JOURNAL OF BONE AND MINERAL RESEARCH Volume 7, Supplement 2, 1992 Mary Ann Liebert, Inc., Publishers

Voltage, Calcium, and Stretch Activated Ionic Channels and Intracellular Calcium in Bone Cells DIRK L. YPEY,' ADAM F. WIDEMA,',' KARIN M. HOLD,3 ARNOUD VAN DER LAARSE,3 JAN H. RAVESLOOT,',' ARIE VAN DER PLAS,' and PETER J. NIJWEIDE2

ABSTRACT Embryonic chick bone cells express various types of ionic channels in their plasma membranes for as yet unresolved functions. Chick osteoclasts (OCL) have the richest spectrum of channel types. Specific for OCL is a K*channel, which activates (opens) when the inside negative membrane potential (V,) becomes more negative (hyperpolarization). This is consistent with findings of others on rat OCL. The membrane conductance constituted by these channels is called the inward rectifying K*conductance (&I), or inward rectifier, because the hyperpolarization-activatedchannels cause cell-inward K*current to pass more easily through the membrane than outward K' current. Besides &i channels, OCL may express two other types of voltage-activated K*channels. One constitutes the transient outward rectifying K*conductance (&& which is activated upon making the membrane potential less negative (depolarization) but has a transient nature. This conductance favors transient K' conduction in the cell-outward direction. The &to also occurs in a small percentage of cells in osteoblast (OBL) and periosteal fibroblast (PFB) cultures. The other OCL K*conductance, the &a, is activated by both membrane depolarization and a rise in [Ca'*]i. &cP channels are also present in the other chick bone cell types, that is, OBL, osteocytes (OCY), and PFB. Furthermore, in excised patches of all bone cell types, channels have been found that conduct anions, including CI- and phosphate ions. These channels are only active around V, = 0 mV. While searching for a membrane mechanism for adaptation of bone to mechanical loading, we found stretch-activatedchannels in chick osteoclasts; other investigators have found stretch-activated cation channels (K*or aselective) in rat and human osteogenic cell lines. In contrast to other studies on cell lines or OBL from other species, we have not found any of the classic macroscopic voltage-activated calcium conductances (Gcr) in any of the chick bone cells under our experimental conditions. However, our fluorescence measurements of [Ca'*]i in single cells indicate the presence of CaZ+conductive pathways through the plasma membrane of osteoblastic cells and osteoclasts, consistent with other studies. We discuss possible roles for &I, &a, and anion channels in acid secretion by OCL and for stretch-activated channels in OCL locomotion.

INTRODUCTION are tunnel-shaped transmembrane proteins, serving as more or less selective conduction pathways for ions across the plasma membrane and across membranes surrounding intracellular organelles.(') Opening and closure of these conduction pathways may be controlled by the transmembrane potential, by the concentra-

I

ONIC CHANNELS

tion of extra- and intracellular substances, and by mechanical stress of the membrane."' Examples of channel-activating substances are neurotransmitters, hormones, autocrine substances, and second messengers.(3) Changes in the number of open channels of a certain type may have two consequences for the cell. First, the membrane potential of the cell may change. This may influence the flow of other ions through other channels as

'Department of Physiology, Leiden University, The Netherlands. 'Department of Cell Biology and Histology, Leiden University, The Netherlands. 'Department of Cardiology, University Hospital, Leiden, The Netherlands.

s377

YPEY ET AL.

!a78

well, since the membrane potential change may cause opening or closure of other channels and alters the driving force for the ions moving through the channels that are already open. Thus, passive transmembrane transport of a given type of ion is principally coupled to transport of other types of ions through changes in the membrane potential, which we find of importance to understand the role of ionic channels in bone cell functions. Second, the intracellular concentration of ions may change, which in turn may provide a signal to the cell to change its functional activity. For example, an influx of Ca2+ions into the cell may control protein secretion,(4)fluid and salt secretion,(s)cell movement,(6)or other metabolic processes. In recent years investigatorshave begun to study the role of ionic channels in bone cell function^(^.'-^^) using a powerful membrane electrophysiologic technique, known as the patch-clamp technique. (l3.l4) This technique, for which Neher and Sakmann recently received the Nobel prize in Medicine and Physiology, allows the measurement of the microscopic, picoampere (l0-l2 A) currents through single ionic channel proteins in small patches of membrane but can also be used to measure the macroscopic currents through or potentials across the plasma membrane of the whole cell. The technique allowed the discovery of various types of cation and anion channels in bone cells. Most channels are voltage activated‘l5)but some are sensitive to membrane stretch. (16) These results open the possibility that biophysical, that is, electrical and mechanical, membrane properties are important determinants in the control of bone cell functions, such as bone formation by osteoblasts, bone resorption by osteoclasts, autocrine communication between osteoblasts and osteoclasts, and intercellular communication between osteocytes and osteoblasts through gap junctions.‘”) The stretch sensitivity of ionic channel activity may even imply a mechanism for bone adaptation to changes in mechanical loading. Another way to study the role of ionic membrane currents in bone cell functions is to measure the resulting intracellular ion concentration changes using fluorescence techniques. This technique has been greatly improved in recent years(’9)and can even be used to measure ion concentration changes in (parts of) single cells (microfluorescence), thus completing patch-clamp measurements on single cells. It is the purpose of the present paper to review the patch-clamp studies on chick bone cells from our group in the light of studies of other groups on other species, to explain the methodology used (the patch-clamp and microfluorescence technique), to add some new comlementary results, and to integrate the data into a functional context. The new results presented include stretch-activated ionic channels in osteoclasts and [Caf+Jimeasurements with fluorescence imaging techniques relevant for the control of [Ca2+],-activatedK+channels and many other cellular functions.

METHODS AND METHODOLOGY Isolation and culture of cells Osteoclasts (OCL) from 18 day chick embryonic long bones and osteoblasts (0BL)-osteocytes ( 0 0 , periosteal

fibroblasts (PFB) from 18 day chick embryonic calvariae were isolated and cultured on glass coverslips as described In some cases pure OCY cultures were used.‘”) OCL were usually taken for experiments a few hours after isolation. OBL/OCY were used after 1-3 days of culture, OCY could be recognized from their slender branched processes, contacting the processes of other OCY, as in vivo.

Patch-clamp technique Details of the technique we used have been given before,(15*20) but for convenience we explain here the methodology, based on Hamill et al.(13)and using the results reviewed in Figs. 1 and 2. The actual measuring probe of the patch-clamp technique is the patch pipette electrode. This is a glass micropipette, drawn from a 1-2 mm glass capillary to an open tip diameter of approximately 1 pn. The pipette is filled with a physiologic salt solution in which an Ag/AgCl wire is inserted to connect the ion-conducting solution inside the pipette to the input of a current or voltage amplifier. To obtain a measurement circuit, a reference Ag/AgCl electrode in the physiologic salt solution bathing the cells is connected to the other input of the amplifier (see insets, Fig. 1). Using a remote-controlled manipulator, the patch pipette can be placed on the membrane of a cell under the microscope. Upon touching the membrane, slight suction is applied to the inside of the pipette to obtain firm attachment and sealing of the membrane to the mouth of the pipette. Once the seal is perfect, as decided from the high value of the seal resistance R, (> 1 GQ 1 Gfl = lo9 fl), measured between the inside of the pipette and the bath, picoampere ionic currents through ionic channels in the patch can be measured while clamping the membrane patch at different voltage displacements from the resting membrane potential with the voltage source in the patchclamp amplifier. The measurement configuration obtained in this way is called the cell-attached patch configuration (CAP; see Fig. 1A). It allows the recording of single-channel activity from a patch while the intact cell is still attached to the patch. Therefore, this configuration is most useful to study the functional role of identified channels under physiologic conditions. By rapidly pulling the sealed patch pipette from the cell, the sealed patch can be “excised” from the cell and becomes an inside-out patch (IOP; see Fig. lC), with the inside face of the membrane exposed to the bathing fluid. This configuration allows well-controlled studies of the effects of cytoplasmic factors on channel activity. Another widely used measurement configuration is the whole-cell (WC) configuration (see Fig. lB), which is obtained by breaking a cell-attached patch by applying an extra suction pulse to the pipette. In the WC the cytoplasrna of the cell is perfused by the pipette solution, which allows study of the collective behavior of all the ionic channels in the plasma membrane under defined intracellular conditions. The perfusion of the whole cell with an artificial solution may be a disadvantage if the ionic channels studied require diffusable intracellular factors. However, this disadvantage can be overcome by using a CAP configuration, the perforated patch, in which the CAP has been strongly

5379

IONIC CHANNELS AND INTRACELLULAR CALCIUM IN BONE CELLS

0

A

-50

C

so

0

lbo

0

FIG.1. Calcium-dependent K' channels and conductance in osteoblasts, measured in the various patch-clamp configurations. (A, B, and D)The bathing solution was a high-sodium extracellular solution (SECS); the pipette was filled with a high-potassium intracellular-like solution (ICS). (A) Current-voltage plot of cell-attached patch channel activity during application of a voltage ramp (6 superimposed sweeps). ICS with pCa = 5.1. (B) Repeated whole-cell recording from the same cell at two different intracellularpCa (indicated), applied by using two successive patch pipettes filled with different calcium-buffered ICS. (C)Increased inside-out patch channel activity upon increasing the Ca'* concentration at the original intracellular face of the membrane patch by changing theof the bathing solution (ICS) from 7 to 3.2. The pipette contained SECS to establish normal Na' and K' concentration asymmetry across the patch membrane. The applied pipette potential was 0 mV. (D)Outside-out patch channel activity, recorded as a current-voltage plot from 17 superimposed sweeps. ICS in pipette with pCa = 5.1. (Reproduced and adapted with permission from Ref. 20, Ravesloot et al., 1990).

permeabilized by incorporation of nystatin, an antibiotic forming monovalent cation channels, into the patch membrane. The fourth frequently used patch-clamp configuration is that of the outside-out patch (OOP; see Fig. 1D).which is obtained by pulling a vesicle from the WC configuration. The OOP is actually a micro-WC and allows the resolution of the single-channel behavior underlying the WC conductance behavior. Since channel opening and closing are reflected by membrane conductance changes and because these openings and closings are usually membrane potential controlled, channel activity is usually measured at different voltages, applied as voltage steps (Fig. 2) or as voltage ramps (see Fig. 1A). By convention, cell outward-directed ionic current, that is, outward cation or inward

anion current, is plotted as positive (upward) current (Figs. 1, 2, and 3). In the present patch-clamp experiments, carried out at room temperature, the cells were usually bathed in a standard extracellular salt solution (SECS) consisting of (mM) 150 NaCl, 5 KCl, 1 CaCll, 1 MgCll, and 10 HEPEW NaOH @H 7.2). Intracellular-likeionic conditions, for example during WC or OOP experiments, were obtained by filling the pipettes with an intracellular salt solution (ICS) consisting of (mM) 8-10 NaCl, 140-145 KCl, 1 MgCl,, 0.024 CaCL, 0.01-10 EGTA to obtain a p C a in the range of 3.2-8, and 10 HEPESIKOH (PH 7.2). CAP pipettes were filled with either SECS or ICS. Other details of measurement and data analysis have already been described. (ls*lo)

YPEY ET AL.

h

.A

,100 mV

-70 FIG.2. Voltage-activated whole-cell conductances in a chick embryonic osteoclast (OCL), consistent with Ravesloot et al.(18)The current records (I,) are responses to hyperpolarizing and depolarizing voltage steps from a holding potential V, = -70 mV to the voltages (V,) indicated. Hyperpolarizing steps activate GKi, depolarizing steps to > -20 mV activate G K ~Steps ~ . to >40 mV activate G K ~Successive . steps were applied with intervals of 5.5 s, starting at -160 mV and increasing with 20 mV increments. The bath contained SECS and the pipette ICS with a pCa = 7.8.

FIG. 3. Stretch-activated CAP channels in an embryonic chick osteoclast. (A) Current records before (left), during -45 cm H,O pressure (middle), and after return to zero pressure (right). The CAP was 100 mV hyperpolarized to increase the channel current amplitude. Increased channel activity is sustained during underpressure, although time-dependent changes may occur. (B). Current-voltage relationship of the stretch-activated channel at -40 cm H,O pressure. Both the bath and the pipette contained SECS.

s381

IONIC CHANNELS AND INTRACELLULAR CALCIUM IN BONE CELLS

Microfluorescence measurements of [Caz+Ji on single cells The concentration of free intracellular calcium ions [Ca1+liwas measured at 37°C on individual cells loaded with the fluorescent calcium indicator fura-2, using a fluorescence microscope and calcium imaging equipment. Fura-2 is a dual-excitation fluorochrome; that is, it is excited at two successive wavelengths (380 and 340 nm) while the emitted fluorescence is measured for both excitations at >480 nm. Emission at 340 nm excitation is proportional to fura-2 bound to Ca”, and emission at 380 nm excitation is proportional to the free fura-2 concentration in the cell. Grynkiewicz et al.(19) developed a formula to calculate [Ca1+Iifrom dual-excitation emitted fluorescence, which is independent of the precise intracellular concentration of fura-2: [Ca’+li = &S(R - R~,.J/(R,,

- R)

(nM)

where Kd is the dissociation constant of fura-2 binding to Ca”, @ = F380,,/F38Omi,, with F380,, the fluorescence F at 380 nm excitation with all fura-2 dissociated from Cal+ions and F38OdnF a t 380 nm excitation with all fura-2 bound to Cal+ ions. R = F340/F380, that is, the ratio of emitted fluorescence excited at 340 nm (F340)and that excited at 380 nm (F380). R, is the maximal R measured after saturation of intracellular fura-2 with Ca” by adding 2 p M ionomycin (Calbiochem, La JoUa, CA) to the incubation solution, which allows equilibration of intracellular with extracellular Ca” (1.5 mM). R ~ ,is, the minimal R measured during complete dissociation of fura-2 from Ca2+by adding 20 mM EGTA just after the measurement of R., Under our conditions Kd = 224 nM,(19)fl = 4.2, R- = 0.34, and R , = 3.44.

During the measurements the cells were adhered to a glass coverslip serving as the bottom of an incubation chamber on the stage of the microscope.(23)The cells were bathed in a standard incubation solution composed of (mM) 125 NaCl, 5 KCl, 1 MgCl,, 1.5 CaCla, 1 KHaPO4, 10 NaHC03, 20 HEPES (PH 7.4), 5.5 glucose, and 2.5 probenecid. Probenecid (Sigma, St. Louis, MO) was used to prevent fura-2 extrusion from the cell and to prevent cornpartmentalization.04)Test solutions contained no added Caa+ or 6 mM Ca”. To load the cells with fura-2, the cells were incubated before the experiment for 30 minutes at 37°C in the standard incubation solution containing 2 pM fura-2AM (Boehringer, Mannheim, Germany), a membrane-permeant derivative that is hydrolyzed to fura-2 by intracellular esterases. Dual-wavelength excitation was controlled by a filter changer box connected to a personal computer. The F380 and F340 microscope images of the cells were captured by a sensitive black-and-white video camera (Hamamatsu, Hersching, Germany), connected to a video frame grabber board (PCVISION; Difa, Breda, The Netherlands) in an IBM-compatible AT personal computer. Frame grabbing to calculate fluorescence ratios and [Caa+Iifrom the F340 and F380 images was controlled by video image analysis software (TIM; Difa, Breda, The Netherlands), supported by custom-made application and commercial (EQCAL; Biosoft, Cambridge, UK) software. Before calculation of R, the F340 and F380 images were corrected for the dark current signal of the video camera. A standard calibration run for the determination of R, and Rd,,was used to obtain better comparison between the cultures. The calibration run was regularly repeated to establish that the measurement conditions were constant. Although this calculation procedure implies some uncertainty on the precise [Caa+Ii,it does not influence our conclusions about the rel-

CONDUCTANCES G FOR VARIOUS IONSIN DIFFERENT BONECELLTYPES AND ANIMAL SPECIES~ TABLE1. CELLMEMBRANE

Conductance (activation by) Cell type

GKi (hyperpol)

GKto

(depol)

GKCo (depol)

Ga (voltage)

GNa

(depol)

GN=

Gco (depol)

Chick osteoblasts Chick osteocytes Chick PFB Chick osteoclasts Rat osteoblasts Rat ROS 17/2.8 Rat UMR-106 Rat osteoclasts Pig chondrocytes Human G292

GxATP

(agonut)

Gxsa (stretch)

+ (81 ~

~

.The first one to two subscript letter@)(K, (I,Na, Ca, x ) indicate the type of ion conducted ((I for anion; x for unknown or unspecified ion selectivity). GK abbreviations(GKi, GKtp, GKCa) are explained in the text. The activation mechanism is indicated as voltage activated (if not further specified), depolarization advated, hyperpolarization activated, agonist activated, or stretch activated (su). The depolarization-activated GNa is the classic TTX (tetrodotoxin)-blockable sodium conductance. The conductances listed under voltage-activated GNaand Gcs and stretch-activated Gs, are heterogeneous. A plus sign means that the conductance or its underlying channels have been rcproducibly found in that cell, although not necessarily with 100% frequency. The minus sign means that the conductance(or channels) has never been found, so far, in that particular cell. Empty places indicate absence of data. A question mark is suggestive, but there is no definitive evidence from voltage clamp analysis. Numbers refer to the reference list, with (0) referring to the present paper and (*) to unpublished data from our group.

YPEY ET AL.

s382

ative changes in intracellular [Cal+] measured within the same cell and compared between cells.

RESULTS Voltage-activatedconductances and channels: A short review In the next two sections we review the voltage-activated channels and conductances in embryonic chick bone cells, based on results from our own studies. This review also serves to explain and illustrate the power of the patchclamp technique in analyzing transmembrane ionic currents in single bone cells. When appropriate, we compare the observations with studies from other groups on other species. Without attempting to be complete, the conductances and channels discussed in this paper are summarized in Table 1.

Calcium-dependent outward-rectifyingK conductance ( G K c ~

+

The G K has ~ been found in all types of embryonic chick bone cells we investigated.(11.15*17.10) It is a K+ conductance activated by depolarization as well as by an increase in [Cal+]i. Its properties and the various patchclamp configurations to establish these properties are reviewed in Fig. 1 for cells from OBL cultures. Figure 1A shows that silent G K channels ~ can be activated in a CAP of an intact OBL by depolarizing the patch more than 50 mV. The voltage across the patch is made more positive by applying a linearly increasing voltage difference (a ramp) to the two electrodes. Thus, the x axis is as voltage as well as a time axis, with 100 mV corresponding to 5 s. The lower, almost horizontal black record is the small baseline current, occurring when no channels are open. Current at increased levels above baseline indicate open channels. Analysis of many records (six in Fig. 1A) made it possible to conclude that the three increased levels correspond to three identical channels. The more the patch is depolarized, the more channels open and the longer they remain in the open state. More depolarization also provides more driving force and increases the current through the channels. The slope between the first-level current (I)-voltage (V) record and baseline is the single-channel conductance g, which was 220 pS for the conditions used (SECS in the bath and ICS in the pipette). A CAP with G Kchannels ~ as in Fig. 1A was excised to make an IOP, and its inside was exposed to ICS with increased [Cal+] to demonstrate [Cal+]i dependence (Fig. 1C). At pCa = 7, no or only a small amount of channel activity was present, but exposure to pCa = 3.2 caused tremendous, but reversible, channel activation. The [Ca1+Iidependence of the GKQ channels was also established from WC experiments, in which the current flowing at each applied voltage was measured (Fig. 1B) at low and high [Can]i in the pipette. At pCai = 7, the current activates at V, > 80 mV. However, application of pCai = 3.2 to the same cell shifts activation to V, > 0 mV. This shift is graded with [Ca1+lifor 3 < pCa < 6. The

observations suggest that [Cal+]iin cells with silent CAP G K channels ~ ~ under resting conditions (Fig. 1A) may be as low as 100 nM. This is confirmed with intracellular [Ca1+Iimeasurements (see later). Figure 1D shows voltage ramp-induced channel currents from an OOP pulled from a WC like that in Fig. 1B. At the applied &ai = 5.1, the channels are active at V, > -50 mV, which makes it possible to observe or estimate the voltage at which the channel current reverses from outward to inward. This reversal potential E, cannot be determined precisely because the inward current becomes smaller than the outward current, but E, must be at < -50 mV. This proves K+conduction through the channel, since the asymmetric K+gradient is the only gradient (with a calculated equilibrium potential of -84 mV) that can provide such a negative E,. In about one-third of CAPSon OBL, OCY, or PFB, the G K achannels were already active under resting conditions ( Vpip = 0), which may indicate an increased [Ca1+Ii.These channels could be made less active by decreasing external [Ca”] to nominally zero by EGTA (data not shown). A recently described K+channel in pig chondrocytes in CAP(1o)is very similar to the chick bone cell %ca channel in CAP. The existence of a GKQ in osteoblast-like cells is also consistent with predictions from earlier microelectrode but our results also raised the question of whether normal intracellular conditions would also require such high Cal+concentrations (> 1 pM) for significant activation of the G K Cchannels. ~

Other voltage-dependent conductances Figure 2 reviews the principal voltage-activated conductances in freshly isolated osteoclasts in WC configuration, described by Ravesloot et al.(I5) The conductances were activated by voltage steps, illustrated in the lower part of Fig. 2. Hyperpolarizing voltage steps from a holding potential V,, = -70 mV evoke large inward currents that become time dependent at extreme hyperpolarization. The conductance underlying these currents was identified as the inward-rectifying K+ conductance (GK~),blocked by 5 mM external Cs+. It is activated below its reversal potential E, (= Ek, the equilibrium or Nernst potential for the K+ion distribution across the membrane), where it conducts K+ ions from outside to inside. Depolarizing steps to V , > -20 mV evoke transient outward currents. The underlying transient outward-rectifying conductance ( G K ~preferen~) tially conducts K+ions and is blocked by 4 mM external 4aminopyridine. Steps to V, > -50 mV activate an outward-rectifying K+conductance ( G K ~ )The . noisy G Kcur~ rents are superimposed on the declining G K currents. ~ ~ The G K has ~ never been observed by us and others in osteogenic cells of any species. The G K has ~ been ~ seen occasionally in chick osteoblast cultures. (I1) The G K occurs ~ in all chick osteoblastic cells we have studied,‘”) and its underlying channels are calcium-dependent K+ chann e l ~ . ‘ ~Thus, ~ . ~ a~ plot ) of the GKo current as a function of the voltage step, after subtraction of the G K current, ~ ~ is similar to that of Fig. 1B at the low [Cal+], which implies that G K =~

IONIC CHANNELS AND INTRACELLULAR CALCIUM IN BONE CELLS Figure lB, as well as Fig. 2, illustrates for our measurement conditions, the absence of the classic types of depolarization-activated Ca" or Na+ conductances in chick embryonic osteoblasts and osteoclasts, consistent with experimental results under optimal recording conditions.(11.16) Others did not find these conductances in rat osteoclasts,(ll*as)but Chesnoy-Marchais and Fritsch") did find these conductances in rat osteoblasts. The final voltage-activated channel in chick embryonic bone cells to be mentioned is a large-conductance (380 pS) channel conducting such anions as C1- and phosphate ions.(a6)This anion conductance (G,) channel has not yet been seen in intact cells (CAP), only in excised (IOP) patches,'") but has been best characterized in OBL, OCY, and PFB, where the channel is active only between -30 and 30 mV. These results raised the question of the physiologic intracellular factors activating the G, channel, which is distinctly different from the WC chloride conductance found in rat oste~blasts(~~ and rat osteoclasts.(lS)Various conductances and channels found in bone cells in different animal species are summarized in Table 1.

Some recent results The results to be discussed here are new but also serve to review similar observations from other studies. Stretch Sensitivity of Ion Channels: In mechanorecep tive sensory cells, mechanoreceptive ion channels are the mechanoelectrical transducers for the generation of recep tor potentials as a fist step in the encoding process for neuronal information processing. Recently it has become clear that the mechanosensitivity of ion channels is a widespread property of cell membranes and is not restricted to specific classes of sensory cells.(2)This notion stimulated the search for mechanosensitivechannels in bone cells that could serve as sensors transducing mechanical stress to changed bone cell activity during adaptation of bone to loading.(a*16) The first results of Duncan and Misler(l') and Davidson et al.(a) are promising and prompted us to extend their observations for the various classes of chicken bone cells in primary culture to arrive at a more general picture of the possible role of mechanosensitivechannels in bone metabolism. During current measurements in patch-clamp experiments, mechanical stress was applied to the cell membrane by applying under- or overpressure (&SO cm water) to the inside of a patch pipette under cell-attached patch or whole-cell conditions. External underpressure on a CAP distends the patch membrane, and the membrane shrinks again during overpressure. The opposite occurs during WC measurements. Since investigators have hypothesized that osteocytes and osteoblasts could have a mechanoreceptor function in the response of bone to loading,("-") we fiirst searched for stretch sensitivity of ionic channels in these two bone cell types. As a control we screened osteoclasts for stretch-sensitive channels. While testing the membrane for stretch sensitivity, the membrane was depolarized or hyperpolarized by 5&100 mV to facilitate finding the channels and

5383

to be able to find channels that are both voltage and stretch activated. In 18 GO-sealed osteocyte CAPs from pure osteocyte cultures and 7 osteoblast CAPs, we have not found channels activated by membrane stretch; neither did we find clear evidence for direct or indirect effects of stretch on G K channel ~ activity, different from the results of Taniguchi and G~ggino,(~') who found results consistent with stretch-induced calcium entry affecting GKQ channels in a rabbit kidney cell line. In later series of experiments on osteoblastic cells we only occasionally found channel activity increased by membrane stretch. The infrequent presence of this type of chmnel is in contrast to the findings of Duncan and Misler(16)and Davidson et al.(@) In WC experiments on osteoclasts, we found no evidence for membrane stretch-induced (40 cm H,O pressure) changes in the WC currents through the G K and ~ G K ~ ~ (three cells). In one of the cells, for example, the effects of 40 cm HIO overpressure and 80 cm underpressure were within the normal range of variation (5%) of control current peaks at -120 mV (GKJ and 30 mV ( G K , ~ )We . also did not fiid significant effects on G Kchannels ~ in cell-attached patches. However, in 11 of 35 CAPs (from 35 OCL), we found 40 pS channels (mean 41 pS, range 29-53 pS, estimated from 12 I-Y plots from seven cells), which were reversibly activated by membrane stretch (Fig. 3). At resting ( Ypip= 0 mv) and hyperpolarized (positive Vpip>conditions, channel currents were inward with SECS in the pipette and in the bath, and the mean reversal potential (E,) was 36.0 mV (range -4 to 70 mV, based on 12 I-Y from seven cells) more positive than the resting membrane potential. Under these conditions, and G K channels ~ reverse at Y , more negative than the resting membrane potential (data not shown). Channel activity was not strongly dependent on the membrane potential. Intracellular Caa+Measurements: The observations on the G Kchannels ~ in the various chick bone cells raised interesting questions with respect to normal values of singlecell [Ca'+]i, its variability in time and in the intracellular space, and its dependence on external [Ca'+]. The microfluorescence experiments showed that [Caa+Ij of OCY in mixed OBL/OCY cultures depends on external Ca" (Fig. 4). At 1.5 mM external Ca", the stationary mean [Ca2+Ijwas 144 nM (range of singleell mean values 94-188 nM, 12 cells from three cultures). Over a period of 5 minutes, singleell mean values could fluctuate by 20%. No calcium oscillations were found under our conditions. After 4-5 minutes of exposure to practically 0 mM external Ca", [Ca1+liwas reduced to 46% (range 35-55%) of its control value. Recovery of [ C a l i after 4-5 minutes at 1.5 mM external Ca'+ was to 84% (range 67-loSVo) of the control. After 4-5 minutes of exposure to 6 mM external Caa+, [Ca1+Ij was slightly further increased to 98% (range 61-172070) of the original 1.5 mM control value. Individual OBL/PFB-like cells and the mean [Caa+Iiof all (100-200) cells imaged from a section of the culture behaved in the same way. The calculated [Ca1+liin single OCYs, both in OBL/OCY and in pure OCY cultures, was often (for ex-

YPEY ET AL.

T

FIG.4. [Caa+Iiin embryonic chick bone cells and the effect of external Ca’+. An osteocyte from a 2 day OBL/OCY culture, first exposed to l .5 mM external Ca’+ (l), then to nominally zero mM Ca2+(2),then back to l .5 mM (3). and finally to 6 mM external Ca’+. The images were taken 5 minutes after exposure to the external [Ca2+].Pixels with fluorescence ratio values above the mean of the OCY under control conditions (1) are darkened in 1-4 to illustrate the effect of external Ca” on [CaZ+li.Pixels with ratio values above the mean in 2 and 3 are hatched to illustrate maintenance of the intracellular Ca” distribution during changed external Ca2+.The ratios in 1-4 were 0.64,0.52, 0.63, and 0.72, respectively. Pixel values are ratio values x 64. The finely branched processes of the OCY are not visible in the imaged cell because of a lack of resolution ( x 20 objective used).

ample, 80% of the cells in a pure OCY culture) seen to be heterogeneous (Fig. 4A). In the example OCY shown in Fii. 4A, [Ca*]i values > 100 nM (the mean value of the cell, range 61181 nM) are clustered midright and along the upper and lower borders of the cell at 1.5 mM external Ca*. This distribution was not drastically changed at zero external [aa+],where the mean [Ca*Ii = 55 nM, range 34142 nM (Fig. 4A2). In OCL, the mean [Ca1+liequaled 104 nM (range of mean singlecell values 59-184 nM, eight cells from five cultures from three culture batches), consistent with microphotometric measurements by Miyauchi et al. After 4-5 minutes of exposure to zero external [Ca2+],[Caa+Iiwas reduced to 51% of its control value (range 39-58%, five cells from two cultures). All five cells had recovered after 4-5 minutes at 1.5 mM external Ca”, to a mean of 121% (range 78-156%) of the original control value. The calculated [Ca2+]iwas not homogeneous within a single OCL (Fig. 4B). In one example cell, the mean [Ca2+Iiwas 103 nM at external [Ca’+] = 1.5 mM and the range was 39-277 nM (>500 intracellular measurement points), with the above-mean values patterned along the cell boundary and in the lamellipods extending from the nuclei-containing

cell body. This distribution was not drastically changed at zero external [Ca2+],as in OCY. The intracellular inhomogeneity in apparent [Ca2+Ii,and the variability and dependence of [Caa+Iion external [Ca2+] are consistent with the observed variability in G K chan~ nel activity under resting conditions and the decrease in channel activity with the removal of external Ca’+. However, pCa < 6, as required for GKC- activation in IOP and WC measurements, were usually not encountered in intact cells, except in a few cells of a culture, which were probably dying.

DISCUSSION The present paper reviews properties of ionic channels in embryonic chick bone cells in comparison to ionic channel properties of bone cell membranes in other species. Furthermore, two new results are presented. First, embryonic chick osteoclasts express stretch-activated channels, and second, microfluorescence measurements revealed that the calculated resting [Ca2+Iiof bone cells may be heterogene-

IONIC CHANNELS AND INTRACELLUIAR CALCIUM IN BONE CELLS ous within bone cell and is dependent on the extracellular [Ca’+]. A survey of the principal conductances and channels found in bone cells is given in Table 1. The various conductances can be discriminated by ionic selectivity, mechanism of activation (by voltage, agonists, or stretch), and kinetics (transient or stationary activity). Voltage-activated channels comprise two classes, depolarization-activated and hyperpolarization-activated channels. An example of an agonist-activated conductance is the extracellular ATPactivated conductance G X ~ with p as yet undefined ionic selectivity. (3a) As Table 1 shows, bone cell ionic channels, both cation and anion channels, are usually strongly dependent on the membrane potential. This implies an important role for the membrane potential in bone cell function regulation. Unfortunately, the role is still unknown but is becoming more appreciated since membrane electrophysiology has become part of the research effort to clarify the cellular physiologic mechanisms of bone tissue metabolism. Thus far, we have been mainly concerned with K’ channels, since these channels were dominant in embryonic chick bone cells. The importance of K+ ions in podosome expression by OCL is suggested by Miyauchi et al.cL8)Other investigators have focused on Caa+ channels in mammalian osteoblastic cells.(4,7.33*34) Richter and F e r ~ i e r , ‘ on ~ ~ )the other hand, found stationary weakly voltage-dependent Na+ channel activity in the osteoblastic cell line ROS 17/2.8, different from the TTX-blockable Na+ channels in rat osteoblasts.ta9)Our patch-clamp experiments did not reveal any of the classic voltage-activated Caa+conductances in chick bone cells, but the microfluorescence experiments suggest the presence of Caa+conduction pathways in the plasma membrane, possibly small voltage-insensitive Caa+channels. (36) It may be expected that some bone cell functions depend on the coordinated action of K+and Ca” channels. In voltage-activated Caa+channels, the K’ channels may provide a hyperpolarizing force to the membrane for deactivation of Ca” channels and recovery of inactivated Ca” channels, and stationary Na+(35)and Caa+channels could provide depolarizing forces to the membrane, leading to increased [Ca2+]i.In voltage-insensitive Ca” channels, the K” channels may set the driving force for Ca” entry through the plasma membrane.(371 The G K of ~ OCL is most suited for inward K+ conduction. This K+uptake by the cell could occur, for example, from the resorption lacuna, in exchange for proton extrusion, when the electrogenic effects of the. proton pump drive the membrane potential to below the resting membrane potential.‘l’) The G K Cmay ~ serve to replenish the resorption lacuna with K+ ions when the [Ca2+Iibecomes high under depolarized conditions. Finally, G, channels could also serve to provide the electrogenic proton pump with counterions for proton extrusion. The membrane potential would then determine which types of ionic channels ( G K ~G. K C ~G, K ~or~ G,) , become active in controlling the pump efficiency. If we learn more about the molecular control of the OCL secretion machinery, we will come closer to a complete biophysical model of osteoclastic proton secretion.

5385

The presence of stretch-activated channels in chick embryonic OCL and its infrequent presence in chick OBL and OCY is surprising, since OBL and OCY have been proposed to possess membrane stress mechanisms to transduce changes in mechanical bone cell loading to changed bone metabolism.(8.16)Therefore, we speculate that the OBL/OCY stretch-activated channels are subject to variable expression and that the OCL stretch-activated channels rather play a role in OCL locomotion during resorptive OCL activity. For example, OCL locomotive movements may stretch part of the cell membrane, which may lead to local stretch-induced opening of channels. If the channels are aselective, consistent with the data, Caa+ could enter into the cell and cause local cell detachment due to decreased podosome expression.(I8) This could even be part of a mechanism for cell displacement to continue bone resorption at another spot after excavation of the previous resorption lacuna. The stretch-activated channel in chick OCL has similarities to that in cultured skeletal muscle of the same species.(38) We observed variability between culture batches in the distensibility of a cell membrane in WC or in CAP with applied pressure. Thus, the absence of stretch effects in some of the OCL and in the OCY and OBL may have been due to stiffness of the membrane. This stiffness may be under the control of cytoskeletal proteins. (38) Furthermore, certain ionic channels and intracellular processes may be sensitive to shear ~ t r e ~ instead ~ ( ~ ~ of *to~planar ~ ) membrane stretching. Morris and Horn‘”’) proposed in their recent study that stretch-activated ion channels in membrane patches derive their stretch sensitivity from patch-pipette membrane interactions, leading to alterations of the cortical cytoskeleton under the plasma membrane. Such alterations may be involved in osteoclast attachment to and detachment from the substratum during osteoclast locomotion. Our measurements of [Caa+Iiin intact cells by image analysis were a rust step in the analysis of the role of G K C ~ channels in bone cell function. The high [Caa+Iirequired for G Kactivation ~ in WC and IOP is still to be explained. It may indicate an intracellular calcium amplification factor for G K activation ~ in the intact cell. Thus far, the observations of the present study are in favor of a local, subcellular role for [Ca’+]i in bone cell activity, as in neurons.‘41) The combined application of patch-clamp and calcium-imaging techniques is a promising approach to study and analyze the local effects of membrane stretching and channel opening on local [Ca’+]i and on cell behavior.

ACKNOWLEDGMENTS We thank Prof. Dr. E.H. Burger (Dept. Oral Cell Biology, Free University, Amsterdam) for her encouragement to search for stretch-sensitive channels. We acknowledge the contributions of Mrs. A. Wiltink and Mr. W. Mulder to the experiments on stretch-activated channels. This work was supported by The Netherlands Organization for

5386 Scientific Research (NWO) through grants (for Ravesloot and Weidema) from the Foundation for Biophysics. We are particularly grateful to Mr. G.J.Verschragen for designing and constructing the filter change controlling dual- 19. wavelength excitation and to Mr. J.P.Buys for writing the application software. 20.

YPEY ET AL.

regulated by voltage-gated calcium channels and extracellular calcium, controls podosome assembly and bone resorption. J Cell Biol 111:2543-2552. Grynkiewicz G, Poenie M, Tsien RY 1985 A new generation of calcium indicators with greatly improved fluorescence properties. J Biol Chem 260:3440-3450. Ravesloot JH, Van Houten RJ, Ypey DL, Nijweide PJ 1990 Identification of Caa*activated K+channels in cells of embryonic chick osteoblast cultures. J Bone Miner Res 512011210. REFERENCES 21. Van der Plas A, Nijweide PJ 1991 Isolation and purification of osteocytes. Calcif Tiss Int Suppl 48, Abstract 159. 1. Hille B 1984 Ionic Channels of Excitable Membranes. 22. Horn R, Marty A 1988 Muscarinic activation of ionic curSinauer Associates, Sunderland, MA. rents measured by a new whole-cell recording method. J Gen 2. Morris C 1990 Mechanosensitive ion channels. J Membr Biol Physiol 92:145-159. 11393-107. 23. Ince C, Van Dissel JT, Diesselhoff MMC 1985 A Teflon cul3. Chesnoy-Marchais D, Fritsch J 1989 Chloride current actiture dish for high-magnification microscopy and measurevated by cyclic AMP and parathyroid hormone in rat osteoments in single cells. Pfluegers Arch 403240-244. blasts. Pfluegers Arch 415:104-114. 24. DiVergilio F, Steinberg TH, Silverstein SC 1990 Inhibition of 4. Guggino SE, Lajeunesse DL, Wagner JA, Snyder SH 1989 fura-2 sequestration and secretion with organic anion transBone remodeling signaled by a dihydropyridyne- and phenylport blockers. Cell Calcium 1157-62. alkylamine-sensitive calcium channel. Proc Natl Acad Sci 25. Sims SM, Kelly MEM, Dixon SJ 1991 K' and C1- currents in USA 86:2957-2960. freshly isolated rat osteoclasts. Pfluegers Arch 419358-370. 5. Petersen OH, Maruyana Y 1984 Calcium-activated potas- 26. Ravesloot JH, Van Houten RJ, Ypey DL, Nijweide PJ 1991 sium channels and their role in secretion. Nature 307:693High-conductance anion channels in embryonic chick osteo6%. genic cells. J Bone Miner Res 6:355-363. 6. Zaidi M, MacIntyre I, Datta H 1990 Intracellular calcium in 27. Bagi C, Burger EH 1989 Mechanical stimulation by intermitthe control of osteoclast function. 11. Paradoxical elevation tent compression stimulates sulfate incorporation and matrix of cytosolic free calcium by verapamil. Biochem Biophys Res mineralization in fetal mouse long-bone rudiments under Commun 162807-812. serum-free conditions. Calcif Tissue Int 49342-347. 7. Chesnoy-Marchais D, Fritsch J 1988 Voltage-gated sodium 28. El Haj AJ, Minter SL. Rawlinson SCF. Suswillo R, Lanyon and calcium currents in rat osteoblasts. J Physiol (Lond) 398: LE 1990 Cellular responses to mechanical loading in vitro. J 291-311. Bone Miner Res 5:923-932. 8. Davidson RM, Tatakis DW, Auerbach AL 1990 Multiple 29. Reich KM, Gay CV, Frangos JA 1990 Fluid shear stress as a forms of mechanosensitive ion channels in osteoblast-like mediator of osteoblast cyclic adenosine monophosphate procells. Pfluegers Arch 416646-651. duction. J Cell Physiol 143:100-104. 9. Ferrier J, Ward-Kesthely A, Homble F, Ross S 1987 Further 30. Skerry TM, Bitensky L, Chayen J, Lanyon LE 1989 Early analysis of spontaneous membrane potential activity and the strain-related changes in enzyme activity in osteocytes folhyperpolarizing response to parathyroid hormone in osteolowing bone loading in vivo. J Bone Miner Res 4:783-788. blastlike cells. J Cell Physiol 130:344-351. 31. Taniguchi J, Guggino WB 1989 Membrane stretch: A physio10. Grandolfo M, Martha M, R u d e r F, Vittur F 1990 A potaslogic stimulator of Ca" activated K' channels in thick assium channel in cultured chondrocytes. Calcif Tissue Int 47: cending limb. Am J Physiol 257:F247-F352. 302-307. 32. Ravesloot JH, Ypey DL, Nijweide PJ, Buisman HP, Vrij11. Sims SM, Dixon SJ 1989 Inwardly rectifying K' currents in heid-Lammers T 1989 Three voltage activated K' conducosteoclasts. Am J Physiol 256:C1277-C1282. tances and an ATP-activated conductance in freshly isolated 12. Ypey DL. Ravesloot JH, Buisman HP, Nijweide PJ 1988 embryonic chick osteoclasts. Pfluegers ARch 414(Suppl. 1): Voltage-activated ionic channels and conductances in embryS166-S167. onic chick osteoblast cultures. J Membr Biol 101:141-150. 33. Grygorczyk C, Grygorczyk R, Ferrier J 1989 Osteoblastic 13. Hamill OP, Marty A, Neher E, Sakmann B, Sigworth B 1981 cells have L-type calcium channels. Bone Miner 2137-148. Improved patch-clamp techniques for high-resolution current 34. Karpinsky E, Wu L, Civitelli R, Avioli LV, Hruska KA. recording from cells and cell-free patches. Pfluegers Arch Pang PKT 1989 A dihydropyridine-sensitivecalcium channel 39k85-100. in rodent osteoblastic cells. Calcif Tissue Int As:54-57. 14. Neher E, Sakmann B 1976 Single-channel current recorded 35. Richter C, Ferrier J 1991 Continuously active sodium chanfrom membrane of denervated frog muscle fibers. Nature nels in osteoblastic ROS 17/2.8 cells. Bone Miner 1557-71. W.779-882. 36. Kuno M, Gardner P 1987 Ion channels activated by inositol 15. Ravesloot JH, Ypey DL, Vrijheid-Lammers T, Nijweide PJ 1,4,5-triphosphate in plasma membrane of human T-lym1989 Voltage-activated K' conductances in freshly isolated phocytes. Nature 326:301-304. embryonic chicken osteoclasts. Proc Natl Acad Sci USA 86: 37. Randriamampita C, Bismuth G, Debre P, Trautmann A 1991 6821-6825. Nitrendipine induced inhibition of calcium influx in a human 16. Duncan R, Misler S 1989 Voltage-activated and stretch-actiT-cell clone: Role of cell depolarization. Cell Calcium 1 2 vated Baa*conducting channels in an osteoblast-like cell line 313-323. (UMR 106). FEBS Lett 251:17-21. 38. Guharay F, Sachs F 1984 Stretch-activated single ion channel 17. Ravesloot JH 1991 Ion channels in bone cells. Ph.D. thesis, currents in tissue-cultured embryonic chick skeletal muscle. J Leiden University. Physiol (Ion) 352:685-701. 18. Miyauchi A, Hruska KA, Greenfield EM, Duncan R, 39. Olesen SP, Clapham DE, Davie PF 1988 Haemodynamic shear stress activates a K' current in vascular endothelial Alvarez J, Barattolo R, Colucci S, Zambonin-Zallone A, Teitelbaum SL, Teti A 1990 Osteoclast cytosolic calcium, cells. Nature 331:168-170.

IONIC CHANNELS AND INTRACELLULAR CALCIUM IN BONE CELLS 40. Morris CE, Horn R 1991 Failure to elicit neuronal macroscopic mechanosensitive currents anticipated by single-channel studies. Science 251:1246-1249. 41. Silver RA, Lamb GL, Bolsover SR 1990 Calcium hotspots caused by L-channel clustering promoter morphological changes in neuronal growth cones. Nature 343:751-754.

Address reprint requests to: Dr. D.L. Ymy, Ph.D. Department of Physiology Faculty of Medicine Leiden University PO Box 9604 2300RC Leiden, The Netherlands

s387