We report here that Msx1 (formerly Hox-7.1) is expressed at high levels in uterine epithelial cells of the non-pregnant adult. These cells undergo pronounced changes in morphology in response to embryo implantation and show a concomitant decrease in Msx1 levels. While Msx1 is restricted to the uterus in adulthood, we observe Msx1 expression throughout the entire perinatal Müllerian duct epithelium in the prospective uterus, cervix and vagina. Through analysis of tissue recombinants, the expression of Msx1 in the epithelium was shown to be dependent upon an interaction with the underlying mesenchyme of uterine origin. The capacity of uterine mesenchyme to support or induce Msx1 expression in Müllerian epithelium is correlated with mesenchymal expression of Wnt-5a. Whereas Msx1 expression in the epithelium results from interaction with uterine mesenchyme, Wnt-5a expression is an intrinsic property of the uterine mesenchyme and does not depend upon the epithelium. The observation that Msx1 is expressed in the adult uterine epithelium and that conversion of the presumptive vaginal epithelium to uterine epithelium can be elicited only during the first week of postnatal development when Msx1 expression is detected suggests that, in addition to regulating various aspects of uterine epithelial morphology and function (e.g. gestation), this homeobox-containing gene plays a role in maintaining the uterus in a morphogenic and developmentally responsive state prerequisite for its unique function.

Homeobox-containing genes represent attractive candidates for regulators of pattern formation during vertebrate development. In this regard, the expression patterns of Hox genes correlate temporarily and spatially with critical morphogenic events. Gene disruptions and gain-of-function mutations generated in mice have shown that improper expression of Hox genes causes developmental defects (Chisaka and Capecchi, 1991; Kessel and Gruss, 1991; Lufkin et al., 1991). In the developing limb, Msx1 (formerly Hox-7.1; new nomenclature, Scott, 1992) is expressed in the proliferative and undifferentiated cells of the ‘progress zone’, which is also the site where cells undergo positional fate assignment (Summerbell et al., 1973; Robert et al., 1989; Hill et al., 1989; Muneoka and Sassoon, 1992). Msx1 expression is dependent upon an interaction between the limb ectoderm and the underlying mesenchyme (Muneoka and Sassoon, 1992; Davidson et al., 1991; Robert et al., 1991). When the ectoderm is removed, limb outgrowth ceases and precocious cell differentiation initiates concomitant with a loss of Msx1 expression (Robert et al., 1991; Muneoka and Sassoon, 1992 and unpublished results). The pattern of Msx1 expression suggests that it is involved in the maintenance of a developmentally plastic or active state.

Although a role for Hox genes in the embryo has been actively examined, the role of such genes in adult tissues has received less attention. In most cases, Hox genes are not expressed at detectable levels in the adult (Duggal et al., 1987; Jackson et al., 1985; Redline et al., 1992). For most tissues, morphological patterning is well advanced by birth. In contrast, the female reproductive tract is rudimentary at birth with extensive morphological and functional differentiation occurring postnatally (see Forsberg, 1973; Brody and Cunha, 1989). Neonatal uterine and vaginal epithelia are developmentally plastic and can be induced by heterotypic mesenchymes to express entirely new programs of morphogenesis and cytodifferentiation (Cunha, 1976; Boutin et al., 1992). In the adult uterus and vagina, epithelial morphology and differentiation change dramatically in response to fluctuating hormone levels during the estrous cycle (Noyes, 1973). Even more striking are the decidual changes in the adult endometrium that are triggered by embryonic implantation (Finn, 1977). Clearly developmental plasticity is a hallmark of both the neonatal and adult uterus.

We report here that Msx1 is expressed at high levels in adult uterine epithelium and decreases during pregnancy following embryonic implantation. In addition, the expression of Msx1 in the epithelium is dependent upon maintenance of mesenchymal-epithelial interactions. Wnt genes, the vertebrate homologues of the Drosophila wingless family of growth factors (Gavin et al., 1990), have been implicated in patterning and, in particular, in establishing regional boundaries during development (McMahon et al., 1992). One member of the Wnt gene family, Wnt-1, is thought to be a secreted glycoprotein, which signals adjacent cells via an unidentified receptor(s). In the developing nervous system, Wnt-1 is involved in the regulation and maintenance of expression of the homeobox-containing gene, engrailed (McMahon et al., 1992). Given the potential role of Wnt genes in governing cell-mediated regulation of homeotic genes and the close overlapping domains of the patterns of Wnt-5a and Msx1, we investigated the pattern of expression of Wnt-5a during the establishment of the uterine, cervical and vaginal compartments (see Gavin et al., 1990; Parr et al., 1993). We observe that Wnt-5a is expressed in the mesenchyme adjacent to the Msx1-positive epithelium and is ultimately restricted to the uterine mesenchyme in the adult. As for Msx1, high expression of Wnt-5a is observed in the adult uterus and a similar down-regulation in Wnt-5a levels following embryonic implantation occurs. Using tissue recombinant assays, we observe that the competence of mesenchyme to induce/maintain Msx1 expression is associated with the mesenchymal expression of Wnt-5a. The observation that Msx1 is expressed and regulated as a function of pregnancy in the adult uterine epithelium suggests that, in addition to regulating various aspects of uterine epithelial morphology and function, this homeobox-containing gene plays a role in maintaining the adult uterus in a morphogenic and developmentally responsive state. We conclude that Msx1 expression in the epithelium of the uterus is a function of its position (adjacent to uterine mesenchyme) whereas Wnt-5a expression in the mesenchyme is an inherent property of the uterus once the uterine compartment is established. It remains to be determined how Wnt-5a is regulated and whether its expression directly underlies the regulation of Msx1.

Preparation of tissues

Uterus (Ut) and vaginal (Va) tissues were isolated from CD-1 (Charles River) and Balb/c (Cancer Research Center, UC Berkeley) mice. Pregnant mice were obtained following timed breedings of CD-1 mice and the morning of the vaginal plug was counted as 0.5 days post coitum (p.c.). At least three animals were examined for each stage reported. Tissues were fixed, dehydrated, embedded in paraffin and sectioned for subsequent in situ hybridization following procedures previously described in detail (Sassoon et al., 1988; Sassoon and Rosenthal, 1993).

Uterine/vaginal recombinants

Uterine/vaginal tissue recombinants were prepared as previously described (Boutin et al., 1991). Briefly, uterine and vaginal tissues were isolated from neonatal (0–1 days postpartum) Balb/c mice and rinsed in Dulbecco’s phosphate-buffered saline. Samples were incubated in 1% trypsin (Difco 1:250) in calcium-magnesium-free Hank’s buffer at 4 °C (1 hour for Va; 1.5 hours for Ut) and rinsed with 10% fetal bovine serum in Hank’s balanced salt solution containing 0.1% deoxyribonuclease 1 (Sigma) and then rinsed in Hank’s without deoxyribonuclease 1 supplemented with 10% serum (Boutin et al., 1992). Samples were separated into an epithelial (E) and mesenchymal (M) components by gentle teasing with forceps or pipetting through a drawn pipette (Bigsby et al., 1986). Epithelial and mesenchymal tissues were combined either homotypically or heterotypically on solidified agar medium (0.5% agar (Difco)) in Dulbecco’s modified Eagle’s medium H-16 containing 0.1% glucose with 10% fetal calf serum, 2 mM glutamine, 100 IU/ml penicillin and 100 μg/ml streptomycin). Tissues were allowed to re-adhere overnight and grafted under the renal capsule of syngeneic adult female mice. In order to assess contamination of the mesenchyme by residual epithelial tissue, vaginal and uterine mesenchymal samples were also implanted under the renal capsule. Of these grafts (n=3), none were observed to be contaminated with residual epithelium. Grafts were allowed to grow in vivo for three to four weeks and then grafts were harvested, fixed and processed for in situ analysis.

In situ analysis and preparation of probes

Techniques for in situ analysis have been previously described in detail elsewhere (Sassoon et al., 1988; Sassoon and Rosenthal, 1993). Probes were generated using either T7 or sp6 polymerase using 35S-radiolabeled UTP (>1000 Ci/mmol, New England Nuclear) and used in a hybridization buffer containing ∼35,000 cts/minute/μl final probe concentration. All probes were prepared under identical reaction conditions. Probes used for analysis of Msx1 (Hox-7.1) were generated from an EcoRI fragment of Msx1 in Bluescript (Bluescribe) linearized with Sac1 as described by Robert et al. (1989) (and see Wang et al., 1992). The probe for Id was generated using a Bluescribe recombinant plasmid containing a cDNA fragment from position 5 to 927 (Benezra et al., 1990). The plasmid was linearized with BamHI and antisense riboprobe was generated using the T7 polymerase (Wang et al., 1992). The Wnt-5a probe (kind gift from A. McMahon) was generated from a cDNA fragment corresponding to amino acid position 260 to 391 and subcloned into Bluescript PGEM3Zf; antisense riboprobe was obtained by sp6 polymerase following linearization of the plasmid by EcoRI to yield a 400-base fragment (and see Parr et al., 1993).

Ribonuclease protection analysis

Probes for nuclease protection analysis were prepared as previously described by Song et al. (1992) for Msx1 and Gavin and McMahon (1991) for Wnt-5a. Briefly, total RNA was isolated using cesium chloride purification following the guanidinium isothiocyanate procedure (Handy et al., 1988; and see Song et al., 1992) and protected fragments were separated by PAGE electrophoresis (6% urea/6% polyacrylamide) using RNA probes of known size and end-labeled pBR322 HinfI fragments as size markers. Antisense RNA probe of Msx1 (577 nucleotides including polylinker) was synthesized using the T7 promoter of a Bluescript clone containing a cDNA fragment of Hox-7.1 linearized at the BglII site (Hill et al., 1989; Song et al., 1992). Preparation of Wnt-5a probe is described as above for in situ. This probe is ∼450 nucleotides including polylinker and yields a protected fragment of ∼390 nucleotides. We note that some heterogeneity is observed in the protected fragments with this probe that presumably reflects some differences in the final size of the message. Heterogeneity of the Wnt-5a transcript has already been reported elsewhere (Gavin and McMahon, 1992).

Msx1 is expressed in the adult uterus and is downregulated following embryonic implantation

Ribonuclease protection analysis was utilized to assay the levels of Msx1 transcripts in the uterus. Msx1 levels in adult tissues generally fall below detectable levels (Song et al., 1992; Robert et al., 1989). In contrast, surprisingly high levels of Msx1 transcripts are observed in the adult uterus (Fig. 1). In uterine tissue from pregnant mice, levels of Msx1 transcripts are markedly lower (Fig. 1). To determine the cellular localization of Msx1 transcripts, we performed in situ hybridizations on adult non-pregnant and pregnant uteri. Results from this analysis reveal that Msx1 transcripts are restricted to luminal and glandular epithelial cells of the uterus (Fig. 1). The endometrial stroma did not show detectable levels of Msx1 expression (Fig. 1; Table 1). In pregnant uteri, the Msx1 hybridization signal is markedly decreased at 4.5 days p.c. (Fig. 1; Table 1) coincident with embryonic implantation and almost no detectable signal was observed at later time points (data not shown). Msx1 was not expressed in the vaginal mesenchyme or epithelium of pregnant or non-pregnant females (Fig. 1; Table 1).

Table 1.
Expression of Msx1 and Wnt-5a
graphic
graphic
Fig. 1.

(Upper panel) Msx1 and Wnt-5a levels decrease upon embryo-implantation. Ribonuclease protection analysis was performed on total RNA obtained from non-pregnant (virgin) and pregnant (5 days postcoitum) uterus. A protected band is clearly detected in the non-pregnant uterus for Msx1 and two bands are observed. Very weak protection of the cRNA probes is seen following embryo implantation (5 days p.c.). Yeast RNA (10 μg) was used as a control for both assays. The doublet observed for Wnt-5a presumably reflects alternate splicing of the Wnt-5a gene as previously observed by Gavin et al. (1992). Size markers are indicated at left side of panel.

(Lower panel) Msx1 levels decrease in uterine epithelium upon embryo-implantation. (A) Bright-field photomicrograph of adult uterus. Arrows denote the simple columnar epithelial cells of the uterine lumen which are columnar in morphology. (B) Same section shown in A using dark-field illumination following hybridization with a cRNA probe corresponding to Msx1. Strong signal is observed over the luminal and glandular epithelial cells. (C) Bright-field photomicrograph of uterus 5 days after conception. Note that the uterine glands have enlarged (open arrow). (D) Same section as shown in C following hybridization with a cRNA probe corresponding to Msx1. Note that epithelial cells do not show the strong and distinct accumulation of transcripts as seen in non-pregnant uterus (A,B). However, the reproducible low level of signal throughout the uterus is consistent with the weak bands detected in the ribonuclease protection analysis shown in Fig. 1. Signal does not appear to be restricted to any one cell type. (E) Bright-field photomicrograph of the adult vagina (non-pregnant). Arrows denote the typical stratified epithelium of the vagina. (F) Same section shown in E using dark-field illumination following hybridization with a cRNA probe corresponding to Msx1. No detectable signal is observed in any cell type. Scale bar, 100 μm.

Fig. 1.

(Upper panel) Msx1 and Wnt-5a levels decrease upon embryo-implantation. Ribonuclease protection analysis was performed on total RNA obtained from non-pregnant (virgin) and pregnant (5 days postcoitum) uterus. A protected band is clearly detected in the non-pregnant uterus for Msx1 and two bands are observed. Very weak protection of the cRNA probes is seen following embryo implantation (5 days p.c.). Yeast RNA (10 μg) was used as a control for both assays. The doublet observed for Wnt-5a presumably reflects alternate splicing of the Wnt-5a gene as previously observed by Gavin et al. (1992). Size markers are indicated at left side of panel.

(Lower panel) Msx1 levels decrease in uterine epithelium upon embryo-implantation. (A) Bright-field photomicrograph of adult uterus. Arrows denote the simple columnar epithelial cells of the uterine lumen which are columnar in morphology. (B) Same section shown in A using dark-field illumination following hybridization with a cRNA probe corresponding to Msx1. Strong signal is observed over the luminal and glandular epithelial cells. (C) Bright-field photomicrograph of uterus 5 days after conception. Note that the uterine glands have enlarged (open arrow). (D) Same section as shown in C following hybridization with a cRNA probe corresponding to Msx1. Note that epithelial cells do not show the strong and distinct accumulation of transcripts as seen in non-pregnant uterus (A,B). However, the reproducible low level of signal throughout the uterus is consistent with the weak bands detected in the ribonuclease protection analysis shown in Fig. 1. Signal does not appear to be restricted to any one cell type. (E) Bright-field photomicrograph of the adult vagina (non-pregnant). Arrows denote the typical stratified epithelium of the vagina. (F) Same section shown in E using dark-field illumination following hybridization with a cRNA probe corresponding to Msx1. No detectable signal is observed in any cell type. Scale bar, 100 μm.

Msx1 is expressed during development in Müllerianderived epithelium and specifically down-regulated in developing vaginal epithelium

Since Msx1 is expressed at high levels during early embryogenesis (Robert et al., 1989; Wang et al., 1992), we examined whether Msx1 was present during early female reproductive tract development. The uterus develops from the urogenital ridge, which consists of undifferentiated mesenchyme surrounding a tube lined by simple epithelial columnar cells. During postnatal development, the uterine epithelium differentiates into luminal and glandular epithelia, both of which retain their simple columnar morphology. The mesenchyme gives rise to smooth muscle cells of the myometrium and fibroblasts of the endometrial stroma (Brody and Cunha, 1989). The anterior portion of the vagina is derived from the Müllerian ducts and follows a similar program of development except that glands do not form and the epithelium becomes stratified (Forsberg, 1973; and see Figs 3D,H,I, 4). At 19 days of gestation (just prior to birth), the epithelium of the Müllerian ducts express Msx1 along their full length (prospective uterus and vagina) (Fig. 2A-C; Table 1). Prenatally, the Müllerian epithelial cells are simple columnar along their entire length and do not yet show any morphological differences between prospective vagina, cervix and uterus (Fig. 2A and lower panel). At birth, high levels of Msx1 are detected in the presumptive uterus, cervix and the Müllerian-derived (upper) portion of the vagina (data not shown) whereas, by 3 days postnatal, levels of Msx1 decline in the vagina. In the cervix, levels of Msx1 remain high at this stage (Fig. 2D,E). High levels of Msx1 expression detected in the developing uterus were maintained in all postnatal stages including adulthood (Fig. 2G,H). By 4–5 days, Msx1 levels declined sharply in the vagina. This coincided with the initial transition of the columnar Müllerian epithelium into a stratified vaginal epithelium (Fig. 2D). By two weeks postnatally, the cervix no longer has detectable levels of Msx1. Thus, Msx1 expression becomes progressively restricted to the rostral (presumptive uterus) portion of the Müllerian ducts during postnatal development suggesting that the segregation of the Müllerian duct into the uterus, cervix and vagina occurs as a gradual process of anterior-posterior differentiation and compartmentalization.

Fig. 2.

(Upper panel) Msx1 and Wnt-5a are developmentally regulated in the female reproductive tract. (A) Bright-field photomicrograph showing a transverse (coronal) section of a 19 day p.c. (fetal) Müllerian tract. The anteroposterior (cranial-caudal) axis is indicated at left. Top set of arrows denote uterine epithelium in the left uterine horn. Lower set of arrows denote vaginal epithelium. Cervix is denoted by dotted lines. (B) Same section shown in A hybridized with a probe corresponding to Msx1 viewed with dark-field optics. At this stage, both the presumptive uterus and vagina show strong labeling for Msx1 in the epithelial cells. (C) Neighboring section to B hybridized with a probe corresponding to Wnt-5a. Strong signal is detected throughout the mesenchymal layer of cells underlying the epithelium (denoted by arrows as on A and B) in both the uterus and vagina. (D) Bright-field photomicrograph of a transverse (coronal) section through the 4 day postnatal cervix and vagina. Arrow at top of panel indicates antero-posterior (cranial-caudal) axis. Thick arrow denotes epithelium near the cervical/vagina boundary used for comparison with E and F. (E) Dark-field photomicrograph of section shown in D hybridized for Msx1. Label is restricted to the epithelium but is slightly weaker in the more posterior portion of the vagina. (F) Dark-field photomicrograph of a neighboring section to E hybridized for Wnt-5a. Note that the orientation is different. Large arrow is placed in the same region as for D and E. As for Msx1, Wnt-5a signal is stronger in the cranial portion of the vagina (cervix), but remains clearly detectable in the posterior portion of the vagina at this stage. (G) Bright-field photomicrograph of a longitudinal section through the 4 day postnatal right uterine horn. Arrow at top of panel indicates cranial-caudal axis. Thick arrows denotes epithelium used for comparison with H and I. (H) Dark-field photomicrograph of section shown in G hybridized for Msx1. As can be seen, label is restricted to the epithelium. (I) Dark-field photomicrograph of a neighboring section to H, hybridized for Wnt-5a. Wnt-5a signal is strong throughout the entire uterine stromal layer. As seen in the 19 day fetus (C), some Wnt-5a signal is present in the uterine epithelial layer whereas the adult vaginal epithelia (Fig. 4) and the 4 day postnatal vaginal epithelia (F) reveal a clear boundary of Wnt-5a expression. Scale bar, 300 μm. (Lower Panel) Schematic summary of Msx1 and Wnt-5a expression during Müllerian tract development. At far left is the Müllerian tract at 19 days p.c. showing Msx1 expression (red) throughout the epithelium of the presumptive uterine horns and vagina and Wnt-5a expression (green) throughout the mesenchymal layer of both uterus and vagina. Division of uterus and vagina at the level of the presumptive cervix is indicated by dotted red line. Shown in middle panel is the reproductive tract at ∼5 days postnatal with Msx1 and Wnt-5a expression greatly reduced in the posterior vaginal portion. Shown at right is the reproductive tract at >2 weeks postnatal development. At this stage, Msx1 expression and Wnt-5a expression are confined to the uterus. Also, the epithelial cells of the vagina have become stratified in morphology whereas the uterine cells remain columnar.

Fig. 2.

(Upper panel) Msx1 and Wnt-5a are developmentally regulated in the female reproductive tract. (A) Bright-field photomicrograph showing a transverse (coronal) section of a 19 day p.c. (fetal) Müllerian tract. The anteroposterior (cranial-caudal) axis is indicated at left. Top set of arrows denote uterine epithelium in the left uterine horn. Lower set of arrows denote vaginal epithelium. Cervix is denoted by dotted lines. (B) Same section shown in A hybridized with a probe corresponding to Msx1 viewed with dark-field optics. At this stage, both the presumptive uterus and vagina show strong labeling for Msx1 in the epithelial cells. (C) Neighboring section to B hybridized with a probe corresponding to Wnt-5a. Strong signal is detected throughout the mesenchymal layer of cells underlying the epithelium (denoted by arrows as on A and B) in both the uterus and vagina. (D) Bright-field photomicrograph of a transverse (coronal) section through the 4 day postnatal cervix and vagina. Arrow at top of panel indicates antero-posterior (cranial-caudal) axis. Thick arrow denotes epithelium near the cervical/vagina boundary used for comparison with E and F. (E) Dark-field photomicrograph of section shown in D hybridized for Msx1. Label is restricted to the epithelium but is slightly weaker in the more posterior portion of the vagina. (F) Dark-field photomicrograph of a neighboring section to E hybridized for Wnt-5a. Note that the orientation is different. Large arrow is placed in the same region as for D and E. As for Msx1, Wnt-5a signal is stronger in the cranial portion of the vagina (cervix), but remains clearly detectable in the posterior portion of the vagina at this stage. (G) Bright-field photomicrograph of a longitudinal section through the 4 day postnatal right uterine horn. Arrow at top of panel indicates cranial-caudal axis. Thick arrows denotes epithelium used for comparison with H and I. (H) Dark-field photomicrograph of section shown in G hybridized for Msx1. As can be seen, label is restricted to the epithelium. (I) Dark-field photomicrograph of a neighboring section to H, hybridized for Wnt-5a. Wnt-5a signal is strong throughout the entire uterine stromal layer. As seen in the 19 day fetus (C), some Wnt-5a signal is present in the uterine epithelial layer whereas the adult vaginal epithelia (Fig. 4) and the 4 day postnatal vaginal epithelia (F) reveal a clear boundary of Wnt-5a expression. Scale bar, 300 μm. (Lower Panel) Schematic summary of Msx1 and Wnt-5a expression during Müllerian tract development. At far left is the Müllerian tract at 19 days p.c. showing Msx1 expression (red) throughout the epithelium of the presumptive uterine horns and vagina and Wnt-5a expression (green) throughout the mesenchymal layer of both uterus and vagina. Division of uterus and vagina at the level of the presumptive cervix is indicated by dotted red line. Shown in middle panel is the reproductive tract at ∼5 days postnatal with Msx1 and Wnt-5a expression greatly reduced in the posterior vaginal portion. Shown at right is the reproductive tract at >2 weeks postnatal development. At this stage, Msx1 expression and Wnt-5a expression are confined to the uterus. Also, the epithelial cells of the vagina have become stratified in morphology whereas the uterine cells remain columnar.

Fig. 3.

Uterine stroma from neonatal mice maintains Msx1 expression in vaginal epithelium. Tissue recombinants were performed between 0-1 day postnatal mesenchyme and epithelium. (A) Recombinant composed of uterine mesenchyme and vaginal epithelium (Mut+Eva). Dark-field photomicrograph of a Mut+Eva tissue recombinant hybridized for Msx1. Epithelial layer reveals Msx1 expression. Scale bar, 300 μm (A-C and E-G). (B) Section through same recombinant hybridized for Wnt-5a. Labeling is not detectable in the epithelium however, strong labeling is present in the stroma immediately subjacent to the epithelium layer. (C) Section through same recombinant hybridized for Id. Labeling is detectable in the epithelial layer. (D) Bright-field at higher magnification of same recombinant shown in A-D showing simple columnar morphology of the epithelial layer (delineated by arrows). Scale bar in D and H, 100 μm. (E) Recombination of vaginal mesenchyme and uterine epithelium (Mva+Eut). Dark-field photomicrograph of a tissue recombinant hybridized for Msx1. No detectable signal is observed in any tissue layer. (F) Section through same recombinant hybridized for Wnt-5a. No detectable signal is observed in any tissue layer. (G) Section through same recombinant hybridized for Id. Labeling is detectable in the epithelial layer. (H) Bright-field photomicrograph at higher magnification of same recombinant shown in E-G showing stratified morphology of the epithelial layer (delineated by arrows). (I) Schema indicating tissue recombinants. Mesenchyme grafted to epithelium from either uterine or vaginal 0-1 day reproductive tracts results in an epithelial morphology reflecting the source of the mesenchyme. In the case of the uterine mesenchyme (lower left), columnar morphology is retained in the epithelium and high Msx1 expression is observed.

Fig. 3.

Uterine stroma from neonatal mice maintains Msx1 expression in vaginal epithelium. Tissue recombinants were performed between 0-1 day postnatal mesenchyme and epithelium. (A) Recombinant composed of uterine mesenchyme and vaginal epithelium (Mut+Eva). Dark-field photomicrograph of a Mut+Eva tissue recombinant hybridized for Msx1. Epithelial layer reveals Msx1 expression. Scale bar, 300 μm (A-C and E-G). (B) Section through same recombinant hybridized for Wnt-5a. Labeling is not detectable in the epithelium however, strong labeling is present in the stroma immediately subjacent to the epithelium layer. (C) Section through same recombinant hybridized for Id. Labeling is detectable in the epithelial layer. (D) Bright-field at higher magnification of same recombinant shown in A-D showing simple columnar morphology of the epithelial layer (delineated by arrows). Scale bar in D and H, 100 μm. (E) Recombination of vaginal mesenchyme and uterine epithelium (Mva+Eut). Dark-field photomicrograph of a tissue recombinant hybridized for Msx1. No detectable signal is observed in any tissue layer. (F) Section through same recombinant hybridized for Wnt-5a. No detectable signal is observed in any tissue layer. (G) Section through same recombinant hybridized for Id. Labeling is detectable in the epithelial layer. (H) Bright-field photomicrograph at higher magnification of same recombinant shown in E-G showing stratified morphology of the epithelial layer (delineated by arrows). (I) Schema indicating tissue recombinants. Mesenchyme grafted to epithelium from either uterine or vaginal 0-1 day reproductive tracts results in an epithelial morphology reflecting the source of the mesenchyme. In the case of the uterine mesenchyme (lower left), columnar morphology is retained in the epithelium and high Msx1 expression is observed.

Fig. 4.

Expression of Msx1 and Wnt-5a the adult uterus and vagina. (A) Bright-field photomicrograph of adult uterus. Arrow denotes the simple columnar epithelial cells of the uterine lumen. (B) Same section shown in A using dark-field illumination following hybridization with a cRNA probe corresponding to Msx1. Strong signal is observed over the uterine luminal and glandular epithelium. (C) Neighboring section to A shown in bright-field optics hybridized with a cRNA probe to Wnt-5a. Arrow indicates epithelium of the uterine lumen and open arrow denotes the boundary of the endometrial stromal and myometrial (*) layers. (D) Dark-field optics showing accumulation of Wnt-5a in the subjacent endometrial stromal layer but not in the myometrium nor epithelium. (E) Bright-field photomicrograph of adult vagina. Arrows denote the epithelial cells of the vaginal lumen which are stratified in morphology. (F) Same section shown in E using dark-field illumination following hybridization with a cRNA probe corresponding to Msx1. No detectable signal is observed. (G) Neighboring section to E shown in bright-field optics hybridized with a cRNA probe to Wnt-5a.. Open arrow denotes the interface between the mesenchymal and epithelial layers. (H) Dark-field optics showing weak accumulation of Wnt-5a in the subjacent stromal layer. We note that signal for Wnt-5a is only localized to disperse regions of stroma in immediate contact with the epithelium, while most of the vaginal stroma is negative. Scale bar, 300 μm (A-H).

Fig. 4.

Expression of Msx1 and Wnt-5a the adult uterus and vagina. (A) Bright-field photomicrograph of adult uterus. Arrow denotes the simple columnar epithelial cells of the uterine lumen. (B) Same section shown in A using dark-field illumination following hybridization with a cRNA probe corresponding to Msx1. Strong signal is observed over the uterine luminal and glandular epithelium. (C) Neighboring section to A shown in bright-field optics hybridized with a cRNA probe to Wnt-5a. Arrow indicates epithelium of the uterine lumen and open arrow denotes the boundary of the endometrial stromal and myometrial (*) layers. (D) Dark-field optics showing accumulation of Wnt-5a in the subjacent endometrial stromal layer but not in the myometrium nor epithelium. (E) Bright-field photomicrograph of adult vagina. Arrows denote the epithelial cells of the vaginal lumen which are stratified in morphology. (F) Same section shown in E using dark-field illumination following hybridization with a cRNA probe corresponding to Msx1. No detectable signal is observed. (G) Neighboring section to E shown in bright-field optics hybridized with a cRNA probe to Wnt-5a.. Open arrow denotes the interface between the mesenchymal and epithelial layers. (H) Dark-field optics showing weak accumulation of Wnt-5a in the subjacent stromal layer. We note that signal for Wnt-5a is only localized to disperse regions of stroma in immediate contact with the epithelium, while most of the vaginal stroma is negative. Scale bar, 300 μm (A-H).

Msx1 expression is maintained via an interaction with uterine mesenchymal cells

Previous studies have established that differentiation of the Müllerian epithelium into the simple columnar glandular phenotype of the uterus or the stratified squamous phenotype of the vagina and cervix is induced and specified by uterine and vaginal mesenchyme, respectively (Cunha, 1976; Boutin et al., 1991, 1992). To determine whether mesenchyme plays a role in regulating epithelial expression of Msx1, heterotypic and homotypic tissue recombinants were prepared using 0 to 1 day old uterine and vaginal mesenchyme (M) and epithelium (E). In tissue recombinants composed of uterine mesenchyme plus uterine epithelium (Mut+Eut), uterine epithelial differentiation was observed and Msx1 expression was maintained in the epithelium (Table 1). Likewise, in tissue recombinants composed of Mut and Eva, uterine epithelial differentiation was induced from the vaginal epithelium and Msx1 was prominently expressed in the newly induced uterine epithelium (see Fig. 3A,D). Thus, the maintenance or induction of uterine epithelial differentiation was accompanied by maintenance of epithelial Msx1 expression (Fig. 3; Table 1). Grafts of uterine mesenchyme alone did not contain identifiable epithelial cells and Msx1 expression was not observed (data not shown). When vaginal mesenchyme was grown in association with either Eva or Eut, a stratified squamous vaginal epithelium developed and in both Mva+Eva and Mva+Eut recombinants, epithelial Msx1 expression was not observed (Fig. 3E,H; Table 1). Therefore, the maintenance or induction of vaginal epithelial differentiation was associated with an absence of epithelial Msx1 expression (Fig. 3E; Table 1), while induction or maintenance of uterine differentiation was associated with epithelial Msx1 expression (Fig. 3A).

The Id gene, which can promote cell proliferation and antagonize cellular processes involved in differentiation (Benezra et al., 1990; Jen et al., 1992), is expressed in the uterine and vaginal epithelium (data not shown). Thus, as a positive control for gene expression in the epithelium of the tissue recombinants, we examined Id expression using in situ hybridization. We observe that Id was detected in the epithelium of Mut/Eva tissue recombinants in which uterine epithelial differentiation was induced (Fig. 3C) and in Mva+Eut or va tissue recombinants both of which expressed vaginal differentiation (Fig. 3G). We conclude that Id expression is a common characteristic of both the vaginal and uterine epithelia and thus serves as a marker in the recombinant assays.

Wnt-5a is expressed in the uterine stroma

One set of genes that may provide important signaling events during development are the vertebrate homologues of the wingless gene family (Gavin and McMahon, 1991; Gavin et al., 1990). Wnt-5a, one member of this family, shows a striking colocalization with Msx1 in embryonic limb, brachial arches and genital eminence (Parr et al., 1993; our observations). Our initial investigation, using ribonuclease protection analysis revealed that Wnt-5a is expressed at high levels in the adult uterus, and as observed for Msx1, levels decrease following embryonic implantation (Fig. 1). Using in situ hybridization, we examined Wnt-5a expression in all situations investigated for Msx1. In contrast to the embryo, which exhibits overlapping domains of Msx1 and Wnt-5a gene expression (Parr et al., 1993), the adult uterus strongly expressed Wnt-5a in the endometrial stroma, but not in the Msx1-positive epithelial cells or in the myometrium (Fig. 4C-F). Thus, in the uterus, Wnt-5a expression differs spatially from that of Msx1. Vaginal stroma reveals barely detectable levels of Wnt-5a and no signal is observed in the Mva+Eut or Mva+Eva tissue recombinants which undergo vaginal differentiation (Figs 4E,F, 3F, 2E; Table 1). We note that in the Mut+Eva tissue recombinants, Wnt-5a levels are very high in the mesenchyme (Fig. 2B), thus, Wnt-5a expression, unlike Msx1, is a property of the mesenchyme source (uterine) and not subject to the same cellular regulation of Msx1. Indeed, in uterine mesenchyme grafted without epithelium, we observe that Wnt-5a expression was maintained (data not shown). Thus, Wnt-5a expression correlates with mesenchymal capacity to induce and/or maintain Msx1 expression in the adjacent epithelium and may be involved in the regulation of Msx1 gene expression via a cell-cell paracrine effect analogous to the situation for engrailed expression and Wnt-1 (McMahon et al., 1992). In agreement with this notion, we note that the pattern of expression of Wnt-5a in the developing female genital tract parallels Msx1 expression in adjacent epithelium. We observe that coincident with Msx1 expression in both presumptive uterine and vaginal epithelia, high levels of Wnt-5a were detected throughout the mesenchyme during late fetal stages (see Fig. 2C). By 4-5 days after birth, expression of Wnt-5a declined to moderate to low levels in the vaginal mesenchyme when Msx1 levels had sharply declined in the vaginal epithelium (Fig. 2 F,I). By 2–3 weeks postnatal, Wnt-5a decline further in vaginal and cervical mesenchyme to very low levels characteristic of the adult (see Fig. 2; Table 1).

In pregnant uteri, we observe that Wnt-5a levels sharply decline (Fig. 1, Table 1) concomitant with a decrease in detectable levels of Msx1. Again, these results are consistent with the notion that Wnt-5a is a good marker for the ability of mesenchymal cells in the female reproductive tract to support epithelial Msx1 expression.

Studies of Msx1 gene expression in the embryo have implicated Msx1 in maintaining an undifferentiated/proliferative cellular phenotype that is also positionally plastic (see Muneoka and Sassoon, 1992; Robert et al., 1989). In the embryo, Msx1 is expressed in migrating neural crest cells, the distal limb progress zone and visceral arches coincident spatially and temporally with cell plasticity in these regions (Wang et al., 1992; Robert et al., 1989; Muneoka and Sassoon, 1992). In particular, Msx1 is expressed in the limb bud mesenchyme prior to any sign of differentiation and is down-regulated as differentiation occurs (Wang et al., 1992). Msx1 expression coincides with the more highly proliferative limb progress zone, which has been demonstrated to be responsible for the proximal-distal limb outgrowth (Davidson et al., 1991; Robert et al., 1991; Summerbell 1974; Summerbell et al., 1973; Muneoka and Sassoon, 1992). Evidence that Msx1 underlies the proliferative cellular phenotype characteristic of its sites of expression was demonstrated by forced-expression of Msx1 in determined myogenic cells, which blocks differentiation and results in a highly proliferative, transformed and tumorigenic cellular phenotype (Song et al., 1992). Our observation that Msx1 is expressed at high levels in uterine epithelial cells suggests that the cellular plasticity of adult uterine epithelial cells may be due to expression of this gene, which is otherwise more typical of embryonic cells. This notion is supported by the observation that Msx1 expression is initiated in the embryonic Müllerian duct prior to overt differentiation of the uterus and is maintained thereafter throughout adulthood. The Müllerian epithelium undergoes cytodifferentiation in response to regionally specific mesenchyme (Cunha, 1976). While these organogenetic cell-cell interactions begin prenatally, morphogenesis and differentiation primarily occur postnatally and are not completed until the end of puberty. In the uterus, epithelial morphology and functional activity change during reproductive cycles (estrous or menstrual) in response to cyclic change in estrogen and progesterone levels. Decidualization of adult uterine epithelium following implantation of the embryo represents an even more dramatic change in uterine epithelial differentiation, which ultimately returns to its original differentiation state postpartum (Noyes, 1973; Finn, 1977). Thus, the uterus has a high degree of developmental plasticity and undergoes extensive morphological and functional changes in the adult. We propose that homeobox-containing genes, which have been clearly implicated in the establishment of patterning of the embryo, must remain active in the adult uterus to retain such a high degree of developmental plasticity.

The expression of homeogenes has been examined in numerous cases of embryonic development. Their homology to Drosophila homeotic genes and their pattern of gene expression in the embryo strongly supports the notion that they play a critical role in guiding patterning during development (see Gehring and Hiromi, 1986; Kessel and Gruss, 1991). In order to understand how homeogenes control patterning, it is also critical to understand how they are regulated. For the case of Msx1, it has been demonstrated that the apical ectodermal ridge (AER), which is required for limb outgrowth, must be in contact with the subjacent limb mesenchymal cells to support the mesenchymal expression of Msx1 in the chick (Robert et al., 1991; Coelho et al., 1991). In the mouse, Msx1 gene expression is observed when limb regeneration is induced, whereas Msx1 is not detectable when a simple wound response is elicited (DiStefano-Reginelli et al., unpublished data; and see Muneoka and Sassoon, 1992). Thus, the ability of the limb to maintain Msx1 gene expression is linked to its capacity to undergo patterning or a regenerative response. Presumably the AER contains and/or releases a factor(s) that regulates Msx1 expression (Robert et al., 1991). One obstacle in identifying a Msx1 inducer has been that the AER is extremely small and cannot be cultured for long periods of time, and thus is not amenable to biochemical and molecular analysis (see for example Boutin and Fallon, 1992). Our observation that uterine mesenchyme is capable of inducing and/or maintaining Msx1 gene expression in Müllerian-derived epithelium may provide a complementary system for characterizing an Msx1 inducer. Msx1 and Wnt-5a expression are found in the limb and the uterus; however, it is not yet clear that their expression is regulated by the same factors present in both sites. It is remarkable nonetheless that these two rather different structures share related regulatory molecules. It has been recently reported that HOX4E, which is also expressed in the developing limb can be detected in the adult uterus (Redline et al., 1992). In many instances, related members of a particular homeobox gene family are expressed in spatially overlapping and/or neighboring domains. Msx2 (formerly Hox-8), another member of the msh gene family, is seen in adjacent but more distal domains to Msx1 in the developing neural tube, eye and limb (Monaghan et al., 1991; Holland, 1991). In contrast to these situations in the embryo, Msx2 is not detected or present at very low levels in the adult uterus (data not shown).

Numerous other examples of mesenchymal-epithelial interaction have been studied in the embryo including the skin (Sawyer, 1983; Dhouailly, 1976), mammary gland (Kratochwil, 1987; Sakakura et al., 1976; Daniel and Silberstein, 1987; Cunha et al., 1992), gastrointestinal tract (Haffen et al., 1987) and tooth (Kollar, 1983; Ruch et al., 1983; Thesleff and Hurmerinta, 1981). Many of these sites of classical epithelialmesenchymal interactions are also sites of Msx1 expression (see Jowett et al., 1993; MacKenzie et al., 1991, 1992 (tooth induction); Robert et al., 1989, 1991; Suzuki et al, 1991; Takahashi et al., 1991 (limb and craniofacial structures)). It is thought that, once the morphology is fully developed, no further induction is required. However, in some adult tissues or organs such as the uterus, it is possible that the morphology is actively maintained (Cunha et al., 1985). Several other gene transcripts or their protein products have also been found to be associated with sites of epithelial-mesenchymal interactions including those sites at which Msx1 is expressed. Most notable among these are members of the FGF and TGFβ superfamilies including FGF4, BMP-4 and BMP-2a (Lyons et al., 1989; Jones et al., 1991; and see Niswander and Martin, 1993). At present, none of these factors have been demonstrated to mediate Msx1 expression and are not present at all sites of Msx1 expression (unpublished results). At present, it seems likely that these factors, in combination with Wnt-5a and other unidentified factors, may underlie the induction or maintenance of Msx1.

In the case of the developing uterus, not only do mesenchymal cells dictate the development and differentiation of the epithelial cells, but it is equally important that the epithelial cells are able to respond to mesenchymal signals. Responsiveness to uterine induction may require the epithelial expression of Msx1. Previous studies have demonstrated that the germ layer origin of the mouse vaginal epithelium restricts its responsiveness to uterine mesenchymal induction (Boutin et al., 1992). In these studies, we restricted our attention to vaginal epithelium derived from the Müllerian duct (mesodermally derived) which is capable of being induced to undergo uterine differentiation. By contrast, the sinus vagina, which constitutes the most caudal portion of the vagina (endodermally derived), is apparently incapable of being induced to express uterine differentiation (Boutin et al., 1992). It may be that the developmental history (germ layer origin and previous Msx1 gene expression) underlie differential sensitivity to uterine induction (Boutin et al., 1992).

Mesenchyme surrounding the embryonic Müllerian ducts, like that of other internal organs, will ultimately differentiate into two major cell types, fibroblasts and smooth muscle cells. In the uterus, these cell types segregate into two distinct layers. The fibroblasts remain associated with the luminal and glandular epithelial cells, while the smooth muscle cells form the thick myometrium, which surrounds the endometrial stroma. The segregation of these two layers begins at about 3 days postpartum in the mouse, but is not completed until about 15 days after birth. It should be noted that, at the end of gestation (before these layers have segregated), the full thickness of the undifferentiated uterine mesenchyme uniformly expresses Wnt-5a whereas, in adulthood, Wnt-5a is expressed solely in the endometrial stromal and not in the myometrium. The endometrial stroma, but not the myometrium, undergoes dramatic phenotypic change during decidualization. Thus, the expression of Wnt-5a in the adult uterus occurs in developmentally plastic stromal cells involved in cell-cell interactions with epithelium. By contrast, adult vaginal stroma, which is developmentally quiescent, has a nearly undetectable level of Wnt-5a expression. It is of interest that Wnt-5a is also expressed in the adult mammary gland, which, like the uterus, undergoes dramatic morphological and functional changes during pregnancy, lactation and weaning (Gavin and McMahon, 1991). Given the close association of Wnt-5a and Msx1 in the uterus and limb bud, it is possible that these two gene products are involved in a regulatory circuit similar to that proposed for Wnt-1 and the homeobox gene engrailed (McMahon et al., 1992). We propose that homeogenes, which have been clearly implicated in the establishment of proper patterning of the embryo, must remain active in specific adult tissues that retain a high degree of developmental plasticity and which undergo further morphogenesis as part of their normal function. Furthermore, the fact that Msx1 has been demonstrated to affect cell proliferation and mediate cellular oncogenesis (Song et al., 1992) raises the intriguing possibility that deregulated expression of Msx1 may underlie abnormal cellular growth in the female reproductive tract.

We thank A. McMahon for the kind gift of the Wnt-5a cDNA and for suggesting there may be a relationship with Msx1. We also thank Drs G. Marazzi and G. Sonenshein for critique of the manuscript as well as Natalia Bogdanova, Esfir Slonimsky, Xun Wang, Yaoqi Wang, Kening Song and John Lincecum for their advice and help. This study was supported in part by grants from The Council for Tobacco Research, RO1 HD27585 and DK44269 to D. S. and CA05388, CA59831to G. C.

Benezra
,
R.
,
Davis
,
R. L.
,
Lockshon
,
D.
,
Turner
,
D. L.
and
Weintraub
,
H.
(
1990
).
The protein Id: a negative regulator of helix-loop-helix DNA binding proteins
.
Cell
61
,
49
59
.
Bigsby
,
R. M.
,
Cooke
,
P. S.
and
Cunha
,
G. R.
(
1986
).
A simple method for separating murine uterine epithelial and mesenchymal cells
.
Am. J. Physiol
.
251
,
E630
636
.
Boutin
,
E. L.
,
Sanderson
,
R. D.
,
Bernfield
,
M.
and
Cunha
,
G. R.
(
1991
)
Epithelia-mesenchymal interactions in uterus and vagina alter the expression of the cell surface proteoglycan, syndecan
.
Dev. Biol
.
148
,
63
74
.
Boutin
,
E. L.
,
Battle
,
E.
and
Cunha
,
G. R.
(
1992
).
The germ layer origin of mouse vaginal epithelium restricts its responsiveness to mesenchymal inductors: uterine induction
.
Differentiation
49
,
101
107
.
Boutin
,
E. L.
and
Fallon
,
J. F.
(
1992
).
Insulin improves survival but does not maintain function of cultured chick wing bud apical ectodermal ridge
.
Anat. Rec
.
233
,
467
477
.
Brody
,
J. R.
and
Cunha
,
G. R.
(
1989
).
Histologic, morphometric and immunocytochemical analysis of myometrial development in rats and mice. 1. Normal development
.
Am. J. Anat
.
186
,
1
20
.
Chisaka
,
O.
and
Capecchi
,
M. R.
(
1991
)
Regionally restricted developmental defects resulting from targeted disruption of the mouse homeobox gene Hox- 1.5
.
Nature
350
,
473
479
.
Coelho
,
C. N. D.
,
Krabbenhoft
,
K.
,
Upholt
,
W. B.
,
Fallon
,
J. F.
and
Kosher
,
R. A.
(
1991
).
Altered expression of the chicken homeobox-containing genes GHox-7 and GHox-8 in the limb buds of limbless mutant chick embryos
.
Development
113
,
1487
1493
.
Cunha
,
G. R.
(
1976
).
Stromal induction and specification of morphogenesis and cytodifferentiation of the epithelia of the Müllerian ducts and urogenital sinus during development of the uterus and vagina in mice
.
J. Exp. Zool
.
196
,
361
370
.
Cunha
,
G. R.
,
Bigsby
,
R. M.
,
Cooke
,
P. S.
and
Sugimura
,
Y.
(
1985
).
Stromal-epithelial interactions in adult organs
.
Cell Diff
.
17
,
137
148
.
Cunha
,
G. R.
,
Young
,
P.
,
Hamamoto
,
S.
,
Guzman
,
R.
and
Nandi
,
S.
(
1992
).
Developmental response of adult mammary epithelial cells to various fetal and neonatal mesenchymes
.
Epithelial Cell Biol
.
1
,
105
118
.
Daniel
,
C. W.
and
Silberstein
,
G. B.
(
1987
).
Postnatal development of the rodent mammary gland
. In
The Mammary Gland Development, Regulation and Function
. (ed.
M. C.
Neville
and
C. W.
Daniel
). pp.
3
36
.
New York
:
Plenum Press
.
Davidson
,
D. R.
,
Crawley
,
A.
,
Hill
,
R. E.
and
Tickle
,
C.
(
1991
).
Position- dependent expression of two related homeobox genes in developing vertebrate limbs
.
Nature
352
,
429
431
.
Dhouailly
,
D.
(
1976
).
Analyse des facteurs de la differenciation specifique de la plume neoptile chez le canard et le poulet
.
J. Embryol. Exp. Morph
.
18
,
389
400
.
Duggal
,
R. N.
,
Zakeri
,
Z. F.
,
Ponzetto
,
C.
and
Wolgemuth
,
D. J.
(
1987
).
Differential expression of the c-abl proto-oncogene and the homeobox-containing gene Hox-1.4 during mouse spermatogenesis
. In
Cell Biology of the Testis and Epididymis
(ed.
M. C.
Orgebin-Crist
and
B. J.
Danzo
) Ann. NY Acad Sci.
513
,
112
127
.
Finn
,
C. A.
(
1977
).
The implantation reaction
. In
Biology of the Uterus
(ed.
R. M.
Wynn
) pp.
246
308
.
New York
:
Plenum Press
.
Forsberg
,
J-G.
(
1973
).
Cevicovaginal epithelium: Its origin and development
.
Am. J. Obstet. Gynecol
.
115
,
1025
1043
.
Gavin
,
B. J.
,
McMahon
,
J. A.
and
McMahon
,
A. P.
(
1990
).
Expression of multiple novel Wnt-1/int-1-related genes during fetal and adult mouse development
.
Genes Dev
.
4
,
2319
2332
.
Gavin
,
B. J.
and
McMahon
,
A. P.
(
1991
).
Differential regulation of the Wnt gene family during pregnancy and lactation suggests a role in postnatal development of the mammary gland
.
Mol. Cell. Biol
.
12
,
2418
2423
.
Gehring
,
W. J.
and
Hiromi
,
Y.
(
1986
).
Homeotic genes and the homeobox
.
Ann. Rev. Genet
.
20
,
147
173
.
Haffen
,
K.
,
Kedinger
,
M.
and
Simon-Assmann
,
P.
(
1987
).
Mesenchymedependent differentiation of epithelial progenitor cells in the gut
.
J. Pediatr. Gastroenterol. Nutr
.
6
,
14
23
.
Handy
,
D. E.
,
McCluskey
,
J.
,
Lew
,
A. M.
,
Coligar
,
J. E.
and
Margulies
,
D. H.
(
1988
).
Signals controlling alternative splicing of major histocompatibility complex H-2 class I pre-mRNA
.
Immunogenetics
28
,
81
90
.
Hill
,
R. E.
,
Jones
,
P. F.
,
Rees
,
A. R.
,
Sime
,
C. M.
,
Justice
,
M. J.
,
Copeland
,
N. G.
,
Jenkins
,
N. A.
,
Graham
,
E.
and
Davidson
,
D. R.
(
1989
).
A new family of mouse homeobox-containing genes: molecular structure, chromosomal location, and developmental expression of Hox7.1
.
Genes Dev
.
3
,
26
37
.
Holland
,
P. W. H.
(
1991
).
Cloning and evolutionary analysis of msh-like genes from mouse, zebrafish and ascidian
.
Gene
98
,
253
257
.
Jackson
,
I. J.
,
Schofield
,
P.
and
Hogan
,
B.
(
1985
).
A mouse homeobox gene is expressed during embryogenesis and in the adult kidney
.
Nature
317
,
745
748
.
Jen
,
Y.
,
Weintraub
,
H.
and
Benezra
,
R.
(
1992
).
Overexpression of Id protein inhibits the muscle differentiation program: in vivo association of Id with E2A proteins
.
Genes Dev
.
6
,
1466
1479
.
Jones
,
C. M.
,
Lyons
,
K. M.
and
Hogan
,
B. L. M.
(
1991
).
Involvement of Bone Morphogenetic Protein-4 (BMP-4) and Vgr-1 in morphogenesis and neurogenesis in the mouse
.
Development
111
,
531
542
.
Jowett
,
A. K.
,
Vainio
,
S.
,
Ferguson
,
M. W. J.
,
Sharpe
,
P. T.
and
Thesleff
,
I.
(
1993
).
Epithelial-mesenchymal interactions are required for Msx1 and Msx- 2 gene expression in the developing murine molar tooth
.
Development
117
,
461
470
.
Kessel
,
M.
and
Gruss
,
P.
(
1991
).
Homeotic transformations of murine vertebrae and concomitant alteration of Hox codes induced by retinoic acid
.
Cell
67
,
89
104
.
Kollar
,
E. J.
(
1983
).
Epithelial-mesenchymal interactions in the mammalian integument: Tooth development as a model for instructive induction
. In
Epithelial-mesenchymal Interactions in Development
. (ed.
R. H.
Sawyer
and
J. F.
Fallon
). pp.
27
49
.
New York
:
Praeger
.
Kratochwil
,
K.
(
1987
).
Tissue combination and organ culture studies in the development of the embryonic mammary gland
. In
Developmental Biology: A Comprehensive Synthesis
(ed. R. B. L. Gwatkin). pp 315-334.
New York
:
Plenum Press
.
Lufkin
,
T.
,
Dierich
,
A.
,
LeMeur
,
M.
,
Mark
,
M.
and
Chambon
,
P.
(
1991
).
Disruption of the Hox-1.6 homeobox gene results in defects in a region corresponding to its rostral domain of expression
.
Cell
66
,
1105
1119
.
Lyons
,
K. M.
,
Pelton
,
R. W.
and
Hogan
,
B. L. M.
(
1989
).
Patterns of expresion of murine Vgr-1 and BMP-2a RNA suggest that transforming growth factor-β-like genes coordinately regulate aspects of embryonic development
.
Genes Dev
.
3
,
1657
1668
.
Mackenzie
,
A.
,
Leeming
,
G. L.
,
Jowett
,
A. K.
,
Ferguson
,
M. W. J.
and
Sharpe
,
P. T.
(
1991
).
The homeobox gene Hox7.1 has specific regional and temporal expression patterns during early murine craniofacial embryogenesis, especially tooth development in vivo and in vitro
.
Development
111
,
269
285
.
MacKenzie
,
A.
,
Ferguson
,
M. W. J.
and
Sharpe
,
P. T.
(
1992
).
Expression patterns of the homeobox gene Hox-8 in the mouse embryo suggest a role in specifying tooth initiation and shape
.
Development
115
,
403
420
.
McMahon
,
A. P.
,
Joyner
,
A. L.
,
Bradley
,
A.
and
McMahon
,
J. A.
(
1992
).
The mid-hindbrain phenotype of Wnt-1/Wnt-1 mice results from a stepwise deletion of engrailed-expressing cells by 9.5 days post-coitum
.
Cell
69
,
581
595
.
Monaghan
,
A. P.
,
Hill
,
R. E.
,
Bhattacharya
,
S. S.
,
Graham
,
E.
and
Davidson
,
D. R.
(
1991
).
Msh-like homeobox genes define domains in the developing vertebrate eye
.
Development
112
,
1053
1061
.
Muneoka
,
K.
and
Sassoon
,
D.
(
1992
).
Molecular aspects of regeneration in developing vertebrate limbs
.
Dev. Biol
.
152
,
37
49
.
Niswander
,
L.
and
Martin
,
G.
(
1993
).
FGF-4 regulates expression of Evx-1 in the developing mouse limb
.
Development
119
,
287
294
.
Noyes
,
R. W.
(
1973
).
Normal phases of the endometrium
. In
The Uterus
. (ed.
H. J.
Norris
,
A. T.
Hertig
, and
M. R.
Abell
), pp.
110
-
135
.
Baltimore
:
Williams & Wilkins.
Parr
,
B. A.
,
Shea
,
M. J.
,
Vassileva
,
G.
and
McMahon
,
A. P.
(
1993
).
Mouse Wnt genes exhibit discrete domains of expression in the early embryonic CNS and limb buds
.
Development
119
,
247
261
.
Redline
,
R. W.
,
Williams
,
A. J.
,
Patterson
,
P.
and
Collins
,
T.
(
1992
).
Human HOX4E: a gene strongly expressed in the adult male and female urogenital tracts
.
Genomics
13
,
425
430
.
Robert
,
B.
,
Sassoon
,
D.
,
Jacq
,
B.
,
Gehring
,
W.
and
Buckingham
,
M.
(
1989
).
Expression of a novel Hox gene, Hox-7, during mouse embryogenesis, is associated with morphogenetic phenomena
.
EMBO J
.
8
,
91
100
.
Robert
,
B.
,
Lyons
,
G.
,
Simandl
,
B.-K.
,
Kuroiwa
,
A.
and
Buckingham
,
M.
(
1991
).
The apical ectodermal ridge regulates Hox-7 and Hox-8 gene expression in developing limb buds
.
Genes Dev
.
5
,
2363
2374
.
Ruch
,
J. V.
,
Lesot
,
H.
,
Karcher-Djuricic
,
V.
,
Meyer
,
J. M.
and
Mark
,
M.
(
1983
).
Epithelial-mesenchymal interactions in tooth germs: mechanisms of differentiation
.
J. Biol. Buccale
11
,
73
93
.
Sakakura
,
T.
,
Nishizuka
,
Y.
and
Dawe
,
C. J.
(
1976
).
Mesenchyme-dependent morphogenesis and epithelium-specific cytodifferentiation in mouse mammary gland
.
Science
194
,
1439
1441
.
Sassoon
,
D.
,
Garner
,
I.
and
Buckingham
,
M. E.
(
1988
).
Transcripts for a-cardiac and a-skeletal actin are early markers for myogenesis in the mouse embryo
.
Development
104
,
155
164
.
Sassoon
,
D.
and
Rosenthal
,
N.
(
1993
).
Detection of messenger RNA by in situ hybridization
. In
Methods in Enzymology
225
, pp.
384
404
.
Sawyer
,
R. H.
(
1983
).
The role of epithelial-mesenchymal interactions in regulating gene expression during avian scale mojrphogenesis
. In
Epithelial-mesenchymal Interactions in Development
. (ed.
R. H.
Sawyer
and J. F. Fallon). pp 115-146
.
New York
:
Praeger
.
Scott
,
M.
(
1992
).
Vertebrate homeobox gene nomenclature
.
Cell
71
,
551
553
.
Song
,
K.
,
Wang
,
Y.
and
Sassoon
,
D. A.
(
1992
).
Expression of Hox-7.1 in myoblasts results in inhibition of differentiation and induces cell transformation
.
Nature
360
,
477
481
.
Summerbell
,
D.
,
Lewis
,
J. H.
and
Wolpert
,
L.
(
1973
).
Positional information in chick limb morphogenesis
.
Nature
244
,
492
496
.
Summerbell
,
D.
(
1974
).
A quantitative analysis of the effect of excision of the AER from the chick limb bud
.
J. Embryol. Exp. Morph
.
32
,
651
660
.
Suzuki
,
H. R.
,
Padanilam
,
B. J.
,
Vitale
,
E.
,
Ramirez
,
F.
and
Solursh
,
M.
(
1991
).
Repeating developmental expression of G-Hox 7, a novel homeoboxcontaining gene in the chicken
.
Dev. Biol
.
148
,
375
388
.
Takahashi
,
Y.
,
Bontoux
,
M.
and
Le Douarin
,
N. M.
(
1991
).
Epithelio-mesenchymal interactions are critical for Quox-7 expression and membrane bone differentiation in the neural crest derived mandibular mesenchyme
.
EMBO J
.
10
,
2387
2393
.
Thesleff
,
I.
and
Hurmerinta
,
K.
(
1981
).
Tissue interactions in tooth development
.
Differentiation
18
,
75
88
.
Wang
,
Y.
,
Benezra
,
R.
and
Sassoon
,
D. A.
(
1992
).
Id expression during mouse development: a role in morphogenesis
.
Devel. Dynamics
194
,
222
230
.