respectively, were obtained from Marc Lippman (National Cancer. Institute, Bethesda, Maryland, USA). R27 cells were maintained in DMEM with 10 % fetal calf ...
Volume 16 Number 5 1988
Nucleic Acids Research
Estrogens and growth factors induce the mRNA of the 52K-pro-cathepsin-D secreted by breast cancer cells
Vincent Cavailles, Patrick Augereau, Marcel Garcia and Henri Rochefort*
Unite d'Endocrinologie Cellulaire et Moleculaire, U 148 INSERM, 60, rue de Navacelles, 34100 Montpellier, France Received January 12, 1988; Accepted February 11, 1988
ABSTRACT The estrogen-induced 52K protein secreted by human breast cancer cells is a lysosomal protease recently identified as a pro-cathepsin D by sequencing several cDNA clones isolated from MCF cells (Augereau et al., Mol. Endocr.). Using one of these cloges, we detected, in MCF cells, a 2.2 kb mRNA whose level was rapidly increased 4- to 70-fold by estradiol, but not by other classes of steroids. Other mitogens, such as epidermal growth factor and insulin, also induced the 2.2 kb mRNA in a dose-dependent manner. Induction with epidermal growth factor was as rapid but was 2- to 3-fold lower than with estradiol. Antiestrogens had no effect on the 52K-cathepsin-D mRNA in MCF7 cells, but became estrogen agonists in two antiestrogenresistant sublines R 7 and LY2. The use of transcription and translation inhibitois and nuclear run-on experiments indicate that estradiol enhances transcription of the 52K-cathepsin-D gene in MCF7 cells.
INTRODUCTION A of large proportion human breast cancers is characterized by its ability to exhibit metastasis and to be regulated by estrogens (1). Estrogens stimulate growth of metastatic breast cancer cell lines containing estrogen receptors (MCF7, T47D...) (2), following the induction of several proteins (3 and ref. therein). The proteins that are secreted, such as growth factors (4) and proteases (5), are particularly interesting since they may stimulate tumor growth and invasion by autocrine and paracrine mechanisms (6). They but not estrogen-regulated, in are also generally produced, estrogen-receptor negative cancers. We have extensively studied a secreted 52K glycoprotein (7) which was found to be mitogenic in vitro (8). The protein has recently been identified as a
© I R L Press Limited, Oxford, England.
1 903
Nucleic Acids Research pro-cathepsin-D-like protease (52K-cath-D) that can degrade extracellular matrix (9,10). The level of regulation of this protease remained unknown, however, and the only gene (pS2) (11), shown to be transcriptionally regulated by estrogens in these cells (12), corresponds to a 7-10K protein of unknown function. Using monoclonal antibodies and a 36-mer oligonucleotide synthesized from the N-terminal sequence of the protein, we have isolated from MCF7 libraries four cDNA clones corresponding to a 2,039 bp coding sequence (13) that is more than 99% identical to that of normal kidney cathepsin D (14). In the present study, we used the 52K-9 cDNA clone to analyze the hormonal regulation of the 52K-cath-D mRNA. We show that estrogens, but not other steroids, rapidly induce 52K-cath-D mRNA by stimulating transcription. Moreover, it was found that other mitogens such as epidermal growth factor (EGF) or insulin can also rapidly increase the level of 52K-cath-D mRNA. MATERIALS AND METHODS Cell culture MCF7 cells (15) were obtained from the Michigan Cancer Foundation and were routinely maintained in T75 flasks in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10 % fetal calf serum (Gibco) and 0.6 4g/ml bovine insulin
(Collaborative Research). To test the effect of hormones on RNA accumulation, slightly confluent MCF7 cells were plated out in T75 flasks (10-fold dilution) in the same medium for 2 days. They were then stripped of hormones with 10 % serum treated with dextran-coated charcoal in phenol-red-free DMEM. The medium was changed every 2 days after two washes with phosphate-buffered saline. After at least 5 days of withdrawal, estradiol or other steroids were added to cells in an ethanol solution (final concentration of ethanol 0.1 %) and solvent alone was added to control cells. For stimulation by insulin and mouse EGF (Collaborative Research), cells were cultured and stripped under the same conditions but without insulin. The two antiestrogen-resistant variants of the MCF7 cell line, R27 (16) and LY2 (17), selected for their resistance to 1904
Nucleic Acids Research the growth inhibitory effects of tamoxifen and LY117018, respectively, were obtained from Marc Lippman (National Cancer Institute, Bethesda, Maryland, USA). R27 cells were maintained in DMEM with 10 % fetal calf serum treated with dextran-coated charcoal, 0.6 gg/ml insulin, and 1 jM tamoxifen. They were stripped of estrogens and tamoxifen by culturing for 14 days in hormone-free medium, as in the case of MCF7 cells. During this time, they were passaged once and the medium was changed every 3 days. LY2 cells were maintained in DMEM with 5% charcoal-stripped calf serum and 0.6 gg/ml insulin. Stimulation by estradiol (1 nM) or antiestrogen (1 jM tamoxifen or its high-affinity metabolite 1 nM 4-hydroxytamoxifen) was performed as described (18). RNA preparation, Northern blot analysis Total RNA was extracted from MCF7 human breast cancer cells by the method of Auffray and Rougeon (19). RNA was electrophoresed on a 1% agarose formaldehyde denaturing gel and then transferred to nitrocellulose. The double-stranded cDNAs in the vectors were 32P-labeled, using random primers (20), to a specific activity of 1 to 3 x 109 dpm per jg. Filters were prehybridized for 24 h at room temperature and then hybridized in 50% formamide for 3 days at 370C (2 x io6 cpm/ml). Hybridization solutions were prepared as described (18). Washing was done in 2 x SSC, 0.1% SDS (1 x SSC is 150 mM NaCl, 15 mM sodium citrate) once for 20 min at room temperature and twice for 1 h at 650C, and the filters were autoradiographed for 5 to 20 h at -700C using intensifying screens. The amount of each RNA was determined by densitometric scanning of different exposures of the autoradiographs. The 36B4 cDNA, which corresponds to an mRNA unaffected by estrogens in MCF7 cells (11), was used to correct for slight variations in the amount of RNA loaded on each track. pS2 RNA, which is transcriptionally regulated by E2 in MCF7 cells (12), was used as a positive control. Measurement of protein synthesis and 52K-cath-D secretion Inhibition of protein synthesis by cycloheximide was estimated by measuring the incorporation of 35SImethionine in the same batches of cells used for-RNA preparation. MCF7 cells 1905
Nucleic Acids Research cultured in the maintenance medium in 8-mm microwells were treated or not with 50 4M cycloheximide (Sigma) for 1 h, and 135Simethionine (Amersham ; SA 800 Ci/mmol ; 10 4Ci/well) was then added to the culture medium. After 5 more hours, the determined was of incorporation radioactivity by trichloroacetic acid precipitation. The cycloheximide treatment reduced radioactivity incorporation by 96-98%. Immunoenzymatic assay of the secreted 52K-cath-D was performed by a double-determinant solid-phase assay (21,22). Nuclear run-on transcription assay Nuclear transcription was performed according to Brown et al. (12) with modifications. Nuclei were isolated from MCF7 cells treated or not with 10 nM estradiol and nascent RNA transcripts initiated in vivo were elongated in vitro in the presence of |32PIUTP (400 Ci/mmole ; Amersham) for 45 min at 300C. Labeled RNA was then extracted with phenol-chloroform The unincorporated after DNAse and proteinase K treatment. nucleotides were eliminated by precipitation twice with ethanol Seven- 4g of denatured plasmids were and ammonium acetate. spotted onto nitrocellulose filters using a BRL Dot blot apparatus. We used 52K-9 cDNA to quantify the newly synthesized of control a positive pS2 cDNA as 52K-cath-D RNA, transcriptional regulation by estradiol, 36B4 and C3 (23) cDNAs (which correspond to poly A+ RNAs unaffected by estradiol in MCF7 cells) as constant controls, and the M13 vector alone to evaluate nonspecific hybridization. Prehybridization was done for 2 days at 370C in 50 mM NaPO4, pH7, 750 mM NaCl, 50% formamide, 0,5% SDS, 2 mM EDTA, lOX Denhart's, 1 4g/ml poly (A) and 500 j.g/ml denatured salmon sperm DNA. Hybridization was done in the same solution for 4 days at 370C, with the same amount of labeled RNA (up to 2 x 107cpm) from control or E2-treated cells. Filters were washed, RNAse A treated and finally autoradiographed for 2 days. The relative intensity of a spot evaluated by densitometric scanning was shown to be proportional
shown).
1906
to
the
amount
of
sample
hybridized
(data not
Nucleic Acids Research RESULTS Effect of estradiol and growth factors on the levels of 52K-cath-D RNA The 52K-9 cDNA probe corresponds to most of the coding sequence of normal cathepsin D mRNA (Fig. la) and hybridized with a 2.2 kb RNA of MCF7 cells, which corresponds to the size of the cathepsin D mRNA in normal tissue (14). The level of this 2.2 kb RNA was increased 6- to 8-fold by estradiol (E2) compared to the estrogen-independent 36B4 mRNA (Fig. lb and c). The 0.6 kb pS2 mRNA (11) was also significantly increased by estradiol as expected (Fig. lc). Similar results were obtained with poly A+ RNA (not shown). The 52K-9 probe also detected a less abundant 52K-cath-D RNA species of about 4.5 kb which was also regulated by estradiol. The significance of this 4.5 kb RNA is not known but it could correspond to a precursor of the 2.2 kb mRNA since its increase was transient and more rapid (data not shown). Other classes of steroid hormones, i.e. progestin (R5020) and glucocorticoid (dexamethasone) increased the 2.2 kb mRNA by no more than 10% (Fig. lb). The androgen dihydrotestosterone was inactive at 10 nM but was active at micromolar concentrations previously shown to induce the secreted 52K-cath-D protein via the estrogen receptor (7). The 52K-cath-D mRNA level was, increased by however, other mitogens EGF than and at estradiol. insulin concentrations previously shown to stimulate the growth of MCF7 cells (24), increased the level of 52K-cath-D RNA 4- and 2-fold, respectively (Fig. lc). These effects were also confirmed at the protein level by the increase in secreted and cellular 52K-cath-D measured by immunoenzymatic assay (21) (Fig. 3 and D. Derocq, unpublished experiments). Using for the we checked that EGF radiocompetition estrogen receptor, and insulin preparations contained no estrogen-like compounds (not shown). Fig. lc also shows that both EGF and insulin which is induced the pS2 also significantly mRNA, estrogen-regulated in MCF7 cells (11). The effect of EGF on the
1907
Nucleic Acids Research a !-,92 29-
g ~ ~ ~ ~ ~ ~ ~
~
-8
~
~
~
~
~
6 .8
~
~
~~~~~.
-6 -8 -6 R.
10 50 100
Figure 1.
4
8
Position of the 52K-9 cDNA clone (a) and effect of steroids (b)an growth factors (c) on the level of 52K-cath-D mRNA In CF. cells a. The 52K-9 cDNA probe isolated from MCF 7 cells iS represented under cathepsin-D mRNA from normal kidney cells,, according to Faust et al. (14). The open boxes stand for the coding sequence and correspond to the signal sequence (1), the pro-sequence (2) and the sequence of the mature enzyme (3 and 4). Coordinates of the coding sequence of normal cathepsin D (from 52 to 1287) and of the terminal nucleotides of the 52K-9 clone are indicated. The internal deletion (dotted line from 192 to 295) in this cDNA clone is a cloning artefact. Sequence analyses of 52K-9 and other clones (13) indicate a 99% homology with normal kidney cathepsin-D (14). b. MCF 7 cells were stripped of estrogens and then incubated for 3 days either in the absence of hormone (C) or with the indicated concentrations of estradiol (E29 dexamethasone (Dex), a synthetic progestin or (R5020), dihydrotestosterone (DHT). Total RNA (40gg) was analyzed by Northern blotting as described in Materials and Methods. Hybridization was done with the 52K-9 cDNA to probe 52K-cath-D mRNAs and with 36B4 cDNA, which detects an RNA species unaffected by estradiol (ii). c. MCF cells were stripped of steroids for 7 days without adding insulin. They were then treated for 3 days withE2 at 10 nM, insulin (Ins) at 50 and 100 nM and EGF at 4 and 8 nM. Total RNA was analyzed as in b and also probed with pS2 cDNA, which corresponds to an estrogen-induced mRNA in MCF7 cells (ii). The lengths of the RNA species detected are indic7ated in kilobases (kb). 1908
Nucleic Acids Research
E E x
E
z
a
I-
ur
Cy
06
24
48
72
Hou-rs of t reat ment Figure 2. Time-course of the effect of estradiol and EGF on the accumulation of 52K-cath-D mRNA. Total RNA (40 gg) was analyzed as described in Fig. lb. The levels of 52K-9 cDNA hybridized to 52K-cath-D RNA (2.2 kb) were determined by densitometric scanning. Values were corrected for slight variations of hybridization in the constant 36B4 RNA, and plotted as percentages of the maximum value. a. Steroid-stripped MCF cells were cultured for the indicated times with (E , A) gr without (C, A) 10 nM estradiol. b. MCF cells weri stripped as described in Fig. lc and treated (EG,A ) or not (C,A ) with 4 nM EGF for the indicated times. level of 52K-cath-D mRNA was greater than that of insulin, whereas both mitogens had a similar effect on pS2 mRNA, suggesting a difference in the hormonal sensitivity of the corresponding genes. We then studied the induction of the 52K-cath-D RNA in MCF7 cells after treatment for different times with estradiol The level of the 2.2 kb 52K-cath-D RNA increased and EGF. rapidly within 2-6 h following the addition of estradiol and 1909
Nucleic Acids Research
0 4
ED E
E E x
a
x
E
C
E
-12 -11 -10 -9 -8 -7 -6 Log ligand concentration (M) -c
a z In
E
-
0 .C
U)
a
LA
C
1.6
4
16
EGF concentration ( nM )
Figure 3. Effects of estradiol, tamoxifen and EGF on the levels of 52K-cath-D mRNA and secreted protein. MCF cells were treated for 3 days as in Fig. 2 with increasiZg concentrations of estradiol (E ), tamoxifen (T) (a) or EGF (b). The levels of 52K-cath-D mRIA (full symbols) were determined after hybridization with 52K-9 and 36B4 cDNAs as The amount of secreted 52K protein described in Fig. 2. accumulated in the medium at the end of treatment (open symbols) was determined using an immunometric assay (21,22). Results are expressed as percentages of the maximum. This increase nearly maximal after 24 h (Fig. 2a). anticipated that of the intracellular protein (25) and the which 52K-cath-D protein secreted into the culture media, increased slowly for 16 h of treatment and more rapidly thereafter (not shown). The half-maximal induction with EGF was also obtained before 10 h of treatment but its effect at 2 h was even higher than that of estradiol (Fig. 2b). Fig. 3 shows that the induction of 52K-cath-D mRNA and the increase in the
was
1910
Nucleic Acids Research G 52K-Cath.D mRNA D Sec reted 52K- Cath.D .
c
100
0
u
-C
a
50
E x
T_tNE
a E
C.
E2
T
OHT
MCF7
T OHT
T OHT
R 27
LY2
Figure 4. Effects of antiestrogens on the induction of 52K-cath-D RNA in NCF7 cells and in two antiestrogen-resistant variants. MCF7 cells and two antiestrogen-resistant variants, R and LY2, were stripped of hormones and treated for 3 days wi{i 1 nM estradiol (E ), 1 liM tamoxifen (T), 1 nM OH-Tamoxifen (OHT) or with solvent alone. The levels of 52K-cath-D mRNAs (
) and secreted 52K-cath-D protein
described in Fig. 3. Results are expressed obtained with estradiol.
as
(U) were determined
percentages
of
the
as
induction
secreted 52K-cath-D were obtained at the same physiological concentrations of estradiol and EGF. These effects are consistent with a progressive occupation and activation of the estrogen and EGF receptor sites, respectively.
Effect
of
antiestrogens
on
NCF7
cells
and
antiestrogen-
resistant variants In MCF7 cells, neither tamoxifen (Fig. 3a) nor hydroxytamoxifen (not shown), which is its high-affinity metabolite, stimulated 52K-cath-D mRNA accumulation while hydroxytamoxifen
antagonized the stimulation of 52K-cath-D mRNA by estradiol (not shown). This is in full agreement with their lack of agonistic effect on the expression of the secreted 52K-cath-D protein (7), shown in parallel (Fig. 3a) and with the complete inhibition of MCF7 cell growth by tamoxifen. By contrast, in two antiestrogen-resistant variants, sublines R27 and LY2, the 2.2 kb mRNA was significantly induced by tamoxifen (1 A M) and
1911
s}
Nucleic Acids Research
b
a
2 3 4 I .&
1,2-~~~~~~~4 CHX
_
2Hy
. + t4
+
Estradiol
Cont rol
kb -2,~~~~~~~~~~~~4 2
X
X
-
s
4 4~~~~~~~~~4 0 . 4 NF*+-s, A.~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Figure 5. Effects of translation and transcription inhibitors. a. MCF7 cells were treated as follows: control cells were incubated in stripped medium Lane 1, for 5 days. Lane 2, cells were treated with 10 nM estradiol (E2) for 8 h. Lanes 3 and 4, cells grown in stripped medium were treated with 50 M cycloheximide (CHX) for 9 h ; after the first hour of CHX treatment, estradiol (lane 4) or only ethanol (lane 3) was added to the cells. b. MCF cells were stripped of estrogens and treated for 2 days with 10 nM estradiol or with solvent alone (control). Cells were then treated with 5 j g/ml actinomycin D for the indicated times. In both cases, the 2.2 kb cathepsin D mRNA level was assayed in total RNA as described in Fig. lb.
1912
Nucleic Acids Research hydroxytamoxifen (1 nM) to a level that was similar or at least 40% of that obtained in parallel with estradiol (Fig. 4). The same antiestrogens had little or no effect on the 2.2 kb mRNA in wild-type MCF7 cells. The secreted 52K-cath-D assayed in the same experiment was induced by antiestrogens in R27 cells but not in LY2 cells, in accordance with previous studies (18,26,27). The dissociated effects observed in LY2 cells, in which the antiestrogens induced the 2.2 kb mRNA but not the secreted 52K-cath-D is not yet explained, however they suggest an additional effect of antiestrogens on protein maturation and/or secretion. Evidence for a regulation of 52K-cath-D gene transcription by estradiol To test whether the induction of 52K-cath-D mRNA is a primary effect of estradiol, we first tested the effect of a protein synthesis inhibitor. MCF7 cells were pretreated for 1 h with 50 4M cycloheximide, which was found to reduce protein synthesis by at least 96% (data not shown), before adding 10 nM estradiol for 8 h with cycloheximide still present. The induction of 2.2 kb RNA was not significantly reduced (Fig. 5a) suggesting that it does not depend on the induction of another protein. We then analyzed the stability of 52K-cath-D mRNA in the presence or absence of estradiol, while its synthesis was blocked by actinomycin D. The level of 2.2 kb mRNA was stable after 10 h of treatment with actinomycin D, in the absence of hormone (control). Pretreatment for 48 h with estradiol did not In another significantly affect this stability (Fig. 5b). experiment, the increased level of 52K-cath-D mRNA produced by 11 h treatment with estradiol was totally abolished when actinomycin D (5 4g/ml) was added together with the hormone (not shown). This effect of actinomycin D suggested a transcriptional regulation by estradiol. We then studied the effect of estradiol on the transcription of the 52K-cath-D gene in isolated nuclei. Following treatment of cells with estradiol, the nascent mRNAs initiated in vivo were elongated in vitro in isolated nuclei in
1913
Nucleic Acids Research the presence of |32PIUTP. Labeled RNAs were then hybridized to pS2, or control cDNAs spotted on nitrocellulose. As 52K-9, shown in Fig. 6a, 52K-cath-D gene transcription was increased 4-fold by incubation with fetal calf serum for 24 h and 2-fold Time-course estradiol. 3 h of by treatment following 52K-cath-D of gene experiments indicated that the stimulation transcription was rapid (30 min), peaked at 1 h and remained pS2 gene In the same experiments, stable until 24 h. transcription was enhanced 4-fold 1 h after estradiol addition. The transcription of two genes known to be unaffected by estradiol in MCF7 cells, i.e. 36B4 (Fig. 6b) and C3 (not shown) was constant with time. These results indicate that estradiol directly stimulates 52K-cath-D gene transcription.
DISCUSSION The hormonal regulation of the 2.2 kb 52K-cath-D mRNA was studied in MCF7 breast cancer cells and three sets of information were obtained : 1) The regulation of 52K-cath-D mRNA by estradiol is at We cannot least partly due to stimulation of transcription. totally exclude a post-transcriptional effect of estradiol since the maximal intensity of the transcription effect (2-3-fold) was lower than the increased accumulation of mRNA (4-10-fold). However, in our run-on experiments, estradiol pS2 transcription which was increased by only 4-fold, previously reported to be increased by 8-fold (12). Cathepsin D is therefore, after the pS2 protein, the second example of a gene transcriptionally regulated by estrogens in human breast Three other mRNAs have been shown to accumulate cancer cells. following estradiol treatment, i.e. progesterone receptor mRNA (E. Milgrom, personal communication), and thymidine kinase and dihydrofolate reductase mRNAs (28), but the mechanism of their induction has not yet been determined. The absence of an effect by cycloheximide indicates that no protein synthesis is required for 52K-cath-D mRNA induction by estradiol, but an of via postindirect gene transcription regulation translationally modified factors cannot be totally excluded. Cloning of the 52K-cath-D gene will make it possible to 1914
Nucleic Acids Research
a
V:
.42K 36B I
9:. .
C FCS
C E2
-o 2.052K -Ccith.D
-o
-o1.5.0
368B4
1.0 U
0
3
Hours
of
5 /. treatment
Figure 6. Nuclear run-on experiments. cells incubated in a. Nuclei were isolated from MCF stripped medium for 5 days (C) or treatgd with 10% fetal calf serum (FCS) for 3 days or with 10 nM estradiol for 3 12 (E ). Nascent RNA chains were elongated in the presence of I PI8TP and hybridized to 52K-9 and 36B4 cDNAs spotted in excess on nitrocellulose as described in Material and Methods. b. Time course of the stimulation of 52K-cath-D gene transcription by estradiol. The nuclear run-on experiment was following MCF cell treatment performed as described in a., with 10 nM estradiol for increasing peZiods of time. Densitometric scanning of the level of newly synthesized 52K-cath-D and 36B4 mRNAs is represented compared to the level of stimulation at time 0.
determine this mechanism and to compare the estrogenresponsive elements with those of other estrogen-regulated genes. There are general differences between the regulation of pS2 and 52K-cath-D in breast cancer cells. The expression of 1915
Nucleic Acids Research whereas pS2 only occurs in estrogen-receptor-positive cells, cathepsin D is also produced in estrogen-receptor-negative breast cancer (13). Both genes are also regulated differently in antiestrogen-resistant cells (see below). 2) The different effects of antiestrogens on the synthesis and of the 2.2 kb-cath-D mRNA in antiestrogen-sensitive on the based -resistant cells confirm previous findings quantification of secreted 52K-cath-D (18,26). The discrepancy where antiestrogen stimulates observed in the LY2 subline, mRNA-cathepsin D accumulation but not 52K-cath-D secretion (27) It suggests an additional level of regulation for secretion. would thus appear that in the three antiestrogen-resistant the RTx6 and LY2 derived from MCF7 cells), sublines (R27, estrogen receptor-antiestrogen complex is able to stimulate the accumulation of 52K-cath-D mRNA, but is unable to do so in MCF7 wild-type cells. The effects of antiestrogens vary according to Progesterone receptors different estrogen-induced responses. are induced in both antiestrogen-sensitive and -resistant cells by tamoxifen whereas pS2 mRNA and the 160K protein are not The induction of affected in any of these cells (18). 52K-cath-D mRNA thus appears to be the only known response The associated with antiestrogen-resistant breast cancer. reasons for the induction of cath-D mRNA by tamoxifen in resistant cells is unknown. Alterations during procedures for subline selection may affect estrogen receptors or the structure of estrogen-responsive elements. 3) The 52K-cath-D mRNA is also induced by other types of suggesting that its such as insulin and EGF, mitogens, regulation may be complex and controlled by different promoter elements that are sensitive to different hormones. EGF-like or IGFI growth factors have also been reported to be induced by estrogens in the same cells (4) and these factors or other mitogens might therefore mediate the induction of 52K-cath-D mRNA by estrogens. However, this possibility is unlikely since the rapid stimulation of 52K-cath-D gene transcription and the both of which suggest that the resistance to cycloheximide, regulation of 52K-cath-D mRNA by estrogens is not mediated by growth factors. The mechanism (direct or indirect) by which 1916
Nucleic Acids Research insulin or EGF operate has not yet been investigated. Further studies are needed to determine the biological significance of induction of protease induction by growth factors. Moreover, 52K-cath-D mRNA by mitogens such as estrogens and growth factors is in agreement with clinical studies indicating that tissue accumulation of this protease is associated with cell proliferation (29). Another example of a growth-factor- induced cathepsin involved in carcinogenesis is cathepsin-L (or major excreted protein : MEP). In mouse BALB/c 3T3 cells, cathepsin-L gene transcription increases after treatment with plateletderived growth factor (PDGF, but not EGF, IGF1 and insulin), tumor promoter, or after transformation by oncogenes (30). In contrast with the effect of estrogens on cathepsin-D mRNA, the induction of cathepsin-L mRNA by PDGF requires de novo protein synthesis. Recently, using a different approach, Westley and May have also shown that estradiol increases the level of cathepsin D RNA in the ZR75-1 estrogen-receptor-positive cell line (31). Their results are in accordance with the present results in MCF7 cells, but the mechanism of induction was not determined and the effect of growth factors was not investigated. Estrogens clearly induce the 52K-cath-D mRNA in human breast but their effect on normal mammary cancer cells in culture, been has not tissues target cells or other estrogen demonstrated. Normal mammary cells in primary culture may not contain a sufficient amount of estrogen receptors to stimulate cathepsin D expression (G. Cavalie, F. Capony, unpublished results). However, in rat uteri, which contain high level of estrogen receptors, progestins rather than estradiol increase cathepsin D activity and synthesis (32). Further studies will be needed to determine whether estrogen induction of the cathepsin D gene characterizes mammary cells or transformed cells. Among the proteins shown to be regulated by estrogens at 52K-cath-D is the the mRNA level in MCF7 cells (4,11,28,33), first example of a protease that is transcriptionally regulated by estrogens. This induction may be important in the process of the resulting mammary carcinogenesis and metastasis since pro-enzyme is also secreted in excess by breast cancer cells
1917
Nucleic Acids Research normal cells (25 and F. Capony et al., in preparation) and may therefore act either directly as a classical growth factor or indirectly by its proteolytic activity. The biological activity of this protease will be more directly demonstrated by transfection experiments. These investigations should increase our understanding of the role of estrogens and of cathepsin D in mammary carcinogenesis, and that of the mechanism by which estrogens and antiestrogens regulate gene expression. compared
to
ACKNOWLEDGEMENTS This research was supported by the "Institut National de la Sante et de la Recherche Medicale", the "Centre National de the Faculty of Medicine of la Recherche Scientifique", Montpellier, and by a fellowship granted to V. Cavailles by the "Ministere de la Recherche et de l'Enseignement Superieur". We Chambon and J.M. Jeltsch for the cDNA are grateful to P. libraries and the pS2 and 36B4 probes, to F. Rougeon for oligonucleotide synthesis, and to M. Egea and E. Barrie for typing the manuscript. We thank D. Derocq and F. Depadova for excellent technical assistance.
*To
whom correspondence should be addressed.
REFERENCES 1. Banbury Reports 8, Hormones and Breast Cancer (1981) Pike,M.C., Siiteri,P.K. and Welsch,C.W. (eds), Cold Spring Harbor Laboratory, Cold Spring Harbor. 2. Lippman,M.E., Bolan,G. and Huff,K. (1976) Cancer Res. 36, 4595-4601. 3. Vignon,F. and Rochefort, H. (1985) In Hollander,V.P. (ed), Hormone Responsive Tumors, Academic Press, New York, pp.135-153. 4. Lippman,M.E., Dickson,R.B., Bates,S.E., Knabbe,C., Huff,K., Swain,S., McManaway,M.E., Bronzert,D., Kasid,A. and Gelmann,E.P. (1986) Breast Cancer Res. Treat. 7, 59-70. 5. Rochefort,H., Capony,F., Garcia,M., Cavailles,V., Freiss,G., Chambon,M., Morisset,M. and Vignon,F. (1987) J. Cell. Biochem. 35, 17-29. 6. Sporn,M.B. and Todaro,G.J. (1980) N. Engl. J. Med. 303, 878-880. 7. Westley,B. and Rochefort,H. (1980) Cell 20, 352-362. 8. Vignon,F., Capony,F., Chambon,M., Freiss,G., Garcia,M. and Rochefort,H. (1986) Endocrinology 118, 1537-1545. 1918
Nucleic Acids Research 9. Morisset,M., Capony,F. and Rochefort,H. (1986) Biochem. Biophys. Res. Commun. 138, 102-109 . 10. Capony,F., Morisset,M., Barrett,A.J., Capony,J.P., Broquet,P., Vignon,F., Chambon,M., Louisot,P. and Rochefort,H. (1987) J. Cell. Biol. 104, 253-262. 11. Masiakowski,P., Breathnach,R., Bloch,J., Gannon,F., Krust,A. and Chambon,P. (1982) Nucl. Acids Res. 10, 7895-7903. 12. Brown,A.M.C., Jeltsch,J.M., Roberts,M. and Chambon,P. (1984) Proc. Natl. Acad. Sci. USA 81, 6344-6348. 13. Augereau,P., Garcia,M., Mattei,M.G., Cavailles,V., Depadova,F., Derocq,D., Capony,F., Ferrara,P. and Rochefort,H. (1988) Mol. Endocrinol., in press. 14. Faust,P.L., Kornfeld,S. and Chirgwin,J.M. (1985) Proc. Natl. Acad. Sci. USA 82, 4910-4914. 15. Soule,H.D., Vazquez,J., Long,A., Albert,S. and Brennan,M.A. (1973) J. Natl. Cancer Inst. 51, 1409-1413. 16. Nawata,H., Bronzert,D. and Lippman,M.E. (1981) J. Biol. Chem. 256, 5016-5021. 17. Bronzert,D.A., Greene,G.L., and Lippman,M.E. (1985) Endocrinology 117, 1409-1417. 18. Westley,B., May,F.E.B., Brown,A.M.C., Krust,A., Chambon,P., Lippman,M.E. and Rochefort,H. (1984) J. Biol. Chem. 259, 10030-10035. 19. Auffray,C. and Rougeon,F. (1980) Eur. J. Biochem. 107, 303-314. 20. Feinberg,A.P. and Vogelstein,B. (1984) Anal. Biochem. 137, 266-267. 21. Garcia,M., Capony,F., Derocq,D., Simon,D., Pau,B. and Rochefort,H. (1985) Cancer Res. 45, 709-716. 22. Rogier,H, Freiss,G., Cavali6-Barthez,G., Garcia,M., Paolucci,F., Rochefort,H. and Pau,B. Submitted for publication. 23. Chalbos,D., Westley,B., May,F., Alibert,C., and Rochefort,H. (1986) Nucl. Acids Res. 14, 965-982. 24. Osborne,C.K., Bolan,G., Monaco,M.E. and Lippman,M.E. (1976) Proc. Natl. Acad. Sci. USA 73, 4536-4540. 25. Morisset,M., Capony,F. and Rochefort,H. (1986) Endocrinology 119, 2773-2783. 26. Vignon,F., Lippman,M.E., Nawata,H., Derocq,D. and Rochefort,H. (1984) Cancer Res. 44, 2084-2088. 27. Davidson,N.E., Bronzert,D.A., Chambon,P., Gelmann,E.P. and Lippman,M.E. (1986) Cancer Res. 46, 1904-1908. 28. Aitken,S.C., Lippman,M.E., Kasid,A. and Schoenberg,D.R. (1985) Cancer Res. 45, 2608-2615.
29. Garcia,M., Lacombe,M.J., Duplay,H., Cavailles,V., Derocq,D., Delarue,J.C., Krebs,B., Contesso,G., Sancho-Garcier,H., Richer,G., Domergue,J., Namer,N. and Rochefort,H. (1987) J. Steroid Biochem. 27, 439-445. 30. Rabin,M.S., Doherty,P.J. and Gottesman,M.M. (1986) Proc. Natl. Acad. Sci. USA 83, 357-360. 31. Westley,B. and May,F.E.B. (1987) Nucl. Acids Res. 15, 3773-3786. 32. Moulton,B.C. and Koenig, B.B. (1983) Am. J. Physiol. 244, E442-E446. 33. Adams,D.J., Edwards,D.P. and McGuire,W.L. (1980) Biochem. Biophys. Res. Commun. 97, 1354-1361. 1919
