Abnormal Epithelial Cell Polarity and Ectopic Epidermal Growth Factor Receptor (EGFR) Expression Induced in Emx2 KO Embryonic Gonads

The gonadal primordium first emerges as a thickening of the embryonic coelomic epithelium, whichhasbeenthoughttomigratemediodorsallytoformtheprimitivegonad.However,theearly gonadal development remains poorly understood. Mice lacking the paired-like homeobox gene Emx2 display gonadal dysgenesis. Interestingly, the knockout (KO) embryonic gonads develop an unusual surface accompanied by aberrant tight junction assembly. Morphological and in vitro cell fate mapping studies showed an apparent decrease in the number of the gonadal epithelial cells migrated to mesenchymal compartment in the KO, suggesting that polarized cell division and subsequent cell migration are affected. Microarray analyses of the epithelial cells revealed signif-icant up-regulation of Egfr in the KO, indicating that Emx2 suppresses Egfr gene expression. This genetic correlation between the two genes was reproduced with cultured M15 cells derived from mesonephric epithelial cells. Epidermal growth factor receptor signaling was recently shown to regulate tight junction assembly through sarcoma viral oncogene homolog tyrosine phosphorylation. We show through Emx2 KO analyses that sarcoma viral oncogene homolog tyrosine phosphorylation, epidermal growth factor receptor tyrosine phosphorylation, and Egfr expression are up-regulated in the embryonic gonad. Our results strongly suggest that Emx2 is required for regulation of tight junction assembly and allowing migration of the gonadal epithelia to the mesenchyme, which are possibly mediated by suppression of Egfr expression. ( Endocrinology 151: 5893–5904, 2010)

The gonadal primordium first emerges as a thickening of the embryonic coelomic epithelium, which has been thought to migrate mediodorsally to form the primitive gonad. However, the early gonadal development remains poorly understood. Mice lacking the paired-like homeobox gene Emx2 display gonadal dysgenesis. Interestingly, the knockout (KO) embryonic gonads develop an unusual surface accompanied by aberrant tight junction assembly. Morphological and in vitro cell fate mapping studies showed an apparent decrease in the number of the gonadal epithelial cells migrated to mesenchymal compartment in the KO, suggesting that polarized cell division and subsequent cell migration are affected. Microarray analyses of the epithelial cells revealed significant up-regulation of Egfr in the KO, indicating that Emx2 suppresses Egfr gene expression. This genetic correlation between the two genes was reproduced with cultured M15 cells derived from mesonephric epithelial cells. Epidermal growth factor receptor signaling was recently shown to regulate tight junction assembly through sarcoma viral oncogene homolog tyrosine phosphorylation. We show through Emx2 KO analyses that sarcoma viral oncogene homolog tyrosine phosphorylation, epidermal growth factor receptor tyrosine phosphorylation, and Egfr expression are up-regulated in the embryonic gonad. Our results strongly suggest that Emx2 is required for regulation of tight junction assembly and allowing migration of the gonadal epithelia to the mesenchyme, which are possibly mediated by suppression of Egfr expression. (Endocrinology 151: 5893-5904, 2010) T he gonadal primordium first appears as a thickening of the coelomic epithelium at the lateral sides of the mesentery. Histological observations suggested that these epithelial cells migrate mediodorsally to form the primitive urogenital ridge (1)(2)(3), which subsequently separates into the future gonad and mesonephros. Sexually dimorphic events occur thereafter in the genital ridge in response to transient expression of sex-determining region on the Y chromosome (Sry). A number of studies has attempted to elucidate the mechanisms underlying gonad development and sex differentiation. Disruption of some genes in mice and humans leads to hypoplastic gonads or complete dysgenesis, whereas disruption of other genes leads to sex reversal (4 -6). Nevertheless, the early phase of gonad development is still poorly understood.
The paired-like homeobox gene Emx2 is a mouse homologue of the Drosophila empty spiracles gene (7). Mouse knockout (KO) studies demonstrated that Emx2 is implicated in the development of multiple tissues, including the kidney, gonad, reproductive tracts, and central nervous system. A detailed study of the kidney defect showed that ureteric bud extension is disrupted in the metanephric mesenchyme of Emx2 KO mice, leading to the disappearance of the kidney (8). In addition, proliferation of neuronal progenitors and area specification of the neocortex are disrupted in Emx2 KO animals (9 -13). Some of the observed abnormalities in these animals appeared to be closely related to defects in the spatial expression of growth factors (14 -17). Although these studies revealed important insights into how Emx2 functions during kidney and central nervous system development, its role in gonad development is still unclear.
Epithelial cells are polarized by establishing functionally specialized apical, lateral, and basal surfaces and adhere tightly one another through tight and adherens junctions at their lateral interfaces. These junctions are comprised of transmembrane proteins, such as junctional adhesion molecules, cadherins, occludin, and claudin, and intracellular membrane-associated proteins, such as ␤-catenin and zonula occludens (ZO) (18,19). Mammalian homologues of the Caenorhabditis elegans partitioning-defective proteins (PARs) PAR3 and PAR6 position tight junctions by forming a complex with atypical protein kinase C (aPKC). Multiple extracellular signals regulate the formation and localization of the PAR3/PAR6/aPKC complex via phosphorylation of the components (20,21).
ErbB (erythroblastic leukemia viral oncogene homolog) receptors comprise a family of four structurally related tyrosine kinase receptors (22) that are activated by a variety of ligand molecules, including epidermal growth factor (EGF). Upon ligand binding, the receptor is dimerized and the kinase activity triggers numerous down-stream signaling pathways (23). One family member, EGF receptor (EGFR), is expressed in the epithelial cells of a variety of tissues, where it plays fundamental roles in tissue development through regulating cell proliferation, cellular polarity formation, and epithelial cell migration (24). A number of studies revealed the presence of phospholipase C-␥, PKC-mediated cascades, mitogen-activated protein cascades, and small GTPase downstream of EGFR (23). Importantly, a recent study demonstrated that EGFR is implicated in the regulation of tight junction assembly via tyrosine phosphorylation of sarcoma viral oncogene homolog (c-Src)/Yamaguchi sarcoma viral oncogene homolog 1 (c-Yes), which subsequently phosphorylates PAR3 to regulate PAR3/PAR6/aPKC complex formation (25).
Here, we examine early stages of gonad development in Emx2 KO embryos and show that tight junction assembly and migration of the epithelial cells of the gonad are significantly affected. Interestingly, microarray analysis of the epithelial cells of the embryonic gonad indicates that Egfr is dramatically up-regulated in Emx2 KO mice. This ectopic Egfr expression is accompanied by aberrant c-Src tyrosine phosphorylation. Our data strongly suggested that Emx2 is required for tight junction assembly and migration of epithelial cells at the early stage of gonadal development possibly through suppression of Egfr expression.

Cell proliferation and apoptosis assays
Pregnant females received an ip injection of BrdU (Sigma) (50 mg/kg body weight) at E10.0, E10.5, and E11.0 (39) and were killed 2 h after injection. Paraffin sections of the embryos were double immunostained for BrdU and GATA4. In brief, after the sections were boiled in 10 mM citrate (pH 6.0) for 20 min (40), they were incubated with the mouse anti-BrdU antibody, and thereafter with the Cy3-labeled antimouse antibody. Subsequently, the sections were incubated with goat anti-GATA4 antibody and with Alexa Fluor 488-labeled antigoat antibody. Nuclei in the sections were stained with propidium iodide (PI) (Molecular Probes). The number of BrdU-immunoreactive gonadal epithelial cells was counted in more than 10 sections for every gonad. Apoptosis in E11.0 and E12.0 embryonic gonads was assayed using the ApopTag Plus Peroxidase kit (CHEMI-CON, Temecula, CA). After apoptotic cells were detected with rhodamine-labeled anti-digoxigenin antibody, the sections were stained by goat anti-GATA4 and Alexa Fluor 488-labeled antigoat antibodies.

Cell fate mapping with organ culture
After the abdominal tissues of E10.25 embryos were removed, the coelomic epithelial cells were labeled with 20 mM 5-(and-6)-carboxy-2Ј,7Ј-dichlorofuluorescein diacetate, succinimidyl ester (CCFSE) (Molecular Probes) in DMEM (Invitrogen) for 1 h. Subsequently, they were cultured for 5 or 24 h in DMEM containing 10% fetal bovine serum and antibiotics under a humidified atmosphere of 5% CO 2 in air at 37 C. After fixation in 4% paraformaldehyde for 5 min, the embryos were frozen sectioned and stained with Ad4BP/SF-1 or laminin antibody. The number of CCFSE-positive cells that migrated through the laminin layer was counted in more than 10 sections for every gonad.

Microarray and quantitative RT-PCR analyses
Microarray analyses were performed essentially as described (44,45). E10.5 wild-type and Emx2 KO embryos were frozen in OCT compound (Sakura Finetechnical, Tokyo, Japan) without fixation. They were sectioned (30 m), stained with hematoxylin, and air dried. Some were used for GATA4 immunostaining to locate gonadal primordia. The epithelial cells of the gonadal primordia were obtained using a Laser Microdissection System (Leica, Wetzlar, Germany). The specimens prepared from three individuals were combined into one group. Total RNA was prepared from three groups, and 15 ng total RNA was subjected to two-cycle amplification and biotin labeling using MessageAmp II aRNA Amplification and MessageAmp II-Biotin Enhanced kit (Ambion, Austin, TX), respectively. The labeled aRNA was fragmented and hybridized to GeneChip Mouse Genome 430 2.0 array according to the manufacturer's instructions (Affymetrix, Santa Clara, CA). Signals were scanned and scaled using Affymetrix GCOS 1.1 software. The scaled values were then analyzed by GeneSpring software (Silicon Genetics, Redwood City, CA). Pairwise comparison analysis was performed with Affymetrix GCOS 1.2 to identify differentially expressed transcripts. Each sample (n ϭ 3) was compared with each reference samples (n ϭ 3), resulting in nine pairwise comparisons. This approach, which is based on Mann-Whitney pairwise comparison test, allows ranking of differentially expressed genes as well as calculation of significance (P Ͻ 0.05) of each identified change. The microarray data have been deposited in the Gene Expression Omnibus of the National Center for Biotechnology Information (accession no. GSE10216; http://www.ncbi.nlm.nih.gov/geo/). Quantitative RT-PCR with TaqMan probes for mouse Egfr (Mm00433021_m1) and Gapdh (TaqMan Rodent glyceraldehyde-3-phosphate dehydrogenase Control Regents; Applied Biosystems, Foster City, CA) was performed using QuantitectProbe RT-PCR kit (QIAGEN, Valencia, CA) and PM9000 cermal cycler (Applied Biosystems). Total RNAs prepared from M15 and M15-Emx2 cells were used for quantitative RT-PCR using Power SYBR Green PCR Master mix (Applied Biosystems). PCR primers for mouse Egfr are mEgfr-S (5Ј-CA-CAGCGCTACCTTGTTATCCA-3Ј) and mEgfr-AS (5Ј-CCTC-CATGTCCTCTTCATCCA-3Ј). Primers for ␤-actin were as previously described (46).

Abnormal surface of Emx2 KO embryonic gonad
Although the gonads of Emx2 KO embryos disappear by E12.5 (8), the mechanism by which this defect arises was not analyzed. As shown in Fig. 1A, the urogenital primordium is observed as a thickening of eplithelia at E10.5 in both wild-type and Emx2 KO embryos. By E11.5, the developing urogenital primordium separates into a sexually indifferent gonad and mesonephros and thereafter develops into either the testis or ovary. The gonads of Emx2 KO embryos are underdeveloped at E11.5 and largely disappear by E12.5 in both sexes. We noted that during the degeneration process, the gonadal surface of Emx2 KO embryos appears to be abnormal at E11.5. Thus, the gonadal surface was examined by SEM (Fig. 1B). The gonads of wild-type embryos exhibit smooth surfaces, whereas those of Emx2 KO embryos show irregularly protruding cellular clusters.
To characterize the Emx2 KO gonad, we examined the expression of various gonadal marker genes. The expressions of Ad4BP/SF-1 required for the gonadal and adrenal development (47,48) and aristaless-related homeobox required for the development of testicular Leydig cells (32) disappeared from the Emx2 KO gonads by E11.5, whereas the expression of Ad4BP/SF-1 in the adrenal primordium was unaffected (Fig. 1C). This loss of Ad4BP/SF-1 expression in the gonads but not the adrenal glands of Emx2 KO mice is consistent with the observation that although Ad4BP/SF-1 KO mice failed to develop both the gonads and adrenal glands (49), Emx2 KO mice failed to develop the adrenal gland. The expression of GATA4, WT1, and CBX2/M33, all of which are expressed in the developing gonad of wild-type and involved in the gonadal development, was unaffected. Interestingly, the epithelial expression of Lhx9 required for the gonadal development (37) was expanded laterally in Emx2 KO.

Abnormal cell proliferation and apoptotic cell death induced in Emx2 KO gonads
The structural abnormalities in the gonadal surface of Emx2 KO strongly suggested that Emx2 is critical for the development of the epithelial cells of the gonad. Therefore, the expression of EMX2 in the developing gonad was examined. Immunohistochemical studies demonstrated that EMX2 is expressed in the gonadal epithelial and adjacent mescenchymal cells, nephric duct, and mesonephric tubule at E10.5 and expressed in all somatic cells in the gonad but not in the mesonephrose except tubular struc-ture at E11.5 ( Fig. 2A). Although the expression was similar between the two sexes before gonadal sex differentiation, the expression became different between the testis and ovary at E12.5. The expression was downregulated in the testis except testicular tunica albuginea and underneath mesenchymal cells, whereas the expression was still evident in the whole ovary at E12.5. Because the testis initiates to synthesize testosterone as early as E12.5, the sexually dimorphic expression of Emx2 established by E12.5 seems to be independent of gonadal sex steroid. Next, we examined whether epithelial cell proliferation is affected in the Emx2 KO gonad. After BrdU was injected into pregnant females, embryos were collected at E10.0, E10.5, and E11.0 and their gonads stained with an antibody against BrdU (Fig. 2B, red). Because the gonads at the stage are structurally primitive, it is difficult to discriminate between the future gonadal and mesonephric areas only by the morphology. For the gonad at the stages, Ad4BP/SF-1 and GATA4 are known to be potential gonadal markers. However, the expression of Ad4BP/ SF-1 was affected significantly in the Emx2 KO gonad, and thus GATA4 immunostaining was performed to evaluate the gonadal area (Fig. 2B, green) (35). In addition, considering that GATA4 is expressed in the mesentery, the gonadal area was eventually determined as GATA4 immunoreactive but not the mesentery cells. The number of BrdU-positive proliferating epithelial cells was similar between wild type and Emx2 KO at E10.0 in both sexes (Fig. 2, B and C). By E10.5, the number of BrdU-positive cells in Emx2 KO gonads was increased by approximately 1.3fold in males and 1.5-fold in females compared with wild type. This increase in epithelial cell proliferation was not observed at E11.5. Apoptosis was examined using terminal deoxynucleotidyltransferase-mediated 2Ј-deoxyuridine 5Ј-triphosphate  nick end labeling (TUNEL) labeling. Although the KO gonad showed obvious structural defects at E11.5, the number of TUNEL-positive apoptotic cells was not increased at E11.0. By E12.0, however, the number of TUNEL-positive cells was significantly increased (Fig. 2D).

Migration of gonadal epithelial cells affected in Emx2 KO
It has long been surmised that the coelomic epithelia at the both sides of the mesentery proliferate and migrate mediodorsally to give rise to the gonadal primordia. As described above, the BrdU incorporation study indicated that epithelial cell proliferation was transiently up-regulated in Emx2 KO gonads. However, if migration of the epithelial cells is affected, it is assumed that the two BrdUpositive daughter cells remain in the epithelial compartment, thus causing an apparent increase in the number of the BrdU-positive epithelial cells. We therefore examined whether migration of epithelial cells to the mesenchymal compartment was affected in the KO.
Gonadal epithelial cells are thought to pass through the basement membrane during their migration into the mesenchymal compartment. The gonads of wild-type and Emx2 KO embryos were sectioned and stained with laminin antibody to visualize the basement membrane (Fig.  3A, green), whereas nuclei were stained with PI (Fig. 3A,  red). As expected, cells were frequently seen in the basement membrane in wild-type embryos (Fig. 3A, arrowheads) but approximately 6-fold less frequently in the KO gonads (Fig. 3B). Epithelial cells can undergo polarized cell division, with one daughter cell retaining epithelial cell features while the other cell losing them and migrating into the mesenchymal compartment through the basement membrane (21,50,51). The decreased number of cells localized to the basement membrane in the Emx2 KO gonad strongly suggests that polarized cell division and cell migration are affected in these mutants. Moreover, we noticed that the basement membrane of the KO gonad is not tightly lining the epithelial cells when compared with wild type. This unusual basement membrane might affect the epithelial cell migration.
Therefore, we examined this migration defect using organ culture. Embryonic trunk tissue containing the developing gonads was prepared at E10.25, and whole coelomic epithelial cells, including the gonadal epithelia, were labeled with a fluorescent dye, CCFSE, to chase the epithelial cells (Fig. 3C, green). After the labeled trunks were cultured for 5 h, the gonads were sectioned and immunostained with antilaminin antibody (Fig. 3C, red). Expectedly, the gonadal epithelia of wild-type embryos migrated through the laminin layer, whereas those of Emx2 KO scarcely migrated. The number of migrating cells seen in wild-type gonads was approximately 9-fold higher than in Emx2 KO (Fig. 3D). This migration defect possibly results in a substantial increase of epithelial cells and, at the same time, decrease of mesenchymal cells. Taken together, the observations above suggest that the loss of Emx2 blocks gonadal development by restricting epithelial cell migration. Based on the observation above, a question arose as to whether the migrating cells differentiate into gonadal somatic cells or not. Thus, the gonad sections were stained with anti-Ad4BP/SF-1 antibody after further incubation up to 24 h. Ad4BP/SF-1 was expressed in many CCFSEpositive migrating cells (Fig. 3C). These results demonstrate directly for the first time that the gonadal epithelial cells at around E10. 25 have the potential to migrate and to form gonadal mesenchyme after migration.

Abnormal tight junction assembly of the gonadal epithelia
Because epithelial cells are characterized by specialized cellular junctions, we examined whether tight junction assembly is affected in the gonadal epithelia of Emx2 KO. The intracellular component ZO-1 normally interacts with the homomeric tight junction protein occludin, and together, they localize to the apicolateral region in epithelial cells. aPKC forms a complex with PAR3 and PAR6 that localizes to tight junctions and regulates tight junction assembly (21,22). These marker proteins were normally localized to the apicolateral region of the gonadal epithelia in wild-type embryos at E10.5 ( Fig. 4) but were disrupted in Emx2 KO. Occludin distribution is expanded deeply to the lateral domain, whereas ZO-1 and aPKC are localized irregularly to whole aspects of the cellular surface. These data suggest that the Emx2 KO gonadal epithelia have lost their cell polarity.

Affected gene expression in Emx2 KO embryonic gonads
To assess the effect of Emx2 gene disruption, gonadal epithelia were microdissected from wild-type and Emx2 KO embryos at E10.5 to prepare total RNA. After biotinylation, these samples were used as probes for microarray analysis. Genes that satisfied a pairwise comparison test and displayed a more than 4-fold change in expression are summarized in Table 1. Because Emx2 KO gonad showed affected cell polarity, genes encoding the components of tight and adherens junctions, as well as Snail and Slug, which regulate transcription of the junction component genes (21,52), were expected to be affected in the KO gonad. However, none of these genes showed differential expression greater than 4-fold.

EGFR ectopically induced in Emx2 KO embryonic gonads
Our microarray data listed Egfr as the top-scored gene induced in the Emx2 KO gonad. Quantitative RT-PCR revealed an approximately 64-fold increase in Egfr expression in Emx2 KO tissue compared with wild type (Fig. 5A). Furthermore, immunohistochemistry demonstrated that EGFR is expressed at a low level in the gonadal epithelia of wild-type embryos, whereas the expression is high in both the epithelial and mesenchymal cells of the Emx2 KO gonad (Fig. 5B). In contrast, EGFR was not induced in the nephric duct and mesonephric tubules of KO.
The microarray study strongly suggested that Egfr gene is suppressed by EMX2. Therefore, we used M15 cells derived from mesonephric epithelial cells to test this hypothesis. Western blot analyses using an anti-EMX2 antibody showed that EMX2 is not expressed in M15 cells (Fig. 5D), whereas immunohistochemistry, Western blotting, and RT-PCR revealed that EGFR is expressed in the cells (Fig. 5, C-E). As expected, when Emx2(HA) was overexpressed, EGFR expression was reduced, as assayed by immunohistochemistry, Western blotting, and RT-PCR. Moreover, when the Emx2(HA)-overexpressing cells were treated with siRNA for Emx2, the expression of EGFR was significantly up-regulated. Such up-regulation was never observed with control siRNA.

Ectopic activation of c-Src in Emx2 KO gonad
Recently, activated EGFR was shown to phosphorylate a tyrosine residue of c-Src/c-Yes, and phosphorylated c-Src/c-Yes in turn phosphorylates a tyrosine residue of PAR3. Through this successive tyrosine phosphorylation, EGFR signaling is thought to fine-tune tight junction assembly (25,53). Therefore, we examined phosphorylation of EGFR and c-Src in the Emx2 KO gonads. Interestingly, tyrosine phosphorylation of c-Src is clearly activated in the Emx2 KO gonad in a pattern that overlaps with that of ectopically induced EGFR. A low level of c-Src phosphorylation is detected in the epithelial cells of wild-type gonads (Fig. 5F). It has been well established that EGFR is phosphorylated at the tyrosine 845 by activated c-Src and autophosphorylated at tyrosine 1068 upon ligand binding followed by dimerization (54). As indicated in Fig. 5G, phosphorylation of EGFR at the tyrosine 845 was clearly elevated in the Emx2 KO gonad, whereas that at the tyrosine 1068 was unlikely elevated.

Migration of coelomic epithelial cells to the mesenchymal compartment during development of the gonadal primordium
Based on histological observations, it has long been surmised that regions of the coelomic epithelia at the both sides of the mesentery proliferate and migrate mediodorsally to give rise to the gonadal primordium (1-3). However, this has not been addressed directly by cell fate map-ping. Here, we chemically labeled coelomic epithelial cells at E10.25 and found that the labeled epithelial cells migrated through the basement membrane. After migration, they began to express Ad4BP/SF-1, a marker gene for Sertoli and Leydig cell lineages (47), strongly suggesting that the migrated epithelial cells differentiated into these gonadal somatic cells.
Transition of the gonadal epithelial cells to mesenchyme was previously demonstrated with mouse embryonic gonad at around E11.5 (55). Interestingly, the migrated cells at E11.2-E11.4 differentiated into Sertoli and interstitial cells, whereas those at E11.5-E11.7 no longer developed into Sertoli cells. Unfortunately, our study with E10.25 embryos failed to culture the gonads until Sertoli and Leydig cells differentiate, and thus it remains unsolved whether the epithelial cells at the earlier stage develop into Sertoli and Leydig cells. A new culture system, which enables to culture the early gonadal primordium for a longer period, is required to resolve the issue.

Emx2 implicated in the maintenance of epithelial polarity and the epithelial-to-mesenchyal transition
Many tissues are known to undergo epithelial-to-mesenchymal transition and/or mesenchymal-to-epithelial transition during the development. These transitions are closely correlated with the assembly and disassembly of tight junctions. In the present study, we have demonstrated that ectopic tight junctions are formed in Emx2 KO gonadal epithelia, and thus it is assumed that the aberrant persistence of tight junctions inhibits the epithelial- to-mesenchymal transition during early gonadal development. As described below, this defect may correlate with ectopic EGFR expression in the developing Emx2 KO gonad.
In addition to the gonad, Emx2 is expressed in the epithelial cells of the ureteric bud, Wolffian duct, Mü llerian duct, and mesonephric tubule (56). Interestingly, extension and branching of the ureteric bud are affected in Emx2 KO. Similarly, Wolffian duct and mesonephric tubules degenerate after the structures initially develop, whereas Mü llerian duct fails to develop (8). Given that these structures are formed via a mesenchymal-to-epithelial transition, Emx2 is thought to be involved in both directions of transitions. However, because the expression of Egfr was unaffected in the tubular structures, Emx2 may regulate two transition processes through differential target gene expression.

Potential function of EGFR in developing gonadal epithelial cells
Microarray studies clearly demonstrated that the Egfr expression was up-regulated in the Emx2 KO gonads, and consequently, the question of why Egfr should be negatively regulated by Emx2 arises from our studies. Given that EGFR is not required throughout gonad development, one would expect the Egfr locus to be silenced, possibly through inactivation by a suppressive chromatin state. However, considering that Egfr expression is activated by Emx2 gene disruption, Egfr gene is not structurally silenced in the early gonad; instead, it is kept in a state that is ready to be activated.
Recently, Wang et al. (25) demonstrated that activated EGFR phosphorylates a tyrosine residue of c-Src/c-Yes, and subsequently, the phosphorylated c-Src/c-Yes phosphorylates a tyrosine residue of PAR3. Because tight junction assembly is delayed but not blocked with a phosphorylation-defective mutant of PAR3, EGFR signaling is thought to fine-tune tight junction assembly through successive tyrosine phosphorylation. Consistent with these observations, our data show that tyrosine-phosphorylated c-Src accumulates in the Emx2 KO embryonic gonad, in which Egfr is ectopically up-regulated.
EGFR is phosphorylated at multiple tyrosine residues. Upon ligand binding and dimerization, autophosphorylation of EGFR occurs at several tyrosine residues, including Y1068. Likewise, activated c-Src phosphorylates several tyrosine residues, including Y845 (54). The present study showed that phosphorylation at Y1086 was not elevated in the Emx2 KO gonad, whereas that at Y845 was elevated. This phosphorylation status strongly suggested that EGFR is phosphorylated by the activated c-Src but not EGFR itself. In fact, c-Src was phosphorylated at ty-rosine residue and thus activated in the KO gonad. Taken together, the overexpressed EGFR was activated possibly by c-Src in the KO gonad, although it remains to be clarified how Emx2 gene disruption causes c-Src activation.
Importantly, our immunohistochemical studies reproducibly detected EGFR and tyrosine-phosphorylated c-Src at low levels in the gonadal epithelial cells of wild type. Considering that the epithelial cells migrate to the mesenchymal compartment at an early stage of gonadal development, tight junctions of the epithelial cells should be disassembled sporadically before migration. EGFR signaling may act as the cue for this disassembly. Taken together, our data demonstrate that Emx2 guarantees proper gonadal development by regulating Egfr expression and thus modulating tight junction assembly.
Microarray analyses identified a number of genes whose expression was up-or down-regulated in the Emx2 KO gonadal epithelial cells. Among the genes that showed up-and down-regulated expressions in the Emx2 KO gonads, we highlighted the up-regulated expression of Egfr in the KO. However, this does not necessarily exclude the possible role of other genes described above in gonad development. Through its regulation of EGFR and possibly other genes, Emx2 appears to play a crucial role in regulating epithelial cellular junctions of the early developing gonad.