Mig-6 modulates uterine steroid hormoneresponsiveness and exhibits altered expressionin endometrial diseaseJae-Wook Jeonga,1, Hee Sun Leeb, Kevin Y. Leea, Lisa D. Whitec, Russell R. Broaddusd, Yu-Wen Zhange,George F. Vande Woudee, Linda C. Giudicef, Steven L. Youngg, Bruce A. Lesseyh, Sophia Y. Tsaia, John P. Lydona,and Francesco J. DeMayoa

aDepartment of Molecular and Cellular Biology, Baylor College of Medicine, Houston, TX 77030; bInfertility and Reproductive Endocrinology Center,Department of Obstetrics and Gynecology, MizMedi Hospital, Seoul, Korea; cMicroarray Core Facility, Department of Molecular and Human Genetics, BaylorCollege of Medicine, Houston, TX 77030; dDepartment of Pathology, University of Texas M. D. Anderson Cancer Center, Houston, TX 77030; eLaboratories ofMolecular Oncology, Van Andel Research Institute, Grand Rapids, MI 49503; fDepartment of Obstetrics, Gynecology and Reproductive Sciences, University ofCalifornia, San Francisco, CA 94143; gDepartment of Obstetrics and Gynecology, University of North Carolina, Chapel Hill, NC 27599; and hDepartment ofObstetrics and Gynecology, University Medical Group, Greenville Hospital System, Greenville, SC 29605

Communicated by Bert W. O’Malley, Baylor College of Medicine, Houston, TX, April 3, 2009 (received for review December 3, 2008)

Normal endometrial function requires a balance of progesterone(P4) and estrogen (E2) effects. An imbalance caused by increasedE2 action and/or decreased P4 action can result in abnormal endo-metrial proliferation and, ultimately, endometrial adenocarcinoma,the fourth most common cancer in women. We have identifiedmitogen-inducible gene 6 (Mig-6) as a downstream target of proges-terone receptor (PR) and steroid receptor coactivator (SRC-1) action inthe uterus. Here, we demonstrate that absence of Mig-6 in miceresults in the inability of P4 to inhibit E2-induced uterine weight gainand E2-responsive target genes expression. At 5 months of age, theabsence of Mig-6 results in endometrial hyperplasia. OvariectomizedMig-6d/d mice exhibit this hyperplastic phenotype in the presence ofE2 and P4 but not without ovarian hormone. Ovariectomized Mig-6d/d

micetreatedwithE2developedinvasiveendometrioid-typeendometrialadenocarcinoma. Importantly, the observation that endometrial carci-nomas from women have a significant reduction in MIG-6 expressionprovidescompellingsupport foran importantgrowthregulatoryrole forMig-6 in the uterus of both humans and mice. This demonstrates theMig-6 is a critical regulator of the response of the endometrium to E2 inregulating tissue homeostasis. Since Mig-6 is regulated by both PR andSRC-1, this identifiesaPR,SRC-1,Mig-6 regulatorypathwaythat is criticalin the suppression of endometrial cancer.

estrogen � endometrial cancer � progesterone � progesterone receptor � SRC-1

The ovarian steroid hormones progesterone (P4) and estrogen(E2) are essential regulators of reproductive events associated

with all aspects involved in the establishment and maintenance ofpregnancy (1, 2). P4 and E2 function both synergistically andantagonistically to regulate appropriate uterine function by actingthrough their cognate nuclear receptors to coordinate uterineepithelial–stromal communication in the regulation of endometrialcell proliferation and differentiation. An imbalance caused byincreased E2 action and/or decreased P4 action can result in abnormalendometrial proliferation and endometrial adenocarcinoma. Endome-trial cancer is the most common gynecological cancer in the UnitedStates (3). Endometrioid-type endometrial adenocarcinoma and itsprecursor lesion, endometrial hyperplasia, are associated with unop-posed estrogen exposure (4, 5). Elucidating the molecular mecha-nisms by which the steroid hormones control uterine physiology isimportant to understanding the pathology of these diseases.

The action of the steroid hormone receptors are modulated inpart by members of the p160 family of steroid receptor coactivators(SRCs). The SRCs facilitate steroid hormone receptor regulationof gene transcription by executing a diverse number of processesincluding chromatin remodeling, RNA processing, and receptordegradation (6). The SRC family is composed of 3 distinct butfunctionally and structurally related members: SRC-1/NcoA1 (7),

SRC-2/TIF2/GRIP1 (8), and SRC-3/RAC3/ACTR/pCIP/AIB1/TRAM1 (9). The SRC family members enhance the transcriptionalactivity of a variety of nuclear receptors, including estrogen receptor� (ER�, also known as ESR1), estrogen receptor � (ER�, alsoknown as ESR2), glucocorticoid receptor (GR), and progesteronereceptor (PR) (10–12) and are expressed in a variety of hormone-responsive tissues including the uterus, brain, prostate, liver, andbreast (7, 13–15). Although female SRC-1�/� mice are fertile,reduced steroid sensitivity in the uterus of SRC-1�/� mice isdemonstrated by a reduction in the ability of the endometrialstroma cells to undergo a decidual transformation (7). This phe-notype indicates that these coregulators may play a necessary rolein coordinating steroid hormone regulation of normal reproductiveuterine function. Using high-density DNA microarray technology,we have identified Mig-6 as a gene whose regulation by P4 isdependent upon SRC-1.

Mig-6 is an immediate early response gene that can be inducedby various mitogens and commonly occurring chronic stress stimuli(16–18). Mig-6 is an adaptor molecule containing a CRIB domain,a src homology 3 (SH3) binding domain, a 14–3-3 binding domain,and an epidermal growth factor receptor (EGFR) binding domain(19, 20). Sustained Mig-6 expression is thought to trigger cells toinitiate hypertrophy in chronic pathological conditions, such asdiabetes and hypertension (21, 22). Ablation of Mig-6 in mice hasled to the development of animals with epithelial hyperplasia,adenoma, and adenocarcinomas in organs, such as the lung, gall-bladder, and bile duct (23, 24). Mig-6 is located on human chro-mosome 1p36, a locus frequently associated with human cancer (25,26). Decreased expression of Mig-6 is observed in human breastcarcinomas that correlate with reduced overall survival of breastcancer patients (27, 28). Mig-6 is mutated in human non-small-celllung cancer (NSCLC) cell lines NCI-H226 and NCI-H 322M and inone primary human lung cancer (24). Recently, altered Mig-6expression has been observed in endometrial RNA taken fromwomen with endometriosis (29). These data point to Mig-6 as atumor suppressor gene in both mice and humans. However, thefunction of Mig-6 in reproductive biology has remained elusive.

Author contributions: J.-W.J., H.S.L., R.R.B., S.Y.T., J.P.L., and F.D. designed research; J.-W.J.,H.S.L., and L.D.W. performed research; Y.-W.Z., G.F.V.W., L.C.G., S.L.Y., and B.A.L. contrib-uted new reagents/analytic tools; J.-W.J., H.S.L., K.Y.L., L.D.W., R.R.B., S.Y.T., J.P.L., and F.D.analyzed data; and J.-W.J., K.Y.L., L.C.G., S.L.Y., B.A.L., and F.D. wrote the paper.

The authors declare no conflict of interest.

Freely available online through the PNAS open access option.

1To whom correspondence should be addressed. E-mail: jjeong@bcm.tmc.edu.

This article contains supporting information online at www.pnas.org/cgi/content/full/0903632106/DCSupplemental.

www.pnas.org�cgi�doi�10.1073�pnas.0903632106 PNAS � May 26, 2009 � vol. 106 � no. 21 � 8677–8682

MED

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Nov

embe

r 3,

202

0

In this study, we used conditional ablation of Mig-6 in mice todemonstrate that Mig-6 is an important molecule in uterinephysiology in part by regulating the ability of P4 to attenuate E2signaling. These mice develop endometrial hyperplasia within 5months, and, if exposed to exogenous E2 for 3 months, developinvasive type I endometrial adenocarcinoma. Analysis of Mig-6in the human endometrium shows that the expression of Mig-6is decreased in human endometrial cancers. Thus, these resultsdemonstrate the importance of Mig-6 in steroid hormone reg-ulation and human endometrial cancer.

ResultsIdentification of P4 and SRC-1 Regulated Genes Using MicroarrayAnalysis. The impact of SRC-1 ablation on uterine mRNA expres-sion profiles in response to P4 was examined by isolating RNA fromovariectomized SRC-1�/� and wild-type mice that were treatedwith either vehicle (sesame oil) or P4 (1 mg) for 4 h (n � 9 pergenotype per treatment). The experimental design not only allowedfor the effect of SRC-1 in SRC-1�/� mice to be determined, but alsoafforded the comparison of the effect of vehicle and P4 on wild-typeand SRC-1�/� mice. Using this experimental design, the identifi-cation of differentially expressed genes was derived from 3 physi-ologically relevant comparisons (Fig. S1A). The summary of thenumber of differentially expressed genes for the 3 comparisons isshown in supporting information (SI) Table S1. Comparison 1identified genes differentially expressed between wild-type andSRC-1�/� mice in the absence of P4 treatment. Comparison 2identified genes regulated by SRC-1 in the presence of P4, andcomparison 3 identified P4 responsive genes in wild-type mice. Acomplete list of the increased and decreased genes that are iden-tified as significant for comparisons 1–3 are presented in Table S2.

To identify the impact of SRC-1 ablation on P4 induction of geneexpression, we identified genes that overlapped between compar-isons 2 and 3. (Fig. S1B) graphically shows the overlap in the analysisof SRC-1-dependent P4 responsive genes (comparison 2) and P4responsive genes (comparison 3). We identified 5 targets of whichthe mRNA expression was induced by P4 in wild-type mice but notin SRC-1�/� mice (Fig. S1C). These genes were Mig-6, Myc, Il13ra2,and an EST. The EST was not analyzed further because it did not

have any known homologous gene and its expression was notPR-dependent, based on our previous microarray data (30).

To confirm whether the gene regulatory effects of P4 aremediated through SRC-1 and PR, we ovariectomized SRC-1�/�

mice, progesterone receptor knockout (PRKO), and WT mice.After P4 treatment for 4 h, the uteri were collected, RNA wasprepared, and total RNA was analyzed by real time RT-PCR.Expression of all 3 genes was highly induced by P4 in the wild-typemice, but induction was significantly decreased in the SRC-1�/�

mice (Fig. S1D). The P4 induction was also not detected in thePRKO mice. Therefore, real time RT-PCR confirmed the resultsof the microarray analysis and further showed that their expressionwas PR dependent.

Ablation of Mig-6 in mice leads to animals with epithelialhyperplasia, adenoma, and adenocarcinoma in organs like the lung,gallbladder, and bile duct (24). This tumor suppressor function inmice (23, 24) is also observed in humans (25, 26). The fact that Mig-6was found to be a SRC-1-dependent PR-responsive gene promptedus to investigate its function in the murine endometrium. Thespatial expression of Mig-6 was examined by in situ hybridization(Fig. S2). Mig-6 transcripts were undetectable in the vehicle-treateduterus. However, Mig-6 mRNAs were strongly expressed in thestroma, luminal epithelium, and glandular epithelium by P4 treat-ment. In contrast, these mRNAs were significantly decreased in theuteri of ovariectomized SRC-1�/� and PRKO mice receiving P4treatment. These results demonstrate that Mig-6 is PR and SRC-1dependently regulated in all compartments of the endometriumincluding the epithelium and stroma but not myometrium.

The Epithelial Hyperplastic Effect in Mice with Conditional Mig-6Ablation in the Uterus. Since Mig-6 ablation results in numerouspathologies and decreased longevity (23, 24, 31, 32), our ability toinvestigate the role of Mig-6 in the mouse uterus is severely limited.To effectively investigate the role of Mig-6 in the regulation ofuterine function and the response to hormonal stimulation, wegenerated a Mig-6 conditional null allele, the Mig-6 flox allele(Mig-6f/f) (33). Mig-6f/f mice were bred to PRCre (34) mice to generateconditional Mig-6 ablation (PRcre/�Mig-6f/f; Mig-6d/d) in the repro-ductive tract. Significant morphological differences in the uteri ofMig-6d/d and Mig-6f/f were not observed at 2 months of age.However, analysis of Mig-6d/d uteri compared to Mig-6f/f mice at 5months of age (n � 10) showed a significant increase in wet weight(Fig. 1 A and B). Histological analysis of these uteri showed anincrease in the number of endometrial glands and in the gland/stroma ratio in the uterus of Mig-6d/d mice (Fig. 1 C–H); however,the myometrium was not enlarged. These histological changesdemonstrate that the uterus of the Mig-6d/d mouse displays endo-metrial hyperplasia, a predisposing factor to endometrial adeno-carcinoma in humans.

To determine if the endometrial epithelial hyperplasia in theMig-6d/d mice is caused by alterations of ER signaling, cell prolif-eration, and/or apoptosis, we performed immunohistochemicalstaining for ER�, Ser 118 phospho-ER�, phosphohistone H3, andcaspase 3. Immunohistochemical staining of caspase 3 was un-changed in the luminal or glandular epithelium of nonpregnantMig-6d/d mice compared to wild-type control mice (data not shown).However, immunohistochemical staining of phosphohistone H3showed that the endometrial glandular cells exhibited a significantincrease in proliferation in the Mig-6d/d mice (Fig. 1D). Supportiveof E2 being the cause of the increase in epithelial proliferation, ER�and phosphorylation of ER� at Ser 118 were also significantlyincreased in the endometrial glands (Fig. 1 F and H).

Steroid Hormone Regulation of Mig-6 in the Murine Uterus. Todetermine if the uteri of Mig-6d/d exhibited an altered responseto steroid hormones, Mig-6d/d and Mig-6f/f mice were ovariecto-mized and treated with vehicle, E2, P4, or E2 � P4 daily for 3days and killed 6 h after the last injection (n � 5 per genotype

Fig. 1. The hyperplastic effect of conditional Mig-6 ablation in the murineuterus. (A) Uteri of 5-month-old PRcre/�Mig-6f/f (Mig-6d/d) and Mig-6f/f mice. (B)Quantitative weights of uteri in Mig-6d/d and Mig-6f/f mice. The results repre-sent the mean � SEM of 5 animals. **, P � 0.01, unpaired t test. (C–H)Immunohistochemical analysis of ER� (C and D), phospho-ER� (E and F), andphosphohistone H3 (G and H) in uteri of 5-month-old Mig-6f/f (C, E, and G) andMig-6d/d (D, F, and H) mice.

8678 � www.pnas.org�cgi�doi�10.1073�pnas.0903632106 Jeong et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 3,

202

0

per treatment). Mig-6d/d and Mig-6f/f uteri showed no differencein weight gain or expression of PR, ER�, or their respectivetarget genes when the mice were treated with vehicle, E2, or P4.However, Mig-6d/d mice treated with E2 � P4 showed a signif-icant increase in uterine weight (Fig. 2A), vascularization (Fig.2B), and expression of endometrial epithelial ER� target genes,

Ltf (lactotransferrin), Clca3 (chloride channel calcium activated3), and C3 (complement component 3), (Fig. 2C) compared toE2 � P4-treated Mig-6f/f uteri. The expression of 2 endometrialepithelial PR target genes, Fst (follistatin) and Areg (amphiregu-lin), is not changed in the Mig-6d/d mice in response to E2 � P4treatment (Fig. 2D). However, the expression of PR mRNA is

Fig. 2. Steroid hormone regulation of Mig-6 in the murine uterus. (A) The ratio of uterine weight to body weight in Mig-6d/d and Mig-6f/f mice treated withE2, P4, or E2 plus P4 for 3 days. (B) Uteri of Mig-6d/d and Mig-6f/f mice treated with E2 plus P4 for 3 days. (C) ER�-regulated gene expression in uteri of Mig-6d/d

and Mig-6f/f mice treated with E2 plus P4. Real-time RT-PCR analysis of Ltf, Clca3, and C3 was performed. (D) PR and PR-regulated gene expression in uteri ofMig-6d/d and Mig-6f/f mice treated with E2 plus P4. Real-time RT-PCR analysis of PR, Fst, and Areg was performed. (E) Immunohistochemical analysis of PR in uteriof Mig-6d/d and Mig-6f/f mice treated with E2 plus P4. (F) Quantification of PR positive cells in stroma of Mig-6d/d and Mig-6f/f mice. The results represent the mean �SEM. ***, P � 0.001.

Fig. 3. Steroid hormone-dependent formation ofendometrial cancer in Mig-6d/d mice. (A) Uteri of5-month-old ovariectomized Mig-6d/d and Mig-6f/f

mice treated with vehicle. H&E stained section show-ing normal uterus in the ovariectomized Mig-6d/d andMig-6f/f mice. (B) Formation of endometrial cancer inovariectomized Mig-6d/d mice treated with E2. Endo-metrial cancer induced in the uteri of ovariectomizedMig-6d/d mice treated with E2 for 3 months. H&Estained section showing simple hyperplasia in the Mig-6f/f mice and endometrioid endometrial cancer in theMig-6d/d mice. (C) Uteri of 5-month-old ovariectomizedMig-6d/d and Mig-6f/f mice treated with E2 � P4 for 3months. H&E stained section showing hyperplasiauterus in the ovariectomized Mig-6d/d mice treatedwith E2 � P4 for 3 months.

Jeong et al. PNAS � May 26, 2009 � vol. 106 � no. 21 � 8679

MED

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Nov

embe

r 3,

202

0

significantly decreased in the Mig-6d/d mice compared to that ofMig-6f/f mice. Immunohistochemical analysis of PR expression inthe Mig-6d/d mice shows that epithelial PR expression is normalbut the expression of PR in the endometrial stroma cells issignificantly reduced (Fig. 2 E and F). Since P4 attenuates E2regulation of proliferation and gene expression by regulating theexpression of a yet-to-be-identified paracrine signal from thestromal cells to the epithelial cells, the regulation of the expres-sion of PR in the endometrial stromal cells by Mig-6 is critical forthe ability of P4 to attenuate the E2-regulated uterine weightgain, vascularization, and expression of ER target genes.

To assess the role of Mig-6 in uterine function, female Mig-6d/d

mice were mated to wild-type male mice for 6 months. Mig-6d/d micewere completely infertile (Table S3). Since the PRCre mouse showsCre recombinase activity in the pituitary, ovary, uterus, and mam-mary gland, the cause of infertility in these mice may be the resultof a defect in any of these tissues (34). To test for an ovarianphenotype, female Mig-6f/f and Mig-6d/d mice were examined fortheir ability to ovulate normally in response to a superovulatoryregimen of gonadotropins. Mig-6f/f and Mig-6d/d yielded 24.75 � 6.61and 24.83 � 7.88 oocytes, respectively. Also, Mig-6f/f and Mig-6d/d

mice did not show any alterations in ovarian morphology andexhibited a normal estrus cycle (data not shown). Finally, cyclingfemale Mig-6f/f and Mig-6d/d mice exhibited normal levels of serumP4 and E2 at 2 and 5 months of age (data not shown). These resultsshow that ovarian morphology, steroidogenesis, and function werenot affected in the Mig-6d/d females. These data suggest that thedefects observed in the Mig-6d/d mice are inherent to the uterus.Thus, to determine if the infertility was in part because of loss of theability of the uterus to support implantation, we investigated theability of the uterus to undergo a decidual reaction. Ovariectomizedfemale Mig-6f/f and Mig-6d/d mice (n � 3) were treated withhormones and the uterus was mechanically stimulated to mimic thesignaling of the embryo at implantation and to induce decidual-ization. Gross anatomy of the decidual and control horn showed anincrease in size for Mig-6d/d mice compared to the Mig-6f/f mice (Fig.S3). However, the ratio of stimulated-to-unstimulated horn weightwas not changed in the Mig-6f/f and Mig-6d/d mice. These resultssuggest that ablation of Mig-6 in PR-expressing cells alters murinefertility because of dysregulation of E2 and P4 but not because ofdefects in ovarian or uterine function.

Tumor Suppressor Function of Mig-6 in the Uterus. As endometrialhyperplasia is an immediate precursor to endometrioid-type endo-metrial carcinoma, the hyperplastic phenotype in the Mig-6d/d micesuggests that Mig-6 has a tumor suppressor role in the tumorigenesisof endometrial cancer. We examined the role of ovarian steroidhormones in the development of the hyperplastic phenotype inMig-6d/d mice. Six-week-old Mig-6f/f and Mig-6d/d mice were ovari-ectomized and treated with vehicle, E2, or E2 � P4 and killed at 5months of age (n � 10 per genotype per treatment). Ovariecto-mized Mig-6d/d mice did not develop endometrial hyperplasia asobserved in intact Mig-6d/d mice (Fig. 3A). This demonstrates thatthe endometrial hyperplasia phenotype of Mig-6d/d mice is depen-dent on ovarian hormone stimulation.

Mig-6d/d mice treated with E2 for 3 months showed a significantincrease in uterine weight compared to Mig-6f/f mice (Fig. 3B).Although the Mig-6f/f mice showed endometrial hyperplasia asexpected from chronic E2 treatment, they did not show thepathology observed in the uteri of intact Mig-6d/d mice or Mig-6d/d

mice treated with E2. All of the Mig-6d/d mice treated with E2developed invasive endometrioid-type endometrial adenocarci-noma. The neoplastic endometrial glands in the Mig-6d/d miceinvaded through the uterine muscle wall and invaded adjacentstructures such as the colon, pancreas, and skeletal muscle. Thisresult demonstrates that Mig-6 may have an estrogen-dependenttumor suppressor function in endometrial cancer. Finally, ovariec-tomized Mig-6d/d mice treated with E2 � P4 for 3 months showed

a significant increase in uterine wet weight and developed endo-metrial hyperplasia (Fig. 3C) but not the endometrial carcinomaobserved in the E2-treated mice. Therefore, P4 treatment was ableto attenuate the pathology observed in the Mig-6d/d mice after E2treatment but not completely block the endometrial hyperplasia.These results demonstrate that Mig-6 is important to regulate theresponse of the uterus to E2 stimulation in part by mediating theprotective action of P4. However, Mig-6 also regulates other path-ways independent of P4 that control endometrial cell proliferation.

Downregulation of MIG-6 in Human Endometrial Cancer. Mig-6 abla-tion shows altered uterine function because of the inability of P4 toattenuate E2 action, which is a common characteristic of endome-trial dysfunction in humans (35, 36). The expression of Mig-6 in thehuman endometrium during the menstrual cycle was determined byreal-time quantitative PCR and immunohistochemistry. The ex-pression of Mig-6 was highest in the early secretory phase of thecycle (Fig. 4A) and this increase in expression was the result of anincrease in the expression of Mig-6 in the endometrial epithelium(Fig. 4B). The increase in expression of Mig-6 during this phase ofthe cycle correlates with P4 regulation as observed in the mouse.We investigated the expression of MIG-6 in endometrial biopsiesfrom patients with endometrioid carcinoma (n � 10) and normalendometrium (n � 5). The level of MIG-6 mRNA is significantlydecreased in patients with endometrioid carcinoma (32.8%) comparedto endometrial biopsies taken from normal women during the secretoryphase of the cycle. (Fig. 4C). Immunohistochemical analysis alsoshows a decrease in the protein level of MIG-6 in patients withendometrial cancer compared to normal women (Fig. 4D).

DiscussionP4, acting through its nuclear receptors, plays important roles inuterine functions associated with the establishment and mainte-nance of pregnancy (37, 38). The identification of P4-regulatedpathways in the uterus is thus crucial for understanding theimpairments that underlie disruption of steroid hormone control ofuterine cell proliferation and differentiation. In previous studies, wehave identified P4-regulated genes using microarray analysis onP4-treated PRKO uteri (30). We have identified Mig-6 as a targetof SRC-1 and PR in the uterus.

Fig. 4. Expression of MIG-6 in endometrial tissue from healthy women andfrom women with endometrioid endometrial cancer. (A) The expression ofMIG-6 in the human endometrium during the menstrual cycle. Real-timeRT-PCR analysis of MIG-6 was performed on human endometrium during themenstrual cycle. (B) Immunohistochemistry analysis of MIG-6 was performedin the human endometrium during the menstrual cycle. (C) Real-time RT-PCRanalysis of Mig-6 was performed on total RNA obtained from women withendometrioid carcinoma. (D) Immunohistochemical analysis of MIG-6 wasperformed on endometrium obtained from women with endometrioid car-cinoma. The results of real-time RT-PCR represent the mean � SEM. *, P � 0.05.

8680 � www.pnas.org�cgi�doi�10.1073�pnas.0903632106 Jeong et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 3,

202

0

Ablation of Mig-6 in mice results in a 50% reduction of theMig-6�/� litter size for an unidentified reason (31, 33). The 50% ofthe Mig-6�/� mice that escape the lethal phenotype, develop jointdeformities, and the majority of mice die within 6 months. Thesemice also develop neoplasias of the lungs and skin (23, 31). Theembryonic lethality and multitissue carcinogenesis makes it difficultto investigate the impact of ablation of Mig-6 in uterine biology. Toeffectively investigate the role of Mig-6 in the uterus, we generateda Mig-6 conditional null allele by introducing LoxP sites (33). Theuterine histology of Mig-6d/d mice demonstrate epithelial hyperpla-sia similar to the 9-month-old survival Mig-6�/� mouse (33). Thishyperplastic phenotype of Mig-6d/d and Mig-6�/� supports the tumorsuppressor role of Mig-6 in the tumorigenesis of endometrial cancer.

Mig-6 mediates the ability of P4 to regulate E2-dependent uterineweight gain. Normally, E2 stimulates uterine growth and epithelialcell proliferation (39). P4 antagonizes E2 actions, such as thestimulation of proliferation of the epithelial cells in the mouseuterus (40). Mig-6d/d and Mig-6f/f mice both respond to E2 treatmentwith an increase in uterine wet weight. This indicates that ablationof Mig-6 does not enhance the effect of E2 treatment alone. Whenwe examined the ability of P4 to inhibit E2-induced uterine hyper-trophy, P4 did not inhibit the E2-induced hypertrophy in Mig-6d/d

mice (Fig. 2). In a separate experiment, we treated wild type,PRcre/�, Mig-6f/f, and Mig-6d/d with E2 and P4 for 3 days andmeasured uterine weight gain. As expected, the response of PRcre/�,Mig-6f/f, and wild-type mice was similar with P4 dampening theE2-induced uterine hypertrophy. The Mig-6d/d mice again demon-strated an increase in uterine weight gain in the presence of P4 andE2. These results demonstrate that Mig-6 mediates the ability of P4to regulate E2-dependent uterine weight gain. Examination ofendometrial epithelial P4 target gene expression showed no changein the ability of PR to regulate the expression of Fst and Areg in theMig-6d/d mouse. Interestingly, the expression of PR mRNA issignificantly decreased in the Mig-6d/d mice compared to that ofMig-6f/f (Fig. 2). Immunohistochemical analysis of PR expression inthe Mig-6d/d mice showed that epithelial PR expression is not alteredbut that PR expression in the endometrial stroma cells is signifi-cantly reduced. P4 attenuates E2 regulation of proliferation andgene expression by regulating the expression of a yet-to-be-identified paracrine signal from the stromal cells to the epithelialcells (41, 42). However, the increase in ER� target gene expressionat 8 weeks of age was not the result of a change in ER� orcoactivator level (data not shown). Interestingly, there is an increasein ER� levels at 5 months of age with the epithelial hyperplasticphenotype, but its impact on gene expression at this time remainsunknown. Nonetheless, we have gained valuable insight into steroidhormone regulation in the uterus and Mig-6’s role in that regulation.

In addition, we have shown that MIG-6 is expressed in thehuman endometrium in a cycle-dependent manner that cor-relates with its being under the control of P4 as observed in themouse. We have also demonstrated that its expression isdecreased in endometrioid carcinoma when compared toexpression in normal endometrium during the secretory phase.Since the endometriod carcinoma samples assayed were ac-quired from postmenopausal women, the decrease in MIG-6expression may be a result of the hormonal status of post-menopausal women. Endometrial diseases such as endometrialcancer and endometriosis are known to be hormone-relatedmalignancies. Exposure to E2 is one of the endocrine riskfactors for developing endometrial cancer and endometriosis(35), and a lower incidence of these diseases is noted in womenwith decreased endogenous E2 production. In contrast, P4exposure is a negative risk factor for these disease (43), andpregnancy or progestin-based therapies can lead to diseaseregression in some women (44). Since endometrial cancer is adisease most often found in postmenopausal women, the lowerlevels of progesterone in these women may result in a lack ofinduction of MIG-6 that may be required to regulate endo-

metrial epithelial proliferations. Recently, published microar-ray gene expression profiles of the endometrium of womenwith or without endometriosis showed that a number of P4target genes incluing MIG-6 were dysregulated during thewindow of implantation, at which time the endometrium isexposed to the highest levels of P4 (29, 45). Mig-6d/d micedeveloped endometrial adenocarcinoma with E2 but not withE2 � P4. Thus, progesterone acting through the PR may bebeneficial for controlling endometrial cancer by inducingMig-6 expression. The Mig-6d/d mice treated with E2 � P4 stilldevelop endometrial hyperplasia, which suggests that Mig-6 isa critical factor involved in P4 protection against the devel-opment of endometrial cancer. However, the ability of P4 toprevent E2-induced endometrial adenocarcinoma despite theabsence of Mig-6 and the reduction in PR levels in the stromaimplicates additional mechanisms of protection. Furthermore,the development of E2-induced endometrial adenocarcinomain Mig-6d/d mice suggests that Mig-6 has an important role asa negative regulator of E2-induced tumorigenesis. However, itis not only the expression of MIG-6 that has been shown to bealtered in cancer. In addition, the MIG-6 gene has been shownto be mutated in the human NSCLC cell lines NCI-H226 andNCI-H 322M, and in 1 primary human lung cancer (24). Thus,when examining cases of P4 resistance in endometrial cancer,aside from examining MIG-6 expression levels, endometrialcarcinoma samples should also be assayed for mutations inMIG-6 as these mutations may be as detrimental as loss ofMIG-6 expression. Regardless, the molecular mechanism bywhich loss of MIG-6 function, either by loss of expression ormutations in the MIG-6 gene, regulates endometrial cancerneeds to be addressed. Further dissection of the intricacies ofthese pathways will lend important insight into the mecha-nisms that regulate endometrial tumorigenesis.

In summation, Mig-6 ablation results in increased ER� activityin the presence of P4, which normally antagonizes ER activity.In humans, we have shown that the expression of MIG-6 isdecreased in human endometrial endometrioid carcinoma. Ourfindings demonstrate that Mig-6 is a novel mediator of steroidhormone signaling in the uterus. Endometrial cancer is a uterinedisease in which hormonal regulation is perturbed. The alteredexpression of MIG-6 in endometrial cancer may serve as apossible cause of these pathologies. Mice with conditionalablation of Mig-6 in the uterus provide a more faithful model forhuman endometrial cancer with respect to pathology and hor-mone sensitivity than any previous models. This model is usefulfor finding new targets for the diagnosis and treatment ofendometrial cancer. Determining how Mig-6 mediates this actionwill be critical in understanding the role of steroid hormonesignaling in endometrial function and dysfunction and in devel-oping therapy for both uterine diseases.

Materials and MethodsAnimals and Hormone Treatments. Mice were maintained in the designatedanimal care facility at Baylor College of Medicine according to the institu-tional guidelines for the care and use of laboratory animals. For microarrayanalysis, ovariectomized SRC-1�/� and wild-type mice were injected withvehicle (sesame oil) or P4 (1 mg/mouse in 100 �L sesame oil) for 4 h (n � 9per genotype per treatment). For the steroid hormone treatment, ovari-ectomized mice Mig-6f/f and Mig-6d/d mice were injected with 1 of thefollowing: vehicle (sesame oil), P4 (1 mg/mouse), E2 (0.1 �g/mouse), P4 plusE2 (n � 5 per genotype per treatment). Hormone injections were repeatedevery day to prevent the effect of hormone degradation by metabolism.Mice were killed 6 h after the third injection. At the time of dissection,uterine tissues were placed in the appropriate fixative or flash frozen andstored at �80 °C. For the endometrial cancer study, ovariectomized Mig-6d/d and Mig-6f/f mice received a pellet of either vehicle (beeswax), E2 (20�g/pellet), or E2 � P4 (20 mg/pellet) at 8 weeks of age. Mice were killed at5 months of age (n � 10 per genotype per treatment).

Jeong et al. PNAS � May 26, 2009 � vol. 106 � no. 21 � 8681

MED

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Nov

embe

r 3,

202

0

Human Samples. Endometrial samples were obtained from 18 normallycycling women, aged 18 –35, after written informed consent, under anapproved protocol by the Institutional Review Board at Baylor College ofMedicine. The endometrial sample was removed from the uterine funduswith a Pipelle (circle R) biopsy catheter. Tissues were fixed in formalin andembedded in paraffin for histological analysis or snap frozen on dry ice.Histological samples were examined blindly by an independent patholo-gist, and phases were assigned according to the Noyes criteria (46). Endo-metrioid carcinoma samples were derived from hysterectomy surgicalspecimens submitted to the Department of Pathology, M. D. AndersonCancer Center following the guidelines approved by the M. D. AndersonCancer Center Committee on Human Research and the Baylor College ofMedicine Committee on the Use of Human Subjects in Medical Research.Classification was verified by light microscopic examination of hematoxylinand eosin-stained slides by gynecologic pathologist Russell R. Broaddusd.Normal endometrial samples were obtained from 5 cycling women (4secretory and 1 atrophic stage) and endometrioid carcinoma samples wereobtained from 8 postmenopausal women and 2 cycling women. Endometri-oid carcinoma samples from patients with a prior history of hormone use,radiation treatment, or chemotherapy were not used for this study.

Microarray Analysis. Microarray analysis was performed by Affymetrix mu-rine genome U74Av2 mouse oligonucleotide arrays (Affymetrix) as previ-ously described (30). All experiments were repeated 3 times. Briefly, weused DNA-Chip analyser dChip version 1.3 (47). We selected differentiallyexpressed genes within each time exposure using 2 sample comparisonsaccording to the following criteria: lower bound of 90% confidence inter-val of fold change greater than 1.2 and absolute value of differencebetween group means greater than 80. After excluding expressed se-quence tags with no functional annotation, differentially expressed genes

were classified according to gene ontology function using Affymetrixannotation, literature search in PubMed and GenMAPP (48).

Quantitative Real-Time RT-PCR. Expression levels of regulated genes werevalidated by real time RT-PCR. Real-time probes and primers were pur-chased from Applied Biosystems). All real-time RT-PCR was done using RNAsamples from 3 separate mice and mRNA quantities were normalizedagainst 18S RNA using ABI rRNA control reagents. Statistical analyses used1-way ANOVA followed by Tukey’s post hoc multiple range test with theInstat package from GraphPad.

Immunohistochemistry. Uterine sections from paraffin-embedded tissue werecut at 5 �m and mounted on silane-coated slides, deparaffinized, and rehydratedin a graded alcohol series. Sections were preincubated with 10% normal serum inPBS (pH 7.5) and then incubated with 1:1,000 anti-Mig-6 antibody (Sigma-Aldrich) in10%normal seruminPBS (pH7.5).Onthefollowingday, sectionswerewashed in PBS and incubated with a secondary antibody (5 �L/mL; Vector Labo-ratories) for 1 h at room temperature. Immunoreactivity was detected using theVectastain Elite ABC kit (Vector Laboratories).

ACKNOWLEDGMENTS. We thank Jinghua Li and Bryan Ngo for technical assis-tance; Heather L. Franco and Janet DeMayo for manuscript preparation. Thiswork was supported by the National Institute of Child Health and Human Devel-opment and the National Institutes of Health (NIH) as part of the CooperativeProgram on Trophoblast-Maternal Tissue Interactions U01HD042311 and NIHGrantU54HD0077495 (toF.J.D.), SPORE inUterineCancerNIH1P50CA098258–01(to R.R.B.), NIH Grant R01HD057873 and pilot grant from Specialized Program ofResearch Excellence in Uterine Cancer NIH 1P50CA098258–01 (to J.W.J.), NIHGrant RO1-CA77530 and the Susan G. Komen Award BCTR0503763 (to J.P.L.), NIHGrant 2U54HD035041–11 (to S.L.Y. and B.A.L.), and by the generosity of the Jayand Betty Van Andel Foundation (to Y.W.Z. and G.V.W.).

1. Clarke CL, Sutherland RL (1990) Progestin regulation of cellular proliferation. EndocrRev 11:266–301.

2. Lydon JP, et al. (1995) Mice lacking progesterone receptor exhibit pleiotropic repro-ductive abnormalities. Genes Dev 9:2266–2278.

3. Jemal A, et al. (2006) Cancer statistics, 2006. CA Cancer J Clin 56:106–130.4. Jick SS, Walker AM, Jick H (1993) Estrogens, progesterone, and endometrial cancer.

Epidemiology 4:20–24.5. Ziel HK, Finkle WD (1975) Increased risk of endometrial carcinoma among users of

conjugated estrogens. N Engl J Med 293:1167–1170.6. Li X, Lonard DM, O’Malley BW (2004) A contemporary understanding of progesterone

receptor function. Mech Ageing Dev 125:669–678.7. Xu J, et al. (1998) Partial hormone resistance in mice with disruption of the steroid

receptor coactivator-1 (SRC-1) gene. Science 279:1922–1925.8. Voegel JJ, Heine MJ, Zechel C, Chambon P, Gronemeyer H (1996) TIF2, a 160 kDa

transcriptional mediator for the ligand-dependent activation function AF-2 of nuclearreceptors. EMBO J 15:3667–3675.

9. Xu J, et al. (2000) The steroid receptor coactivator SRC-3 (p/CIP/RAC3/AIB1/ACTR/TRAM-1) is required for normal growth, puberty, female reproductive function, andmammary gland development. Proc Natl Acad Sci USA 97:6379–6384.

10. Anzick SL, et al. (1997) AIB1, a steroid receptor coactivator amplified in breast andovarian cancer. Science 277:965–968.

11. Onate SA, Tsai SY, Tsai MJ, O’Malley BW (1995) Sequence and characterization of acoactivator for the steroid hormone receptor superfamily. Science 270:1354–1357.

12. Wong CW, Komm B, Cheskis BJ (2001) Structure-function evaluation of ER alpha and betainterplay with SRC family coactivators. ER selective ligands. Biochemistry 40:6756–6765.

13. McKenna NJ, O’Malley BW (2002) Combinatorial control of gene expression by nuclearreceptors and coregulators. Cell 108:465–474.

14. Han SJ, et al. (2005) Dynamic cell type specificity of SRC-1 coactivator in modulatinguterine progesterone receptor function in mice. Mol Cell Biol 25:8150–8165.

15. Jeong JW, et al. (2006) The genomic analysis of the impact of steroid receptorcoactivators ablation on hepatic metabolism. Mol Endocrinol 20:1138–1152.

16. Saarikoski ST, Rivera SP, Hankinson O (2002) Mitogen-inducible gene 6 (MIG-6),adipophilin and tuftelin are inducible by hypoxia. FEBS Lett 530:186–190.

17. van Laar T, Schouten T, van der Eb AJ, Terleth C (2001) Induction of the SAPK activatorMIG-6 by the alkylating agent methyl methanesulfonate. Mol Carcinog 31:63–67.

18. Wick M, Burger C, Funk M, Muller R (1995) Identification of a novel mitogen-induciblegene (mig-6): regulation during G1 progression and differentiation. Exp Cell Res219:527–535.

19. Burbelo PD, Drechsel D, Hall A (1995) A conserved binding motif defines numerous candidatetarget proteins for both Cdc42 and Rac GTPases. J Biol Chem 270:29071–29074.

20. Pirone DM, Carter DE, Burbelo PD (2001) Evolutionary expansion of CRIB-containingCdc42 effector proteins. Trends Genet 17:370–373.

21. Mahgoub MA, Abd-Elfattah AS (1998) Diabetes mellitus and cardiac function. Mol CellBiochem 180:59–64.

22. Makkinje A, et al. (2000) Gene 33/Mig-6, a transcriptionally inducible adapter proteinthat binds GTP-Cdc42 and activates SAPK/JNK. A potential marker transcript for chronicpathologic conditions, such as diabetic nephropathy. Possible role in the response topersistent stress. J Biol Chem 275:17838–17847.

23. Ferby I, et al. (2006) Mig6 is a negative regulator of EGF receptor-mediated skinmorphogenesis and tumor formation. Nat Med 12:568–573.

24. Zhang YW, et al. (2007) Evidence that MIG-6 is a tumor-suppressor gene. Oncogene26:269–276.

25. Girard L, Zochbauer-Muller S, Virmani AK, Gazdar AF, Minna JD (2000) Genome-wideallelotyping of lung cancer identifies new regions of allelic loss, differences betweensmall cell lung cancer and non-small cell lung cancer, and loci clustering. Cancer Res60:4894–4906.

26. Nomoto S, et al. (1998) Search for mutations and examination of allelic expressionimbalance of the p73 gene at 1p36.33 in human lung cancers. Cancer Res 58:1380–1383.

27. Amatschek S, et al. (2004) Tissue-wide expression profiling using cDNA subtraction andmicroarrays to identify tumor-specific genes. Cancer Res 64:844–856.

28. Anastasi S, et al. (2005) Loss of RALT/MIG-6 expression in ERBB2-amplified breastcarcinomas enhances ErbB-2 oncogenic potency and favors resistance to Herceptin.Oncogene 24:4540–4548.

29. Burney RO, et al. (2007) Gene expression analysis of endometrium reveals progester-one resistance and candidate susceptibility genes in women with endometriosis.Endocrinology. 148:3814–3826.

30. Jeong JW, et al. (2005) Identification of murine uterine genes regulated in a ligand-dependent manner by the progesterone receptor. Endocrinology 146:3490–3505.

31. Zhang YW, et al. (2005) Targeted disruption of Mig-6 in the mouse genome leads toearly onset degenerative joint disease. Proc Natl Acad Sci USA 102:11740–11745.

32. Zhang YW, Vande Woude GF (2007) Mig-6, signal transduction, stress response andcancer. Cell Cycle 6:507–513.

33. Jin N, Gilbert JL, Broaddus RR, Demayo FJ, Jeong JW (2007) Generation of a Mig-6conditional null allele. Genesis 45:716–721.

34. Soyal SM, et al. (2005) Cre-mediated recombination in cell lineages that express theprogesterone receptor. Genesis 41:58–66.

35. Stovall DW, Halme J (1991) Endometriosis and associated pathology. Curr Opin ObstetGynecol 3:853–858.

36. Surrey ES, Halme J (1990) Effect of peritoneal fluid from endometriosis patients onendometrial stromal cell proliferation in vitro. Obstet Gynecol 76:792–797.

37. Conneely OM, Mulac-Jericevic B, Lydon JP, De Mayo FJ (2001) Reproductive functionsof the progesterone receptor isoforms: Lessons from knock-out mice. Mol Cell Endo-crinol 179:97–103.

38. Conneely OM, Mulac-Jericevic B, DeMayo F, Lydon JP, O’Malley BW (2002) Reproduc-tive functions of progesterone receptors. Recent Prog Horm Res 57:339–355.

39. Lubahn DB, et al. (1993) Alteration of reproductive function but not prenatal sexualdevelopment after insertional disruption of the mouse estrogen receptor gene. ProcNatl Acad Sci USA 90:11162–11166.

40. Huet-Hudson YM, Andrews GK, Dey SK (1989) Cell type-specific localization of c-mycprotein in the mouse uterus: Modulation by steroid hormones and analysis of theperiimplantation period. Endocrinology 125:1683–1690.

41. Cunha GR, Cooke PS, Kurita T (2004) Role of stromal-epithelial interactions in hormonalresponses. Arch Histol Cytol 67:417–434.

42. Tibbetts TA, Mendoza-Meneses M, O’Malley BW, Conneely OM (1998) Mutual andintercompartmental regulation of estrogen receptor and progesterone receptor ex-pression in the mouse uterus. Biol Reprod 59:1143–1152.

43. Grosskinsky CM, Halme J (1993) Endometriosis: The host response. Baillieres Clin ObstetGynaecol 7:701–713.

44. Kaunitz AM (1998) Injectable depot medroxyprogesterone acetate contraception: Anupdate for U.S. clinicians. Int J Fertil Womens Med 43:73–83.

45. Kao LC, et al. (2003) Expression profiling of endometrium from women with endome-triosis reveals candidate genes for disease-based implantation failure and infertility.Endocrinology 144:2870–2881.

46. Noyes RW, Hertig AT, Rock J (1975) Dating the endometrial biopsy. Am J ObstetGynecol 122:262–263.

47. Li C, Wong WH (2001) Model-based analysis of oligonucleotide arrays: Expression indexcomputation and outlier detection. Proc Natl Acad Sci USA 98:31–36.

48. DahlquistKD,SalomonisN,VranizanK,LawlorSC,ConklinBR(2002)GenMAPP,anewtoolfor viewing and analyzing microarray data on biological pathways. Nat Genet 31:19–20.

8682 � www.pnas.org�cgi�doi�10.1073�pnas.0903632106 Jeong et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 3,

202

0