Reconstitution of mammary gland development in vitro: requirement of c-met and c-erbB2 signaling for branching and alveolar morphogenesis

Abstract

We have established a cell culture system that reproduces morphogenic processes in the developing mammary gland. EpH4 mouse mammary epithelial cells cultured in matrigel form branched tubules in the presence of hepatocyte growth factor/scatter factor (HGF/SF), the ligand of the c-met tyrosine kinase receptor. In contrast, alveolar structures are formed in the presence of neuregulin, a ligand of c-erbB tyrosine kinase receptors. These distinct morphogenic responses can also be observed with selected human mammary carcinoma tissue in explant culture. HGF/SF-induced branching was abrogated by the PI3 kinase inhibitors wortmannin and LY294002. In contrast, neuregulin- induced alveolar morphogenesis was inhibited by the MAPK kinase inhibitor PD98059. The c-met-mediated response could also be evoked by transfection of a c-met specific substrate, Gab1, which can activate the PI3 kinase pathway. An activated hybrid receptor that contained the intracellular domain of c-erbB2 receptor suffices to induce alveolar morphogenesis, and was observed in the presence of tyrosine residues Y1028, Y1144, Y1201, and Y1226/27 in the substrate-binding domain of c-erbB2. Our data demonstrate that c-met and c-erbB2 signaling elicit distinct morphogenic programs in mammary epithelial cells: formation of branched tubules relies on a pathway involving PI3 kinase, whereas alveolar morphogenesis requires MAPK kinase.

Figure 1

Effect of HGF/SF on the morphology of EpH4/K6 mammary epithelial cells on matrigel. Cells were cultured in medium containing hormones in the absence (a and d) or in the presence of HGF/SF (b and e). Small aggregates with (arrow) or without lumen were formed in the absence of factor. HGF/SF induced branching tubules with prominent lumina (arrows in b) and end buds (arrowheads in b). Gab1-transfected EpH4/K6 cells exhibited a similar morphogenic response in the absence of HGF/SF (c and f). Top, micrographs using Nomarski optics; bottom, semithin sections of Epon-embedded cultures. Bars: (a) 50 μm; (c) 25 μm.

Figure 2

Statistical analysis of the effects of HGF/SF (B) and neuregulin (C) on the morphology of EpH4/K6 cells on matrigel. Control aggregates in the absence of the factors are analyzed in A. From three sets of experiments, a total of 1,050 structures were examined. The different groups of structures are indicated in the box, i.e., small aggregates without or with lumen (smaller than 50 μm), elongated tubes, and round large alveoli (larger than 50 μm).

Figure 2

Statistical analysis of the effects of HGF/SF (B) and neuregulin (C) on the morphology of EpH4/K6 cells on matrigel. Control aggregates in the absence of the factors are analyzed in A. From three sets of experiments, a total of 1,050 structures were examined. The different groups of structures are indicated in the box, i.e., small aggregates without or with lumen (smaller than 50 μm), elongated tubes, and round large alveoli (larger than 50 μm).

Figure 3

Effect of HGF/SF on the expression of β-casein in EpH4 mammary epithelial cells cultured on matrigel in the presence of lactogenic hormones. (A) Representative autoradiograph of a Northern blot showing the inhibition of β-casein expression by HGF/SF after 6 d of culture (lane b); control in the absence of HGF/SF (lane a). RNA from mammary glands of a pregnant mouse served as a control for β-casein expression (lane c). (B) Time course of expression of β-casein in EpH4 cells over 6 d of culture in the absence (○) or presence of HGF/ SF (•), as analyzed by Northern blotting. Radioactive counts are expressed using a phosphoimager. (C) Concentration dependence on HGF/SF of the inhibition of β-casein expression. Control in the absence of HGF/SF was set to 100%.

Figure 4

Effect of neuregulin on morphogenesis of mammary epithelial cells EpH4/K6 on matrigel. Cells were cultured for 6 d in medium containing hormones in the presence of neuregulin (a and c; for the control in the absence of neuregulin refer to Fig. 1 a). Large alveolar structures were formed which were predominantly built up by a single layer of columnar epithelial cells. Identical structures were observed in EpH4/K6 cells transfected with a trk/c-erbB2 hybrid receptor and stimulated with NGF (b and d). Top, micrograph using Nomarski optics; bottom, semithin sections of Epon-embedded cultures. Bars: (b) 50 μm; (d) 25 μm.

Figure 5

Statistical analysis of NGF-induced morphological responses of EpH4/K6 cells transfected with the hybrid receptors trk/c-erbB2 (A), trk/c-erbB2-P1 (B) or trk/c-erbB2Y1253F (C). For each set of experiments, a total of 450 structures from each of four different stably expressing clones were analyzed. The aggregates are defined as in Fig. 2 (inset).

Figure 5

Statistical analysis of NGF-induced morphological responses of EpH4/K6 cells transfected with the hybrid receptors trk/c-erbB2 (A), trk/c-erbB2-P1 (B) or trk/c-erbB2Y1253F (C). For each set of experiments, a total of 450 structures from each of four different stably expressing clones were analyzed. The aggregates are defined as in Fig. 2 (inset).

Figure 6

Effects of inhibitors of the PI3 kinase (wortmannin and LY294002) and MAPK kinase (PD98059) on the formation of elongated structures (A) and the formation of round, large alveoli (B). Experiments were terminated after 3 d of culture. From two sets of experiments, 350 cell aggregates each were analyzed.

Figure 7

Effect of HGF/SF and neuregulin on the expression of β-casein in mammary epithelial cells EpH4 after their sequential application. (A) A representative Northern blot. (B) Quantification of Northern blot by phosphoimager analysis. Radioactive counts are expressed as PSL. Cells were cultured in medium containing hormones for 4 d in the presence of HGF/SF (lane b) and for an additional 4 d in the absence (lane f) or presence of HGF/SF (lane e) or neuregulin (lane d). Control cells were cultured in medium containing hormones without addition of factors (lanes a and c). Inset in B indicates the respective culture conditions.

Figure 8

Ultrastructural analysis of EpH4/K6 cells grown under control conditions (a) and in the presence of HGF/SF (b) or neuregulin (c). Control aggregates often had small lumina lined by flattened cells containing numerous electron-dense granules (arrows). HGF/SF-induced tubules were lined by cuboidal cells with round nuclei and contained a reduced number of secretory granules. Neuregulin-induced alveoli had columnar epithelial cells with nuclei often elongated along the apical–basal axis and with abundant secretory vesicles. The lumina in most of neuregulin-treated structures were spacious; for better demonstrability, a small alveolus was selected here.

Figure 9

Scanning confocal micrograph of EpH4/K6 aggregates treated with HGF/SF (b and e) or neuregulin (c and f); controls are shown in a and d. Green, nuclei; red, E-cadherin; blue, β-casein. In control- and HGF/SF-treated cultures, E-cadherin is evenly distributed along all cell surfaces (arrows in d and e). Neuregulin-treated cells exhibit a stronger polarity resulting in a predominantly lateral localization of E-cadherin. β-Casein expression is strongly inhibited in HGF/SF-treated cells.

Figure 10

Effect of HGF/SF on morphogenesis of explants of the human mammary carcinoma 3366. Explants were cultured for 8 d in the absence (a and b) or presence of HGF/SF (c, c′, d, and d′). Hematoxylin/eosin-stained paraffin sections of the cultured explants are shown. Bars: (a, c, and c′) 100 μm; (b, d, and d′) 50 μm.

Figure 11

Effect of neuregulin on morphogenesis on explants of human mammary carcinoma 3366. Explants were cultured for 8 d in the absence (a–c) or presence of neuregulin (d–f). Hematoxylin/eosin-stained paraffin sections of explants are shown. Bars: (a and d) 50 μm; (b, c, e, and f) 25 μm.