This cell development step describes commitment of mammary stem cells (MaSCs) to the myoepithelial lineage by differentiating into unipotent mammary myoepithelial progenitors.

Characteristics of MaSCs

MaSCs have been studied in detail in mouse development, and have been much less characterized in humans. Most data available on human MaSCs is derived from adult stem-like cells in human mammary glands that are able to differentiate in vitro into both myoepithelial and ductal (luminal) epithelial cells and thus are also known as mammary bipotent progenitors (MBiPs). It is uncertain how much these adult human MaSCs/MBiPs differ from embryonic and fetal MaSCs. In the mouse, MaSCs are defined as those cells that are able to generate a functional mammary gland when transplanted in vivo (Shackleton et al. 2006; Stingl et al. 2006), but there is evidence that expression of certain markers in mouse MaSCs changes during different stages of mammary gland morphogenesis e.g. during pregnancy (Desgrosellier et al. 2014). Multiple reports agree on MaSCs residing within the basal epithelial cell subpopulation in both humans and mice, but the definitive list of markers that can be used to isolate pure MaSCs from within the basal epithelial subset has not been defined (reviewed in Phillips and Kuperwasser 2014). Mouse MaSCs are first apparent in mammary placodes. Lef1 (Lymphoid Enhancer Binding Factor 1) is a transcription factor that binds to beta-catenin (Ctnnb1) after activation by Wnt/β-catenin signaling. The Lef1:Ctnnb1 complex plays a critical role in controlling the transcription of genes involved in mouse mammary placode formation by driving epithelial cell proliferation and differentiation during the early stages of mammary gland development (reviewed in Lindvall et al. 2007). WNT ligand Wnt10b plays an important role in the embryonic stages of mouse mammary gland development, including placode and bud formation (reviewed in Watson and Khaled 2020, Slepicka et al. 2021). Besides Wnt10b, Wnt6 and Wnt10a are also implicated in mouse mammary bud formation (reviewed in McNally and Stein 2017). Wnt4 plays a role in mouse mammary gland development during puberty and pregnancy (reviewed in Slepicka et al. 2021). Wnt4 expression is regulated by progesterone, as it is expressed in progesterone receptor-positive cells (reviewed in Tanos et al. 2012), and it acts as a paracrine mediator of progesterone signaling in the mammary gland (reviewed in McNally and Stein 2017). FGF signaling also plays a role in mammary placode formation, in particular Fgf10-mediated activation of Fgfr2b (reviewed in Mailleux et al. 2002, McNally and Stein 2017), but Fgfr1 and ligands Fgf4, Fgf8, Fgf7, and Fgf17 are also expressed in the developing placode (reviewed in McNally and Stein 2017). Other signaling pathways studied in mouse that contribute to development of mammary glands, maintenance and differentiation of MaSC include Hedgehog signaling (reviewed in Lewis and Veltmaat 2004), NOTCH signaling (Bouras et al. 2008, reviewed in Edwards and Brennan 2021), and estrogen and progesterone signaling (Feng et al. 2007, reviewed in Tanos et al. 2012).

In normal adult human breast epithelium mammary stem cells are rare and are located in the ductal part of terminal ductal lobular units (Villadsen et al. 2007; Ginestier et al. 2007).

Markers of human mammary stem cells/bipotent progenitor cells (MaSC/MBiP) are listed in the table below (in the table, "No" means that the marker is not listed for the specified cell type while "N/A" means that the specified cell type is not annoted in the cited external marker database). Recent studies indicate that MaSC/MBiP cell population is heterogeneous, and it is likely that early and late MaSCs exist with not completely overlapping sets of markers (Scheele et al. 2017, reviewed in Visvader and Stingl 2014 and Slepicka et al. 2021).

ALDH1A1 (also known as ALDH1 or Aldehyde dehydrogenase 1A1) and ALDH1A3 (also known as ALDH6 or Retinaldehyde dehydrogenase 3) belong to the ALDH family of cytosolic NADP+-dependent dehydrogenases that catalyze the irreversible oxidation of a various endogenous and exogenous aldehydes to their corresponding carboxylic acids (reviewed in Tomita et al. 2016). The ALDH1 subfamily has three main isotypes, ALDH1A1, ALDH1A2 and ALDH1A3, and is a marker of normal tissue stem cells and cancer stem cells (reviewed in Tomita et al. 2016). ALDH1 catalytic activity, which plays an important role in oxidation of retinal to retinoic acid (RA), is mainly attributed to ALDH1A1 (reviewed in Tomita et al. 2016). Both ALDH1A1 and ALDH1A3 are expressed in human breast tissue (reviewed in Tomita et al. 2016). The study by Prat et al. 2013 did not detect increased ALDH1A1 transcript levels in MaSC/MBiPs compared to other mammary gland cell types.

CD44 is a transmembrane glycoprotein that is a common component of the stem cell niche and exists in a range of alternative splicing-generated isoforms (reviewed in Williams et al. 2013). CD44 modulates signal transduction in the stem cell niche through interactions with hyaluronan, extracellular matrix molecules and growth factors and their cognate receptor tyrosine kinases, thus regulating adhesion, differentiation, homing and migration of stem cells (reviewed in Williams et al. 2013).

CDH1 (E-cadherin) mRNA was reported to be significantly enriched in MaSCs/MBiPs isolated from normal human adult breast tissue (Prat et al. 2013), but a study using human mammary epithelial cell (HMEC)-derived MaSC-like cells implied that the stem-like phenotype is associated with reduction in CDH1 mRNA and protein levels, although CDH1 expression is still detectable (Lindley and Briegel 2010). Reduction in CDH1 levels observed in the study by Lindley and Briegel 2010 could be an artefact of upregulation of TGF-beta signaling and activation of wound healing- and EMT-related transcription factors that are known negative regulators of CDH1 gene expression (Phillips et al. 2014, see below).

EPCAM (Epithelial cell adhesion molecule) is expressed on a subset of normal epithelial cells, including human and mouse embryonic stem cells (ESCs), where it is necessary for the maintenance of pluripotency and self-renewal (reviewed in Imrich et al. 2012). EPCAM is overexpressed on malignant cells from a variety of cancers, the expression being particularly high on the surface of tumor initiating cells (reviewed in Imrich et al. 2012). EPCAM is a glycosylated protein with a function in homophilic cell adhesion (reviewed in Imrich et al. 2012). EPCAM undergoes gamma-secretase-mediated cleavage, releasing its intracellular domain that can translocate to the nucleus and participate in WNT signaling-regulated transcription by associating with CTNNB1 (beta-catenin), LEF1, and FHL2 (reviewed in Imrich et al. 2012). EPCAM levels are higher in luminal progenitors than in BiPs (Villadsen et al. 2007). In some studies, MaSC cells are considered to be EPCAM negative and positive for ITGA6 (CD49f) (Prat et al. 2013, Martin Carli et al. 2020) or to express low-levels of EPCAM (Bachelard-Cascales et al. 2010).

ERBB2 is a receptor tyrosine kinase of the EGFR subfamily that does not bind a ligand but forms heterodimers with other ligand-binding members of the EGFR subfamily, EGFR, ERBB3, and ERBB4 (reviewed in Korkaya and Wicha 2013). ERBB2 plays an important role in the stemness maintenance of both normal mammary stem cells and breast cancer stem cells, and also in the normal breast development and breast carcinogenesis (reviewed in Korkaya and Wicha 2013, Pupa et al. 2021).

ITGA6 (alpha6-integrin, also known as CD49f) is part of a stemness signature found on the plasma membrane of more than 30 stem cell populations, particularly in complex with ITGB1 (beta1-integrin, also known as CD29) and ITGB4 (beta4-integrin) (reviewed in Zhou et al. 2018). The heterodimer of ITGA6 and ITGB1 serves as a cellular receptor for laminin, which may play a role in anchoring of stem cells to their niches (reviewed in Zhou et al. 2018). High levels of ITGA6 and ITGB1 are expressed on MaSCs (reviewed in Zhou et al. 2018). In breast cancer tumor-initiating cells, ITGA6 and ITGB1 were shown to promote self-renewal (reviewed in Zhou et al. 2018).

Keratins are the largest group of intermediate filaments, primarily expressed in epithelial tissues and integumentary appendages, including mammary glands (reviewed in Toivola et al. 2015). In humans, there are 54 keratin-coding genes, divided into type I (KRT9-KRT40) and type II (KRT1-KRT8, KRT71-KRT86) subfamilies (reviewed in Toivola et al. 2015). Keratins form obligate heterodimers/heteropolymers with preferential dimerization partners varying between different cell types and developmental stages (reviewed in Toivola et al. 2015). Keratins expressed in normal breast tissue include KRT5, KRT7, KRT8, KRT14, KRT15, KRT17, KRT18, and KRT19 (Böcker et al. 2002). While KRT6 is sometimes reported to be expressed in normal breast tissue and MaSCs/MBiPs (Villadsen et al. 2007, Zhao et al. 2010, Morel et al. 2017), it has been reported that some antibodies used for immunohistochemistry cross-react with KRT5 and KRT6, and that KRT5 but not KRT6 is expressed in normal breast (Böcker et al. 2002, Boecker et al. 2002). While multiple studies reported KRT5 as a marker of MaSCs, initially based on the existence of rare cells that are positive for KRT5 but negative for luminal cell keratins KRT8 and KRT18, a study by Clarke et al. 2004 showed that this might be an artefact of paraffin embedding, because immunostaining of frozen breast tissue sections shows that all KRT5-positive cells expressed at least low levels of KRT8 and KRT18. KRT14 is detectable in MaSCs at protein (Villadsen et al. 2007, Zhao et al. 2010, Hilton et al. 2012, Prat et al. 2013, Morel et al. 2017) and mRNA (Hilton et al. 2012, Prat et al. 2013) levels, but detailed analyses have shown that only a small percentage of MaSC express KRT14 (Stingl et al. 2001, Ginestier et al. 2007) or it is undetectable (Phillips et al. 2014). MaSCs positive for KRT14 may represent later stage MaSCs, in the process of committing to the myoepithelial lineage. KRT17 transcript has been reported to be specific to MaSCs (Prat et al. 2013, Phillips et al. 2014).

TP63 is a p53 family member with six different splicing isoforms. The two main splicing isoform groups differ in the presence of the N-terminal transactivation (TA) domain and are named TAp63 (group of longer splicing isoforms with the TA domain) and deltaNp63 (group of shorter splicing isoforms with no TA domain) (reviewed in Crum and McKeon 2010). Within each group, there are three splicing isoforms that differ in their C-termini, alpha, beta, and gamma, with alpha having the longest and gamma the shortest C-terminus, so that the six isoforms are TAp63alpha, TAp63beta, TAp63gamma, deltaNp63alpha, deltaNp63beta, and deltaNp63gamma (reviewed in Crum and McKeon 2010). Translation of the TAp63 isoforms appears to be restricted primarily to the female-derived germ cell populations in humans (reviewed in Crum and McKeon 2010). deltaNp63 isoforms are highly expressed in basal cells, immature squamous epithelium, and epithelial stem cells (reviewed in Crum and McKeon 2010). Due to unavailability of antibodies that can distinguish between alpha, beta, and gamma isoforms, the knowledge about protein expression of different deltaNp63 isoforms is limited, but the analysis of transcripts shows that deltaNp63alpha is the dominant isoform (reviewed in Crum and McKeon 2010). deltaNp63alpha is therefore annotated as a marker of MaSCs.

MME (Neprilysin, also known as Membrane metalloendopeptidase, Enkephalinase or CD10) is a zinc-dependent metalloendoprotease that inactivates a number of signaling peptides and is implicated in stem cell niche maintenance in various normal tissues, including mammary glands, but also in cancer (reviewed in Maguer-Satta et al. 2011). MME has been reported to be expressed in MaSC by several studies at protein (Stingl et al. 2005, Bachelard-Cascales et al. 2010, Morel et al. 2017) and mRNA levels (Prat et al. 2013). Based on the fact that MME is not expressed in the primary bud during human fetal development (Jolicoeur et al. 2003) and that the majority of studies consider it to be a marker of mature myoepithelial cells of the mammary gland (Jolicoeur et al. 2003, Ginestier et al. 2007, Zhao et al. 2010, Hilton et al. 2012, Mukhopadhyay et al. 2013), MME is not included as a MaSC marker.

Potential MaSC markers that are not yet corroborated by enough data to be annotated include CLDN4, CLDN7, OCLN, PROM1 (CD133), SNAI2 (SLUG), TWIST1, TWIST2 and ZEB1.

CLDN4, CLDN7, and OCLN mRNAs were reported to be significantly enriched in MaSCs/MBiPs isolated from normal human adult breast tissue (Prat et al. 2013).

PROM1 (CD133) mRNA was reported to be significantly enriched in MaSCs/MBiPs isolated from normal human adult breast tissue (Prat et al. 2013) but other studies identify PROM1 as a marker of mammary luminal progenitors (Stingl et al. 2005, Hilton et al. 2012), and mature luminal cells (Stingl et al. 2005).

Several transcription factors regulated by TGF-beta signaling and implicated in epithelial-to-mesenchymal (EMT) transition, namely SNAI2, TWIST1, TWIST2, and ZEB1, a transcription factor regulated by TGF-beta signaling, has been identified as highly expressed at the mRNA (Prat et al. 2013) and protein (Morel et al. 2017) levels in MaSCs-like cells isolated from adult human breast. Based on human (Lim et al. 2010, Phillips et al. 2014) and mouse (Guo et al. 2012) studies (reviewed in Phillips and Kuperwasser 2014) SNAI2 is thought to promote stemness of MaSCs by repressing genes associated with the luminal phenotype. However, due to the role SNAI2, TWIST1, TWIST2, and ZEB1 play in EMT and wound healing, which raises the possibility that their upregulation could be the artefact of isolation or culture conditions that increase TGF-beta signaling (Lindley and Briegel 2010), they will be annotated as MaSC markers if more data becomes available.

SOX family transcription factors Sox9 (Guo et al. 2012), Sox10 (Guo et al. 2012, Dravis et al. 2018), and Sox11 (Lim et al. 2010) are implicated in mouse MaSC stemness maintenance. SOX10 is also implicated in stem-like features in human breast cancer (Dravis et al. 2018). These SOX transcription factors will be annotated as human MaSC markers if more data becomes available.

Development and characteristics of myoepithelial progenitor cells

Mammary myoepithelial progenitor cells are characterized by expression of some of the mature myoepithelial cells markers but also expression of some of the MaSC markers no longer detectable in fully differentiated myoepithelial cells (Bocker et al. 2002, Boecker et al. 2002, Jolicoeur et al. 2003).

Markers of mammary myoepithelial progenitor cells are listed in the table below. Only markers reported in at least two studies have been included.

EGFR ligands EGF and TGFA stimulate in vitro differentiation of MaSCs into myoepithelial lineage, while EGFR ligand AREG inhibits it (Pasic et al. 2011, Mukhopadhyay et al. 2013). EGF and TGFA stimulate EGFR signaling and activation of downstream RAS/RAF/MAPK and PI3K/AKT signaling more strongly than AREG (Pasic et al. 2011, Mukhopadhyay et al. 2013). While reports on EGFR expression in myoepithelial progenitor cells are concordant (Stingl et al. 2001, Phillips et al. 2004), reports on expression of ERBB2, one of heterodimerization partners of EGFR, are conflicting, with ERBB2 being reported as expressed at the protein level in mammary myoepithelial progenitor cells (Stingl et al. 2001) and ERBB2 expression not reported as differentially expressed at the mRNA level (Phillips et al. 2014).

EPCAM protein is expressed at a very low level in myoepithelial progenitors (Stingl et al. 2001) and is not annotated as a marker. Findings in myoepithelial progenitors derived in vitro from immortalized human mammary epithelial cells suggest that levels of ITGA6, KRT5, KRT14 and TP63 progressively decrease from the MaSC through the myoepithelial progenitor state and that at later cell divisions of myoepithelial progenitors expression of these markers may become undetectable, while expression of ACTA2 and TGF-beta regulated genes involved in mesenchymal phenotype increases (Zhao et al. 2012, Mukhopadhyay et al. 2003).

ACTA2 is part of the contractile apparatus that enables contractility of myoepithelial cells, necessary for milk secretion in response to oxytocin (reviewed in Watson and Khaled 2020).