Unique Roles of Estrogen-Dependent Pten Control in Epithelial Cell Homeostasis of Mouse Vagina

ORIGINAL ARTICLE Unique roles of estrogen-dependent Pten control in epithelial cell homeostasis of mouse vagina

S Miyagawa1, M Sato1, T Sudo2, G Yamada3 and T Iguchi1

Numerous studies support a role of phosphatase and tensin homolog deleted from chromosome 10 (Pten) as a tumor suppressor gene that controls epithelial cell homeostasis to prevent tumor formation. Mouse vaginal epithelium cyclically exhibits cell proliferation and differentiation in response to estrogen and provides a unique model for analyzing homeostasis of stratified squamous epithelia. We analyzed vaginal epithelium-specific Pten conditional knockout (CKO) mice to provide new insights into Pten/phosphoinositide-3-kinase (PI3K)/Akt function. The vaginal epithelium of ovariectomized (OVX) mice (control) was composed of 1–2 layers of cuboidal cells, whereas OVX CKO mice exhibited epithelial hyperplasia in the suprabasal cells with increased cell mass and mucin production. This is possibly due to misactivation of mammalian target of rapamycin and mitogen-activated protein kinase. Intriguingly, estrogen administration to OVX Pten CKO mice induced stratification and keratinized differentiation in the vaginal epithelium, as in estrogen-treated controls. We found that Pten is exclusively expressed in the suprabasal cells in the absence of estrogens, whereas estrogen administration induced Pten expression in the basal cells. This suggests that Pten acts to prevent excessive cell proliferation as in the case of other squamous tissues. Thus, Pten exhibits a dual role on the control of vaginal homeostasis, depending on whether estrogens are present or absent. Our results provide new insights into how Pten functions in tissue homeostasis.

Oncogene (2015) 34, 1035–1043; doi:10.1038/onc.2014.62; published online 17 March 2014

INTRODUCTION moving up through the epithelium. Apical cells exhibit keratiniza- Phosphatase and tensin homolog deleted from chromosome 10 tion. The fully stratified and keratinized vaginal epithelium (Pten) is a lipid phosphatase that functions as a tumor suppressor, resembles the typical stratified and keratinized epidermis found and mutations in Pten are frequently found in both sporadic and in the skin and other organs. Thus, the vaginal epithelium provides hereditary cancers. Pten acts in opposition to phosphoinositide-3- a unique model to study homeostasis in stratified squamous kinase (PI3K) function. In the absence of Pten activity, concentra- epithelia. There are several case reports of vaginal cysts in the ’ tions of phosphatidylinositol 3,4,5-trisphosphate (PIP3), a lipid patients with Cowden s disease, which is associated with Pten second messenger produced by PI3K, are increased, leading to the mutation.2,3 In addition, there is one reported description of a enhancement of phosphorylation and activation of the v-akt vaginal squamous cell carcinoma in Pten mutant mice.4 However, murine thymoma viral oncogene homolog (Akt). Akt kinase little is known about the usual function of Pten/PI3K/Akt signaling activity exerts antiapoptotic and proproliferative functions; there- or its relationship with estrogen signaling in the vagina. fore, mice with Pten deletion and/or loss-of-function mutations are In rodent reproductive organs such as vagina and uterus, highly susceptible to tumor induction by abnormal Akt activation. the effects of estrogen on the epithelia are mediated primarily Pten has a pivotal role in maintaining stratified squamous via stromally expressed estrogen receptor α (ERα).5,6 Estrogen- epithelia because conditional knockout (CKO) of Pten in the induced growth factors secreted from the stroma promote epithelium leads to hyperplasia, hyperkeratosis and tumor epithelial cell proliferation,7 resulting in the activation of cellular formation in the skin, esophagus and stomach.1 signal transduction via PI3K/Akt and mitogen-activated Cell proliferation and differentiation of stratified squamous protein kinase (MAPK) signaling cascades.8,9 EGF (epidermal epithelia must be tightly regulated and coordinated during growth factor)-like growth factors and insulin-like growth factor-I homeostasis. Vaginal epithelium, despite its similarity to other have mitogenic effects similar to estrogens, and administration of stratified squamous epithelia, is exceptional in one major way—- these growth factors to OVX adult mice induced cell proliferation the vaginal epithelium exhibits cyclical, estrogen-dependent cell and differentiation in the female reproductive tract.10–12 These proliferation and differentiation during the estrous cycle. The results suggest that Akt and MAPK signalings are functional vaginae of ovariectomized (OVX) mice contain an atrophied mediators of estrogen-induced cell proliferation and differentia- epithelium of 2–3 cell layers; estrogen administration rapidly tion. Importantly, aberrant PI3K/Akt activation results in complex induces epithelial cell proliferation in the basal layer. The atypical hyperplasia and endometrioid carcinoma in the suprabasal cells are no longer mitogenic but differentiate while uterus.13–16 Therefore, repression of PI3K/Akt signaling is

1Okazaki Institute for Integrative Bioscience, National Institute for Basic Biology, National Institutes of Natural Sciences, and Department of Basic Biology, The Graduate University for Advanced Studies, Okazaki, Aichi, Japan; 2Section of Translational Research and Department of Gynecologic Oncology, Hyogo Cancer Center, Akashi, Hyogo, Japan and 3Department of Developmental Genetics and Laboratory Animal Center, Institute of Advanced Medicine, Wakayama Medical University, Wakayama, Wakayama, Japan. Correspondence: Professor T Iguchi, Okazaki Institute for Integrative Bioscience, National Institute for Basic Biology, National Institutes of Natural Sciences, and Department of Basic Biology, The Graduate University for Advanced Studies, Okazaki, Aichi 444-8787, Japan. E-mail: [email protected] Received 25 August 2013; revised 24 December 2013; accepted 19 January 2014; published online 17 March 2014 Pten function in vagina S Miyagawa et al 1036 considered to be essential during the absence of estrogen in the indicating mucin production. Such mucin production was vagina as well. observed restrictedly in the superficial layer of cells, but not in In the present study, we analyzed epithelium-specific the basal cells, of the vaginal epithelium. PAS- and Alcian blue- Pten CKO mouse vagina to provide new insights into Pten/PI3K/ positive cells were limited in the superficial layer of the control Akt function. We found that Pten is expressed in suprabasal vaginae. Thus, the phenotypes of Pten CKO mice probably result epithelial cells, where it prevents abnormal cell proliferation and from defects in the control of epithelial cell proliferation and differentiation in the absence of estrogen. On the other hand, in differentiation, causing increased mucin production in epithelial the presence of estrogen, Pten is predominantly expressed in cells. Intriguingly, 17β-estradiol (E2) administration in Pten CKO basal epithelial cells, where it may aid in preventing tumor mice induced stratification and keratinized differentiation in the induction. Thus, Pten exhibits a dual role on the control of vaginal vaginal epithelium comparable with that in E2-treated controls homeostasis, depending on the presence and the absence of (Figures 2g and h). The epithelium in the Pten CKO mice estrogen. administered with E2 is thicker than that of controls (Figure 2i). PAS- and Alcian blue-positive cells were not observed in E2- treated control or Pten CKO mouse vagina (data not shown). At RESULTS 2 weeks after the last E2 administration, Pten CKO mouse vaginae Effect of epithelial cell-specific inactivation of Pten on the mouse exhibited a recurrence of the abnormal phenotypes (Figure 2j), vagina and thus the vaginal changes in the Pten CKO mouse are CK5 is generally expressed in the stratified and squamous dependent on the absence of estrogen. We also investigated epithelium.17 Hence, Pten CKO induced by CK5-Cre-mediated the effects of a prolonged E2 exposure. Two months after deletion progressively developed hyperplasia and keratosis implantation with E2 pellet, Pten CKO mice exhibited frequent in the stratified and keratinized organs, which include epidermis invagination of the vaginal epithelium into the underlying stroma (Figures 2k and l), which might be associated with tumor of the skin, tongue, esophagus and forestomach (Supplementary 4 Figure S1). We first explored whether Pten CKO induced formation later in life. hyperplasia and/or hyperkeratinization in the vagina as well. The epithelial cell-specific expression of Cre in the K5-Cre line was confirmed in the Rosa reporter mouse (Supplementary Figure S2). Cell differentiation in the Pten CKO mouse vaginal epithelial cells Accordingly, the epithelium-specific deletion of Pten was To further characterize the phenotypes of the Pten CKO mouse evident in the Pten CKO mouse vagina (Figure 1; note that vagina, several immunohistochemical stainings were performed. stromally expressed Pten is still evident). In the current study, CK14 is a whole epithelial marker in the vagina.18 In both control OVX mice were used to avoid any effects of hypothalamus–- and Pten CKO mice, CK14 was expressed throughout all layers of pituitary–gonadal axis, and to simply analyze the effects the vaginal epithelium (Figures 3a and b). In the control OVX mice, of Pten/PI3K/Akt in the absence of estrogen signaling (all the stratified and squamous differentiation marker, CK1, was information for intact Pten CKO phenotypes are included in expressed in several suprabasal cells, whereas its expression was Supplementary Figure S3). not observed in the OVX Pten CKO mice (Figures 3c and d). CK8 is The vaginal epithelium of 8-week-old OVX mice (control) was a marker for undifferentiated cells and was expressed in some composed of 1–2 layers of cuboidal cells (Figure 2a). In the CKO suprabasal cells in the control OVX mouse vaginae (Figure 3e). mice, epithelial hyperplasia was evident, which was accompanied Intriguingly, CK8 was strongly expressed in the OVX Pten CKO with distended cells in the superficial layer of the vagina (Figures mouse vagina (Figure 3f). We infer from these results that 2b and b'). We also found multiple abnormal gland-like pits in the suprabasal cells in the Pten CKO mouse vaginae are maintained epithelium. The epithelial cells lining the pits were positive for in an undifferentiated, proliferating state. P63 is normally periodic acid-Schiff (PAS) and Alcian blue staining (Figures 2c–f), expressed in the basal layer of the epithelium and is a crucial

Control (E2) Pten CKO (E2)

Control (OVX) Pten CKO (OVX)

Figure 1. Epithelium-speciﬁc deletion of Pten in the mouse vagina. Pten is predominantly expressed in the basal cells in the E2-administered control mouse vagina (a), whereas Pten is exclusively expressed in the suprabasal cells in the control OVX mice (b). In both E2-administered and OVX Pten CKO mouse vagina, Pten expression is not detected in the epithelium (c and d). Scale bar: 100 μm.

* PAS PAS m) 80 μ 60 40 20 Thickness of Alcian Blue Alcian Blue 0 the epithelium ( Control Pten Control (E2) Pten CKO (E2) CKO Pten CKO (E2 withdrawal)

Control Pten CKO (E2, 2 months) (E2, 2 months)

Figure 2. Effect of epithelial cell-specific inactivation of Pten on the mouse vagina. The vaginal epithelium of 8-week-old OVX control (a) and Pten CKO mice (b). Some parts of the Pten CKO vagina exhibit severe hyperplasic phenotypes with multiple abnormal gland-like pits (b’). PAS (c and d) and Alcian blue staining (e and f) indicate mucin production in the control (c and e) and Pten CKO vagina (d and f). E2 administration induces stratification and keratinized differentiation in the control (g) and Pten CKO mice (h). The epithelium in the Pten CKO mice administered with E2 is thicker than that of controls (i). The recurring abnormal phenotypes of vagina in Pten CKO mice at 2 weeks after the last E2 administration (j; n = 4). Prolonged E2 exposure results in papillomatosis, invagination of the vaginal epithelium into the underlying stroma in the CKO mice (k and l; n = 4). Scale bar: 100 μm. regulator of squamous differentiation.19,20 P63 was expressed in in the Pten CKO mouse vagina. Three successive administrations of cells adjacent to the basement membrane in OVX control mice, E2 strongly stimulated cell proliferation in both control and Pten whereas its expression in OVX Pten CKO mice was not restricted in CKO mice (Figures 4c–e). the basal cells and was observed some suprabasal cells away from We also examined whether reduced apoptosis in the epithelial the basement membrane (Figures 3g and h). cell contributed to the phenotypes in the Pten CKO mouse vagina. After E2 administration, these marker proteins were similarly Apoptotic indices as measured by terminal deoxynucleotidyl expressed in both control and the mutant mouse vagina (Figures transferase dUTP nick-end labeling staining did not differ 3i–n), but p63 expression in mutant animals was more extensive in significantly between control and mutant vaginal epithelia the upper layers (Figures 3o and p). (Figures 4f–h). Apoptotic cells were almost undetectable in the vagina of E2-treated control and mutant mice (data not shown). These results indicate that Pten loss-of-function enhanced cell Cell proliferation, but not apoptosis, is increased in the Pten CKO proliferation in the OVX mice but does not significantly alter mouse vaginae apoptosis. We next investigated the proliferation indices of vaginal epithelial cells in control and Pten CKO mice using bromodeoxyuridine (BrdU) incorporation. Proliferation indices were higher in the OVX Activation of mTOR and MAPK pathway in the mutants CKO than in OVX controls (Figures 4a, b and e). In both control and To identify the potential molecular mechanisms that could lead to mutants, BrdU-positive cells were mostly found adjacent to or abnormal cell proliferation and differentiation in the vaginal attached to the basement membrane, but some cells in the upper epithelium, we examined phosphorylation levels of known layer were positive for BrdU staining (red arrowheads in Figure 4b) downstream effector of the Akt signaling pathway, mammalian

CK14

CK1

CK8

p63

Figure 3. Expression pattern of cell differentiation marker in the mouse vagina. Immunohistochemical staining for CK14 (a, b, i and j), CK1 (c, d, k and l), CK8 (e, f, m and n) and p63 (g, h, o and p). In OVX Pten CKO mice, an undifferentiating cell marker, CK8, is expressed, but a stratiﬁed and squamous differentiating cell marker, CK1, is not (c–f), suggesting the undifferentiation status of the epithelia. P63 expression is only expressed in the basal layers of OVX control, whereas some cells away from the basal layer are also positive in the OVX Pten CKO mice (g and h). P63 expression area appeared to be extended to the upper layer upon E2 administration in the Pten CKO mouse (o and p). Scale bar: 100 μm.

Control Pten CKO (OVX)

30 * 20

index (%) 10 BrdU-labeling

Pten Pten Control Control CKO CKO OVX +E2

2.0

1.0 index (%) Appoptotic

0 Control Pten CKO

OVX Figure 4. Pten CKO leads to cell proliferation but does not contribute to epithelial cell survival in the mouse vagina. BrdU immunostaining in control (a and c) and Pten CKO mouse vagina (b and d) is shown. Note that BrdU-positive cells in the upper layer of the OVX K5-Cre; Ptenf/f mice (red arrowheads in (b)). Black arrowheads indicate normal cell proliferation in the basal layer. Proliferation indices in the OVX Pten CKO mouse vagina were higher than those of OVX controls (e). Terminal deoxynucleotidyl transferase dUTP nick-end labeling assay indicates that Pten mutation dose not inﬂuence cell apoptosis in the vaginal epithelium (f–h). Black arrowheads indicate apoptotic cells. Scale bar: 100 μm.

target of rapamycin (mTOR).21–23 As expected, enhanced Akt rapamycin, a speciﬁc inhibitor of mTOR, was administered to phosphorylation was observed in the mutant mouse vagina control and Pten CKO OVX mice. Rapamycin administration for (Figures 5a and b). In addition, mTOR phosphorylation was 3 weeks resulted in regression of cells in the suprabasal layer of signiﬁcantly increased in Pten mutant vaginae (Figures 5a and b). most part of the vagina, although mucus-like products still To explore the contribution of mTOR to the mutant phenotype, remained (Figures 5c–f). Levels of cell proliferation in mutant

Phospho- Phospho- Phospho- Akt (T308) Akt (S473) mTOR (S2448) CKO 1.2 0.8 1.2 Control Pten Phosphp- * Akt (T308) 0.6 * * Phosphp- 0.8 0.8 Akt (S473) 0.4 Akt 0.2 0.4 Phospho- 0.2 Relative expression

mTOR (S2448) (Phospho/total protein) 0 0 0 mTOR Control Pten Control Pten Control Pten CKO CKO CKO

Control (OVX) + rapamycin

CKO CKO Pten HE Alcian Blue Pten + vehicle + rapamycin Phosphp- Pten CKO (OVX) + rapamycin Akt (T308) Phosphp- Akt (S473)

Akt HE Alcian Blue

* *

10 * 8 6

(%) 4 2

BrdU-labeling index 0 Control Pten CKO Pten CKO (OVX) (OVX) (OVX) + rapamycin + rapamycin + vehicle Figure 5. mTOR activation is associated with hyperplasia in the mouse vagina. Activation status of Akt and mTOR (a). In addition to Akt, mTOR in the mutants show significant increase of phosphorylation when calculated with Image J program (b). Rapamycin administration can partially rescue the hyperplasic phenotype of Pten CKO mice, but mucous production remains (c–f; n = 5). The level of cell proliferation in the mutant is significantly reduced by rapamycin treatment (g). Rapamycin treatment did not influence the phosphorylation levels of Akt (h). Scale bar: 100 μm. rapamycin-treated group were slightly increased compared with mice administered with rapamycin (Figure 6e and f), supporting control rapamycin-treated mice (Figure 5g). Importantly, the index the fact that rapamycin did not repress mucin production in the mutants with rapamycin was reduced by half (5.1% vs (Figures 5d and f). 10.3%), compared with that of OVX Pten CKO mice without It is known that activated MAPK and/or Akt can stimulate the rapamycin (Figure. 5g). Rapamycin administration did not transcriptional activity of ERα (in a ligand-independent manner) by influence the Akt phosphorylation (Figure 5h), indicating specific phosphorylating serines 118 and 167 of this receptor (located inhibition of mTOR complex 1 pathway. These results indicate that within the AF-1 region), respectively.27–29 ERα protein was increased cell proliferation in Pten mutant vaginal epithelium at observed in the vaginal epithelium and in the stroma of both least partly requires the mTOR pathway. control and Pten CKO animals (Figures 6g and h). Both residues of Activation of MAPK has been implicated for mucin production the ERα were phosphorylated in the Pten CKO compared with in the respiratory tract, intestine and prostate.24–26 Indeed, we control animals (Figure 6i). On the other hand, expression of observed increased expression of phosphorylated (phospho) mito- growth factor mRNAs that are considered to be targets of ERα8 gen-activated protein kinase kinase and extracellular signal- was not significantly changed between controls and Pten CKO regulated kinase 1/2 in the suprabasal cells from Pten CKO mice (Figure 6j). Although phosphorylation of ERα has been animals (Figures 6a–d). Activated MAPK remained in the Pten CKO implicated in estrogen-independent cell proliferation and

pMEK pMEK ERα ERα

CKO ControlPten pERK1/2 pERK1/2 pERα (S118) pERα (S167) Pten CKO (OVX) Pten CKO (OVX) ERα + vehicle + rapamycin

Control 2.0 Pten CKO

pERK1/2 pERK1/2 1.0

Relative expression α

EGF APR EPR IGF-I TGF HB-EGF Figure 6. Conditional Pten deletion results in MAPK and ERα phosphorylation. Phospho-mitogen-activated protein kinase kinase (MEK) and extracellular signal-regulated kinase (ERK1/2) were detected in the suprabasal cells in the mutants (a–d). Activated MAPK remains in the Pten CKO mice administered with rapamycin (e and f). ERα protein was similarly expressed in both control and Pten CKO mouse vagina (g, and h), whereas phosphorylated ERα was detected only in the Pten CKO mice (i). Expression proﬁles of growth factor mRNAs in OVX control and Pten CKO mouse vagina (j). Scale bar: 100 μm.

differentiation,8,9 these results suggest that phenotypes induced DISCUSSION by Pten loss-of-function are independent of ERα signaling. Estrogens have a central role in female reproductive organ biology. Despite the importance of this organ system for fertility and women’s health, the signaling pathways that regulate cell Estrogen-dependent localization of phospho-Akt in the mouse proliferation and differentiation remain poorly understood. Pten vagina tumor suppressor gene lipid phosphatase activity acts in Lastly, we examined potential differences in Pten function among opposition to PI3K function, and Pten loss-of-function mutations several stratified and squamous epithelia in mice. Numerous frequently result in sporadic and hereditary cancers accompanied studies have shown that Pten exerts its tumor suppressive by Akt activation.1,4,30 Akt was shown to be phosphorylated as a function through inhibition of Akt activation; therefore, it is result of estrogen treatment and was described as a functional generally believed that a major Pten function appears to be in mediator of estrogen-induced cell proliferation and differentiation the regulation of cell proliferation in such organs. We found that in mouse uterus and vagina.9,31 Therefore, Akt signaling is Pten is mainly expressed in the basal cells in stratified and currently thought to be an essential mediator for estrogen- squamous epithelia, such as epidermis of the skin, tongue, induced events. If this is the case, activation of Akt, per se, may esophagus and forestomach, although Pten expression levels result in cell proliferation and squamous differentiation in the vary among these tissues (Figures 7a–d). Consequently, the mouse vaginal epithelium even in the absence of estrogen. To test deletion of Pten augmented the phosphorylation of Akt in the this hypothesis, we used an epithelial cell-specific Pten CKO mice. corresponding regions of the epithelia in the Pten CKO mice First, we expected cell proliferation and squamous hyperplasia (Figures 7e–h). in the vaginal epithelium, because Pten mutation results in The stratified and keratinized vagina of control mice squamous hyperplasia and tumor formation in other stratified and administered with E2 expressed Pten predominantly in the basal squamous tissues, such as the skin, esophagus and stomach.1 cells as well (see Figure 1a). Concomitantly, phospho-Akt Despite the intrinsic similarity among such differentiated stratified, was detected in the lower layers of epithelia including basal cells squamous epithelia, the phenotypes observed in the Pten CKO in the control (Figure 7i) and was augmented throughout the mouse vagina are unique. Cell proliferation in the Pten CKO mouse epithelium, particularly in the basal to suprabasal cell layers in the vaginal epithelium is indeed increased; however, a remarkable Pten CKO mice (Figure 7j). Intriguingly, in the OVX control mice, hyperplasia was observed in the suprabasal cells accompanied Pten was exclusively expressed in the suprabasal cells (see with abnormal mucin production. On the basis of the CK1, CK8 Figure 1b). The phosphorylated Akt was not detected in the and p63 expression pattern, the suprabasal cells in the Pten CKO OVX control mice (Figure 7k). When Pten was deleted, phospho- mice are maintained in a largely undifferentiated condition. We Akt was found in the suprabasal cells but not in the basal cells infer that the suprabasal cells in the Pten CKO mice fail to (Figure 7l). Thus, Pten expression and its regulation of Akt in the differentiate completely into squamous cells and retain an mouse vagina depend on the presence or the absence of undifferentiated status with the potential for cell proliferation. estrogen, and this might explain the unique phenotypes of the Immunohistochemical analysis in OVX control mouse vaginae Pten CKO mouse vagina. revealed the unique expression of Pten in the suprabasal layers.

Control

PTEN PTENPTEN PTEN

Pten CKO PTEN PTEN PTEN PTEN

Pten CKO pAkt pAkt pAkt pAkt

Control

pAkt pAkt pAkt pAkt

Vagina (E2) Vagina (OVX)

sbc bc

Control pAkt Pten CKOpAkt Control pAkt Pten CKO pAkt

Figure 7. Estrogen-dependent localization of phospho-Akt expression in the mouse vagina. Pten is mainly expressed in the basal cells in stratiﬁed and squamous epithelia in epidermis of the skin (a), tongue (b), esophagus (c) and forestomach (d) of control mice (note no such signals in the Pten CKO mice; a'–d'). In the Pten CKO mice, phospho-Akt expression is similarly detected in these tissues (e–h). Phospho-Akt expression in the control mice is not prominent (e'–h'). Phospho-Akt is predominantly expressed in the basal to suprabasal epithelial cells in the E2-administered control mouse vagina (i). In the Pten CKO mice, the phosphorylated Akt expression is augmented (j). Akt is not phosphorylated in the absence of estrogen in the control mice (k). In the OVX Pten CKO mouse vagina, phospho-Akt is evident in the suprabasal epithelial cells (sbc), but not in the basal epithelial cells (bc) (l). See also Figure 1 for Pten expression. Dotted lines indicate the basal layer of the epithelium. Scale bar: 100 μm.

Hence, the primary role for Pten in the mouse vagina seems to Thus, once epithelial cells undergo squamous differentiation, Pten block aberrant cell proliferation in the suprabasal cells rather than could also function as a tumor suppressor in the vagina. to control basal cell proliferation in the absence of estrogen. Intriguingly, the p63-expressing region is extended apically in It was unknown whether the vaginae in the OVX Pten CKO the Pten CKO mice. P63 is a marker for basal cell characteristics mouse develop tumors. It has been reported that 90% of OVX Pten and is associated with proliferative capacity in the stratified heterozygote mice developed uterine epithelial hyperplasia.32 This epithelium. Activation of Akt induces increase of p63 expression suggests that Pten loss-of-function is sufficient for the develop- and is associated with transformation of keratinocytes.35 ment of complex atypical hyperplasia in the absence of estrogen. Our mouse model revealed several downstream effectors of As complex atypical hyperplasia often precedes endometrial PI3K/Akt in the mouse vagina. One of the targets is mTOR, which carcinoma,32 this may be consistent with a role for Pten in vaginal showed increased phosphorylation levels together with Akt cancers. Therefore, we could not exclude the possibility of phosphorylation, resulting from Pten deletion. mTOR regulates spontaneous tumorigenesis in the Pten CKO mouse vagina translation and cell growth as a central component of raptor without estrogenic stimulation. Several lines of epithelial cell- (mTORC1) or rictor (mTORC2) complexes via phosphorylation of specific Pten mutants have shown decreased survival rate with their substrates (e.g., p70 ribosomal S6 kinase and 4E binding severe growth retardation and developments of multiple dis- protein). The PI3K/Akt/mTOR pathway also represents a prolif- orders as age progressed.1,30 Further experiments using alter- erative pathway activated by PI3K and Pten loss, and contributes native Pten CKO mice with Cre expression restricted to the vaginal to the formation of multiple tumor types.30,36 Notably, tumors epithelium, and higher survivorship will be required to address involving Pten deletions or Akt activations are highly sensitive to a this issue. specific inhibitor of mTOR, rapamycin.37,38 In our mouse model, Pten expression shifts from the suprabasal layer to the basal the levels of cell proliferation, particularly in the suprabasal layers, layer upon estrogen administration, as was observed in other were decreased by rapamycin treatment, accompanied with stratified and squamous tissues. Ours is the first demonstration of partial rescue of their hyperplasia, supporting a role for the Akt/ an estrogen-dependent role for Pten. Pten appears to regulate cell mTOR pathway in the mutant phenotypes. Rapamycin treatment proliferation in the basal layer to prevent squamous hyperplasia clearly diminished cell proliferation, although mucin production in and/or tumors in the presence of estrogens. In addition, the suprabasal layer cells was retained. This supports the prolonged E2 exposure induced invagination of the vaginal involvement of an additional alternative pathway for such epithelium into the underlying stroma in the mutants, which abnormal cell differentiation. We found MAPK activation in the phenotypes are associated with vaginal and cervical cancer.33,34 suprabasal cells in the mutant mouse vagina, and activation of

© 2015 Macmillan Publishers Limited Oncogene (2015) 1035 – 1043 Pten function in vagina S Miyagawa et al 1042 Pten function in vagina vagina. For a long treatment, OVX mice were implanted with E2 pellets (0.01 mg per pellet, 2 months; Innovative Research of America, Sarasota, FL, USA) into subcutaneous tissue for 2 months. Rapamycin (Toronto Research Inhibition of abnormal cell proliferation and differentiation Chemicals, North York, ON, Canada) was reconstituted in dimethyl sulfoxide at 10 mg/ml, and then diluted in saline containing 5% PEG-400 Inhibition of phosphorylation of (Sigma) and 5% Tween-80 (Sigma). OVX mice were administered Akt, MAPK,mTOR and ERα rapamycin (1 mg/kg body weight per day, intraperitoneally) for 3 weeks. All procedures and protocols were approved by the Institutional Animal Care and Use Committee at the National Institute for Basic Biology.

Inhibition of excess cell proliferation and tumorigenesis Histology and immunohistochemistry Hematoxylin and eosin staining was performed by a standard procedure. Pten expressing cell PAS and Alcian blue staining was performed with PAS Staining Kit (Muto Chemical, Tokyo Japan) and acidic Alcian blue. For immunohistochemistry, Figure 8. A possible scheme depicting Pten expression and its paraformaldehyde-fixed, paraffin-embedded sections were incubated with regulation of homeostasis for mouse vagina in the absence (a) and the following primary antibodies: ERα (H184, 1:200), p63 (4A4, 1:200; Santa in the presence of estrogen (b). Cruz, Santa Cruz, CA, USA), CK8 (1:50; Progen, Heidelberg, Germany), CK1 (1:500), CK14 (1:1000; Covance, Emeryville, CA, USA), Pten (138G6, 1:100), phospho-Akt (D9E, 1:100), phospho-mitogen-activated protein kinase extracellular signal-regulated kinase 1/2 has been implicated for kinase (D26, 1:160) and phospho-extracellular signal-regulated kinase 1/2 (D13.14.4E, 1:400; Cell Signaling, Danvers, MA, USA). The sections were mucin production in the respiratory tract, intestine and 24–26 stained with the Vectastain ABC Kit (Vector Laboratories, Burlingame, CA, prostate. This supports a model wherein aberrant mucin USA). Immunofluorescence analysis was performed with Alexa Fluor production results from misactivation of the MAPK pathway. protein-conjugated secondary antibodies (Life Technology, Carlsbad, Among vaginal carcinoma, approximately 80% are squamous CA, USA). For BrdU immunostaining, mice were injected with BrdU cell carcinomas, with 13% being adenocarcinoma.39 There (Sigma) at 100 mg/kg body weight. At 1 h after the injection, tissues were are several types of vaginal adenocarcinoma reported, including collected. BrdU-incorporated cells were detected with anti-BrdU antibody clear-cell, serous and mucinous types.40–42 The histogenesis of (1:20; Roche, Mannheim, Germany). Terminal deoxynucleotidyl transferase dUTP nick-end labeling assay for the detection of apoptotic cells was primary vaginal adenocarcinoma, particularly which is not fi associated with in utero diethylstilboestrol exposure,43 is not well performed with the in situ apoptosis detection kit (Roche). More than ve animals were used for each histological analysis. Error bars represent the understood. The current results suggest a contribution of the standard error. Pten/Akt/mTOR signaling pathway to the aforementioned histogenesis; in particular, the vaginal mucinous adenocarcinoma might be involved in MAPK signaling cascade as well. In addition, Quantitative reverse transcriptase-polymerase chain reaction this study also showed that vaginal phenotypes in the Pten CKO Total RNA (2.5 μg), isolated with an RNeasy Kit (Qiagen, Velno, The mice depend on the estrogen levels. It is currently thought that Netherlands) from each group, was used in reverse transcription–- excess estrogen stimulation has a role in the etiology of polymerase chain reactions carried out with SuperScript III reverse hyperplasia and/or tumors in female reproductive organs. How- transcriptase and SYBR Green Master Mix (Life Technologies) according ever, even in the absence of estrogen, Pten mutation induced a to manufacturer’s instructions. Polymerase chain reaction conditions deleterious effect on mouse vagina. Taken together, the results and sequences of primer sets are given in the previous report.8 Relative presented herein have important implications for treatment of RNA equivalents for each sample were obtained by standardization lesions in the reproductive organ in women. of ribosomal protein L8 levels. More than three pools of samples per Pten has a critical role in epithelial cell homeostasis and acts as group were run in triplicate to determine sample reproducibility. Error a tumor suppressor in various tissues. Here we took advantage of bars represent the standard error, with all values represented as fold change compared with the control treatment group normalized to an unique features of the mouse vagina as a model for analyzing the average of 1.0. role of Pten in epithelial cell homeostasis. In the absence of estrogen, Pten is expressed in the suprabasal cells, where it inhibits ectopic cell proliferation (Figure 8a). In the presence of Protein preparation and immunoblotting estrogen, Pten functions in the basal cells, where it may prevent excessive cell proliferation (Figure 8b). Thus, Pten is indispensable Isolated mouse vaginae were homogenized in a buffer (20 mM HEPES, 2 mM EDTA, 2 mM EGTA, 100 mM β-glycerophosphate, 250 mM NaCl, 1% Triton for homeostasis in the vaginal epithelium, where its function X-100, protease inhibitor cocktail (Roche) and phosphatase inhibitor depends on estrogen levels. This provides new insight into the cocktail (Roche), pH 7.5) and spun at 15 000 r.p.m. for 10 min. The role of Pten in tissue homeostasis. supernatant was collected and the protein content was determined using the Bradford Assay (BioRad, Hercules, CA, USA). Proteins (10 μg) were electrophoresed on sodium dodecyl sulfate–polyacrylamide gels and MATERIALS AND METHODS transferred onto a nitrocellulose membrane. The membranes were Mouse and chemical treatment incubated with the following primary antibodies at a dilution of 1:1000 overnight at 4 °C: ERα (H184; Santa Cruz), phospho-ERαs (2514 and 2515), C57BL/6J (CLEA, Tokyo, Japan), K5-Cre44 and Pten-floxed mice 45 were – mTOR (7C10), phospho-mTOR (D9C2), Akt (C67E7), phospho-Akt (C31E5E maintained under 12 h light/12 h dark at 23 25 °C and fed laboratory chow and D9E; Cell Signaling). Signals were detected with the ECL Prime Kit (GE (CA-1, CLEA) and tap water ad libitum. K5-Cre line (B6 background) and fl flox/flox Healthcare, Buckinghamshire, UK). Digital images were taken and Pten- oxed line (Pten ; 129 background) were crossed to produce densitometry analysis was performed using the NIH Image J software male K5-Cre/+;Ptenflox/+ mice. Pten CKO mice (K5-Cre/+;Ptenflox/flox) were flox/+ flox/flox (NIH, Bethesda, MD, USA), downloaded from http://rsbweb.nih.gov/ij/. obtained by crossing K5-Cre/+;Pten male mice and Pten female More than three pools of samples per group were used. mice. Cre-negative and Ptenflox/flox siblings were used as control. In such conditions, no prominent phenotypic variation was observed within each experimental group. In most experiments, mice were OVX at 6 weeks of age and killed at 8 weeks of age. For examining effects of estrogen, a Statistical analysis single injection of 0.1 μg E2 (Sigma, St Louis, MO, USA) was given to OVX For BrdU labeling and apoptotic indices, and gene expression analyses, mice for 3 days and killed 24 h after the last injection. This timing is statistical analyses were performed using Student's t-test or Welch's t-test sufficient to induce stratified and fully keratinized epithelium in the mouse followed by F-test; differences with Po0.05 were considered significant.

Oncogene (2015) 1035 – 1043 © 2015 Macmillan Publishers Limited Pten function in vagina S Miyagawa et al 1043 CONFLICT OF INTEREST 21 Bodine SC, Stitt TN, Gonzalez M, Kline WO, Stover GL, Bauerlein R, et al. Akt/mTOR The authors declare no conflict of interest. pathway is a crucial regulator of skeletal muscle hypertrophy and can prevent muscle atrophy in vivo. Nat Cell Biol 2001; 3: 1014–1019. 22 Rommel C, Bodine SC, Clarke BA, Rossman R, Nunez L, Stitt TN, et al. Mediation of ACKNOWLEDGEMENTS IGF-1-induced skeletal myotube hypertrophy by PI(3)K/Akt/mTOR and PI(3)K/Akt/ GSK3 pathways. Nat Cell Biol 2001; 3:1009–1013. We thank Dr J Takeda, Osaka University, Suita, Japan, for providing K5-Cre mice. We 23 Pende M, Kozma SC, Jaquet M, Oorschot V, Burcelin R, Le Marchand-Brustel Y, also thank Dr Y Mikami, Kyoto University, Kyoto, Japan, and Dr Bruce Blumberg, et al. Hypoinsulinaemia, glucose intolerance and diminished beta-cell size in University of California, Irvine, CA, USA, for their support. This study was supported by S6K1-deficient mice. Nature 2000; 408:994–997. fi Grant-in-Aid for Young Scientists B, Grant-in-Aid for Scienti c Research B from the 24 Kurita T, Medina RT, Mills AA, Cunha GR. Role of p63 and basal cells in the Ministry of Education, Culture, Sports, Science and Technology of Japan, and Health prostate. Development 2004; 131: 4955–4964. Sciences Research Grant from the Ministry of Health, Labour and Welfare, Japan. 25 Lee HW, Ahn DH, Crawley SC, Li JD, Gum JRJr., Basbaum CB, et al. Phorbol 12- myristate 13-acetate up-regulates the transcription of MUC2 intestinal mucin via Ras. ERK, and NF-kappa B. J Biol Chem 2002; 277: 32624–32631. REFERENCES 26 Li JD, Feng W, Gallup M, Kim JH, Gum J, Kim Y, et al. Activation of NF-kappaB via a 1 Suzuki A, Itami S, Ohishi M, Hamada K, Inoue T, Komazawa N, et al. Keratinocyte- Src-dependent Ras-MAPK-pp90rsk pathway is required for Pseudomonas aerugi- specific Pten deficiency results in epidermal hyperplasia, accelerated hair follicle nosa-induced mucin overproduction in epithelial cells. Proc Natl Acad Sci USA morphogenesis and tumor formation. Cancer Res 2003; 63:674–681. 1998; 95:5718–5723. 2 Blanco V, Keochgerian V. Cowden's syndrome. Case report, with reference to an 27 Vilgelm A, Lian Z, Wang H, Beauparlant SL, Klein-Szanto A, Ellenson LH, et al. Akt- affected family. Med Oral Patol Oral Cir Bucal 2006; 11:E12–E16. mediated phosphorylation and activation of estrogen receptor alpha is required 3 Leao JC, Batista V, Guimaraes PB, Belo J, Porter SR. Cowden's syndrome affecting for endometrial neoplastic transformation in Pten+/ − mice. Cancer Res 2006; 66: the mouth, gastrointestinal, and central nervous system: a case report and review 3375–3380. of the literature. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2005; 99: 28 Lannigan DA. Estrogen receptor phosphorylation. Steroids 2003; 68:1–9. 569–572. 29 Campbell RA, Bhat-Nakshatri P, Patel NM, Constantinidou D, Ali S, Nakshatri H. 4 Yoo LI, Liu DW, Le Vu S, Bronson RT, Wu H, Yuan J. Pten deficiency activates Phosphatidylinositol 3-kinase/AKT-mediated activation of estrogen receptor distinct downstream signaling pathways in a tissue-specific manner. Cancer Res alpha: a new model for anti-estrogen resistance. J Biol Chem 2001; 276: 2006; 66: 1929–1939. 9817–9824. 5 Buchanan DL, Kurita T, Taylor JA, Lubahn DB, Cunha GR, Cooke PS. Role of stromal 30 Squarize CH, Castilho RM, Gutkind JS. Chemoprevention and treatment of and epithelial estrogen receptors in vaginal epithelial proliferation, stratification, experimental Cowden's disease by mTOR inhibition with rapamycin. Cancer Res and cornification. Endocrinology 1998; 139: 4345–4352. 2008; 68:7066–7072. 6 Cooke PS, Buchanan DL, Young P, Setiawan T, Brody J, Korach KS, et al. Stromal 31 Klotz DM, Hewitt SC, Ciana P, Raviscioni M, Lindzey JK, Foley J et al. Requirement estrogen receptors mediate mitogenic effects of estradiol on uterine epithelium. of estrogen receptor-alpha in insulin-like growth factor-1 (IGF-1)-induced uterine Proc Natl Acad Sci USA 1997; 94: 6535–6540. responses and in vivo evidence for IGF-1/estrogen receptor cross-talk. J Biol Chem 7 Nelson KG, Sakai Y, Eitzman B, Steed T, McLachlan J. Exposure to diethylstilbestrol 2002; 277: 8531–8537. during a critical developmental period of the mouse reproductive tract leads to 32 Joshi A, Wang H, Jiang G, Douglas W, Chan JS, Korach KS, et al. Endometrial persistent induction of two estrogen-regulated genes. Cell Growth Differ 1994; 5: tumorigenesis in Pten(+/ − ) mice is independent of coexistence of estrogen and 595–606. estrogen receptor alpha. Am J Pathol 2012; 180: 2536–2547. 8 Miyagawa S, Katsu Y, Watanabe H, Iguchi T. Estrogen-independent activation of 33 Brake T, Lambert PF. Estrogen contributes to the onset, persistence, and malig- erbBs signaling and estrogen receptor alpha in the mouse vagina exposed nant progression of cervical cancer in a human papillomavirus-transgenic neonatally to diethylstilbestrol. Oncogene 2004; 23:340–349. mouse model. Proc Natl Acad Sci USA 2005; 102: 2490–2495. 9 Miyagawa S, Suzuki A, Katsu Y, Kobayashi M, Goto M, Handa H, et al. Persistent 34 Ozawa S, Iguchi T, Sawada K, Ohta Y, Takasugi N, Bern HA. Postnatal vaginal gene expression in mouse vagina exposed neonatally to diethylstilbestrol. J Mol nodules induced by prenatal diethylstilbestrol treatment correlate with later Endocrinol 2004; 32: 663–677. development of ovary-independent vaginal and uterine changes in mice. Cancer 10 Ghahary A, Chakrabarti S, Murphy LJ. Localization of the sites of synthesis and action Lett 1991; 58:167–175. of insulin-like growth factor-I in the rat uterus. Mol Endocrinol 1990; 4:191–195. 35 Segrelles C, Moral M, Lara MF, Ruiz S, Santos M, Leis H, et al. Molecular deter- 11 Ignar-Trowbridge DM, Nelson KG, Bidwell MC, Curtis SW, Washburn TF, McLachlan minants of Akt-induced keratinocyte transformation. Oncogene 2006; 25: JA, et al. Coupling of dual signaling pathways: epidermal growth factor action 1174–1185. involves the estrogen receptor. Proc Natl Acad Sci USA 1992; 89: 4658–4662. 36 Shaw RJ, Ras Cantley LC. PI(3)K and mTOR signalling controls tumour cell growth. 12 Nelson KG, Takahashi T, Bossert NL, Walmer DK, McLachlan JA. Epidermal growth Nature 2006; 441:424–430. factor replaces estrogen in the stimulation of female genital-tract growth and 37 Majumder PK, Febbo PG, Bikoff R, Berger R, Xue Q, McMahon LM, et al. mTOR differentiation. Proc Natl Acad Sci USA 1991; 88:21–25. inhibition reverses Akt-dependent prostate intraepithelial neoplasia through 13 Daikoku T, Hirota Y, Tranguch S, Joshi AR, DeMayo FJ, Lydon JP, et al. Conditional regulation of apoptotic and HIF-1-dependent pathways. Nat Med 2004; 10: loss of uterine Pten unfailingly and rapidly induces endometrial cancer in mice. 594–601. Cancer Res 2008; 68: 5619–5627. 38 Neshat MS, Mellinghoff IK, Tran C, Stiles B, Thomas G, Petersen R, et al. Enhanced 14 Joshi A, Ellenson LH. Adenovirus mediated homozygous endometrial epithelial sensitivity of PTEN-deficient tumors to inhibition of FRAP/mTOR. Proc Natl Acad Pten deletion results in aggressive endometrial carcinoma. Exp Cell Res 2011; 317: Sci USA 2001; 98: 10314–10319. 1580–1589. 39 Platz CE, Benda JA. Female genital tract cancer. Cancer 1995; 75: 270–294. 15 Kim TH, Franco HL, Jung SY, Qin J, Broaddus RR, Lydon JP, et al. The synergistic 40 Mudhar HS, Smith JH, Tidy J. Primary vaginal adenocarcinoma of intestinal type effect of Mig-6 and Pten ablation on endometrial cancer development and pro- arising from an adenoma: case report and review of the literature. Int J Gynecol gression. Oncogene 2010; 29: 3770–3780. Pathol 2001; 20:204–209. 16 Memarzadeh S, Zong Y, Janzen DM, Goldstein AS, Cheng D, Kurita T, et al. Cell- 41 McCluggage WG, Price JH, Dobbs SP. Primary adenocarcinoma of the vagina autonomous activation of the PI3-kinase pathway initiates endometrial cancer arising in endocervicosis. Int J Gynecol Pathol 2001; 20:399–402. from adult uterine epithelium. Proc Natl Acad Sci USA 2010; 107: 17298–17303. 42 Ebrahim S, Daponte A, Smith TH, Tiltman A, Guidozzi F. Primary mucinous ade- 17 Moll R, Franke WW, Schiller DL, Geiger B, Krepler R. The catalog of human cyto- nocarcinoma of the vagina. Gynecol Oncol 2001; 80:89–92. keratins: patterns of expression in normal epithelia, tumors and cultured cells. Cell 43 Herbst AL, Ulfelder H, Poskanzer DC. Adenocarcinoma of the vagina. Association 1982; 31:11–24. of maternal stilbestrol therapy with tumor appearance in young women. N Engl J 18 Gimenez-Conti IB, Lynch M, Roop D, Bhowmik S, Majeski P, Conti CJ. Expression of Med 1971; 284: 878–881. keratins in mouse vaginal epithelium. Differentiation 1994; 56: 143–151. 44 Tarutani M, Itami S, Okabe M, Ikawa M, Tezuka T, Yoshikawa K, et al. Tissue-specific 19 Yang A, Schweitzer R, Sun D, Kaghad M, Walker N, Bronson RT, et al. P63 is knockout of the mouse Pig—a gene reveals important roles for GPI-anchored essential for regenerative proliferation in limb, craniofacial and epithelial devel- proteins in skin development. Proc Natl Acad Sci USA 1997; 94: 7400–7405. opment. Nature 1999; 398:714–718. 45 Groszer M, Erickson R, Scripture-Adams DD, Lesche R, Trumpp A, Zack JA, et al. 20 Mills AA, Zheng B, Wang XJ, Vogel H, Roop DR, Bradley A. P63 is a p53 homologue Negative regulation of neural stem/progenitor cell proliferation by the Pten required for limb and epidermal morphogenesis. Nature 1999; 398: 708–713. tumor suppressor gene in vivo. Science 2001; 294: 2186–2189.

Supplementary Information accompanies this paper on the Oncogene website (http://www.nature.com/onc)