Both Dr. Veltmaat and W. van Veelen contributed equally to this work.
Patterns & Phenotypes
Identification of the mammary line in mouse by Wnt10b expression
Version of Record online: 8 JAN 2004
Copyright © 2004 Wiley-Liss, Inc.
Volume 229, Issue 2, pages 349–356, February 2004
How to Cite
Veltmaat, J. M., Van Veelen, W., Thiery, J. P. and Bellusci, S. (2004), Identification of the mammary line in mouse by Wnt10b expression. Dev. Dyn., 229: 349–356. doi: 10.1002/dvdy.10441
- Issue online: 20 JAN 2004
- Version of Record online: 8 JAN 2004
- Manuscript Accepted: 24 AUG 2003
- Manuscript Revised: 12 AUG 2003
- Manuscript Received: 29 MAR 2003
- Fifth Framework of the European Commission. Grant Number: QLGA-CT-2000-52130
- Dutch Cancer Fund/KWF
- French Association for Cancer Research/ARC. Grant Number: 4223
- French Ministry of Research/ACI. Grant Number: 81/2001
- California Breast Cancer Research program. Grant Number: 9PB-0082
- mouse embryo;
- mammary line;
- mammary gland development
Mammogenesis in rabbit, rat, and human begins with the formation of an elevated ectodermal ridge in the embryo. Attempts to demonstrate a morphologically or histologically equivalent mammary line in the mouse have yielded controversial results. We show here that a mammary line exists in the mouse embryo at embryonic day (E) 11.25 as a concise line of Wnt10b expression and a broader band of Wnt6 expression in the surface ectoderm, between the subaxillary and suprainguinal region of each flank. Additional streaks of Wnt10b expression in the axillary and inguinal region join the mammary line on the flank slightly later. Expression of Wnt10b and Wnt6 becomes restricted to the placodes within 1.5 days. The ectoderm of the mammary line is organized as a pseudostratified epithelium connecting the developing mammary placodes at around E11.5, whereas all other surface ectoderm is single-layered. Thus, the mammary line expressing Wnt10b defines a distinct ectodermal region that we consider the equivalent of the ectodermal ridge in, for example, rabbit. To date, the formation of the mammary line expressing Wnt10b is the earliest discernible ectodermal event in murine embryonic mammary gland development. Developmental Dynamics 229:349–356, 2004. © 2004 Wiley-Liss, Inc.
The onset of mammogenesis in mammals such as rabbit, rat, and human becomes apparent as the mammary line, an ectodermal ridge between the forelimb and hindlimb of each flank that is elevated above the surrounding surface ectoderm (Myers, 1917; Raynaud, 1961; Propper, 1976). In rabbit, these mammary ridges gradually fragment and subside in anteroposterior order, leaving only the mammary buds behind as elevations (Propper, 1976). In mouse embryos, such an elevated ridge has never been detected and the existence of an equivalent structure has been much debated (reviewed in Veltmaat et al., 2003). Currently, the mammary line in mouse is considered a theoretical concept, as one can draw a slightly curved line through these placode positions, at or very near the dorsoventral boundary on each flank of the embryo (reviewed in Robinson et al., 1999; reviewed in Veltmaat et al., 2003).
The first evidence of mammogenesis in mouse is the appearance of five left–right symmetric pairs of mammary placodes seen at embryonic day (E)11.5 by scanning electron microscopy as elevations above the surrounding surface ectoderm (Mailleux et al., 2002). The morphologic appearance of placodes coincides with the high expression of the transcription factor Lef1 (Van Genderen et al., 1994; Mailleux et al., 2002), an effector of canonical WNT signaling (Oosterwegel et al., 1993). The first placode pair is located in the axillary region, giving rise to the pectoral pair of mammary glands. The second, third, and fourth pairs are located on the flank; in the subaxillary region, thoracic region, and suprainguinal region, respectively. The fifth pair is localized in the inguinal region. As the position of each of the mammary rudiments shifts more ventrally later in life, pairs 1, 2, and 3 are generally considered thoracic pairs and pairs 4 and 5 inguinal pairs. Peculiarly, unlike in the rabbit (Propper, 1976), the mammary placodes in the mouse, as detected by scanning electron microscopy and Lef1 expression, do not appear in an anteroposterior order, but in the order of pair number 3, then 4, followed by 5 and 1, and finally pair number 2 (Mailleux et al., 2002). Furthermore, in wild-type strains, different placode pairs have a different susceptibility to spontaneous induction defects, such as the induction of supernumerary placodes, or mispositioning of the placode (Little and McDonald, 1945; reviewed in Veltmaat et al., 2003). Even more striking is the failure of induction of all but the number 4 placode pair in absence of FGF10 function (Mailleux et al., 2002) and the relatively high susceptibility of placode pairs 3 and 4 to a variety of induction aberrancies in Ska mice (Howard and Gusterson, 2000a, b). This once more raises the question of whether the placodes originate from a common structure, i.e., a mammary line.
Guided by the knowledge that Lef1 (Van Genderen et al., 1994; Mailleux et al., 2002) and Wnt10b (Christiansen et al., 1995) are expressed in the mammary rudiments at E11.5–E12.5 and that the mammary rudiments are absent in E13.5 mouse embryos either lacking Lef1 expression (Van Genderen et al., 1994) or overexpressing the WNT inhibitor Dkk1 in the skin (Andl et al., 2002), we decided to examine the expression of Wnt genes in the mammary region of the mouse embryo by in situ hybridization. We show here that expression of Wnt10b precedes the appearance of mammary placodes. Wnt6 and Wnt10b are expressed along the whole flank in a band and concise line, respectively, through the positions of the prospective placodes. Specifically, the line of Wnt10b expression is reminiscent of a mammary line. Analysis of E11.25 and E11.75 embryos revealed that this mammary line is histologically discernible as a small band of pseudostratified epithelium that connects the mammary placodes with each other on the boundary of dorsal and ventral ectoderm.
Wnt10b Expression Reveals the Existence of a Mammary Line in the Mouse Embryo
The expression of Wnt2a, Wnt2b, Wnt3a, Wnt4, Wnt5a, Wnt5b, Wnt6, Wnt7b, Wnt10a, and Wnt10b was assessed by whole-mount in situ hybridization on wild-type mouse embryos at E11.5–E12.0. Of these genes, Wnt6, Wnt10a, and Wnt10b were expressed in the mammary placodes. In contrast to the concise expression of Wnt10a in the placodes only, Wnt10b and Wnt6 expression also appeared in the region in between the placodes on the flank, and we, therefore, decided to study the expression patterns of these two genes over a time frame from E10.25 (30–32 somites) through E12.0 (48–49 somites).
In embryos between the 30- and 35-somite stage, Wnt10b expression was not detected in the region where mammary glands develop (data not shown). Figure 1 shows the expression of Wnt10b in embryos between 36 somites and 49 somites. With high variability between the 36- and 38-somite stage, a streak of Wnt10b expression appears immediately dorsally to the forelimb and splits into a dorsal and a ventral line (Fig. 1A,A′) that descend toward the hindlimb at later stages. Whereas this initial expression of Wnt10b is not located in the region where mammary glands develop, in the figures we refer to the ventral line as the mammary line because this line shows a relative increase of Wnt10b expression at the position of the prospective placode number 3 at the 38- to 40-somite stage (Fig. 1B,B′,C,C′). Moreover, at the 42-somite stage, expression is also increased at the position of the prospective placode number 4 (Fig. 1D,D′). Meanwhile, the ventral line has lost Wnt10b expression dorsally to the forelimb, whereas the dorsal line shows increased Wnt10b expression in this region, while extending toward the hindlimb (Fig. 1B,B′,C,C′). Between the position of placodes 3 and 4, Wnt10b expression in both the ventral and the dorsal line first appears as fragments that align with the somites at the 40-somite stage (Fig. 1C′,D′), but subsequently as a continuous line along the flank at around the 45-somite stage (Fig. 1E,E′). In this time frame between the 40- and 45-somite stage, Wnt10b expression on the ventral line progressively increases at the positions of the developing placodes 3 and 4, while disappearing between those placodes (Fig. 1E,E′,F,F′). Moreover, Wnt10b expression on the ventral line extends anteriorly congruently with the dorsal line toward the subaxillary position of the prospective placode 2 (Fig. 1E,E′,F,F′), where it subsequently becomes elevated at the 49-somite stage (Fig. 1G,G′).
Vibratome sectioning of Wnt10b-probed embryos revealed that, at the 43-somite stage, Wnt10b is not expressed in the somites, whereas it is expressed ubiquitously in the dermis and the surface ectoderm (Fig. 1J,J′). Wnt10b expression is particularly high in the mammary placode 3 (Fig. 1J,J′) and in the line of ectoderm extending between placodes 3 and 4 (data not shown). Additionally, Wnt10b expression is slightly elevated in the surface ectoderm and dermis of the dorsal line that appears to be located at the dorsal indentation of the body wall. At the 47-somite stage, Wnt10b expression is no longer detectable in the dermis and has also disappeared from most of the surface ectoderm (Fig. 1K,K′). High levels of expression are seen in the mammary placodes. Given the appearance of Wnt10b positive mammary placodes in the surface ectoderm of the ventral line of Wnt10b expression but not on the dorsal line, we consider the ventral line to be the mammary line.
Slightly after the appearance of the mammary line on the flank, additional streaks of Wnt10b expression appear at the ventral border of each of the limbs: first in the axilla at the 38- to 40-somite stage and then in the inguinus at the 40- to 42-somite stage. By the 49-somite stage, these streaks have extended toward the extremities of the mammary line. With some variability between the 38- and 40-somite stage, Wnt10b expression first appears as a thin line throughout the axilla or only at the position of the prospective placode 1 (Fig. 1B′′,C′′) and progressively elevates at the placode position (Fig. 1D′′). A thin line extends downward from this position, but it does not reach the position of the prospective placode 2 until the 45- to 46-somite stage (Fig. 1E′′,F′′). At the 49-somite stage, Wnt10b expression is restricted to placode 1 with a very thin stripe remaining connected to the developing placode 2 (Fig. 1G′′).
In the inguinal region, Wnt10b expression first appears at the 40-somite stage as a very thin line in the surface ectoderm from the posterior end to the position of the prospective placode 5 (Fig. 1C′′′). Expression becomes broader and elevated and extends anteriorly (Fig. 1D′′′,E′′′) but does not reach the prospective placode 4 until between the 46- and 49-somite stages (Fig. 1F′′′,G′′′). Unlike placodes 3, 4, and 1, placode 5 is never distinguished as a dense dot of Wnt10b expression during the time frame studied (Fig. 1G′′′), although a local thickening of the inguinal streak of Wnt10b expression at the 45-somite stage (Fig. 1E′′′) is suggestive of placode formation. A recapitulation of the spatiotemporal order of appearance of Wnt10b expression in the mammary line on the flank and in the mammary streaks in the axilla and inguinus (Fig. 1H), as well as the restriction of Wnt10b expression to the different foci suggests that the mammary placodes are formed in the temporal order of number 3 then 4, followed by 1 and 5, and finally number 2.
Wnt6 Is Expressed in a Broad Band at and Dorsal From the Mammary Line
In embryos between the 30- and 39-somite stage, no expression of Wnt6 was observed in the region where the mammary glands develop. Between the 39- and 42-somite stage, Wnt6 is expressed at a low level ubiquitously in the surface ectoderm. At the 43-somite stage, Wnt6 expression is detected at a low level almost ubiquitously in the surface ectoderm and the underlying dermis and slightly higher in the dorsal part of the embryo (Fig. 2A,E,E′). At this stage, the developing placode 3 highly expresses Wnt6, while expression is only slightly elevated at the mammary line itself and seems to follow the progression toward the limbs in a manner similar to Wnt10b expression. Similar to Wnt10b but with differences in timing and pattern, Wnt6 expression is fragmented along the flank aligned with the somites (Fig. 2A′,B′,C′,D′) and between the 43- and 46-somite stage also is evident in the axillary (Fig. 2A′′,B′′,C′′) and inguinal (Fig. 2A′′′,B′′′,C′′′,D′′′) regions. In one exceptional case, Wnt6 expression was observed in the axillary region as early as at the 39-somite stage (not shown). With high variability among the various developmental stages, Wnt6 expression is either confined to the position of placodes 1 and 5 or is found throughout a large part of the axillary and inguinal regions, respectively, with higher expression at the placodes. At the 45- to 46-somite stage, placodes 1 and 5 are connected to the mammary line by a thin line of Wnt6 expression (Fig. 2B′′,B′′′,C′′,C′′′). However, at the 48-somite stage, Wnt6 expression has disappeared completely from placode 1, which is by then well detectable on a morphologic basis (Fig. 2D′′) and the expression connecting placode 5 with placode 4 has become fainter (Fig. 2D′′′). Of interest, Wnt6 expression is restricted to the placode (number 3) at the 43-somite stage when Wnt10b expression is less defined as a fragmented streak. Conversely, Wnt10b expression defines the placodes (numbers 2 and 4) at the 48-somite stage when Wnt6 expression is weaker and more diffuse (Fig. 2D–D′′′). Vibratome sectioning of Wnt6-probed embryos revealed that Wnt6 expression is highly reminiscent of Wnt10b expression: at the 43-somite stage, Wnt6 is expressed ubiquitously in the dermal and ectodermal region of the skin and is highly elevated in mammary placode 3 (Fig. 2E,E′), whereas it is confined to the surface ectoderm and mammary placode epithelium at the 46-somite stage (Fig. 2F,F′).
Mammary Line Is Discernible as a Multilayering of the Ectoderm at E11.25
We next wanted to determine whether the mammary line as seen by Wnt10b expression is also morphologically detectable. We, therefore, carried out a histologic analysis of the mammary region of E11.75 (47-somite stage) mouse embryos, in which the mammary placodes have been formed and serve as a positional reference in the sections. The surface ectoderm of the area between the two Wnt10b-expressing lines (the mammary line and the dorsal line in Fig. 1) exists of cuboidal epithelial cells, in contrast to the squamous morphology of the epithelium located ventrally and dorsally to this region (Fig. 3A). At the boundary with the squamous ventral ectoderm, above the ventral bud of the somite, the cuboidal epithelium is multilayered in a pseudostratified organization (Fig. 3A,B). Similar results were obtained with E11.25 (43-somite stage) embryos, in which placode number 3 has started to develop. The position of this band of multilayered surface ectoderm corresponds to the region of the mammary line as defined by Wnt10b expression in Figure 1 and demonstrated more clearly in a thin transversal vibratome section of a Wnt10b-probed 47-somite stage embryo (Fig. 3D). The mesenchyme of the dermis underlying this multilayered ectoderm is densely populated. In contrast, in the region of the dorsal line, the mesenchymal cells of the dermis are more sparsely distributed. Whereas the dense mesenchyme is specific for the region of the mammary line, multilayering of the surface ectoderm is also observed in the more rostral region of the dorsal line. However, the dorsal line is never as thick as the mammary line.
Mammary Line in Mouse Is Identified by Wnt10b Expression
Morphologic studies in the past have yielded controversial results concerning the existence of a mammary line in the mouse embryo (reviewed in Veltmaat et al., 2003). By taking a molecular approach, we demonstrate here that the mammary line indeed exists in the mouse. This line is identified as a thin line of Wnt10b expression on each flank of E11.25 (42- to 43-somite stage) mouse embryos, connecting the positions of the prospective placodes 2, 3, and 4. Additional streaks of Wnt10b expression appear slightly later in the axillary and inguinal regions, defining the positions of placodes 1 and 5, respectively. These mammary streaks connect with the mammary line on the flank only at the 46- to 49-somite stage. Additionally, Wnt6 is expressed in the dorsal skin extending ventrally up to the region where the line of Wnt10b expression appears and seems slightly elevated at the level of the mammary line. As Wnt6 expression in the region where the mammary glands develop is much broader than Wnt10b expression and, moreover, not as evident as Wnt10b expression in all developing placodes, we consider only the expression of Wnt10b to be a true marker for the mammary line.
Mammary Gland Development in the Mouse Is Initiated at Three Positions
Wnt10b expression in the mammary region correlates with the apparently random order in which mammary placodes arise in the mouse. The elevation and restriction of Wnt10b expression at the positions of the mammary placodes occurs in the order of placode 3, followed by 4, then 1, and 5 at approximately the same time, and finally 2. This finding corresponds to the order of placode appearance that we have observed previously by electron microscopy and Lef1 expression (Mailleux et al., 2002).
This temporal order of placode appearance is not along an anteroposterior axis, despite what one might expect on the basis of appearance of the mammary rudiments in rabbit. There, the mammary rudiments “bud off” sequentially from the mammary ridge on the flank from an anterior position to progressively more posterior positions (Propper, 1976). Of interest, the first rudiment to bud off from the mammary ridge in rabbit is situated at approximately the same position as placode 3 in the mouse, the first to appear. Furthermore, more anterior to this rudiment exists another rudiment in rabbit, the pectoral mammary gland, which evolves without an apparent connection to the elevated mammary ridge (Propper, 1976). Similarly, the two inguinal mammary rudiments in rat appear without being connected to the elevated mammary ridge at the flank (Henneberg, 1900). This finding suggests that the emergence of the pectoral mammary placode (number 1 and maybe number 2) and the inguinal placode (number 5 and maybe number 4) in mouse from inductive streaks that are initially unrelated to the mammary line may not be uncommon to other mammals. Furthermore, we demonstrate by histologic analysis that the mammary line in mouse is clearly distinct from the neighboring tissue by the multilayering of the epithelium and by the dense dermis underneath. Taken together, our data suggest that, despite the differences in morphologic appearance of the mammary line in mouse and, for example, rabbit, the initiation of mammogenesis may essentially occur in a similar manner in these species.
Fragmentation of the Mammary Line Suggests an Involvement of the Somites in the Induction of the Mammary Line
The mammary line in the mouse embryo is initially fragmented between placode positions 3 and 4 in alignment with the segmentation of the somites and overlying the ventral bud of the somites. This position is in agreement with our previous hypothesis that the dermamyotomal region of the somites emits signals that initiate mammary gland development (Mailleux et al., 2002). This hypothesis was based on (1) the lack of induction of placode numbers 1, 2, 3, and 5 in Fgf10-/- embryos; (2) the lack of Fgf10 expression in the surface ectoderm or mammary rudiments of wild-type littermates until several days beyond the induction of the mammary placodes; (3) the expression of Fgf10 in the dermamyotome at E10.5; and (4) the knowledge that the dermamyotomal region in the ventral bud of the somites gives rise to the dermis (Sengel, 1976), which in turn is instructive for the differentiation of surface ectoderm to mammary epithelium (Sakakura et al., 1987; Cunha and Hom, 1996; reviewed in Veltmaat et al., 2003). Recently, mice expressing transgenic Eda-A1 under the Keratin14 promoter in the surface ectoderm have been reported to have supernumerary mammary glands along a line on the flank at adulthood (Mustonen et al., 2003). It would now be interesting to analyze the exact location of these supernumerary glands and the embryonic stages of mammary gland development in these mice, to gain more insight in the putative interaction between somitic signals and ectodermal response factors.
Wnt10b and Wnt6 Expression Correlates With Ectodermal Cell Fate Change in the Mouse Embryo
Our data on Wnt10b and Wnt6 expression reveal the earliest ectodermal molecular signals associated with murine mammogenesis to date. An ectodermal response to WNT signaling is important during embryonic mammogenesis in the mouse, as previously demonstrated by the absence of mammary glands in E13.5 Lef1-/- embryos (Van Genderen et al., 1994) and in E13.5 embryos expressing transgenic Dkk1 under the Keratin14 promoter in the ectoderm (Andl et al., 2002). Lef1 is expressed in the mammary placodes at the time of their induction (Mailleux et al., 2002). The analysis of mammary placodes in E11.5–E12 Lef1-/- embryos revealed that they are smaller in diameter and invaginate the underlying dermis to a lesser extent than mammary placodes of wild-type littermates (K. Kratochwil, personal communication, and our observations). Together, these data strongly suggest the involvement of an ectodermal response to WNT signaling by means of LEF1 in mammogenesis at the time of placode formation. Furthermore, as the spatiotemporal expression of Wnt10b and Wnt6 in the mammary line coincides with cell fate decisions that locally transform surface ectoderm into mammary epithelium, we hypothesize that WNT10b and WNT6 may be instrumental in this process.
Whole-Mount In Situ Hybridizations on Mouse Embryos
Mice on a C57BL/6 genetic background were intercrossed. Pregnant females were sacrificed to obtain embryos between E10.5 and E12.0 (midday of the day of the vaginal plug was considered E0.5). After finestaging the embryos by counting of the somites, the embryos were fixed overnight in 4% paraformaldehyde, dehydrated, and processed for whole-mount in situ hybridization as described previously (Mailleux et al., 2002). pGEM7-Wnt2a (bp 810-1500 of the coding region of Wnt2a), pGEM3zf-Wnt3a (750 bp 3′-untranslated region (UTR) of Wnt3a), pGEM7zf-Wnt4 (bp 851-1215 of the coding region of Wnt4), pGEM3zf-Wnt5a (bp 1260-1620 of the coding region of Wnt5a), pGEM7zf-Wnt5b (bp 1343-1662 of the 3′-UTR of Wnt5b), pBS-SK-Wnt6 (full-length Wnt6), pGEM3zf-Wnt7b (bp 1183-1485 of the coding region of Wnt7b) (all kindly provided by Dr. A. McMahon), pCR2.1-Wnt2b (kindly provided by Dr. Izpisua Belmonte), pBCG-2-Wnt10a (full-length Wnt10a) and pSK-Wnt10b (full-length Wnt10b; both kindly provided by Dr. K. Kratochwil) were used to generate riboprobes labeled with digoxigenin-dUTP (Roche) according to the manufacturer's protocol. Two or three embryos were used for each of the shown stages. Some embryos were embedded in 5% bovine serum albumin/0.4% gelatin in 0.12 M sodium phosphate buffer, pH 7.4, and cut to 70 μm (Figs. 1, 2) or 35 μm (Fig. 3) thick sections in a vibratome.
C57BL/6 mouse embryos were dissected at around E11.5 and fixed overnight in Carnoy's fixative (60% MeOH, 30% chloroform, 10% acetic acid). After dehydration in EtOH, embryos were paraffin-embedded and cut to 7-μm thick sections, which were stained with hematoxylin/Biebrich Scarlett.
We thank Drs. J. Izpisua Belmonte, K. Kratochwil, and A. McMahon for kindly providing the probes, A. Mailleux for discussion and technical advice, Dr. I Taddei for help with somite counts, and Drs. M. Morgan and R. Mudge for critically reading the manuscript. J.V. was funded by the Fifth Framework of the European Commission, W.v.V. was funded by the Dutch Cancer Fund/KWF, and S.B. was funded by the California Breast Cancer Research program.
- 2002. WNT signals are required for the initiation of hair follicle development. Dev Cell 2: 643–653. , , , .
- 1995. Murine Wnt-11 and Wnt-12 have temporally and spatially restricted expression patterns during embryonic development. Mech Dev 51: 341–350. , , , , , .
- 1996. Role of mesenchymal-epithelial interactions in mammary gland development. J Mammary Gland Biol Neoplasia 1: 21–35. , .
- 1900. Die erste Entwicklung der mammarorgane bei der Ratte. Anat Hefte 13: 1–68. .
- 2000a. The characterization of a mouse mutant that displays abnormal mammary gland development. Mamm Genome 11: 234–237. , .
- 2000b. Mammary gland patterning in the AXB/BXA recombinant inbred strains of mouse. Mech Dev 91: 305–309. , .
- 1945. Abnormalities of the mammae in the house mouse. J Hered 36: 285–288. , .
- 2002. Role of FGF10/FGFR2b signaling during mammary gland development in the mouse embryo. Development 129: 53–60. , , , , , , , , , .
- 2003. Stimulation of ectodermal organ development by Ectodysplasin-A1. Dev Biol 259: 123–136. , , , , , , , .
- 1917. Studies on the mammary gland II. The fetal development of the mammary gland of the female albino rat. Am J Anat 22: 195–223. .
- 1993. Differential expression of the HMG box factors TCF-1 and LEF-1 during murine embryogenesis. Development 118: 439–448. , , , , , , .
- 1976. Modalités et déterminisme du développement embryonnaire de la glande mammaire. Senologia 1: 19–26. .
- 1961. Morphogenesis of the mammary gland. In : KonSK, CowieAT, editors. Milk: the mammary gland and its secretion. New York: Academic Press. p 3–46. .
- 1999. Regulation of mammary gland development by tissue interaction. J Mammary Gland Biol Neoplasia 4: 9–19. , , .
- 1987. Mammary embryogenesis. In: NevilleMC, DanielCW, editors. The mammary gland: development, regulation, and function. New York: Plenum Press. p 37–66. .
- 1976. Morphogenesis of skin. Cambridge: Cambridge University Press. 277 p .
- 1994. Development of several organs that require inductive epithelial-mesenchymal interactions is impaired in LEF-1-deficient mice. Genes Dev 8: 2691–2703. , , , , , , .
- 2003. Mouse embryonic mammogenesis as a model for the molecular regulation of pattern formation. Differentiation 71: 1–17. , , , .