PTH-Induced Actin Depolymerization Increases MechanosensitiveChannel Activity to Enhance Mechanically Stimulated Ca2+ Signaling

in Osteoblasts*

Jinsong Zhang,1,2,3 Kimberly D Ryder,2,4 Jody A Bethel,1 Raymund Ramirez,1 and Randall L Duncan1,3

ABSTRACT: Disruption of the actin cytoskeleton with cytochalasin D enhanced the mechanically inducedincrease in intracellular Ca2+ ([Ca2+]i) in osteoblasts in a manner similar to that of PTH. Stabilization of actinwith phalloidin prevented the PTH enhanced [Ca2+]i response to shear. Patch-clamp analyses show that theMSCC is directly influenced by alterations in actin integrity.

Introduction: PTH significantly enhances the fluid shear-induced increase in [Ca2+]i in osteoblasts, in part,through increased activation of both the mechanosensitive, cation-selective channel (MSCC) and L-typevoltage-sensitive Ca2+ channel (L-VSCC). Both stimuli have been shown to produce dynamic changes in theorganization of the actin cytoskeleton. In this study, we examined the effects of alterations in actin polymer-ization on [Ca2+]i and MSCC activity in MC3T3-E1 and UMR-106.01 osteoblasts in response to shear ± PTHpretreatment.Materials and Methods: MC3T3-E1 or UMR-106.01 cells were plated onto type I collagen–coated quartzslides, allowed to proliferate to 60% confluency, and mounted on a modified parallel plate chamber andsubjected to 12 dynes/cm2. For patch-clamp studies, cells were plated on collagen-coated glass coverslips,mounted on the patch chamber, and subjected to pipette suction. Modulators of actin cytoskeleton polymer-ization were added 30 minutes before the experiments, whereas channel inhibitors were added 10 minutesbefore mechanical stimulation. All drugs were maintained in the flow medium for the duration of the experi-ment.Results and Conclusions: Depolymerization of actin with 1–5 �M cytochalasin D (cyto D) augmented the peak[Ca2+]i response and increased the number of cells responding to shear, similar to the increased responsesinduced by pretreatment with 50 nM PTH. Stabilization of actin with phalloidin prevented the PTH enhanced[Ca2+]i response to shear. Inhibition of the MSCC with Gd3+ significantly blocked both the peak Ca2+ responseand the number of cells responding to shear in cells pretreated with either PTH or cyto D. Inhibition of theL-VSCC reduced the peak [Ca2+]i response to shear in cells pretreated with PTH, but not with cyto D.Patch-clamp analyses found that addition of PTH or cyto D significantly increased the MSCC open probabilityin response to mechanical stimulation, whereas phalloidin significantly attenuated the PTH-enhanced MSCCactivation. These data indicate that actin reorganization increases MSCC activity in a manner similar to PTHand may be one mechanism through which PTH may reduce the mechanical threshold of osteoblasts.J Bone Miner Res 2006;21:1729–1737. Published online on July 31, 2006; doi: 10.1359/JBMR.060722

Key words: mechanotransduction, PTH, intracellular Ca2+, actin cytoskeleton, fluid shear, osteoblasts

INTRODUCTION

SKELETAL INTEGRITY AND bone homeostasis are depen-dent on the physical forces exerted on bone during

movement. Frost(1) has proposed that discrete thresholds ofstrain magnitude exist where bone mass is lost, maintained,

or increased. Martin and Burr(2) have shown that thesethresholds are well defined and are species and size inde-pendent, but that the strain threshold where formation isgreater than resorption is at or beyond the limit of physi-ologic strain. Frost also proposed that these strain thresh-olds may be lowered by biochemical factors so that lesser,more physiologic, strains could promote bone formation.Several potential candidates for the modulation of me-chanical thresholds have been proposed. Of these, PTHmay be the strongest contender, producing a synergisticeffect on loading-induced bone formation.(3,4) Because

Parts of these data were presented at the 50th Annual Meetingof The Orthopaedic Research Society, San Francisco, CA, March7–10, 2004.

The authors state that they have no conflicts of interest.

1Department of Orthopaedic Surgery, Indiana University School of Medicine, Indianapolis, Indiana, USA; 2These authors contributedequally to this study; 3Department of Biological Sciences, University of Delaware, Newark, Delaware, USA; 4Cellular and IntegrativePhysiology, Indiana University School of Medicine, Indianapolis, Indiana, USA.

JOURNAL OF BONE AND MINERAL RESEARCHVolume 21, Number 11, 2006Published online on July 31, 2006; doi: 10.1359/JBMR.060722© 2006 American Society for Bone and Mineral Research

1729

JO605354 1729 1737 November

both PTH and mechanical stimulation activate similar sec-ond messenger pathways in osteoblasts, we hypothesizethat PTH may sensitize one of these pathways to lower thethreshold for mechanical stimulation.

Both PTH and mechanical stimulation produce a rapidincrease in intracellular Ca2+ ([Ca2+]i) that is dependent onextracellular Ca2+ entry and intracellular Ca2+ release.(5,6)

Ca2+ entry requires the activation of multiple ion channelsin osteoblasts. One of these, the mechanosensitive cation-selective channel (MSCC), is activated by membrane de-formation,(7,8) and when inhibited by the nonspecificblocker, gadolinium (Gd3+), the shear-induced increase in[Ca2+]i is significantly reduced.(5,9) We have shown that ad-dition of PTH to UMR 106.01 cells increases the MSCCstretch sensitivity and channel open probability,(10) suggest-ing that by increasing these kinetic parameters of theMSCC, PTH may increase Ca2+ signaling in osteoblasts inresponse to mechanical stimulation. L-type voltage-sensitive channels (L-VSCCs) have also been implicated inthe [Ca2+]i response to mechanical stimulation.(5) L-VSCCsare voltage-gated and dihydropyridine-sensitive, but do notseem to be responsive to membrane perturbation. How-ever, we have shown that the PTH-enhanced [Ca2+]i re-sponse to fluid shear requires the activation of both MSCCsand VSCCs in osteoblasts.(9) Furthermore, PTH stimula-tion alone has been shown to modulate L-VSCC activity inosteoblasts(6,11) and that inhibition of this channel signifi-cantly reduces Ca2+ signaling induced by PTH in osteo-blasts.(11)

Osteoblasts also respond to both PTH and mechanicalstimulation with an alteration in the actin cytoskeletal or-ganization. The actin cytoskeleton is a dynamic structuralnetwork that is essential for the regulation of a number ofcellular events, including mechanotransduction.(12,13) Me-chanical stimulation rapidly reorganizes the actin cytoskel-eton into stress fibers in the osteoblast,(14,15) and disruptionof actin stress fibers can lead to changes in the response ofosteoblasts to mechanical stimulation.(15) PTH also altersthe organization of actin in osteoblasts resulting in changesin cell morphology within minutes of stimulation,(16,17)

leading to the postulate that actin cytoskeletal reorganiza-tion may contribute to the functional response of osteo-blasts to PTH.(18)

Several ion channels have been shown to be directlylinked to, or modulated by, the actin cytoskeleton, includ-ing the epithelial Na+ channel,(19) mechanosensitive chan-nels,(20) and the L-VSCC.(21,22) In this study, we postulatethat PTH sensitizes osteoblasts to mechanical stimulationby modulating channel kinetics of the MSCC through re-polymerization of the actin cytoskeleton. To test this hy-pothesis, we disrupted or stabilized the actin network withcytochalasin D (cyto D) or phalloidin before PTH treat-ment or application of fluid shear and determined the[Ca2+]i response and the activation of the MSCC.

MATERIALS AND METHODS

Cell culture

The mouse osteoblast-like cell line, MC3T3-E1 (passages6–19), and the rat osteosarcoma cell-line, UMR106.01 (pas-

sages 27–45), were kind gifts from Dr Mary C Farach-Carson (University of Delaware) and Dr Nicola Partridge(St Louis University), respectively. These cells were grownin �-MEM (MC3T3-E1) and DMEM (UMR-106.01) con-taining 10% FCS (Gibco, New York, NY, USA), 100 U/mlpenicillin G, and 100�g/ml streptomycin. Cells were main-tained in a humidified incubator at 37°C with 5% CO2/95%air and subcultured every 72 h. All cell culture media andantibiotics were purchased from Sigma Chemical, St Louis,MO, USA.

Materials

The 1-34 fragment of bovine PTH [bPTH(1-34)](Bachem, Torrance, CA, USA) was dissolved in distilledwater and used at a final concentration of 50 nM. Gadolin-ium chloride (GdCl3), an inhibitor of MSCCs, was dissolvedin water at a stock concentration of 1 mM and used at afinal concentration of 10 �M. Nifedipine, an inhibitor ofL-VSCCs, was dissolved in 100% ethanol at a stock con-centration of 3 mM and used at a final concentration of 5�M. Cytochalasin D (cyto D), an actin cytoskeleton depo-lymerizing agent, phalloidin (phall), an actin cytoskeletonstabilizer, and nocodazole (noca), a microtubule disruptor,were dissolved in dimethyl sulfoxide at stock concentrationsso that the final concentration of the solvent was �0.1%.All drugs were obtained from Sigma, unless otherwise in-dicated.

Ca2+ imaging fluid flow experiments

MC3T3-E1 cells were grown for 4 days on type I colla-gen–coated (10 �g/cm2; Collaborative Biomedical, Bed-ford, MA, USA) quartz slides. For flow experiments, cellswere rinsed two times with Hanks’ balanced saline solution(HBSS). Cells were loaded with 3 �M fura-2/AM (Molecu-lar Probes, Eugene, OR, USA), a fluorescent Ca2+ probe,in HBSS for 45 minutes at 37°C. Loaded cells were incu-bated for an additional 15 minutes with HBSS alone toensure complete de-esterification of the fluorescent mol-ecule, yet minimize intracellular compartmentalization.

A parallel-plate flow chamber with a uniform flow chan-nel height of 250 �m was used to subject the cells to fluidshear, as previously described.(9) Flow was introduced tothe chamber through a syringe mounted on a Harvard Sy-ringe Pump (PHD Programmable; Harvard Apparatus,Holliston, MA, USA) that controlled the flow rate. To es-tablish a fluid-flow [Ca2+]i baseline, cells were exposed tofluid shear of 1 dyne/cm2 for 3 minutes. Fluid shear mag-nitude remained at 1 dyne/cm2 or was increased to 12 or 25dynes/cm2 for 3 minutes. Corresponding flow rates for eachof the fluid shear levels were 1, 15, and 30 ml/min, respec-tively.

A ratiometric video-image analysis apparatus (Intracel-lular Imaging, Cincinnati, OH, USA) was used to recordchanges in [Ca2+]i. Fura-2 fluorescence was visualized witha Nikon inverted microscope using a Nikon 30× fluor ob-jective. The cells were illuminated with a Xenon lampequipped with quartz collector lenses. A shutter and filterchanger containing the two different interference filters(340 and 380 nm) was computer controlled. In this system,

ZHANG ET AL.1730

emitted light is passed through a 430-nm dichroic mirror,filtered at 510 nm, and imaged with an integrating CCDvideo camera. The ratio of emitted light at 340- and 380-nmexcitation was determined (F340/F380) from consecutiveframes, and the [Ca2+]i for each cell was calculated fromthis ratio by comparison with fura-2 free acid standards.Computer-generated individual Ca2+ traces are populationmeans derived from simultaneous recording of Ca2+ in the4–12 single cells in the field of view.

Immunocytochemistry and fluorescence microscopy

MC3T3 cells were seeded on type I collagen–coated (10�g/cm2; Collaborative Biomedical) coverslips for 4 days.Cells were washed with PBS (Sigma) and treated for 0.5 hwith bPTH(1-34) or cytoskeletal modifiers in �-MEM. Af-ter treatment, cells were washed in PBS and fixed in 4%paraformaldehyde in PBS for 15 minutes. Cells were per-meabilized with 0.2% Triton-X100 (Sigma) in PBS for 5minutes. After permeabilization, cells were rinsed for 5minutes with PBS. Cells were incubated with 10 �g/mlFITC-phalloidin (Molecular Probes) in PBS for 30 minutesand washed three times for 5 minutes with PBS. Imageswere recorded using a Nikon Optiphot II microscopethrough a ×100 objective.

Patch-clamp studies

UMR106.01 cells were plated on glass coverslips and in-cubated for 48 h. Coverslips were transferred to the patchchamber and bathed in an isotonic Na+ Ringers consistingof (in mM) 137 NaCl, 5.5 KCl, 1 CaCl2, 1 MgCl2, 3 glucose,and 20 HEPES, titrated to a pH of 7.3 with NaOH. Trieth-ylammonium chloride (TEA; 1 mM) was added to the bathto block K channel activity. Fire polished, 5- to 10-M�borosilicate glass pipettes were backfilled with Na+ Ringer,and a conventional, cell-attached seal (>15 G�) was ob-tained. The membrane voltage was clamped at −40 mV, andbasal MSCC single channel activity and kinetics were de-termined by application of suction (15 mmHg, 1 Hz) to thebackside of the pipette. After determination of basal MSCCactivity, either 50 nM bPTH(1-34), 1 �M cytochalasin D, or1 �M phalloidin was directly added to the chamber, andchanges in MSCC kinetics were monitored for 30 minutes.The osmolality of all solutions was checked with a freezing-point depression osmometer (Precision Systems) and ad-justed to 300 ± 5 mosmol/kgH2O. Single channel currentswere recorded with a List EPC-7 amplifier (Medical Sys-tems, Great Neck, NY, USA), filtered at 1 kHz, and digi-tized at a sampling frequency of 2–5 kHz using pCLAMP8.0 software (Axon Instruments).

Statistical analysis

Mean peak [Ca2+]i response was expressed as a percent-age increase in [Ca2+]i over baseline [Ca2+]i levels. The per-centage of cells responding was determined by dividing thenumber of cells that responded with a 100% or greaterincrease in [Ca2+]i by the total number of cells. Data arepresented as mean ± SE and were obtained from at leastfive separate passages of cells. The data were pooled be-cause there was no significant difference for a treatment

between different passages of cells. Significance of all ex-periments was determined using one-way ANOVA, andthe Bonferroni posthoc test was used to determine signifi-cance when multiple comparisons in the study were made.Differences were considered significant when p < 0.05.

RESULTS

Effects of PTH and actin organization on theshear-induced Ca2+ increase

MC3T3-E1 osteoblasts respond to fluid shear with arapid increase in [Ca2+]i as shown in the representativetrace shown in Fig. 1A. Baseline [Ca2+]i levels were deter-mined during application of 1-dyn/cm2 shear. Fluid shearwas stepped to 12 dynes/cm2 for 3 minutes. We defined aresponding cell as one that had an increase in [Ca2+]i of atleast 100% over its baseline [Ca2+]i level. Defined by thiscriterion, the percent increase in the mean peak [Ca2+]i

response in sheared control MC3T3-E1 cells was 185 ± 28%(Fig. 2A), with 62 ± 3% (66 of 105 cells) of the cells re-sponding with this increase (Fig. 2B). Pretreatment ofMC3T3-E1 cells for 10 minutes before shear with 50 nMbPTH(1-34) produced a mean peak [Ca2+]i response of 295± 44% (p < 0.05) above baseline. PTH pretreatment alsosignificantly increased the number of cells responding to 82± 4% (55/67 cells; p < 0.05).

To determine the role of the actin cytoskeleton in theresponse of mean peak [Ca2+]i and the number of cellsresponding, we pretreated MC3T3-E1 cells with cyto D,which disrupts the actin cytoskeleton and promotes the for-mation of short chained actin filaments. Pretreatment ofMC3T3-E1 cells with cyto D increased the mean peak[Ca2+]i response to shear to 271 ± 32% (p < 0.01 comparedwith sheared control), with 80 ± 5% (45/56 cells; p < 0.05)of the cells responding to shear (Fig. 2). Whereas the ab-solute value of the peak [Ca2+]i response to shear + cyto Dwas much higher than that with PTH pretreatment (Fig.1C), cyto D also significantly increased baseline [Ca2+]i

from control levels of 87 ± 7 to 147 ± 15 nM (p < 0.01), thusmaking the percent increase in mean peak [Ca2+]i responseapproximately the same as PTH pretreatment. When theactin cytoskeleton was disrupted with cyto D, followed byaddition of PTH before shear, we observed no significantincrease in the mean peak [Ca2+]i response or the numberof cells responding compared with the PTH + shear group.We next stabilized the actin cytoskeleton during shear withphall, an agent that binds to actin and prevents reorganiza-tion (Fig. 1D). Pretreatment of MC3T3-E1 cells with phallfor 30 minutes before shear reduced the mean peak [Ca2+]i

response to shear (100 ± 15%; p < 0.05) and PTH + shear(151 ± 22%; p < 0.01). Phall also decreased the number ofcells responding to shear to 37 ± 3% (18/48; p < 0.05) andPTH + shear to 39 ± 4% (15/38; p < 0.01). These datasuggest that the actin cytoskeleton plays an important rolein the [Ca2+]i response to both shear alone and PTH +shear.

To determine the effects of disruption of microtubules onthe [Ca2+]i response to shear and shear + PTH, we pre-treated MC3T3-E1 cells for 30 minutes with nocodazole (5

CYTOSKELETAL CONTROL OF CA2+ SIGNALING IN OSTEOBLASTS 1731

�M) before application of shear. Figure 2 shows that noco-dazole did not significantly alter either the mean peak[Ca2+]i response to shear alone (166 ± 15%) or PTH + shear(228 ± 25%). Nocodazole also did not change the percent-age of the number of cells responding to shear alone orPTH + shear. These data suggest that disruption of themicrotubule structure has little effect on the Ca2+ signalingresponse to shear or shear + PTH.

Role of ion channels in the [Ca2+]i response toshear, PTH, and cytoskeletal reorganization

We have previously shown that inhibition of the MSCCwith Gd3+ significantly reduced the [Ca2+]i response andthe number of cells responding to shear and shear + PTH,but that inhibition of the L-VSCC with nifedipine was onlyable to significantly block the [Ca2+]i response in shearedcells pretreated with PTH.(9) To determine which of thesechannels are important to the increase in the [Ca2+]i re-sponse to shear during actin cytoskeletal disruption, weadded gadolinium or nifedipine separately to cells pre-treated with cyto D. Gadolinium (Gd3+; 10 �M) signifi-cantly reduced the mean peak [Ca2+]i response to shear andshear + PTH (Fig. 3A). Gd3+ also significantly reduced thenumber of cells responding to shear in cells pretreated withcyto D (Fig. 3B). However, inhibition of the L-VSCC withnifedipine only partially blocked the mean peak [Ca2+]i re-sponse in the shear + PTH group and failed to significantlyreduce the peak [Ca2+]i response to shear alone or shear +cyto D. These data suggest that the MSCC is important tothe [Ca2+]i response induced by shear and that PTH is ableto control the activity of this channel by altering the actin

cytoskeleton. These data further show that part of the PTH-enhanced response is mediated through the L-VSCC andthat this channel does not seem to be directly controlled bythe actin cytoskeleton.

Effect of PTH and cytoskeletal modifiers on theactin cytoskeleton

We examined the time-course of the effect of PTH oncell morphology and actin cytoskeletal organization inMC3T3-E1 cells (Fig. 4). MC3T3-E1 cells grown on type Icollagen are characterized by a flattened morphology witha prevalence of actin stress fibers traversing the cell. Within10 minutes of addition of 50 nM PTH, MC3T3-E1 cellsexhibited a more stellated morphology, and the stress fiberswere dramatically reduced compared with control cells. Cy-toskeletal changes were maximal at 30 minutes with heavystaining for f-actin appearing in the perinuclear region ofthe cell. Cyto D (1 �M) produced similar changes inMC3T3-E1 cell morphology and actin organization as PTHtreatment. However, after 30 minutes of cyto D treatment,f-actin staining was punctuated throughout the cell ratherthan focused in the perinuclear region as with PTH treat-ment. When PTH was added to cells after 30-minute treat-ment with cyto D, PTH did not further alter the cellularmorphology or actin organization or localization (data notshown). Addition of phalloidin (1 �M), which stabilizesactin and prevents reorganization, did not change the cellmorphology or actin stress fiber organization. Furthermore,addition of phalloidin before stimulation of MC3T3-E1

FIG. 1. Representative traces showing theintracellular Ca2+ response to 12 dynes/cm2

fluid shear in MC3T3-E1 osteoblasts exposedto the following: (A) untreated cells, (B) 50nM PTH pretreatment for 10 minutes, (C)cytochalasin D (1 �M) pretreatment for 30minutes, and (D) phalloidin (1 �M) pretreat-ment for 30 minutes followed by 50 nM PTHfor 10 minutes. All drugs were maintained inthe medium during shear. After mounting onthe flow chamber, MC3T3-E1 cells were sub-jected to a preflow of 1 dyne/cm2 for 3 min-utes to establish a baseline before the in-crease to 12 dynes/cm2 shear. PTH produceda significant increase in peak [Ca2+]i com-pared with untreated sheared controls. Ad-dition of cytochalasin D also significantly in-creased [Ca2+]i; however, this increase wasnot different from PTH pretreatment be-cause the baseline Ca2+ levels were elevated.Addition of phalloidin before PTH pretreat-ment completely blocked the enhanced[Ca2+]i response to shear.

ZHANG ET AL.1732

cells with PTH prevented the PTH-induced morphologicalchanges in cell shape and reduction of stress fiber within thecell.

Effects of PTH and the cytoskeleton onMSCC kinetics

To determine the effects of cytoskeletal reorganizationon MSCC kinetics, we used patch-clamp analyses to deter-

mine changes in channel activity in UMR-106.01 cells inresponse to PTH stimulation and actin cytoskeletal organi-zation. Using the cell-attached patch configuration to rec-ord single channel activities, we activated MSCC channelsby application of 15 mmHg suction to the back of the pi-pette. After determination of basal activity in untreatedcells, UMR-106.01 cells were treated with either PTH (50nM) for 10 minutes, cyto D (1 �M) for 30 minutes, phal-

FIG. 2. Effects of cytoskeletal disruption and stabilization on the(A) mean peak [Ca2+]i response and (B) the number of respond-ing cells in MC3T3-E1 osteoblasts exposed to 12 dynes/cm2 fluidshear alone or with 50 nM PTH pretreatment. PTH pretreatmentsignificantly increased both the mean peak [Ca2+]i response andthe number of cells responding to shear. Cytochalasin D (cyto D;1 �M) also significantly increased peak [Ca2+]i and the number ofcells responding to shear compared with shear alone. However,PTH did not enhance the effects of actin disruption by cyto Dabove the effects of cyto D and shear. Stabilization of actin withphalloidin (phall; 1 �M) before shear or shear + PTH significantlyreduced both the mean peak [Ca2+]i and number of cells respond-ing, suggesting that disruption of actin is required not only forPTH enhancement of these parameters but also for the responsesto shear alone. Addition of the microtubule disruption agent, no-codazole (noco; 5 �M), did not significantly alter either the meanpeak [Ca2+]i response or the number of cells responding to eithershear alone or shear + PTH. (ap < 0.05 compared with shearcontrols; bp < 0.01 compared with shear controls; cp < 0.05 com-pared with shear + PTH controls).

FIG. 3. Effects of MSCC and L-VSCC inhibition on the (A)mean peak [Ca2+]i response and (B) the number of respondingMC3T3-E1 osteoblasts to shear, shear + PTH, or shear + cyto D.Addition of the MSCC inhibitor, gadolinium chloride (10 �M),significantly reduced both the mean peak [Ca2+]i response and thenumber of cells responding compared with the shear control ineach of the groups. Addition of the L-VSCC specific inhibitor,nifedipine (5 �M), significantly reduced the mean peak [Ca2+]iresponse and the number of responding cells in the shear + PTHgroup, but failed to inhibit the mean peak [Ca2+]i response ineither the shear alone or the shear + cyto D groups. Nifedipine didsignificantly inhibit the number responsive cells in the shear +cyto D group compared with the shear + cyto D shear control. (ap< 0.05 compared with shear + cyto D control; bp < 0.05 comparedwith shear controls in each group; cp < 0.01 compared with shearcontrols in each group).

CYTOSKELETAL CONTROL OF CA2+ SIGNALING IN OSTEOBLASTS 1733

loidin (1 �M) for 30 minutes, or phalloidin for 30 minutesfollowed by PTH addition for 10 minutes. Representativetraces from each group are shown in Fig. 5A. To determinechannel open probability (NPo), we measured the time thechannel was open over the time that suction was applied.Thus, if one channel was open the duration of suction ap-plication, the NPo value would be 1.0. We observed little, ifany, spontaneous MSCC activity; however, suction in-creased NPo in control cells to 0.41 ± 0.03. As we havepreviously shown,(10) PTH pretreatment increased sponta-neous MSCC activity and increased NPo during suction to1.44 ± 0.08 (p < 0.0001). Disruption of actin cytoskeletalorganization with cyto D also increased NPo to 1.15 ± 0.13(p < 0.0001; Fig. 5B). When PTH was added to cytoD–treated cells, NPo was increased above that of eithercyto-D or PTH alone to 1.80 ± 0.22. Stabilization of theactin cytoskeleton with phalloidin reduced the MSCC NPo

to 0.30 ± 0.07 (p < 0.05 compared with untreated controls)and completely blocked the increased NPo elicited by PTH(0.37 ± 0.08; p < 0.05 compared with PTH-treated group).These data indicate that disruption of the actin cytoskeletonsignificantly increases MSCC activity similar to the increasein activity with PTH treatment. However, the NPo valuesfrom both the PTH alone and the cyto D + PTH groupswere significantly greater than the NPo from the cyto Dalone group (p < 0.01 and p < 0.001, respectively).

DISCUSSION

The mechanical forces placed on the skeleton throughlocomotion define the architecture and mass of bone. Whena novel mechanical load is encountered, bone cells have theability to perceive these vectorial changes in force and ini-tiate a cascade of cellular events that result in alterations inbone architecture to adapt to these new loads by resorbingbone in areas of low strain and forming new bone in regionsof increased strain. Whereas bending of a bone during lo-comotion undoubtedly produces strains on bone cells, themovement of fluid within the canaliculi and Haversian ca-nals of bone has also been proposed to be a significantmechanical signal.(23,24) Estimates of the magnitudes offluid shear forces generated in bone as the result of physi-ologic loads range from 8 to 30 dyn/cm2.(24) Whereas in vivoand in vitro studies have yet to determine what mechanicalsignal plays the dominant role in mechanotransduction,most mechanical forces activate many of the same secondmessenger pathways and signaling molecules that may beimportant to the osteogenic response to mechanical load-ing.(25) However, we have shown that fluid shear, but notphysiologic levels of mechanical stretch, increases the ex-pression of osteopontin, c-fos, cyclooxygenase 2, and TGF-�.(26) Therefore, in this study, the principal mechanicalstimulus used was laminar fluid flow.

Well-defined thresholds for the magnitudes of force re-

FIG. 4. Effects of PTH and cytoskeletalagents on actin filaments in MC3T3-E1 cellsusing rhodamine phalloidin to stain f-actinfilaments. Control static MC3T3-E1 cells ex-hibit a high degree of f-actin organized intostress fibers. Addition of 50 nM PTH over 30minutes decreased stress fibers in MC3T3-E1cells in a time-dependent manner, so that by30 minutes, few stress fibers remained andmost of the staining was localized to the peri-nuclear region of the cell. Cytochalasin D(cyto D) also disrupted the f-actin network;however, the staining pattern differed fromthat of PTH in that actin appeared to col-lapse to the points of attachment of the cell.If actin was stabilized with phalloidin for 30minutes before 30-minute treatment withPTH, f-actin organization did not appear dif-ferent from static control cells.

ZHANG ET AL.1734

quired for net bone resorption and formation have beendetermined, however the threshold for net bone formationexceeds the levels of strain that occur under physiologicconditions.(1,2) We, and others,(1,3,4,9) have postulated thatparathyroid hormone (PTH) can interact with the signalingmechanisms of mechanotransduction to lower the mechani-cal threshold and prime the osteogenic cells of bone torespond to lesser magnitudes of mechanical stimulation topromote bone formation. The effects of PTH on bone areparadoxical in that PTH is released in response to low se-rum Ca2+ and stimulates bone resorption to increase theserum Ca2+ levels. However, when given in low, intermit-

tent doses, PTH increases bone formation in intactrats,(27,28) ovariectomized rats,(29) and humans.(30) Theseanabolic effects are quite similar to the effects of mechani-cal stimulation on osteoblasts and bone. PTH has beenshown to induce a number of genomic responses, but per-haps most relevant to the mechanical effects on bone is thestimulation of prostaglandin synthesis by increasing COX-2production.(31) Inhibition of COX-2 with NS398, a specificCOX-2 blocker, has been shown to completely abrogatemechanically induced bone formation in vivo.(32) PTH hasalso been shown to reverse the effects of mechanical un-loading in hindlimb suspended rats and even produce tra-becular bone formation greater than controls.(3) Chow etal.(4) have also shown that PTH is essential for the mechani-cal responsiveness of bone to mechanical stimuli. Applica-tion of mechanical loading to rat tail vertebrae had no effectin thyroparathyroidectomized rats, yet a single dose of PTHre-established mechanical responsiveness. These observa-tions indicate that a synergistic interaction between PTHand mechanical loading of bone exists to stimulate boneformation, and we hypothesize that this interaction may bethe mechanism behind the anabolic response of bone to lowintermittent doses of PTH.

PTH and mechanical loading also activate many of thesame second messenger pathways and the two stimuli to-gether produce even greater activation of these pathways.Addition of PTH to rat dentoalveolar cells before mechani-cal stimulation significantly elevated inositol trisphosphate,cyclic AMP, and protein kinase C levels compared withmechanical stimulation alone.(33) One of the initial re-sponses of osteogenic cells to either mechanical stimulationor PTH treatment is a rapid rise in [Ca2+]i.

(5,6,34) Like othersecond messengers, this increase in [Ca2+]i in response toshear is significantly enhanced when osteoblasts are pre-treated with PTH.(9) Because the [Ca2+]i response to eitherstimuli has been linked to both changes in gene expressionand release of factors associated with signal amplification,the mechanisms involved in the increase in intracellularCa2+ are likely candidates for determining the mechanicalthresholds for mechanotransduction and that PTH can alterthese thresholds by sensitizing these mechanisms. This in-crease in [Ca2+]i to either stimulus is dependent on bothextracellular Ca2+ entry and intracellular Ca2+ release.(5,6)

Two ion channels have been shown to be involved in ex-tracellular Ca2+ entry in response to fluid shear: the MSCCand the L-VSCC.(5,9) Activation of these channels by fluidshear has been linked to release of a number of autocrine/paracrine factors,(35–38) suggesting that these channels areimportant in amplification of the mechanical stimulus toaugment the number of cells responding to the stimulus.

Both MSCC and L-VSCC exhibit increased activation inresponse fluid shear in MC3T3-E1 osteoblasts pretreatedwith PTH compared with cells subjected to fluid shearalone.(9) The mechanisms behind this increased activationare still unclear. However, ion channels can be phosphory-lated or dephosphorylated by a number of second messen-ger pathways that alter channel activity and kinetics.(39)

PTH activates several protein kinases that are known tophosphorylate L-VSCC channels, including protein kinaseA (PKA) and protein kinase C (PKC).(40) We showed that

FIG. 5. (A) Representative single channel traces of MSCC ac-tivity in control MC3T3-E1 cells or cells pretreated with 50 nMPTH, 1 �M cytochalasin D, 1 �M phalloidin, or 1 �M phalloidinfor 30 minutes followed by 10-minute treatment with PTH. MSCCchannels were activated by application of suction to the back ofthe pipette. (B) Average open probability (NPo) of MSCC from atleast 12 patched cells from each group. NPo was only determinedduring application of suction to the pipette. PTH significantlyincreased NPo 3-fold during application of mechanical stimulusand induced spontaneous activity of the channel. Introduction ofcyto D increased NPo, as well, and produced spontaneous activityof the MSCC. However, addition of PTH to cyto D–treated cellsdid not significantly increase NPo above that of cyto D alone.Phalloidin did not prevent activation of the MSCC on suction tothe pipette, but did prevent PTH enhancement of channel activity.(ap < 0.05 vs. control untreated; bp < 0.05 vs. PTH control; cp <0.01 vs. PTH control).

CYTOSKELETAL CONTROL OF CA2+ SIGNALING IN OSTEOBLASTS 1735

the enhanced increase in [Ca2+]i in response to fluid shearin PTH pretreated MC3T3-E1 osteoblasts is predominantlythe result of PKA phosphorylation of the L-VSCC.(9) How-ever, this enhanced response required the activation of theMSCC. This synchronization between the MSCC and L-VSCC in response to mechanical loading is similar to thesynergistic interaction of 1,25(OH)2-vitamin D3 andPTH.(41) We also showed that the kinetics of the MSCC arealso regulated by PKA. Addition of PTH before mechani-cal stimulation increased both the open probability and theconductance of the MSCC in UMR106.01 cells.(10) Further-more, pretreatment of UMR cells with 8br cAMP increasedthe conductance of the channel but did not alter the openprobability, suggesting that the increase in open probabilitywas controlled through another mechanism.

Although phosphorylation from other kinases, such astyrosine kinase, could be involved in this increase in openprobability of the MSCC, the nature of the gating mecha-nism of this channel suggests a structural control by the cell.Whereas deformation of the lipid bilayer has been impli-cated in the activation of mechanically gated channels, suchas the MscL,(42) the cytoskeleton of the cell has been linkedto regulation of several types of ion channels in differenttissues, including the epithelial Na+ channel (ENaC),(43)

L-VSCC,(44) voltage and ATP gated K+ channels,(45,46) andstretch-activated cation channels.(47) Prat et al.(43) foundthat disruption of actin with cyto D increased ENaC chan-nel activity. Furthermore, if monomeric g-actin was addedto excised patches at a concentration that promoted short-chained actin filaments, channel activity was also increased.Cyto D collapses the actin cytoskeleton without signifi-cantly altering the g-actin/f-actin ratio,(48) suggesting that,like the ENaC channel, the MSCC becomes more respon-sive when actin is disrupted into shorter filaments.

PTH has been shown to significantly alter osteoblastmorphology(16) and disrupt the actin cytoskeleton.(18) Aswe show here, the disruption of actin filaments by PTH israpid and correlates with the increase in MSCC activity.How PTH alters actin integrity is still unclear; however,both PKA and PKC are activated by PTH and have beenimplicated in the control of actin organization.(49,50)

Whereas PKA has been associated with loss of actin asso-ciation with integrins,(49) we find that activation of PKAonly increases single channel conduction of the MSCCs anddoes not increase open probability. PKC has been shown totarget several proteins associated with the cytoskeleton, in-cluding integrin- and actin-associated proteins that can capor sever actin filaments.(50) Thus, PTH activation of PKCmay be responsible for the increase in MSCC activation andis the subject of ongoing studies.

In summary, we found that loss of actin filament integritymimics the effects of PTH pretreatment on the intracellularCa2+ response to fluid shear in MC3T3-E1 osteoblasts.Whereas we have previously shown that much of this en-hanced response can be blocked when the L-VSCC is in-hibited, we believe that activation of the L-VSCC is depen-dent on the depolarization event initiated by the MSCC.(10)

Thus, increased MSCC activity will increase L-VSCC acti-vation that, in turn, leads to an increased intracellular Ca2+

response. We have previously shown that PTH pretreat-

ment will increase both MSCC single channel conductanceand open probability and that the increase in single channelconductance resulted from PTH-induced activation ofPKA. In this study, we showed that the increase in MSCCopen probability correlates with actin filament disruptionand that stabilization of the actin cytoskeleton before PTHstimulation prevents this increase. Whereas we can not con-clude a direct linkage of the cytoskeleton to the MSCC,these data indicate that PTH can prime osteoblasts to re-spond to mechanical stimulation through alteration in cel-lular structural integrity.

ACKNOWLEDGMENTS

This work was supported by NIH Grants NIDDKDK058246 and NIAMS AR043222 (RLD) and NASA Pre-doctoral Fellowship Grant NGT5-5023 (KDR).

REFERENCES

1. Frost HM 1987 The mechanostat: A proposed pathogenicmechanism of osteoporosis and the bone mass effect of me-chanical and non-mechanical agents. Bone Miner 2:73–85.

2. Martin RB, Burr DB 1989 Structure, Function, and Adaptationof Compact Bone. Raven Press, New York, NY, USA.

3. Ma Y, Jee WS, Yuan Z, Wei W, Chen H, Pun S, Liang H, LinC 1999 Parathyroid hormone and mechanical usage have asynergistic effect in rat tibial diaphyseal cortical bone. J BoneMiner Res 14:439–449.

4. Chow JW, Fox S, Jagger CJ, Chambers TJ 1998 Role for para-thyroid hormone in mechanical responsiveness of rat bone.Am J Physiol Endocr 274:E146–E154.

5. Hung CT, Allen FD, Pollack SR, Brighton CT 1996 Intracel-lular calcium stores and extracellular calcium are required inthe real-time calcium response of bone cells experiencing fluidflow. J Biomech 29:1411–1417.

6. Reid IR, Civitelli R, Halstead LR, Avioli LV, Hruska KA 1987Parathyroid hormone acutely elevates intracellular calcium inosteoblast-like cells. Am J Physiol 253:E45–E51.

7. Duncan RL, Misler S 1989 Voltage-activated and stretch-activated Ba2+ conducting channels in an osteoblast-like cellline (UMR-106). FEBS Lett 251:17–21.

8. Davidson RM, Tatakis DW, Auerbach AL 1990 Multipleforms of mechanosensitive ion channels in osteoblast-like cells.Pflugers Arch 416:646–651.

9. Ryder KD, Duncan RL 2001 Parathyroid hormone enhancesfluid shear-induced Ca2+ signaling in osteoblastic cells throughactivation of mechanosensitive and voltage-sensitive Ca2+

channels. J Bone Miner Res 16:240–248.10. Duncan RL, Hruska KA, Misler S 1992 Parathyroid hormone

activation of stretch-activated cation channels in osteosarcomacells (UMR-106.01). FEBS Lett 307:219–223.

11. Yamaguchi DT, Hahn TJ, Iida-Klein A, Kleeman CR, Mual-lem S 1987 Parathyroid hormone-activated calcium channelsin an osteoblast-like osteosarcoma cell line. J Biol Chem262:7711–7718.

12. Wang N, Butler JP, Ingber DE 1993 Mechanotransductionacross the cell surface and through the cytoskeleton. Science260:1124–1127.

13. Maniotis Andrew J, Chen Christopher S, Ingber Donald E1997 Demonstration of mechanical connections between inte-grins, cytoskeletal filaments, and nucleoplasm that stabilizenuclear structure. Proc Natl Acad Sci USA 94:849–854.

14. Buckley MJ, Banes AJ, Levin LG, Sompio BE, Sato M, JordanR, Gilbert J, Link GW, Tran Son Tay R 1988 Osteoblastsincrease their rate of division and align in response to cyclic,mechanical tension, in vitro. Bone Miner 4:225–236.

15. Pavalko FM, Chen NX, Turner CH, Burr DB, Atkinson S,Hsieh Y-F, Qiu J, Duncan RL 1998 Fluid shear-induced me-

ZHANG ET AL.1736

chanical signaling in MC3T3-E1 osteoblasts requires cytoskel-eton-integrin interactions. Am J Physiol Cell Physiol275:C1591–C1601.

16. Miller SS, Wolf AM, Arnaud CD 1976 Bone cells in culture:Morphologic transformation by hormones. Science 192:1340–1343.

17. Lomri A, Marie PJ 1988 Effect of parathyroid hormone andforskolin on cytoskeletal protein synthesis in cultured mouseosteoblastic cells. Biochim Biophys Acta 970:333–342.

18. Lomri A, Marie PJ 1990 Changes in cytoskeletal proteins inresponse to parathyroid hormone and 1,25-dihydroxyvitaminD in human osteoblastic cells. Bone Miner 10:1–12.

19. Cantiello HF, Stow JL, Prat AG, Ausiello DA 1991 Actinfilaments regulate epithelial Na+ channel activity. Am J PhysiolCell Physiol 261:C882–C888.

20. Kamkin A, Kiseleva I, Isenberg G 2003 Ion selectivity ofstretch-activated cation currents in mouse ventricular myo-cytes. Pflugers Arch 446:220–231.

21. Nakamura M, Sunagawa M, Kosugi T, Sperelakis N 2000 Actinfilament disruption inhibits L-type Ca2+ channel current in cul-tured vascular smooth muscle cells. Am J Physiol Cell Physiol279:C480–C487.

22. Pascarel C, Brette F, Cazorla O, Le Guennec JY 1999 Effectson L-type calcium current of agents interfering with the cyto-skeleton of isolated guinea pig ventricular myocytes. Exp Phys-iol 84:1043–1050.

23. Reich KM, Gay CV, Frangos JA 1990 Fluid shear stress as amediator of osteoblast cyclic adenosine monophosphate pro-duction. J Cell Physiol 143:100–104.

24. Weinbaum S, Cowin SC, Zeng Y 1994 A model for the exci-tation of osteocytes by mechanical loading-induced bone fluidshear stresses. J Biomech 27:339–360.

25. Duncan RL, Turner CH 1995 Mechanotransduction and thefunctional response of bone to mechanical strain. Calcif TissueInt 57:344–358.

26. Owan I, Burr DB, Turner CH, Qiu J, Yu T, Onyia JE, DuncanRL 1997 Mechanotransduction in bone: Are osteoblast-likecells more responsive to mechanical strain or extracellular fluidforces? Am J Physiol Cell Physiol 273:C810–C815.

27. Gunness M, Hock JM 1993 Anabolic effect of parathyroid hor-mone on cancellous and cortical bone histology. Bone 14:277–281.

28. Hock JM, Gera I 1992 Effects of continuous and intermittentadministration and inhibition of resorption on the anabolic re-sponse of bone to parathyroid hormone. J Bone Min Res 7:65–72.

29. Mosekilde L, Danielsen CC, Gasser J 1994 The effect on ver-tebral bone mass and strength of long term treatment withantiresorptive agents (estrogen and calcitonin), human para-thyroid hormone-(1-38), and combination therapy, assessed inaged ovariectomized rats. Endocrinology 134:2126–2134.

30. Dempster DW, Cosman F, Parisien M, Shen V, Lindsay R 1993Anabolic actions of parathyroid hormone on bone. EndocrRev 14:690–709.

31. Tetradis S, Pilbeam CC, Liu Y, Herschmann HR, Kream BE1997 Parathyroid hormone increases prostaglandin G/H syn-thase-2 transcription by a cAMP medicated pathway in murineosteoblastic MC3T3-E1 cells. Endocrinology 138:3594–3600.

32. Forwood MR 1996 Inducible cyclo-oxygenase (COX-2) medi-ates the induction of bone formation by mechanical loading, invivo. J Bone Miner Res 11:1688–1693.

33. Carvalho RS, Scott JE, Suga DM, Yen EHK 1994 Stimulationof signal transduction pathways in osteoblasts by mechanicalstrain potentiated by parathyroid hormone. J Bone Miner Res9:999–1011.

34. Jones DB, Nolte H, Scholubbers J-G, Turner E, Veltel D 1991Biochemical signal transduction of mechanical strain in osteo-blast-like cells. Biomaterials 12:101–110.

35. Ajubi NE, Klein-Nulend J, Alblas MJ, Burger EH, Nijweide PJ1999 Signal transduction pathways involved in fluid flow-induced PGE2 production by cultured osteocytes. Am. J. Phys-iol. Endocrine 276:E171–E178.

36. Rawlinson SCF, Pitsillides AA, Lanyon LE 1996 Involvementof different ion channels in osteoblasts and osteocytes earlyresponses to mechanical strain. Bone 19:609–614.

37. Sakai K, Mohtai M, Iwamoto Y 1998 Fluid shear stress in-creases transforming growth factor beth-1 expression in humanosteoblast-like cells: Modulation by cation channel blockers.Calcif Tissue Int 63:515–520.

38. Genetos DC, Geist DJ, Liu D, Donahue HJ, Duncan RL 2005Fluid shear-induced ATP secretion mediates prostaglandin re-lease in MC3T3-E1 osteoblasts. J Bone Miner Res 20:41–49.

39. Levitan IB 1994 Modulation of ion channels by protein phos-phorylation and dephosphorylation. Ann Rev Physiol 56:193–212.

40. Kamp TJ, Hell JW 2000 Regulation of cardiac L-type calciumchannels by protein kinase A and protein kinase C. Circ Res87:1095–1102.

41. Li W, Duncan RL, Karin NJ, Farach-Carson MC 19971,25(OH)2D3 enhances parathyroid hormone-induced Ca2+

transients in pre-osteoblasts by activating L-type Ca2+ chan-nels. Am. J. Physiol. Endocrine 273:E599–E605.

42. Hamill OP, Martinac B 2001 Molecular basis of mechanotrans-duction in living cells. Physiol Rev 81:685–740.

43. Prat AG, Bertorello AM, Ausiello DA, Cantiello HF 1993Activation of epithelial Na+ channels by protein kinase A re-quires actin filaments. Am J Physiol Cell Physiol 265:C224–C233.

44. Hohaus A, Person V, Behlke J, Schaper J, Morano I, Han-nelore H 2002 The carboxyl-terminal region of ahnak providesa link between cardiac L-type Ca2+ channels and the actin-based cytoskeleton. FASEB J. 16:1205–1216.

45. Cukovic D, Lu GW, Wible B, Steele DF, Fedida D 2001 Adiscrete amino terminal domain of Kv1.5 and Kv1.4 potassiumchannels interacts with the spectrin repeats of alpha-actinin-2.FEBS Lett 498:87–92.

46. Terzic A, Kurachi Y 1996 Actin microfilament disrupters en-hance K(ATP) channel opening in patches from guinea pigcardiomyocytes. J Physiol 492:395–404.

47. Guharay F, Sachs F 1984 Stretch-activated single ion-channelscurrents in tissue-cultured embryonic chick skeletal muscle. JPhysiol 352:685–701.

48. Norvell SM, Ponik SM, Bowen DK, Gerard R, Pavalko FM2004 Fluid shear stress induction of COX-2 protein and pros-taglandin release in cultured MC3T3-E1 osteoblasts does notrequire intact microfilaments of microtubules. J Appl Physiol96:957–966.

49. Howe AK 2004 Regulation of actin-based migration by cAMP/PKA. Biochim Biophys Acta 1692:159–174.

50. Larrson C 2006 Protein kinase C and the regulation of the actincytoskeleton. Cell Signal 18:276–284.

Address reprint requests to:Randall L Duncan, PhD

Departments of Biological Sciences and MechanicalEngineering

University of Delaware319 Wolf Hall

Newark, DE 19716, USAE-mail: rlduncan@udel.edu

Received in original form May 28, 2006; revised form July 10, 2006;accepted July 26, 2006.

CYTOSKELETAL CONTROL OF CA2+ SIGNALING IN OSTEOBLASTS 1737