Culture of Mammary Epithelial cells from Bovine Milk GERTRUDE CASE BUEHRING School of Public Health University of California Berkeley 94720 ABSTRACT
mary gland. Although mammary gland tissue may be obtained by biopsy from healthy cows at any phase of reproductive life, the procedure is time-consuming and inconvenient. Like other epithelial cells. mammary epithelial cells exfoliate and are shed into secretions. The somatic cells described in dairy manuals as part of typical milk sediment are actually leukocytes and exfoliated epithelial cells (14). In humans and in nonhuman primates, mammary epithelial cells shed into milk w~re .viable an,d grew in culture (2, 19). The objectIve of thiS study was to determine if cows' milk could ~ an additional source of BMEC for culture. This report described the growth and characterization of BMEC cultured from cows' milk.
Mammary epithelial cells are exfoliated into cows' milk and comprise a component of the somatic cells enumerated for milk quality control. This paper described the finding that these epithelial cells can be recovered from milk and grown in culture. The cultured cells were characterized for growth rate and growth potential, morphology by light and electron microscopy. presence of esterases and cytokeratins, and sensitivity to storage at 4°C. Cultured mammary epithel.ial cells from milk may be useful to dairy scientists and mammary gland biologists. (Key words: culture. milk. mammary epithelium)
MATERIALS AND METHODS
INTRODUCTION
Cell Culture
The first successful in vitro growth of bovine mammary epithelial cells (BMEC) occurred in 1961 as a cell culture (7). Since then, numerous investigators have cultured BMEC either as organ culture (6, 12, 13, 16, 17) or cell culture (16, 20. 22). The primary interest has been to develop in vitro systems for the study of lactogenesis (6. 12, 13, 16, 17). although one group (20) has used BMEC in vitro as a system for basic research on the cytoskeleton and another (22) for research on virology. Investigators studying lactogenesis have evaluated the role of hormones (6. 12, 13. 17) and collagen gels (16) in increasing lactogenesis by BMEC in vitro. Bovine mammary epithelial cells for culture have been obtained from tissue biopsies (6. 13, 15. 17) or glands excised after slaughter (7. 12. 13, 16, 20). Healthy mammary tissue obtained at necropsy is at a premium. It is not economical to slaughter a cow solely to obtain a mam-
Received July 5, 1989. Accepted November 6, 1989. 1990 J Dairy Sci 73:956-963
Raw cows' milk, packaged in quart (.95-L) cartons, was purchased from a local grocery store on the day of milking and delivery. Cells were pelleted from the milk in 50-011 tubes by centrifugation in a table-top centrifuge (500 x g) for 5 to 10 min. Supernatants were decanted, and pellets were pooled by resuspending each in approximately 5 011 of Dulbecco's phosphate-buffered saline (DPBS) (OmCO Laboratories, Orand Island, NY), then pipetting each resuspension into one tube. The tube was filled to 50 011 with DPBS, then recentrifuged 5 min to rinse and repellet. The final pellet was resuspended in 5 011 of Dulbecco's modified Eagles medium (DME) (OmCO) supplemented with antimicrobials and 15% fetal bovine serum (FBS) (Hyclone, Logan, UT) to facilitate attachment of cells. Concentrations of antimicrobials were lOx the usual concentrations for culture work (see end of this paragraph). Cells were inoculated into various types of culture vessels depending on the experiment. After 24 to 48 h, the medium was replaced by DME10% FBS with the same antibiotics but at the usual concentrations for culture work: gentarni-
956
CULTURE OF MAMMARY CELLS FROM MILK
dn (Schering, Kenilworth, NJ; 50 Ilg/ml), penicillin (ICN Nutritional Biochemicals, Cleveland, OH; 250 V/ml), streptomycin (Pfizer, New York, NY; .1 mg/ml), polymixin B (ICN, 50 Vim I), and Fungizone (E. R. Squibb, Princeton. NJ, 2.5 Ilg/ml). Control cultures of fibroblasts were obtained from bovine tendon. Tendon was excised from the forefoot immediately after slaughter (Hoehner Meat Co., Oakland, CA) and placed into sterile culture medium for transport to the laboratory. The tendon was minced into approximately 10-mm3 pieces and digested with shaking for 2 d at 3TC in a solution of hyaluronidase (Sigma Chemical Co., St. Louis, MO, #H-3502; 98.5 V/ml) and collagenase (lCN. Plainview, NY, #100502; 19.6 V/ml). After removing undigested tissue, the remaining solution was centrifuged at 500 x g to pellet cells. Control cultures of macrophages were obtained from bovine blood collected immediately after slaughter into 3.8% sodium citrate in water as an anticoagulant (9 parts blood: 1 part sodium citrate). Blood was then mixed 1: 1 with DPBS and layered over Ficoll-paque (Pharmacia Fine Chemicals, Piscataway. NJ) (2 parts blood: 1 part Ficoll-paque) and centrifuged (500 x g) for 45 min. Leukocytes were harvested from the band immediately above the erythrocyte sediment. rinsed, and pelleted once in DPBS. Cell pellets of both fibroblasts and leukocytes were resuspended in DME-I0% FBS with the antimicrobials listed above (Ix concentration) and inoculated into 96-well culture plates (Corning, Corning, NY). After 24-h, cultures of blood cells were fluid changed. removing all nonadherent cells (lymphocytes, granulocytes, erythrocytes) and leaving only adherent cells (monocytes). Control cultures of BMEC were a cell line (BMEC23) derived from a bovine mammary explant and characterized as epithelial (3). Cultures of all cell types were maintained at 3TC in a humidified atmosphere of 5% C02 and the medium (DME-lO% FBS) changed once per week. Growth Determination
Cells constituting each epithelioid island were counted every 24 to 48 h in situ at l00x magnification with a grid in the microscope eyepiece to facilitate counting. Growth curves
957
for each island were constructed by plotting the logarithm of cell number on the y axis against time in culture on the x axis. The population doubling time (PDT) as an expression of growth rate was calculated from the steepest slope of the logarithmic portion of the curve. Population doubling level (PDL) as an expression of growth potential, was calculated for each island according to the formula: PDL = log (maximum cell number reached/starting cell number)/log 2. Cytochemistry
Presence of keratin in cultured cells was detected with an avidin-biotin immunoperoxidase system. The buffer used throughout was DPBS without Ca2+ and Mg2+ (GlBCO). After an initial wash with buffer and a lO-min fixation with 95% ethanol, nonspecific binding sites were blocked 20 min with normal goat serum (Vector Laboratories, Burlingame, CA; I: 74 in buffer). The primary antibody, rabbit antibovine keratin (Dako. Carpinteria, CA, #Ll824, prediluted by manufacturer I :250 in 10% .5 M Tris, pH 7.6, 90% .2 M PBS), was reacted for 30 min, and the secondary antibody, biotinylated goat antirabbit IgG (Vector Laboratories; 1:222 in buffer) was reacted for 30 min. Signal was amplified with avidin-biotinperoxidase complex (Vector) for 30 min and detected with the chromogen diamino-benzidine (DAB) (Sigma) 7 min. The DAB was prepared just before use by adding equal parts of a 1 mglml solution of DAB in .1 M Tris buffer, pH 7.2 to .02% of hydrogen peroxide prepared fresh in distilled water from a 30% stock. Cells were counterstained for 1 min with undiluted double strength hematoxylin (Lerner Laboratories. New Haven, CT, #S 3141-2). Buffer was used for rinses after primary and secondary antibody and amplification steps. The DAB and hematoxylin were rinsed with tap water. Presence of esterase enzymes was detected according to a standard method described in detail by Koski et al. (I5). Cells were fixed in an acetone and formaldehyde mixture, rinsed, dried, and reacted with an esterase substrate solution of pararosaniline, sodium nitrite, and a-naphthyl butyrate. Cells were rinsed and lightly counterstained with hematoxylin. Journal of Dairy Science Vol. 73, No.4, 1990
958
BUEHRING
Electron Microscopy
After growth in culture for 2 to 4 wk, cells' were fixed for thin-section electron microscopy (EM) in 2.5% glutaraldehyde (Polysciences. Wanington. PA; EM grade; vol/vol in .1 M of sodium cacodylate buffer. pH 7.2) for 3 h; postfixed in 1% osmium tetroxide (Polysciences; vol/vol in .1 M of sodium cacodylate buffer) for 1 h; stained with .5% uranyl acetate (Electron Microscopy Sciences, Fort Washington, PA; wt/wt in distilled water) for 2 h; passed through a graded series of ethanol and resin in ethanol; and embedded in a resin mixture (1 :21.6:25.2) of benzyldimethylamine. dodecenylsuccinic anhydride, and Araldite 6005 (Polysciences). Cells were prepared in situ and sectioned perpendicular to the substrate. All sections were examined with a Zeiss 109 electron microscope (Thornwood. NY). RESULTS
When inoculated into culture, pellets from centrifuged milk exhibited single cells, clumps of cells, bacteria. crystals. and debris. It was possible to see cells with segmented nuclei typical of granulocytes as well as cells and clumps of various sizes (2 to 15 cells) with nonsegmented nuclei. However. the debris was so extensive (Figure I), it was impossible to count individual cells accurately to detennine the mean number of viable cells per liter of milk. After the initial medium change 24 h later, most bacteria and debris were eliminated, and single cells and small epithelioid islands of cells were apparent, spread upon the substrate. Attachment was dependent upon the presence of FBS and did not occur in serum-free medium. even when supplemented with fibronectin (2 Ilg/ml; Collaborative Research, Waltham, MA). The epithelioid cells attached readily to all tissue culture vessels tried: 25-cm 2 flasks and 96-well and 24-well multitest plates (Corning. Corning, NY). In initial experiments, the appearance and quantity of epithelioid islands derived from nonfat milk did not differ subjectively from islands derived from whole milk. Therefore. nonfat milk was used for all subsequent experiments to avoid tenacious cream layers during centrifugation. All attached cells had nonsegmented nuclei. Occasional single cells with the granular. denJournal of Dairy Science Vol. 73. No.4. 1990
Figure 1. Pellet material centrifuged from milk. Note large rod-shaped crystals (e). pigmented plaques (P), and other debris obscuring individual cells (phase contrast, x313; bar 32 Il.
=
dritic appearance of macrophages were present. These made up
mary gland. Although mammary gland tissue may be obtained by biopsy from healthy cows at any phase of reproductive life, the procedure is time-consuming and inconvenient. Like other epithelial cells. mammary epithelial cells exfoliate and are shed into secretions. The somatic cells described in dairy manuals as part of typical milk sediment are actually leukocytes and exfoliated epithelial cells (14). In humans and in nonhuman primates, mammary epithelial cells shed into milk w~re .viable an,d grew in culture (2, 19). The objectIve of thiS study was to determine if cows' milk could ~ an additional source of BMEC for culture. This report described the growth and characterization of BMEC cultured from cows' milk.
Mammary epithelial cells are exfoliated into cows' milk and comprise a component of the somatic cells enumerated for milk quality control. This paper described the finding that these epithelial cells can be recovered from milk and grown in culture. The cultured cells were characterized for growth rate and growth potential, morphology by light and electron microscopy. presence of esterases and cytokeratins, and sensitivity to storage at 4°C. Cultured mammary epithel.ial cells from milk may be useful to dairy scientists and mammary gland biologists. (Key words: culture. milk. mammary epithelium)
MATERIALS AND METHODS
INTRODUCTION
Cell Culture
The first successful in vitro growth of bovine mammary epithelial cells (BMEC) occurred in 1961 as a cell culture (7). Since then, numerous investigators have cultured BMEC either as organ culture (6, 12, 13, 16, 17) or cell culture (16, 20. 22). The primary interest has been to develop in vitro systems for the study of lactogenesis (6. 12, 13, 16, 17). although one group (20) has used BMEC in vitro as a system for basic research on the cytoskeleton and another (22) for research on virology. Investigators studying lactogenesis have evaluated the role of hormones (6. 12, 13. 17) and collagen gels (16) in increasing lactogenesis by BMEC in vitro. Bovine mammary epithelial cells for culture have been obtained from tissue biopsies (6. 13, 15. 17) or glands excised after slaughter (7. 12. 13, 16, 20). Healthy mammary tissue obtained at necropsy is at a premium. It is not economical to slaughter a cow solely to obtain a mam-
Received July 5, 1989. Accepted November 6, 1989. 1990 J Dairy Sci 73:956-963
Raw cows' milk, packaged in quart (.95-L) cartons, was purchased from a local grocery store on the day of milking and delivery. Cells were pelleted from the milk in 50-011 tubes by centrifugation in a table-top centrifuge (500 x g) for 5 to 10 min. Supernatants were decanted, and pellets were pooled by resuspending each in approximately 5 011 of Dulbecco's phosphate-buffered saline (DPBS) (OmCO Laboratories, Orand Island, NY), then pipetting each resuspension into one tube. The tube was filled to 50 011 with DPBS, then recentrifuged 5 min to rinse and repellet. The final pellet was resuspended in 5 011 of Dulbecco's modified Eagles medium (DME) (OmCO) supplemented with antimicrobials and 15% fetal bovine serum (FBS) (Hyclone, Logan, UT) to facilitate attachment of cells. Concentrations of antimicrobials were lOx the usual concentrations for culture work (see end of this paragraph). Cells were inoculated into various types of culture vessels depending on the experiment. After 24 to 48 h, the medium was replaced by DME10% FBS with the same antibiotics but at the usual concentrations for culture work: gentarni-
956
CULTURE OF MAMMARY CELLS FROM MILK
dn (Schering, Kenilworth, NJ; 50 Ilg/ml), penicillin (ICN Nutritional Biochemicals, Cleveland, OH; 250 V/ml), streptomycin (Pfizer, New York, NY; .1 mg/ml), polymixin B (ICN, 50 Vim I), and Fungizone (E. R. Squibb, Princeton. NJ, 2.5 Ilg/ml). Control cultures of fibroblasts were obtained from bovine tendon. Tendon was excised from the forefoot immediately after slaughter (Hoehner Meat Co., Oakland, CA) and placed into sterile culture medium for transport to the laboratory. The tendon was minced into approximately 10-mm3 pieces and digested with shaking for 2 d at 3TC in a solution of hyaluronidase (Sigma Chemical Co., St. Louis, MO, #H-3502; 98.5 V/ml) and collagenase (lCN. Plainview, NY, #100502; 19.6 V/ml). After removing undigested tissue, the remaining solution was centrifuged at 500 x g to pellet cells. Control cultures of macrophages were obtained from bovine blood collected immediately after slaughter into 3.8% sodium citrate in water as an anticoagulant (9 parts blood: 1 part sodium citrate). Blood was then mixed 1: 1 with DPBS and layered over Ficoll-paque (Pharmacia Fine Chemicals, Piscataway. NJ) (2 parts blood: 1 part Ficoll-paque) and centrifuged (500 x g) for 45 min. Leukocytes were harvested from the band immediately above the erythrocyte sediment. rinsed, and pelleted once in DPBS. Cell pellets of both fibroblasts and leukocytes were resuspended in DME-I0% FBS with the antimicrobials listed above (Ix concentration) and inoculated into 96-well culture plates (Corning, Corning, NY). After 24-h, cultures of blood cells were fluid changed. removing all nonadherent cells (lymphocytes, granulocytes, erythrocytes) and leaving only adherent cells (monocytes). Control cultures of BMEC were a cell line (BMEC23) derived from a bovine mammary explant and characterized as epithelial (3). Cultures of all cell types were maintained at 3TC in a humidified atmosphere of 5% C02 and the medium (DME-lO% FBS) changed once per week. Growth Determination
Cells constituting each epithelioid island were counted every 24 to 48 h in situ at l00x magnification with a grid in the microscope eyepiece to facilitate counting. Growth curves
957
for each island were constructed by plotting the logarithm of cell number on the y axis against time in culture on the x axis. The population doubling time (PDT) as an expression of growth rate was calculated from the steepest slope of the logarithmic portion of the curve. Population doubling level (PDL) as an expression of growth potential, was calculated for each island according to the formula: PDL = log (maximum cell number reached/starting cell number)/log 2. Cytochemistry
Presence of keratin in cultured cells was detected with an avidin-biotin immunoperoxidase system. The buffer used throughout was DPBS without Ca2+ and Mg2+ (GlBCO). After an initial wash with buffer and a lO-min fixation with 95% ethanol, nonspecific binding sites were blocked 20 min with normal goat serum (Vector Laboratories, Burlingame, CA; I: 74 in buffer). The primary antibody, rabbit antibovine keratin (Dako. Carpinteria, CA, #Ll824, prediluted by manufacturer I :250 in 10% .5 M Tris, pH 7.6, 90% .2 M PBS), was reacted for 30 min, and the secondary antibody, biotinylated goat antirabbit IgG (Vector Laboratories; 1:222 in buffer) was reacted for 30 min. Signal was amplified with avidin-biotinperoxidase complex (Vector) for 30 min and detected with the chromogen diamino-benzidine (DAB) (Sigma) 7 min. The DAB was prepared just before use by adding equal parts of a 1 mglml solution of DAB in .1 M Tris buffer, pH 7.2 to .02% of hydrogen peroxide prepared fresh in distilled water from a 30% stock. Cells were counterstained for 1 min with undiluted double strength hematoxylin (Lerner Laboratories. New Haven, CT, #S 3141-2). Buffer was used for rinses after primary and secondary antibody and amplification steps. The DAB and hematoxylin were rinsed with tap water. Presence of esterase enzymes was detected according to a standard method described in detail by Koski et al. (I5). Cells were fixed in an acetone and formaldehyde mixture, rinsed, dried, and reacted with an esterase substrate solution of pararosaniline, sodium nitrite, and a-naphthyl butyrate. Cells were rinsed and lightly counterstained with hematoxylin. Journal of Dairy Science Vol. 73, No.4, 1990
958
BUEHRING
Electron Microscopy
After growth in culture for 2 to 4 wk, cells' were fixed for thin-section electron microscopy (EM) in 2.5% glutaraldehyde (Polysciences. Wanington. PA; EM grade; vol/vol in .1 M of sodium cacodylate buffer. pH 7.2) for 3 h; postfixed in 1% osmium tetroxide (Polysciences; vol/vol in .1 M of sodium cacodylate buffer) for 1 h; stained with .5% uranyl acetate (Electron Microscopy Sciences, Fort Washington, PA; wt/wt in distilled water) for 2 h; passed through a graded series of ethanol and resin in ethanol; and embedded in a resin mixture (1 :21.6:25.2) of benzyldimethylamine. dodecenylsuccinic anhydride, and Araldite 6005 (Polysciences). Cells were prepared in situ and sectioned perpendicular to the substrate. All sections were examined with a Zeiss 109 electron microscope (Thornwood. NY). RESULTS
When inoculated into culture, pellets from centrifuged milk exhibited single cells, clumps of cells, bacteria. crystals. and debris. It was possible to see cells with segmented nuclei typical of granulocytes as well as cells and clumps of various sizes (2 to 15 cells) with nonsegmented nuclei. However. the debris was so extensive (Figure I), it was impossible to count individual cells accurately to detennine the mean number of viable cells per liter of milk. After the initial medium change 24 h later, most bacteria and debris were eliminated, and single cells and small epithelioid islands of cells were apparent, spread upon the substrate. Attachment was dependent upon the presence of FBS and did not occur in serum-free medium. even when supplemented with fibronectin (2 Ilg/ml; Collaborative Research, Waltham, MA). The epithelioid cells attached readily to all tissue culture vessels tried: 25-cm 2 flasks and 96-well and 24-well multitest plates (Corning. Corning, NY). In initial experiments, the appearance and quantity of epithelioid islands derived from nonfat milk did not differ subjectively from islands derived from whole milk. Therefore. nonfat milk was used for all subsequent experiments to avoid tenacious cream layers during centrifugation. All attached cells had nonsegmented nuclei. Occasional single cells with the granular. denJournal of Dairy Science Vol. 73. No.4. 1990
Figure 1. Pellet material centrifuged from milk. Note large rod-shaped crystals (e). pigmented plaques (P), and other debris obscuring individual cells (phase contrast, x313; bar 32 Il.
=
dritic appearance of macrophages were present. These made up
