- Email: [email protected]
Intracellular Calcium Puffs in Osteoclasts
Experimental Cell Research 253, 689 – 696 (1999) Article ID excr.1999.4714, available online at http://www.idealibrary.com on
Intracellular Calcium Puffs in Osteoclasts Wilson Radding,* S. Elizabeth Jordan,† Raymond B. Hester,‡ and Harry C. Blair† ,§ ,1 *Keck Center for Computational Biology, Rice University, Houston, Texas; †Veterans’ Affairs Medical Center, Birmingham, Alabama; ‡Department of Microbiology and Immunology and the Division of Biotechnical Services, University of South Alabama, Mobile, Alabama; and §Center for Metabolic Bone Diseases and Department of Pathology, University of Alabama School of Medicine, Birmingham, Alabama
We studied intracellular calcium ([Ca 21] i) in acidsecreting bone-attached osteoclasts, which produce a high-calcium acidic extracellular compartment. Acid secretion and [Ca 21] i were followed using H 1-restricted dyes and fura-2 or fluo-3. Whole cell calcium of acid-secreting osteoclasts was ;100 nM, similar to cells on inert substrate that do not secrete acid. However, measurements in restricted areas of the cell showed [Ca 21] i transients to 500 –1000 nM consistent with calcium puffs, transient (millisecond) localized calcium elevations reported in other cells. Spot measurements at 50-ms intervals indicated that puffs were typically less than 400 ms. Transients did not propagate in waves across the cell in scanning confocal measurements. Calcium puffs occurred mainly over regions of acid secretion as determined using lysotracker red DND99 and occurred at irregular periods averaging 5–15 s in acid secreting cells, but were rare in lysotracker-negative nonsecretory cells. The calmodulin antagonist trifluoperazine, cell-surface calcium transport inhibitors lanthanum or barium, and the endoplasmic reticulum ATPase inhibitor thapsigargin had variable acute effects on the mean [Ca 21] i and puff frequency. However, none of these agents prevented calcium puff activity, suggesting that the mechanism producing the puffs is independent of these processes. We conclude that [Ca 21] i transients in osteoclasts are increased in acid-secreting osteoclasts, and that the puffs occur mainly near the acid-transporting membrane. Cell membrane acid transport requires calcium, suggesting that calcium puffs function to maintain acid secretion. However, membrane H 1ATPase activity was insensitive to calcium in the 100 nM–1 mM range. Thus, any effects of calcium puffs on osteoclastic acid transport must be indirect. © 1999 Academic Press
Key Words: LaCl 3; trifluoperazine; BaCl 2; thapsigargin; V-type H 1-ATPase.
1 To whom correspondence and reprint requests should be addressed at 535 LHRB, 701 S. 19th Street, Birmingham, AL 352940007. Fax: (205) 934-9927. E-mail: [email protected].
INTRODUCTION
Osteoclasts secrete HCl onto the surface of bone, which is composed mainly of acid-soluble calcium salts, to maintain serum calcium and for periodical replacement of the skeleton. Osteoclasts use an expanded (ruffled) cell membrane that contains a massive concentration of vacuolar-type H 1-ATPase to produce this acid [1]. The acid is secreted into an extracellular compartment formed by the ring-shaped osteoclastic attachment, and consequently this space has a very high calcium activity [2]. Calcium is removed from this compartment by vacuolar transcytosis [3, 4]. There are high concentrations of calmodulin at the osteoclast ruffled border [5] consistent with calcium-modulated control processes. Several studies have suggested that osteoclasts respond to calcium, and intracellular calcium signals are generated in osteoclasts by a variety of stimuli. Osteoclast– bone attachment causes transient changes in intracellular calcium activity 2 ([Ca 21] i) [6], and similar calcium signals are evoked by integrin receptors [7]. On the other hand, some intracellular calcium signals do not depend on integrin binding [8]. Elevated extracellular calcium outside the usual bone– osteoclast attachment site alters the bone resorption rate inversely to [Ca 21] i [9] and causes cell retraction via ryanodinesensitive receptors [10]. However, attachment to bone is required for full osteoclast differentiation, including the membrane specialization for acid secretion at the bone attachment, and bone-attached osteoclasts are relatively insensitive to extracellular calcium [11]. This suggests that calcium-sensitive membrane proteins are modified or redistributed in bone-attached osteoclasts. Further, calcium transients have been observed in osteoclasts in the absence of extracellular calcium [12], calling into question whether extracellular calcium is important to the bone-resorbing osteoclast. Thus, it is unclear whether bone degradation affects 2 Fluorescent dye-binding measures calcium activity. For simplicity, dye-binding activity measurements are indicated as [Ca 21] i or as calcium concentration where the meaning is unaffected.
689
0014-4827/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
690
RADDING ET AL.
intracellular calcium or vice versa. We explored the hypothesis that osteoclastic intracellular calcium is increased during bone resorption, which would be expected to make the cell more sensitive to calciumactivation via membrane-associated calmodulin. We used the fluorescent weak base probe lysotracker red DND 99 to locate acid secreting osteoclasts attached to bone. Then, using fluorescent calcium probes, we compared intracellular calcium in these osteoclasts to calcium in osteoclasts attached to inert substrates where no acid is secreted. We were unable to distinguish the average intracellular calcium of acid-secreting and nonsecreting cells. However, transient calcium puffs, localized millisecond increases in calcium, were seen, which occurred much more frequently in acid secreting cells. These calcium transients occurred mainly near the site of acid secretion. METHODS Avian osteoclast cultures. Osteoclast-rich medullary bone from calcium starved chickens was harvested and osteoclast-rich preparations were made using bone affinity [13]. Cells on bone were cultured on glass coverslips in tissue culture plates to allow them to be used for fluorometry. Cells were cultured in Dubelco’s modified Eagle’s minimal essential medium with 10% heat inactivated serum at 37°C in 5% CO 2, except during fluorescence studies, which used a modified Krebs–Ringer’s solution with bicarbonate and glucose (NaCl 120 mM, KCl 5 mM, Ca 1.3 mM, phosphate 1.2 mM, MgSO 4 1.2 mM, Hepes 20 mM, glucose 2 mM, bicarbonate 22 mM). To minimize the possible effects of upward drift in pH of bicarbonate buffers exposed to air during fluorescence studies, pH was monitored and results shown were obtained with pH between 7.35 and 7.59. Identification of acid secreting cells. Acid lakes produced by osteoclasts were identified with the fluorescent polyamine–pyrrole dye lysotracker red DND-99 (lysotracker) (Molecular Probes, Eugene, OR). This dye diffuses through membranes at neutral pH since a major fraction of their phenazine/pyrrole rings and amine side groups are uncharged; the amines are protonated at low pH and accumulate in acid compartments because membranes are impermeable to ions. Lysotracker, rather than neutral red or acridine [14, 15], was used in calcium studies because its long-wavelength excitation and emission do not overlap fura-2 or fluo-3. However, neutral red, 25 mM, added 30 min prior to observations [14], was used to calibrate Lysotracker for use in demonstrating osteoclast acid secretion (Fig. 1B, right). Lysotracker, 2–3 mM, was added 3–5 min before observation using a 560-nm low-pass excitation filter and a 600-nm emission barrier filter. Intracellular calcium measurement. Fura-2 (1-[2-(5-carboxyoxazole)-6-amino-5-benzofuranoxy]-2-(2-amino-4-methylphenoxy)ethane-N,N,N9,N9-tetraacetic acid) or fluo-3 (1-[2-amino-5-(2,7dichloro-6-hydroxy-3-oxy-9-xanthenyl)phenoxy]-2-[2-amino-5methylphenoxy]ethane-N,N,N9,N9-tetraacetic acid) was used as their membrane-permeable acetoxymethyl esters (-AM) as described [6]. Fluors were from Molecular Probes. Cells were incubated in 10 mM fura-2-AM or fluo-3-AM for 60 –90 min, allowing cell esterases to release the membrane-impermeant free calcium-binding polycarboxylic acids. After washing to remove unreacted probe, fluorescence was measured. Fura-2 was used for dual excitation measurements in a defined microscopic field, and fluo-3 for single wavelength scanning measurements to indicate variability within a scanned area. For fura-2, measurements used excitation at 340 and 380 nm to differentiate Ca 21-bound and unbound dye using the Spex (Edison,
NJ) AR-CM-MAC dual excitation fluorescence on a Diaphot inverted epifluorescence microscope (Nikon, Melville, NY). Cell regions were isolated using an optical mask. Emission fluorescence was monitored by photomultiplier with computer recording. A typical signal was ;10 4 counts per 250 ms of a 3 3 3-mm region. Calcium activities were calculated by the ratio method of Grynkiewicz et al. [16], using a fura–2 K Ca of 224 nM. Fluo-3 measurements were single-wavelength, relative Ca 21 activities by dye fluorescence emission at 525 nM with excitation at 488 nm; dye emission intensity varies approximately fourfold between 40 nM and 3 mM calcium. These measurements were 30 pixel/line (900 point) scans at 3-s intervals on an ACAS 570 laser scanning confocal microscope (Meridian Instruments, Okemos, MI) as described [17]. Pixel step was 0.8 mm, with the scanned area 24 mm across. Fluo-3 was used to indicate calcium-dependent relative fluorescence intensity, shown using linear scaling. Permeabilization of Fluo-3 loaded cells was used to verify maximum dye response 630% across the fields shown to assure that relative differences reported substantially exceed background fluorescence variations in dye loading. Effect of calcium on membrane acid transport activity. Osteoclast membranes were obtained by nitrogen cavitation and sequential centrifugation [18]; membrane vesicles were reconstituted at 1 mg/ml protein by incubation for 30 min at 4°C in 120 mM KCl, 20 mM NaCl, 10 mM Hepes, pH 7.4 [18]. Calcium activity was adjusted using buffers with 0.5 mM EGTA and 2 to 400 mM calcium. Calcium activity was calculated using Ca z EGTA K D of 380 nM [16]. Mg 21ATP-dependent HCl transport was determined by acridine orange (3 mM) fluorescence quenching, recording dye emission at 540 nm with excitation at 468 nm [18] in a SLM/Aminco-Bowman (Urbana, IL) Series 2 luminescence spectrophotometer. Reactions were performed at 37°C with 15 mg vesicle protein in 2.5 ml of 400 mM ATP, 4 mM acridine orange in 120 mM KCl, 20 mM NaCl, 10 mM Hepes, pH 7.4; transport was initiated with 500 mM MgCl 2. Washout with 1 mM NH 4Cl was used to control for nonspecific changes in fluorescence.
RESULTS
Intracellular calcium activity in bone-attached and glass attached osteoclasts. Calibrated whole-cell Fura-2 measurements were used in these measurements; optical windows contained one isolated cell (;20-mm square fields), and 1-s integration was used. Measurements of [Ca 21] i in several acid-secreting osteoclasts on bone averaged ;100 nM. This was not significantly different from [Ca 21] i of osteoclasts off bone (Fig. 1A). Appearance of a typical field with lysotracker, showing that the bone-attached cells produced acid but the unattached cells did not is also shown (Fig. 1B). Because it had been hypothesized that calcium would increase in acid-secreting cells, a second series of calcium measurements was done in osteoclasts on and off bone, with similar results. Differences of 20 –30 nM would not be resolved due to experimental variability, but it was clear that any differences in the average intracellular calciums of osteoclasts on and off bone would be small. Whole-cell calcium studies on additional cell preparations, with 30 individual calibrated measurements, gave similar results. Instability of [Ca 21] i of acid secreting osteoclasts. Analysis of fura-2 measurements at 1-s intervals suggested that calcium of osteoclasts on bone was more
INTRACELLULAR CALCIUM PUFFS IN OSTEOCLASTS
highly variable than in cells on glass. Occasional spikes of ;500 nM were seen, which were too large to attribute to noise. Further, Xia and Ferrier [21] had reported osteoclast calcium fluctuations. To determine whether the variability in measurements was due to spatially limited or short-duration calcium events, high-resolution studies were performed. Fluorometric measurements were refined by recording at 250-ms integration intervals and by limiting the studied area to a 3 3 3-mm region of the osteoclast. These measurements showed transient micrometer-scale calcium pulse activity with duration ,1 s (calcium puffs) (Fig. 2A). These calcium puffs (or spikes) to ;500 nM occurred at irregular intervals, in Fig. 2A averaging ;16 s. The frequency of puffs varied between experiments. This may reflect regional differences in the cells studied or variation between cells. Fourier analysis of traces was performed to determine whether there was an underlying periodicity, but none was found. In scanning confocal measurements, the calcium puffs were seen as ;2-mm regions of elevated calcium, which varied in location between frames at 3-s intervals (Fig. 2B). Comparison with lysotracker analysis indicated that the calcium puffs correlated very well with the region of acid secretion. Similar results were obtained when bone-attached and basolateral regions of a cell were studied using calibrated measurements in spots ;2 mm square (Fig. 2C). The nuclear area and membrane away from bone showed little activity, and was similar to that seen in glass-attached cells (not illustrated) while membrane adjacent to the acidified bone attachment had much higher activity. Recordings us-
691
ing minimum instrument integration intervals of 50 ms showed that the duration of the puffs was typically 200 – 400 ms (Fig. 2C, inset). Effects of calcium transport inhibitors. The effect of the calmodulin antagonist trifluoperazine, 10 mM, to inhibit the Ca 21-ATPase [19], which caused a slow increase in the average intracellular calcium, is shown in Figs. 3A and 3B. Both glass-attached and acid-secreting cells showed increasing average intracellular calcium after addition of trifluoperazine. However, there was much more variation in the traces obtained from cells on bone. The effects of 30 mM thapsigargin, an inhibitor of endoplasmic reticulum ATPases (Figs. 3C– 3D), and of blocking cell-surface calcium channels with 100 mM lanthanum (Fig. 3E) or barium (Fig. 3F) were tested. None of the treatments abolished the calcium puffs, and indeed the number and size of calcium puffs typically increased over several minutes, although these late effects may reflect the toxicity of these agents. These results are selected frames from 50 –90 sequential 3-s time points obtained in each study. Each agent was studied in three or more time-course measurement series, with similar results in each study. These findings suggest that the calcium transients do not depend wholly on intracellular calcium stores or on lanthanum-sensitive cell surface calcium channels; these agents may not affect membranes at the acidification site, however (see Discussion). Effect of calcium activity on osteoclast membrane ATP-dependent HCl transport. Osteoclast acid secretion is known to be abolished by calmodulin antago-
FIG. 1. Intracellular calcium and acid secretion in osteoclasts on and off bone. (A) Intracellular calcium in individually calibrated fura-2 whole-cell measurements, eight osteoclasts on bone and six osteoclasts off bone. Mean 6 standard deviation. The difference is not significant. (B) Localization of acid transporting cells. Lysotracker (left) was used to determine acid-transporting cells in further studies. Acid lakes are indicated; these are 10- to 20-mm diameter regions on the bone surface. They appear oval when seen face-on and linear when seen edge-on. Faint granular staining represents lysosomal reaction or nonspecific fluorescence from bone particles. Transmitted light (right) shows another coverslip from the same preparation, stained with neutral red [14] to visualize the acid lakes. Note that glass-attached cells do not form acid lakes and are hence not seen in the fluorescent assay. Cells bound to the edge of the bone allow the acid-secreting region to be distinguished from other cell regions by point measurements (see Fig. 2C). In either case, inclusion of 1 mM ammonium chloride, a nonfluorescent weak base that also accumulates in acidic compartments, blunted dye uptake to essentially undetectable levels (not illustrated). FIG. 2. Intracellular calcium in bone-attached osteoclasts. Acid secreting cells were located using lysotracker as in Fig. 1B (left). (A) Fluorometric measurements with 250-ms integration in a 3 3 3-mm acid-secreting region. Transient calcium pulse activity to ;500 –1000 mM, with duration ,1 s is seen (calcium puffs), two of which are indicated by arrows. In these high-resolution measurements, fluctuations in the 50 –200 nM range are indistinguishable from photon-counting noise. The spikes indicated represent consistent differences of a much larger magnitude (see also (C), below, which resolves the calcium puffs into multiple measurements showing elevated calcium using 50-ms measurements). A difficulty in these measurements is that visual determination of cellular regions is subject to observer interpretation of phase images; however, similar results were obtained using confocal measurements over the whole cell or when subcellular regions were selected (B and C, following). (B) Nine successive scanning confocal measurements of fluo-3 activity in an acid-secreting osteoclast, at 3-s intervals. These are false-color illustrations, which convert the photon counts of the green fluorescence of the fluo-3 into intensity readings: white . yellow . green . light blue . dark blue. Black, no signal. The arrows are examples of calcium elevations that move from frame to frame (many other examples are also present). The calcium puffs are ;2-mm regions of elevated calcium that vary in location between frames, but correlate with the region of acid secretion, which is outlined in frame 1. (C) Similar results were obtained when bone-attached and basolateral regions of a cell were studied using calibrated point measurements. The nuclear area and membrane away from bone (similar to regions 2 and 1 in the cell outlined in Fig. 1B, right) showed little activity, while membrane adjacent to bone had much higher activity (a region similar to region 3 in the cell outlined in Fig. 1B, right). A 50-ms trace of a similar region shows the duration of the puffs, 200 – 400 ms (inset).
692
RADDING ET AL.
INTRACELLULAR CALCIUM PUFFS IN OSTEOCLASTS
693
FIG. 3. Effect of calcium transport inhibitors. In each study shown, the acid-secreting region and cells were localized using lysotracker dye, as in Fig. 1, and phase microscopy. The effect of the calmodulin antagonist trifluoperazine (10 mM), an inhibitor of endoplasmic reticulum ATPases, thapsigargin (30 mM), and cell-surface calcium channel blocking cations, lanthanum or barium (100 or 50 mM) on calcium puffs in acid-transporting regions are shown using calibrated point measurements (A, C, and E) and scanning confocal measurement of relative calcium activity (B, D, and F). None of the treatments abolished the calcium puffs, and the number and size of calcium puffs tended to increase over several minutes after the calcium transport antagonists were added. Representative traces using acid-secreting cells with basal calcium of ;100 nM are shown, although calcium activity varied severalfold between cells, in keeping with variability noted within and between cells (Figs. 2B and 2C). Here, this effect is most apparent after lanthanum addition, although in some traces thapsigargin markedly increased activity. Trifluoperazine caused an upward drift in calcium, since this paralyzes the key calcium ATPase; this effect typically required longer times, 5–10 min, and was also associated with increased in calcium transients. In the scanning measurements, the antagonists were added after the third frame shown. In the first frame, the position of the cell and its acid secreting site, determined as in Fig. 1B, is indicated. The fluo-3 loading of the cells shown varied resulting in different basal fluorescence return. However, as with the calibrated spot measurements, the treatments did not reduce, and typically increased, calcium puff activity.
694
RADDING ET AL.
FIG. 4. ATP-dependent acid transport by osteoclast membranes in buffers with 1 nM–10 mM free calcium. Acid transport by 15 mg of osteoclast cell-membrane vesicles was measured as 4 mM acridine uptake over 200 s in 400 mM ATP and 500 mM MgCl 2. Activity shown is arbitrary fluorescence units at 540 nm, determined by washout of acridine uptake with 3 mM NH 4Cl. Buffers included 500 mM EGTA except at the lowest calcium, containing 1 mM EGTA. One of two experiments with similar results is shown.
nists. Further, osteoclast acid secretion produces a high calcium concentration, and calcium puffs were observed mainly at the acid secretion site. Thus we studied the direct effects of calcium on membrane HCl transport. Acid transport activity was abolished by incubation of membranes in 1 mM EGTA, 400 mM ATP, 120 mM KCl, 20 mM NaCl, 10 mM Hepes, pH 7.4, 37°C, for even a few minutes. However, in buffers with calcium activity greater than 2 nM, there was no apparent variation in H 1-ATPase activity (Fig. 4). This indicates that calcium puffs appear unlikely to modify the H 1-ATPase directly. This experiment was repeated with three different osteoclast membrane vesicle preparations with essentially identical results. DISCUSSION
Calcium is a modulator of osteoclastic activity, although how calcium varies during physiological bone resorption is uncertain. Zheng et al. [22] used inhibitors to show that calcium-calmodulin kinase affects osteoclast retraction. The osteoclastic Ca 21 ATPase has a high ATP-affinity component that is inhibited by the calmodulin antagonist trifluoperazine [19]. Miyauchi et al. [23], studying [Ca 21] i decrement with a vb 3 integrin binding, attributed reduced [Ca 21] i to this pump; the calmodulin antagonist blocked Ca 21 efflux. In plastic attached osteoclasts [Ca 21] i variation has been related to the postintegrin receptor mechanisms [24]. However, intracellular calcium signals are complex and respond to signals independent from integrin binding [8]. Osteoclast membranes also contain an outwardly rectifying voltage-activated K 1 channel that is calcium
dependent [25], and there are reports of purinergic [26] and mechanical [27] calcium responses in osteoclasts. Our results are in keeping with these previous studies and further suggest that there are regional differences in osteoclast calcium activity (Fig. 2), although these differences are transient rather than continuous. Importantly, the results in this study were obtained using bone-attached osteoclasts; the properties of osteoclasts, including acid secretion, may vary after the cells are disaggregated from bone and allowed to reattach to tissue-culture plastic or glass (Fig. 1B). While macrophages and foreign body giant cells may have attachment sites that are to varying degrees acidic, such as seen in microprobe studies (reviewed in [28]), the bone-attachment site has a much greater accumulation of the HCl secreting pumps [28], and this attachment is critical to maintenance of the extremely active acid secretion apparatus of the osteoclast [1]. Osteoclast intracellular calcium has been measured under various conditions , but only recently have data been produced in bone attached cells. On plastic, osteoclast intracellular [Ca 21] varies with extracellular [Ca 21] and causes cell retraction via ryanodine-sensitive receptors [10]. However, Lakkakorpi et al. [11] showed that resorbing osteoclasts are much less sensitive to extracellular calcium, suggesting that boneattached cells have membrane calcium response mechanisms that may vary from plastic-attached cells. Our results are in accord with this, and suggest that there are regional differences in osteoclast calcium activity. We used acid-specific dyes to define osteoclasts that were secreting acid (Fig. 1B), and observed near the acidic, high-calcium bone-attachment transient micrometer-scale, 200- to 400-ms-duration calcium pulse activity (calcium puffs) (Figs. 2 and 3). However, we did not observe calcium waves. Calcium pulses, fluctuations, and repetitive spikes have been reported in rabbit osteoclasts [21], although the resolution of the measurements reported here suggest more finely localized, short-duration events. Further, although activity was observed in osteoclasts off bone, calcium puffs were much more frequently observed in osteoclasts on bone, and particularly in association with the acidic bone resorption site. However, it must be emphasized that the colocalization would not resolve calcium puffs originating near, but not in, the ruffled border, such as activity associated with vesicles or the cytoskeleton near the acid-secreting membrane. Acid secretion-related local calcium activity suggested a positive feedback stimulus for acid secretion, but osteoclast membrane vesicle HCl transport was not directly modulated by variation in calcium activity (Fig. 4). Thus, if the role of calcium puffs at the acid secretion site functions in a feedback mechanism, this is unlikely to act directly on the ruffled membrane HCl secreting apparatus. Indeed, the sensitivity of the
INTRACELLULAR CALCIUM PUFFS IN OSTEOCLASTS
membrane at very low calcium may reflect changes in the membrane that are not physiological, and the calcium-responsive elements may be largely lost during membrane preparation at intermediate calcium activity levels. Recent reports do relate osteoclastic intracellular calcium to bone resorption, however. Ritchie et al. [9] showed that calcium channel-blockers reduced osteoclastic bone resorption. General inhibitors of calcium or calmodulin also reduce osteoclastic bone degradation [5]. Thus, it is clear that calcium modifies bone resorption, but the mechanism is not a simple feedback effect at the level of the acid-secreting membrane. It is possible that transient increases in calcium near the high-calcium bone-resorption compartment reflect “leaks” in the ruffled border. The osteoclast’s secretory membrane has a large surface area and separates regions of very high and low calcium activity. However, observations in other cells suggest that, whatever the mechanism producing the calcium puffs, they are likely to have physiologically meaningful regulatory effects. In this regard, calmodulin kinase II activity changes with the frequency of calcium elevations [29]. Since osteoclast acid secretion activity is maintained by calmodulin [5], but there is no direct calcium sensitivity of membrane HCl transport, this mechanism is a likely mediator of whole-cell effects. Calcium puffs have been described in other cell types. Puffs of calcium of 300 –500 ms have been described in Xenopus oocytes [30]. Parker and Yao [31] summarize data on calcium puffs in Xenopus as consistent with concerted opening of several inositol trisphosphate gated calcium channels. The spatial size (2–3 mm) of calcium puffs in Xenopus is consistent with the calcium puffs observed in osteoclasts. In myocytes, similar calcium puffs occur, with frequency dependent on extracellular calcium. Our data are consistent generally with the fire– diffuse–fire model of propagation of calcium sparks in myocytes [32]. However, in this case the calcium spark sites are either not distributed in a simple array, or the array is arranged on a convoluted membrane structure. Nevertheless, the relation of puffs to proximity to the high calcium site suggest similar roles of osteoclastic calcium puffs and calcium puffs in muscle. Other observations have included calcium puffs originating mostly within 2–3 mm of nuclei of HeLa cells [33]. We cannot rule out an association of perinuclear membranes with calcium elevations in osteoclasts, but a major association in osteoclasts with the acid secretion site is much more likely from our observations; this difference probably reflects cellular specialization. In osteoclasts treated with the calmodulin antagonist trifluoperazine calcium increased as expected. This may reflect effects on the calcium ATPase or other calmodulin-sensitive proteins. Calcium transients per-
695
sisted and increased in frequency, indicating that the mechanism producing the calcium puffs is not calmodulin dependent. Blocking cell surface calcium channels with lanthanum or barium did not abolish calcium puffs, although it is unlikely that the cations would affect intracellular membranes or the isolated osteoclast-bone attachment. The specificities of La and Ba are not precisely the same, but both ions block major calcium channels, and between the two, if cell-surface calcium channels were important for the phenomenon observed, some effect would have been expected, but was not observed. Calcium puffs in osteoclasts were not abolished by thapsigargin, but these short-term experiments may have been inadequate to deplete calcium stores, so an endoplasmic reticulum-dependent mechanism is not excluded. Our results indicate that intermittent localized puffs of calcium occur in osteoclasts. These were seen predominately at or near the acid-secreting membrane. Osteoclast internal calcium was maintained robustly in the face of inhibitors, suggesting that the osteoclast expresses significant amounts of multiple plasma membrane and endoplasmic reticulum pumps, which is consistent with its role as a calcium-transporting cell [3]. We speculate that the calcium spikes or puffs observed may be related to calcium entering the cell from the adjacent high-calcium acid degradation site and that the calcium activity near the acid transport site may be used as a feedback signal to continue acid secretion. However, osteoclast membrane HCl secretion was not directly regulated by calcium, so if calcium puffs function in positive feedback for acid secretion, the mechanism is indirect. We thank Drs. John P. Williams and Robert W. Hardy, University of Alabama, Birmingham, for valuable discussions and assistance with Spex measurements. Supported in part by National Institutes of Health Grants AG12951 and AR43225, the University of Alabama Health Services Foundation, the Department of Veterans’ Affairs, and National Library of Medicine Grant LM07093.
REFERENCES 1.
Schlesinger, P. H., Mattsson, J. P., and Blair, H. C. (1994). Osteoclastic acid transport: mechanism and implications for physiological and pharmacological regulation. Miner. Electrolyte Metab. 20, 31–39.
2.
Silver, A., Murrills, R. J., and Etherington, D. J. (1988). Microelectrode studies on the acid microenvironment beneath adherent macrophages and osteoclasts. Exp. Cell Res. 175, 266 –276.
3.
Salo, J., Lehenkari, P., Metsikko, K., and Va¨a¨na¨nen, H. K. (1997). Removal of osteoclast bone resorption products by transcytosis. Science 276, 270 –273.
4.
Nesbitt, S. A., and Horton, M. A. (1997). Trafficking of matrix collagens through bone-resorbing osteoclasts. Science 276, 266 – 269.
5.
Radding, W., Williams, J. P., Hardy, R. W., McDonald, J. M., Whitaker, C. H., Turbat-Herrera, E. A., and Blair, H. C. (1994).
696
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
RADDING ET AL. Calmodulin concentrated at the osteoclast ruffled border modulates acid secretion. J. Cell Physiol. 160, 17–28. Teti, A., Blair, H. C., Schlesinger, P. Grano, M., ZamboninZallone, A., Kahn, A., Teitelbaum, S. L., and Hruska, K. A. (1989). Extracellular protons acidify osteoclasts, reduce cytosolic calcium and promote expression of cell-matrix attachment structures. J. Clin. Invest. 84, 773–780. Shankar, G., Davison, I., Helfrich, M. H., Mason, W. T., and Horton, M. A. (1993). Integrin receptor-mediated mobilisation of intranuclear calcium in rat osteoclasts. J. Cell Sci. 105, 61– 68. Shankar, G., Gadek, T. R., Burdick, D. J., Davison, I., Mason, W. T., and Horton, M. A. (1995). Structural determinants of calcium signaling by RGD peptides in rat osteoclasts: Integrindependent and -independent actions. Exp. Cell Res. 219, 364 – 371. Ritchie, C. K., Maercklein, P. B., and Fitzpatrick, L. A. (1994). Direct effect of calcium channel antagonists on osteoclast function: Alterations in bone resorption and intracellular calcium concentrations. Endocrinology 135, 996 –1003. Zaidi, M., Shankar, V. S., Tunwell, R., Adebanjo, O. A., Mackrill, J., Pazianas, M., O’Connell, D., Simon, B. J., Rifkin, B. R., Venkitaraman, A. R., Huang, C.L.-H., and Lai, F. A. (1995). A ryanodine receptor-like molecule expressed in the osteoclast plasma membrane functions in extracellular Ca 21 sensing. J. Clin. Invest. 96, 1582–1590. Lakkakorpi, P. T., Lehenkari, P. P., Rautiala, T. J., and Va¨a¨na¨nen, H. K. (1996). Different calcium sensitivity in osteoclasts on glass and on bone and maintenance of cytoskeletal structures on bone in the presence of high extracellular calcium. J. Cell Physiol. 168, 668 – 677. Shankar, V. S., Huang, C. L., Adebanjo, O. A., Pazianas, M., and Zaidi, M. (1994). Calcium influx and release in isolated rat osteoclasts. Exp. Physiol. 79, 537–545. Williams, J. P., Blair, H. C., McKenna, M. A., Jordan, S. E., and McDonald, J. M. (1996). Regulation of avian osteoclastic H 1ATPase and bone resorption by tamoxifen and calmodulin antagonists. J. Biol. Chem. 271, 12488 –12495. Barnicot, N. A. (1947). The supravital staining of osteoclasts with neutral-red: Their distribution on the parietal bone of normal growing mice, and a comparison with the mutants greylethal and hydrocephalus-3. Proc. R. Soc. B 134, 467– 485. Baron, R., Neff, L., Louvard, D., and Courtoy, P. J. (1985). Cell-mediated extracellular acidification and bone resorption: Evidence for a low pH in the resorbing lacunae and localization of a 100-kD lysosomal membrane protein at the osteoclast ruffled border. J. Cell Biol. 101, 2210 –2222. Grynkiewicz, G., Poenie, M., and Tsien, R. Y. (1985). A new generation of Ca 12 indicators with greatly improved fluorescence properties. J. Biol. Chem. 260, 3440 –3450. Hardy, R. W., Ladenson, J. H., Hruska, K. A., Jiwa, A. H., and McDonald, J. M. (1992). The effects of extracellular calcium and epinephrine on cytosolic-free calcium in single rat adipocytes. Endocrinology 130, 3694 –3702. Blair, H. C., Teitelbaum, S. L., Koziol, C. M., and Schlesinger, P. H. (1991). Passive chloride permeability charge-coupled to
Received July 20, 1999 Revised version received September 30, 1999
the electrogenic H 1-ATPase of avian osteoclast ruffled membrane. Am. J. Phys. 260, C1315–C1324. 19.
Bekker, P. J., and Gay, C. V. (1990). Characterization of a Ca 21-ATPase in osteoclast plasma membrane. J. Bone Miner. Res. 5, 557–567.
20.
Arkle, S., Wormstone, I. M., Bascal, Z. A., and Dacke, C. G. (1994). Estimation of intracellular calcium activity in confluent monolayers of primary cultures of quail medullary bone osteoclasts. Exp. Physiol. 79, 975–982.
21.
Xia, S. L., and Ferrier J. (1996). Localized calcium signaling in multinucleated osteoclasts. J. Cell Physiol. 167, 148 –155.
22.
Zheng, M. H., Wood, D. J., Papadimitriou, J. M., and Nicholson, G. C. (1992). Evidence that protein kinase-A, calcium-calmodulin kinase and cytoskeletal proteins are involved in osteoclast retraction induced by calcitonin. Exp. Mol. Pathol. 57, 105–115.
23.
Miyauchi, A., Alvarez, J. Greenfield, E. M., Teti, A., Grano, M., Colucci, S., Zambonin-Zallone, A., Ross, F. P., Teitelbaum, S. L., Cheresh, D., and Hruska, K. A. (1991). Recognition of osteopontin and related peptides by an avb 3 integrin stimulates immediate cell signals in osteoclasts. J. Biol. Chem. 266, 20369 – 20374.
24.
Paniccia, R., Riccioni, T., Zani, B. M., Zigrino, P., Scotlandi, K., and Teti, A. (1995). Calcitonin down-regulates immediate cell signals induced in human osteoclast-like cells by the bone sialoprotein-IIA fragment through a postintegrin receptor mechanism. Endocrinology 136, 1177–1186.
25.
Weidema, A. F., Ravesloot, J. H., Panyi, G., Nijweide, P. J., and Ypey, D. L. (1993). A Ca 21-dependent K 1-channel in freshly isolated and cultured chick osteoclasts. Biochim. Biophys. Acta 1149, 63–72.
26.
Yu, H., and Ferrier, J. (1993). ATP induces an intracellular calcium pulse in osteoclasts. Biochem. Biophys. Res. Commun. 191, 357–363.
27.
Xia, S. L., and Ferrier, J. (1995). Calcium signal induced by mechanical perturbation of osteoclasts. J. Cell Physiol. 163, 493–501.
28.
Blair, H. C., and Schlesinger, P. H. (1992). The mechanism of osteoclast acidification. in “The Biology and Physiology of the Osteoclast” (C. V. Gay and B. R. Rifkin, Eds.), pp. 259 –288. CRC Press, Boca Raton, FL.
29.
De Koninck, P., and Schulman, H. (1998). Sensitivity of CaM kinase II to the frequency of Ca 21 oscillations. Science 279, 227–230.
30.
Yao, Y., Choi, J., and Parker, I. (1995). Quantal puffs of intracellular Ca 21 evoked by inositol trisphosphate in Xenopus oocytes. J. Physiol. 482, 533–553.
31.
Parker, I., and Yao, Y. (1995). Calcium puffs in Xenopus oocytes. Ciba Found. Symp. 188, 50 – 65.
32.
Keizer, J., Smith, G. D., Ponce-Dawson, S., and Pearson, J. E. (1998). Saltatory propagation of Ca 21 waves by Ca 21 sparks. Biophys. J. 75, 595– 600.
33.
Lipp, P., Thomas, D., Berridge, M. J., and Bootman, M. D. (1997). Nuclear calcium signalling by individual cytoplasmic calcium puffs. EMBO J. 16, 7166 –7173.
Intracellular Calcium Puffs in Osteoclasts Wilson Radding,* S. Elizabeth Jordan,† Raymond B. Hester,‡ and Harry C. Blair† ,§ ,1 *Keck Center for Computational Biology, Rice University, Houston, Texas; †Veterans’ Affairs Medical Center, Birmingham, Alabama; ‡Department of Microbiology and Immunology and the Division of Biotechnical Services, University of South Alabama, Mobile, Alabama; and §Center for Metabolic Bone Diseases and Department of Pathology, University of Alabama School of Medicine, Birmingham, Alabama
We studied intracellular calcium ([Ca 21] i) in acidsecreting bone-attached osteoclasts, which produce a high-calcium acidic extracellular compartment. Acid secretion and [Ca 21] i were followed using H 1-restricted dyes and fura-2 or fluo-3. Whole cell calcium of acid-secreting osteoclasts was ;100 nM, similar to cells on inert substrate that do not secrete acid. However, measurements in restricted areas of the cell showed [Ca 21] i transients to 500 –1000 nM consistent with calcium puffs, transient (millisecond) localized calcium elevations reported in other cells. Spot measurements at 50-ms intervals indicated that puffs were typically less than 400 ms. Transients did not propagate in waves across the cell in scanning confocal measurements. Calcium puffs occurred mainly over regions of acid secretion as determined using lysotracker red DND99 and occurred at irregular periods averaging 5–15 s in acid secreting cells, but were rare in lysotracker-negative nonsecretory cells. The calmodulin antagonist trifluoperazine, cell-surface calcium transport inhibitors lanthanum or barium, and the endoplasmic reticulum ATPase inhibitor thapsigargin had variable acute effects on the mean [Ca 21] i and puff frequency. However, none of these agents prevented calcium puff activity, suggesting that the mechanism producing the puffs is independent of these processes. We conclude that [Ca 21] i transients in osteoclasts are increased in acid-secreting osteoclasts, and that the puffs occur mainly near the acid-transporting membrane. Cell membrane acid transport requires calcium, suggesting that calcium puffs function to maintain acid secretion. However, membrane H 1ATPase activity was insensitive to calcium in the 100 nM–1 mM range. Thus, any effects of calcium puffs on osteoclastic acid transport must be indirect. © 1999 Academic Press
Key Words: LaCl 3; trifluoperazine; BaCl 2; thapsigargin; V-type H 1-ATPase.
1 To whom correspondence and reprint requests should be addressed at 535 LHRB, 701 S. 19th Street, Birmingham, AL 352940007. Fax: (205) 934-9927. E-mail: [email protected].
INTRODUCTION
Osteoclasts secrete HCl onto the surface of bone, which is composed mainly of acid-soluble calcium salts, to maintain serum calcium and for periodical replacement of the skeleton. Osteoclasts use an expanded (ruffled) cell membrane that contains a massive concentration of vacuolar-type H 1-ATPase to produce this acid [1]. The acid is secreted into an extracellular compartment formed by the ring-shaped osteoclastic attachment, and consequently this space has a very high calcium activity [2]. Calcium is removed from this compartment by vacuolar transcytosis [3, 4]. There are high concentrations of calmodulin at the osteoclast ruffled border [5] consistent with calcium-modulated control processes. Several studies have suggested that osteoclasts respond to calcium, and intracellular calcium signals are generated in osteoclasts by a variety of stimuli. Osteoclast– bone attachment causes transient changes in intracellular calcium activity 2 ([Ca 21] i) [6], and similar calcium signals are evoked by integrin receptors [7]. On the other hand, some intracellular calcium signals do not depend on integrin binding [8]. Elevated extracellular calcium outside the usual bone– osteoclast attachment site alters the bone resorption rate inversely to [Ca 21] i [9] and causes cell retraction via ryanodinesensitive receptors [10]. However, attachment to bone is required for full osteoclast differentiation, including the membrane specialization for acid secretion at the bone attachment, and bone-attached osteoclasts are relatively insensitive to extracellular calcium [11]. This suggests that calcium-sensitive membrane proteins are modified or redistributed in bone-attached osteoclasts. Further, calcium transients have been observed in osteoclasts in the absence of extracellular calcium [12], calling into question whether extracellular calcium is important to the bone-resorbing osteoclast. Thus, it is unclear whether bone degradation affects 2 Fluorescent dye-binding measures calcium activity. For simplicity, dye-binding activity measurements are indicated as [Ca 21] i or as calcium concentration where the meaning is unaffected.
689
0014-4827/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
690
RADDING ET AL.
intracellular calcium or vice versa. We explored the hypothesis that osteoclastic intracellular calcium is increased during bone resorption, which would be expected to make the cell more sensitive to calciumactivation via membrane-associated calmodulin. We used the fluorescent weak base probe lysotracker red DND 99 to locate acid secreting osteoclasts attached to bone. Then, using fluorescent calcium probes, we compared intracellular calcium in these osteoclasts to calcium in osteoclasts attached to inert substrates where no acid is secreted. We were unable to distinguish the average intracellular calcium of acid-secreting and nonsecreting cells. However, transient calcium puffs, localized millisecond increases in calcium, were seen, which occurred much more frequently in acid secreting cells. These calcium transients occurred mainly near the site of acid secretion. METHODS Avian osteoclast cultures. Osteoclast-rich medullary bone from calcium starved chickens was harvested and osteoclast-rich preparations were made using bone affinity [13]. Cells on bone were cultured on glass coverslips in tissue culture plates to allow them to be used for fluorometry. Cells were cultured in Dubelco’s modified Eagle’s minimal essential medium with 10% heat inactivated serum at 37°C in 5% CO 2, except during fluorescence studies, which used a modified Krebs–Ringer’s solution with bicarbonate and glucose (NaCl 120 mM, KCl 5 mM, Ca 1.3 mM, phosphate 1.2 mM, MgSO 4 1.2 mM, Hepes 20 mM, glucose 2 mM, bicarbonate 22 mM). To minimize the possible effects of upward drift in pH of bicarbonate buffers exposed to air during fluorescence studies, pH was monitored and results shown were obtained with pH between 7.35 and 7.59. Identification of acid secreting cells. Acid lakes produced by osteoclasts were identified with the fluorescent polyamine–pyrrole dye lysotracker red DND-99 (lysotracker) (Molecular Probes, Eugene, OR). This dye diffuses through membranes at neutral pH since a major fraction of their phenazine/pyrrole rings and amine side groups are uncharged; the amines are protonated at low pH and accumulate in acid compartments because membranes are impermeable to ions. Lysotracker, rather than neutral red or acridine [14, 15], was used in calcium studies because its long-wavelength excitation and emission do not overlap fura-2 or fluo-3. However, neutral red, 25 mM, added 30 min prior to observations [14], was used to calibrate Lysotracker for use in demonstrating osteoclast acid secretion (Fig. 1B, right). Lysotracker, 2–3 mM, was added 3–5 min before observation using a 560-nm low-pass excitation filter and a 600-nm emission barrier filter. Intracellular calcium measurement. Fura-2 (1-[2-(5-carboxyoxazole)-6-amino-5-benzofuranoxy]-2-(2-amino-4-methylphenoxy)ethane-N,N,N9,N9-tetraacetic acid) or fluo-3 (1-[2-amino-5-(2,7dichloro-6-hydroxy-3-oxy-9-xanthenyl)phenoxy]-2-[2-amino-5methylphenoxy]ethane-N,N,N9,N9-tetraacetic acid) was used as their membrane-permeable acetoxymethyl esters (-AM) as described [6]. Fluors were from Molecular Probes. Cells were incubated in 10 mM fura-2-AM or fluo-3-AM for 60 –90 min, allowing cell esterases to release the membrane-impermeant free calcium-binding polycarboxylic acids. After washing to remove unreacted probe, fluorescence was measured. Fura-2 was used for dual excitation measurements in a defined microscopic field, and fluo-3 for single wavelength scanning measurements to indicate variability within a scanned area. For fura-2, measurements used excitation at 340 and 380 nm to differentiate Ca 21-bound and unbound dye using the Spex (Edison,
NJ) AR-CM-MAC dual excitation fluorescence on a Diaphot inverted epifluorescence microscope (Nikon, Melville, NY). Cell regions were isolated using an optical mask. Emission fluorescence was monitored by photomultiplier with computer recording. A typical signal was ;10 4 counts per 250 ms of a 3 3 3-mm region. Calcium activities were calculated by the ratio method of Grynkiewicz et al. [16], using a fura–2 K Ca of 224 nM. Fluo-3 measurements were single-wavelength, relative Ca 21 activities by dye fluorescence emission at 525 nM with excitation at 488 nm; dye emission intensity varies approximately fourfold between 40 nM and 3 mM calcium. These measurements were 30 pixel/line (900 point) scans at 3-s intervals on an ACAS 570 laser scanning confocal microscope (Meridian Instruments, Okemos, MI) as described [17]. Pixel step was 0.8 mm, with the scanned area 24 mm across. Fluo-3 was used to indicate calcium-dependent relative fluorescence intensity, shown using linear scaling. Permeabilization of Fluo-3 loaded cells was used to verify maximum dye response 630% across the fields shown to assure that relative differences reported substantially exceed background fluorescence variations in dye loading. Effect of calcium on membrane acid transport activity. Osteoclast membranes were obtained by nitrogen cavitation and sequential centrifugation [18]; membrane vesicles were reconstituted at 1 mg/ml protein by incubation for 30 min at 4°C in 120 mM KCl, 20 mM NaCl, 10 mM Hepes, pH 7.4 [18]. Calcium activity was adjusted using buffers with 0.5 mM EGTA and 2 to 400 mM calcium. Calcium activity was calculated using Ca z EGTA K D of 380 nM [16]. Mg 21ATP-dependent HCl transport was determined by acridine orange (3 mM) fluorescence quenching, recording dye emission at 540 nm with excitation at 468 nm [18] in a SLM/Aminco-Bowman (Urbana, IL) Series 2 luminescence spectrophotometer. Reactions were performed at 37°C with 15 mg vesicle protein in 2.5 ml of 400 mM ATP, 4 mM acridine orange in 120 mM KCl, 20 mM NaCl, 10 mM Hepes, pH 7.4; transport was initiated with 500 mM MgCl 2. Washout with 1 mM NH 4Cl was used to control for nonspecific changes in fluorescence.
RESULTS
Intracellular calcium activity in bone-attached and glass attached osteoclasts. Calibrated whole-cell Fura-2 measurements were used in these measurements; optical windows contained one isolated cell (;20-mm square fields), and 1-s integration was used. Measurements of [Ca 21] i in several acid-secreting osteoclasts on bone averaged ;100 nM. This was not significantly different from [Ca 21] i of osteoclasts off bone (Fig. 1A). Appearance of a typical field with lysotracker, showing that the bone-attached cells produced acid but the unattached cells did not is also shown (Fig. 1B). Because it had been hypothesized that calcium would increase in acid-secreting cells, a second series of calcium measurements was done in osteoclasts on and off bone, with similar results. Differences of 20 –30 nM would not be resolved due to experimental variability, but it was clear that any differences in the average intracellular calciums of osteoclasts on and off bone would be small. Whole-cell calcium studies on additional cell preparations, with 30 individual calibrated measurements, gave similar results. Instability of [Ca 21] i of acid secreting osteoclasts. Analysis of fura-2 measurements at 1-s intervals suggested that calcium of osteoclasts on bone was more
INTRACELLULAR CALCIUM PUFFS IN OSTEOCLASTS
highly variable than in cells on glass. Occasional spikes of ;500 nM were seen, which were too large to attribute to noise. Further, Xia and Ferrier [21] had reported osteoclast calcium fluctuations. To determine whether the variability in measurements was due to spatially limited or short-duration calcium events, high-resolution studies were performed. Fluorometric measurements were refined by recording at 250-ms integration intervals and by limiting the studied area to a 3 3 3-mm region of the osteoclast. These measurements showed transient micrometer-scale calcium pulse activity with duration ,1 s (calcium puffs) (Fig. 2A). These calcium puffs (or spikes) to ;500 nM occurred at irregular intervals, in Fig. 2A averaging ;16 s. The frequency of puffs varied between experiments. This may reflect regional differences in the cells studied or variation between cells. Fourier analysis of traces was performed to determine whether there was an underlying periodicity, but none was found. In scanning confocal measurements, the calcium puffs were seen as ;2-mm regions of elevated calcium, which varied in location between frames at 3-s intervals (Fig. 2B). Comparison with lysotracker analysis indicated that the calcium puffs correlated very well with the region of acid secretion. Similar results were obtained when bone-attached and basolateral regions of a cell were studied using calibrated measurements in spots ;2 mm square (Fig. 2C). The nuclear area and membrane away from bone showed little activity, and was similar to that seen in glass-attached cells (not illustrated) while membrane adjacent to the acidified bone attachment had much higher activity. Recordings us-
691
ing minimum instrument integration intervals of 50 ms showed that the duration of the puffs was typically 200 – 400 ms (Fig. 2C, inset). Effects of calcium transport inhibitors. The effect of the calmodulin antagonist trifluoperazine, 10 mM, to inhibit the Ca 21-ATPase [19], which caused a slow increase in the average intracellular calcium, is shown in Figs. 3A and 3B. Both glass-attached and acid-secreting cells showed increasing average intracellular calcium after addition of trifluoperazine. However, there was much more variation in the traces obtained from cells on bone. The effects of 30 mM thapsigargin, an inhibitor of endoplasmic reticulum ATPases (Figs. 3C– 3D), and of blocking cell-surface calcium channels with 100 mM lanthanum (Fig. 3E) or barium (Fig. 3F) were tested. None of the treatments abolished the calcium puffs, and indeed the number and size of calcium puffs typically increased over several minutes, although these late effects may reflect the toxicity of these agents. These results are selected frames from 50 –90 sequential 3-s time points obtained in each study. Each agent was studied in three or more time-course measurement series, with similar results in each study. These findings suggest that the calcium transients do not depend wholly on intracellular calcium stores or on lanthanum-sensitive cell surface calcium channels; these agents may not affect membranes at the acidification site, however (see Discussion). Effect of calcium activity on osteoclast membrane ATP-dependent HCl transport. Osteoclast acid secretion is known to be abolished by calmodulin antago-
FIG. 1. Intracellular calcium and acid secretion in osteoclasts on and off bone. (A) Intracellular calcium in individually calibrated fura-2 whole-cell measurements, eight osteoclasts on bone and six osteoclasts off bone. Mean 6 standard deviation. The difference is not significant. (B) Localization of acid transporting cells. Lysotracker (left) was used to determine acid-transporting cells in further studies. Acid lakes are indicated; these are 10- to 20-mm diameter regions on the bone surface. They appear oval when seen face-on and linear when seen edge-on. Faint granular staining represents lysosomal reaction or nonspecific fluorescence from bone particles. Transmitted light (right) shows another coverslip from the same preparation, stained with neutral red [14] to visualize the acid lakes. Note that glass-attached cells do not form acid lakes and are hence not seen in the fluorescent assay. Cells bound to the edge of the bone allow the acid-secreting region to be distinguished from other cell regions by point measurements (see Fig. 2C). In either case, inclusion of 1 mM ammonium chloride, a nonfluorescent weak base that also accumulates in acidic compartments, blunted dye uptake to essentially undetectable levels (not illustrated). FIG. 2. Intracellular calcium in bone-attached osteoclasts. Acid secreting cells were located using lysotracker as in Fig. 1B (left). (A) Fluorometric measurements with 250-ms integration in a 3 3 3-mm acid-secreting region. Transient calcium pulse activity to ;500 –1000 mM, with duration ,1 s is seen (calcium puffs), two of which are indicated by arrows. In these high-resolution measurements, fluctuations in the 50 –200 nM range are indistinguishable from photon-counting noise. The spikes indicated represent consistent differences of a much larger magnitude (see also (C), below, which resolves the calcium puffs into multiple measurements showing elevated calcium using 50-ms measurements). A difficulty in these measurements is that visual determination of cellular regions is subject to observer interpretation of phase images; however, similar results were obtained using confocal measurements over the whole cell or when subcellular regions were selected (B and C, following). (B) Nine successive scanning confocal measurements of fluo-3 activity in an acid-secreting osteoclast, at 3-s intervals. These are false-color illustrations, which convert the photon counts of the green fluorescence of the fluo-3 into intensity readings: white . yellow . green . light blue . dark blue. Black, no signal. The arrows are examples of calcium elevations that move from frame to frame (many other examples are also present). The calcium puffs are ;2-mm regions of elevated calcium that vary in location between frames, but correlate with the region of acid secretion, which is outlined in frame 1. (C) Similar results were obtained when bone-attached and basolateral regions of a cell were studied using calibrated point measurements. The nuclear area and membrane away from bone (similar to regions 2 and 1 in the cell outlined in Fig. 1B, right) showed little activity, while membrane adjacent to bone had much higher activity (a region similar to region 3 in the cell outlined in Fig. 1B, right). A 50-ms trace of a similar region shows the duration of the puffs, 200 – 400 ms (inset).
692
RADDING ET AL.
INTRACELLULAR CALCIUM PUFFS IN OSTEOCLASTS
693
FIG. 3. Effect of calcium transport inhibitors. In each study shown, the acid-secreting region and cells were localized using lysotracker dye, as in Fig. 1, and phase microscopy. The effect of the calmodulin antagonist trifluoperazine (10 mM), an inhibitor of endoplasmic reticulum ATPases, thapsigargin (30 mM), and cell-surface calcium channel blocking cations, lanthanum or barium (100 or 50 mM) on calcium puffs in acid-transporting regions are shown using calibrated point measurements (A, C, and E) and scanning confocal measurement of relative calcium activity (B, D, and F). None of the treatments abolished the calcium puffs, and the number and size of calcium puffs tended to increase over several minutes after the calcium transport antagonists were added. Representative traces using acid-secreting cells with basal calcium of ;100 nM are shown, although calcium activity varied severalfold between cells, in keeping with variability noted within and between cells (Figs. 2B and 2C). Here, this effect is most apparent after lanthanum addition, although in some traces thapsigargin markedly increased activity. Trifluoperazine caused an upward drift in calcium, since this paralyzes the key calcium ATPase; this effect typically required longer times, 5–10 min, and was also associated with increased in calcium transients. In the scanning measurements, the antagonists were added after the third frame shown. In the first frame, the position of the cell and its acid secreting site, determined as in Fig. 1B, is indicated. The fluo-3 loading of the cells shown varied resulting in different basal fluorescence return. However, as with the calibrated spot measurements, the treatments did not reduce, and typically increased, calcium puff activity.
694
RADDING ET AL.
FIG. 4. ATP-dependent acid transport by osteoclast membranes in buffers with 1 nM–10 mM free calcium. Acid transport by 15 mg of osteoclast cell-membrane vesicles was measured as 4 mM acridine uptake over 200 s in 400 mM ATP and 500 mM MgCl 2. Activity shown is arbitrary fluorescence units at 540 nm, determined by washout of acridine uptake with 3 mM NH 4Cl. Buffers included 500 mM EGTA except at the lowest calcium, containing 1 mM EGTA. One of two experiments with similar results is shown.
nists. Further, osteoclast acid secretion produces a high calcium concentration, and calcium puffs were observed mainly at the acid secretion site. Thus we studied the direct effects of calcium on membrane HCl transport. Acid transport activity was abolished by incubation of membranes in 1 mM EGTA, 400 mM ATP, 120 mM KCl, 20 mM NaCl, 10 mM Hepes, pH 7.4, 37°C, for even a few minutes. However, in buffers with calcium activity greater than 2 nM, there was no apparent variation in H 1-ATPase activity (Fig. 4). This indicates that calcium puffs appear unlikely to modify the H 1-ATPase directly. This experiment was repeated with three different osteoclast membrane vesicle preparations with essentially identical results. DISCUSSION
Calcium is a modulator of osteoclastic activity, although how calcium varies during physiological bone resorption is uncertain. Zheng et al. [22] used inhibitors to show that calcium-calmodulin kinase affects osteoclast retraction. The osteoclastic Ca 21 ATPase has a high ATP-affinity component that is inhibited by the calmodulin antagonist trifluoperazine [19]. Miyauchi et al. [23], studying [Ca 21] i decrement with a vb 3 integrin binding, attributed reduced [Ca 21] i to this pump; the calmodulin antagonist blocked Ca 21 efflux. In plastic attached osteoclasts [Ca 21] i variation has been related to the postintegrin receptor mechanisms [24]. However, intracellular calcium signals are complex and respond to signals independent from integrin binding [8]. Osteoclast membranes also contain an outwardly rectifying voltage-activated K 1 channel that is calcium
dependent [25], and there are reports of purinergic [26] and mechanical [27] calcium responses in osteoclasts. Our results are in keeping with these previous studies and further suggest that there are regional differences in osteoclast calcium activity (Fig. 2), although these differences are transient rather than continuous. Importantly, the results in this study were obtained using bone-attached osteoclasts; the properties of osteoclasts, including acid secretion, may vary after the cells are disaggregated from bone and allowed to reattach to tissue-culture plastic or glass (Fig. 1B). While macrophages and foreign body giant cells may have attachment sites that are to varying degrees acidic, such as seen in microprobe studies (reviewed in [28]), the bone-attachment site has a much greater accumulation of the HCl secreting pumps [28], and this attachment is critical to maintenance of the extremely active acid secretion apparatus of the osteoclast [1]. Osteoclast intracellular calcium has been measured under various conditions , but only recently have data been produced in bone attached cells. On plastic, osteoclast intracellular [Ca 21] varies with extracellular [Ca 21] and causes cell retraction via ryanodine-sensitive receptors [10]. However, Lakkakorpi et al. [11] showed that resorbing osteoclasts are much less sensitive to extracellular calcium, suggesting that boneattached cells have membrane calcium response mechanisms that may vary from plastic-attached cells. Our results are in accord with this, and suggest that there are regional differences in osteoclast calcium activity. We used acid-specific dyes to define osteoclasts that were secreting acid (Fig. 1B), and observed near the acidic, high-calcium bone-attachment transient micrometer-scale, 200- to 400-ms-duration calcium pulse activity (calcium puffs) (Figs. 2 and 3). However, we did not observe calcium waves. Calcium pulses, fluctuations, and repetitive spikes have been reported in rabbit osteoclasts [21], although the resolution of the measurements reported here suggest more finely localized, short-duration events. Further, although activity was observed in osteoclasts off bone, calcium puffs were much more frequently observed in osteoclasts on bone, and particularly in association with the acidic bone resorption site. However, it must be emphasized that the colocalization would not resolve calcium puffs originating near, but not in, the ruffled border, such as activity associated with vesicles or the cytoskeleton near the acid-secreting membrane. Acid secretion-related local calcium activity suggested a positive feedback stimulus for acid secretion, but osteoclast membrane vesicle HCl transport was not directly modulated by variation in calcium activity (Fig. 4). Thus, if the role of calcium puffs at the acid secretion site functions in a feedback mechanism, this is unlikely to act directly on the ruffled membrane HCl secreting apparatus. Indeed, the sensitivity of the
INTRACELLULAR CALCIUM PUFFS IN OSTEOCLASTS
membrane at very low calcium may reflect changes in the membrane that are not physiological, and the calcium-responsive elements may be largely lost during membrane preparation at intermediate calcium activity levels. Recent reports do relate osteoclastic intracellular calcium to bone resorption, however. Ritchie et al. [9] showed that calcium channel-blockers reduced osteoclastic bone resorption. General inhibitors of calcium or calmodulin also reduce osteoclastic bone degradation [5]. Thus, it is clear that calcium modifies bone resorption, but the mechanism is not a simple feedback effect at the level of the acid-secreting membrane. It is possible that transient increases in calcium near the high-calcium bone-resorption compartment reflect “leaks” in the ruffled border. The osteoclast’s secretory membrane has a large surface area and separates regions of very high and low calcium activity. However, observations in other cells suggest that, whatever the mechanism producing the calcium puffs, they are likely to have physiologically meaningful regulatory effects. In this regard, calmodulin kinase II activity changes with the frequency of calcium elevations [29]. Since osteoclast acid secretion activity is maintained by calmodulin [5], but there is no direct calcium sensitivity of membrane HCl transport, this mechanism is a likely mediator of whole-cell effects. Calcium puffs have been described in other cell types. Puffs of calcium of 300 –500 ms have been described in Xenopus oocytes [30]. Parker and Yao [31] summarize data on calcium puffs in Xenopus as consistent with concerted opening of several inositol trisphosphate gated calcium channels. The spatial size (2–3 mm) of calcium puffs in Xenopus is consistent with the calcium puffs observed in osteoclasts. In myocytes, similar calcium puffs occur, with frequency dependent on extracellular calcium. Our data are consistent generally with the fire– diffuse–fire model of propagation of calcium sparks in myocytes [32]. However, in this case the calcium spark sites are either not distributed in a simple array, or the array is arranged on a convoluted membrane structure. Nevertheless, the relation of puffs to proximity to the high calcium site suggest similar roles of osteoclastic calcium puffs and calcium puffs in muscle. Other observations have included calcium puffs originating mostly within 2–3 mm of nuclei of HeLa cells [33]. We cannot rule out an association of perinuclear membranes with calcium elevations in osteoclasts, but a major association in osteoclasts with the acid secretion site is much more likely from our observations; this difference probably reflects cellular specialization. In osteoclasts treated with the calmodulin antagonist trifluoperazine calcium increased as expected. This may reflect effects on the calcium ATPase or other calmodulin-sensitive proteins. Calcium transients per-
695
sisted and increased in frequency, indicating that the mechanism producing the calcium puffs is not calmodulin dependent. Blocking cell surface calcium channels with lanthanum or barium did not abolish calcium puffs, although it is unlikely that the cations would affect intracellular membranes or the isolated osteoclast-bone attachment. The specificities of La and Ba are not precisely the same, but both ions block major calcium channels, and between the two, if cell-surface calcium channels were important for the phenomenon observed, some effect would have been expected, but was not observed. Calcium puffs in osteoclasts were not abolished by thapsigargin, but these short-term experiments may have been inadequate to deplete calcium stores, so an endoplasmic reticulum-dependent mechanism is not excluded. Our results indicate that intermittent localized puffs of calcium occur in osteoclasts. These were seen predominately at or near the acid-secreting membrane. Osteoclast internal calcium was maintained robustly in the face of inhibitors, suggesting that the osteoclast expresses significant amounts of multiple plasma membrane and endoplasmic reticulum pumps, which is consistent with its role as a calcium-transporting cell [3]. We speculate that the calcium spikes or puffs observed may be related to calcium entering the cell from the adjacent high-calcium acid degradation site and that the calcium activity near the acid transport site may be used as a feedback signal to continue acid secretion. However, osteoclast membrane HCl secretion was not directly regulated by calcium, so if calcium puffs function in positive feedback for acid secretion, the mechanism is indirect. We thank Drs. John P. Williams and Robert W. Hardy, University of Alabama, Birmingham, for valuable discussions and assistance with Spex measurements. Supported in part by National Institutes of Health Grants AG12951 and AR43225, the University of Alabama Health Services Foundation, the Department of Veterans’ Affairs, and National Library of Medicine Grant LM07093.
REFERENCES 1.
Schlesinger, P. H., Mattsson, J. P., and Blair, H. C. (1994). Osteoclastic acid transport: mechanism and implications for physiological and pharmacological regulation. Miner. Electrolyte Metab. 20, 31–39.
2.
Silver, A., Murrills, R. J., and Etherington, D. J. (1988). Microelectrode studies on the acid microenvironment beneath adherent macrophages and osteoclasts. Exp. Cell Res. 175, 266 –276.
3.
Salo, J., Lehenkari, P., Metsikko, K., and Va¨a¨na¨nen, H. K. (1997). Removal of osteoclast bone resorption products by transcytosis. Science 276, 270 –273.
4.
Nesbitt, S. A., and Horton, M. A. (1997). Trafficking of matrix collagens through bone-resorbing osteoclasts. Science 276, 266 – 269.
5.
Radding, W., Williams, J. P., Hardy, R. W., McDonald, J. M., Whitaker, C. H., Turbat-Herrera, E. A., and Blair, H. C. (1994).
696
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
RADDING ET AL. Calmodulin concentrated at the osteoclast ruffled border modulates acid secretion. J. Cell Physiol. 160, 17–28. Teti, A., Blair, H. C., Schlesinger, P. Grano, M., ZamboninZallone, A., Kahn, A., Teitelbaum, S. L., and Hruska, K. A. (1989). Extracellular protons acidify osteoclasts, reduce cytosolic calcium and promote expression of cell-matrix attachment structures. J. Clin. Invest. 84, 773–780. Shankar, G., Davison, I., Helfrich, M. H., Mason, W. T., and Horton, M. A. (1993). Integrin receptor-mediated mobilisation of intranuclear calcium in rat osteoclasts. J. Cell Sci. 105, 61– 68. Shankar, G., Gadek, T. R., Burdick, D. J., Davison, I., Mason, W. T., and Horton, M. A. (1995). Structural determinants of calcium signaling by RGD peptides in rat osteoclasts: Integrindependent and -independent actions. Exp. Cell Res. 219, 364 – 371. Ritchie, C. K., Maercklein, P. B., and Fitzpatrick, L. A. (1994). Direct effect of calcium channel antagonists on osteoclast function: Alterations in bone resorption and intracellular calcium concentrations. Endocrinology 135, 996 –1003. Zaidi, M., Shankar, V. S., Tunwell, R., Adebanjo, O. A., Mackrill, J., Pazianas, M., O’Connell, D., Simon, B. J., Rifkin, B. R., Venkitaraman, A. R., Huang, C.L.-H., and Lai, F. A. (1995). A ryanodine receptor-like molecule expressed in the osteoclast plasma membrane functions in extracellular Ca 21 sensing. J. Clin. Invest. 96, 1582–1590. Lakkakorpi, P. T., Lehenkari, P. P., Rautiala, T. J., and Va¨a¨na¨nen, H. K. (1996). Different calcium sensitivity in osteoclasts on glass and on bone and maintenance of cytoskeletal structures on bone in the presence of high extracellular calcium. J. Cell Physiol. 168, 668 – 677. Shankar, V. S., Huang, C. L., Adebanjo, O. A., Pazianas, M., and Zaidi, M. (1994). Calcium influx and release in isolated rat osteoclasts. Exp. Physiol. 79, 537–545. Williams, J. P., Blair, H. C., McKenna, M. A., Jordan, S. E., and McDonald, J. M. (1996). Regulation of avian osteoclastic H 1ATPase and bone resorption by tamoxifen and calmodulin antagonists. J. Biol. Chem. 271, 12488 –12495. Barnicot, N. A. (1947). The supravital staining of osteoclasts with neutral-red: Their distribution on the parietal bone of normal growing mice, and a comparison with the mutants greylethal and hydrocephalus-3. Proc. R. Soc. B 134, 467– 485. Baron, R., Neff, L., Louvard, D., and Courtoy, P. J. (1985). Cell-mediated extracellular acidification and bone resorption: Evidence for a low pH in the resorbing lacunae and localization of a 100-kD lysosomal membrane protein at the osteoclast ruffled border. J. Cell Biol. 101, 2210 –2222. Grynkiewicz, G., Poenie, M., and Tsien, R. Y. (1985). A new generation of Ca 12 indicators with greatly improved fluorescence properties. J. Biol. Chem. 260, 3440 –3450. Hardy, R. W., Ladenson, J. H., Hruska, K. A., Jiwa, A. H., and McDonald, J. M. (1992). The effects of extracellular calcium and epinephrine on cytosolic-free calcium in single rat adipocytes. Endocrinology 130, 3694 –3702. Blair, H. C., Teitelbaum, S. L., Koziol, C. M., and Schlesinger, P. H. (1991). Passive chloride permeability charge-coupled to
Received July 20, 1999 Revised version received September 30, 1999
the electrogenic H 1-ATPase of avian osteoclast ruffled membrane. Am. J. Phys. 260, C1315–C1324. 19.
Bekker, P. J., and Gay, C. V. (1990). Characterization of a Ca 21-ATPase in osteoclast plasma membrane. J. Bone Miner. Res. 5, 557–567.
20.
Arkle, S., Wormstone, I. M., Bascal, Z. A., and Dacke, C. G. (1994). Estimation of intracellular calcium activity in confluent monolayers of primary cultures of quail medullary bone osteoclasts. Exp. Physiol. 79, 975–982.
21.
Xia, S. L., and Ferrier J. (1996). Localized calcium signaling in multinucleated osteoclasts. J. Cell Physiol. 167, 148 –155.
22.
Zheng, M. H., Wood, D. J., Papadimitriou, J. M., and Nicholson, G. C. (1992). Evidence that protein kinase-A, calcium-calmodulin kinase and cytoskeletal proteins are involved in osteoclast retraction induced by calcitonin. Exp. Mol. Pathol. 57, 105–115.
23.
Miyauchi, A., Alvarez, J. Greenfield, E. M., Teti, A., Grano, M., Colucci, S., Zambonin-Zallone, A., Ross, F. P., Teitelbaum, S. L., Cheresh, D., and Hruska, K. A. (1991). Recognition of osteopontin and related peptides by an avb 3 integrin stimulates immediate cell signals in osteoclasts. J. Biol. Chem. 266, 20369 – 20374.
24.
Paniccia, R., Riccioni, T., Zani, B. M., Zigrino, P., Scotlandi, K., and Teti, A. (1995). Calcitonin down-regulates immediate cell signals induced in human osteoclast-like cells by the bone sialoprotein-IIA fragment through a postintegrin receptor mechanism. Endocrinology 136, 1177–1186.
25.
Weidema, A. F., Ravesloot, J. H., Panyi, G., Nijweide, P. J., and Ypey, D. L. (1993). A Ca 21-dependent K 1-channel in freshly isolated and cultured chick osteoclasts. Biochim. Biophys. Acta 1149, 63–72.
26.
Yu, H., and Ferrier, J. (1993). ATP induces an intracellular calcium pulse in osteoclasts. Biochem. Biophys. Res. Commun. 191, 357–363.
27.
Xia, S. L., and Ferrier, J. (1995). Calcium signal induced by mechanical perturbation of osteoclasts. J. Cell Physiol. 163, 493–501.
28.
Blair, H. C., and Schlesinger, P. H. (1992). The mechanism of osteoclast acidification. in “The Biology and Physiology of the Osteoclast” (C. V. Gay and B. R. Rifkin, Eds.), pp. 259 –288. CRC Press, Boca Raton, FL.
29.
De Koninck, P., and Schulman, H. (1998). Sensitivity of CaM kinase II to the frequency of Ca 21 oscillations. Science 279, 227–230.
30.
Yao, Y., Choi, J., and Parker, I. (1995). Quantal puffs of intracellular Ca 21 evoked by inositol trisphosphate in Xenopus oocytes. J. Physiol. 482, 533–553.
31.
Parker, I., and Yao, Y. (1995). Calcium puffs in Xenopus oocytes. Ciba Found. Symp. 188, 50 – 65.
32.
Keizer, J., Smith, G. D., Ponce-Dawson, S., and Pearson, J. E. (1998). Saltatory propagation of Ca 21 waves by Ca 21 sparks. Biophys. J. 75, 595– 600.
33.
Lipp, P., Thomas, D., Berridge, M. J., and Bootman, M. D. (1997). Nuclear calcium signalling by individual cytoplasmic calcium puffs. EMBO J. 16, 7166 –7173.