance (Kvl .3). Neither of these conductances were altered to any great extent by ATP (over the course of 20–40 s). The pattern of ATP-induced currents described above was observed more frequently in rounded than in spread osteoclasts. All the figures illustrate responses recorded in rounded osteoclasts plated on collagen-coated coverslips.
Non-selective cation current
To study the initial inward current in isolation, we blocked K+ currents with Cs+ in the electrode solution and shifted ECl to –40 mV by reducing [Cl−] in the recording electrode solution. Under these conditions, ATP caused activation of the inward current at –60 mV with a very short latency (50–200 ms), that inactivated within several seconds (Fig. 2A). The I–V relationship of the ATP-induced current was determined using voltage ramp commands, by subtracting the current recorded under control conditions. The ATP-induced current showed inward rectification and reversed direction close to 0 mV with 135 him Na+ in the bath (Fig. 2B; mean reversal potential +3 ± 4 mV, n= 14), similar to that observed with K+ in the electrode solution (Fig. 1B). Taken together, these data indicate that ATP activates a non-selective cation current that is not blocked by Cs+ and rule out the involvement of Cl− current. Further evidence for cation selectivity was obtained by reducing [Na+]o to 10 mM (Na+ replaced by N-methyl-D-glucamine+), which shifted the reversal potential to –21 ± 3 mV (n= 8; Fig. 2B). Based on rapid activation and inactivation, cation selectivity and inward rectification, the inward current activated by ATP in osteoclasts has properties consistent with involvement of the P2X class of ligand-gated ion channels (North, 1996). Although we did not examine in detail the potency of various agonists, inward currents were activated by ATP (from 1 to 100 μM) as well as 2-MeSATP (10μM) and the poorly hydrolysable analogue ATPγS (50 μM, see below). Sapid inactivation of the current and variability of repeated responses in many cells precluded the accurate determination of dose–response relationships.
The later phase of the current activated by ATP was outward (see Fig. 1), and was blocked by Cs+ in the electrode solution, suggesting it was due to K+ channels. To study the voltage dependence of the current, we used K+ electrode solutions and voltage ramps. During these experiments, we set the holding potential at 0 mV to inactivate most of the Kv1.3 channels (Arkett et al. 1994). Under these conditions, ATP induced an outward current that peaked and then declined (Fig. 3A). The I–V relationship of the ATP-induced current was linear and reversed direction close to –75 mV with [K+]o at 5 mm (Fig. 3B). When [K+]o was increased (by substitution of Na+) to shift the equilibrium potential for K+ (EK) positively, the reversal potential also shifted positively (Fig. 3C). The reversal potential shifted 51 mV per 10-fold elevation of [K+]o (Fig. 3D) suggesting that ATP activated a current that was largely selective for K+. The linear K+ current (IK), activated by nucleotides, was observed in thirty of fifty-five rounded osteoclasts studied, and is distinct from all K+ currents previously described in mammalian osteoclasts (Dixon, Arkett & Sims, 1993).
Nucleotides induce transient elevation of [Ca2+]i
Nucleotides, acting on the P2Y class of purinoceptors, mediate an elevation of [Ca2+]i in many cell systems (Burnstock, 1996), including rabbit osteoclasts (Yu & Ferrier, 1993). Many K+ channels are Ca2+-dependent, so we considered the possibility that nucleotide-induced changes of [Ca2+]i contribute to the activation of IK. Using fura-2-loaded rat osteoclasts, ATP was found to cause rapid elevation of [Ca2+]i with transient and sustained phases (Fig. 4A). Following the initial peak response, oscillations of [Ca2+]i were frequently observed in rounded osteoclasts (Fig. 4B). In the example shown, oscillations were elicited by 2-MeSATP. ATP and ATPγS also elicited oscillations in other osteoclasts (not shown). In contrast, neither AMP nor adenosine (10–50μM) caused any Ca2+ response in cells which subsequently responded to ATP. These data argue against the involvement of PI receptors activated by adenosine, formed from breakdown of ATP by ectonucleotidases.
To investigate the involvement of intracellular Ca2+-stores, cells were superfused with Ca2+-free bath solution (with 0.5mM EGTA) to eliminate Ca2+ influx. Under these conditions, the nucleotide-induced rise of [Ca2+]i was still present, and oscillations persisted (Fig. 4B, right-hand side), providing evidence for the release of Ca2+ from stores. Nucleotide-induced elevation of [Ca2+]i was seen in greater than 90% of spread and rounded osteoclasts, with more than forty cells tested of each morphology. However, the amplitude of nucleotide-induced elevation of [Ca2+]i was generally smaller in spread cells.
Relationship between nucleotide-induced [Ca24]i transients and K+ current
The elevations of [Ca2+]i closely resembled the patterns of nucleotide-induced outward currents shown above (Figs 1 and 3), consistent with Ca2+ activating IK. We used simultaneous patch clamp and fluorescence recordings to investigate the relation between [Ca2+]i and IK in individual fura-2-loaded osteoclasts. When nucleotides induced a transient elevation of [Ca2+]i followed by a plateau, IK activated with virtually the same time course (Fig. 5A). The correspondence between [Ca2+]i and IK was even more apparent in cells exhibiting oscillatory changes of [Ca2+]i (Fig. 5B). Close correlation between nucleotide-induced changes of [Ca2+]i and IK was seen in eleven cells, and provides evidence that this K+ current is activated by elevation of [Ca2+]i If this is correct, elevation of [Ca2+]i would be expected to precede activation of IK. As seen on an expanded time scale (Fig. 6A), elevation of [Ca2+]i did become apparent prior to the outward K+ current. Furthermore, IK was not observed in cells where nucleotides failed to elevate [Ca2+]i in combined fluorescence and patch clamp experiments. In contrast to the good correlation between Ca2+ elevation and activation of IK, the inward cation current (Ication) did not correspond well with Ca2+ elevation. Activation of Ication preceded elevation of [Ca2+]i and the current decayed while [Ca2+]i remained high (Fig. 6B).
Ca2+ dependence of IK
To confirm the role of [Ca2+]i in activation of IK, we used the calcium ionophore 4Br-A23l87 to elevate [Ca2+]i in the absence of nucleotides. Using combined patch clamp and fluorescence recordings, 4Br-A23l87 induced a slow increase of [Ca2+]i accompanied by an increase of outward current (Fig. 7A). This outward current was identified as IK based on its linear I–V relationship and reversal at –70 ± 3 mV (n= 5). The Ca2+ sensitivity was determined by plotting IK at 0 mV as a function of [Ca2+]i (Fig. 7B). For this analysis, the amplitude of IK was corrected for the basal current level at 0 mV (which represented persistent current through Kv1.3 channels) and normalized for cell capacitance. The response shown in Fig. 7B is represented by open circles in Fig. 7B, and was fitted with a sigmoidal function (continuous line) with half-maximal activation at 550 nM [Ca2+]i and a Hill coefficient of 5.4. Mean values for four cells revealed half-maximal activation at 580 ± 50 nM, and a Hill coefficient of 5.5 ± 1.5. The correlation between IK and [Ca2+]i was determined during the rising phase of responses to 4Br-A23l87, but qualitatively similar relationships were observed during the recovery period (data not shown). We confirmed that IK activated by ATP showed a similar dependence on [Ca2+]i The open squares in Fig. 7B represent the initial phase of the response shown in Fig. 5B, and correspond well with that seen in response to the ionophore. In nine spread osteoclasts studied, ATP did not elicit an outward K+ current as described above for rounded osteoclasts. However, fluorescence studies revealed that in the spread cells the peak of the Ca2+ transient elicited by nucleotides rarely exceeded 400 nM. Since this may have been insufficient to activate IK, 4Br-A23l87 was used to elevate [Ca2+]i above 800 nM, under which condition IK still did not activate in five spread osteoclasts tested.
Single channel conductance of IK
To examine the channels underlying nucleotide-induced IK, we studied channel activity in cell-attached patches, while monitoring [Ca2+]i in the same cell. These experiments were carried out with 135 mm K+ in the pipette solution, so that the activation of Kv1.3 channels with depolarization, as well as their reversal at EK gave an indication of the resting membrane potential of the cell. In unstimulated cells, there were typically few channel openings at negative potentials. However, upon application of nucleotides to the cell membrane outside the recording pipette, a transient increase in channel activity occurred, giving rise to inward current even at very negative potentials (Fig. 8). Two current sweeps are superimposed in Fig. 8A, showing the control I–V relationship with no channel activity, and the channel current recorded during stimulation with ATPγS (traces truncated at +20 mV to eliminate Kv1.3 channel currents). The unitary I–V relationship had a slope conductance of 50 pS and an extrapolated reversal near EK (as indicated by reversal of current through Kv1.3; Fig. 8A). Similar single channel currents were observed in eight of fourteen cells, which showed nucleotide-induced elevation of [Ca2+]i, with a mean channel conductance of 48 ± 6 pS.
Channel open probability was quantified from the average current above baseline divided by unitary current amplitude (NPo, where N is number of channels and Po is open probability). NPo increased transiently in response to ATPγS (Fig. 8B, upper panel), and was correlated with the rise of [Ca2+]i (Fig. 8B, lower panel; representative of 3 cells). A 50 pS channel was also apparent in some patches accompanying seal formation, possibly reflecting a local elevation of [Ca2+]i In some patches studied, nucleotides caused hyperpolarization of cells, which was apparent as a shift in the reversal of current through Kv1.3. Changes of membrane potential were studied in current clamp, where ATP was found to cause brief depolarization followed by hyperpolarization (3 cells, data not shown). Taken together, these observations support the interpretation that an intermediate conductance K+ channel underlies the whole-cell IK activated by nucleotides. IK was not blocked by apamin (100 nM, 3 cells), a blocker of small conductance Ca2+-activated K+ channels; charybdotoxin (100 nM, 4 cells), a blocker of large conductance Ca2+-activated K+channels; or kaliotoxin (1 μM, 3 cells), a blocker of intermediate and large conductance Ca2+-activated K+ channels. The effectiveness of charybdotoxin and kaliotoxin was confirmed by their ability to inhibit Kv1.3 current (Arkett et al. 1994).
We have demonstrated that extracellular ATP activates two distinct conductances, which have not been described previously in mammalian osteoclasts. The initial inward cation current occurred with short latency, consistent with activation of ligand-gated P2X receptor channels. The later current was outward, more prolonged, and associated with P2Y receptor-mediated Ca2+ release, consistent with activation of Ca2+-dependent K+ channels.
Comparison of nucleotide-induced cation current with P2X currents in other cell types
Several observations establish that the initial current elicited by ATP in osteoclasts is non-selective for cations. This current reversed direction close to 0 mV, with either K+ or Cs+ in the pipette solution and Na+ in the bath. Since the equilibrium potential for Cl− was –40 mV in these experiments, reversal close to 0 mV indicates that the current is not selective for Cl−. Further evidence for cation selectivity was provided by the negative shift of the reversal potential seen upon decreasing extracellular Na+ concentration. Another feature of the ATP-induced inward current was strong inward rectification. Taken together, these observations are consistent with the presence of P2X receptors in rat osteoclasts. As expected for a ligand-gated P2X channel, the inward current activated rapidly, with a delay of 50–200 ms, although this delay is probably overestimated because of a delay in the system used to apply ATP. Finally, in the continued presence of ATP, the inward current inactivated within seconds. When expressed in heterologous systems, the cloned P2X channels exhibit inactivation rates both faster (P2X1 P2X3) and slower (P2X2, P2X4, P2X5, P2X6) than the native channel in osteoclasts (Collo et al. 1996). It should be noted that expression of multiple P2X subtypes in a single cell can result in the formation of heteromultimeric channels, which differ in their functional characteristics from homo-multimeric P2X channels (North, 1996). Accordingly, classification of the osteoclastic P2X receptor subtype(s) must await molecular studies.
ATP, in the absence of extracellular Mg2+, has been shown to induce a large non-selective conductance in marine osteoclasts (Modderman et al. 1994). This is consistent with activation of P2Z receptors, which form pores permeable to large molecules up to 800 Da in macrophages and macrophage-like cell lines (Hickman, Semrad & Silverstein, 1996). The P2Z receptor has recently been identified as a member of the P2X receptor family (P2X7), which mediates formation of cytolytic pores in the absence of Mg2+ (Surprenant, Rassendren, Kawashima, North & Buell, 1996). However, the P2X current observed in rat osteoclasts differs from the murine P2Z current in a number of respects. The murine P2Z conductance does not inactivate, has a linear I–V relationship, and is only activated by ATP at millimolar concentrations in the absence of Mg2+(Modderman et al. 1994).
ATP-activated non-selective cation current in osteoclasts exhibits some features in common with currents previously described in macrophages, a cell type that arises in the bone marrow from a common hematopoietic progenitor. In peritoneal murine and rat macrophages, ATP induces a non-selective cation current that rapidly activates but does not completely inactivate (Alonso-Torre & Trautmann, 1993; Naumov, Kaznacheyeva, Kiselyov, Kuryshev, Mamin & Mozhayeva, 1995). The ATP-activated channel in rat macrophages exhibits a relatively high permeability for Ca2+, and thus is an important Ca2+ influx pathway (Alonso-Torre & Trautmann, 1993). A number of P2X receptor subtypes have been found to be permeable to Ca2+, with a permeability ratio of Ca2+: Na+ of 4:1, when ion activities are taken into account (Collo et al. 1996; North, 1996). The contribution of P2X receptors to Ca2+ influx in osteoclasts remains to be determined. The P2X current in rat osteoclasts also has properties resembling those of ATP-induced currents in megakaryocytes, based on ionic selectivity, and rapid activation and inactivation (Somasundaram & Mahaut-Smith, 1994).
Nucleotide-induced activation of K+ conductance
The later phase of the ATP-induced current in osteoclasts has a linear I–V relationship and is selective for K+, based on its reversal close to EK over a range of K+ gradients and a 51 mV shift per 10-fold change in [K+]o. The deviation from the predicted 59 mV shift for a pure K+ current may reflect limited permeability of the channel to other cations, such as Na+, some residual P2X current or activation of a small chloride current. In this regard, outwardly rectifying chloride conductances have been described in osteoclasts, activated by osmotic swelling (Kelly et al. 1992) or extracellular Ca2+ (Fujita, Matsumoto, Kawashima, Ogata, Fujita & Yamashita, 1996). However, in experiments where IK was inhibited by Cs+, there was no evidence for a contribution of a chloride conductance to the initial current activated by ATP (e.g. Fig. 2).
IK activated with a longer delay than the P2X current, suggesting the involvement of second messengers. Extracellular nucleotides, acting on P2Y receptors, induce an elevation of [Ca2+]i in many cells (Dubyak & El-Moatassim, 1993; Burnstock, 1996), including rabbit osteoclasts (Yu & Ferrier, 1993). We have confirmed this observation in rat osteoclasts, and shown that nucleotide-induced elevation of [Ca2+]i persists in Ca2+-free bath solution. Accordingly, we conclude that rat osteoclasts possess two classes of purinoceptors, P2X and P2Y (also confirmed in rabbit osteoclasts, data not shown). Using combined patch clamp and fluorescence techniques, we found that IK closely paralleled changes in [Ca2+]i Furthermore, in cells where oscillations of [Ca2+]i were apparent, oscillations of IK were also evident. Such oscillations persisted when cells were held under voltage clamp, indicating that changes of membrane potential are not required. In contrast, oscillations of [Ca2+]i in the Jurkat T cell line require changes in membrane potential (Grissmer, Lewis & Cahalan, 1992).
The calcium sensitivity of the osteoclast IK was studied using the ionophore 4Br-A23l87 in combined patch clamp and fluorescence experiments. Ionophore-induced elevation of [Ca2+]i activated a linear current which reversed close to EK, similar to the ATP-induced current. The K+ current was strongly dependent on [Ca2+]i, with half-maximal activation at 580 nM [Ca2+]i and a Hill coefficient of 5.5. Similar Hill coefficients have been reported for Ca2+-activated K+ current in other systems (Grissmer et al. 1992; Köhler et al. 1996), and suggest that multiple Ca2+-binding sites are involved in channel opening. Channel activation by ionophore in the absence of nucleotides emphasizes the primary role of [Ca2+]i in activation of IK.
Several properties of the Ca2+-dependent K+ current provide an indication of the type of channels expressed in osteoclasts. The linear I–V relationship indicates a lack of voltage dependency, a feature that distinguishes this current from many other Ca2+-dependent K+ currents (Haylett & Jenkinson, 1990). A large conductance Ca2+-dependent K+ channel (150 pS) has been characterized in avian osteoclasts (Weidema, Ravesloot, Panyi, Nijweide & Ypey, 1993), but the avian channel is strongly voltage dependent, with open probability increasing at more positive potentials. Other K+ currents reported in osteoclasts, such as Kv1.3 and IRK1, have non-linear I–V relationships and are not activated by Ca2+ (Ravesloot, Ypey, Vrijheid-Lammers & Nijweide, 1989; Kelly et al. 1992; Arkett et al. 1992, 1994; Hammerland et al. 1994).
An indication of the type of channel underlying the whole-cell IK came from studies in the cell-attached patch configuration. Under these conditions, stimulation of osteoclasts with ATP caused elevation of [Ca2+]i which was accompanied by increased channel activity. This channel had a unitary conductance of 50 pS and no apparent voltage dependence. Furthermore, the channel current reversed near the predicted, EK, as expected for a K+-selective channel. Since nucleotides were applied to the rest of the cell membrane, and not directly to channels in the recording pipette, activation of this channel must involve a diffusible cytoplasmic signal. Based upon these observations, this intermediate conductance Ca2+-dependent K+ channel is a likely candidate to account for the whole-cell IK activated by nucleotides in osteoclasts. A similar current is activated in macrophages by ATP (Alonso-Torre & Trautmann, 1993). Studies of Ca2+-dependent K+ channels in lymphocytes revealed two classes of channels, of small and intermediate conductance (Grissmer et al. 1992). The small conductance channel (4–7 pS) is voltage independent and apamin sensitive, whereas the intermediate conductance channel (40–60pS) is insensitive to apamin but blocked by charybdotoxin. Within the class of small conductance Ca2+-dependent K+ channels, two major groups can be distinguished based on their pharmacology (Hanselmann & Grissmer, 1996). The first is blocked by charybdotoxin, which also blocks many large conductance Ca2+-dependent K+ channels. The second group is blocked by apamin. The IK activated by ATP in rat osteoclasts was not blocked by apamin, charybdotoxin or kaliotoxin, but did show Ca2+sensitivity similar to the small conductance channel described by Grissmer et al. (1992). Kaliotoxin blocks intermediate (60 pS) and large conductance Ca2+-dependent K+ channels in neurons (Crest et al. 1992).
Two small conductance Ca2+-activated K+ channels (9–11 pS) were recently cloned from rat and human brain (Köhler et al. 1996). Expression of the mRNA for these homologous channels results in Ca2+-dependent K+ currents that are voltage independent and differ in their sensitivity to apamin. The Ca2+ sensitivity of these cloned channels is similar to that reported for a number of native K+ channels, including those in lymphocytes (Grissmer et al. 1992) and osteoclasts. Thus, nucleotides activate a Ca2+-dependent K+ channel not previously reported in mammalian osteoclasts. The insensitivity of the osteoclast K+ channel to charybdotoxin and apamin distinguishes it from Ca2+-dependent K+ channels in other cell types.
Possible role of nucleotides in regulation of osteoclast function
ATP accumulates in the extracellular fluid at sites of inflammation and tissue injury (Born & Kratzer, 1984; Osipchuk & Cahalan, 1992), where it may function as a local mediator. In this regard, osteoclastic resorption is enhanced at sites of inflammation (Wiebe, Hafezi, Sandhu, Sims & Dixon, 1996). Under physiological conditions, nucleotides may act as autocrine/paracrine signalling molecules following their secretion by members of the ATP-binding cassette family of transport proteins, such as P-glycoprotein (Al-Awqati, 1995). P-glycoprotein has recently been identified in calcifying chondrocytes and osteoblasts (Mangham et al. 1996), suggesting an additional source of ATP to act on osteoclasts. Osteoclastic bone resorption has been shown to be inhibited by the non-selective P2-antagonist suramin (Yoneda, Williams, Rhine, Boyce, Dunstan & Mundy, 1995), suggesting that purinoceptors are involved in regulation of osteoclast number or activity.
It has been suggested that purinoceptors are involved in mechanosensation (Nakamura & Strittmatter, 1996). Heterologous expression of P2Y1 receptors in oocytes renders them mechanosensitive through an autocrine mechanism involving release of ATP. Since bone is known to remodel in response to mechanical stimuli, it is tempting to speculate that purinoceptors mediate mechanotransduction in bone cells such as osteoclasts. It has been suggested that in other systems P2 purinoceptors play a role in cell fusion and apoptosis (Di Virgilio, 1995), processes that are important in the formation and death of osteoclasts (Kameda, Ishikawa & Tsutsui, 1995; Roodman, 1996).
We have shown that ATP activates P2Xand P2Y purinoceptors on rat osteoclasts. Responses were predominantly observed in rounded osteoclasts, rather than spread cells, suggesting that the functional expression of these receptors is regulated. Activation of these receptors results in opening of non-selective cations channels, transient elevation of [Ca2+]i and complex changes in membrane potential. It is possible that P2X channels provide a Ca2+-influx pathway, which contributes to elevation of [Ca2+]i Subsequent activation of IK hyper-polarizes the membrane potential, further enhancing Ca2+ influx. How these complex responses to extracellular ATP contribute to the regulation of osteoclast number or activity is yet to be determined.
- 1995). Regulation of ion channels by ABC transporters that secrete ATP. Science 269, 805–806. (
- 1993). Calcium responses elicited by nucleotides in macrophages. Interaction between two receptor subtypes. Journal of Biological Chemistry 268, 18640–18647. & (
- 1992). Substrate influences rat osteoclast morphology and expression of potassium conductances. Journal of Physiology 458, 633–653. , & (
- 1994). Mammalian osteoclasts express a transient potassium channel with properties of Kv1.3. Receptors and Channels 2, 281–293. , , , , & (
- 1984). Source and concentration of extracellular adenosine triphosphate during haemostasis in rats, rabbits and man. Journal of Physiology 354, 419–429. & (
- 1995). Identification and cloning of human P2U purinoceptor present in osteoclastoma, bone, and osteoblasts. Journal of Bone and Mineral Research 10, 1137–1145. , , & (
- 1996). P2 purinoceptors: historical perspective and classification. In P2 Purinoceptors: Localization, Function and Transduction Mechanisms, ed. Chadwick, D. J. & Goode, J. A., Ciba Foundation Symposium, vol. 198, pp. 1–34. Wiley, Chichester . (
- 1996). Cloning of the P2X5 and P2X6 receptors and the distribution and properties of an extended family of ATP-gated ion channels. Journal of Neuroscience 16, 2495–2507. , , , , , & (
- 1992). Kaliotoxin, a novel peptidyl inhibitor of neuronal BK-type Ca2+-activated K+ channels characterized from Androctonus mauretanicus mauretanicus venom. Journal of Biological Chemistry 267, 1640–1647. , , , , , , & (
- 1995). The P2Z purinoceptor: an intriguing role in immunity, inflammation and cell death. Immunology Today 16, 524–528.DOI: 10.1016/0167-5699(95)80045-X (
- 1993). Electrophysiology of osteoclasts and macrophages. In Blood Cell Biochemistry, vol. 5, Macrophages and Related Cells, ed. Horton, M. A., pp. 203–222. Plenum, New York . , , (
- 1993). Signal transduction via P2-purinergic receptors for extracellular ATP and other nucleotides. American Journal of Physiology 265, C577–606. & (
- 1996). Activation of Cl− channels by extracellular Ca2+ in freshly isolated rabbit osteoclasts. Journal of Cellular Physiology 169, 217–225.DOI: 10.1002/(SICI)1097-4652(199610)169:1<217::AID-JCP22>3.0.CO;2-8 , , , , & (
- 1992). Ca2+-activated K+ channels in human leukemic T cells. Journal of General Physiology 99, 63–84.DOI: 10.1085/jgp.99.1.63 , & (
- 1985). A new generation of Ca2+ indicators with greatly improved fluorescence properties. Journal of Biological Chemistry 260, 3440–3450. , & (
- 1994). Voltage-activated potassium currents of rabbit osteoclasts: effects of extracellular calcium. American Journal of Physiology 267, C1l03–1111. , , & (
- 1996). Characterization of apamin-sensitive Ca2+-activated potassium channels in human leukaemic T-lymphocytes. Journal of Physiology 496, 627–637. & (
- 1990). Calcium-activated potassium channels In Potassium Channels. Structure, Classification, Function and Therapeutic Potential. Series in Pharmaceutical Technology, ed. Cooke, N., pp. 71–95. Ellis Horwood Limited, New York . & (
- 1996). P2Z purinoceptors In P2 purinoceptors: Localization, Function and Transduction Mechanisms, ed. Chadwick, D. J. & Goode, J. A., Ciba Foundation Symposium, vol. 198, pp. 71–90. Wiley, Chichester . , & (
- 1995). Detection and characterization of apoptosis in osteoclasts in vitro. Biochemical and Biophysical Research Communications 207, 753–760.DOI: 10.1006/bbrc.1995.1251 , & (
- 1992). Inwardly rectifying potassium current in rabbit osteoclasts: A whole-cell and single-channel study. Journal of Membrane Biology 126, 171–181. , & (
- 1994). Outwardly rectifying chloride current in rabbit osteoclasts is activated by hyposmotic stimulation. Journal of Physiology 475, 377–389. , & (
- 1996). Small-conductance, calcium-activated potassium channels from mammalian brain. Science 273, 1709–1714. , , , , , & (
- 1996). P-Glycoprotein is expressed in the mineralizing regions of the skeleton. Calcified Tissue International 58, 186–191.DOI: 10.1007/s002239900031 , , , , , , & (
- 1994). Permeabilization of cells of hemopoietic origin by extracellular ATP4-: Elimination of osteoclasts, macrophages and their precursors from isolated bone cell populations and fetal bone rudiments. Calcified Tissue International 55, 141–150. , , , & (
- 1996). P2Y1, purinergic receptors in sensory neurons: contribution to touch-induced impulse generation. Proceedings of the National Academy of Sciences of the USA 93, 10465–10470.DOI: 10.1073/pnas.93.19.10465 & (
- 1995). ATP-activated inward current and calcium-permeable channels in rat macrophage plasma membranes. Journal of Physiology 486, 323–337. , , , , & (
- 1995). Regulation of cytoplasmic pH in osteoclasts. Contribution of proton pumps and a proton-selective conductance. Journal of Biological Chemistry 270, 2203–2212.DOI: 10.1074/jbc.270.5.2203 , , , , , , & (
- 1996). P2X receptors: a third major class of ligand-gated ion channels. In P2 Purinoceptors: Localization, Function and Transduction Mechanisms, ed. Chadwick, D. J. & Goode, J. A., Ciba Foundation Symposium, vol. 198, pp. 91–109. Wiley, Chichester . (
- 1992). Cell-to-cell spread of calcium signals mediated by ATP receptors in mast cells. Nature 359, 241–244.DOI: 10.1038/359241a0 & (
- 1989). Voltage activated K+ conductances in freshly isolated embryonic chicken osteoclasts. Proceedings of the National Academy of Sciences of the USA 86, 6821–6825. , , & (
- 1996). Advances in bone biology: the osteoclast. Endocrine Reviews 17, 308–332.DOI: 10.1210/er.17.4.308 (
- 1994). Three cation influx currents activated by purinergic receptor stimulation in rat megakaryocytes. Journal of Physiology 480, 225–231. & (
- 1996). The cytolytic P2Z receptor for extracellular ATP identified as a P2X receptor (P2×7). Science 272, 735–738. , , , & (
- 1993). A Ca2+-dependent K+ channel in freshly isolated and cultured chick osteoclasts. Biochimica et Biophysica Acta 1149, 63–72. , , , & (
- 1996). Extracellular nucleotides activate non-selective cation and K+ conductances in rat osteoclasts. Biophysical Journal 70, A97 (abstract). , , & (
- 1996). Osteoclast activation in inflammatory periodontal diseases. Oral Diseases 2, 167–180. , , , & (
- 1994). Inhibition of inwardly rectifying K+ current by external Ca2+ ions in freshly isolated rabbit osteoclasts. Journal of Physiology 480, 217–224. , , & (
- 1993). ATP induces an intracellular calcium pulse in osteoclasts. Biochemical and Biophysical Research Communications 191, 357–363.DOI: 10.1006/bbrc.1993.1225 & (
- 1995). Suramin suppresses hypercalcemia and osteoclastic bone resorption in nude mice bearing a human squamous cancer. Cancer Research 55, 1989–1993. , , , , & (