| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Department of Medicine, University of Colorado Health Sciences Center (M.M.M., R.A.J., J.K.R., D.F.G., W.M.W., K.B.H.), Denver, Colorado 80262; Departments of Obstetrics and Gynecology and Anatomy, McGill University (M.M.M.), Montreal, Quebec, Canada H3A 1A1; and Department of Medicine, University of Newcastle-Upon-Tyne (R.A.J.), Newcastle, NE2 4HH, United Kingdom
Address all correspondence and requests for reprints to: Kathryn B. Horwitz, Department of Medicine, Division of Endocrinology, Box B-151, University of Colorado Health Sciences Center, Denver, Colorado 80262.
| Abstract |
|---|
|
|
|---|
| Introduction |
|---|
|
|
|---|
Analysis of progesterone action is complicated by the fact that two progesterone receptor (PR) isoforms are coexpressed in human target cells (7): 120-kilodalton B-receptors, and N-terminally truncated 94-kilodalton A-receptors. Thus, when equimolar levels of the two isoforms are expressed, A:A, A:B, or B:B homo- and heterodimers, form at 1:2:1 molar ratios. This is important, as the two receptor isoforms regulate transcription unequally when occupied by progesterone agonists (8). with A:A homodimers usually weaker than B:B homodimers. Additionally, in the A:B heterodimer, the inhibitory transcriptional phenotype of A-receptors is dominant. On the other hand, only B-receptors can paradoxically activate transcription when bound by antagonists (9, 10, 11). All known progesterone-dependent target tissues and cells express both PR isoforms, but recent data in breast cancers (12) and uteri (13) suggest that the ratio of B- to A-receptors can flucuate widely. This structural variability would influence the responsiveness of tissues to progestational signals. To study the function of the two PR isoforms independently, we isolated a PR-negative subline (Y cells) of the A- plus B-receptor-positive T47D breast cancer cell line, and then stably reintroduced expression vectors encoding either A- or B-receptors into Y cells to create, respectively, YA and YB cells (14). The protein levels of each PR isoform expressed in these new cell lines is equivalent to the levels of that isoform present in wild-type T47D cells (14). We have used these new cells to study isoform-specific regulation of PR target genes by the method of differential display (15).
The present study demonstrates that the overall pattern of expressed messenger RNAs (mRNAs) in YA and YB cells is remarkably similar, that some messages may be regulated in a ligand-independent manner, and that the progestin agonist R5020 modulates transcription of a small but unique subset of messages. Among the messages regulated by R5020, but only in cells expressing the B-isoform, are those encoding flavin-containing monooxygenase 5 (FMO5). In the liver, studies suggest that FMOs can metabolize drugs and xenobiotics to reactive intermediates that bind covalently with microsomal proteins and DNA (16, 17). Tamoxifen, for example, can be metabolized to a reactive intermediate by these enzymes (16), and such tamoxifen adducts have been described in the uterine DNA of breast cancer patients (18). Our data suggest the interesting hypothesis, that in target cells that overexpress PR B-receptors, progesterone can induce FMO5 and enhance the carcinogenicity of tamoxifen, and therefore, that an excess of B-receptors may serve as a risk marker.
| Materials and Methods |
|---|
|
|
|---|
Construction and characterization of YA and YB breast cancer cell lines that stably express either PR A-receptors or B-receptors, respectively, has been previously described (14). Cells were plated at 1 x 107 cells/75 cm2 tissue culture flask and allowed to grow for 6 days in MEM supplemented with 5% FCS until they were approximately 75% confluent. YA and YB cells were then treated with 20 nM of the progestin R5020 (New England Nuclear, Dupont, Boston, MA) or vehicle alone for 18 h, before harvesting as previously described (14).
Isolation and purification of total RNA and poly A+ mRNA
Total cellular RNA was isolated with guanidinium isothiocyanate followed by centrifugation through a CsCl cushion. DNA was eliminated from total RNA using RNase-free DNase I (Message Clean, GenHunter, Brookline, MA). Poly(A)+ RNA was separated from total RNA by affinity chromatography on oligo(dT)-cellulose (type 7; Pharmacia; Piscataway, NJ) according to our previously published procedures (19).
Differential display of mRNA
We evaluated approximately half (
5000 mRNAs) of the mRNA
species estimated to be present in eukaryotic cells (15). Ten arbitrary
primers were examined with each of four anchored primers (RNAmap kit
II, GenHunter, Brookline, MA). RT-PCR was performed on purified total
RNA as described, using Moloney Murine leukemia virus reverse
transcriptase. Complementary DNAs (cDNAs) were then selectively
amplified by PCR with the appropriate primer pairs in the presence of
[35S]deoxycytidine ATP. Aliquots of separate duplicate
PCR reactions containing amplified cDNA fragments were then
electrophoretically resolved on adjacent lanes of a 6%
polyacrylamide/urea denaturing gel. Following transfer, drying, and
autoradiography, bands were evaluated for differentially displayed
candidate messages.
Recovery, reamplification, and cloning of cDNAs
The autoradiographed film was accurately aligned with the dried gel and differentially regulated PCR fragments were excised, eluted by boiling, ethanol precipitated, and reamplified by PCR with the original primer pairs. The PCR product of the expected size was excised from a 1% agarose gel, and the DNA was purified by silica gel adsorption (QIAEX II kit, Qiagen, Chatsworth, CA). PCR products were ligated into pCRII and transformed into competent INF F E. coli (One Shot Invitrogen TA Cloning, Invitrogen, San Diego, CA). Following blue/white selection, appropriate clones were grown and plasmid DNA was isolated. Sequencing was performed by the Cancer Center Core Laboratory at the University of Colorado Health Sciences Center, Denver, CO. Sequences were searched against the NIH GenBANK database using a BLAST algorithm.
Northern blotting
Total RNA (20 µg) was separated on a 1% agarose/6% formaldehyde gel according to our published protocol (19). The RNA was transferred overnight to a nylon membrane (TurboBlotter; Schleicher and Schuell, Keene, NH) and fixed by UV cross-linking (Stratalinker; Stratagene, La Jolla, CA). Blots were probed with 2.0 x 107 cpm [32P]cDNA labeled by nick translation from the cloned candidate PCR fragments. Northern blots were subsequently probed with 32P-labeled ß-actin to assess loading uniformity.
Rapid amplication of cDNA ends (RACE) was used to obtain further 5'
sequence information from the 360-bp cDNA fragment regulated by R5020
in YB cells (Fig. 3
). A 26-bp primer complementary to a region of the
360-bp sequence was used for first-strand synthesis using Moloney
Murine leukemia virus reverse transcriptase (Clontech Marathon cDNA
Amplification Kit; Clontech Laboratories, Palo Alto, CA) using
poly(A)+ RNA from YB cells treated with R5020. Following
second-stand synthesis, double-stranded cDNA was blunt ended with T4
DNA polymerase I and ligated to an adapter primer supplied with the
Marathon Kit. Adapter ligated cDNAs were amplified by long-distance PCR
(Boehringer Mannheim, Indianapolis, IN). The PCR conditions consisted
of 30 sec at 94 C for denaturation, 30 cycles at 68 C, and 30 sec for
annealing and extension. Amplifed PCR products were size separated on a
1% agarose gel, transferred to a Nytran membrane, and hybridized to a
36-bp oligonucleotide just 5' of the 26 nucleotide RT primer to verify
the specificity of the amplification. Positively hybridizing PCR
products were subcloned into PCR 2.1 (Invitrogen), and miniprep DNAs
were screened for internal sequences by Southern blot analysis (20)
using the radiolabeled 36 mer. An 800-bp positively hybridizing insert
was sequenced from both ends using M13 forward and reverse primers, and
searched against the GenBANK database. To reconfirm its differential
expression pattern, an approximately 300-bp
EcoRI/BamHI fragment from the 5' end of the
800-bp clone was labeled and used for Northern blotting.
|
| Results |
|---|
|
|
|---|
|
The Northern blot in Fig. 2A
shows a
3200-bp RNA transcript labeled by a 380-bp band excised from a
differential display gel. The same Northern blot hybridized with a
labeled ß-actin cDNA demonstrates the relative amounts of RNA loaded
(Fig. 2B
). A BLAST search with the sequencing data indicates that this
message encodes a novel protein. This Northern blot confirms an unusual
ligand-independent up-regulation restricted to B-receptor-containing
cells. Gene regulation by ligand-free steroid receptors has recently
been described (21). The authors speculate that the receptors are
activated by means of cross-talk with other signaling pathways. A
second example of this type of regulation by ligand-free B-receptors is
demonstrated in Fig. 1
(starred product) in which the
product is down-regulated (lanes 5 and 6). It has not been
characterized further.
|
B-receptor-specific and progestin-dependent regulation of FMO5
An approximately 360-bp mRNA fragment was amplified by the
combination of arbitrary primer 13 and anchor primer T12 MA, from
R5020-treated YB cells (Fig. 1
, arrow, lanes 7 and 8) and
represents a classically hormone-regulated message. Interestingly, it
is only regulated by PR B-receptors. This cDNA fragment was excised
from the duplicate lanes of the differential display gel, reamplified,
and cloned. The insert was sequenced, radiolabeled, and used as a probe
on a Northern blot (Fig. 4A
) using total
RNA from the basal and R5020-treated YA and YB cell lines. The labeled
probe hybridized to a major transcript of approximately 3.8 kb and a
minor one of approximately 2.6 kb that were expressed only in
R5020-treated YB cells (Fig. 4A
, lane 4). The 360-bp fragment was
extended in the 5' direction by RACE using RNA from YB+ cells, to yield
an 800-bp product. After the probe was removed, the Northern blot shown
in Fig. 3A
was reprobed with an approximate 300-bp
EcoRI/BamHI fragment corresponding to the 5' end
of the 800-bp RACE product. This Northern blot yielded the same two
3.8-kb and 2.6-kb transcripts, and an additional 1.8-kb species (Fig. 4B
). The blot was then probed with a [32P]cDNA encoding
ß-actin (Fig. 4C
). This demonstrates relatively uniform RNA loading
in lanes 13, with lane 4 somewhat underloaded. Thus, the extent of
mRNA induction in lane 4 is probably underestimated.
|
Figure 5
shows a study with YA and YB
cells, similar to that in Fig. 4
, which also includes a Northern blot
of wild-type T47D cells. These cells express approximately equimolar
amounts of A- and B-receptors (7). As shown, FMO5 mRNAs are
up-regulated 9- to 10-fold in R5020-treated YB cells, but not in YA
cells. Interestingly, in wild-type T47D cells, this up-regulation is
greatly attenuated. This is to be expected if only 25% of the
receptors bind the FMO5 promoter as B:B homodimers. The remaining 75%
bind either as A:A homodimers or A:B heterodimers, both of which would
fail to activate transcription of this promoter (10, 11).
|
| Discussion |
|---|
|
|
|---|
Because PR are important in breast cancers (22), and the A- and B-isoforms of PR are functionally different (8, 9), we have constructed model T47D breast cancer cell lines in which the physiological role of each isoform can be independently evaluated (14). These cells stably express each isoform separately, but at the same levels seen in wild-type T47D cells (14). Differential display of a subset of the messages expressed in these T47D cells suggests that, for the most part, the two stable cell lines, A-receptor containing YA cells and B-receptor containing YB cells, contain the same mRNA populations, confirming their value for analysis of isoform-specific effects. We report here data for only 50% of possible primer pairs and for only one progesterone treatment time. Thus it is clear that we have analyzed only a subset of the messages present in these cells that are capable of being regulated by progesterone and that can be displayed by the method used. This is supported by the fact that several messages known to be regulated by progesterone in breast cancer cells, including ones encoding lactate dehydrogenase (4), alkaline phosphatase (5), and fatty acid synthetase (6) were not identified in our analysis. Of course some of the differentially displayed bands among those that have not yet been sequenced could, in theory, encode these proteins. Also of interest is the fact that all three of the progestin-regulated messages we detected were under B-receptor control. This PR isoform is usually, but not always (8), the stronger transcriptional activator in experimental model systems. Nevertheless, there are conditions under which A-receptor effects predominate (9), and at least one gene, the multidrug resistance gene encoding P-glycoprotein, is under progesterone control by the A-isoform (23). However, our observation that only a low frequency of transcripts are progesterone-regulated is supported by earlier data (24).
Progesterone regulation of FMO5 in breast cancer cells
Our data unequivocally show that three mRNAs containing FMO5 sequences are progesterone regulated in breast cancer cells specifically under the control of the PR B-isoform. To our knowledge no other B-receptor-specific regulated gene has been previously described. Additionally, we found no previous reports showing that FMO5 messages are expressed in normal or malignant breast cells. However, evidence is accumulating for both developmental and progesterone regulation of the five known FMOs in several tissues. Early studies showed that the enzyme is induced in rabbit lung during pregnancy (25) and in hepatic microsomes of CD1 mice during late gestation (26). More recently, modulation of FMO isoform B levels has been demonstrated in rabbit lung and kidney during pregnancy and after progesterone administration (27). Similarly, elevated levels of plasma progesterone, but not of cortisol, correlate with elevated FMO2 enzyme levels in the maternal and fetal rabbit lung, both during gestation and postpartum (28). On the other hand, FMO2 is induced by both progesterone and cortisol in the rabbit kidney (28). Interestingly, the FMO2 isoform is also induced by 17ß-estradiol in the lung and by glucocorticoids in the liver (27), demonstrating tissue specific modulation of these enzymes by other steroid hormones. Regulation by progesterone and/or cortisol may be a direct consequence of the binding of PR B-receptors or glucocorticoid receptors to the FMO5 promoter, which contains a progesterone/glucocorticoid response element (R.M. Philpot, personal communication). Why PR A-receptors do not activate this promoter will be the subject of future studies.
We observed three transcripts (3.8 kb, 2.6 kb, and 1.8 kb) in poly(A)+ RNA isolated from human breast cancer cells that were labeled with the FMO5-specific probe. A recent study using human, rabbit, and guinea pig liver RNA demonstrated only two FMO5 products (3.8 kb and 2.6 kb) (17). The third, a 1.8-kb band that we observe may be caused by tissue-specific differences among FMO5 mRNAs produced in the breast compared with the liver, to the detection of a third FMO5 message that has not previously been reported, or to the presence of an unrelated mRNA that has a sequence similar to that of the FMO5 message. At present we cannot distinguish among these possibilities.
Our finding that FMO5 is progesterone regulated in breast cancer cells has interesting implications regarding drug metabolism in the breast and other progesterone target tissues. FMOs are flavin-containing microsomal monooxygenases that use NADPH as a cofactor. They catalyze the oxidation of a diverse array of substrates including hydrophobic foreign molecules and drugs (29, 30), and have a metabolic role akin to that of the cytochrome P450s (27). Indeed, FMOs as a group appear to be important in the detoxification of drugs, pesticides, and a variety of industrial, chemical, and other xenobiotics.
However, in addition to detoxification, under the influence of FMOs many xenobiotic compounds undergo metabolic activation and produce highly reactive and toxic intermediates that bind covalently to proteins or DNA (16, 31). For example, it has been suggested that the antiestrogen tamoxifen is a potential substrate for FMOs (16). The isoform of liver FMO that is responsible for the formation of the reactive tamoxifen intermediate has not yet been reported, and the FMO5 isoform of the enzyme was only recently cloned and sequenced from a human hepatic library (17). Although the antiestrogenic activity of tamoxifen is the major indicator for its current therapeutic and prophylactic use in breast cancer (32), concerns have been raised about site-specific second cancers arising after long-term therapy with this drug (33, 34). Specifically, tamoxifen-induced DNA adducts were recently described in five of seven endometrial samples of breast cancer patients (18). Our data offer the intriguing possibility that metabolic activation of a variety of drugs mediated by FMOs, including tamoxifen (16, 17), can be controlled by progesterone in target tissues like the breast and the uterus, that overexpress PR B-receptors. If so, in conditions of B-receptor excess, progesterone might accelerate tamoxifen-induced carcinogenesis. This also implies that heightened uterine B- to A-ratios might identify those women at greatest risk of developing tamoxifen-induced malignancies.
| Acknowledgments |
|---|
| Footnotes |
|---|
2 Funded by grants NCI CA47411 (to W.M.W.) and NIH 1 R01 DK47407 (to
D.F.G.). ![]()
3 Funded by NIH through Grants DK48238 and CA26869 and by Grant DAMD
1794-J-4026 from the United States Army. ![]()
Received December 9, 1996.
Revised April 15, 1997.
Accepted June 16, 1997.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
J. Wu, S. Brandt, and S. M. Hyder Ligand- and Cell-Specific Effects of Signal Transduction Pathway Inhibitors on Progestin-Induced Vascular Endothelial Growth Factor Levels in Human Breast Cancer Cells Mol. Endocrinol., February 1, 2005; 19(2): 312 - 326. [Abstract] [Full Text] [PDF] |
||||
![]() |
B. Furnes and D. Schlenk Evaluation of Xenobiotic N- and S-Oxidation by Variant Flavin-Containing Monooxygenase 1 (FMO1) Enzymes Toxicol. Sci., April 1, 2004; 78(2): 196 - 203. [Abstract] [Full Text] [PDF] |
||||
![]() |
B. Furnes, J. Feng, S. S. Sommer, and D. Schlenk Identification of Novel Variants of the Flavin-Containing Monooxygenase Gene Family in African Americans Drug Metab. Dispos., February 1, 2003; 31(2): 187 - 193. [Abstract] [Full Text] [PDF] |
||||
![]() |
O. M. Conneely, J. P. Lydon, F. De Mayo, and B. W. O'Malley Reproductive Functions of the Progesterone Receptor Reproductive Sciences, January 1, 2000; 7(1_suppl): S25 - S32. [Abstract] [PDF] |
||||
![]() |
J. K. Richer, C. A. Lange, N. G. Manning, G. Owen, R. Powell, and K. B. Horwitz Convergence of Progesterone with Growth Factor and Cytokine Signaling in Breast Cancer. PROGESTERONE RECEPTORS REGULATE SIGNAL TRANSDUCERS AND ACTIVATORS OF TRANSCRIPTION EXPRESSION AND ACTIVITY J. Biol. Chem., November 20, 1998; 273(47): 31317 - 31326. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. R. Hovland, R. L. Powell, G. S. Takimoto, L. Tung, and K. B. Horwitz An N-terminal Inhibitory Function, IF, Suppresses Transcription by the A-isoform but Not the B-isoform of Human Progesterone Receptors J. Biol. Chem., March 6, 1998; 273(10): 5455 - 5460. [Abstract] [Full Text] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |