Mammary ductal morphogenesis requires paracrine activation of stromal EGFR via ADAM17-dependent shedding of epithelial amphiregulin

Abstract

Epithelial-mesenchymal crosstalk is essential for tissue morphogenesis, but incompletely understood. Postnatal mammary gland development requires epidermal growth factor receptor (EGFR) and its ligand amphiregulin (AREG), which generally must be cleaved from its transmembrane form in order to function. As the transmembrane metalloproteinase ADAM17 can process AREG in culture and Adam17(-/-) mice tend to phenocopy Egfr(-/-) mice, we examined the role of each of these molecules in mammary development. Tissue recombination and transplantation studies revealed that EGFR phosphorylation and ductal development occur only when ADAM17 and AREG are expressed on mammary epithelial cells, whereas EGFR is required stromally, and that local AREG administration can rescue Adam17(-/-) transplants. Several EGFR agonists also stimulated Adam17(-/-) mammary organoid growth in culture, but only AREG was expressed abundantly in the developing ductal system in vivo. Thus, ADAM17 plays a crucial role in mammary morphogenesis by releasing AREG from mammary epithelial cells, thereby eliciting paracrine activation of stromal EGFR and reciprocal responses that regulate mammary epithelial development.

Fig. 1

Relative expression of potential interacting genes in 5-week-old mammary glands. Cy5-labeled cDNAs from microdissected TEB- or duct-containing regions and Cy3-labeled cDNAs from distal stroma were hybridized to long-oligonucleotide microarrays. TEB/stroma (Cy5/Cy3) expression ratios and duct/stroma ratios are each shown for six independent experiments (lanes 1–6) and their respective means (M) using the color scale shown, with black indicating no difference in expression, red indicating relative enrichment within the TEB or duct regions and green representing relative enrichment in the pure distal stroma. Thus, intense red indicates a greater than fourfold enrichment in and around TEBs or ducts. Relative expression levels for ADAM10 are shown for two independent oligonucleotides, whereas all other genes were represented by a single oligonucleotide. ^†Genes for which the mean overall fluorescence signal A=0.5[log₂(Cy5) + log₂(Cy3)] was consistently more than eight, the level below which it was difficult distinguish true signal from noise.

Fig. 2

Effect of AREG and EGFR deficiency on postnatal mammary development. (A) Carmine-stained whole mounts of paired wild-type, Areg^−/− (upper left and lower right panels) and recombined mammary transplants grown for 3 weeks in the presence of estradiol pellets. Scale bar: 1 mm. (B) Epithelial areas of 3-week-old, estradiol-stimulated transplants with the indicated epithelial and stromal genotypes (A^−/−, Areg^−/−; WT, wild type). Lines connect paired contralateral transplants from individual host mice and single data points represent unpaired transplants. (C) Whole mounts of wild-type, Egfr^−/− and recombined 3-week-old transplants. Scale bar: 1 mm. (D) Epithelial areas of 3-week-old transplants with the indicated epithelial and stromal genotypes (R^−/−, Egfr^−/−). (E) Histological appearance of paired transplants. Similar alveolar morphologies and accumulation of luminal secretory products (arrowheads) were seen in paired transplants from individual hosts, regardless of the transplant genotype. Scale bar: 50 μm.

Fig. 3

Effect of ADAM17 deficiency on prenatal mammary development. (A) Carmine-stained E18.5 mammary gland/skin whole mounts. Scale bar: 500 μm. (B) Average number of branches in the thoracic (2 and 3) mammary glands of Adam17^−/− embryos and neonates (white circles), Egfr^−/− neonates (white squares) and their respective wild-type littermates (black symbols). (C) Overall lengths of the thoracic mammary ductal trees of Adam17^−/− embryos and neonates, Egfr^−/− neonates and their wild-type littermates.

Fig. 4

Effect of ADAM17 deficiency on postnatal mammary development. (A–F) Whole mounts of paired wild-type (A–E), Adam17^−/− (A′–E′) and surgically recombined (F, F′) mammary glands grown under contralateral kidney capsules (A–C, F) or in surgically cleared fat pads (D, E) for 2 (A), 3 (B, D, F) or 5 (C, E) weeks with (B, C, F) or without (A, D, E) slow-release estradiol pellets. Adam17^−/− outgrowths (arrowheads) were consistently smaller than contralateral wild-type outgrowths, whereas the extent of estradiol-induced alveolar differentiation was similar in each respective host (e.g. B versus B′). (G, H) Paired Adam17^−/− transplants were also grown for 3 weeks in the presence of estradiol and adjacent AREG (a) or placebo (p) pellets. The farthest AREG pellet to yield greater-than-normal growth was ~0.75 mm from the epithelium (double-headed arrow in G), whereas the closest pellet to have no effect on growth as compared to its placebo control was ~1.4 mm from the epithelium (double-headed arrow in H). Scale bar: 500 μm in A′; 2 mm in D, D′; 2.5 mm in E, E′; 1 mm in all other panels and inserts. (I) Overall ductal lengths for newborn wild-type and Adam17^−/− mammary glands, 2-week-old renal transplants, and 3- and 5-week-old cleared fat pad grafts in the absence of added estradiol. Error bars at birth and 2 weeks are hidden by the mean data points. (J) Epithelial areas of paired (connected) and unpaired renal transplants with the indicated epithelial and stromal genotypes (KO, Adam17^−/−; WT, wild type). The non-recombined pairs on the left were grown for 3–6 weeks, whereas all other transplants were grown for 3 weeks in the presence of added estradiol. Bracketed data points are those for which AREG pellets were within ~0.75 mm of the Adam17^−/− epithelium. (K) Mean epithelial areas of non-recombined mammary glands at various times after renal transplantation in the presence of E2.

Fig. 5

Effect of growth factors on genetically defined mammary organoids in Matrigel. Hoffman modulation contrast images were taken after 7 days of growth in the presence of the indicated supplements. Scale bar: 400 μm for KGF-treated organoids; 200 μm for all other panels.

Fig. 6

EGFR phosphorylation in recombined mammary glands. Tissue extracts were immunoprecipitated (IP) with anti-EGFR antibodies, immunoblotted (IB) for phosphotyrosine (pTyr) or phosphorylated EGFR (pEGFR; Y1068), and re-probed for EGFR. Western blots (WB) for keratin 14 were run on the original extracts and re-probed for β-actin. Renal transplants with the indicated epithelial and stromal genotypes (wt, wild type; 17^−/−, Adam17^−/−; A^−/−, Areg^−/−; R^−/−, Egfr^−/−) were harvested 2 weeks after transplantation without exogenous estradiol. Control tissues were harvested from 6-week-old wild-type male (M) and female (F) mice, and Egfr^−/− neonates (R^−/− p0) 15 minutes after intraperitoneal injection of EGF (2.5 mg/kg body weight) or phosphate-buffered saline (PBS).

Fig. 7

Model depicting epithelial-mesenchymal crosstalk and potential modifiers of ADAM17-AREG-EGFR signaling in mammary development.