Effect of calcium on oxytocin-induced contraction of mammary gland

myoepithelium as visualized by NBD-phallicidin

DAVID M. MOORE1, A. WAYNE VOGL1, KENNETH BAIMBRIDGE2 and JOANNE T. EMERMAN1*

'Departments of Anatomy and 2Physiology, The University of British Columbia, Vancouver, BC V6T l\V5, Canada

* Author for correspondence

Summary

The effect of calcium on oxytocin-induced con-traction of myoepithelial cells was visualizedwith NBD-phallacidin, a fluorescent stain forfilamentous actin. In the absence of oxytocin, thecells appeared relaxed; long, branching processesradiated from the cell bodies. In the presenceof 50nM-oxytocin, myoepithelial cells contractedinto smaller spoke-shaped bodies in which thearms were shorter and thicker. Electron mi-croscopy confirmed the morphological differ-ences between oxytocin-treated and untreatedmyoepithelium. To determine a role for extra-cellular calcium, tissue was incubated in EGTA,then exposed to oxytocin, with or without addedcalcium. Contraction occurred in the presence

of oxytocin plus additional calcium but not inthe absence of calcium. When the tissue wasincubated with the calmodulin antagonist trifluo-perazine (TFP) in calcium-containing medium,oxytocin did not induce myoepithelial cell con-traction. These data support previous resultsobtained with a myosin light-chain phosphoryl-ation assay implicating calcium and calmodulinin oxytocin-induced contraction. Furthermore,NBD-phallicidin visualization of myoepithelialcells demonstrates that the effect of calcium oncontraction is physiologically significant.

Key words: oxytocin, myoepithelial cells, calcium,calmodulin.

Introduction

Two types of epithelial cells comprise the parenchymaof the mammary gland: secretory epithelial cells, whichare involved in milk synthesis and secretion, andmyoepithelial cells, which are involved in milk ejection.The myoepithelial cells are contractile and their modeof contraction is thought to be somewhat analogous tothat of smooth muscle (Bremel & Shaw, 1978). Thereis evidence to suggest that myoepithelial cells as well assmooth muscle require calcium-dependent phosphoryl-ation of the 20 000 A/,, myosin light chain to prime themolecules for actin-induced ATPase activity (Chackoet al. 1977). However, contraction in these two celltypes is stimulated by different factors. Smooth musclecontraction is initiated by a variety of neurotransmit-ters released by the autonomic nervous system, whilethe myoepithelium contracts following binding of thehormone oxytocin to surface receptors (Soloff, 1976).

Journal of Cell Science 88, 563-569 (1987)Printed in Great Britain © The Company of Biologists Limited 1987

Oxytocin is a nonapeptide released, through a neuro-endocrine arc, from the posterior pituitary in responseto suckling of the nipple.

While much is known about the cellular mechanismsinvolved in smooth muscle contraction (Rasmussen,1981; Conti & Adelstein, 1981), the mechanisms in-volved in myoepithelial cell contraction remain to befully elucidated. Utilizing myosin light-chain phos-phorylation as an indication of myoepithelial cell con-traction, Olins & Bremel (1982, 1984) concluded thatoxytocin-induced contraction is dependent on cal-modulin activity and the presence of extracellularcalcium. However, it is not evident from these exper-iments what degree of myosin light-chain phos-phorylation is necessary for physiologically significantcontraction.

The fluorescent stain NBD-phallacidin has beenshown to bind specifically to filamentous actin (Baraket al. 1980). Actin occupies much of the myoepithelialcell and the stain has been utilized to visualize the

563

three-dimensional arrangement of myoepithelial cellsin whole tissue (Emerman & Vogl, 1986). In thisstudy, we used NBD-phallacidin to visualize oxytocin-induccd contraction of myoepithelial cells in pieces ofmouse mammary gland tissue. In addition, we usedthis technique to assess the role of extracellular calciumin oxytocin-induced contraction, utilizing the divalentcation chelator EGTA. The possible role of the multi-functional calcium receptor protein, calmodulin, wasstudied by utilizing trifluoperazine (TFP). Aside fromacting as a neuroleptic agent, TFP is also capableof blocking the activation of a variety of enzymesby calcium-calmodulin, including myosin light-chainkinase (Weiss & Wallace, 1980). We also attempted toassess the basic calcium- and ATP-dependent contrac-tile mechanism by solubilization of the cell membraneswith Triton X-100. These investigations show thatNBD-phallacidin is useful for visualizing the effects ofprocesses involved in the myoepithelial cell contractilepathway during physiologically significant contraction.

Materials and methods

Oxytocin-induced contractionMammary glands were obtained from lactating (7-10 days)DD/S mice following cervical dislocation. The tissue wasfinely minced in a control buffer (136mM-NaCl, l-2mM-NaH2PO4, l-2niM-MgSO4) SmM-KCl, 17mM-CaCl2,llmM-glucose, 0-03 mM-Na2EDTA, lOmM-Hepes, pH7-2)for 10 min and then challenged with 50 nM-oxytocin for 2minat 37°C. All chemicals and oxytocin were purchased fromSigma Chemical Co., St Louis, MO, USA, unless otherwiseindicated.

Role of intracellular calcium and ATPMammary glands were minced for 10 min in a control buffercontaining 25mM-Pipes, 50mM-KCl, 5mM-MgCl2, 1 mM-EGTA, soybean trypsin inhibitor (1 nig 1 ~') and 0-125 mM-PMSF (phenylmethylsulphonyl fluoride), p H 6 9 . Tissuewas then incubated for 1 h in 1 % Triton X-100 at 4°C, duringwhich time the medium was changed twice. Tissue was thenincubated for 30 min at 37°C in one of the following: (1)control buffer, (2) control buffer plus 2mM-ATP, (3) controlbuffer plus 2mM-CaCl2 or (4) control buffer, 2mM-CaCl2

plus 2 mM-ATP. The additional calcium would yield, in thepresence of 1 mM-EGTA, a free Ca2+ concentration ofapproximately 1 mM.

Role of extracellular calcium and calmodulinMammary glands were minced in a Ca2+-free buffer(136niM-NaCl, l-2mM-NaH2PO4, l-2mM-MgSO4, SmM-KCl, llmM-glucose, 003 mM-Na2EDTA, 0-3 mM-EGTA,lOmM-Hepes, pi 17-2) for lOmin and then washed witheither TFP plus Ca2+-free buffer or Ca2+-free buffer alonefor 21 min; the medium was changed twice during this time.Tissue was then incubated for 2 min at 37°C in the presenceor absence of TFP in one of the following: (1) control buffer,(2) control buffer plus 50 nM-oxytocin, (3) Ca2+-free buffer,

or (4) Ca +-free buffer plus 50 nM-oxytocin. TFP was fromSmith Kline & French (Canada) Ltd.

Fluorescence microscopyThe tissue was prepared for fluorescence microscopy asdescribed (Emerman & Vogl, 1986). Briefly, it was fixed for30min in 3 7 % paraformaldehyde in phosphate-bufferedsaline (PBS) at room temperature, washed three times withPBS, set on polylysine-coated slides, dehydrated in -20°Cacetone for 5 min and rehydrated in PBS for 10 min. Thetissue was then stained for 20 min in either: (1) PBS (controlfor autofluorescence), (2) PBS plus 165x 10"6M-NBD-phallacidin (stain for filamentous actin), (3) PBS plusl-65xlO"6M-NBD-phallacidin plus 104X 10~4M-phalloidin(competitive specificity control), or (4) PBS plus 1-04X10~4M-phalloidin (control for phalloidin treatment in (3)).The tissue was washed twice with PBS, mounted in glycerol:PBS (1:1, v/v) and squashed under a coverslip.

Slides were photographed on a Zeiss Photomicroscope 111fitted with fluorescein^detecting filters. The Kodak Tri-Xfilm was pushed to 1600 ASA and developed in Acufine.

Electron microscopyTissue was prepared for electron microscopy as described(Emerman & Vogl, 1986). Briefly, tissue was fixed in 1-5%paraformaldehyde/1-5 % glutaraldehyde for 2 5 h, washedthree times with 0*1 M-sodium cacodylate, pH7-3, andfurther processed using standard techniques. Thin sectionswere photographed on a Philips 300 operated at 60 kV.

Results

Oxytocin-induced contraction

Myoepithelial cells from the mammary glands of lactat-ing mice were incubated in the presence or absence ofSO nM-oxytocin for 2 min, then fixed and processed forfluorescence microscopy. The hormone concentrationof 50 nM was chosen because it has been shown to causemaximal phosphorylation of myosin light-chains (Olins& Bremel, 1982). An exposure time of 2 min was usedfor convenience, since we found no difference inmyoepithelial cell appearance when stimulated withoxytocin over a range of 1 to 10 min (data not shown).In the absence of oxytocin, myoepithelial cells did notcontract. They appeared star-shaped, with elongated,thin processes branching out from the cell bodies(Fig. 1A). The processes were of unequal length andsome extensions split into two or three tracts near theirdistal ends. The number and length of the radial armsvaried for different cells. The processes of severalmyoepithelial cells defined the sac-like structure of thealveolus, leaving large spaces between adjacent cells.Staining was uniform throughout the cell, but wassomewhat less intense in the nuclear region. Also,changing the plane of focus revealed that staining wasgenerally uniform throughout the tissue. Althoughstaining was also found in the junctional complexes ofsecretory epithelial cells and blood vessels, these were

564 D. M. Moore et al.

Fig. 1. Myocpithelial cell morphology in mammary gland tissue of a 7- to 10-day lactating mouse as visualized by thefluorescent stain NBD-phallicidin. In A the tissue was minced in a control buffer for lOmin prior to fixation. The cellsdisplay long, thin processes that radiate from the cell body, characteristic of relaxed myoepithelial cells. In B and C, thetissue was exposed to 50nM-oxytocin following mincing and prior to fixation. Myoepithelial cells appear contracted withshort, thick processes extending just beyond the cell body. A, X890; B, X760; C, X810.

easily distinguished from the unique myoepithelialstaining pattern (Emerman & Vogl, 1986). This fluor-escence gave evidence of the well-preserved state of thetissue as the junctions and blood vessels appeareduniform and healthy.

No specific fluorescence was observed in the com-petitive specificity control or in controls for autofluor-escence or treatment with phalloidin.

In contrast to the elongated arms of the relaxed cells,myoepithelial cells appeared as compact spoke-shapedbodies following oxytocin stimulation (Fig. 1B,C).The cells consisted predominantly of brightly fluor-escing cell bodies with short, thick extensions. Pro-cesses near their origin became thickened, so that arapid decrease in arm diameter was observed as onefollowed the extensions away from the cell body. Insome cells, where the terminal portion of the armbranched, there was more intense staining. The de-creased amount of fluorescence in the nuclear regionwas not as obvious in contracted cells. Alveoli appearedto be smaller than those embraced by relaxed myoepi-thelial cells and the cells that defined alveolar shapeappeared much closer together. However, not all cellsin the oxytocin-treated tissue contracted.

Oxytocin-induced myoepithelial cell contraction wasalso observed in thin sections of mammary gland tissueprepared for electron microscopy. Uncontracted myo-epithelial cells were found to underlie the secretoryepithelium at intervals along the basal lamina (Fig. 2).The basal lamina curved gradually with no inden-tations around the half-moon-shaped sections of myo-epithelial extensions. However, in oxytocin-stimulatedtissue, the secretory epithelium appeared to bulge outat its basal pole, leaving deep invaginations in the basallamina where the myoepithelial cell processes were

found (Fig. 3A). Higher magnification revealed thatmyoepithelial cell processes were packed with myofila-ments oriented parallel to the long axis of the processes.Small vesicles were evident immediately adjacent to theplasma membrane (Fig. 3B).

The basic contractile mechanism

Attempts were made to demonstrate a basic ATP- andCa2+-dependent contractile system using methodologydescribed for myoid cells in the seminiferous tubules ofthe testes (Vogl & Soucy, 1985) and intestinal epi-thelium (Hirokawa<?f al. 1983). Triton X-100 was usedto solubilize myoepithelial cell membranes and permitdirect access to the cell interior. The results of theseexperiments proved inconclusive, since cells contractedin the presence of ATP regardless of the presence ofCa2+ (data not shown). Although similar results havebeen reported (Vogl & Soucy, 1985), they radicallycontradict results of other investigators on actomyosincontractile systems (Rasmussen, 1981; Conti el al.1981) and they probably result from an artifact of themethodology. Described below are other studies thatconfirm a role for Ca2+ in myoepithelial cell contrac-tion.

Role of calcium in oxytocin-induced contraction

Myoepithelial cells were incubated in a Ca2+-frcebuffer containing EGTA, to chelate any contaminatingCa2+, and challenged with 50nM-oxytocin. If Ca2+ wasadded back to the medium prior to oxytocin stimu-lation, the tissue had an overall contracted morphology(Fig. 4A). However, morphological differences be-tween tissues in the presence or absence of oxytocinwere negligible if Ca + was not added back to themedium. Cells in both oxytocin-stimulated and un-stimulated tissue in the absence of Ca2+ appeared

Myoepithelial cell contraction 565

relaxed, possessing the characteristic elongated fineprocesses (Fig. 4B). The results suggest that the failureto contract in the presence of oxytocin was due to thedepletion of Ca + in the medium. It must be noted thatthere was some tissue damage, possibly due to thepresence of EGTA, which has been reported to stripproteins from cell membranes (Apgar & Mescher,1986). Consequently, only cells that possessed anordered arrangement of processes and relatively uni-form fluorescence, similar to that of cells not exposed toEGTA, were considered in the results. The prevalenceof contracted cells in Ca2+-resupplied tissue in thepresence of oxytocin was not as great as in stimulatedtissue not exposed to EGTA.

The calmodulin inhibitor, TFP, was also found toblock oxytocin-induced contraction of myoepithelialcells. When TFP was present the myoepithelium ap-

peared relaxed (Fig. 4C). Myoepithelial cells not ex-posed to TFP contracted as described above (Fig. 4A).Consequently, the effect of Ca2+ on oxytocin-inducedcontraction of myoepithelial cells appears to requirecalmodulin to produce a physiologically significantcontraction.

Discussion

Lactating mammals release significant quantities ofoxytocin from the posterior pituitary when suckled bytheir young, causing myoepithelial cell contraction(Lincoln & Paisley, 1982). Yet, when mammary glandsare removed from mice immediately following nursing,no myoepithelial contraction is observed (Emerman &Vogl, 1986). The short interval (12min) between tissueremoval and fixation is enough to allow contracted cells

Fig. 2. Transmission electron micrograph (TEM) of mammary epithelial cells from a 7- to 10-day lactating mouse. Theapical surfaces of cuboidal secretory epithelium (se) face the alveolar lumen (/). Myoepithelial cell processes are interposedbetween the basal surface and the basal lamina (arrowheads). X 10500.

566 D. M. Moore et al.

to relax. However, we have found that myoepithelialcells can be induced to contract following a 1-minexposure to 50 nM-oxytocin (data not shown) whentissue fixation occurs immediately after oxytoein stimu-lation. Contraction is still observed after a 10-minincubation in 50nM-oxytocin. It has been demon-strated that myosin light-chain phosphorylation dropsbelow 50 % of the initial rise when exposed to oxytoeinfor lOmin (Olins & Bremel, 1984). This discrepancy

points to the basic difficulty with the phosphorylationassay. It is not clear what level of myosin light-chainphosphorylation is required for myoepithelial cell con-traction to occur. NBD-phallacidin staining of cells intissue blocks, however, seems an accurate indicator fordetermining what processes are important in milkejection. Using the stain, we can visualize myoepi-thelial cell contraction. These observations are con-firmed by electron microscopy.

Fig. 3. TEM of mammary epithelial cells from a lactatingmouse following exposure to 50 nM-oxytocin. In A, thesecretory epithelium (,v<?) bulges out basally, leavinginvaginations along the basal lamina where myoepithelial cellprocesses are found (arrowheads). At higher magnification(B), a myoepithelial cell process is shown to be packed withmicrofilaments and possess pinocytotic vesicles near the cellsurface. A, X5800; B, X39 800.

Myoepithelial cell contraction 56?

Not all cells within oxytocin-stimulated tissue con-tract, even when not exposed to EGTA. Several factorscould account for this observation: (1) oxytocin is notreaching these cells; (2) these cells are dead or dying;or (3) these unresponsive cells lack available oxytocinreceptors. The first explanation seems unlikely becauseuncontracted cells can be found at the tissue peripherywhere exposure to oxytocin is certain. While we cannotcompletely rule out the second possibility, it seemsunlikely, since the conditions to which the tissue wasexposed were quite mild and would not be expected tocause such extensive cell death. Consequently, it islikely that the cells that failed to contract followingexposure to oxytocin lack the appropriate receptors.

It has been shown that, while all cells within a singlealveolus synthesize milk components in synchrony,different alveoli produce milk at different times (Saake& Ilcald, 1974). If the changing plasma level ofoxytocin were the only means of regulating myoepi-thelial cell contraction, all cells would contract at once,even if all the alveoli had not produced significantquantities of milk. To avoid this unnecessary expenseof energy, myoepithelial cell contraction may also beregulated by the presence of available oxytocin recep-tors, so that only those cells embracing alveoli that haveproduced milk components will contract. Proof of thishypothesis awaits further study.

When myoepithelial cell membranes are solubilizedin Triton X-100, contraction can be induced merely byexposure to ATP, regardless of the presence or absenceof calcium. Similar observations have been reportedfrom experiments on myoid cells of the seminiferoustubule (Vogl & Soucy, 1985) and do not seem to be dueto contaminating calcium in Ca +-free solutions. Evenafter exposure to 20mM-EDTA, contraction withoutadded calcium can be found. We suspect that this maybe due to proteolytic cleavage of the mvosin light-chainkinase by proteases released from cellular lysosomesfollowing membrane dissolution. Proteolysis has beenshown to render the kinase independent of calcium-calmodulin (Tanaka et al. 1980). However, theprotease inhibitors, PMSF and soybean trypsin inhibi-tor, have no effect in preventing contraction. Becausethese observations clearly contradict other investi-gations, a different approach to studying a calciumeffect was undertaken.

Olins & Bremel (1982, 1984) reported that onlytransient myosin light-chain phosphorylation occurs

Fig. 4. Myoepithelial cell appearance following exposure toEGTA. In A, tissue was exposed to oxytocin in thepresence of Ca2+, permitting the cells to contract. In B,the cells, exposed to oxytocin in Caz+-free medium, do notcontract. The cells also appear relaxed in C, when cellswere exposed to oxytocin in the presence of Ca2+, but inthe presence of the calmodulin antagonist TFP. A, X730;B, X810; C, X800.

when myoepithelial cells are challenged with oxytocinin Ca2+-free media. NBD-phallacidin visualizationillustrates the fact that in the absence of extracellularcalcium the cells do not contract. The presence ofEGTA, which has been shown to strip membraneproteins (Agpar & Mescher, 1986), causes some tissue

568 D. M. Moore et al.

damage. However, the majority of myoepithelial cellsremain healthy and are relaxed in Ca2+-free mediawhether challenged by oxytocin or not. It appears,therefore, that intracellular stores of calcium are notadequate to induce a contractile response in myoepi-thelium.

TFP has been reported to block myosin light-chainphosphorylation in oxytocin-exposed myoepithelialcells (Olins & Bremel, 1982, 1984). In these studies italso prevents the contractile morphology. This indi-cates a vital role for calmodulin in the myoepithelial cellcontractile pathway. The influx of extracellular calciumby itself does not cause contraction; calmodulin bind-ing is necessary for calcium to exert its effect. Thisremains consistent with the smooth muscle system,where calmodulin activation by an influx of extracellu-lar calcium causes myosin light-chain kinase activation(Rasmussen, 1981). The activated kinase then pro-ceeds to catalyse myosin light-chain phosphorylation sothat actin-induced ATPase activity will result in con-traction. Whether the latter steps of the pathway arealso found in the myoepithelial cell remains to beclarified. NBD-phallicidin may be a useful tool forvisualization of these processes.

This research was supported by grants from the NationalCancer Institute of Canada and the Medical Research Coun-cil. J. T. Emerman is a research scholar of the NCIC. D. M.Moore was a recipient of an NSERC studentship. Theauthors thank Dr Nadine Wilson and Dr Vladimir Palaty fortheir helpful advice during the course of this work and MrsJudy Bysouth for preparation of this manuscript.

References

APGAR, J. R. & MESCHER, M. F. (1986). Agorins: majorstructural proteins of the plasma membrane skeleton ofP815 tumour cells. J . Cell Biol. 103, 351-360.

BARAK, L. S., YOCUM, R. R., NOTHNAGEL, E. A. & WEBB,

W. W. (1980). Fluorescence staining of the actincytoskeleton in living cells with 7-nitrobenz-2-oxa-l,3-diazole-phallicidin. Proc. naln. Acacl. Sci. U.SA. 77,980-984.

BREMEL, R. D. & SHAW, M. E. (1978). Actomyosin frommammary myoepithelial cells and phosphorylation bymyosin light chain kinase. J. Daily Sci. 61, 1561.

CHACKO, S., CONTI, M. A. & ADELSTEIN, R. S. (1977).

Effect of phosphorylation of smooth muscle myosin onactin activation and calcium regulation. Pmc. natn.Acad. Sci. U.SA. 74, 129.

CONTI, M. A. & ADELSTEIN, R. S. (1981). The relationship

between calmodulin binding and phosphorylation ofsmooth muscle myosin kinase by the catalytic subunit of3', 5'-cAMP dependent protein kinase. jf. biol. Cliem.256, 3178-3181.

EMERMAN, J. T. & VOGL, A. W. (1986). Cell size andshape changes in the myoepithelium of the mammarygland during differentiation. Anal. Rec. 216, 405-415.

HIROKAWA, N., KELLER, T. C. S. Ill, CHASAN, R. &

MOOSEKER, M. S. (1983). Mechanism of brush bordercontactility studied by quick freeze, deep etch method.,J.Cell Biol. 96, 1325-1336.

LINCOLN, D. W. & PAISLEY, A. C. (1982). Neuroendocrine

control of milk ejection. J. Repmcl. Fert. 65, 571-586.OLINS, G. M. & BREMEL, R. D. (1982). Phosphorylation of

myosin in mammary myoepithelial cells in response tooxytocin. Endocrinology 110, 1933-1938.

OLINS, G. M. & BREMEL, R. D. (1984). Oxytocinstimulated myosin phosphorylation in mammarymyoepithelial cells: Roles of calcium ions and cyclicnucleotides. Endocrinology 114, 1617-1626.

RASMUSSEN, H. (1981). Smooth muscle contraction. InCalcium and cAlVlP as Synarchic Messengers, pp.257-265. Toronto: Wiley Interscience.

SAAKE, R. G. & HEALD, C. W. (1974). Cytological aspectsof milk formation and secretion. In Lactation: AComprehensive Treatise, vol. II (ed. B. L. Larson & R.S. Vearl), pp. 147-189. New York: Academic Press.

SOLOFF, M. S. (1976). Oxytocin receptors in the mammarygland and uterus. In Methods in Receptor Research (ed.M. Belcher), p. 511. New York: Marcel Decker.

TANAKA, T., NAKA, M. & HIDAKA, H. (1980). Activationof myosin light chain kinase by trypsin. Biochem.biophys. Res. Commun. 92, 313-318.

VOGL, A. W. & SOUCY, L. J. (1985). Arrangement andpossible function of actin filament bundles in ectoplasmicspecializations of ground squirrel Sertoli cells. .7. CellBiol. 100, 814-828.

WEISS, B. & WALLACE, T. L. (1980). Mechanisms andpharmacological implications of altering calmodulinactivity. In Calcium and Cell Function, vol. 1 (ed. W.Y. Cheung), pp. 329-379. New York: Academic Press.

(Received 28 Mav 1987 - Accepted, in revised fonn,14 August 1987)

Myoepithelial cell contraction 569