casein secretion by mammary gland epithelia from collagen gel cultures and lactating glands

JOURNAL OF CELLULAR PHYSIOLOGY 132:343-348 (1987)

Casein Secretion by Mammary Gland Epithelia From Collagen Gel Cultures and Lactating

Glands VICTOR ROCHA,* S00-IN HWANC, AND C. LEO ORTlZ

Biology Board of Studies, University of California, Santa Cruz, California 95064

Amino acid incorporation experiments show that epithelial cells from lactating mouse mammary glands and from collagen gel culture both synthesize and secrete four principal phosphocaseins (p45, p40, p27, and p23 kD). In both cases, however, t h e casein production is largely dominated by the p27 species. The average percentage distribution of the above casein species in medium from cultured epithelia is approximately 13%, 6%, 68%, and 14%, respectively; for milk the distribution is approximately 23%, 7%, 54%, and 16%. The predominance of the p27 species is not a consequence of extensive extracellular differential degradation of the secreted caseins since no significant casein degradation was observed in culture medium, either in contact or isolated from epithelial cell monolayers. Synthesis and secretion of all the caseins by cultured epithelia is dependent upon insulin, prolactin, and hydrocortisone. Presumably some intracellular events result in the secretion of p27 as the principal casein in mouse milk. Apparently, some selection factor(s) operate to make p27 a major nitrogenous nutritional component for a newborn mouse. In addition, on a quantitative basis, t h e relative levels of various caseins secreted by epithelia from lactating mammary glands is essentially duplicated by epithelia in collagen gel culture.

It is generally agreed that caseins are the principal nitrogenous nutritional component of milk since in most species they constitute the vast majority of the protein found in milk (Davies et al., 1983). Cultivation of mammary epithelia provides an opportunity to study the mechanism of casein synthesis and secretion and its regulation. In particular, mammary epithelia maintained on collagen gels appear to provide an attractive model system with which to study the process of mammary epithelial cell secretory differentiation (Emerman and Pitelka, 1977; Emerman et al., 1977).

Recently, we reported an immunochemical comparison of caseins synthesized and secreted by lactating mammary glands and those produced by epithelia cultured on collagen gels (Rocha et al., 1985). We found that in culture and in vitro the pattern of casein synthesis and secretion is virtually identical. Because of the qual- itative nature of immunoblotting, these studies did not provide a quantitative comparison of secreted caseins. Using radioisotope incorporation techniques, we have made a quantitative comparison of the caseins secreted by lactating mammary glands and epithelia cultured on collagen gels. We report here that the relative levels of various caseins secreted by lactating mouse mammary glands and by collagen-gel-cultivated epithelia are essentially identical. However, the ratio of the secreted caseins is quite different; casein secreted by epithelia from both sources is largely dominated by the 27 kD phosphocasein (@casein). Apparently, of the various mouse caseins, p27 has been selected for, and con-

stitutes a principal nitrogenous nutritional protein for a newborn mouse. A preliminary report of this work has been presented elsewhere (Rocha et al., 1986).

MATERIALS AND METHODS Tissue culture

Collagen gel cultures were prepared according to methods of Emerman and Pitelka (1977), with modifica- tions previously described (Ringo and Rocha, 1983; Ro- cha et al., 1985).

After 48 hr in culture, when complete monolayers were formed, gels were gently released to float in the culture medium. These monolayers undergo gel contrac- tion and cell shape change, and begin to secrete caseins into the medium (released gels). Cells were fed fresh medium every 24 hr; harvested medium was stored at -20°C.

Labeling of caseins Cell culture studies. Caseins were labeled in cell cul-

ture experiments on day 5 following release of the monolayers. With 3H-proline labeling, monolayers were washed with proline-free culture medium. The cultures were then maintained in culture medium containing 3H- proline (31 pC/ml, sp. act. 102 Ci/mmole) plus 0.5 pg/ml proline for 24 hr. Labeling with other isotopes followed

Received May 9, 1986; accepted March 10,1987

*To whom reprint requestsicorrespondence should be addressed.

0 1987 ALAN R. LISS, INC.

344 ROCHA, HWANG, AND ORTIZ

A

p 4 5 2 p 4 0 >

p27s

p 2 3 ~

1 2 1 2

Fig. 1. Isotope-labeled caseins from culture medium and milk. Epithe- lial cells on collagen gels or in lactating mammary glands were labeled, and TCA-precipitated culture medium or milk was resolved in 12% SDS-PAGE. The labeled caseins were visualized by fl,urography. A) 1. 3H-Proline-labeled caseins from culture medium; 2. 'H-Proline- labeled from caseins mouse milk. B) Isotope-labeled caseins from culture medium. l. %Leucine labeled caseins; 2. 14C-proline-labeled caseins. Arrows on the left indicate the migrational position of authentic mouse caseins as control markers. Numbers refer to minimum molecular weights (kD).

a similar protocol where the levels of radioactivit were:

cine (33 pC/ml, sp. act. 40-60 Ci/mmole), and 35S-methionine (35 pC/ml, sp. act. 1,000-1,300 Ci/mmole). All isotopes were obtained from Amersham Corp., Arling- ton Heights, IL. Total protein in culture medium was quantified by counting label in TCA precipitates (10%).

Animal studies. Balb/c mice in the fifth day of lactation were injected with 3H-proline (ca. 1 mCi) and maintained with sucklings overnight. Following a 2-hr period of isolation from the newborn mice, the lactating moth- ers were injected with oxytocin (ca. 10 units), and milk was obtained using mild suction. The labeled milk was treated with ether extraction to remove lipid and then frozen.

14C-proline (33 pC/ml, sp. act. 290 mCi/mmole), 3 H-leu-

Gel electrophoresis Acrylamide gel electrophoresis was performed accord-

ing to Laemmli (1970) using 12% slabs. Culture medium and milk samples were precipitated with cold 10% tri- chloroacetic acid (TCA). TCA precipitates were washed three times with cold 5% TCA and resuspended in sample loading buffer. Sample loading buffer consisted of 125 mM Tris, pH 6.8, 25% glycerol, 5% sodium dodecyl sulfate (SDS), and 3% B-mercaptoethanol. Samples were then thermally denatured a t 100°C for 3 min.

Quantitation of SDS-PAGE fluorographs The position of labeled caseins on the dried gels was

determined from fluorographs. Bands were carefully marked, excised with sharp microdissecting scissors, and placed in glass scintillation vials with 300-500 p1 of 30% Hz02. Vials were tightly capped and incubated overnight a t 80°C. Excised bands (< lo0 mg) were com- pletely solubilized by this procedure. Samples were allowed to cool and prepared for counting by addition of 50-100 p1 acidifying agent (Delume, Isolab; Akron, OH) to eliminate chemiluminescence, followed by 10-ml counting cocktail (Biofluor, NEN; Boston, MA). Control studies indicated that radioactivity recovery was > 95% and that quenching was negligible.

RESULTS Isotope-labeled caseins from culture medium and milk

Following 5 days of cultivation as released cultures, monolayers were cultured in medium containing 3H- proline for 24 hr. Labeled mouse milk was obtained by injecting a 1-week lactating mouse with 3H-proline. Both labeled culture medium and labeled mouse milk were resolved by SDS-PAGE, visualized with fluorography, and compared to authentic mouse caseins (Fig. 1A). Both labeled culture medium and milk demonstrated four principal casein species of approximate molecular weight, 45, 40, 27, and 23 kD. These correspond to the a-1, a-2, 0, and y caseins described by Hennighausen et al. (1982). Similar cultured epithelial cell labeling experiments were conducted with 3H-leucine and 14C-proline (Fig. 1B). With each isotope, while four labeled caseins were obtained, the principal casein labeled was the p27 species.

Quantitation of labeled secreted caseins Labeled caseins in the gels used to generate the differ-

ent fluorographs from Figure 1 were quantified by cut- ting out and eluting the various species, then counting them with liquid scintillation spectrometry (Table 1). 3H-proline-labeled caseins from epithelia cultivated on collagen gels and from lactating mammary glands were

TABLE 1. Relative amount of labeled casein in culture medium and milk (9%)

Culture medium Milk Casein "H-proline 14C-proline 3H-leucine - 'H-proline species in = 4) (n = 5) (n = 1) X in = 3)

P45 10.4 f 1.7 13.4 f 3.3 14.1 12.6 + 2.0 23.2 f 3.0 P40 4.0 + 1.0 6.6 f 1.1 7.3 6.0 + 1.7 6.5 0.6

P23 17.9 + 5.2 16.0 k 5.6 6.7 13.5 k 6.0 15.9 k 1.2 P27 67.8 f 5.3 63.8 f 8.5 72.0 67.9 f 4.1 53.6 f 3.7

CASEIN SECRETION PATTERNS 345

@

4 o t P

30

16 12 0

0 4 0 Time (hrsl

Fig. 2. Stability of secreted caseins in culture medium of mammary epithelial cell cultivated on collagen gels. A) Change in total precipitable labeled protein in culture medium. Cells were labeled for 24 hr with 14C-proline and three paired cultures were set up. At this point (t = 0) the degradation experiment commenced. T: A culture allowed to continue accumulation of labeled caseins (0); a culture treated in an identical manner, but aprotonin was added at 0 hr (W. C: a culture treated with 1 mgiml proline (metabolic chase) at 0 hr and allowed to accumulate labeled and non-labeled caseins (0); a culture treated in an identical manner, but aprotonin was added at 0 hr (0). M: Labeled culture medium transfered to a sterile culture dish and monitored for

significantly dominated by the p27 species. With cultured epithelia, p27 was approximately 68% of the total secreted casein; with intact-gland-derived casein, p27 constituted approximately 54% of the total secreted casein. In addition, p27 was the dominant casein when total caseins from cultured mammary epithelia were labeled with 3H-leucine ( - 72%), and 14C-proline (- 64%).

Stability of caseins in culture system Degradation of newly secreted caseins was examined

in collagen gel culture. Following five days of maintenance as released cultures, the monolayer gels were labeled with medium containing 14C-proline. After 24 hr of labeling, three paired cultures were set up as follows: 1) a control culture and a culture containing the protease inhibitor, aprotonin; 2) two cultures containing 1 mg/ml sterile proline (chase experiment), and one of the cultures containing aprotonin; 3) labeled culture medium transferred to two sterile plastic dishes, with one dish containing aprotonin. We defined this time point to be 0 hr of protein degradation, and the culture medium from these six cultures was sampled at various times and examined for total TCA precipitable labeled

18hr

T C

Ohrs 18hr

M

level of labeled casein (A); labeled culture medium treated in an identical manner, but aprotonin was added at 0 hr (A). B) Quantitation of labeled caseins from SDS-PAGE fluorographs of samples presented in panel A. Labeled p27 caseins (viz. the dominant species) were cut from SDS-PAGE slabs, eluted, and counted by liquid scintillation spectrometry; two time points were examined. Labeled p27 casein at 0 hr for cultures, T, C, and M, respectively. Labeled caseins at 18 hr following the treatments described in panel A for cultures T, C, and M. Only the cultures not treated with aprotonin were examined, since addition of this inhibitor had no effect on casein degradation in panel A.

protein (Fig. 2A). In addition, culture medium from 0 hr and 18 hr of the degradation experiment was examined by SDS-PAGE fluorography (data not shown); the p27 casein in the gels was quantified as described above (Fig. 2B). The data clearly show that significant degradation of total precipitable protein or casein does not occur whether or not culture medium is in contact with epithelial cell monolayers. The presence of aprotonin had no effect on casein degradation. In a second experiment we obtained essentially identical results as those presented in Fig. 2A. In addition, in this experiment we measured the ratios of all four caseins in culture medium at 0 hr degradation, at 24 h r of accumulation, and in isolated medium over the same period. All three samples had nearly identical casein ratios (Table 2). Clearly, differential degradation of caseins does not occur.

Hormonal requirement for casein secretion Initially, cells were plated on serum-conditioned colla-

gen gels in culture medium containing insulin. Follow- ing cell attachment to the gels, cultures were maintained for the remainder of the experiment in medium containing insulin only, or insulin, hydrocortisone, and prolac-


TABLE 2. Stability of casein species ratios in culture media with and without cells (%;I)

Control With cells Without cells (t = 24 hr) (t = 24 hr) Species (t = 0 hr)

P45 15 11 15 6 7 7 n40

A B

58 59 62 20 19 19

r--

P27 P23

p 4 5 5

p 4 0 >

tin. Cultures were maintained as released monolayer gels in their appropriate medium, and on day 5, the

proline or 35S-methionine. Culture medium was harvested following 24 hr and examined by SDS-PAGE fluorography (Fig. 3). Four labeled caseins were recovered only when the complete hormone combination of insulin, hydrocortisone, and prolactin was provided to the

35S-methionine labeling of cells in insulin medium alone

ture medium that migrated to a Position in 12% SDS- cultivated mouse mammary epithelial cells. Cells were cultured in PAGE that partially overlaps p45 (Fig. 3B-1). The SDS- medium containing various hormone combinations, and caseins were PAGE positions occupied by p40, p27, and p23 had no labeled. Culture medium was precipitated with TCA and resolved in

12% SDS-PAGE. The labeled caseins were visualized by fluorography. labeled proteins in medium from insulin-only To provide for appropriate comparisons, all sample wells received 100 cultivated cultures. A comparable protein was not ob- p~ of cell lysate (i.e,, secretory material from approximately 3 x lo5 served with 'H-proline labeling (Fig. 3A-1). cells. A) Casein labeling experiment using 3H-proline. The hormone

combinations examined were: 1) insulin only, 2) insulin, h drocorti- DISCUSSION sone, and prolactin. B) Casein labeling experiment using S-methio-

nine. The same hormone combinations used in panel A were used here.

veal that p27 casein is produced in quantities greater species. Arrows on the left and right indicate the migrational position than the other three major mouse caseins combined. of authentic mouse caseins as control markers. Numbers refer to min- This was observed for caseins secreted by lactating imum

mammary glands and by secretory mammary epithelia cultivated on collagen gels. Since the mole fraction of any amino acid in the various caseins is not constant, the least biased representation of the data is the average casein ratios from each amino acid incorporation exper- Differential degradation of caseins in the culture me- iment. With collagen-gel-cultured epithelia, incorpora- dium could provide an explanation for the predominance tion experiments with 'H-proline, 3H-leucine, and 14C- of p27 in our labeling experiments. However, in our proline show that the p45, p40, p27, and p23 caseins experiments the extent of casein accumulation, as mea- comprise approximately 13%, 6%, 68%, and 14% of the sured by TCA precipitable counts, in cultures was iden- newly synthesized casein, respectively. In a similar tical regardless of the presence or absence of a general manner, milk from lactating mammary glands labeled protease inhibitor. In addition, culture medium contain- with 'H-proline showed that the p45, p40, p27, and p23 ing labeled casein, when isolated from the epithelial cell caseins comprise approximately 23%, 7%, 54%, and 16% monolayers, maintained constant casein level whether of the newly synthesized casein, respectively. protease inhibitor was present or absent. Thus, we have

We do not know how this level of differential casein not been able to observe any significant extracellular secretion occurs. Exchange of protons with 'H-labeling protein or casein degradation in the collagen gel culture is not the explanation since 14C-proline also shows p27 system. In addition, the ratios of the various caseins to be the principal labeled species. Unusual proline con- species at 0 hr and 24 hr of monitoring were essentially tent in the individual caseins is excluded since 3H-leu- identical, indicating that no differential degradation of cine labeling also showed p27 to be the dominant labeled the casein species occurred during our period of analysis. casein. To insure that the intracellular amino acid pool We do not suggest that no casein degradation occurs; in of the cultured epithelia was in equilibrium with the fact we would be surprised if such were the case. How- radio-amino acid used, we only examined cultures after ever, we do conclude that the magnitude of such degra- linear output of casein had been achieved. Thus, we do dation is negligible compared to the rate of casein not believe that the p27-dominated casein fluorograph secretion. Otherwise net accumulation of casein in the pattern is an artifact of differential amino acid pool culture medium could not occur. At this time we do not labeling. know anything about intracellular casein degradation;

Degradation of caseins has been reported in various however, we would not be surprised to find significant mammary gland and epithelial cell preparations (Wilde and, perhaps, differential degradation of the various et al., 1980, 1984; Hasan et al., 1982; Lee et al., 1985). casein species within mammary epithelia.

p27 z

monolayers were cultured in medium containing 3H- p 2 3 >

1 2 cultures (Fig. 3A-2; 3B-2). It is important to note that 1 2

demonstrated a labeled protein in the cul- Fig, 3, Hormonal requirement for casein secretion in collagen-gel.

2 Our amino acid isotope incorporation experiments re- X marks the position of a non-casein that migrates close to the p45

weights in kD.

CASEIN SECRETION PATTERNS 347

Lack of extracellular casein degradation suggests that some intracellular process influences the quantitative level of the caseins secreted by epithelia, both in the collagen gel system and in the lactating mammary gland.

Currently, there is no evidence that the mouse casein genes are amplified (Rosen and Barker, 1976), and, therefore, some form of differential casein mRNA transcription or translation may provide a reasonable explanation. There is some data concerning differential transcription of the casein genes when mammary epithelia are maintained in altered cell shapes (Haeuptle et al., 1983; Suard et al., 1983; Lee et al. 1984, 1985). However, those data are not comparable to our study since our collagen gel cultures are maintained in the preferred morphology for maximum secretory activity (Shannon and Pitelka, 1981).

The most likely possibility for this observation is that differential intracellular degradation of the various caseins might occur to produce a secretion slate of caseins that is highly enriched for p27. It is possible that the level of intracellular casein phosphorylation might play a role in controlling the degradation of caseins prior to their secretion. In this regard, we have previously shown that non-phosphorylated casein species are present within the epithelia and that only mature fully phosphorylated caseins are recovered in the culture medium (Ro- cha et al., 1985). Perhaps differential phosphorylation of the various caseins is a n intracellular signal to degrade them. The predominance of p27 in the culture medium and milk may reflect its efficient intracellular phosphorylation, thus protecting it from degradation.

Durban et al. (1985) have reported that virgin mammary epithelial cells maintained in embedded collagen gels for prolonged periods prior to secretion, and then induced to produce casein, show a transition of casein production from p45 (a1-casein) to p27 (&casein). They suggest, that more mature cultures may have engaged in synthetic events that make them more sensitive to developmental cues. While our cultures were treated in a very different manner, the results of Durban et al. with prolonged collagen gel cultures and differential casein production could have some bearing on our observation. In this regard, labeled mouse milk does show a slightly greater amount of al-casein and a lower level of 0-casein when compared to the same caseins in labeled culture medium (Table 1).

As has been shown in several other systems (Banerjee, 1978), casein production in this culture system requires insulin prolactin and hydrocortisone. Our experiments clearly show that cultivation of epithelia in medium lacking prolactin and hydrocortisone resulted in failure to secrete any of the casein species produced with the full lactogenic hormone combination. The fact that re- moval of hydrocortisone and prolactin resulted in the absence of p27 and the other caseins indicates that our quantitative assay for caseins is both specific and direct. In quantifying the p45 casein (a1-casein), especially with 35S-methionine, this becomes particularly important because a labeled non-casein species migrates very close to this casein species in SDS-PAGE. Its non-casein charac- ter is based on the fact that it does not perfectly CO- migrate with the p45 casein, that it is synthesized during the cell-spreading phase of this culture system (data

not shown), that it continues to be synthesized in hormone-deficient culture medium, and that it fails to react with casein antibodies in immunoblot analysis (data not shown).

An explanation why p27 is the dominant casein secreted is not obvious. Perhaps the amino acid composition of the p27 casein has been selected because of its superior nutritional benefit for the newborn mouse. It is also possible that p27 may be able to form micelles more efficiently and, therefore, has been selected as the casein that can accumulate to the highest concentration in mouse milk.

Whatever the reason, it is probably not a coincidence that p27 casein is the predominant secreted casein in mouse milk. Some selection pressure has operated on the protein to render it a principle nitrogenous nutritional component for a newborn mouse. We are currently pursuing experiments to provide an explanation for this interesting observation.

ACKNOWLEDGMENTS The authors wish to acknowledge the expert technical

assistance of Mr. Duk Pi1 Chong in the early phase of this project. This work was supported in part by the University of California Faculty Research Committee, University of California Cancer Research Coordinating Committee, NIH MBRS Program Grant RR08132, and by NIH NIGMS T34-07910.

LITERATURE CITED Banerjee, M.R. (1978) Responses of mammary cells to hormones. Int.

Rev. Cytol. 47r1-97. Davies, D.T., Holt, C., and Christie, W.W. (1983) The composition of

milk. In: Biochemistry of Lactation. T.B. Mepham, ed. Elsevier, New York, pp. 71-117.

Durban, E.M., Medina, D., and Butel, J.S. (19851 Comparative analysis of casein synthesis during mammary cell differentiation in collagen and mammary gland development in uiuo. Dev. Biol. 109t288-298.

Emerman, J.T., Enami, J., Pitelka, D.R., and Nandi, S., (1977) Hor- mone effects on intracellular and secreted casein in cultures of mouse mammary epithelial cells on floating collagen membranes. Proc. Natl. Acad. Sci. USA 74:4466-4470.

Emerman, J.T., and Pitelka, D.R. (1977) Maintenance and induction of morphological differentiation in dissociated mammary epithelium on floating collagen membranes. In Vitro 13t316-328.

Haeuptle, M. -T., Suard, Y.L. M., Bogenmann, D., Reggio, H., Racine, L., and Kraehenbuhl, J.-P. (1983) Effect of cell shape change on the function and differentiation of rabbit mammary cell in culture. J. Cell Biol. 961425-1434.

Hasan, H.R., White, D.A., and Mayer, R.J. (1982) Extensive destruc- tion of newly synthesized casein in mammary explants in organ culture. Biochem. J. 202t133-138.

Hennighausen, L. G., and Sippel, A. E. (1982) Characterization and cloning of the mRNAs specific for the lactating mouse mammary gland. Eur. J . Biochem. 125:131-141.

Laemmli, U.K. (1970) Cleavage of structural proteins during the as- sembly of the head of bacteriophage T4. Nature 227:680-685.

Lee, E.Y.-P., Lee, W., Kaetzel, C.S., Parry, G., and Bissell, M.J. (1985) Interaction of mouse mammarv euithelial cells with collaeen sub- < . " strata: Regulation of casein gene expression and secretion. Proc. Natl. Acad. Sci. USA 82r1419-1423.

Lee, E. Y.-H., Parry, G., and Bissell, M.J. (1984) Modulation of secreted proteins of mouse mammary epithelial cells by the collagenous substrata. J . Cell Biol. 98t146-155.

Ringo, D.L., and Rocha, V. (1983) Xanthine oxidase, an indicator of secretory differentiation in mammary cells. Exp. Cell Res. 147t216- 220.

Rocha, V., Hwang, S., and Ortiz, C.L. (1986) B-Casein is the principal secreted casein by collagen gel and lactating mouse mammary gland epithelia. Fed. Proc. 45r1453.

Rocha, V., Ringo, D.L., and Read, D. B. (1985) Casein production during differentiation of mammary epithelial cells in collagen gel culture.


Exp. Cell Res. 159:201-210. mammary cell cultures. J. Cell Biol. 96:1435-1442. Rosen. J.M.. and Barker. S.W. (1976) Quantitation of casein messenger Wilde. C.J.. Paskin. N. Saxton. F. and Maver. R.J. (1980) Protein I , , I

ribonucIeic acid sequences using a specific complementary DNA yhy- bridization probe. Biochemistry 15.5272-5280. 320.

Shannon, M. M., and Pitelka, D.R. (1981) The influence of cell shape

degradation during terminal cy&differentia<on: Biochem. J. 192:311-

Wilde, C.J., Hasan, H.R., and Mayer, R.J. (1984) Comparison of colla- on the induction of functional differentiation in mouse mammary cells in vitro. In Vitro 17;1016-1028.

Suard, Y.M.L., Haeuptle, M.-T., Farinon, E., and Kraehenbuhl, J.-P. (1983) Cell proliferation and milk protein gene expression in rabbit

gen gels and mammary extracellular matrix as substrata for study of terminal differentiation in rabbit mammary epithelial cells. Exp. Cell Res. 151:519-532.

casein secretion by mammary gland epithelia from collagen gel cultures and lactating glands

Documents