Milk lipid and protein traffic in mammary epithelial cells

Review

Milk lipid and protein traffic in mammary epithelialcells: joint and independent pathways

Michèle OLLIVIER -BOUSQUET*

Unité de Biologie Cellulaire et Moléculaire, INRA, 78352 Jouy-en-Josas Cedex, France

Abstract — In mammary epithelial cells, milk lipids and proteins are synthesised in the same com-partment, the endoplasmic reticulum. Lipids, carried through the cytoplasm, associate with the api-cal membrane which then pinches off and releases the lipid globule. Proteins, carried through mem-brane compartments are released in the lumen after fusion of secretory vesicles with the apicalmembrane. These processes assure a relatively constant composition of milk but it is not knownwhether lipid and protein secretion are linked. The protein composition of the milk fat globule mem-brane and the stimulatory effects of prolactin and oxytocin on lipid and protein secretion suggestthat these processes are coupled and co-regulated. However, it is possible to observe a dissociationbetween the formation and the secretion of the two constituents, during differentiation and in variousexperimental conditions, and this suggests that coupling is not strictly required.

milk lipids / mammary epithelial cells / secretion

Résumé — Transport des lipides et des protéines du lait dans les cellules épithélialesmammaires.Dans la cellule épithéliale mammaire, les lipides et les protéines du lait sont synthéti-sés dans le même compartiment, le reticulum endoplasmique. Les lipides transportés à travers lecytoplasme s’accolent à la membrane plasmique apicale qui les enveloppe avant leur sécrétion. Lesprotéines, transportées dans des compartiments membranaires jusqu’aux vésicules de sécrétion, sontlibérées dans la lumière après fusion des vésicules avec la membrane apicale. Ces processus assurentune composition relativement constante du lait. Cependant, on ne sait pas si ces deux types de sécré-tion sont liés. La composition protéique de la membrane du globule gras et l’effet stimulant de la pro-lactine et de l’ocytocine sur la sécrétion des deux constituants suggèrent que synthèse et sécrétion sontliées et co-régulées. Cependant, la possibilité d’observer une dissociation entre la formation et lasécrétion des deux constituants suggère que ce couplage n’est pas obligatoire.

lipides du lait / cellule épithéliale mammaire / sécrétion

Reprod. Nutr. Dev. 42 (2002) 149–162 149© INRA, EDP Sciences, 2002DOI: 10.1051/rnd:2002014

* Correspondence and reprintsE-mail: [email protected]

M. Ollivier-Bousquet150

1. INTRODUCTION

Unresolved questions continue to intriguemammary cell biologists, dairy scientistsand technologists: Are the intracellularsecretion processes of lipids and proteins(the two main components of milk) linked,and are they controlled by the same mecha-nisms.

Milk composition has been the subjectof numerous studies, and in fact it is moreappropriate to speak of the composition of“milks” in the plural to reflect the great vari-ety of milks found in different mammals[26]. For example, average fat and caseinconcentrations are 183 g.L–1 and 104 g.L–1

in the rabbit, 38 g.L–1 and 4 g.L–1 in humansand 19 g.L–1 and 13 g.L–1 in the horse,respectively [16]. Milk composition alsovaries according to time since birth(colostrum versus mature milk). Breed,number of lactations, diet and season alsocause variations in milk composition [1, 44,45]. However, these variations are adaptedto the physiological status of the offspringand, as long as the mammary gland ishealthy, they are strictly regulated. Duringestablished lactation, the concentration ofcaseins gradually increases and, in mostspecies, milk fat concentration increases inparallel [16]. In the rat, fat concentrationdeclines in early lactation, but after the firstfive days following birth, once lactation isestablished, both fat and protein concentra-tions remain constant [45].

One approach to the questions of inter-actions between the secretion of milk lipidsand of proteins has been to describe the mor-phological aspects of each type of productsecretion (Fig. 1). In mammary epithelialcells (MEC), lipid droplets are surroundedby a layer that is largely composed of pro-teins and polar lipids. Caseins, in filamen-tous form or aggregated in micelles, are con-tained in secretory vesicles. Sometimes,lipid droplets are completely surrounded bysecretory vesicles. Based on this observa-tion, it has been proposed that there is inter-

action between the secretory vesicle and thelipid droplet [69]. But this mechanism mightbe limited to periods when milk secretionis inhibited (discussed in [27]). Lipid glob-ules in milk differ from intracellular lipiddroplets in that they have a membrane, themilk fat globule membrane (MFGM), whichis derived from apical membrane. Caseinmicelles have the same honeycomb appear-ance in milk as in secretory vesicles.

Another approach to the question of inter-actions between secretion of lipids and pro-teins has been to study their secretory path-ways. The cellular mechanisms for milkcomponent transport by the MEC have beenrecently reviewed [7, 27, 64]. Three mainintracellular pathways have been described:

(1) A pathway ending with exocytosis:This pathway involves the transport ofnewly synthesised proteins from the endo-plasmic reticulum (ER) to the Golgi appa-ratus and through the trans Golgi network(TGN) to the secretory vesicles. Secretoryvesicles containing mature caseins, wheyproteins, lactose, citrate and soluble salts,release their contents into the lumen of theacinus by exocytosis [7, 38]. In addition, asmall quantity of newly synthesised pro-teins, including milk proteins, hormonereceptors and enzymes, is carried to thebasolateral membrane [17].

(2) An endocytotic and transcytotic path-way involving internalisation of plasma-borne components into the cell and theirtransport into the milk [50]: during trans-port, a sorting of internalised moleculesstrictly controls the destination of the tran-scytosed products [41, 61].

(3) Milk lipid droplets are carried fromthe ER where they are synthesised to theapical region; they then bud from the api-cal membrane and are released into thelumen enveloped in the MFGM [27].

Despite all this information, it is not yetpossible to describe how the MEC coordi-nates the synthesis and secretion of milkcomponents to produce milk with a constantcomposition.

Milk lipid and protein traffic

Control of synthesis of milk lipids andproteins may occur at the transcriptionallevel and involve control of gene activation.There remains the question of whether ornot the genes involved in milk lipid and milkprotein synthesis are activated simultane-ously during final differentiation of theMEC. The expression of genes coding forproteins involved in the synthesis of pro-teins in the ER may have repercussions onthe synthesis of lipids in that organelle. Onthe other hand, metabolic control of theexocytosis of secretory vesicles may beresponsible for replenishing the apical mem-brane by secretory vesicle membranes and

The important role of one gene, whichcodes for a key enzyme in lipid synthesis,has recently been reported [65]: inactiva-tion of the gene for acyl-CoA: diacylglycerolacyltransferase (DGAT), an enzyme thatcatalizes the final esterification reaction inthe biosynthesis of triacylglycerol (TAG),results in mice which do not lactate.Whereas these DGAT-null mice have nor-mal serum lipid levels, no TAG synthesisis detectable in mammary tissue during lac-togenesis. These results suggest that expres-sion of one gene coding for a key enzyme inlipid synthesis is necessary for lactogene-sis and subsequent milk secretion.

151

Figure 1.Morphological aspects of lactating mammary epithelial cells, milk lipid globule and caseinmicelles. (a) Lactating rabbit mammary epithelial cell contains well developed endoplasmic reticu-lum (ER), Golgi apparatus (G), secretory vesicles containing casein micelles (sv), and lipid droplets(ld). Some secretory vesicles are associated with lipid droplets (arrows). In the lumen (L), lipid glob-ules and casein micelles are detectable. N: nucleus; my: myoepithelial cells. Bar 2 mm. (b) Scanningelectron micrograph of sheep milk lipid globule. Bar 0.5 mm. (c) Scanning electron micrograph of sheepmilk casein micelles. Bar 0.5 mm.

M. Ollivier-Bousquet

consequently be favourable to the releaseof lipid globules enveloped in the MFGM.

In this review, after a brief descriptionof the formation of milk lipid globules, wewill compare the expression of milk lipidsand proteins during differentiation of MECand the regulation of their intracellulartransport.

2. FORMATION OF MILK LIPIDGLOBULES

Recent reviews have given exhaustivehistorical accounts of research on milk fatglobules, their surrounding membranes, andtheir origin, growth and secretion [27, 39].In brief, the first to observe fat globules ofmicroscopic dimensions in milk was vanLeeuwenhoeck in 1674. The description ofa membrane composed of phospholipidsand proteins, stabilising the milk lipid glob-ule, was confirmed by electron microscopeobservations, which showed the lipid dropletenveloped in the plasma membrane in theapical cell region [5]. The composition ofMFGM was described in biochemical stud-ies [56]. Since these pioneering studies,numerous data have made it possible todescribe general aspects of how milk lipidglobules are formed and secreted. The fattyacids that compose the TAG, are either sup-plied exogenously from plasma lipids orsynthesised de novo via acetyl-CoA car-boxylase and fatty acid synthetase [12].Small lipid droplets, known as microlipiddroplets, may originate from the ER whereacyltransferases have been localised by cellfractionation. Microlipid droplets form inor on the ER and are released into the cyto-plasm (reviewed in [39]). A protein identi-fied as adipocyte differentiation-related pro-tein (ADRP) was found in homogenates oflactating mammary gland, in ER and in lipiddroplet fractions [23]. It was then suggestedthat this protein might be involved in thedeposition of TAG in droplets as it occurs inadipocytes. For other cell types, the mech-anisms of lipid droplet formation in the

bilayer of the ER membrane and the pro-cesses of storage in the cytoplasm have beendescribed [10, 42]. For example, inadipocytes and certain plant seeds, lipiddroplets are believed to form by coalescenceof neutral lipids into discs inside the bilayerof the ER membranes. The discs grow intospheres and may bud from the ER into thecytoplasm as droplets surrounded by a phos-pholipid monolayer derived from the ER[10]. In adipocytes and steroidogenic cells,ADRP is mainly expressed during differ-entiation and another protein, perilipin, isexpressed during the lipid storage phase. Ithas been suggested that perilipin might ster-ically block access of the hormone-sensi-tive lipase, which catalyses the rate-limit-ing step in TAG hydrolysis, to the dropletsurface [9, 33]. It is not yet known whetherthere are similar mechanisms in MEC forregulating the growth or lipolysis of the lipiddroplet. However, ADRP is always presentin the MFGM after lipid globules have beensecreted into milk [23], suggesting that thelipid droplet storage mechanisms here arenot the same as those described inadipocytes.

Caveolins, a family of membrane-pro-teins, form the coatings of caveolae that aregenerally restricted to cell surface and Golgiapparatus. Caveolins contain NH2- andCOOH-terminal domains that project intothe cytosol with a central domain in the lipidphase of the membrane. It has been shownthat, when caveolin-1 and, to a lesser extent,caveolin-2 accumulate in the ER, they canbe found in association with the membranessurrounding lipid droplets [54] and may playa part in maintaining cholesterol balance[59]. Moreover, the presence of caveolin-2in a raft-like membrane domain on the lipiddroplet surface may be involved in intra-cellular signalling [19]. These authors positthat caveolin plays a role in controlling lipiddroplet transport and storage. Caveolin hasbeen detected in lactating mouse MEC [55].It will be interesting to localise caveolin inMEC under different physiological or experi-mental conditions, to see whether there is a

152


hypothetical. Droplet diameters vary from0.5 mm or less to 8 mm or more. Microlipiddroplets may fuse with preformed cyto-plasmic lipid droplets [18, 67]. Fusionbetween large droplets has not beenobserved. Incorporation of fatty acid pre-cursors of TAG from rabbit and sheep MEC,monitored by electron microscope autora-diography, shows that fatty acids are incor-porated in a few minutes in lipid dropletsof large diameter and are very quicklyreleased into the lumen [15]. Stembergerand Patton [66] have suggested that controlof the growth process may occur after thedroplet has made contact with the apicalmembrane. Altogether, these results sup-port the hypothesis of continuous growth ofdroplets simultaneously with release intothe alveolar lumen.

Lipid globules in the milk of at least somespecies vary considerably in size [28]. It isnot known whether some smaller dropletshave been secreted or whether largerdroplets have broken up during or aftersecretion. However, each mammalianspecies tends to accumulate milk lipids indroplets of a relatively narrow size range,suggesting that size is regulated.

The fusion of droplets with the apicalmembrane is a key mechanism in the pro-cess of milk lipid secretion. In cytoplasm,lipid droplets are associated with ADRP. Inmilk, lipid globules contain abundant pro-teins associated with the MFGM, includingbutyrophilin and xanthine oxidase [39].Butyrophilin, a membrane glycoprotein,accumulates close to the apical membrane ofthe MEC [3]. It has been suggested that xan-thine oxidase, a soluble protein, binds to thecytoplasmic domain of butyrophilin asso-ciated with the plasma membrane, and thatthis scaffold may act as a receptor for bind-ing the cytosolic droplet through interac-tion with ADRP until the lipid droplet iscompletely enveloped in the apical mem-brane, which then pinches off and releasesthe milk lipid globule [39]. There is pre-sumed to be coordination between the two

similar interaction between caveolin andlipid droplet surfaces in MEC.

Many other ER membrane and luminalproteins have been identified in MFGMsand in lipid droplets [22, 70]. Some ERluminal proteins (calreticulin, protein disul-fide isomerase, immunoglobulin-bindingprotein), are associated with intracellularlipid droplets but are not detectable inMFGM, suggesting that they are lost dur-ing secretion [22]. Taken together, theseexamples suggest that ER proteins may beinvolved in lipid droplet formation or trans-port. Interestingly, a reduced level of caseinand lipid secretion has been described ingoats with mutations of the as1-casein gene([43] in this issue and reference therein). Inthese animals, very low levels of as1-caseinsare synthesised and the transport of the othercaseins through the ER is greatly delayed[11]. It is not known whether these abnor-malities in the ER are responsible for thelow milk lipid concentration found in theseanimals.

Cytosolic proteins known to be relatedto membrane trafficking have also beendetected in lipid droplets and MFGM [70].A cytoplasmic protein called TIP 47, whichfunctions as a receptor for mannose 6-phos-phate receptors by binding to the cytoplas-mic domain of these receptors, has a struc-ture homologous to that of ADRP (43%identical). It has therefore been suggestedthat this protein might be involved in bothmembrane vesicle traffic and lipid droplettraffic [68]. Controversial results obtained indifferent cell types have been published [4].However, using a specific antibody raisedagainst a N-terminal peptide sequence ofTIP 47, Heid and Keenan (unpublishedresults) showed that TIP 47 was associatedwith MFGM. A cytoplasmic protein inter-acting both with the lipid droplet surfaceand with secretory vesicles might explainthe association of fat droplets and secretoryvesicles described above.

The mechanisms of lipid droplet growthand translocation within the cell remain

153


secretion processes – that of the aqueousphase constituents of milk (lactose, water,ions, whey proteins and caseins), by exo-cytosis after the secretory vesicles havefused with the plasma membrane, and that ofthe fat globules, by envelopment in the api-cal plasma membrane. Among the proteinmarkers of the apical membrane, MUC1, aglycoprotein of the mucin family, is presentin large amounts in MFGM [35, 36]. Immu-noelectron microscopy localisation ofMUC1 in lactating guinea pig MEC showsthat this mucin is concentrated on the apicalsurface of the mammary epithelial cells andon secreted fat globules. By contrast, secre-tory vesicle membranes show relatively lowconcentrations of MUC1 [37]. These obser-vations suggest that secretory vesicle mem-brane does not directly give rise to the api-cal membrane from which MFGM isderived. To know whether proteins are redis-tributed after fusion of the secretory vesi-cles and apical plasma membrane, the pro-tein composition of these membranes willhave to be described.

3. FORMATION OF LIPIDDROPLETS IN MAMMARYEPITHELIAL CELLS AT VARIOUSSTAGES OF DIFFERENTIATION

3.1. Differentiation during pregnancy

The timing and sequence of specificdifferentiation events differ according tospecies and breed [34, 71]. In vivo, it seemsthat MEC are able to express casein mRNAsas soon as they become organised in buds. Inthe rabbit mammary gland, the level ofmRNAs coding for as1-casein, b-casein,k-casein and WAP increases as pregnancyproceeds and is relatively high during latepregnancy [25, 60]. Milk products are syn-thesised in MEC at the end of pregnancy,although milk synthesis remains moderate.However, the time at which morphologi-cally detectable secretory products appearin the cytoplasm of the MEC varies from

species to species [58]. The changes observedin rabbit mammary gland during pseudo-pregnancy and pregnancy illustrate well thesequence in which milk products appear inthe MEC. The morphological appearanceof the MECs shows that in early pregnancy(9th day), the cytoplasm contains a fewsmall lipid droplets. The Golgi apparatus isreduced to small stacks and no secretoryvesicles or casein micelles are detectable.The lumen of the acini contains granularand filamentous products. At days 14–19of pregnancy or at 12–14 day of pseudo-pregnancy, the cytoplasm has a similarappearance but the lipid droplets are largerand more abundant and the lumen is filledwith an homogenous electron-dense product[8]. In MEC from some rabbits, at this stage,a few casein micelles are detectable in thecytoplasm. However, instead of beinglocalised in secretory vesicles filled with alucent content as observed during lactation,they are surrounded by a closely apposedmembrane. The Golgi apparatus is com-posed of a small number of stacked sac-cules. These observations show that duringthe early stages of pregnancy, although milkprotein mRNAs are present, caseins are notdetectable in micellar form. Secondly, itappears that lipid droplets are present earlierin pregnancy than casein micelles. It is worthnoting that, at the end of pregnancy whenthe secretory processes are not yet wellestablished, very large lipid droplets accu-mulate inside the cell and are also found inthe lumen of the acini.

3.2. In vitro differentiation

HC11 cells, a clonal derivative of theCOMMA-1D cell line, have been used as amodel for studying hormonal regulation ofmammary cell differentiation [2]. Prolactin(PRL) and cortisol regulate the expression ofb-casein and of xanthine oxidoreductase, aconstituent of the milk lipid globule [40].Epidermal growth factor blocks the stimu-lation of b-casein by PRL and cortisol, but

154


in the presence of these hormones itincreases the expression of xanthine oxi-doreductase. Lactogenic hormones stimu-late xanthine oxidoreductase by a signallingpathway involving the MAP kinase path-way, whereas the JAK/Stat pathway is themain pathway involved in the activation ofb-casein expression [40].

Mammary epithelial cells from a 12-daypseudo-pregnant rabbit, cultured in twochambered cell culture plates, becamepolarised. After 12 days of culture in thepresence of PRL, insulin and glucocorti-coids, numerous large lipid droplets werepresent in the cytoplasm and the apical regionof the cells, partly surrounded by the apicalmembrane (Fig. 2). No casein micelles weredetectable in these cells. Thus, in vitro underhormonal control, milk lipid synthesis andmilk lipid droplet formation occur beforecasein micelle formation, just as in vivo dur-ing pregnancy. It may be that lipid dropletsand casein micelles are formed indepen-dently. The lipid droplets reach very largediameters in these conditions, suggestingthat they continue to grow and are not effi-ciently secreted.

3.3. Inhibition of protein synthesisand transport in fully differentiatedand actively secreting MEC

One way to study whether milk proteinand lipid secretion are associated is to inhibitsynthesis or transport of one and measureany effect this has on the secretion of theother. Unfortunately, to our knowledge, nodata on this are available in the literatureso far. Inhibition of protein synthesis bycycloheximide added in vitro to mammary

155

Figure 2.Mammary epithelial cells from a 12 day pseudo-pregnant rabbit cultured on two chamberedcell culture plates. (a) Cells were cultured for three days in the presence of insulin plus cortisol. A fewsmaller lipid droplets were detectable in the cytoplasm (arrow). (b) Cells were cultured for threedays in the presence of insulin plus prolactin plus cortisol. Large lipid droplets are present in the cyto-plasm (arrow). Numerous lipid droplets bind at the apical membrane (arrowheads). No caseinmicelles are detectable. In some regions of the sheet, cells are surrounding a matrix-like formation(asterisk). Bar 10 mm. (c) Scanning electron micrograph of protruding lipid droplets on the apical sideof the cells. Bar 5 mm.


fragments induced an inhibition of 80–90%of casein secretion in one hour. Morpho-logical observations revealed that, in theseconditions, very few secretory vesicles con-taining casein micelles remained detectablewhereas numerous small, empty vesicleswere present in the cytoplasm. It can be sup-posed that pre-existing secretory vesicleshad released their contents during the one-hour treatment. On the other hand, lipiddroplets of various sizes were still present inthe cytoplasm (Ollivier-Bousquet, unpub-lished results). This suggests either that lipidsynthesis continues in the absence of pro-tein synthesis or that previously synthesisedlipids accumulate in the cytoplasm and arenot released into the lumen when proteinsynthesis is inhibited.

Protein secretion is inhibited in MECin vitro by adding brefeldin A (BFA), a drugthat disorganises the Golgi region [57]. After60 min’ incubation in the presence of BFA,Golgi stacks and secretory vesicles havecompletely disappeared and the ER fills thecytoplasm. Lipid droplets of various sizesare present in the cytoplasm. Large lipiddroplets are in close contact with the apicalmembrane. Numerous lipid globules aredetectable in the lumen (Fig. 3). Thus,although the secretory vesicles have disap-peared, lipid droplets are still present, aswas observed after protein synthesis inhi-bition. To draw any conclusions as to theimportance of the integrity of Golgi appa-ratus for lipid droplet transport, one wouldhave to determine whether these lipids pre-existed the action of the BFA or are newlysynthesised.

4. REGULATIONOF INTRACELLULARTRANSPORT AND SECRETIONOF LIPID DROPLETS

Between two nursings or milkings, milkconstituents are released from their sites ofsynthesis in MEC into the lumen of the aciniby continuous secretion. In addition, at each

nursing or milking, pituitary hormones suchas PRL and oxytocin are discharged. Thisregular peaking of PRL and oxytocin cir-culation is necessary to maintain the syn-thetic activities of the MEC and empty theacini.

The binding of PRL to its receptors hasseveral biological effects, including the con-trol of gene expression [24] and a metaboliceffect, the secretagogue effect, which has

156

Figure 3. Morphological aspects of rabbit lac-tating mammary epithelial cells incubated in thepresence of brefeldin A. (a) Section of an aci-nus after 5 min incubation in the presence of5 mM brefeldin A. The cytoplasm is filled withendoplasmic reticulum. Golgi apparatus andsecretory vesicles are not detectable. Small sizelipid droplets are present in the cytoplasm(arrow). Bar 10 mm. (b) Electron micrograph ofrabbit lactating mammary epithelial cells after60 min of incubation in the presence of5 mm brefeldin A. The cytoplasm is filled withendoplasmic reticulum (ER). Lipid droplets arepresent in the cytoplasm (arrow), in the apicalregion of cell (arrowheads) and in the lumen (L).Bar 1 mm.


Oxytocin provokes a contraction ofmyoepithelial cells and ejection of milk fromthe mammary acini. Regular removal of milkfrom the mammary gland due to an increasein the number of milkings or to exogenousoxytocin administration, increases milk pro-duction ([31] in this issue and referencestherein). This increase is accompanied byqualitative changes in milk composition dueto better alveolar milk transfer. Oxytocinstimulates release of TAG from rat mam-mary gland slices [14]. It has been suggestedthat oxytocin may directly act on intracel-lular transport of newly synthesised proteinsin the MEC [48] by an effect on epithelialcells [32].

These results show that PRL, and possi-bly oxytocin, exert direct effects on themammary epithelium, stimulating milk lipidsecretion. The mechanisms of these hor-mones’ action on lipid secretion are notknown. Monomeric GTP-binding proteinsare key regulator of vesicular transport ineukaryotic cells [63]. A class of low molec-ular mass GTP-binding proteins are tightlyassociated with the lipid droplet surface andwith the MFGM. These GTP-binding pro-teins co-enrich with xanthine oxidase andbutyrophilin during purification, suggest-ing that they may be associated with thecoat structure located on the inner face ofthe MFGM. These GTP-binding proteinsalso play a role in the transduction of varioussignals. It has been postulated that they mayalso play role in transduction of the signalneeded for regulating vectorial transport oflipid droplets. In agreement with thishypothesis, adding GTPgs to the incubationmedium of permeabilised acini increasesthe release of newly synthesised lipids [21].Identifying the factors that regulate the activ-ity of GTP-binding proteins associated withlipid droplets and MFGM will provide anunderstanding of the role of these proteins.

A b subunit of a trimeric G proteinhas been found in MFGM but not in lipiddroplets [70]. Trimeric G proteins areclassically known to be involved in the

been described in vitro [51]. This secreta-gogue effect produces a very rapid increaseof milk protein secretion [17, 29, 49]. PRLsignalling for the secretagogue effectinvolves a rapid but transient release, andmetabolism, of arachidonic acid [6]. Lipoxy-genase metabolites of arachidonic acid maybe involved in interactions and fusionsbetween intracellular membranes. Conse-quently, lipid secretion, which depends onthe lipid globule being enveloped by theapical membrane, may also be affected bythe increased metabolism of arachidonicacid induced by PRL. In fact, in vitro, PRLstimulates lipid secretion in the same way asprotein secretion [15]. Metabolic labelling ofmilk lipids in a 3-min pulse, followed by a120-min chase, revealed that 98% of theTAG in the intracellular lipids was radioac-tive compared to 71% of the TAG in thelipids secreted into the medium. PRL addedafter the pulse increased the radioactivityof TAG in lipids secreted into the medium.Morphological evaluation of the distribu-tion of lipid globules between the cytoplasmand the lumen of the acinus showed that inthe presence of PRL, more lipid globulesare secreted into the lumen than into a con-trol medium [52]. These results suggest thatPRL stimulates the secretion of growinglipid droplets enriched with newly synthe-sised radioactive triglycerides, whereas, inMEC incubated in a control medium, grow-ing lipid droplets accumulated in the cyto-plasm for a longer period. Incorporation intothese lipids of non radioactive, newly syn-thesised triglycerides may continue duringthe chase; this would explain the lower per-centage of radioactivity in the triglycerides.This hypothesis is in accordance with thesuggestion that growth, envelopment in theapical membrane and release of the lipiddroplet occur simultaneously. If this is thecase, PRL, by stimulating lipid secretion,might increase the release of smaller lipidglobules. To verify this hypothesis, mor-phological studies to assess the diametersof released lipid globules will be required.

157


transduction of various signalling systems,and could play a regulatory role in the laststages of fusion at the apical plasma mem-brane.

Transport within MECs involves the ele-ments of the cytoskeletal system. Actin andmicrotubules constitute a network that isrelated to the polarity of the epithelium [30].Actin is located at the periphery of the cell,in close apposition to the basolateral andapical membrane. Microtubules are foundin the apical and medial portions of cyto-plasm, associated with vesicles [46].Colchicine and other microtubule-alteringdrugs inhibit the secretion of milk compo-nents both in vivo and in vitro (referencesin [30]). Colchicine produces cytoplasmicdisorganisation and loss of microtubulesand causes intracellular accumulation ofsecretory vesicles followed by a decreasein protein, lactose and lipid secretion [47].When measured in vitro, colchicine added toan incubation medium did not modify thesecretion of radioactive triglycerides labelledin a 3-min pulse followed by a 120-minchase, whereas secretion of newly synthe-sised proteins was completely inhibited insimilar experimental conditions [15, 53].This suggests that at least during the time-window studied, and with the concentra-tions of colchicine used in these experi-ments, intracellular transport of proteinsrequires the microtubule system, while thatof lipids does not. In agreement with this,it is worth noting that no important associ-ation between microtubules and lipiddroplets has been detected in morphologicalstudies [46]. On the other hand, proteinssuch as dynein, motor protein, gelsolin andgephyrin, which are known to be associatedwith microtubules and actin and to beinvolved in the binding and movement ofvesicular cargoes, have been found in iso-lated MFGMs and mammary lipid droplets[70]. It is not clear by what mechanism thecytoskeletal system might be involved inthe regulation of lipid droplet transport.

5. CONCLUSIONSAND PERSPECTIVES

Lipid droplets and casein micelles in theirmature forms do not appear simultaneouslyduring pregnancy and during differentiationof MEC in vitro. During established lacta-tion, secretion of the two constituents is par-allelly regulated. However, it may be pos-sible to dissociate protein secretion and lipidsecretion in various experimental condi-tions. Answering questions as to possibleinteraction between these two types of secre-tion requires knowledge of the intracellulartransport mechanisms involved.

The secretory pathways of milk con-stituents have been well described. The lac-tating MEC conveys newly synthesised milkproteins by a process involving consider-able membrane flow, which is used now asa model of cisternal maturation trafficking[13, 20, 38]. Another type of endomembranetrafficking involves endocytosis of mem-brane-associated molecules, and a meetingof the endocytic and exocytic pathway hasbeen described [50, 61, 62]. Such membranetraffic may be required at each stage of milklipid droplet transport: formation in the ER,growth and transit, associated or not withsecretory vesicles, and secretion after envel-opment in the apical plasma membrane.

Milk components are secreted into thelumen of the acinus by a constitutive secre-tory pathway, which depends on the hor-monal state of the animal. In addition, ateach nursing or milking there is a peak inoxytocin and PRL release into the blood,giving transient high concentrations of thesehormones. These high concentrations maybe responsible in vivo for the secretagogueeffect described in vitro.

The mechanisms involved in these pro-cesses have yet to be understood, to dis-cover whether or not lipid and protein secre-tions are associated in the MEC.

Several questions be can raised and partlyanswered using molecular cell biologicalapproaches.

158


known that a lipid droplet grows in size as itmoves to the apical region of the cell [66]and it has been suggested that it continues togrow while protruding from the apical mem-brane. So the rate of release may determinethe size of the lipid globules secreted. Thesecretagogue effect of PRL may increasethe release of secretory vesicles and lipidglobules. Oxytocin may also induce releaseof milk constituents. Do these hormonesplay a part in determining the size of thelipid globules secreted? What mechanismsare involved in determining envelopment,fusion and detachment of the MFGM? Toanswer these questions, comparative analy-sis of genes expressed in MEC during preg-nancy and lactation will be informative.Cytoplasmic and membrane proteins medi-ate fusions between vesicles. A large num-ber of these proteins are known [63]. Withproteomic analysis of the protein composi-tion of MEC it will be possible to describethe proteins involved in these fusion pro-cesses under different experimental condi-tions of hormonal stimulation. However, tointerpret the complex interactions betweenthe proteins active in these processes, knowl-edge of their spatial and temporal recruit-ment at specific locations will be needed.For this, precise detection of these moleculesin their cellular environment will be neces-sary.

ACKNOWLEDGEMENTS

The author is grateful to Dr T.W. Keenan forcritically reading the manuscript, to S. Delpalfor his assistance in performing the electronmicroscopy analysis, to Drs J.-L. Servely andR. Malienou N’Gassa in performing cell culturesand to M.-E. Marmillod for preparing themanuscript.

REFERENCES

[1] Allen J.C., Keller R.P., Archer P., Neville M.C.,Studies in human lactation: milk compositionand daily secretion rates of macronutrients inthe first year of lactation, Am. J. Clin. Nutr. 54(1991) 69–80.

(1) How do nascent lipid droplets budoff from their site of synthesis in the ER?Recent findings show that a neutral lipidcore surrounded by a phospholipid mono-layer bulges as lipids accumulate in the cen-tre of the ER membrane, excluding trans-membrane proteins but containing proteinswith long hydrophobic amino acidsequences [10]. In this way, newly synthe-sised proteins can enter the droplets. Inmouse MEC, the expression of caveolin-1 isdown-regulated during late pregnancy andlactation whereas caveolin-2 decreases onlyslightly [55]. Do caveolins associate withlipid droplets in the MEC? Since caveolin-1and caveolin-2 have the property of forminghetero-oligomers, do variations in theexpression of caveolin-1, in relation to phys-iological stage, play a part in a regulatoryfunction of caveolin-2? These questions cannow be partly answered by combining pro-teomic analysis of the protein compositionof MEC with immunocytochemical locali-sation of caveolins and other proteins asso-ciated with the lipid droplets.

(2) What determines the polarised trans-port of lipid droplets? As mentioned above,lipid droplet transport has been ascribed toassociation with secretory vesicles. How-ever, lipid droplets are able to reach the api-cal region and be released into the lumenof the acini in MEC in which no secretoryvesicles are detectable in the cytoplasm oraround the lipid droplets. Thus, associationwith secretory vesicles does not seem strictlyrequired for vectorial transport. Cytoskele-tal components may play a role in vectorialmovements of lipid droplets to the apicalplasma membrane. Controversial resultssuggest that the mechanisms involving thecytoskeletal system are complex. Comple-mentary investigations will be needed todescribe more precisely the possible inter-actions between microtubules, actin and pro-teins in association with lipid droplets.

(3) What determines the mature size ofthe lipid globule and the release of thedroplet into the lumen? It has long been

159


[2] Ball R.K., Friis R.R., Schoenberger C.A.,Doppler W., Groner B., Prolactin regulation ofbeta casein gene expression and of a cytosolic120-kD protein in a cloned mouse mammaryepithelial cell line, EMBO J. 7 (1988)2089–2095.

[3] Banghart L.R., Chamberlain C.W., Vellarde J.,Korobko I.V., Ogg S.L., Jack L.J.W., VakhariaV.N., Mather I.H., Butyrophilin is expressed inmammary epithelial cells from a single-sizedmessenger RNA as a type I membrane glyco-protein, J. Biol. Chem. 237 (1998) 4171–4179.

[4] Barbero P., Buell E., Zulley S., Pfeffer S.R.,TIP 47 is not a component of lipid droplets,J. Biol. Chem. 276 (2001) 24348–24351.

[5] Bargmann W., Knoop A., Uber die morphologieder milchsekretion. Light- und electronmikro-scopische studie an der milchdrüse der ratte,Z. Zellforsch. Mikrosk. Anat. 49 (1959)344–388.

[6] Blachier F., Lacroix M.C., Ahmed-Ali M., LégerC., Ollivier-Bousquet M., Arachidonic acidmetabolism and casein secretion in lactatingrabbit mammary epithelial cells: Effects ofinhibitors of prostaglandins and leukotrienessynthesis, Prostaglandins 35 (1988) 259–276.

[7] Boisgard R., Chanat E., Lavialle F., Pauloin A.,Ollivier-Bousquet M., Roads taken by milk pro-teins in mammary epithelial cells, Livest. Prod.Sci. 70 (2001) 49–61.

[8] Bousquet M., Fléchon J.E., Denamur R., Aspectsultrastructuraux de la glande mammaire delapine pendant la lactogenèse, Z. Zellforsch. 96(1969) 418–436.

[9] Brasaemle D.L., Rubin B., Harten I.A.,Gruia-Gray J., Kimmel A.R., Londos C., Per-ilipin A increases triacylglycerol storage bydecreasing the rate of triacylglycerol hydroly-sis, J. Biol. Chem. 275 (2000) 38486–38493.

[10] Brown D.A., Lipid droplets: Proteins floatingon a pool of fat, Curr. Biol. 11 (2001)R446–R449.

[11] Chanat E., Martin P., Ollivier-Bousquet M.,as1-casein is required for the efficient transportof b- and k-casein from the endoplasmic retic-ulum to the Golgi apparatus of mammary epithe-lial cells, J. Cell Sci. 112 (1999) 3399–3412.

[12] Clegg R.A., Barber M.C., Pooley L., Ernens I.,Larondelle Y., Travers M.T., Milk fat synthe-sis and secretion: molecular and cellular aspects,Livest. Prod. Sci. 70 (2001) 3–14.

[13] Clermont Y., Xia L., Rambourg A., Turner J.D.,Hermo L., Transport of casein submicelles andformation of secretion granules in the Golgiapparatus of epithelial cells of the lactating mam-mary gland of rat, Anat. Rec. 235 (1993)353–373.

[14] Da Costa T.H.M., Taylor K., Ilic V., WilliamsonD.H., Regulation of milk lipid secretion: effectsof oxytocin, prolactin and ionomycin on tria-cylglycerol release from rat mammary glandslices, Biochem. J. 308 (1995) 975–985.

[15] Daudet F., Augeron C., Ollivier-Bousquet M.,Effet rapide in vitro de la colchicine, du chlorured’ammonium et de la prolactine sur la sécrétiondes lipides du lait dans la glande mammaire,Eur. J. Cell Biol. 24 (1981) 197–202.

[16] Davis D.T., Holt C., Christie W.W., The com-position of milk. Biochemistry of lactation, in:Mepham T.B. (Ed.), Biochemistry of Lactation,Elsevier Science Publishers BV, 1983, pp.71–117.

[17] Devinoy E., Stinnakre M.G., Lavialle F., ThépotD., Ollivier-Bousquet M., Intracellular routingand release of caseins and growth hormone pro-duced into milk from transgenic mice, Exp. CellRes. 221 (1995) 272–280.

[18] Dylewski D.P., Dapper C.H., Valivullan H.M.,Deeney J.T., Keenan T.W., Morphological andbiochemical characterization of possible intra-cellular precursors of milk lipid globules, Eur. J.Cell Biol. 35 (1984) 99–111.

[19] Fujimoto T., Kogo H., Ishiguro K., Tauchi K.,Nomura R., Caveolin is targeted to lipid droplets,a new “membrane domain” in the cell, J. CellBiol. 152 (2001) 1079–1085.

[20] Glick B., Organisation of the Golgi apparatus,Curr. Op. Cell Biol. 12 (2000) 450–456.

[21] Ghosal D., Ankrapp D., Keenan T.W., Lowmolecular mass GTP-binding proteins aresecreted from mammary epithelial cells in asso-ciation with lipids globules, Biochim. Biophys.Acta 1168 (1993) 299–306.

[22] Ghosal D., Shappell N.W., Keenan T.W., Endo-plasmic reticulum lumenal proteins of rat mam-mary gland. Potential involvement in lipiddroplet assembly during lactation, Biochim. Bio-phys. Acta 1200 (1994) 175–181.

[23] Heid H.W., Schnölzer M., Keenan T.W.,Adipocyte differentiation-related protein issecreted into milk as a constituent of milk lipidglobule membrane, Biochem. J. 320 (1996)1025–1030.

[24] Henninghausen L., Robinson G.W., WagnerK.U., Liu W., Prolactin signaling in mammarygland development, J. Biol. Chem. 272 (1997)7567–7569.

[25] Houdebine L.M., The regulation of milk pro-tein synthesis, in: Martinet J., Houdebine L.M.,Head H.H. (Eds.), Biology of Lactation, INRAEditions, Chap. 16, 1999, pp. 401–427.

[26] Jensen R.G., Introduction, in: Jensen R.G. (Ed.),Handbook of milk composition, Academic Press,Chap. I, 1995, pp. 1–3.

[27] Keenan T.W., Milk lipid globules and their sur-rounding membrane: A brief history and per-spectives for future research, J. Mam. GlandBiol. Neoplasia 6 (2001) 365–371.

[28] Keenan T.W., Patton S., The structure of milk:implications for sampling and storage, in: JensenR.G. (Ed.), Handbook of milk composition,Academic Press, Chap. 2, 1995, pp. 5–50.

160


[41] Monks J., Uelmen Huey P., Hanson L., EckelR.H., Neville M.C., Gavigan S., A lipoprotein-containing particle is transferred from the serumacross the mammary epithelium into the milkof lactating mice, J. Lipid Res. 42 (2001)686–696.

[42] Murphy D.J., Vance J., Mechanism of lipid-body formation, Trends Biochem. Sci. 24 (1999)109–115.

[43] Neveu C., Riaublanc A., Miranda G., ChichJ.-F., Martin P., Is the apocrine milk secretionprocess observed in the goat species rooted in theperturbation of the intracellular transport mech-anism induced by defective alleles at the as1-Cnlocus? Reprod. Nutr. Dev. 42 (2002) 163–172.

[44] Neville M.C., Allen J.C., Archer P.C., CaseyC.E., Seacat J., Keler R.P., Lutes V., Rasbach J.,Neifert M., Studies in human lactation: milkvolume and nutrient composition during wean-ing and lactogenesis, Am. J. Clin. Nutr. 54(1991) 81–92.

[45] Nicholas K.R., Hartmann P.E., Milk secretion inthe rat: Progressive changes in milk composi-tion during lactation and weaning and the effectof diet, Comp. Biochem. Physiol. A 98 (1991)535–542.

[46] Nickerson S.C., Keenan T.W., Distribution andorientation of microtubules in milk secretingepithelial cells of rat mammary gland, Cell Tis-sue Res. 202 (1979) 303–312.

[47] Nickerson S.C., Smith J.J., Keenan T.W., Roleof microtubules in milk secretion – Action ofcolchicine on microtubules and exocytosis ofsecretory vesicles in rat mammary epithelialcells, Cell Tissue Res. 207 (1980) 361–376.

[48] Ollivier-Bousquet M., Effet de l’ocytocinein vitro sur le transit intracellulaire et la sécrétiondes protéines du lait, C.R. Acad. Sci. Paris 282(1976) 1433–1436.

[49] Ollivier-Bousquet M., Early effects of prolactinon lactating rabbit mammary gland, CellTissue Res. 187 (1978) 25–43.

[50] Ollivier-Bousquet M., Transferrin and prolactintranscytosis in lactating mammary epithelialcell, J. Mam. Gland Biol. Neoplasia 3 (1998)303–313.

[51] Ollivier-Bousquet M., Hormonal control ofcasein secretion, in: Martinet J., HoudebineL.M., Head H.H. (Eds.), Biology of Lactation,INRA Editions, Chap. 17, 1999, pp. 429–451.

[52] Ollivier-Bousquet M., Demarne Y., Effets dela prolactine sur la sécrétion des lipides du laitdans les cellules épithéliales mammaires delapine en lactation, Reprod. Nutr. Dev. 26 (1986)815–826.

[53] Ollivier-Bousquet M., Denamur R., Inhibitionpar la colchicine de la sécrétion des protéinesdu lait, C.R. Acad. Sci. Paris 276 (1973)2183–2186.

[29] Lkhider M., Petridou B., Aubourg A., Ollivier-Bousquet M., Prolactin signalling to milk proteinsecretion but not to gene expression depends onthe integrity of the Golgi region, J. Cell Sci. 114(2001) 1883–1891.

[30] Loizzi R.F., Mammary cytoskeleton and the reg-ulation of microtubules, in: Neville M.C., DanielC. (Eds.), The Mammary gland, Plenum Press,Chap. 5, 1987, pp. 147–179.

[31] Lollivier V., Guinard-Flament J., Ollivier-Bousquet M., Marnet P.-G., Oxytocin and milkremoval: two important sources of variation inmilk production and milk quality during andbetween milkings, Reprod. Nutr. Dev. 42 (2002)173–186.

[32] Lollivier V., Rainteau D., Marnet P.G., Letort S.,Delpal S., Ollivier-Bousquet M., Early oxytocineffects on secretory events in rabgit and cowmammary cells, in: Baldi A., Stelwagen K.(Eds.), Fifth International Workshop on the Biol-ogy of Lactation in farm animals, Elsevier, 70,2001, p. 180.

[33] Londos C., Brasaemle D.L., Schultz C.J.,Segrest J.P., Kimmel A.R., Perilipins, ADRP,and other proteins that associate with intracel-lular neutral lipid droplets in animal cells, SeminCell Dev. Biol. 10 (1999) 51–58.

[34] Martinet J., Houdebine L.M., Mammary gland,mammogenesis, growth factors, lactogenesis,in: Martinet J., Houdebine L.M., Head H.H.(Eds.), Biology of Lactation, INRA Editions,Chap. I , 1999, pp. 1–27.

[35] Mather I.H., Proteins of the milk-fat globulemembrane as markers of mammary epithelialcells and apical plasma membrane, in: NevilleM.C., Daniel C.W. (Eds.), The mammary gland,development, regulation and function, PlenumPress New York, 1987, pp. 217–267.

[36] Mather I.H., A review and proposed nomencla-ture for major proteins of the milk-fat globulemembrane, J. Dairy Sci. 83 (2000) 203–247.

[37] Mather I.H., Jack L.J.W., Madara P.J., JohnsonV.G., The distribution of MUC1, an apical mem-brane glycoprotein, in mammary epithelial cellsat the resolution of the electron microscope:Implications for the mechanism of milk secre-tion, Cell Tissue Res. 304 (2001) 91–101.

[38] Mather I.H., Keenan T.W., Function ofendomembranes and the cell surface in the secre-tion of organic milk constituents, in: MephanT.B. (Ed.), Biochemistry of lactation, ElsevierScience Publishers, 1983, pp. 231–283.

[39] Mather I.H., Keenan T.W., Origin and secre-tion of milk lipids, J. Mam. Gland Biol. Neo-plasia 3 (1998) 259–273.

[40] McManamam J.L., Hanson L., Neville M.C.,Wright R.M., Lactogenic hormones regulatexanthine oxidoreductase and b-casein levels inmammary epithelial cells by distinct mecha-nisms, Arch. Biochem. Biophys. 373 (2000)318–327.

161


[54] Ostermeyer A.G., Paci J.M., Zeng Y., LublinD.M., Munro S., Brown D.A., Accumulation ofcaveolin in the endoplasmic reticulum redirectsthe protein to lipid storage droplets, J. Cell Biol.152 (2001) 1071–1078.

[55] Park D.S., Lee H., Riedel C., Hulit J., SchererP.E., Pestell R.G., Lisanti M.P., Prolactin neg-atively regulates caveolin-1 gene expression inthe mammary gland during lactation, via a Ras-dependent mechanism, J. Biol. Chem. 276(2001) 48389–48397.

[56] Patton S., Keenan T.W., The milk fat globulemembrane, Biochim. Biophys. Acta 415 (1975)273–309.

[57] Pauloin A., Delpal S., Chanat E., Lavialle F.,Aubourg A., Ollivier-Bousquet M., Brefeldin Adifferently affects basal and prolactin-stimulated milk protein secretion in lactatingrabbit mammary epithelial cells, Eur. J. CellBiol. 72 (1997) 324–336.

[58] Pitelka D.R., Hamamoto S.T., Ultrastructure ofthe mammary secretory cell, in: Mepham T.B.(Ed.), Biochemistry of Lactation, Elsevier, 1983,pp. 29–70.

[59] Pol A., Luetterforst R., Lindsay M., Heino S.,Ikonnen E., Parton R.G., A caveolin dominantnegative mutant associates with lipid bodies andinduces intracellular cholesterol imbalance,J. Cell Biol. 152 (2000) 1057–1070.

[60] Puissant C., Bayad-Sarmadi M., Devinoy E.,Houdebine L.M., Variation of transferrin mRNAconcentration in the rabbit mammary gland dur-ing the pregnancy-lactation-weaning cycle andin cultured mammary cells. A comparison withthe other major milk protein, Eur. J. Endocrinol.130 (1994) 522–529.

[61] Seddiki T., Delpal S., Aubourg A., Ollivier-Bousquet M., Endocytic prolactin routes to thesecretory pathway in lactating mammary epithe-lial cells, Biol. Cell 94 (2002) 173–185.

[62] Seddiki T., Ollivier-Bousquet M., Temperaturedependence of prolactin endocytosis and caseinexocytosis in epithelial mammary cells, Eur. J.Cell Biol. 55 (1991) 60–70.

[63] Segev N., Ypt and Rab GTPases: insight intofunctions through novel interactions, Curr. Op.Cell Biol. 13 (2001) 500–511.

[64] Shennan D.B., Peaker M., Transport of milkconstituents by the mammary gland, Physiol.Rev. 80 (2000) 925–951.

[65] Smith S.J., Cases S., Jensen D.R., Chen H.C.,Sande E., Tow B., Sanan D.A., Raber J., EckelR.H., Farese R.V., Obesity resistance and mul-tiple mechanisms of triglyceride synthesis inmice lacking Dgat, Nat. Genet. 25 (2000) 87–90.

[66] Stemberger B.H., Patton S., Relationships ofsize, intracellular location, and time requiredfor secretion of milk fat droplets, J. Dairy Sci. 64(1981) 422–426.

[67] Valivullah H.M., Bevan D.R., Peat A., KeenanT.W., Milk lipid globules: Control of their sizedistribution, Proc. Natl Acad. Sci. USA 85(1988) 8775–8779.

[68] Wolins N.E., Rubin B., Brasaemle D.L., TIP 47associates with lipid droplets, J. Biol. Chem.278 (2001) 5101–5108.

[69] Wooding F.B.P., The mechanism of secretionof the milk fat globule, J. Cell Sci. 9 (1971)805–821.

[70] Wu C.C., Howell K.E., Neville M.C., YatesJ.R., McManaman J.L., Proteomics reveal a linkbetween the endoplasmic reticulum and lipidsecretory mechanisms in mammary epithelialcells, Electrophoresis 21 (2000) 3470–3482.

[71] Zwierzchowski L., Hormonal control of celldivision and DNA synthesis in the mammarygland, in: Martinet J., Houdebine L.M., HeadH.H. (Eds.), Biology of Lactation, INRA édi-tions, Chap. III, 1999, pp. 57–76.

162

To access this journal online:www.edpsciences.org

Documents

Milk lipid and protein traffic in mammary epithelial cells