Notch3 promotes mammary luminal cell specification and forced Notch3 activation can induce mammary tumor formation. However, recent studies suggest a tumor-suppressive role for Notch3. Here, we report on Notch3 expression and functional analysis in the mouse mammary gland. Notch3 is expressed in the luminal compartment throughout mammary gland development, but switches to basal cells with initiation of post-lactational involution. Deletion of Notch3 caused a decrease of Notch activation in luminal cells and diminished luminal progenitors at puberty, as well as reduced alveolar progenitors during pregnancy. Parous Notch3−/− mammary glands developed hyperplasia with accumulation of CD24hiCD49flo cells, some of which progressed to invasive tumors with luminal features. Notch3 deletion abolished Notch activation in basal cells during involution, accompanied by altered apoptosis and reduced brown adipocytes, leading to expansion of parity-identified mammary epithelial cells (PI-MECs). Interestingly, the postpartum microenvironment is required for the stem cell activity of Notch3−/− PI-MECs. Finally, high expression of NOTCH3 is associated with prolonged survival in patients with luminal breast cancer. These results highlight an unexpected tumor-suppressive function for Notch3 in the parous mammary gland through restriction of PI-MEC expansion.

Breast cancer is one of the most commonly diagnosed malignancies and a leading cause of cancer-related death among women worldwide. Epidemiological studies suggest that parity has a complex impact on the risk of breast cancer. Early full-term pregnancy reduces the lifetime risk, yet women also experience a transient increase in breast cancer risk that peaks approximately 5-6 years after giving birth and may persist for up to three decades (Albrektsen et al., 2005; Chie et al., 2000; Lambe et al., 1994; Liu et al., 2002). The paradoxically opposing effects of parity have been attributed to differentiation of mammary stem cells (MaSC) and the proliferation of precancerous cells, respectively, following pregnancy-associated hormonal changes (Polyak, 2006). Pregnancy also leads to a transient expansion of the MaSC compartment, likely mediated through paracrine signaling from RANK ligand (Asselin-Labat et al., 2010). In addition to the impact on the mammary epithelial cells, pregnancy, lactation and post-lactational involution profoundly alter the mammary stromal compartment, which may promote initiation and progression of postpartum breast cancer. The involuting mammary gland, which undergoes tissue remodeling with leukocyte infiltration and adipocyte repopulation, has been shown to accelerate tumor growth and metastasis (Guo et al., 2017; Lyons et al., 2014, 2011; Martinson et al., 2015; Schedin et al., 2007). Further understanding of parity-related alterations may lead to identification of preventative therapies for postpartum breast cancer.

A parous mammary gland is different from a nulliparous mammary gland, resulting in part from the formation of a new mammary epithelial cell population during pregnancy (Wagner et al., 2002). These parity-identified mammary epithelial cells (PI-MECs) are defined as a subset of pregnancy hormone-responsive cells that activate the promoter of late milk protein genes during pregnancy and lactation and bypass programmed cell death during post-lactational remodeling of the gland (Matulka et al., 2007). PI-MECs are able to proliferate and produce new secretory acini in subsequent pregnancies (Wagner et al., 2002). Virtually all PI-MECs express CD24, with the vast majority co-expressing CD49f (also known as Itga6), and a portion of PI-MECs are found within the MaSC-enriched CD24+CD49fhi mammary epithelial cell fraction (Matulka et al., 2007). Indeed, isolated PI-MECs were able to form mammospheres in vitro, generate both luminal and myoepithelial lineages to establish a fully functional mammary gland upon transplantation, and were able to self-renew over several transplant generations (Boulanger et al., 2005; Matulka et al., 2007). Interestingly, PI-MECs serve as targets of MMTV-neu/ErbB2-induced tumorigenesis (Henry et al., 2004; Jeselsohn et al., 2010).

Notch, a conserved signaling pathway, plays an important role in regulation of mammary epithelial differentiation and proliferation. In vitro studies using primitive cells from normal human mammary tissue showed activation of NOTCH3 to be essential for the restriction of bipotent progenitors to the luminal lineage (Raouf et al., 2008). In mice, Notch3 is expressed in a highly clonogenic and transiently quiescent luminal progenitor cell population that is capable of surviving multiple pregnancies. Also, Notch3 restricts the proliferation and consequent clonal expansion of these cells (Lafkas et al., 2013). Pregnant mice expressing an activated intracellular form of Notch3 (MMTV-Notch3ICD) show expansion of a premalignant luminal progenitor population (Ling et al., 2013). Following parity, these mice develop luminal-subtype mammary tumors in a cyclin D1-dependent manner (Ling et al., 2013). Here, we used Notch3β-geo mutant mice to define its function in mammary epithelium. Consistent with its function in restricting expansion of PI-MECs, we identified a tumor-suppressive role for Notch3 in the postpartum mammary gland. In addition, we tested for NOTCH3 expression in association with survival in human breast cancer. The results corroborate our finding in mice that Notch3 suppresses luminal breast cancer.

Deletion of Notch3 caused decreased Notch-dependent transcription in luminal epithelium associated with reduced accumulation of luminal progenitors in the pubescent mammary gland

We used a Notch3β-geo knock-in allele, harboring a transmembrane-anchored β-geo fused in-frame with sequences coding for the first 21 EGF repeats of Notch3 (Xu et al., 2010), to define expression and functions of Notch3 in the mouse mammary gland. X-gal staining in pubescent Notch3β-geo/+ mice showed robust expression in body cells of the terminal end buds (TEBs) and luminal cells of mature ducts (Fig. 1A). To determine alterations in Notch signaling associated with Notch3 deletion, we crossed a Transgenic Notch Reporter (TNR), containing an eGFP expression cassette under the regulation of an artificial Notch-responsive promoter with multiple RBPJκ-binding sites (Duncan et al., 2005), into Notch3β-geo/β-geo (hereinafter referred to as Notch3−/−) mice. Lineage-depleted mammary epithelial cells isolated from TNR and Notch3−/−;TNR mice were stained for surface markers CD24, CD49f, CD61 (Itgb3) and Sca1 (Ly6a) to analyze Notch signaling in various epithelial subsets of the mammary gland. In TNR mice, GFP expression was detected in ∼10% of CD24+CD49fhi cells, a population known to contain MaSC (Stingl et al., 2006), and in nearly 30% of CD24hiCD49flo cells, a population enriched for lobule progenitors (Jeselsohn et al., 2010). Notch3−/−;TNR mice showed a reduced number of GFP+ cells, especially within the CD24hiCD49flo population, suggesting that deletion of Notch3 causes decreased Notch-dependent transcription in lobule progenitors (Fig. 1B,C). The Notch3−/−;TNR mutants also had a significantly reduced number of luminal progenitors (CD24hiCD49floCD61+) (Asselin-Labat et al., 2007; Visvader and Stingl, 2014) compared with TNR mice (Fig. 1B,C). Of note, GFP expression was detected in ∼20% of the CD24hiCD49flo cells in the Notch3−/−;TNR mice, suggesting that other Notch receptors are still active in these cells. Therefore, Notch3 receptor may activate expression of a unique set of Notch target genes that are important for the luminal cell fate determination during mammary gland maturation. Anti-GFP immunostaining confirmed a reduced number of GFP-expressing cells within the TEB body cell compartment, as well as in luminal cells of mature ducts in Notch3−/−;TNR mammary glands (Fig. 1D,E). Although Notch3 is expressed predominantly in the luminal cells, weak Notch3 expression is noted in a few basal cells. Notch3−/−;TNR mice also showed a slightly decreased percentage of GFP+ cells in CD24+CD49fhi population (Fig. 1B) and a reduced number of GFP+ TEB cap cells (Fig. 1D), suggesting that Notch3 may be activated in a subset of basal cells. Interestingly, Notch3 mutants exhibited increased keratin 14 (myoepithelial cell marker) expression in TEB cap cells but decreased keratin 8 (luminal cell marker) expression in mature ducts, suggesting skewed differentiation of bipotent progenitors towards myoepithelial lineage (Fig. 1F,H). However, ductal morphogenesis in these mice appeared to be indistinguishable from that seen in wild-type mice (Fig. 1G,I).

Fig. 1.

The pubescent Notch3−/− mammary gland exhibits decreased Notch signaling in the luminal compartment associated with diminished CD24hiCD49floCD61+ luminal progenitors. (A) X-Gal staining of mammary tissue from a 6-week-old Notch3β-geo/+ mouse showing the TEB (left) and mature duct (right). (B) Representative flow cytometry analysis of lineage-depleted mammary cells isolated from TNR and Notch3−/−;TNR mice at 7-8 weeks of age. (C) Quantitation of GFP-positive cells in CD24hiCD49flo population and comparison of the CD24hiCD49floCD61+ subpopulation. (D) Anti-GFP immunostaining in the TEBs (upper) and mature ducts (lower) of TNR and Notch3−/−;TNR mice. (E) Percentage of TEB cells with strong GFP staining and percentage of ductal cells with strong GFP staining in TNR and Notch3−/−;TNR mice. (F) Keratin 8 (K8) and keratin 14 (K14) immunofluorescence staining in the TEBs (upper) and mature ducts (lower) of wild-type and Notch3−/− mice. (G) Representative whole-mount mammary glands from 6-week-old wild-type and Notch3−/− virgins. (H) Percentage of K14+ cap cells and percentage of K8+ ductal cells in the wild-type and Notch3−/− mammary glands. (I) Relative length of ductal elongation and relative density of branching points in the wild-type and Notch3−/− mammary glands. Scale bars: 50 µm in A,D,F; 5 mm in G. Data are mean±s.e.m. *P<0.05, **P<0.01, ***P<0.001 (paired two-tailed Student's t-test).

Fig. 1.

The pubescent Notch3−/− mammary gland exhibits decreased Notch signaling in the luminal compartment associated with diminished CD24hiCD49floCD61+ luminal progenitors. (A) X-Gal staining of mammary tissue from a 6-week-old Notch3β-geo/+ mouse showing the TEB (left) and mature duct (right). (B) Representative flow cytometry analysis of lineage-depleted mammary cells isolated from TNR and Notch3−/−;TNR mice at 7-8 weeks of age. (C) Quantitation of GFP-positive cells in CD24hiCD49flo population and comparison of the CD24hiCD49floCD61+ subpopulation. (D) Anti-GFP immunostaining in the TEBs (upper) and mature ducts (lower) of TNR and Notch3−/−;TNR mice. (E) Percentage of TEB cells with strong GFP staining and percentage of ductal cells with strong GFP staining in TNR and Notch3−/−;TNR mice. (F) Keratin 8 (K8) and keratin 14 (K14) immunofluorescence staining in the TEBs (upper) and mature ducts (lower) of wild-type and Notch3−/− mice. (G) Representative whole-mount mammary glands from 6-week-old wild-type and Notch3−/− virgins. (H) Percentage of K14+ cap cells and percentage of K8+ ductal cells in the wild-type and Notch3−/− mammary glands. (I) Relative length of ductal elongation and relative density of branching points in the wild-type and Notch3−/− mammary glands. Scale bars: 50 µm in A,D,F; 5 mm in G. Data are mean±s.e.m. *P<0.05, **P<0.01, ***P<0.001 (paired two-tailed Student's t-test).

Notch3 deletion mice exhibit a reduced number of alveolar progenitors and show altered mammary alveolar morphogenesis during pregnancy

Notch3+ cells contribute to the formation of alveolar buds during pregnancy (Lafkas et al., 2013). Indeed, X-gal staining of Notch3β-geo/+ mammary glands at pregnancy day 17.5 showed that Notch3 expression is high in alveolar cells but weak in ducts (Fig. S1A). Consistent with a role for Notch3 in this context, fewer CD24hiCD49floCD61+Sca1 cells, a population enriched for alveolar progenitors (Visvader and Stingl, 2014), were seen in Notch3 mutants at mid-pregnancy (Fig. S1B). Also, whole-mount mammary gland preparation at late-pregnancy showed that Notch3−/− mutants had smaller alveoli compared with wild-type mice (Fig. S1C,D). Expressions of keratin 8 and 14 (luminal and myoepithelial markers, respectively) appeared to be normal in Notch3 mutant glands (Fig. S1C). Thus, deletion of Notch3 had no effect on luminal versus myoepithelial cell fate specification during pregnancy; however, it did lead to a reduced number of alveolar progenitors with defective alveolar morphogenesis.

Deletion of Notch3 led to excess accumulation of PI-MECs accompanied by a decreased number of brown adipocytes

Despite the slight defect in alveolar morphogenesis during pregnancy, pups of Notch3−/− mice had similar size at weaning compared with pups from wild-type mice (Fig. S1E), suggesting normal lactation in Notch3 mutants. Notch3 expression remained high in mammary alveolar cells during lactation (Fig. 2A). Interestingly, forced weaning of pups at lactation day 10 caused a drastic downregulation of Notch3 in secretory alveolar cells (Fig. 2B,U). By post-weaning involution day 2, Notch3 expression was mostly restricted to myoepithelial cells (Fig. 2B). Expressions of the Notch ligands Jag1 and Dll1 were also detected in myoepithelial cells, and Jag1 expression was upregulated during involution (Fig. 2C-F,U). Consistent with the expression data, TNR mice displayed GFP+ myoepithelial cells at the onset of involution (Fig. 2G), whereas Notch3−/−;TNR mice had no or very weak GFP expression at this stage (Fig. 2H,V), indicating that deletion of Notch3 abolished Notch signaling in myoepithelial cells of the involuting alveoli. Using whole-mount analysis, post-lactational involution appeared to proceed normally in Notch3−/− mice (Fig. S2). However, the number of TUNEL+ apoptotic cells was increased in mutant glands at involution day 1 but decreased at involution days 2 and 3 (Fig. 2I-N,W). In addition, Notch3−/− mammary glands showed a significantly increased number of p53+ alveolar cells compared with the wild type at involution day 1 (Fig. 2O,P,X). It has been shown that p53 (Trp53) mRNA is upregulated at the start of involution (Strange et al., 1992), and that p53 participates in the first stage of involution initiated by the epithelium itself, but does not affect the second phase which involves stromal proteases (Jerry et al., 1998). These results suggest that deletion of Notch3 caused accelerated apoptosis at the start of involution but reduced cell death in the second phase of involution and remodeling. PI-MECs, a subset of alveolar cells that survive involution and post-lactation remodeling, exhibit stem cell characteristics and can serve as cellular targets for transformation (Henry et al., 2004; Matulka et al., 2007). Reduced cell death in the second phase of involution may lead to increased PI-MECs in Notch3−/− mice. We next analyzed PI-MEC levels by labeling mammary epithelium during pregnancy using R26YFP;WAP-Cre and then by staining for YFP at involution day 28 post lactation. Indeed, Notch3−/−;R26YFP;WAP-Cre mice contained a significantly increased number of YFP+ cells compared with R26YFP;WAP-Cre mice, suggesting that deletion of Notch3 caused an increase in PI-MEC accumulation (Fig. 2Q,R,Y). Remodeling of the mammary gland during post-lactational involution is accompanied by adipocyte repopulation. Interestingly, brown adipocytes, which stain positive for Ucp1, emerged in close vicinity to alveolar epithelial cells at the start of involution in wild-type gland but not in Notch3−/− mice (Fig. 2S,T). Brown fat and Ucp1 expression in the mammary gland were previously reported to be highest during prepuberty, decreased upon puberty and almost undetectable in the adult gland (Gouon-Evans and Pollard, 2002). We noted a rapid decrease in Ucp1 staining following the initial stage of involution (data not shown). However, using western blot, Ucp1 expression was still detectable in the mammary gland 3 months after completion of post-lactational involution. In this context, parous Notch3−/− mice exhibited lower levels of Ucp1 than their wild-type counterparts (Fig. 2Z). Taken together, deletion of Notch3 led to excess PI-MECs accumulation with an accompanying decrease in brown adipocytes.

Fig. 2.

Decreased Notch activity in myoepithelial cells, altered apoptosis and defective brown adipocyte differentiation during involution, leading to increased PI-MECs in Notch3−/− mammary glands. (A-F) X-Gal staining of mammary tissues from Notch3β-geo/+, Jag1β-geo/+, and Dll1lacZ/+ mice at lactation day 10 and involution day 2. (G,H) Anti-GFP immunostaining in the mammary tissues of TNR and Notch3−/−;TNR mice at involution day 1. Insets in B and G are higher magnifications. (I-N) Representative photomicrographs of TUNEL assays in the mammary tissues of wild-type (WT) and Notch3−/− mice at involution days 1, 2 and 3. (O,P) Anti-p53 immunostaining in mammary tissues of wild-type and Notch3−/− mice at involution day 1. (Q,R) Anti-YFP immunostaining in mammary tissues from R26YFP;WAP-Cre and Notch3−/−;R26YFP;WAP-Cre mice at involution day 28. (S,T) Immunostaining for Ucp1 in the wild-type and Notch3−/− mice at involution day 1. Arrowheads indicate positive staining. (U) Percentage of lacZ+ alveolar cells in Notch3β-Geo/+ mice and relative number of lacZ+ cells in Jag1β-Geo/+ mice at lactation day 10 and involution day 2. (V) Relative number of GFP+ cells in TNR and Notch3−/−;TNR mice at involution day 1. (W) Quantitation of TUNEL positive cells. (X) Percentage of p53+ alveolar cells in the wild-type and Notch3−/− mice at involution day 1. (Y) Quantitation of YFP-positive mammary epithelial cells in R26YFP;WAP-Cre and Notch3−/−;R26YFP;WAP-Cre mice at involution day 28. (Z) Western blot analysis of E-cadherin and Ucp1 in mammary tissues from wild-type and Notch3−/− mice, and relative levels of Ucp1 normalized with the level of β-actin in parous animals. Scale bars: 50 µm. Data are mean±s.e.m. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001 (unpaired two-tailed Student's t-test).

Fig. 2.

Decreased Notch activity in myoepithelial cells, altered apoptosis and defective brown adipocyte differentiation during involution, leading to increased PI-MECs in Notch3−/− mammary glands. (A-F) X-Gal staining of mammary tissues from Notch3β-geo/+, Jag1β-geo/+, and Dll1lacZ/+ mice at lactation day 10 and involution day 2. (G,H) Anti-GFP immunostaining in the mammary tissues of TNR and Notch3−/−;TNR mice at involution day 1. Insets in B and G are higher magnifications. (I-N) Representative photomicrographs of TUNEL assays in the mammary tissues of wild-type (WT) and Notch3−/− mice at involution days 1, 2 and 3. (O,P) Anti-p53 immunostaining in mammary tissues of wild-type and Notch3−/− mice at involution day 1. (Q,R) Anti-YFP immunostaining in mammary tissues from R26YFP;WAP-Cre and Notch3−/−;R26YFP;WAP-Cre mice at involution day 28. (S,T) Immunostaining for Ucp1 in the wild-type and Notch3−/− mice at involution day 1. Arrowheads indicate positive staining. (U) Percentage of lacZ+ alveolar cells in Notch3β-Geo/+ mice and relative number of lacZ+ cells in Jag1β-Geo/+ mice at lactation day 10 and involution day 2. (V) Relative number of GFP+ cells in TNR and Notch3−/−;TNR mice at involution day 1. (W) Quantitation of TUNEL positive cells. (X) Percentage of p53+ alveolar cells in the wild-type and Notch3−/− mice at involution day 1. (Y) Quantitation of YFP-positive mammary epithelial cells in R26YFP;WAP-Cre and Notch3−/−;R26YFP;WAP-Cre mice at involution day 28. (Z) Western blot analysis of E-cadherin and Ucp1 in mammary tissues from wild-type and Notch3−/− mice, and relative levels of Ucp1 normalized with the level of β-actin in parous animals. Scale bars: 50 µm. Data are mean±s.e.m. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001 (unpaired two-tailed Student's t-test).

Parous Notch3 null mice show mammary hyperplasia with mammary tumor formation

Given expansion of the PI-MEC compartment in postpartum Notch3 mutant mice, we tested for mammary tumor development in these mice. Using whole-mount analysis, mammary glands from parous Notch3−/− mutants older than 10 months of age showed severe hyperplasia (n=12). In contrast, few parous mammary glands from wild-type or Notch3+/− mice of the same age showed hyperplasia (Fig. 3A,B). In agreement with mammary epithelial hyperplasia, Notch3 mutant glands showed drastically elevated levels of E-cadherin compared with wild-type mice (Fig. 2Z). In addition, Notch3 mutant glands exhibited expansion of the lobule progenitor-enriched CD24hiCD49flo population (Fig. 3C). Multiparous Notch3−/− mice of advanced age developed tumorous lesions arising from hyperplastic alveoli or from inside dilated ducts. Some of these progressed to invasive mammary tumors, whereas age-matched nulliparous Notch3−/− mice were hyperplasia and tumor free (Fig. 3D). Among animals under tumor watch, 10 out of 33 parous Notch3−/− mice, 0 out of 13 nulliparous Notch3−/− mice and 0 out of 37 parous controls (including wild type and Notch3+/−) developed full-blown mammary tumors (Fig. 3E). Of note, 3 out of 10 Notch3−/− mice that developed mammary tumors showed metastasis to the lung (see below, Fig. 4U-Y). Interestingly, X-gal staining was seen in benign tissues and in tumor-associated stroma from Notch3β-geo/β-geo glands, but not within tumor cells (Fig. 3F). We also used flow cytometry to compare tumor-free mammary glands with tumors isolated from Notch3−/− mice, and found that >70% of tumor cells were CD24hiCD49flo compared with ∼20% in epithelial cells in tumor-free glands (Fig. 3G). These data suggest that loss of Notch3 caused inhibition of luminal/alveolar differentiation and accumulation of lobule progenitors (CD24hiCD49flo) in the parous gland, ultimately leading to malignant transformation of these cells.

Fig. 3.

Mammary hyperplasia and tumor development in parous Notch3−/− mice. (A) Whole-mount mammary glands from 10-month-old parous Notch3+/− and Notch3−/− littermates. (B) Whole-mount mammary glands of parous wild-type (WT) and Notch3−/− mice at 13 months of age. (C) CD24/CD49f flow cytometry analysis and quantitation of CD24hiCD49flo cells among linage-depleted mammary cells from 8-month-old parous wild-type and Notch3−/− mice. (D) Representative photomicrographs of whole-mount mammary glands from nulliparous and multiparous Notch3−/− mice at 14-16 months of age. Panels on right show magnification of boxed areas in bottom left panel. (E) Kaplan–Meier mammary tumor-free survival curve for parous and nulliparous Notch3−/− mice and parous control (wild type and Notch3+/−) mice. (F) X-Gal staining of non-tumorous mammary tissue and mammary tumor from a Notch3β-geo/β-geo mouse. (G) Representative CD24/CD49f flow cytometry analysis in linage-depleted non-tumorous mammary cells and tumor cells from parous Notch3−/− mice. Scale bars: 50 µm. LN, lymph node. Data are mean±s.e.m. *P<0.05 (paired two-tailed Student's t-test).

Fig. 3.

Mammary hyperplasia and tumor development in parous Notch3−/− mice. (A) Whole-mount mammary glands from 10-month-old parous Notch3+/− and Notch3−/− littermates. (B) Whole-mount mammary glands of parous wild-type (WT) and Notch3−/− mice at 13 months of age. (C) CD24/CD49f flow cytometry analysis and quantitation of CD24hiCD49flo cells among linage-depleted mammary cells from 8-month-old parous wild-type and Notch3−/− mice. (D) Representative photomicrographs of whole-mount mammary glands from nulliparous and multiparous Notch3−/− mice at 14-16 months of age. Panels on right show magnification of boxed areas in bottom left panel. (E) Kaplan–Meier mammary tumor-free survival curve for parous and nulliparous Notch3−/− mice and parous control (wild type and Notch3+/−) mice. (F) X-Gal staining of non-tumorous mammary tissue and mammary tumor from a Notch3β-geo/β-geo mouse. (G) Representative CD24/CD49f flow cytometry analysis in linage-depleted non-tumorous mammary cells and tumor cells from parous Notch3−/− mice. Scale bars: 50 µm. LN, lymph node. Data are mean±s.e.m. *P<0.05 (paired two-tailed Student's t-test).

Fig. 4.

Histological and immunohistochemical analyses of mammary tumors in parous Notch3−/− mice. (A-D) Representative histology of Notch3−/− mammary tumors showing solid sheets (A), glandular (B) and papillary (C) patterns, and squamous differentiation (D). (E-H) Immunostaining for ERα in Notch3−/− mammary tumors. (I-L) Immunostaining for cyclin D1 (CCND1) in Notch3−/− mammary tumors. (M-R) Immunofluorescence staining for keratin 8 (K8) and keratin 14 (K14) in Notch3−/− mammary tumors. (S,T) Immunostaining for K10 and EGFR in Notch3−/− mammary tumors with squamous differentiation. (U) Gross pathology of the lung with metastasis (arrowheads). (V) Histology of the tumor metastasized to the lung. (W-Y) Immunostaining for K8, K14 and SPC in lung metastasis (boxed area in V). Scale bars: 50 µm.

Fig. 4.

Histological and immunohistochemical analyses of mammary tumors in parous Notch3−/− mice. (A-D) Representative histology of Notch3−/− mammary tumors showing solid sheets (A), glandular (B) and papillary (C) patterns, and squamous differentiation (D). (E-H) Immunostaining for ERα in Notch3−/− mammary tumors. (I-L) Immunostaining for cyclin D1 (CCND1) in Notch3−/− mammary tumors. (M-R) Immunofluorescence staining for keratin 8 (K8) and keratin 14 (K14) in Notch3−/− mammary tumors. (S,T) Immunostaining for K10 and EGFR in Notch3−/− mammary tumors with squamous differentiation. (U) Gross pathology of the lung with metastasis (arrowheads). (V) Histology of the tumor metastasized to the lung. (W-Y) Immunostaining for K8, K14 and SPC in lung metastasis (boxed area in V). Scale bars: 50 µm.

Mammary tumors in Notch3 deletion mice exhibit luminal characteristics

Histological analysis of mammary tumors in Notch3−/− mice identified solid sheets of neoplastic cells (Fig. 4A), as well as glandular (Fig. 4B) and papillary (Fig. 4C) patterns with occasional squamous differentiation (Fig. 4D). The majority (6/9) of Notch3−/− mammary tumors were estrogen receptor (ER)-positive (Fig. 4E-H). Given that deletion of Notch3 caused expansion of PI-MECs and that cyclin D1 activity was shown to be required for PI-MEC self-renewal and activity (Jeselsohn et al., 2010), we speculated that cyclin D1 may be essential in these tumors. Indeed, all Notch3−/− mammary tumors showed high level of cyclin D1 expression (Fig. 4I-L). Also, the vast majority of Notch3−/− mammary tumor cells stained positive for the luminal marker keratin 8, with a small subset showing co-expression of keratin 8 and keratin 14, a basal marker (Fig. 4M-O). In a few tumors, almost all cells co-expressed keratin 8 and keratin 14 (Fig. 4P-R), suggesting that such tumors may have originated from a bipotent progenitor. Tumors with squamous differentiation stained positive for keratin 10 as well as for EGFR, which is expressed in human breast tumors with squamous differentiation (Bossuyt et al., 2005) (Fig. 4S,T). As with primary tumors, lung metastasis in these mice stained positive for keratin 8, with a small subset of cells expressing keratin 14. Consistent with the mammary tumor origin of these lesions, these lesions did not express surfactant protein C (SPC), a marker for lung adenocarcinoma (Fig. 4W-Y). Thus, loss of Notch3 induced luminal subtype mammary tumors with the potential for metastatic dissemination to the lung.

Expression from the Notch3 locus is repressed by Wnt signaling through β-catenin binding and maintenance of a repressed chromatin state. This effect functions as part of Wnt-mediated suppression of luminal/alveolar differentiation in MaSC-enriched basal cells (Gu et al., 2013). In agreement with previous findings, X-gal staining was not seen in tumor cells nor in benign mammary ductal cells from Notch3β-geo/+;MMTV-Wnt1 mice (Fig. S3A). Deletion of Notch3 had no effect on Wnt-induced mammary tumor development (Fig. S3B), further supporting the idea that repression of Notch3 occurs downstream of Wnt activation. Although both have lost Notch3 expression, mammary tumors in Notch3−/− mice have a luminal phenotype, whereas MMTV-Wnt1 lesions are mostly keratin 14+ cells (Fig. S3C). Different tumor phenotypes may be attributed to distinct cells-of-origin. MMTV-Wnt1 tumors are thought to originate from CD24medCD49fhi ductal progenitors (Jeselsohn et al., 2010), whereas Notch3−/− tumors likely arise through transformation of CD24hiCD49flo lobule progenitors.

The postpartum microenvironment is required for repopulating activity of mammary epithelial cells isolated from parous Notch3−/− mice

The microenvironment is emerging as an important regulator of MaSC (Fu et al., 2020), and the postpartum mammary gland microenvironment has been shown to mediate breast cancer progression (Schedin et al., 2007). For this reason, we tested whether postpartum stroma plays a role in regulation of mammary epithelial specification from PI-MECs. Specifically, we injected mammary epithelial cells from parous Notch3β-geo/β-geo mice into mammary fat pads of nulliparous wild-type mice or age-matched wild-type mice at involution day 21 (all on an FVB background). Whole-mount X-gal staining at 8 weeks post-injection showed incorporation of Notch3β-geo/β-geo PI-MECs into mammary glands of the postpartum but not nulliparous host mice (in three independent experiments) (Fig. 5A,B). Interestingly, the vast majority of X-gal+ cells were found in alveolar buds, indicating that Notch3β-geo/β-geo PI-MECs contribute primarily to these structures. Alveolar buds in the mouse mammary glands are somewhat related to terminal ductal lobular units in humans, the primary anatomical source of most breast cancers. For comparison, both nulliparous and parous (3 months post-weaning) Notch3β-geo/β-geo mice showed X-gal staining throughout mammary ducts (Fig. 5C), and Notch3 protein levels were similar in age-matched nulliparous and parous wild-type mice (Fig. 5D), indicating that parity does not alter Notch3 expression in the mammary gland. Thus, negative X-gal staining in the Notch3β-geo/β-geo cell-injected nulliparous hosts is most likely due to failed incorporation of injected cells into the host glands, rather than silencing of the Notch3 locus in the nulliparous microenvironment. These results suggest that postpartum microenvironment may affect PI-MEC stem cell activity.

Fig. 5.

The postpartum microenvironment is required for stem cell activity of mammary epithelial cells isolated from parous Notch3−/− mice. (A) Whole-mount X-gal staining of a nulliparous wild-type (WT) mammary gland harvested at 8 weeks post transplantation of mammary epithelial cells isolated from a parous Notch3β-Geo/β-Geo donor. (B) Whole-mount X-gal staining of a parous wild-type mammary gland 8 weeks after injection of parous Notch3β-Geo/β-Geo mammary epithelial cells at involution day 21. X-Gal staining in A and B were counterstained with Hematoxylin. (C) Whole-mount X-gal staining of mammary glands from nulliparous and parous Notch3β-Geo/β-Geo mice. (D) Western blot analysis for Notch3 in mammary tissues from age-matched nulliparous and parous wild-type mice. Scale bars: 100 µm.

Fig. 5.

The postpartum microenvironment is required for stem cell activity of mammary epithelial cells isolated from parous Notch3−/− mice. (A) Whole-mount X-gal staining of a nulliparous wild-type (WT) mammary gland harvested at 8 weeks post transplantation of mammary epithelial cells isolated from a parous Notch3β-Geo/β-Geo donor. (B) Whole-mount X-gal staining of a parous wild-type mammary gland 8 weeks after injection of parous Notch3β-Geo/β-Geo mammary epithelial cells at involution day 21. X-Gal staining in A and B were counterstained with Hematoxylin. (C) Whole-mount X-gal staining of mammary glands from nulliparous and parous Notch3β-Geo/β-Geo mice. (D) Western blot analysis for Notch3 in mammary tissues from age-matched nulliparous and parous wild-type mice. Scale bars: 100 µm.

High NOTCH3 expression is associated with better survival in patients with luminal subtype breast cancer

As deletion of Notch3 caused hyperplasia and luminal-like mammary tumors in multiparous mice, we analyzed NOTCH3 expression in human breast cancer. Survival analysis using publicly available patient data revealed a significant association between high level NOTCH3 gene expression and prolonged relapse-free survival in luminal A, luminal B and HER2-positive tumors, but not in basal-like disease. High NOTCH3 expression also correlates with better overall survival in luminal A subtype (Fig. 6A). As noted above, postpartum microenvironment is required for stem cell activity of Notch3−/− PI-MECs. Notch3 mutants had decreased brown adipocytes during post-lactational involution and lower Ucp1 expression compared with wild-type mice. Interestingly, UCP1 was recently found to inhibit tumor progression through the suppression of the ALDH-positive breast cancer stem cell population in basal-like breast cancer (Zhang et al., 2020). Therefore, we analyzed UCP1 expression related to patient survival in human breast cancer datasets. Indeed, UCP1 expression was positively related to relapse-free survival in all four subtypes as well as overall survival in luminal A subtype (Fig. 6B). Although patient parity status is unknown, these data suggest that NOTCH3 and UCP1 play an important role in human breast cancer, especially in the luminal subtypes of the disease.

Fig. 6.

High expressions of NOTCH3 and UCP1 are associated with prolonged survival in patients with luminal breast cancer. (A) Survival analyses in breast cancer patients using univariate Cox regression and Kaplan–Meier methods. Relapse-free survival rates in patients with high NOTCH3 expression are significantly higher than those in patients with low NOTCH3 expression in all except the basal subtype of breast cancer. Overall survival rate in patients with high NOTCH3 expression is significantly higher than that in patients with low NOTCH3 expression in luminal A subtype. (B) Relapse-free survival rates in patients with high UCP1 expression are significantly higher than those in patients with low UCP1 expression in all breast cancer subtypes. Overall survival rate in patients with high UCP1 is significantly higher than that in patients with low UCP1 in luminal A subtype.

Fig. 6.

High expressions of NOTCH3 and UCP1 are associated with prolonged survival in patients with luminal breast cancer. (A) Survival analyses in breast cancer patients using univariate Cox regression and Kaplan–Meier methods. Relapse-free survival rates in patients with high NOTCH3 expression are significantly higher than those in patients with low NOTCH3 expression in all except the basal subtype of breast cancer. Overall survival rate in patients with high NOTCH3 expression is significantly higher than that in patients with low NOTCH3 expression in luminal A subtype. (B) Relapse-free survival rates in patients with high UCP1 expression are significantly higher than those in patients with low UCP1 expression in all breast cancer subtypes. Overall survival rate in patients with high UCP1 is significantly higher than that in patients with low UCP1 in luminal A subtype.

Signaling through Notch receptors plays a crucial role in the regulation of the mammary epithelial hierarchy. Notch3 mRNA expression is highest among Notch receptor genes in the mouse mammary gland (Raafat et al., 2011). Subsequent analysis found that Notch3 is expressed in a highly clonogenic but quiescent progenitor population that gives rise to luminal cells during mammary gland development. These cells can survive multiple cycles of pregnancy and involution, and Notch3 activation restricts proliferation and clonal expansion of these cells (Lafkas et al., 2013). Interestingly, mice expressing an activated form of Notch3 (MMTV-Notch3ICD) show expansion of premalignant CD24+CD29lo luminal progenitors during pregnancy, ultimately leading to luminal mammary tumors in parous mice (Ling et al., 2013). In this paper, we show that Notch3 expression is predominantly restricted to luminal compartment cells of the mammary gland, throughout development, except briefly during early-stage involution, when it is expressed in basal cells. In agreement with a role for Notch3 in promoting luminal lineage specification (Lafkas et al., 2013; Raouf et al., 2008), we found that deletion of Notch3 results in decreased number of common luminal progenitors (CD24hiCD49floCD61+) at puberty and early alveolar progenitors (CD24hiCD49floCD61+Sca1) at mid-pregnancy. To our surprise, parous Notch3−/− mice showed expansion of the CD24hiCD49flo population, leading to mammary hyperplasia with mammary tumor formation. It appears to be paradoxical that overexpression and loss of Notch3 both result in expansion of luminal progenitors and mammary tumors. Notably, expansion of CD24+CD29lo luminal progenitors started from pregnancy in MMTV-Notch3ICD mice, evidenced by higher bromodeoxyuridine incorporation in these cells (Ling et al., 2013). To the contrary, Notch3−/− mice showed decreased alveolar progenitors during pregnancy, and the expansion of the CD24hiCD49flo population was only observed after the post-lactational involution. CD24hiCD49flo cells have been shown to enrich for PI-MECs and lobule progenitor cells (Jeselsohn et al., 2010). Lineage tracing of WAP-Cre-labeled cells revealed an increased accumulation of PI-MECs in Notch3−/− mice, which may contribute to, at least in part, expansion of CD24hiCD49flo population in these mice. Finally, MMTV-Notch3ICD tumors are negative for keratin 18 and keratin 14 (Ling et al., 2013), whereas Notch3−/− tumors express keratin 8, with a subset co-expressing keratin 14, suggesting distinct tumor phenotypes.

Notch3 has previously been associated with the highly aggressive triple-negative breast cancer, including basal-like and claudin-low tumors (Choy et al., 2017; Chung et al., 2017; Turner et al., 2010; Xu et al., 2012). Signaling mediated by IL6 and Notch3 was shown to promote endocrine resistance in metastatic luminal breast cancer through generation of CD133hi/ERlo/IL6hi cancer stem cells (Sansone et al., 2016). However, recent studies suggest that Notch3 maintains a luminal phenotype and may suppress mammary tumorigenesis and metastasis via transactivation of estrogen receptor-α (ERα; ESR1) and PTEN gene expression (Dou et al., 2017; Zhang et al., 2021). Given the formation of mammary tumors in multiparous but not nulliparous Notch3−/− mice, it appears to be clear that Notch3 functions as a tumor suppressor in the postpartum gland. Notably, Notch3−/− mammary glands show normal expression of ERα compared with wild-type animals, in both nulliparous and parous contexts (Fig. S4), suggesting that Notch3 does not regulate ERα expression under normal physiological conditions.

Most of the Notch3−/− mammary tumors are ERα-positive. The tumor cells express luminal marker keratin 8, with a subset co-expressing the basal marker keratin 14, suggesting a luminal phenotype. In corroboration with our finding that deletion of Notch3 leads to luminal tumors in mice, high NOTCH3 mRNA expression is associated with better survival in patients with luminal subtype breast cancer. Interestingly, all Notch3−/− tumors show high-level expression of cyclin D1. It has been reported that cyclin D1 activity is required for self-renewal of CD24hiCD49flo mammary stem and progenitor cells in PI-MECs that are targets of MMTV-ErbB2 tumorigenesis (Jeselsohn et al., 2010; Wagner et al., 2013). In light of PI-MEC expansion and CD24hiCD49flo cell accumulation in parous Notch3−/− mice, PI-MECs may also serve as the cell-of-origin in loss-of-Notch3-induced mammary tumorigenesis.

In concert with the switch in Notch3 expression from luminal to basal cells upon involution, Notch is activated specifically in myoepithelial cells of involuting alveoli. Notch3−/− mice show drastically decreased Notch activation in these cells, suggesting a non-redundant function of Notch3 in this context. A recent study reveals that myoepithelial cells do not die concomitantly with the luminal cells during involution; instead, they reorganize into parallel strands of ducts or remain as small outpouchings (Hitchcock et al., 2020). It is tantalizing to speculate that some PI-MECs detected in parous mice may represent myoepithelial cells that survived involution, and loss of Notch3 in these cells may cause decreased apoptosis and/or increased proliferation.

Brown adipocytes may regulate mammary epithelial differentiation (Gouon-Evans and Pollard, 2002). We observed a transient emergence of brown adipocytes in wild-type but not Notch3−/− mice during early involution. Mammary alveolar epithelial cells can convert into brown adipocytes in post-lactational mice (Giordano et al., 2017), and myoepithelial cells can be induced to differentiate into beige/brite adipocytes in vitro (Li et al., 2017), raising the possibility of Notch3-regulated alveolar myoepithelial cell to brown adipocyte transdifferentiation during involution. Despite this, quantitative lineage tracing showed that less than 1% of brown adipocytes were derived from mammary epithelium in the post-weaning gland (Li et al., 2017). Notch3 was recently identified as the major Notch receptor involved in brown adipogenesis from multipotent mesenchymal cells (Rodríguez-Cano et al., 2020). As Notch3β-geo is likely to represent a null allele, loss-of-function in the stroma should disrupt brown adipocyte differentiation. Intriguingly, breast cancer cells co-cultured with Ucp1-deficient adipocytes showed an increased proportion of cancer stem cells, suggesting a potential tumor-suppressive role for Ucp1 in the tumor microenvironment (Zhang et al., 2020). Indeed, like NOTCH3, high expression of UCP1 is associated with prolonged relapse-free survival, especially in patients with luminal subtype tumors.

Finally, transplantation of parous Notch3−/− mammary epithelial cells into the mammary fat pad of wild-type mice showed that mutant cells were able to incorporate into alveolar buds in parous but not in age-matched nulliparous hosts, indicating a requirement of postpartum microenvironment for clonal activity of Notch3−/− PI-MECs. Mammary gland-associated adipocytes were shown to be required for proper epithelial remodeling and differentiation during involution (Zwick et al., 2018). Other components of the parous gland microenvironment, including immune cells, fibroblasts, and blood and lymphatic vasculatures, may also control PI-MEC cell behavior, and thereby regulate mammary tumorigenesis.

Mice

Generation of the Notch3β-geo mouse strain has been previously described (Xu et al., 2010). TNR mice (Duncan et al., 2005) were kindly provided by Dr Nicholas Gaiano (Johns Hopkins University School of Medicine, Baltimore, MD, USA). R26YFP, WAP-Cre and MMTV-Wnt1 were purchased from the Jackson Laboratory. Mice were housed under standard condition and all mouse procedures were performed in accordance with protocols approved by Institutional Animal Care and Use Committees at the University of Mississippi Medical Center and the Hospital for Sick Children.

Mammary gland whole-mount preparation, morphometric analysis and X-gal staining

All analyses were performed using number four (inguinal) mammary glands. Whole-mount mammary glands were stained with Hematoxylin according to standard procedure. Morphometric analysis was performed using ImageJ software. The length of ductal elongation was assessed by measuring the mean distance from nipple to the three most distal TEBs in each mammary gland. The density of branch points was the mean number of branch points on the three longest primary ducts divided by their mean length. For the size of alveoli (area), at least ten representative alveoli in each gland were measured to calculate the mean value. The mean size of alveoli is compared between three pairs of wild-type and Notch3−/− littermates, and presented as fold change. For X-gal staining, mammary tissues were fixed in 1.0% formaldehyde and 0.02% Nonidet P-40 (in PBS) overnight, and stained with X-gal between 4 and 16 h. X-gal staining of frozen sections was counterstained with 0.5% Eosin. Some of the whole-mount X-gal staining was counterstained with Hematoxylin, or paraffin-embedded for sectioning and counterstaining with Neutral-red.

Histology and immunohistochemistry

Formalin-fixed paraffin-embedded mouse mammary tissues were processed for Hematoxylin and Eosin staining by standard procedure. For immunostaining, 5 µm sections were rehydrated, processed by microwave antigen retrieval and stained according to standard protocol. Representative photomicrographs were acquired using a Nikon Eclipse 80i microscope. Primary antibodies used for immunostaining were: GFP (Abcam, ab6673, 1:200), Cytokeratin 8 (Fitzgerald, 10R-C177AX, 1:10), Cytokeratin 14 (Panomics, E2624, 1:200), YFP (Invitrogen, A-11122, 1:200), Ucp1 (ProteinTech, 23673-1-AP, 1:200), ERα (Santa Cruz Biotechnology, sc-542, 1:200), Cyclin D1 (Santa Cruz Biotechnology, sc-753, 1:200), Cytokeratin 10 (Santa Cruz Biotechnology, sc-23877, 1:200), EGFR (Abcam, ab32077, 1:200) and SPC (Santa Cruz Biotechnology, sc-13979, 1:200).

Western blot analysis

Mammary tissues were homogenized and lysed in RIPA buffer (Boston BioProducts) supplemented with protease inhibitor (Roche). Supernatants were clarified by centrifugation (20,000 g), and total amount of proteins were quantified. Equivalent amounts of proteins from each sample were loaded for western blots, performed according to standard methodology. Antibodies for probing specified proteins are as follows: E-cadherin (Cell Signaling Technology, 3195, 1:1000), Ucp1 (ProteinTech, 23673-1-AP, 1:1000), Notch3 (ProteinTech, 55114-1-AP, 1:1000), and β-Actin (Santa Cruz Biotechnology, sc-81178, 1:5000).

Flow cytometry

Mouse mammary tissues were dissociated in Collagenase/Hyaluronidase solution (StemCell Technologies) and prepared for single cell suspensions. Lineage-depleted mammary epithelial cells were generated using EasySep mouse mammary stem cell enrichment kit (StemCell Technologies) according to the manufacturer's instructions. Flow cytometry of lineage-depleted cells was performed using a BD LSR-II Flow Cytometer (BD Biosciences) as per the manufacturer's protocol with the following antibodies: PE-Cy5 rat anti-human CD49f (BD Pharmingen, 551129, 1:500), PE-Cy7 anti-mouse Ly-6A/E (Sca-1) (eBioscience, 25-5981, 1:500), eFluor 450 anti-mouse CD24 (eBioscience, 48-0242, 1:400) and R-PE hamster anti-mouse/rat CD61 (Invitrogen, MCD6104, 1:400). Fluorescence was recorded using BD LSR-II flow cytometer and analyzed with FlowJo 9.1 (Treestar). Dead cells were excluded based on propidium iodide staining. Mammary tissues from three animals per genotype were analyzed with similar results.

Mammary epithelial cell transplantation

Lineage-depleted mammary epithelial cells were prepared from parous Notch3β-geo/β-geo mice (having been backcrossed to FVB for at least six generations). Approximately 2×105 lineage-depleted cells were resuspended in DMEM medium and mixed at 1:1 ratio with 20 μl Matrigel (BD Bioscience) on ice. A total of 40 μl cell sample was then injected into each of the inguinal mammary gland in syngeneic FVB mice at involution day 21 or age-matched nulliparous FVB mice, under anesthesia with isoflurane. Host mammary glands were harvested for analysis at 8 weeks after transplantation.

Gene expression analysis of human data sets

For the survival analysis related to NOTCH3 (203238_s_at) and UCP1 (221384_at) expressions, we used online tool (https://kmplot.com/analysis/) to perform univariate Cox regression analysis and Kaplan–Meier survival curves. The breast cancer database on this site was set up by downloading transcriptome-level gene expression datasets with available clinical information from the GEO (https://www.ncbi.nlm.nih.gov/geo/) and EGA (https://ega-archive.org/) repositories. Only datasets with at least 30 samples and only those which were generated using the GEO platforms GPL96, GPL570 and GPL571 were included (Győrffy, 2021). The high versus low gene expression was determined using best cut-off (Lánczky and Győrffy, 2021). The hazard ratio with 95% confidence and P-values were calculated from 846 patients with basal subtype of breast cancer, 2277 patients with luminal A subtype, 1491 patients with luminal B subtype, and 315 patients with HER2+ subtype that have relapse-free survival information, and from 404 patients with basal subtype, 794 patients with luminal A subtype, 515 patients with luminal B subtype and 166 patients with HER2+ subtype having overall survival information.

Statistics

Statistical analyses were performed using Prism version 9.2.0 (GraphPad Software). All data are presented as the mean with standard error of the mean (s.e.m.). Two-group comparisons were analyzed using two-tailed Student's t-test. Mouse tumor-free survival was calculated by the Kaplan–Meier method and compared using nonparametric log-rank test. P-value of 0.05 or less was considered statistically significant.

The authors thank Dr Nicholas Gaiano for providing us with TNR mice and Dr Robert D. Cardiff for the pathological evaluation of mouse mammary tumors.

Author contributions

Conceptualization: S.E.E., K.X.; Methodology: W.-C.C.; Validation: W.-C.C., S.E.E., K.X.; Formal analysis: W.-C.C., K.X.; Investigation: W.-C.C., K.X.; Resources: S.E.E., K.X.; Data curation: W.-C.C.; Writing - original draft: K.X.; Writing - review & editing: W.-C.C., S.E.E., K.X.; Supervision: S.E.E., K.X.; Project administration: K.X.; Funding acquisition: S.E.E., K.X.

Funding

This work was supported by a U.S. Department of Defense grant W81XWH-19-1-0031 to K.X. and a grant from the Canadian Cancer Society Research Institute to S.E.E.

Albrektsen
,
G.
,
Heuch
,
I.
,
Hansen
,
S.
and
Kvåle
,
G.
(
2005
).
Breast cancer risk by age at birth, time since birth and time intervals between births: exploring interaction effects
.
Br. J. Cancer
92
,
167
-
175
.
Asselin-Labat
,
M. L.
,
Sutherland
,
K. D.
,
Barker
,
H.
,
Thomas
,
R.
,
Shackleton
,
M.
,
Forrest
,
N. C.
,
Hartley
,
L.
,
Robb
,
L.
,
Grosveld
,
F. G.
,
van der Wees
,
J.
et al.
(
2007
).
Gata-3 is an essential regulator of mammary-gland morphogenesis and luminal-cell differentiation
.
Nat. Cell Biol.
9
,
201
-
209
.
Asselin-Labat
,
M. L.
,
Vaillant
,
F.
,
Sheridan
,
J. M.
,
Pal
,
B.
,
Wu
,
D.
,
Simpson
,
E. R.
,
Yasuda
,
H.
,
Smyth
,
G. K.
,
Martin
,
T. J.
,
Lindeman
,
G. J.
et al.
(
2010
).
Control of mammary stem cell function by steroid hormone signalling
.
Nature
465
,
798
-
802
.
Bossuyt
,
V.
,
Fadare
,
O.
,
Martel
,
M.
,
Ocal
,
I. T.
,
Burtness
,
B.
,
Moinfar
,
F.
,
Leibl
,
S.
and
Tavassoli
,
F. A.
(
2005
).
Remarkably high frequency of EGFR expression in breast carcinomas with squamous differentiation
.
Int. J. Surg. Pathol.
13
,
319
-
327
.
Boulanger
,
C. A.
,
Wagner
,
K. U.
and
Smith
,
G. H.
(
2005
).
Parity-induced mouse mammary epithelial cells are pluripotent, self-renewing and sensitive to TGF-β1 expression
.
Oncogene
24
,
552
-
560
.
Chie
,
W.-C.
,
Hsieh
,
C.
,
Newcomb
,
P. A.
,
Longnecker
,
M. P.
,
Mittendorf
,
R.
,
Greenberg
,
E. R.
,
Clapp
,
R. W.
,
Burke
,
K. P.
,
Titus-Ernstoff
,
L.
,
Trentham-Dietz
,
A.
et al.
(
2000
).
Age at any full-term pregnancy and breast cancer risk
.
Am. J. Epidemiol.
151
,
715
-
722
.
Choy
,
L.
,
Hagenbeek
,
T. J.
,
Solon
,
M.
,
French
,
D.
,
Finkle
,
D.
,
Shelton
,
A.
,
Venook
,
R.
,
Brauer
,
M. J.
and
Siebel
,
C. W.
(
2017
).
Constitutive NOTCH3 signaling promotes the growth of basal breast cancers
.
Cancer Res.
77
,
1439
-
1452
.
Chung
,
W. C.
,
Zhang
,
S.
,
Challagundla
,
L.
,
Zhou
,
Y.
and
Xu
,
K.
(
2017
).
Lunatic fringe and p53 cooperatively suppress mesenchymal stem-like breast cancer
.
Neoplasia
19
,
885
-
895
.
Dou
,
X.-W.
,
Liang
,
Y.-K.
,
Lin
,
H.-Y.
,
Wei
,
X.-L.
,
Zhang
,
Y.-Q.
,
Bai
,
J.-W.
,
Chen
,
C.-F.
,
Chen
,
M.
,
Du
,
C.-W.
,
Li
,
Y.-C.
et al.
(
2017
).
Notch3 maintains luminal phenotype and suppresses tumorigenesis and metastasis of breast cancer via trans-activating estrogen receptor-alpha
.
Theranostics
7
,
4041
-
4056
.
Duncan
,
A. W.
,
Rattis
,
F. M.
,
DiMascio
,
L. N.
,
Congdon
,
K. L.
,
Pazianos
,
G.
,
Zhao
,
C.
,
Yoon
,
K.
,
Cook
,
J. M.
,
Willert
,
K.
,
Gaiano
,
N.
et al.
(
2005
).
Integration of Notch and Wnt signaling in hematopoietic stem cell maintenance
.
Nat. Immunol.
6
,
314
-
322
.
Fu
,
N. Y.
,
Nolan
,
E.
,
Lindeman
,
G. J.
and
Visvader
,
J. E.
(
2020
).
Stem cells and the differentiation hierarchy in mammary gland development
.
Physiol. Rev.
100
,
489
-
523
.
Giordano
,
A.
,
Perugini
,
J.
,
Kristensen
,
D. M.
,
Sartini
,
L.
,
Frontini
,
A.
,
Kajimura
,
S.
,
Kristiansen
,
K.
and
Cinti
,
S.
(
2017
).
Mammary alveolar epithelial cells convert to brown adipocytes in post-lactating mice
.
J. Cell. Physiol.
232
,
2923
-
2928
.
Gouon-Evans
,
V.
and
Pollard
,
J. W.
(
2002
).
Unexpected deposition of brown fat in mammary gland during postnatal development
.
Mol. Endocrinol.
16
,
2618
-
2627
.
Gu
,
B.
,
Watanabe
,
K.
,
Sun
,
P.
,
Fallahi
,
M.
and
Dai
,
X.
(
2013
).
Chromatin effector Pygo2 mediates Wnt-notch crosstalk to suppress luminal/alveolar potential of mammary stem and basal cells
.
Cell Stem Cell
13
,
48
-
61
.
Guo
,
Q.
,
Minnier
,
J.
,
Burchard
,
J.
,
Chiotti
,
K.
,
Spellman
,
P.
and
Schedin
,
P.
(
2017
).
Physiologically activated mammary fibroblasts promote postpartum mammary cancer
.
JCI Insight
2
,
e89206
.
Győrffy
,
B.
(
2021
).
Survival analysis across the entire transcriptome identifies biomarkers with the highest prognostic power in breast cancer
.
Comput. Struct. Biotechnol. J.
19
,
4101
-
4109
.
Henry
,
M. D.
,
Triplett
,
A. A.
,
Oh
,
K. B.
,
Smith
,
G. H.
and
Wagner
,
K. U.
(
2004
).
Parity-induced mammary epithelial cells facilitate tumorigenesis in MMTV-neu transgenic mice
.
Oncogene
23
,
6980
-
6985
.
Hitchcock
,
J. R.
,
Hughes
,
K.
,
Harris
,
O. B.
and
Watson
,
C. J.
(
2020
).
Dynamic architectural interplay between leucocytes and mammary epithelial cells
.
FEBS J.
287
,
250
-
266
.
Jerry
,
D. J.
,
Kuperwasser
,
C.
,
Downing
,
S. R.
,
Pinkas
,
J.
,
He
,
C.
,
Dickinson
,
E.
,
Marconi
,
S.
and
Naber
,
S. P.
(
1998
).
Delayed involution of the mammary epithelium in BALB/c-p53null mice
.
Oncogene
17
,
2305
-
2312
.
Jeselsohn
,
R.
,
Brown
,
N. E.
,
Arendt
,
L.
,
Klebba
,
I.
,
Hu
,
M. G.
,
Kuperwasser
,
C.
and
Hinds
,
P. W.
(
2010
).
Cyclin D1 kinase activity is required for the self-renewal of mammary stem and progenitor cells that are targets of MMTV-ErbB2 tumorigenesis
.
Cancer Cell
17
,
65
-
76
.
Lafkas
,
D.
,
Rodilla
,
V.
,
Huyghe
,
M.
,
Mourao
,
L.
,
Kiaris
,
H.
and
Fre
,
S.
(
2013
).
Notch3 marks clonogenic mammary luminal progenitor cells in vivo
.
J. Cell Biol.
203
,
47
-
56
.
Lambe
,
M.
,
Hsieh
,
C.
,
Trichopoulos
,
D.
,
Ekbom
,
A.
,
Pavia
,
M.
and
Adami
,
H. O.
(
1994
).
Transient increase in the risk of breast cancer after giving birth
.
N. Engl. J. Med.
331
,
5
-
9
.
Lánczky
,
A.
and
Győrffy
,
B.
(
2021
).
Web-based survival analysis tool tailored for medical research (KMplot): development and implementation
.
J. Med. Internet Res.
23
,
e27633
.
Li
,
L.
,
Li
,
B.
,
Li
,
M.
,
Niu
,
C.
,
Wang
,
G.
,
Li
,
T.
,
Król
,
E.
,
Jin
,
W.
and
Speakman
,
J. R.
(
2017
).
Brown adipocytes can display a mammary basal myoepithelial cell phenotype in vivo
.
Mol. Metab.
6
,
1198
-
1211
.
Ling
,
H.
,
Sylvestre
,
J. R.
and
Jolicoeur
,
P.
(
2013
).
Cyclin D1-dependent induction of luminal inflammatory breast tumors by activated notch3
.
Cancer Res.
73
,
5963
-
5973
.
Liu
,
Q.
,
Wuu
,
J.
,
Lambe
,
M.
,
Hsieh
,
S. F.
,
Ekbom
,
A.
and
Hsieh
,
C. C.
(
2002
).
Transient increase in breast cancer risk after giving birth: postpartum period with the highest risk (Sweden)
.
Cancer Causes Control
13
,
299
-
305
.
Lyons
,
T. R.
,
O'Brien
,
J.
,
Borges
,
V. F.
,
Conklin
,
M. W.
,
Keely
,
P. J.
,
Eliceiri
,
K. W.
,
Marusyk
,
A.
,
Tan
,
A. C.
and
Schedin
,
P.
(
2011
).
Postpartum mammary gland involution drives progression of ductal carcinoma in situ through collagen and COX-2
.
Nat. Med.
17
,
1109
-
1115
.
Lyons
,
T. R.
,
Borges
,
V. F.
,
Betts
,
C. B.
,
Guo
,
Q.
,
Kapoor
,
P.
,
Martinson
,
H. A.
,
Jindal
,
S.
and
Schedin
,
P.
(
2014
).
Cyclooxygenase-2-dependent lymphangiogenesis promotes nodal metastasis of postpartum breast cancer
.
J. Clin. Invest.
124
,
3901
-
3912
.
Martinson
,
H. A.
,
Jindal
,
S.
,
Durand-Rougely
,
C.
,
Borges
,
V. F.
and
Schedin
,
P.
(
2015
).
Wound healing-like immune program facilitates postpartum mammary gland involution and tumor progression
.
Int. J. Cancer
136
,
1803
-
1813
.
Matulka
,
L. A.
,
Triplett
,
A. A.
and
Wagner
,
K. U.
(
2007
).
Parity-induced mammary epithelial cells are multipotent and express cell surface markers associated with stem cells
.
Dev. Biol.
303
,
29
-
44
.
Polyak
,
K.
(
2006
).
Pregnancy and breast cancer: the other side of the coin
.
Cancer Cell
9
,
151
-
153
.
Raafat
,
A.
,
Goldhar
,
A. S.
,
Klauzinska
,
M.
,
Xu
,
K.
,
Amirjazil
,
I.
,
McCurdy
,
D.
,
Lashin
,
K.
,
Salomon
,
D.
,
Vonderhaar
,
B. K.
,
Egan
,
S.
et al.
(
2011
).
Expression of Notch receptors, ligands, and target genes during development of the mouse mammary gland
.
J. Cell. Physiol.
226
,
1940
-
1952
.
Raouf
,
A.
,
Zhao
,
Y.
,
To
,
K.
,
Stingl
,
J.
,
Delaney
,
A.
,
Barbara
,
M.
,
Iscove
,
N.
,
Jones
,
S.
,
McKinney
,
S.
,
Emerman
,
J.
et al.
(
2008
).
Transcriptome analysis of the normal human mammary cell commitment and differentiation process
.
Cell Stem Cell
3
,
109
-
118
.
Rodríguez-Cano
,
M. M.
,
González-Gómez
,
M. J.
,
Sánchez-Solana
,
B.
,
Monsalve
,
E. M.
,
Díaz-Guerra
,
M. M.
,
Laborda
,
J.
,
Nueda
,
M. L.
and
Baladrón
,
V.
(
2020
).
NOTCH receptors and DLK proteins enhance brown adipogenesis in mesenchymal C3H10T1/2 cells
.
Cells
9
,
2032
.
Sansone
,
P.
,
Ceccarelli
,
C.
,
Berishaj
,
M.
,
Chang
,
Q.
,
Rajasekhar
,
V. K.
,
Perna
,
F.
,
Bowman
,
R. L.
,
Vidone
,
M.
,
Daly
,
L.
,
Nnoli
,
J.
et al.
(
2016
).
Self-renewal of CD133(hi) cells by IL6/Notch3 signalling regulates endocrine resistance in metastatic breast cancer
.
Nat. Commun.
7
,
10442
.
Schedin
,
P.
,
O'Brien
,
J.
,
Rudolph
,
M.
,
Stein
,
T.
and
Borges
,
V.
(
2007
).
Microenvironment of the involuting mammary gland mediates mammary cancer progression
.
J. Mammary Gland Biol. Neoplasia
12
,
71
-
82
.
Stingl
,
J.
,
Eirew
,
P.
,
Ricketson
,
I.
,
Shackleton
,
M.
,
Vaillant
,
F.
,
Choi
,
D.
,
Li
,
H. I.
and
Eaves
,
C. J.
(
2006
).
Purification and unique properties of mammary epithelial stem cells
.
Nature
439
,
993
-
997
.
Strange
,
R.
,
Li
,
F.
,
Saurer
,
S.
,
Burkhardt
,
A.
and
Friis
,
R. R.
(
1992
).
Apoptotic cell death and tissue remodelling during mouse mammary gland involution
.
Development
115
,
49
-
58
.
Turner
,
N.
,
Lambros
,
M. B.
,
Horlings
,
H. M.
,
Pearson
,
A.
,
Sharpe
,
R.
,
Natrajan
,
R.
,
Geyer
,
F. C.
,
van Kouwenhove
,
M.
,
Kreike
,
B.
,
Mackay
,
A.
et al.
(
2010
).
Integrative molecular profiling of triple negative breast cancers identifies amplicon drivers and potential therapeutic targets
.
Oncogene
29
,
2013
-
2023
.
Visvader
,
J. E.
and
Stingl
,
J.
(
2014
).
Mammary stem cells and the differentiation hierarchy: current status and perspectives
.
Genes Dev.
28
,
1143
-
1158
.
Wagner
,
K.-U.
,
Boulanger
,
C. A.
,
Henry
,
M. D.
,
Sgagias
,
M.
,
Hennighausen
,
L.
and
Smith
,
G. H.
(
2002
).
An adjunct mammary epithelial cell population in parous females: its role in functional adaptation and tissue renewal
.
Development
129
,
1377
-
1386
.
Wagner
,
K.-U.
,
Booth
,
B. W.
,
Boulanger
,
C. A.
and
Smith
,
G. H.
(
2013
).
Multipotent PI-MECs are the true targets of MMTV-neu tumorigenesis
.
Oncogene
32
,
1338
.
Xu
,
K.
,
Nieuwenhuis
,
E.
,
Cohen
,
B. L.
,
Wang
,
W.
,
Canty
,
A. J.
,
Danska
,
J. S.
,
Coultas
,
L.
,
Rossant
,
J.
,
Wu
,
M. Y.
,
Piscione
,
T. D.
et al.
(
2010
).
Lunatic Fringe-mediated Notch signaling is required for lung alveogenesis
.
Am. J. Physiol. Lung Cell. Mol. Physiol.
298
,
L45
-
L56
.
Xu
,
K.
,
Usary
,
J.
,
Kousis
,
P. C.
,
Prat
,
A.
,
Wang
,
D. Y.
,
Adams
,
J. R.
,
Wang
,
W.
,
Loch
,
A. J.
,
Deng
,
T.
,
Zhao
,
W.
et al.
(
2012
).
Lunatic fringe deficiency cooperates with the Met/Caveolin gene amplicon to induce basal-like breast cancer
.
Cancer Cell
21
,
626
-
641
.
Zhang
,
F.
,
Liu
,
B.
,
Deng
,
Q.
,
Sheng
,
D.
,
Xu
,
J.
,
He
,
X.
,
Zhang
,
L.
and
Liu
,
S.
(
2020
).
UCP1 regulates ALDH-positive breast cancer stem cells through releasing the suppression of Snail on FBP1
.
Cell Biol. Toxicol.
37
,
277
-
291
.
Zhang
,
Y.-Q.
,
Liang
,
Y.-K.
,
Wu
,
Y.
,
Chen
,
M.
,
Chen
,
W.-L.
,
Li
,
R.-H.
,
Zeng
,
Y.-Z.
,
Huang
,
W.-H.
,
Wu
,
J.-D.
,
Zeng
,
D.
et al.
(
2021
).
Notch3 inhibits cell proliferation and tumorigenesis and predicts better prognosis in breast cancer through transactivating PTEN
.
Cell Death Dis.
12
,
502
.
Zwick
,
R. K.
,
Rudolph
,
M. C.
,
Shook
,
B. A.
,
Holtrup
,
B.
,
Roth
,
E.
,
Lei
,
V.
,
Van Keymeulen
,
A.
,
Seewaldt
,
V.
,
Kwei
,
S.
,
Wysolmerski
,
J.
et al.
(
2018
).
Adipocyte hypertrophy and lipid dynamics underlie mammary gland remodeling after lactation
.
Nat. Commun.
9
,
3592
.

Competing interests

The authors declare no competing or financial interests.

Supplementary information