DEVELOPMENTAL
BIOLOGY
137,425-433 (1990)
Inhibition of Mouse Mammary Ductal Morphogenesis and DownRegulation of the EGF Receptor by Epidermal Growth Factor S. COLEMAN AND C. W. DANIEL Department of Biology, University
Calzfomzia, Santa
of
CTUZ, California
95064
Accepted October 18, 1989 EGF, initially demonstrated to be a potent mitogen for a variety of cell types, has more recently been shown to inhibit proliferation of several cell lines. Few studies, however, have addressed the effects of EGF on growth and morphogenesis of tissues in vivo, particularly with regard to EGF as a possible inhibitor. We now demonstrate that EGF treatment of vigorously growing mammary ducts, administered directly to the glands by slow release plastic implants, inhibited normal ductal growth. Inhibition was restricted to the region around the implant and untreated glands in the same animal were normal, indicating direct effects of EGF. EGF-treated end buds were smaller and demonstrated reduced levels of DNA synthesis, although remnants of a stem (cap) cell layer persisted. Full inhibition of growth occurred within 3 days of implantation and required extended exposure to EGF, since treatment of 5 hr or less had no effect on ductal growth. At the lower inhibitory doses tested, growth resumed within 8 days, indicating reversibility of inhibition. No lobuloalveolar or hyperplastic response was seen. ‘%I-EGF autoradiography revealed that ductal growth inhibition was preceded by the disappearance of EGF receptors located in the cap cell layer of the end bud epithelium and in stromal cells adjacent to the buds. These results, in conjunction with our previous evidence demonstrating the growthstimulatory effect by EGF on nonproliferating mammary ducts, suggest a growth regulatory role for EGF in mouse mammary duetal morphogenesis. o 1990 Academic PRSS. IIIC. INTRODUCTION
Growth and morphogenesis of mammary ducts in the juvenile mouse are controlled by systemic and local regulators acting on bulbous terminal end buds (Topper and Freeman, 1980). Classical experiments demonstrated that ovarian and pituitary mammogens are necessary for end bud maintenance and growth, while recent data indicate that peptide growth factors may also be involved (Dembinski and Shiu, 1987; Silberstein and Daniel, 198’7b). EGF is known to promote growth of a variety of nontransformed epithelial and connective tissue cell types in vitro (Carpenter and Cohen, 1979). Recently, EGF has been shown to inhibit proliferation of several cell lines (Imai et aL, 1982; Knowles et aZ., 1985; Kamata et all, 1986; Tsao and Liu, 1988; Koyasu et a& 1988), and a few studies have demonstrated EGF inhibition of growth and morphogenesis of tissues. In tooth organ cultures, EGF interfered with organ morphogenesis and inhibited DNA synthesis in dental mesenchymal cells (Partanen et ah, 1985). EGF injection into mouse embryos in viva inhibited DNA synthesis and down-regulated the EGF receptor (Adamson et al., (1981), Adamson and Warshaw, 1982). Mitogenic effects of EGF on cultured mammary gland have been previously reported. EGF stimulated proliferation of.epithelial cells isolated from the glands of nulliparous (Richards et a.!., 1982, Imagawa et uL, 1985; Kawamura et uL, 1986) and pregnant/lactating
mice (Taketani and Oka, 1983a,b). In organ cultures of mammary gland from estrogen/progesterone-primed virgin mice, EGF induced lobuloalveolar growth (Tonelli and Sorof, 1980) and inhibited the synthesis of casein (Taketani and Oka, 1983b). We previously demonstrated that growth-quiescent glands in endocrine-ablated animals responded to EGF treatments in viva with normal growth and morphogenesis of mammary ducts (Coleman et uL, 1988). EGF receptors associated with cells in and around growing ducts were described. In the present study, we have used slow release plastic implants (EVAc) capable of releasing undenatured hormone, to locally treat, in situ, naturally proliferating ducts in endocrine-intact mice. In sharp contrast with earlier reports using endocrine-ablated mice, we found that implanted EGF caused a pronounced, prolonged, but fully reversible inhibition of ductal growth. Inhibitory doses of EGF resulted in the down-regulation of glandular EGF receptor, suggesting this as a mode of action. MATERIALS
AND METHODS
Muteriuk. Ethylene/vinyl acetate copolymer (EVAc, 40% vinyl acetate by weight) was a gift of DuPont. Epidermal growth factor (receptor grade) was purchased from Biomedical Technologies, Inc. Tritiated thymidine (68 Ci/mmole) was from ICN Biomedicals and iodine-125 (100 mCi/ml) was from Amersham. Tissue. Hormonally intact virgin female C5’i’ mice were 5 weeks of age and weighed 16 g at the time of use.
425
06121666/90 $3.00 Copyright All rights
0 1990 by Academic Press, Inc. of reproduction in any form reserved.
426
DEVELOPMENTALBIOLOGY
Preparation of implants. Implant preparation is described in detail elsewhere (Silberstein and Daniel, 1987a). Briefly, a lyophilized mixture of 25 mg BSA and various amounts of EGF were dispersed in 0.125 ml of EVAc that had been dissolved in dichloromethane (20% (wt/vol)). This mixture was quick-frozen and dried and the polymer matrix with entrapped chemical was then cut to size (typical implant weight, 0.2-l mg) and surgically implanted. BSA was used in all implants as a nonbioactive carrier protein; the BSA concentration determines the release rate of EGF (Murray et aL, 1983). EGF release from EVAc. The release kinetics of EGF was studied in vitro. A 35-mg pellet containing 100 ng ‘251-EGF and 25 mg BSA was cut into pieces ranging from 0.5 to 1.0 mg, sealed in porous nylon (Nitex) bags, and then placed in 2 ml of phosphate-buffered saline (PBS) at 37°C on a rotator (1 rpm). At each time point the bags were removed and the saline was counted in a Beckman Gamma 8000. Bags were then rinsed and placed in fresh saline. Counts obtained were adjusted for radioactive decay and compared to total counts added. Surgical implantation. The abdominal skin of Nembutal-anesthetized mice (60 pg/g body wt) was cut and pinned back exposing the No. 3 mammary glands. Iris scissors and Dumont forceps were used to place the implant in the No. 3 gland at the ductal tips. Histology. Glands were fixed overnight in Tellyesniczky’s fixative. For whole mount preparations, glands were defatted in three changes of acetone, hydrated through graded alcohols, stained with hematoxylin, dehydrated through graded alcohols to xylene, and photographed. After whole mount examination, pieces of tissue were embedded in paraffin and sectioned for DNA autoradiography. DNA autoradiography. Animals were injected intraperitoneally with 100 PCi of [3H]thymidine. After 40 min, the mammary glands were removed, fixed, processed, and paraffin sectioned by standard methods. Slides were dipped in Kodak NTB-2 emulsion diluted 1:l with water, exposed for 11 days, developed, and stained with hematoxylin/eosin. DNA labeling index. The labeling indices for end buds and surrounding stroma were determined by counting cells of representative end buds in at least three alternate 5-pm sections. Sections from EGF-treated glands and BSA-treated glands from 5-week old animals were counted. Injection of EGF. Small volumes (5-10 ~1) of EGF in 1% BSA:PBS were injected directly into the gland using a Hamilton microsyringe. EGF iodination. ‘%I-EGF was prepared by the lactoperoxidase method of Thorell and Johansson (1971) with minor modifications as described by Smith and Talamantes (1986). Labeled hormone was separated
VOLUME 137,199O
from unreactive Na1251by passage through a Sephadex G-25 column which had been preequilibrated with 0.1 M PO1, 1% BSA. Specific activity was determined by measuring the percentage incorporation of 1251into trichloroacetic acid-precipitable protein and was about 5 X lo5 cpm/ng. lz51-EGF autoradiography. The method of Green et al. (1983) was used. Briefly, fragments of tissue were incubated in Hank’s balanced salt solution containing 0.2 nM ‘251-EGF for 90 min at room temperature, rinsed, fixed overnight at 4°C in 4% paraformaldehyde:PBS, and dehydrated through graded alcohols to xylene. After paraffin embedding, 5-pm sections were then processed for autoradiography, as in DNA autoradiography methods, and exposed for 14 days. 1251-EGF autcrradiograph data analysis. Alternate serial sections were examined in three areas: (1) cap cells of end buds, (2) stromal cells adjacent to ducts, and (3) stromal cells distant from ducts. Silver grains over 100 cells were counted for each area in alternate serial sections from two different experiments. This was done for both experimental sections and control sections where tissues were incubated in labeled plus excess (1 pg) unlabeled EGF. Specific binding was determined by subtracting control values, using excess unlabeled EGF, from experimental values for each area. Specific binding for areas 1,2, and 3 were compared using a Student t test. RESULTS
The effect of implanted EGF on proliferating ducts in mammary glands of 5-week-old mice is shown in Fig. 1. In control glands implanted with EVAc containing only carrier protein, end buds appeared large and bulbous, even when in close proximity to the implant (Fig. lA, solid arrows). After 3 days of exposure to an inhibitory dose of EGF, end buds near the implant had involuted (Fig. lB), and the ductal tips resembled mammary ducts which had ceased to proliferate due to normal processes of tissue regulation that occur when ductal structures approach the edge of the fat pad (Fig. lA, open arrows). The effect of EGF on end bud growth was restricted to the implanted gland, since other glands in the same animal contained ducts of normal shape and size. Doses of EGF between 2 and 10 /*g per gland, the lower doses of EGF tested, resulted in ductal growth inhibition just around the implant; end buds were still present in the EGF-treated gland distant from the implant (not shown). The effect of EGF treatment on glandular histomorphology and DNA synthesis is shown in Fig. 2. Glands treated with BSA showed typical end bud morphology, with a multilayered, bulbous epithelial end bud covered by a distinct cap cell layer (solid arrow). End buds con-
COLEMAN AND DANIEL
i&use
Mammary
Dudal
Morphogenesis
427
FIG. 1. Photomicrographs of mammary glands illustrating the inhibitory effect of EGF on whole gland morphology. Five-week-old virgin mice were implanted with EVAc pellets containing EGF in the right No. 3 mammary gland and BSA controls in the left (contralateral) No. 3 gland for 3 days (bar = 2 mm). (A) BSA control. Normally developing ductal structures are present around implant (*) with large, bulbous end buds apparent at the ductal tips (solid arrows). Blunt tipped, growth-quiescent ducts are present along the edge of either side of the fat pad (open arrows). (B) Fifteen micrograms EGF treatment. Note the absence of large end buds, and the presence of only slightly rounded ductal tips (arrows) similar to the growth-quiescent ducts seen in A.
tained numerous DNA synthetic epithelial cells (open arrows), particularly in the cap cell layer. DNA synthetic cells were also apparent in the stroma. EGF treatment for 3 days resulted in a much smaller structure, with a reduction in DNA synthetic epithelial cells. In other respects the miniature end buds appeared normal, with a multilayered epithelium and the remnants of a cap cell layer. EGF-treated end bud tips were free of stromal condensation normally present around the mature, nonproliferating duct. DNA labeling indices were determined for treated and untreated glands (Table 1). EGF treatment reduced the percentage of DNA synthetic epithelial cells by 83% between 24 and 74 hr after EVAc implantation. Contralateral control glands showed no significant change in labeling indices of the epithelial cells (Student’s t test, P < 0.01). EGF treatment did not affect the indices of the stromal cell population within 3 days, although the percentage labeled stromal cells was slightly increased on the contralateral side. The loss of end buds 3 days after EGF treatment was dependent upon the concentration of EGF in the EVAc implant (Fig. 3). Doses up to 1 pg EGF per implant had little or no effect on the number of end buds present per gland. Amounts of EGF greater than 2 pg resulted in a dose-related decrease in end bud numbers per gland. At
20 pg no end buds were present within a one-half-cmradius of the implant. End bud inhibition by EGF occurred within 3 to 4 days, and the length of inhibition depended upon the amount of implanted EGF (Fig. 4). At 4 pg EGF per implant, end buds were again present 8 days after initial exposure to EGF, with full recovery by 14 days. At 21 pg EGF, end buds were just beginning to reappear 14 days after implantation. We were interested in determining whether the inhibition of ductal growth required long-term release of EGF or if the observed effect could be attributed solely to the initial surge of EGF released directly after implantation (Fig. 5). Instead of incorporating the EGF into EVAc, a saline solution containing either BSA or EGF/BSA was injected directly into the gland, immediately ahead of the array of end buds. Injections of 3 to 15 fig EGF into the mammary gland showed no effect on end bud numbers as compared with the BSA controls. Since the majority of protein is released from the EVAc implant after 5 hr (Fig. 5, graph), experiments were designed to determine whether the inhibitory EGF dose was delivered in the initial 5-hr treatment or by the low level of EGF released after 5 hr. Implants containing 15 fig EGF were placed in the gland and then removed after 5 hr. Glands were removed for examina-
428
DEVELOPMENTALBIOLOGY
VOLUME137,199O
FIG. 2. DNA autoradiographs of end bud sections from EGF-treated and control glands in 5 week-old animals (bar = 50 pm). (A) End bud treated with BSA for 3 days. Ductal stem (cap) cell layer is present (solid arrow), and both cap and lumenal epithelial cells show extensive label characteristic of normal end buds (open arrow). (B) Ductal tip from gland treated with implant containing 20 pg EGF for 3 days. Note that the overall size is greatly reduced, remnants of a cap cell layer (solid arrow) are still apparent, however. Several DNA synthetic cells are present in the epithelium and stroma (open arrow).
tion after 3 days. EGF treatment for 5 hr followed by EGF removal, had little effect on end bud number when compared with BSA controls. However, glands treated with implants in which EGF had been prereleased into saline solution for 5 hr prior to implantation showed a significant decrease in end bud number after 3 days when compared with contralateral BSA controls. The effect of EGF treatment on EGF receptor distribution was studied in sections of glands from hormonally intact, 5-week-old animals, using lz51autoradiography (Fig. 6). Glands with BSA control implants showed the heaviest concentration of label over stromal cells adjacent to ducts, where the cells and extracellular matrix form a tunic around the flank region of the end bud (Figs. 6A and 6B). Label was also present over some stromal cells located at a distance from ducts. Epithelial cell label was restricted to the cap cells of end buds, as previously reported (Coleman et ak, 1988). Controls, where tissue was incubated with labeled plus lOOO-fold excess unlabeled EGF showed few silver grains over either stromal or epithelial cells (Figs. 6E and 6F). EGF treatment of end buds in vivo resulted in a loss of label over stromal cells and end bud cap cells within 12 hr of treatment (Figs. 6C and 6D). Label did not reappear up to 3 days after EGF implantation (results not shown). The labeling pattern observed in the ‘%I-EGF autora-
diographs was quantified by counting cell-associated silver grains over an average of 100 cells/area/section in multiple alternate tissue sections and is reported as the number of silver grains per cell (Table 2). In control tissue, the heaviest label was present over duct-associated stromal cells and, to a significantly lesser degree, over stromal cells distant from ducts and epithelial cap cells. After EGF treatment of tissue in vivo, no significant label (as determined by Student’s t test) was present over the cap cells of end buds and stromal cells distant from ducts. Label over tunic-associated stromal cells was decreased by 85% (P < 0.01). DISCUSSION
Results from our present study indicate that exogenous EGF, locally delivered, influences the mammary gland in a bifunctional manner. The capacity for EGF to both stimulate or inhibit tissue growth depending upon conditions has been demonstrated in several cell lines in culture, as well as in embryonic development. A role for EGF as a natural growth regulator in vivo has yet to be reported, however. In an earlier study, we described the reappearance of proliferating ductal structures in growth-quiescent glands after treatment with EGF in situ, demonstrating the ability of EGF to initiate and maintain a com-
COLEMAN AND DANIEL
Mouse Mammary
Ductal
429
Morph,ogenesis
TABLE 1 TIME COURSE OF EGF EFFECB ON MAMMARY CELL DNA SYNTHESIS Percentage
labeled cells Stroma”
End bud epithelium Time (hr) 12 24 74
Control”
23.6 f
21.3 f 7.8
6.5&"
Control
Treatment
Treatmentb
cwf
(27)
26.4 + 13.3 (24 3.7 rf: 2.7 (24
19.8 f 4.6 (18) 24.3 f 5.9 cw
(21) 2.5 f 0.9 (24)
plex morphogenetic pathway (Coleman et aL, 1988). This EGF-induced ductal growth was characterized by numerous DNA synthetic mammary cells, apparently normal tissue growth, and typical morphogenetic patterning. Here, we have extended these findings to describe how EGF treatment of rapidly proliferating ducts in mammary glands from hormonally intact, subadult mice resulted in substantial, prolonged growth inhibition, with no indication of hyperplastic growth. 14
4
6
6
10
12
14
16
1614-
1.5 + 0.8 (21) 3.3 + 1.5 (15) 6.8 + 2.0 (24
3.0 f 1.4 (24) 3.9 + 1.6
a Stromal cells were counted within a 0.2-mm radius of the relevant end bud. b Right No. 3 mammary glands implanted with EVAc pellets containing 20 pg EGF. ’ Contralateral glands implanted with EVAc pellets containing BSA. d Results are expressed as the percentage of labeled cells. See Materials and Methods for labeling index determination. e Mean + SD. fThe total number of sections examined.
2
2 4 0
18
20
Pug EGF
FIG. 3. Dose dependency of the ductal response to EGF. All points represent mean number of end buds f SE with an average of five glands per dose. (0) Growing glands were treated with EVAc pellets containing 0.4 to 20 pg EGF for 3 days. At the 3-day time point, a 20 pg EGF dose resulted in complete inhibition. (0) Control glands (contralateral to treatment) were treated with a BSA implant.
1
2
3
4
5
6
7 6 DAY
9
10
11
12
13
14
15
FIG. 4. Time course of the EGF response in 5-week-old animals. All points represent mean number of end buds + SE with an average of four glands per time point. (A) Twenty-one micrograms EGF/gland; end buds disappeared by Day 3 and began to reappear at Day 14. (0) Four micrograms EGF/gland; end buds were completely inhibited by Day 4 recovering to normal numbers by Day 14. (0) BSA control; increase in end bud number at the later time points was due to the growth and increased branching of the ducts during the course of the experiment.
EGF-induced inhibition of ductal growth appeared to occur through normal morphogenetic pathways present in the hormonally intact, virgin animal, as determined by the following criteria: (1) EGF treatment of proliferating ducts for 3 days resulted in end bud-like structures of reduced size, (2) a remnant cap (stem) cell layer persisted in many of the small end buds, (3) EGF-induced inhibition did not result in terminal differentiation of ducts since growth resumed 7-14 days after implantation, and (4) no alveolar or hyperplastic growth was observed. The effects on end bud growth by EGF can be compared with TGF-& another putative ductal growth regulator. Unlike EGF, TGF-pl does not stimulate growth of nonproliferative ducts in endocrine-ablated animals, but like EGF, it is a potent inhibitor of rapidly growing ducts in hormonally intact, virgin mice (Silberstein and Daniel, 198713;Daniel et CL!.,1989). Interestingly, the characteristics of ductal growth inhibition by the two hormones are quite different, suggesting separate mechanisms of inhibition. Morphologically, EGF-treated end buds were reduced in size, but without heavy stromal accumulations present at the ductal tips as seen within 12 hr of TGF-& treatment. The cap cell layer, still present in a vestigial form 3 days after EGF treatment, is lost within 24 hr of TGF-01 treatment. The effect of TGF-& on end bud DNA syn-
430
DEVELOPMENTAL BIOLOGY
VOLUME137, 1990 TREATMENT EGF
Inject
1 CONTROL
Implant
I ;E
q/ z
, 0
12
+-,+--; 3
4
5
24
46
9
76
TIME (HOURS)
FIG. 5. Treatment required for ductal growth inhibition.
(Diagram) Treatment values indicate mean number of end buds f SE present in glands either (A) injected with 3 pg EGF, (B) implanted for 5 hr with EVAc containing 15 wg EGF (implants were removed surgically after 5 hr), or (C) treated with prereleased implants containing 15 pg EGF (implants had been incubated in PBS for 5 hr prior to implantation). All glands were removed from the animal after 3 days. Control values represent BSA-treated glands. Glands treated with implants containing EGF which had been prereleased in vitro had a mean number of 0.7 end buds per gland and were significantly lower than the BSA control. Injection or 5-hr EGF treatment (A,B) had no significant effect on end bud numbers when compared to their controls, nor were they different from each other (as determined by Student’s t test, P = 0.01). (n) Represents number of glands tested. (Graph) EGF release from EVAc was quantified in vitro with lz51-labeled EGF. Maximum release from EVAc occurred during the first hour; less than 1% of the total dose of EGF was released per hour after Hour 4.
thesis is rapid; within 12 hr, TGF-&-treated ducts contained sharply reduced numbers of DNA synthetic cells, whereas EGF-induced inhibition of DNA synthesis occurs after 24 hr. In contrast to the inhibitory response to EGF, TGF-& inhibition of end buds did not require extended exposure; a single injection of the growth factor was effective (Coleman et al, 1988; Silberstein and Daniel, 198713;Daniel et aH, 1989). One noteworthy similarity between TGF+& and EGF inhibition was the restriction of inhibition of DNA synthesis to the epithelial cells. Stromal cells did not respond to treatment by either growth factor with changes in the number of DNA synthetic cells. Stromal cells adjacent to ducts respond to TGF-P1 with increased synthesis of collagen and sulfated glycosaminoglycans (Silberstein et CCL,submitted). The effects of EGF on extracellular matrix components were not investigated, but EGF has been shown to stimulate the synthesis of fibronectin (Gospodarowicz, 1981) and hyaluronate (Lembach, 1976). Receptors for EGF are located in both epithelial and stromal cell populations, indicating the ability of these cell populations to bind
EGF, whereas the location of mammary receptors for & are unknown at present. The mechanism of action of EGF-induced ductal growth inhibition is suggested by the disappearance of the EGF receptor observed prior to growth inhibition. In mammary glands undergoing rapid ductal elongation, EGF receptors are distributed in stromal cells located along the mammary ducts, in stromal cells located at a distance from ductal structures, and in epithelium forming the end bud cap cell layer (Coleman et aL, 1988). Addition of EGF resulted in nearly complete ablation of EGF receptors in the cap cells and an 85% reduction in duct-associated stromal cells within 12 hr of treatment. A correlation between down-regulation of the EGF receptor and EGF-induced inhibition of DNA synthesis has been demonstrated in two other systems, the bud stage of tooth development and in mouse embryos (Abbott and Pratt, 1988; Adamson and Warshaw, 1982). The studies of Adamson and co-workers present an interesting analogy to the effect of EGF in the mammary gland. Treatment of mouse embryos with up to 5 ng EGF
FIG. 6. Autoradiographs of ‘%I-EGF binding in glands from B-week-old animals implanted with EGF in tivo 12 hr prior to organ culture autoradiography. Mammary gland was incubated with 0.2 nM’251-EGF ( see Materials and Methods, bar = 50 am). (A,B) Light- and dark-field micrographs of end bud not treated with exogenous EGF. Label is heaviest over the stromal cells forming the fibrous ductal tunic (open arrows). The cap cell layer also labels (solid arrow), while little label is present over the rest of the epithelial cells. (C,D) End bud from gland treated with implant containing 15 fig EGF for 12 hr. Note absence of label over condensed stromal cells at the flank region of the end bud (open arrow) and over the cap cell layer of the epithelium (solid arrow). (E,F) End bud as in A and B except 1 pg (one thousand-fold excess) unlabeled EGF was added to the incubation medium. Label over the stromal cells and the cap cells is absent. 431
432
DEVELOPMENTALBIOLOGY
TABLE 2 EGF RECEPTORDOWN-REGULATIONMEASUREDBY ‘%I-EGF
BINDING
Treatment Zone Cap cells Stromal cells” (tunic-associated) Stromal cells (non-tunic-associated)
EGF” 0.1 f 0.5”rd 1.5 + 0.2 0.3 * 0.1
BSAb 3.4 + 0.5 10.5 f 0.9 1.8 + 0.1
a Glands were implanted with 15 fig EGF and removed for ‘%I-EGF autoradiography 12 hr later. bGlands were implanted with BSA control. c Silver grains per cell. Alternate serial sections were examined in the three areas. Silver grains over 100 cells were counted per section for each area. Sections were examined from two different experiments for both experimental and control tissues. Specific binding was determined by subtracting values obtained from tissues incubated in labeled EGF plus excess (1 fig) unlabeled EGF from experimental values. d Mean +_SE. “Stromal cells associated with the fibrotic tunic surrounding end buds.
caused an increase in DNA synthesis, yet doses of EGF greater than 10 ng resulted in reduced DNA synthetic activity in addition to down-regulation of the EGF receptor (Adamson et aZ., 1981; Adamson and Warshaw, 1982). Adamson and Warshaw (1982) proposed that embryonic tissue was already maximally responding to endogenous EGF, and prolonged exposure to increased concentrations of the growth factor down-regulated the receptor, thereby preventing a maximal mitogenic response. Our results suggest an analogous explanation for the bifunctionality of EGF in the mammary gland. Additional support for this hypothesis is the requirement for prolonged exposure to EGF for ductal growth inhibition; short-term exposure to exogenous EGF did not elicit an inhibitory response. Recovery of EGF receptors after down-regulation in mouse embryos can occur within 14 hr after injection of EGF (Adamson and Warshaw, 1982), and EGF receptor biosynthesis in A431 cells can occur within 60 min (Carlin and Knowles, 1984). To achieve and maintain the growth-quiescent state for a period of days in the mammary gland, it would be expected that extended exposure inducing prolonged down-regulation of the receptor would be required, preventing full recovery of the EGF receptor population. EVAc implants are capable of releasing low levels of incorporated hormone for at least 7 days. Sources of endogenous EGF in the ductal stage of mammary gland development are speculative at this point. EGF is present in the serum of virgin, female mice at levels reported between 0.17 and 1.2 rig/ml (Byyny et al., 1974; Kurachi and Oka, 1985). Autocrine or
VOLUME137,199O
paracrine sources of EGF in the mammary gland have not been identified. Doses of EGF inhibitory for end bud growth are not accurately described by the amounts of hormone incorporated into an EVAc pellet. Estimates of true inhibitory doses of EGF must take into account factors influencing the concentration of EGF around the implant such as the heavily vascularized stroma, in which released EGF may escape into the vascular system. These variables have been discussed in greater detail in relation to release of TGF-/3i (Daniel et aL, 1989). Estimation of inhibitory dose must also reflect the observation that injection of EGF or the doses of hormone released within the first 5 hr after implantation of the EVAc pellet were not inhibitory for end bud growth. Calculations must be based on the less than 2% released from the implant between Hour 5 and Hour 24 (approx, 0.1% per hour). For example, a pellet containing 20 pg EGF would release an estimated 20 ng EGF per hour 5 hr after implantation. Inhibitory doses of EGF in tissue culture studies and administered exogenously in viva range from 10 to 50 rig/ml. (Imai et aL, 1982, Knowles et aZ., 1985; Kamata et ak, 1986; Koyasu et cd, 1988; Tsao and Liu, 1988). Other possible mechanisms of EGF-induced ductal growth inhibition must be considered. Exogenous EGF may be eliciting synthesis and/or secretion of an unknown inhibitor acting on a pathway other than through the EGF receptor, or EGF itself may be toxic to ductal cells. Neither of these possibilities seem likely in light of the results showing that EGF promotes ductal growth and morphogenesis in endocrine-ablated mice. It is apparent that the mammary gland is extremely sensitive to EGF, in that the hormone can act as either a mitogen or a growth inhibitor depending on the growth state of the gland. In both cases, increasing concentrations of EGF over endogenous levels resulted in a normal ductal morphology rather than abnormal growth. It is very probable that the influence of other mammotrophic hormones plays a role in determining the magnitude and nature of the EGF response. This sensitivity, together with the apparent linkage between down-regulation of the EGF receptor and ductal growth inhibition, suggests that EGF has a normal role in mammary growth and morphogenesis. We thank Gary Silberstein for scientific and editorial advice, and Phyllis Strickland and Kathy Van Horn for skillful technical assistance. This work was supported by PHS Grant CA-45231 from the National Cancer Institute. REFERENCES
ABBOTT,B. D., and PRATT, R. M. (1988). EGF receptor expression in the developing tooth is altered by exogenous retinoic acid and EGF. Dev. Biol. 128,300-304.
COLEMANAND DANIEL
MOUSQMammary Dudal Marphogenesis
ADAMSON,E. D., DELLER, M. J., and WARSHAW,J. B. (1981). Functional EGF receptors are present on mouse embryo tissues. Nature (London) 291,656-659. ADAMSON, E. D., and WARSHAW, J. B. (1982). Down-regulation of epidermal growth factor receptors in mouse embryos. De-v.Biol SO, 430-434.
BYYNY, R. L., ORTH, COHEN, S., and DOYNE, E. S. (1974). Epidermal growth factor: Effects of androgens and adrenergic agents. En& crinology 95,776-782. CARLIN, C. R., and KNOWLES,B. B. (1984). Biosynthesis of the epiderma1 growth factor receptor in human epidermoid carcinoma-derived A431 cells. J. Biol. Chem 259( 12). ‘7902-7908. CARPENTER, G., and COHEN, S. (1979). Epidermal growth factor. Annu. Rev. Biochem 48,193-216. COLEMAN,S., SILBERSTEIN,G. B., and DANIEL, C. W. (1988). Ductal morphogenesis in the mouse mammary gland: Evidence supporting a role for epidermal growth factor. Dev. Biol. 127,304-315. DANIEL, C. W., SILBERSTEIN,G. B., VAN HORN, K., STRICKLAND,P., and ROBINSON, S. (1989). TGF-&-induced inhibition of mouse mammary ductal growth: Developmental specificity and characterization. Dev. Biol. 135,20-30. DEMBINSKI, T. C., and SHIU, R. P. C. (1987). Growth factors in mammary gland development and function. In “The Mammary Gland: Development, Regulation, and Function” (M. C. Neville and C. W. Daniel, Eds.), pp. 355-381. Plenum, New York. GOsPODAROwICZ,D. (1981). Epidermal and nerve growth factors in mammalian development. Annu. Rev. PhysioL 43,251-263. GREEN, M. R., BASKETTER,D. A., COUCHMAN,J. R., and REES, D. A. (1983). Distribution and number of epidermal growth factor receptors in skin is related to epidermal cell growth. Dew. Biol. 100, 506-512.
IMAGAWA, W., TOMOOKA,Y., HAMAMOTO, S., and NANDI, S. (1985). Stimulation of mammary epithelial cell growth in vitro: Interaction of epidermal growth factor and mammogenic hormones. End,ocrinologg 116,1514-1524.
IMAI, Y., LEUNG, C. K. H., FRIESEN, H. G., and SHIU, R. P. C. (1982). Epidermal growth factor receptors and effect of epidermal growth factor on growth of human breast cancer cells in long-term culture. Cancer Res. 42,4394-4398. KAMATA, N., CHIDA, K., RIKIMARU, K., HORIKOSHI,M., ENOMOTO,S., and KUROKI, T. (1986). Growth-inhibitory effects of epidermal growth factor and overexpression of its receptors on human squamous cellcarcinomas in culture. Cancer Res. 46,1648-1653. KAWAE~~URA, K., ENA~I, J., ENAMI, S., KOEZUKA,M., KOHMOTO,K., and KOGA, M. (1986). II. Growth and morphogenesis of mouse mammary epithelial cells cultured in collagen gels: Stimulation by hormones, epidermal growth factor, and mammary fibroblast-conditioned medium factor. Proc. Japan Acad Ser. B 62,5+X KNOWLES,A., SALAS-PRATO,M., and VILLELA, J. (1985). Epidermal growth factor inhibits growth while increasing the expression of an
433
ecto-Ca’+-ATPase of a human hepatoma cell line. Biochem. Biophys. Res. Commun 126.8-14. KOYASU, S., KADOWAKI, T., NISHIDA, E., TOBE, K., ABE, E., KASUGA, M., SAKAI, H., and YAHARA, I. (1988). Alteration in growth, cell morphology, and cytoskeletal structures of KB cells induced by epidermal growth factor and transforming growth factor 8. Exp. Cell Res. 176,107-116. KURACHI, H., and OKA, T. (1985). Changes in epidermal growth factor concentrations of submandibular gland, plasma, and urine of normal and sialoadenectomized female mice during various reproductive stages. J. Endocrinol. 106,197-202. LEMBACH, K. J. (1976). Enhanced synthesis and extracellular accumulation of hyaluronic acid during stimulation of quiescent human fibroblasts by mouse epidermal growth factor. J. Cell Physiol. 89, 277-288.
MURRAY,J. B., BROWN,L., LANGER, R., and KLAGSBURN,M. (1983). A micro sustained release system for epidermal growth factor. In vitro 19,743-748. PARTANEN, A-M., EKBLOM, P., and THESLEFF, I. (1985). Epidermal growth factor inhibits morphogenesis and cell differentiation in cultured mouse embryonic teeth. Deu. Biol. 111,84-94. RICHARDS, J., GUZMAN, R., KONRAD, M., YANG, J., and NANDI, S. (1982). Growth of mouse mammary gland end buds cultured in a collagen gel matrix. Exp. Cell Res. 141,433-443. SILBERSTEIN,G. B., and DANIEL, C. W. (1987a). Investigation of mouse mammary ductal growth regulation using slow-release plastic implants. J. Dairy Sci. 70, 1981-1990. SILBERSTEIN,G. B., and DANIEL, C. W. (1987b). Reversible inhibition of mammary gland growth by transforming growth factor-& Science 237,291-293. SMITH, W. C., and TALAMANTES, P. (1986). Characterization of the mouse placental epidermal growth factor receptor: Changes in receptor number with day of gestation. Placenta 7,511-522. TAKETANI, Y., and OKA, T. (1983a). Possible physiological role of epidermal growth factor in the development of the mouse mammary gland. FEBS I&. 152,256-260. TAKETANI, Y., and OKA, T. (198313).Epidermal growth factor stimulates cell proliferation and inhibits functional differentiation of mouse mammary epithelial cells in culture. Endocrinology 113, 871-877.
THORELL,J. I., and JOHANSSON,B. G. (1971). Enzymatic iodination of polypeptides with I251to high specific activity. B&him Biophys. Acta. 251,363-369. TONELLI,Q. J., and SOROF,S. (1980). Epidermal growth factor requirement for development of cultured mammary gland. Nature (London) 285,250-252. TOPPER,Y. J., and FREEMAN,C. S. (1980). Multiple hormone interactions in the development biology of the mammary gland. Phgsiol. Rev. So;,1049-1106. TSAO, M.-S., and LIU, C. (1988). Inhibition of growth of early passage normal rat liver epithelial cell lines by epidermal growth factor. Lab. Invest. 58,636-642.
BIOLOGY
137,425-433 (1990)
Inhibition of Mouse Mammary Ductal Morphogenesis and DownRegulation of the EGF Receptor by Epidermal Growth Factor S. COLEMAN AND C. W. DANIEL Department of Biology, University
Calzfomzia, Santa
of
CTUZ, California
95064
Accepted October 18, 1989 EGF, initially demonstrated to be a potent mitogen for a variety of cell types, has more recently been shown to inhibit proliferation of several cell lines. Few studies, however, have addressed the effects of EGF on growth and morphogenesis of tissues in vivo, particularly with regard to EGF as a possible inhibitor. We now demonstrate that EGF treatment of vigorously growing mammary ducts, administered directly to the glands by slow release plastic implants, inhibited normal ductal growth. Inhibition was restricted to the region around the implant and untreated glands in the same animal were normal, indicating direct effects of EGF. EGF-treated end buds were smaller and demonstrated reduced levels of DNA synthesis, although remnants of a stem (cap) cell layer persisted. Full inhibition of growth occurred within 3 days of implantation and required extended exposure to EGF, since treatment of 5 hr or less had no effect on ductal growth. At the lower inhibitory doses tested, growth resumed within 8 days, indicating reversibility of inhibition. No lobuloalveolar or hyperplastic response was seen. ‘%I-EGF autoradiography revealed that ductal growth inhibition was preceded by the disappearance of EGF receptors located in the cap cell layer of the end bud epithelium and in stromal cells adjacent to the buds. These results, in conjunction with our previous evidence demonstrating the growthstimulatory effect by EGF on nonproliferating mammary ducts, suggest a growth regulatory role for EGF in mouse mammary duetal morphogenesis. o 1990 Academic PRSS. IIIC. INTRODUCTION
Growth and morphogenesis of mammary ducts in the juvenile mouse are controlled by systemic and local regulators acting on bulbous terminal end buds (Topper and Freeman, 1980). Classical experiments demonstrated that ovarian and pituitary mammogens are necessary for end bud maintenance and growth, while recent data indicate that peptide growth factors may also be involved (Dembinski and Shiu, 1987; Silberstein and Daniel, 198’7b). EGF is known to promote growth of a variety of nontransformed epithelial and connective tissue cell types in vitro (Carpenter and Cohen, 1979). Recently, EGF has been shown to inhibit proliferation of several cell lines (Imai et aL, 1982; Knowles et aZ., 1985; Kamata et all, 1986; Tsao and Liu, 1988; Koyasu et a& 1988), and a few studies have demonstrated EGF inhibition of growth and morphogenesis of tissues. In tooth organ cultures, EGF interfered with organ morphogenesis and inhibited DNA synthesis in dental mesenchymal cells (Partanen et ah, 1985). EGF injection into mouse embryos in viva inhibited DNA synthesis and down-regulated the EGF receptor (Adamson et al., (1981), Adamson and Warshaw, 1982). Mitogenic effects of EGF on cultured mammary gland have been previously reported. EGF stimulated proliferation of.epithelial cells isolated from the glands of nulliparous (Richards et a.!., 1982, Imagawa et uL, 1985; Kawamura et uL, 1986) and pregnant/lactating
mice (Taketani and Oka, 1983a,b). In organ cultures of mammary gland from estrogen/progesterone-primed virgin mice, EGF induced lobuloalveolar growth (Tonelli and Sorof, 1980) and inhibited the synthesis of casein (Taketani and Oka, 1983b). We previously demonstrated that growth-quiescent glands in endocrine-ablated animals responded to EGF treatments in viva with normal growth and morphogenesis of mammary ducts (Coleman et uL, 1988). EGF receptors associated with cells in and around growing ducts were described. In the present study, we have used slow release plastic implants (EVAc) capable of releasing undenatured hormone, to locally treat, in situ, naturally proliferating ducts in endocrine-intact mice. In sharp contrast with earlier reports using endocrine-ablated mice, we found that implanted EGF caused a pronounced, prolonged, but fully reversible inhibition of ductal growth. Inhibitory doses of EGF resulted in the down-regulation of glandular EGF receptor, suggesting this as a mode of action. MATERIALS
AND METHODS
Muteriuk. Ethylene/vinyl acetate copolymer (EVAc, 40% vinyl acetate by weight) was a gift of DuPont. Epidermal growth factor (receptor grade) was purchased from Biomedical Technologies, Inc. Tritiated thymidine (68 Ci/mmole) was from ICN Biomedicals and iodine-125 (100 mCi/ml) was from Amersham. Tissue. Hormonally intact virgin female C5’i’ mice were 5 weeks of age and weighed 16 g at the time of use.
425
06121666/90 $3.00 Copyright All rights
0 1990 by Academic Press, Inc. of reproduction in any form reserved.
426
DEVELOPMENTALBIOLOGY
Preparation of implants. Implant preparation is described in detail elsewhere (Silberstein and Daniel, 1987a). Briefly, a lyophilized mixture of 25 mg BSA and various amounts of EGF were dispersed in 0.125 ml of EVAc that had been dissolved in dichloromethane (20% (wt/vol)). This mixture was quick-frozen and dried and the polymer matrix with entrapped chemical was then cut to size (typical implant weight, 0.2-l mg) and surgically implanted. BSA was used in all implants as a nonbioactive carrier protein; the BSA concentration determines the release rate of EGF (Murray et aL, 1983). EGF release from EVAc. The release kinetics of EGF was studied in vitro. A 35-mg pellet containing 100 ng ‘251-EGF and 25 mg BSA was cut into pieces ranging from 0.5 to 1.0 mg, sealed in porous nylon (Nitex) bags, and then placed in 2 ml of phosphate-buffered saline (PBS) at 37°C on a rotator (1 rpm). At each time point the bags were removed and the saline was counted in a Beckman Gamma 8000. Bags were then rinsed and placed in fresh saline. Counts obtained were adjusted for radioactive decay and compared to total counts added. Surgical implantation. The abdominal skin of Nembutal-anesthetized mice (60 pg/g body wt) was cut and pinned back exposing the No. 3 mammary glands. Iris scissors and Dumont forceps were used to place the implant in the No. 3 gland at the ductal tips. Histology. Glands were fixed overnight in Tellyesniczky’s fixative. For whole mount preparations, glands were defatted in three changes of acetone, hydrated through graded alcohols, stained with hematoxylin, dehydrated through graded alcohols to xylene, and photographed. After whole mount examination, pieces of tissue were embedded in paraffin and sectioned for DNA autoradiography. DNA autoradiography. Animals were injected intraperitoneally with 100 PCi of [3H]thymidine. After 40 min, the mammary glands were removed, fixed, processed, and paraffin sectioned by standard methods. Slides were dipped in Kodak NTB-2 emulsion diluted 1:l with water, exposed for 11 days, developed, and stained with hematoxylin/eosin. DNA labeling index. The labeling indices for end buds and surrounding stroma were determined by counting cells of representative end buds in at least three alternate 5-pm sections. Sections from EGF-treated glands and BSA-treated glands from 5-week old animals were counted. Injection of EGF. Small volumes (5-10 ~1) of EGF in 1% BSA:PBS were injected directly into the gland using a Hamilton microsyringe. EGF iodination. ‘%I-EGF was prepared by the lactoperoxidase method of Thorell and Johansson (1971) with minor modifications as described by Smith and Talamantes (1986). Labeled hormone was separated
VOLUME 137,199O
from unreactive Na1251by passage through a Sephadex G-25 column which had been preequilibrated with 0.1 M PO1, 1% BSA. Specific activity was determined by measuring the percentage incorporation of 1251into trichloroacetic acid-precipitable protein and was about 5 X lo5 cpm/ng. lz51-EGF autoradiography. The method of Green et al. (1983) was used. Briefly, fragments of tissue were incubated in Hank’s balanced salt solution containing 0.2 nM ‘251-EGF for 90 min at room temperature, rinsed, fixed overnight at 4°C in 4% paraformaldehyde:PBS, and dehydrated through graded alcohols to xylene. After paraffin embedding, 5-pm sections were then processed for autoradiography, as in DNA autoradiography methods, and exposed for 14 days. 1251-EGF autcrradiograph data analysis. Alternate serial sections were examined in three areas: (1) cap cells of end buds, (2) stromal cells adjacent to ducts, and (3) stromal cells distant from ducts. Silver grains over 100 cells were counted for each area in alternate serial sections from two different experiments. This was done for both experimental sections and control sections where tissues were incubated in labeled plus excess (1 pg) unlabeled EGF. Specific binding was determined by subtracting control values, using excess unlabeled EGF, from experimental values for each area. Specific binding for areas 1,2, and 3 were compared using a Student t test. RESULTS
The effect of implanted EGF on proliferating ducts in mammary glands of 5-week-old mice is shown in Fig. 1. In control glands implanted with EVAc containing only carrier protein, end buds appeared large and bulbous, even when in close proximity to the implant (Fig. lA, solid arrows). After 3 days of exposure to an inhibitory dose of EGF, end buds near the implant had involuted (Fig. lB), and the ductal tips resembled mammary ducts which had ceased to proliferate due to normal processes of tissue regulation that occur when ductal structures approach the edge of the fat pad (Fig. lA, open arrows). The effect of EGF on end bud growth was restricted to the implanted gland, since other glands in the same animal contained ducts of normal shape and size. Doses of EGF between 2 and 10 /*g per gland, the lower doses of EGF tested, resulted in ductal growth inhibition just around the implant; end buds were still present in the EGF-treated gland distant from the implant (not shown). The effect of EGF treatment on glandular histomorphology and DNA synthesis is shown in Fig. 2. Glands treated with BSA showed typical end bud morphology, with a multilayered, bulbous epithelial end bud covered by a distinct cap cell layer (solid arrow). End buds con-
COLEMAN AND DANIEL
i&use
Mammary
Dudal
Morphogenesis
427
FIG. 1. Photomicrographs of mammary glands illustrating the inhibitory effect of EGF on whole gland morphology. Five-week-old virgin mice were implanted with EVAc pellets containing EGF in the right No. 3 mammary gland and BSA controls in the left (contralateral) No. 3 gland for 3 days (bar = 2 mm). (A) BSA control. Normally developing ductal structures are present around implant (*) with large, bulbous end buds apparent at the ductal tips (solid arrows). Blunt tipped, growth-quiescent ducts are present along the edge of either side of the fat pad (open arrows). (B) Fifteen micrograms EGF treatment. Note the absence of large end buds, and the presence of only slightly rounded ductal tips (arrows) similar to the growth-quiescent ducts seen in A.
tained numerous DNA synthetic epithelial cells (open arrows), particularly in the cap cell layer. DNA synthetic cells were also apparent in the stroma. EGF treatment for 3 days resulted in a much smaller structure, with a reduction in DNA synthetic epithelial cells. In other respects the miniature end buds appeared normal, with a multilayered epithelium and the remnants of a cap cell layer. EGF-treated end bud tips were free of stromal condensation normally present around the mature, nonproliferating duct. DNA labeling indices were determined for treated and untreated glands (Table 1). EGF treatment reduced the percentage of DNA synthetic epithelial cells by 83% between 24 and 74 hr after EVAc implantation. Contralateral control glands showed no significant change in labeling indices of the epithelial cells (Student’s t test, P < 0.01). EGF treatment did not affect the indices of the stromal cell population within 3 days, although the percentage labeled stromal cells was slightly increased on the contralateral side. The loss of end buds 3 days after EGF treatment was dependent upon the concentration of EGF in the EVAc implant (Fig. 3). Doses up to 1 pg EGF per implant had little or no effect on the number of end buds present per gland. Amounts of EGF greater than 2 pg resulted in a dose-related decrease in end bud numbers per gland. At
20 pg no end buds were present within a one-half-cmradius of the implant. End bud inhibition by EGF occurred within 3 to 4 days, and the length of inhibition depended upon the amount of implanted EGF (Fig. 4). At 4 pg EGF per implant, end buds were again present 8 days after initial exposure to EGF, with full recovery by 14 days. At 21 pg EGF, end buds were just beginning to reappear 14 days after implantation. We were interested in determining whether the inhibition of ductal growth required long-term release of EGF or if the observed effect could be attributed solely to the initial surge of EGF released directly after implantation (Fig. 5). Instead of incorporating the EGF into EVAc, a saline solution containing either BSA or EGF/BSA was injected directly into the gland, immediately ahead of the array of end buds. Injections of 3 to 15 fig EGF into the mammary gland showed no effect on end bud numbers as compared with the BSA controls. Since the majority of protein is released from the EVAc implant after 5 hr (Fig. 5, graph), experiments were designed to determine whether the inhibitory EGF dose was delivered in the initial 5-hr treatment or by the low level of EGF released after 5 hr. Implants containing 15 fig EGF were placed in the gland and then removed after 5 hr. Glands were removed for examina-
428
DEVELOPMENTALBIOLOGY
VOLUME137,199O
FIG. 2. DNA autoradiographs of end bud sections from EGF-treated and control glands in 5 week-old animals (bar = 50 pm). (A) End bud treated with BSA for 3 days. Ductal stem (cap) cell layer is present (solid arrow), and both cap and lumenal epithelial cells show extensive label characteristic of normal end buds (open arrow). (B) Ductal tip from gland treated with implant containing 20 pg EGF for 3 days. Note that the overall size is greatly reduced, remnants of a cap cell layer (solid arrow) are still apparent, however. Several DNA synthetic cells are present in the epithelium and stroma (open arrow).
tion after 3 days. EGF treatment for 5 hr followed by EGF removal, had little effect on end bud number when compared with BSA controls. However, glands treated with implants in which EGF had been prereleased into saline solution for 5 hr prior to implantation showed a significant decrease in end bud number after 3 days when compared with contralateral BSA controls. The effect of EGF treatment on EGF receptor distribution was studied in sections of glands from hormonally intact, 5-week-old animals, using lz51autoradiography (Fig. 6). Glands with BSA control implants showed the heaviest concentration of label over stromal cells adjacent to ducts, where the cells and extracellular matrix form a tunic around the flank region of the end bud (Figs. 6A and 6B). Label was also present over some stromal cells located at a distance from ducts. Epithelial cell label was restricted to the cap cells of end buds, as previously reported (Coleman et ak, 1988). Controls, where tissue was incubated with labeled plus lOOO-fold excess unlabeled EGF showed few silver grains over either stromal or epithelial cells (Figs. 6E and 6F). EGF treatment of end buds in vivo resulted in a loss of label over stromal cells and end bud cap cells within 12 hr of treatment (Figs. 6C and 6D). Label did not reappear up to 3 days after EGF implantation (results not shown). The labeling pattern observed in the ‘%I-EGF autora-
diographs was quantified by counting cell-associated silver grains over an average of 100 cells/area/section in multiple alternate tissue sections and is reported as the number of silver grains per cell (Table 2). In control tissue, the heaviest label was present over duct-associated stromal cells and, to a significantly lesser degree, over stromal cells distant from ducts and epithelial cap cells. After EGF treatment of tissue in vivo, no significant label (as determined by Student’s t test) was present over the cap cells of end buds and stromal cells distant from ducts. Label over tunic-associated stromal cells was decreased by 85% (P < 0.01). DISCUSSION
Results from our present study indicate that exogenous EGF, locally delivered, influences the mammary gland in a bifunctional manner. The capacity for EGF to both stimulate or inhibit tissue growth depending upon conditions has been demonstrated in several cell lines in culture, as well as in embryonic development. A role for EGF as a natural growth regulator in vivo has yet to be reported, however. In an earlier study, we described the reappearance of proliferating ductal structures in growth-quiescent glands after treatment with EGF in situ, demonstrating the ability of EGF to initiate and maintain a com-
COLEMAN AND DANIEL
Mouse Mammary
Ductal
429
Morph,ogenesis
TABLE 1 TIME COURSE OF EGF EFFECB ON MAMMARY CELL DNA SYNTHESIS Percentage
labeled cells Stroma”
End bud epithelium Time (hr) 12 24 74
Control”
23.6 f
21.3 f 7.8
6.5&"
Control
Treatment
Treatmentb
cwf
(27)
26.4 + 13.3 (24 3.7 rf: 2.7 (24
19.8 f 4.6 (18) 24.3 f 5.9 cw
(21) 2.5 f 0.9 (24)
plex morphogenetic pathway (Coleman et aL, 1988). This EGF-induced ductal growth was characterized by numerous DNA synthetic mammary cells, apparently normal tissue growth, and typical morphogenetic patterning. Here, we have extended these findings to describe how EGF treatment of rapidly proliferating ducts in mammary glands from hormonally intact, subadult mice resulted in substantial, prolonged growth inhibition, with no indication of hyperplastic growth. 14
4
6
6
10
12
14
16
1614-
1.5 + 0.8 (21) 3.3 + 1.5 (15) 6.8 + 2.0 (24
3.0 f 1.4 (24) 3.9 + 1.6
a Stromal cells were counted within a 0.2-mm radius of the relevant end bud. b Right No. 3 mammary glands implanted with EVAc pellets containing 20 pg EGF. ’ Contralateral glands implanted with EVAc pellets containing BSA. d Results are expressed as the percentage of labeled cells. See Materials and Methods for labeling index determination. e Mean + SD. fThe total number of sections examined.
2
2 4 0
18
20
Pug EGF
FIG. 3. Dose dependency of the ductal response to EGF. All points represent mean number of end buds f SE with an average of five glands per dose. (0) Growing glands were treated with EVAc pellets containing 0.4 to 20 pg EGF for 3 days. At the 3-day time point, a 20 pg EGF dose resulted in complete inhibition. (0) Control glands (contralateral to treatment) were treated with a BSA implant.
1
2
3
4
5
6
7 6 DAY
9
10
11
12
13
14
15
FIG. 4. Time course of the EGF response in 5-week-old animals. All points represent mean number of end buds + SE with an average of four glands per time point. (A) Twenty-one micrograms EGF/gland; end buds disappeared by Day 3 and began to reappear at Day 14. (0) Four micrograms EGF/gland; end buds were completely inhibited by Day 4 recovering to normal numbers by Day 14. (0) BSA control; increase in end bud number at the later time points was due to the growth and increased branching of the ducts during the course of the experiment.
EGF-induced inhibition of ductal growth appeared to occur through normal morphogenetic pathways present in the hormonally intact, virgin animal, as determined by the following criteria: (1) EGF treatment of proliferating ducts for 3 days resulted in end bud-like structures of reduced size, (2) a remnant cap (stem) cell layer persisted in many of the small end buds, (3) EGF-induced inhibition did not result in terminal differentiation of ducts since growth resumed 7-14 days after implantation, and (4) no alveolar or hyperplastic growth was observed. The effects on end bud growth by EGF can be compared with TGF-& another putative ductal growth regulator. Unlike EGF, TGF-pl does not stimulate growth of nonproliferative ducts in endocrine-ablated animals, but like EGF, it is a potent inhibitor of rapidly growing ducts in hormonally intact, virgin mice (Silberstein and Daniel, 198713;Daniel et CL!.,1989). Interestingly, the characteristics of ductal growth inhibition by the two hormones are quite different, suggesting separate mechanisms of inhibition. Morphologically, EGF-treated end buds were reduced in size, but without heavy stromal accumulations present at the ductal tips as seen within 12 hr of TGF-& treatment. The cap cell layer, still present in a vestigial form 3 days after EGF treatment, is lost within 24 hr of TGF-01 treatment. The effect of TGF-& on end bud DNA syn-
430
DEVELOPMENTAL BIOLOGY
VOLUME137, 1990 TREATMENT EGF
Inject
1 CONTROL
Implant
I ;E
q/ z
, 0
12
+-,+--; 3
4
5
24
46
9
76
TIME (HOURS)
FIG. 5. Treatment required for ductal growth inhibition.
(Diagram) Treatment values indicate mean number of end buds f SE present in glands either (A) injected with 3 pg EGF, (B) implanted for 5 hr with EVAc containing 15 wg EGF (implants were removed surgically after 5 hr), or (C) treated with prereleased implants containing 15 pg EGF (implants had been incubated in PBS for 5 hr prior to implantation). All glands were removed from the animal after 3 days. Control values represent BSA-treated glands. Glands treated with implants containing EGF which had been prereleased in vitro had a mean number of 0.7 end buds per gland and were significantly lower than the BSA control. Injection or 5-hr EGF treatment (A,B) had no significant effect on end bud numbers when compared to their controls, nor were they different from each other (as determined by Student’s t test, P = 0.01). (n) Represents number of glands tested. (Graph) EGF release from EVAc was quantified in vitro with lz51-labeled EGF. Maximum release from EVAc occurred during the first hour; less than 1% of the total dose of EGF was released per hour after Hour 4.
thesis is rapid; within 12 hr, TGF-&-treated ducts contained sharply reduced numbers of DNA synthetic cells, whereas EGF-induced inhibition of DNA synthesis occurs after 24 hr. In contrast to the inhibitory response to EGF, TGF-& inhibition of end buds did not require extended exposure; a single injection of the growth factor was effective (Coleman et al, 1988; Silberstein and Daniel, 198713;Daniel et aH, 1989). One noteworthy similarity between TGF+& and EGF inhibition was the restriction of inhibition of DNA synthesis to the epithelial cells. Stromal cells did not respond to treatment by either growth factor with changes in the number of DNA synthetic cells. Stromal cells adjacent to ducts respond to TGF-P1 with increased synthesis of collagen and sulfated glycosaminoglycans (Silberstein et CCL,submitted). The effects of EGF on extracellular matrix components were not investigated, but EGF has been shown to stimulate the synthesis of fibronectin (Gospodarowicz, 1981) and hyaluronate (Lembach, 1976). Receptors for EGF are located in both epithelial and stromal cell populations, indicating the ability of these cell populations to bind
EGF, whereas the location of mammary receptors for & are unknown at present. The mechanism of action of EGF-induced ductal growth inhibition is suggested by the disappearance of the EGF receptor observed prior to growth inhibition. In mammary glands undergoing rapid ductal elongation, EGF receptors are distributed in stromal cells located along the mammary ducts, in stromal cells located at a distance from ductal structures, and in epithelium forming the end bud cap cell layer (Coleman et aL, 1988). Addition of EGF resulted in nearly complete ablation of EGF receptors in the cap cells and an 85% reduction in duct-associated stromal cells within 12 hr of treatment. A correlation between down-regulation of the EGF receptor and EGF-induced inhibition of DNA synthesis has been demonstrated in two other systems, the bud stage of tooth development and in mouse embryos (Abbott and Pratt, 1988; Adamson and Warshaw, 1982). The studies of Adamson and co-workers present an interesting analogy to the effect of EGF in the mammary gland. Treatment of mouse embryos with up to 5 ng EGF
FIG. 6. Autoradiographs of ‘%I-EGF binding in glands from B-week-old animals implanted with EGF in tivo 12 hr prior to organ culture autoradiography. Mammary gland was incubated with 0.2 nM’251-EGF ( see Materials and Methods, bar = 50 am). (A,B) Light- and dark-field micrographs of end bud not treated with exogenous EGF. Label is heaviest over the stromal cells forming the fibrous ductal tunic (open arrows). The cap cell layer also labels (solid arrow), while little label is present over the rest of the epithelial cells. (C,D) End bud from gland treated with implant containing 15 fig EGF for 12 hr. Note absence of label over condensed stromal cells at the flank region of the end bud (open arrow) and over the cap cell layer of the epithelium (solid arrow). (E,F) End bud as in A and B except 1 pg (one thousand-fold excess) unlabeled EGF was added to the incubation medium. Label over the stromal cells and the cap cells is absent. 431
432
DEVELOPMENTALBIOLOGY
TABLE 2 EGF RECEPTORDOWN-REGULATIONMEASUREDBY ‘%I-EGF
BINDING
Treatment Zone Cap cells Stromal cells” (tunic-associated) Stromal cells (non-tunic-associated)
EGF” 0.1 f 0.5”rd 1.5 + 0.2 0.3 * 0.1
BSAb 3.4 + 0.5 10.5 f 0.9 1.8 + 0.1
a Glands were implanted with 15 fig EGF and removed for ‘%I-EGF autoradiography 12 hr later. bGlands were implanted with BSA control. c Silver grains per cell. Alternate serial sections were examined in the three areas. Silver grains over 100 cells were counted per section for each area. Sections were examined from two different experiments for both experimental and control tissues. Specific binding was determined by subtracting values obtained from tissues incubated in labeled EGF plus excess (1 fig) unlabeled EGF from experimental values. d Mean +_SE. “Stromal cells associated with the fibrotic tunic surrounding end buds.
caused an increase in DNA synthesis, yet doses of EGF greater than 10 ng resulted in reduced DNA synthetic activity in addition to down-regulation of the EGF receptor (Adamson et aZ., 1981; Adamson and Warshaw, 1982). Adamson and Warshaw (1982) proposed that embryonic tissue was already maximally responding to endogenous EGF, and prolonged exposure to increased concentrations of the growth factor down-regulated the receptor, thereby preventing a maximal mitogenic response. Our results suggest an analogous explanation for the bifunctionality of EGF in the mammary gland. Additional support for this hypothesis is the requirement for prolonged exposure to EGF for ductal growth inhibition; short-term exposure to exogenous EGF did not elicit an inhibitory response. Recovery of EGF receptors after down-regulation in mouse embryos can occur within 14 hr after injection of EGF (Adamson and Warshaw, 1982), and EGF receptor biosynthesis in A431 cells can occur within 60 min (Carlin and Knowles, 1984). To achieve and maintain the growth-quiescent state for a period of days in the mammary gland, it would be expected that extended exposure inducing prolonged down-regulation of the receptor would be required, preventing full recovery of the EGF receptor population. EVAc implants are capable of releasing low levels of incorporated hormone for at least 7 days. Sources of endogenous EGF in the ductal stage of mammary gland development are speculative at this point. EGF is present in the serum of virgin, female mice at levels reported between 0.17 and 1.2 rig/ml (Byyny et al., 1974; Kurachi and Oka, 1985). Autocrine or
VOLUME137,199O
paracrine sources of EGF in the mammary gland have not been identified. Doses of EGF inhibitory for end bud growth are not accurately described by the amounts of hormone incorporated into an EVAc pellet. Estimates of true inhibitory doses of EGF must take into account factors influencing the concentration of EGF around the implant such as the heavily vascularized stroma, in which released EGF may escape into the vascular system. These variables have been discussed in greater detail in relation to release of TGF-/3i (Daniel et aL, 1989). Estimation of inhibitory dose must also reflect the observation that injection of EGF or the doses of hormone released within the first 5 hr after implantation of the EVAc pellet were not inhibitory for end bud growth. Calculations must be based on the less than 2% released from the implant between Hour 5 and Hour 24 (approx, 0.1% per hour). For example, a pellet containing 20 pg EGF would release an estimated 20 ng EGF per hour 5 hr after implantation. Inhibitory doses of EGF in tissue culture studies and administered exogenously in viva range from 10 to 50 rig/ml. (Imai et aL, 1982, Knowles et aZ., 1985; Kamata et ak, 1986; Koyasu et cd, 1988; Tsao and Liu, 1988). Other possible mechanisms of EGF-induced ductal growth inhibition must be considered. Exogenous EGF may be eliciting synthesis and/or secretion of an unknown inhibitor acting on a pathway other than through the EGF receptor, or EGF itself may be toxic to ductal cells. Neither of these possibilities seem likely in light of the results showing that EGF promotes ductal growth and morphogenesis in endocrine-ablated mice. It is apparent that the mammary gland is extremely sensitive to EGF, in that the hormone can act as either a mitogen or a growth inhibitor depending on the growth state of the gland. In both cases, increasing concentrations of EGF over endogenous levels resulted in a normal ductal morphology rather than abnormal growth. It is very probable that the influence of other mammotrophic hormones plays a role in determining the magnitude and nature of the EGF response. This sensitivity, together with the apparent linkage between down-regulation of the EGF receptor and ductal growth inhibition, suggests that EGF has a normal role in mammary growth and morphogenesis. We thank Gary Silberstein for scientific and editorial advice, and Phyllis Strickland and Kathy Van Horn for skillful technical assistance. This work was supported by PHS Grant CA-45231 from the National Cancer Institute. REFERENCES
ABBOTT,B. D., and PRATT, R. M. (1988). EGF receptor expression in the developing tooth is altered by exogenous retinoic acid and EGF. Dev. Biol. 128,300-304.
COLEMANAND DANIEL
MOUSQMammary Dudal Marphogenesis
ADAMSON,E. D., DELLER, M. J., and WARSHAW,J. B. (1981). Functional EGF receptors are present on mouse embryo tissues. Nature (London) 291,656-659. ADAMSON, E. D., and WARSHAW, J. B. (1982). Down-regulation of epidermal growth factor receptors in mouse embryos. De-v.Biol SO, 430-434.
BYYNY, R. L., ORTH, COHEN, S., and DOYNE, E. S. (1974). Epidermal growth factor: Effects of androgens and adrenergic agents. En& crinology 95,776-782. CARLIN, C. R., and KNOWLES,B. B. (1984). Biosynthesis of the epiderma1 growth factor receptor in human epidermoid carcinoma-derived A431 cells. J. Biol. Chem 259( 12). ‘7902-7908. CARPENTER, G., and COHEN, S. (1979). Epidermal growth factor. Annu. Rev. Biochem 48,193-216. COLEMAN,S., SILBERSTEIN,G. B., and DANIEL, C. W. (1988). Ductal morphogenesis in the mouse mammary gland: Evidence supporting a role for epidermal growth factor. Dev. Biol. 127,304-315. DANIEL, C. W., SILBERSTEIN,G. B., VAN HORN, K., STRICKLAND,P., and ROBINSON, S. (1989). TGF-&-induced inhibition of mouse mammary ductal growth: Developmental specificity and characterization. Dev. Biol. 135,20-30. DEMBINSKI, T. C., and SHIU, R. P. C. (1987). Growth factors in mammary gland development and function. In “The Mammary Gland: Development, Regulation, and Function” (M. C. Neville and C. W. Daniel, Eds.), pp. 355-381. Plenum, New York. GOsPODAROwICZ,D. (1981). Epidermal and nerve growth factors in mammalian development. Annu. Rev. PhysioL 43,251-263. GREEN, M. R., BASKETTER,D. A., COUCHMAN,J. R., and REES, D. A. (1983). Distribution and number of epidermal growth factor receptors in skin is related to epidermal cell growth. Dew. Biol. 100, 506-512.
IMAGAWA, W., TOMOOKA,Y., HAMAMOTO, S., and NANDI, S. (1985). Stimulation of mammary epithelial cell growth in vitro: Interaction of epidermal growth factor and mammogenic hormones. End,ocrinologg 116,1514-1524.
IMAI, Y., LEUNG, C. K. H., FRIESEN, H. G., and SHIU, R. P. C. (1982). Epidermal growth factor receptors and effect of epidermal growth factor on growth of human breast cancer cells in long-term culture. Cancer Res. 42,4394-4398. KAMATA, N., CHIDA, K., RIKIMARU, K., HORIKOSHI,M., ENOMOTO,S., and KUROKI, T. (1986). Growth-inhibitory effects of epidermal growth factor and overexpression of its receptors on human squamous cellcarcinomas in culture. Cancer Res. 46,1648-1653. KAWAE~~URA, K., ENA~I, J., ENAMI, S., KOEZUKA,M., KOHMOTO,K., and KOGA, M. (1986). II. Growth and morphogenesis of mouse mammary epithelial cells cultured in collagen gels: Stimulation by hormones, epidermal growth factor, and mammary fibroblast-conditioned medium factor. Proc. Japan Acad Ser. B 62,5+X KNOWLES,A., SALAS-PRATO,M., and VILLELA, J. (1985). Epidermal growth factor inhibits growth while increasing the expression of an
433
ecto-Ca’+-ATPase of a human hepatoma cell line. Biochem. Biophys. Res. Commun 126.8-14. KOYASU, S., KADOWAKI, T., NISHIDA, E., TOBE, K., ABE, E., KASUGA, M., SAKAI, H., and YAHARA, I. (1988). Alteration in growth, cell morphology, and cytoskeletal structures of KB cells induced by epidermal growth factor and transforming growth factor 8. Exp. Cell Res. 176,107-116. KURACHI, H., and OKA, T. (1985). Changes in epidermal growth factor concentrations of submandibular gland, plasma, and urine of normal and sialoadenectomized female mice during various reproductive stages. J. Endocrinol. 106,197-202. LEMBACH, K. J. (1976). Enhanced synthesis and extracellular accumulation of hyaluronic acid during stimulation of quiescent human fibroblasts by mouse epidermal growth factor. J. Cell Physiol. 89, 277-288.
MURRAY,J. B., BROWN,L., LANGER, R., and KLAGSBURN,M. (1983). A micro sustained release system for epidermal growth factor. In vitro 19,743-748. PARTANEN, A-M., EKBLOM, P., and THESLEFF, I. (1985). Epidermal growth factor inhibits morphogenesis and cell differentiation in cultured mouse embryonic teeth. Deu. Biol. 111,84-94. RICHARDS, J., GUZMAN, R., KONRAD, M., YANG, J., and NANDI, S. (1982). Growth of mouse mammary gland end buds cultured in a collagen gel matrix. Exp. Cell Res. 141,433-443. SILBERSTEIN,G. B., and DANIEL, C. W. (1987a). Investigation of mouse mammary ductal growth regulation using slow-release plastic implants. J. Dairy Sci. 70, 1981-1990. SILBERSTEIN,G. B., and DANIEL, C. W. (1987b). Reversible inhibition of mammary gland growth by transforming growth factor-& Science 237,291-293. SMITH, W. C., and TALAMANTES, P. (1986). Characterization of the mouse placental epidermal growth factor receptor: Changes in receptor number with day of gestation. Placenta 7,511-522. TAKETANI, Y., and OKA, T. (1983a). Possible physiological role of epidermal growth factor in the development of the mouse mammary gland. FEBS I&. 152,256-260. TAKETANI, Y., and OKA, T. (198313).Epidermal growth factor stimulates cell proliferation and inhibits functional differentiation of mouse mammary epithelial cells in culture. Endocrinology 113, 871-877.
THORELL,J. I., and JOHANSSON,B. G. (1971). Enzymatic iodination of polypeptides with I251to high specific activity. B&him Biophys. Acta. 251,363-369. TONELLI,Q. J., and SOROF,S. (1980). Epidermal growth factor requirement for development of cultured mammary gland. Nature (London) 285,250-252. TOPPER,Y. J., and FREEMAN,C. S. (1980). Multiple hormone interactions in the development biology of the mammary gland. Phgsiol. Rev. So;,1049-1106. TSAO, M.-S., and LIU, C. (1988). Inhibition of growth of early passage normal rat liver epithelial cell lines by epidermal growth factor. Lab. Invest. 58,636-642.
