Mullerian inhibiting substance regulates NFkappaB signaling and growth of mammary epithelial cells in vivo.

Müllerian inhibiting substance (MIS) inhibits breast cancer cell growth in vitro through interference with cell cycle progression and induction of apoptosis, a process associated with NFkappaB activation and up-regulation of one of its important target genes, IEX-1S (Segev, D. L., Ha, T., Tran, T. T., Kenneally, M., Harkin, P., Jung, M., MacLaughlin, D. T., Donahoe, P. K., and Maheswaran, S. (2000) J. Biol. Chem. 275, 28371-28379). Here we demonstrate that MIS activates the NFkappaB signaling cascade, induces IEX-1S mRNA, and inhibits the growth of MCF10A, an immortalized human breast epithelial cell line with characteristics of normal cells. In vivo, an inverse correlation was found to exist between various stages of mammary growth and MIS type II receptor expression. Receptor mRNA significantly diminished during puberty, when the ductal system branches and invades the adipose stroma and during the expansive growth at lactation, but it was up-regulated during involution, a time of regression and apoptosis. Peripartum variations in MIS type II receptor expression correlated with NFkappaB activation and IEX-1S mRNA expression. Administration of MIS to female mice induced NFkappaB DNA binding and IEX-1S mRNA expression in the breast. Furthermore, exposure to MIS in vivo increased apoptosis in the mouse mammary ductal epithelium. Thus, MIS may function as an endogenous hormonal regulator of NFkappaB signaling and growth in the breast.

The importance of MIS, 1 a sexually dimorphic member of the transforming growth factor ␤ family of hormones, in regression of the Mü llerian duct in male embryos is well established. MIS is produced by Sertoli cells of the testis even after regression of the Mü llerian duct and continues to be made throughout adulthood. In females, synthesis by granulosa cells of the ovary commences after birth and persists until menopause (2,3). The continued production of MIS throughout adolescence and adulthood in males and females implies other functional roles for this hormone after causing regression of the Mü llerian duct.
The MIS type II receptor, a highly conserved single transmembrane serine threonine kinase, is homologous to members of the transforming growth factor ␤ family of type II receptors (4 -6). The binding of MIS ligand to its receptor initiates a signaling cascade that is dependent on recruitment of type I receptors, ALK2 and ALK6, which also signal for bone morphogenetic proteins (7)(8)(9). The MIS type II receptor gene contains 11 exons and encodes a 1.8-kilobase mRNA. It is expressed at high levels in the Mü llerian duct, the Sertoli and granulosa cells of embryonic and adult gonads (6). However, the status of receptor expression and MIS responsiveness in other tissues has yet to be clarified.
Using several different techniques, we recently demonstrated MIS type II receptor expression in normal breast, human breast fibroadenomas, ductal carcinomas, and cancer cell lines (10). MIS inhibited the growth of both estrogen receptorpositive and estrogen receptor-negative human breast cancer cells in vitro by interfering with cell cycle progression and inducing apoptosis. The effect of MIS on breast cell proliferation correlated with its ability to induce the NFB family of transcription factors and to up-regulate IEX-1S (10), an immediate early gene known to be induced following NFB activation by other extracellular signals. PRG1 and gly96 represent the rat and mouse homologues, respectively, of human IEX-1 (11). Overexpression of IEX-1S in breast cancer cells inhibited their growth, indicating a negative growth regulatory role for this newly identified NFB-inducible gene (10).
The NFB family consists of transcriptional activators, including p65, p50, p52, and c-Rel, that share a Rel homology domain and form either homo-or heterodimers that bind to DNA in a sequence-specific manner. NFB in its inactive state exists in the cytosol bound to the inhibitory IB family of molecules. Activation of the pathway by extracellular signals leads to phosphorylation and degradation of IB with subsequent nuclear localization of NFB (12,13). Expression of a dominant negative inhibitor of NFB (IB␣-DN) in breast cancer cells ablated MIS-mediated induction of IEX-1S, inhibition of growth, and induction of apoptosis, indicating that activation of the NFB pathway was required for these processes (10).
In order to determine whether MIS-mediated growth inhibition through activation of the NFB signaling cascade, previously characterized using human breast cancer cell lines, is also functional in normal breast tissue, we analyzed the effect of MIS on MCF10A, a nontumorigenic breast epithelial cell line (1), and on murine mammary glands in vivo. Furthermore, to evaluate whether these events are developmentally regulated, we analyzed endogenous MIS type II receptor expression, NFB activity, and IEX-1 expression in the mammary gland during postnatal morphogenesis. In this report, we demonstrate that MIS activates the NFB signaling cascade, induces IEX-1S expression, and inhibits the growth of MCF10A cells. Peripartum expression of MIS type II receptor in the rat breast correlated with the level of NFB DNA binding activity and expression of IEX-1S mRNA. In addition, exogenous MIS activated NFB DNA binding and induced IEX-1S expression in the mammary glands of adult mice and increased the number of apoptotic cells in the ductal epithelium of the breast in vivo. Thus, MIS might be a hormonal regulator of the NFB signaling cascade in vivo and a negative regulator of normal breast growth.

EXPERIMENTAL PROCEDURES
Cell Lines and Growth Inhibition Assays-Human breast cancer cell line T47D was grown in Dulbecco's modified medium supplemented with 10% female fetal bovine serum, glutamine, and penicillin/streptomycin. MCF10A cells were grown in mammary epithelial growth medium (Clonetics) supplemented with 100 ng/ml cholera toxin (Calbiochem). 184A1 cells (a gift from Dr. Martha Stampfer) were grown in mammary epithelial growth medium supplemented with isoproterenol and transferrin.
To test the growth inhibitory effect of exogenous MIS, MCF 10A cells were seeded in 100-mm tissue culture flasks in the absence of MIS. MIS was added after 24 h, cultures were grown in the presence or absence of 35 nM MIS for 3 days, and cell numbers were compared by Coulter counter.
Animals, MIS, and MIS Treatment-MIS type II receptor expression analyses in the rat breast during development and peripartum stages were done using Harlan Sprague-Dawley rats. Recombinant human MIS (rhMIS) was collected from growth medium of Chinese hamster ovary cells transfected with the human MIS gene and purified as described in Ref. 14.
To study the effects of rhMIS on the mammary gland, adult female C3H mice (8 weeks old; average weight, 25 g) were obtained from the Edwin L. Steele Laboratory, Massachusetts General Hospital (Boston, MA). All animals were cared for and experiments performed in this facility under guidelines approved by the Assessment and Accreditation of Laboratory Animal Care using protocols approved by the Institutional Review Board-Institutional Animal Care and Use Committee of the Massachusetts General Hospital. All experiments were performed using ketamine/xylazine (100/10 mg/kg) for anesthesia. Each animal was injected intraperitoneally with 100 g of rhMIS or phosphatebuffered saline (vehicle control). Breast tissue was harvested bilaterally from each animal for RNA isolation and gel shift assays. Blood was drawn from the animals at the time of tissue harvest to determine the circulating level of rhMIS using MIS-enzyme-linked immunosorbent assay.
Six-week-old female Rag 2 knockout mice (16) were injected intraperitoneally with 100 g of rhMIS (n ϭ 4) or vehicle (PBS, n ϭ 4) twice daily for 7 days. At the end of this period, breast tissue was harvested bilaterally from each animal, and serum was collected to determine the circulating levels of rhMIS. Part of the tissue was embedded in paraffin for ApopTag assay and fluorescein-labeled in situ cell death detection.
NFB Electrophoretic Mobility Shift Assays-MCF10A and T47D cells were grown to 70% confluence and treated with 35 nM rhMIS for 1 h. Cells were harvested in cold PBS, resuspended in 1 ml of TKM (10:10:1) (10 mM Tris, pH 8.0, 10 mM KCl, and 1 mM MgCl 2 ), and lysed with 0.1% Triton X-100. Nuclei were pelleted by centrifugation at 5000 rpm at 4°C, and proteins were extracted in buffer containing 10 mM HEPES, pH 7.0, 350 mM NaCl, and 1 mM EDTA. 3 g of protein was used in 25-l binding reactions containing 10 mM HEPES, pH 7.0, 70 mM NaCl, 0.1% Triton X-100, and 4% glycerol. The oligonucleotide containing the consensus DNA binding sequence for NFB proteins (Promega) was 32 P-end-labeled, and DNA-protein complexes were resolved on 4% native polyacrylamide gels. Supershift experiments were performed by adding 0.1 g of rabbit anti-p65 or p50 antibodies (Santa Cruz Biotechnology) to the binding reactions. Nuclear protein extraction from tissues was performed as described by Sovak et al. (27).
RNase Protection Assay and PCR Analysis to Detect MIS Type II Receptor Expression-RNase protection assays to detect MIS type II receptor expression were done as previously described (10). To generate a riboprobe to detect MIS type II receptor expression in the rat, a DNA fragment containing part of exon 11 and the 3Ј untranslated region of the rat MIS type II receptor was amplified using the following primers: sense, 5Ј-CCCCGAATTCCCTGGCTTATCCTCAGG-3Ј; antisense, 5Ј-CCCCCTCGAGTCAGCCTGTACAGAGTTCATATGA-3Ј. The rat MIS type II receptor cDNA was used as the template. The receptor fragment was cloned in reverse orientation into XhoI-EcoRI sites of the pCDNA 3.1(-) plasmid. The resulting construct was sequenced to confirm the boundaries of the insert and linearized with HindIII, and the antisense transcript was obtained using T7 polymerase (MAXIscript in vitro transcription kit, Ambion). RNase protection assays were done with 90 -100 g of total RNA isolated from the rat breast at various stages of development using the RPA III ribonuclease protection assay kit (Ambion). RNA samples derived from both individual animals or pooled from groups of three animals were analyzed as indicated in the figure legends. Briefly, RNA was hybridized with 75-80 pg of radiolabeled probe overnight at 50 -55°C and digested with a mixture of RNase A and RNase T1 for 30 min at 37°C. The protected fragments were precipitated and analyzed on a denaturing 6% polycrylamide/6 M urea gel. The same amount of yeast tRNA were used as a positive control for the function of RNase, and another sample containing the same amount of yeast tRNA was incubated without RNase to control for probe integrity.
The riboprobe to detect MIS type II receptor expression in the mouse was generated by PCR amplification using the following primers: sense, 5Ј-CCC CGA ATT CTG CCC AGA GAA CTC CCT T-3Ј; antisense, 5Ј-CCC CCT CGA GTT CCT GAG CAT ATC TAC CCC-3Ј. cDNA generated from RNA isolated from the mouse testis was used as the template. RNase protection was done as described above.
The riboprobe to detect the long and short forms of PRG1/IEX-1 mRNA in the rat was generated by PCR amplification using the following primers: sense, 5Ј-AAC CAC CTC CAC ACC ATG ACT G-3Ј; antisense, 5Ј-CCT TCT TCA GCC ATC AAA ATC TGG-3Ј. Rat genomic DNA was used as the template. The resulting fragment was cloned into SrfI sites of the pPCR-script Amp sk (ϩ) plasmid. The construct was linearized with SalI, and the antisense transcript was obtained with T3 polymerase (MAXIscript in vitro transcription kit, Ambion). RNase protection was done as described above.
Primers for detecting MIS type II receptor expression in MCF 10A and 184A1 cells were as follows: sense, 5Ј-GCT GGC TTA TGC TCT TCT CCT TC-3Ј; antisense, 5Ј-ACC TCG CAC TCT GTA GTT CTT TCG-3Ј. Total RNA was converted to cDNA and amplified with the above primers by PCR.
Northern Blot and PCR Analyses-RNA was isolated from cells or tissue samples using RNA STAT-60 total RNA isolation kit (Tel-Test, Inc.). Indicated amounts of RNA were separated on a formaldehyde gel, transferred to HyBond membrane (Amersham Pharmacia Biotech), and probed with human IEX-1 or mouse gly96/IEX-1 as indicated in the figure legends.
The human IEX-1 probe for Northern analysis was derived by PCR amplification as previously described (10). The probe for detecting both the long and short forms of mouse gly96/IEX-1 was derived by PCR amplification using the following primers: sense, 5Ј-AAC CAC CTC CAC ACC ATG ACT G-3Ј; antisense, 5Ј-CCT TCT TCA GCC ATC AAA ATC TGG-3Ј. Primers for detecting the presence of IEX-1L and IEX-1S transcripts in MCF10A cells have been described (10).
Apoptosis Assays-Breast tissue was harvested, fixed, and embedded in paraffin. After sectioning and deparaffinization, apoptotic cells were detected using a fluorescein in situ cell death detection kit (Roche Molecular Biochemicals) as indicated in the user manual. Images were obtained at a magnification of ϫ 60. For confirmation, the same tissue was sectioned and stained using an ApopTag peroxidase in situ apoptosis detection kit (Intergen) using the protocol provided in the user's manual. Number of apoptotic cells on each slide was compared with number of mammary ducts seen in cross section and expressed as a ratio normalized to the average ratio in the controls.

MIS Inhibits the Growth of Human Mammary Epithelial Cells in Vitro-
We had previously demonstrated that MIS inhibits the growth of both estrogen receptor-positive and estrogen receptor-negative human breast cancer cells in vitro through activation of the NFB signaling cascade. In order to determine whether MIS could also inhibit the growth of nontumorigenic breast epithelial cells, we analyzed MCF10A cells, a human mammary epithelial cell line with normal karyotype derived from a patient with fibrocystic breast disease (1), as well as 184A1 cells (17). Reverse transcription-PCR analysis demonstrated the presence of receptor in both MCF10A and 184A1 cells (Fig. 1A). Sequence analysis of the 582-base pair DNA fragment was identical to exons 1, 2, 3, 4, and 5 of the human MIS type II receptor (data not shown).
Treatment of MCA10A cells with MIS induced three NFB DNA-protein complexes following 1 h of treatment (Fig. 1B, left  panel). The heavier complex comigrated with the NFB DNAprotein complex stimulated in T47D cells following MIS treatment (Fig. 1B, right panel). Simultaneous addition of either rabbit anti-p50 or anti-p65 antibodies to the binding reaction demonstrated that the heaviest complex consisted predominantly of p50 and p65 subunits, as was demonstrated in T47D cells (10). The faster migrating complexes, however, were unique to MCF10A cells and contained p50 subunits. Incubation of nuclear lysates with anti-c-Rel antibody did not supershift the complexes, suggesting that c-Rel protein was not present in them We next investigated whether MIS-mediated increase in NFB DNA binding activity in MCF10A cells correlated with the induction of its target gene IEX-1S. An estimated 5-fold induction of IEX-1 mRNA was observed following 1 h of treatment with MIS, suggesting that the NFB binding activity induced by MIS was functionally active (Fig. 1C, top panel). Reverse transcription-PCR analysis of the MIS-treated samples demonstrated that MIS selectively up-regulated the IEX-1S transcript (Fig. 1C, bottom panel). We had previously demonstrated that MIS inhibited the growth of breast cancer cells in vitro through a NFB-mediated mechanism. As with breast cancer cell lines, treatment of MCF10A cells in vitro with exogenous MIS inhibited growth by 60% (Fig. 1D).
Regulation of MIS-mediated Signaling during Mammary Gland Development in Vivo-To confirm that MIS-mediated signaling events and its effects on breast epithelial cell growth identified using the in vitro cell systems are functional in vivo, we investigated whether MIS type II receptor is expressed in the normal breast. Total RNA isolated from the mammary glands of 8-week-old mice was analyzed by RNase protection assay using an antisense riboprobe specific for exon 11 and 3Ј untranslated region of the mouse MIS type II receptor ( Fig. 2A, left panel). The protected fragment was 89 base pairs shorter than the probe due to unrelated sequences at the 5Ј and 3Ј ends. Detection of a protected fragment of the expected size in the breast, which comigrated with that from the testis, confirmed that the MIS type II receptor mRNA was expressed in normal breast but at a level much lower than that in the testis ( Fig. 2A, right panel).
MIS type II receptor was also detected in rat breast by RNase protection assay using an antisense riboprobe that contained exon 11 and the 3Ј untranslated region specific to the rat MIS type II receptor DNA sequence. Because mammary tissue undergoes the majority of its development in the adult and robust expansion of the breast epithelium occurs during pregnancy and continues into lactation, MIS type II receptor mRNA levels were analyzed using total RNA isolated from mammary glands of virgin, pregnant, lactating, and weaned rats (Fig. 2B). Phosphorimaging of band intensities from three independent experiments demonstrated an 80% decrease in MIS type II receptor expression 2 days after delivery during early lactation. The receptor mRNA rebounded to higher levels 2 days after removal of pups, a period of ductal regression (Fig. 2B, bottom  panel).
Analysis of MIS type II receptor expression during breast development in Harlan Sprague-Dawley rats revealed a grad- Total RNA (90 g) isolated from 8-week-old female mouse mammary glands was analyzed by RNase protection assay. The protected fragments after digestion with RNase were analyzed on a 6% denaturing polyacrylamide gel. 90 g of yeast tRNA was hybridized with the probe and incubated with or without RNase to test the activity of RNases and probe integrity, respectively. A small aliquot of the probe was loaded directly on the gel to control for integrity of the probe. Equal amount of total RNA from mouse testis was analyzed as a positive control. Positions of the 341-nucleotide (nt) probe and the protected fragment (252 nt) are indicated. B, regulation of MIS type II receptor expression during postnatal morphogenesis of the rat ual increase up to postnatal day 30 and a decrease in three individual animals older than 30 days (Fig. 2C, top panel). Quantification of transcript levels by phophorimaging analysis demonstrated a 2.5-fold decrease in MIS type II receptor between animals of postnatal days 14 -30 and postnatal days 40 -60 (Fig. 2C, bottom panel). Onset of pubertal changes in Harlan Sprague-Dawley rats occur at an average age of 35 days after birth (18,19). Interestingly, lowering of MIS type II receptor expression during breast development in the rat coincides with puberty, when the ductal system branches and invades the fat pad (20). The inverse correlation between MIS type II receptor expression and growth in the breast during puberty and peripartum stages was compatible with the hypothesis that MIS-mediated signaling may exert an inhibitory effect on proliferation.
Several transcription factors, including NFB, are regulated during tissue remodeling of the breast. In mice, NFB DNA binding activity in the breast increases slightly during pregnancy, with undetectable levels during lactation (21,22), and is robustly induced in the involuting mammary gland, with the highest levels of binding evident at 2-3 days of involution (21). In order to determine whether activation of NFB DNA binding correlated with changes in MIS type II receptor expression during breast development in rats, electromobility gel shift assays were performed. NFB DNA binding activity was quite readily detectable in the mammary glands of both virgin and pregnant rats. Consistent with the results in mice, reported by Clarkson et al. (21), very little NFB activation was observed in the mammary glands of rats during early lactation, when type II receptor expression was lowest, but was strongly induced 2 days after removal of pups, a period of ductal involution, when type II receptor expression rebounds, indicating a correlation of NFB activation with peripartum MIS type II receptor levels (Fig. 2D).
Because MIS-induced NFB activation in human breast cancer cell lines led to specific up-regulation of its target gene IEX-1S, we tested whether activation of NFB during different phases of mammary morphogenesis in the rat correlated with the levels of IEX-1 mRNA. As demonstrated in Fig. 2E, top panels, PRG1/IEX-1S mRNA expression diminished in the lactating breast but returned to levels observed in the breasts of virgin animals after 2 days of postlactational involution, par-alleling NFB activation during this period of breast morphogenesis. Ethidium bromide staining of the gel demonstrated equal loading of RNA. PCR primers that permit the differentiation of the long and short PRG1 transcripts predominantly detected the latter in the normal breast, with a minor band of 403 base pairs (Fig. 2E, middle panel). PRG1/IEX-1L has an in-frame insertion of 37 amino acids resulting from the presence of the entire intronic sequence within the coding region of the PRG1/IEX-1S transcript (23). To rule out the possibility that the 403-base pair band could have resulted from contaminating genomic DNA being amplified by PCR, a nonquantitative DNA amplification technique, RNase protection assay using an antisense riboprobe specific for rat PRG1/IEX-1L (Fig.  2E, Scheme), was performed. The results confirmed that PRG1/ IEX-1S was the predominant transcript expressed throughout development (Fig. 2E, bottom panel). A protected band of 403 nucleotides that would correspond to PRG1/IEX-1L was not detected. The developmental regulation of PRG1/IEX-1S coincided with NFB DNA binding activity, suggesting that it might indeed be one of the downstream effector genes of activated NFB in vivo in the mammary gland.
MIS Induces NFB DNA Binding Activity and IEX-1S mRNA in Mammary Glands of Mice in Vivo-Because MIS type II receptor levels, NFB DNA binding activity, and IEX-1S expression demonstrated a compelling correlation during postnatal breast morphogenesis, we analyzed whether exogenous rhMIS could induce NFB DNA binding activity and IEX-1S mRNA in the mammary glands of mice in vivo (n ϭ 3). Exposure of mammary tissue to MIS resulted in the induction of NFB DNA binding activity (Fig. 3A, top panel). Analysis of DNA-protein complexes demonstrated the presence of both p50 and p65 subunits; c-Rel was not present in the complex. The specificity of NFB induction in vivo was demonstrated by incubating nuclear protein lysates with an oligonucleotide specific for OCT-1 (Fig. 3A, bottom panel). These experiments identify MIS as one of the first ligands that can induce NFB DNA binding activity in the mammary gland in vivo. The levels of circulating rhMIS in the injected animals were estimated to be 2-4 g/ml by MIS-enzyme-linked immunosorbent assay (2).
To determine whether MIS-mediated induction of NFB DNA binding in vivo correlated with gly96/IEX-1 induction, Northern blot analysis was performed. Exogenous rhMIS inbreast. Top panels, total RNA (90 g) isolated from 8-week-old virgin, pregnant (15 day gestation), lactating (2 days), and weaned (2 days) mammary glands of rats (n ϭ 3 for each sample) was analyzed by RNase protection assay. A representative set is shown on the left; samples pooled from all animals (n ϭ 3) are shown on the right. 90 g of yeast tRNA was hybridized with the probe and incubated with or without RNase to test the activity of RNases and probe integrity, respectively. Positions of the probe (341 nt) and the protected fragment (227 nt) are indicated. RNA from testis (4 g) is shown as a positive control. Bottom left panel, to quantify the changes in MIS type II receptor expression during peripartum, RNase protection assay was performed with three samples from each stage and quantified using a phosphorimager and iQMac data analysis software. The differences in intensities between virgin and lactating, virgin and weaned, and lactating and weaned samples were statistically significant (p Ͻ 0.05) as defined by unpaired Student's t test. Bottom right panel, hybridization to a riboprobe for specific for 18 S rRNA was performed to control for loading (n ϭ 3). C, top panel, MIS type II receptor expression in the breast during postnatal development. Total RNA obtained from mammary glands of Harlan Sprague-Dawley rats during the indicated periods of development were analyzed by RNase protection assay (n Ͼ 2 for postnatal days 2-30; n ϭ 1 for all other ages). Yeast tRNA was hybridized with the probe and incubated with or without RNase. Position of MIS type II receptor (MIS RII) is indicated. Bottom panel, phosphorimaging analysis of MIS type II receptor expression during breast development. The peripubertal group represents animals from postnatal days 14, 21, and 30, and the postpubertal group represents animals 40, 50, and 60 days old (unpaired Student's t test, p Ͻ 0.005). D, peripartum regulation of NFB DNA binding activity. Nuclear proteins (3 g) isolated from the mammary glands of rats at the various peripartum stages described in B were analyzed by electrophoretic mobility shift assay using a 32 P-labeled NFB oligonucleotide probe. Antibody supershift experiments were done with nuclear proteins extracted from the weaned breast. The positions of the NFB DNA-protein complexes (closed arrow) and supershifted complexes (open arrows) are indicated. E, PRG1/IEX-1 expression during peripartum development of the rat mammary gland. Top panel, equal amounts of total RNA (20 g) isolated from the mammary glands of rats during indicated periods of peripartum breast development were analyzed by Northern blot using a PCR-derived probe specific for mouse gly96/IEX-1, which also hybridizes to rat PRG1/IEX-1. The arrow indicates the relative position of 18 S rRNA. A representative set of one animal for each stage is shown on the left; samples pooled from all animals (n ϭ 3) are shown on the right. Ethidium bromide staining is shown to control for loading. Below the top panels is a PCR analysis of cDNA derived from weaned rat breast RNA performed with oligonucleotide primers that detect both the long and short forms of PRG1/IEX-1. The presence of a 296-base pair fragment revealed expression of PRG1-short mRNA (closed arrow). Also shown is a schematic representation of the rat PRG1/IEX-1L antisense riboprobe used for RNase protection assay. Total RNA (10 g) isolated from the mammary glands of rats during indicated periods of peripartum breast development (n ϭ 3 for each sample) was analyzed by RNase protection assay. Yeast tRNA was hybridized with the probe and incubated with or without RNase to test the activity of RNases and probe integrity, respectively. Positions of the probe (499 nt) and the two protected fragments that result from exons 1 and 2 of the rat PRG1/IEX-1S transcript are indicated. duced gly96/IEX-1 expression in the mammary glands of mice within 1 h of treatment compared with untreated controls and remained elevated up to 6 h (Fig. 3B, top panel). The serum rhMIS levels averaged 2-4 g/ml in the animals. Glyceraldehyde-3-phosphate dehydrogenase levels demonstrated equal loading and indicated that the increase in gly96/IEX-1 was not due to general elevation of mRNA expression following treatment with rhMIS. Consistent with results in human cells and rat tissue, reverse transcription-PCR analysis detected predominantly induction of gly96S/IEX-1S (Fig. 3B, bottom panel). Thus, IEX-1S/PRG1S/gly96S represents the first identified target gene to be up-regulated by rhMIS in vivo.
MIS Induces Apoptosis of Mammary Epithelial Cells in Vivo-MIS mediated inhibition of breast cancer cell growth in vitro, as manifested by an increase in the G 1 phase of the cell cycle; was accompanied by apoptosis; and was mediated through activation of NFB DNA binding activity. To determine whether rhMIS induced apoptosis of normal mammary epithelial cells in vivo, tissue isolated from 6-week-old RAG-2-null female mice (16) injected with 100 g of rhMIS (14) twice daily for 7 days was analyzed by a fluorescein in situ cell death detection assay. This system detects DNA strand breaks by labeling the free 3Ј-OH termini with fluorescein-dUTP in a reaction catalyzed by terminal deoxynucleotidyl transferase. An increase in the number of apoptotic cells was observed in the mammary epithelium of mice exposed to rhMIS compared with phosphate-buffered saline-treated control animals (Fig.  4A). This observation was confirmed and quantified using a peroxidase in situ apoptosis detection kit, in which digoxigeninlabeled nucleotides are enzymatically added to ends of DNA and detected using an anti-digoxigenin antibody fragment (Fig.  4B). The ratio of apoptotic cells per duct, normalized to controls, increased 8-fold in mammary epithelial cells exposed to MIS treatment compared with animals injected with vehicle control. DISCUSSION Our previous demonstration that breast cancer cell lines express the MIS type II receptor and are a likely target for the growth inhibitory effect of MIS through a NFB dependent Nuclear proteins were analyzed using a radiolabeled NFB oligonucleotide probe. Positions of the DNA-NFB protein complexes (closed arrow) and the supershifted complexes (open arrows) are indicated. Gel shift analysis performed with OCT-1 oligonucleotide is shown as control. B, mammary glands of 8-week-old female mice were harvested 1, 3, and 6 h after intraperitoneal injections of 100 g of MIS/animal and total RNA was analyzed for gly96/IEX-1 expression. RNA isolated from mammary glands of mice 6 h after intraperitoneal injection of PBS was used as control. Hybridization to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) is shown to control for loading. Bottom panel, cDNA derived from RNA isolated from mammary glands exposed to MIS for 6 h was analyzed using IEX-1-specific oligonucleotide primers. The presence of a single 296-base pair (bp) fragment revealed expression of gly96S/IEX-1S mRNA alone (closed arrow).

FIG. 4. MIS induces apoptosis of mammary epithelial cells in vivo.
A, mammary tissue sections from 6-week-old female Rag2 mutant mice (n ϭ 4) exposed to MIS or vehicle control (n ϭ 4) were analyzed using fluorescein-labeled in situ cell death detection kit. Representative sections are shown at a magnification of ϫ 60. B, the number of peroxidase-positive cells/duct in sections stained with ApopTag peroxidase apoptosis detection kit are represented as the ratio between apoptotic cells and ducts seen in mammary tissue exposed to MIS and normalized to vehicle-treated control. The unpaired Student's t test demonstrated significant differences (p Ͻ 0.005) in apoptosis between mammary glands exposed to vehicle control and MIS. signaling pathway (10) were confirmed using MCF10A cells, a human breast epithelial cell line. MCF10A cells are immortalized but otherwise normal, and when seeded on Matrigelcoated plates, they form duct-like structures and secrete milk proteins into the lumen (1,15).
Although development of the mammary gland begins in the embryo and continues after birth, the major developmental changes occur during puberty and pregnancy. At the onset of puberty, the terminal end buds rapidly proliferate, and the ducts elongate into the mammary fat pad (20,24,25). In 28-day-old Harlan Sprague-Dawley rats, most of the mammary fat pad is epithelium-free, and substantial ductal branching and invasion of the adipose stroma occurs by 50 days after birth (20). However, it is pregnancy that leads to complete maturation and functional activity of the breast with phases of proliferation, differentiation, and apoptosis during pregnancy, lactation, and involution (20,24,25). The inverse correlation between the abundance of MIS type II receptor mRNA expression and growth and functional differentiation of the breast during puberty, and at lactation and involution, suggests that MIS may mediate a negative growth regulatory signaling pathway. This hypothesis is supported by MIS-induced apoptosis in the breast epithelial cell compartment in vivo, which also confirms the induction of apoptosis observed in breast cancer cell lines in vitro. Mammary gland development during puberty occurs under the synergistic influence of growth hormone and estrogen (26), whereas progesterone and prolactin appear to be essential for glandular expansion during pregnancy and lactation. Whether repression of MIS type II receptor expression in the breast during pubertal changes and lactation could result from the action of these hormones provides an exciting area of future investigation.
The dynamic pattern of NFB expression and activity in the breast epithelium during pregnancy, lactation, and involution (21,22) and its aberrant DNA binding activity in breast cancer (27) suggest a role for this family of transcription factors in development and differentiation of the breast. Furthermore, a role for NFB activation in functional differentiation of the breast epithelium is suggested by its ability to inhibit prolactin-induced STAT 5 phosphorylation, resulting in negative regulation of the ␤-casein gene expression (22). Brantley et al. (28) demonstrated that IB-␣ is expressed uniformly in the cytoplasm of virgin, pregnant, and early lactating glands but decreases during late lactation and involution, which was ascribed to degradation resulting from NFB activating signals. Targeted expression to the mammary glands of mice of a luciferase construct driven by a NFB-responsive promoter demonstrated that luciferase expression strongly correlated with the oscillating NFB DNA binding activity observed during postnatal breast morphogenesis. Thus, the NFB DNA complexes activated during mammary gland development appear to be transcriptionally active (28). Little is understood about how transcription factors, such as NFB, are influenced by extracellular hormones and local growth factors known to affect breast development in vivo or which downstream effector genes are induced as a result of NFB activation. Our studies identify MIS as an in vivo hormonal activator of this transcription factor in the breast.
The homo-or heterodimers arising from the NFB family of transcriptional activators p65, p50, p52, and c-Rel have been shown to display differences in DNA binding specificity (29). Furthermore, transcriptional activation of genes involved in Drosophila immunity is differentially regulated by Rel-related protein dimers (29,30). Interestingly, the NFB complexes induced by MIS in breast cancer cell lines, MCF10A cells, and the mammary gland in vivo consisted predominantly of p50/ p65 heterodimers, demonstrating a consistent cell-specific pattern of activation.
MIS-induced activation of NFB in breast cancer cell lines resulted in up-regulation of the immediate early gene IEX-1S; overexpression of this gene inhibited breast cancer cell growth by 50% in colony inhibition assays, identifying IEX-1S as a putative NFB inducible growth inhibitory gene in the breast (10). Arlt et al. (31) recently demonstrated that inducible expression of IEX-1/PRG1 in HeLa cells interferes with cell cycle progression and increases susceptibility to apoptosis. The correlation between NFB DNA binding activity and PRG1S/ IEX-1S expression in the mammary tissue of rats during pregnancy-related changes suggests that IEX-1S/PRG1S/gly96S may be one of the targets activated by NFB nuclear localization in vivo. This is also suggested by the increase in NFB DNA binding activity after exposure to MIS and the induction of IEX-1S/PRG1S/gly96S expression in human breast epithelial cells in vitro and in the mammary glands of mice in vivo. Thus, MIS may be an important normal in vivo hormonal signal that regulates the activity of this transcription factor and the expression of its downstream effectors, including IEX-1S/PRG1S/gly96S.
The lack of developmental abnormalities in mammary gland morphology and function in both MIS ligand and MIS type II receptor-null mice may be due to existence of redundant signals, such as local secretion of both transforming growth factor ␤1 and transforming growth factor ␤3, which induces apoptosis of the mammary epithelial cells (32)(33)(34). Moreover, the possibility that MIS knockout mice could exhibit delayed involution following the removal of suckling pups still remains to be investigated. These studies encourage further characterization of the effects of MIS-mediated signaling events in both normal and neoplastic breast and would require the generation of transgenic mice with targeted expression of MIS or its receptor to the breast.