HomeWeb-linksRussianEnglish  Forum

Cancer biology 
Cancer stem cells 
Cancer angiogenesis 
Tumor invasion & metastasis 
Principles of cancer screening 
Tumor markers 
Circulating tumor cells 
Breast cancer 
CNS tumors 
Lung cancer 
Gastric cancer 
Urologic tumors 
Liver tumors 
Pancreatic cancer 
Hereditary cancer syndromes 
Cancer complications 
Web resources 





Breast developmental biology and tumorigenesis

The current view on breast cancer as a stem cell disease is founded on compelling evidence that many breast cancers may arise as clonal expansions from epithelial progenitors with an infinite lifespan. It has been hypothesized that unique properties of mammary stem cells, such as self-renewal, make this population a prime target for transformation and tumorigenesis. Several experimental breast cancer models support this hypothesis. The most venerable is the mammary tumor virus model in mice, where MMTV proviral insertions produce mutated mammary cells, which attain immortality (escape from growth senescence) and produce clones of mammary cells with increased propensity to develop into mammary cancer. Serial transplantations of these preneoplastic lesions result in the formation of hyperplastic/dysplastic ductal trees, suggesting that multipotent cells are affected by MMTV transformation and that they pass on their neoplastic properties to their descendants. Morphologically undifferentiated cells, reminiscent of stem/progenitor cells, are present in both premalignant and malignant mammary populations (Fig. 1). Reproductive history has a profound impact on breast tumorigenesis; thus, it is reasonable to assume that pregnancy and lactation have enduring effects on the cancer susceptibility of multipotent stem/progenitor cells.

Fig. 1.  This electron micrograph depicts an ultrathin section through one of the acini in an MMTV-induced alveolar hyperplasia. There is evidence of virus replication (MMTV) of secretory activity leading to secretory granule formation in the apical cytoplasm of the luminal cells and release into the lumen. An undifferentiated suprabasal cell (SLC) is present and proximal to it, a differentiated myoepithelial cell (arrow). Bar equals 1.0 mM.

Evidence that cancer stem cells sustain solid neoplasms has recently emerged. Whether these “cancer stem cells” arise de novo or result from mutations within normal tissue stem/progenitor cells is presently unknown. A shift in the microenvironment of mammary epithelial cells as the result of pregnancy is a plausible mechanism by which to explain the greater refractivity of mammary tissue after early parity to cancer induction or progression. In a rat chemical carcinogenesis model, Nandi and his colleagues have argued that there is no difference in the susceptibility of the mammary epithelium between nulliparous and parous females to initiation (malignant transformation) by NMU; rather, it is a reduction in the incidence of progression of the “initiated” cells to frank malignancy. This difference in “progression” is completely reversible when the parous rodents are subjected to various hormonal regimens or given growth factors such as IGF-1. If this is correct, then epithelial (stem/progenitor) cell targets for carcinogenesis are the same but behave differently in their respective microenvironments (niches) during homeostatic tissue maintenance in the parous female.

Pregnancy mediates permanent changes in mammary epithelial cells

The basic principle for the dual phenomenon of pregnancy and breast cancer is that a gestation cycle induces massive proliferation and an endpoint differentiation of epithelial subtypes. Either permanent systemic changes following a full-term pregnancy (such as a decrease in circulating levels of hormones) or the alteration of the mammary tissue itself could explain the difference in breast cancer risk between nulliparous and parous women. Sivaraman and coworkers suggested that the hormonal milieu of pregnancy affects the developmental state of a subset of mammary epithelial cells and their progeny.

Microarray evidence suggests that pregnancy mediates persistent changes in the gene expression profile in parous females. These pregnancy-induced changes can be imitated through a transient administration of hormones, in particular estrogen and progesterone or human chorionic gonadotropin. Ginger et al. used subtractive hybridization as a method to identify differentially expressed genes between hormone-treated Wistar–Furth rats and their untreated controls. Twenty-eight days after the last treatment, they identified approximately 100 differentially expressed loci. In a more comprehensive study, D’Cruz and colleagues utilized oligonucleotide arrays to examine differences in the expression profile of approximately 5,500 genes between parous mice and their nulliparous controls. These initial results were verified by more laborious methods (northern blot analysis and in situ hybridization) and across several mouse strains as well as in two rat models.

The origination of parity-induced mammary epithelial cells during late pregnancy and lactation

Using the Cre-lox technology, a mammary epithelial subtype, which is abundant in nonlactating and nonpregnant parous mice, was recently described. These parity-induced mammary epithelial cells (PI-MECs) then permanently reside at the terminal ends of ducts (i.e., lobuloalveolar units) after postlactational remodeling. Two lines of evidence exist showing that the presence of PI-MECs in the involuted mammary gland is not an artifact caused by a deregulated activation of the promoter of our randomly integrated WAP-Cre construct. First, the WAP-Cre transgenic expression closely follows the activation of the endogenous WAP locus, and Ludwig and coworkers have reported similar observations in genetically engineered mice that express Cre recombinase under the endogenous Wap gene promoter (i.e., WAP-Cre knock-in mutants). Second, limiting dilution transplantation assays with dispersed epithelial cells from nulliparous female mice demonstrate the existence of lobule-limited and duct-limited progenitors. These studies were carried out with epithelial cells from WAP-LacZ transgenic mice, where LacZ is expressed from the whey acidic protein promoter in late pregnant mice. Lobule-limited outgrowths positive for LacZ expression were observed in the implanted fat pads at parturition. Similar lobule-limited outgrowths were developed when PI-MECs were inoculated in limiting dilution into the cleared mammary fat pads of subsequently impregnated hosts (Fig. 2). These structures like those described earlier comprised both secretory luminal cells and myoepithelial cells and were 100% positive for LacZ activity indicating that they were developed entirely from PI-MECs. Therefore, it is likely that PI-MECs arise from the lobule-limited progenitor population discovered by Smith among the mammary epithelial cells present in nulliparous unbred females. In addition to luminal and myoepithelial progeny, PI-MECs produced both small (SLC) and large undifferentiated light cells (ULLC) in the lobules. SLCs and ULLCs have essential roles in mammary stem/progenitor cell function. The existence of committed mammary alveolar precursors in mice and rats has been proposed earlier.

Fig. 2.  The image shows a section through a lobule-limited LacZ-positive outgrowth in full-term pregnant host composed entirely of progeny from PI-MEC. The growth comprises both luminal and myoepithelial (long arrows) cells and small-undifferentiated light cells (short arrows). Bar equals 20 mM.

PI-MECs are self-renewing and pluripotent

When fragments from glands containing PI-MECs were transplanted into epithelium-free fat pads in nulliparous hosts, PI-MECs contributed to ductal elongation in a very significant manner. The vast majority of resulting outgrowths contained LacZ-positive cells, and in >75% of the transplants, PI-MEC-derived cells were present throughout the entire ductal tree. These results clearly demonstrated that PI-MECs exhibit two important features of multipotent stem cells: self-renewal and contribution to diverse epithelial populations in ducts and alveoli. We demonstrated, for the first time, that the progeny from cells previously expressing an alveolar differentiation marker (i.e., WAP) could contribute to the formation of primary and secondary ducts. When the transplanted hosts were impregnated, the self-renewed PI-MECs at the tips of duct side branches proliferated during early pregnancy to form the new secretory acini. The transplantation procedure itself had no effect on the activation of the WAP-Cre and Rosa-LacZ transgenes because mammary fragments from nulliparous double transgenic donors never produced outgrowths with uniformly distributed LacZ-positive cells. To establish an estimate of the self-renewing ability of PI-MECs, mammary fragments containing LacZ-positive cells were transferred through four transplant generations. Each successful transplant resulted in a 400-fold increase of the implanted epithelial population, which represents roughly an 8–9 (8.65)-fold doubling of the implanted cells.

To determine to what extent the presence of neighboring LacZ-negative epithelial cells contributed to the self-renewing capacity of labeled PI-MECs, dispersed mammary epithelial cells from multiparous WAP-Cre/Rosa-LacZ females were inoculated at limiting dilutions into cleared fat pads, and the hosts were subsequently impregnated. All outgrowths contained LacZ-expressing cells, even though PI-MECs represented only 20% of the inoculated epithelial cells. Notably, no epithelial outgrowths were comprised entirely from unlabeled (LacZ-negative) cells. Both lobule-limited and duct-limited outgrowths were, however, entirely comprised from PI-MECs (and their LacZ-expressing descendents), as determined by serial sections through these structures. These results indicate that all luminal, myoepithelial and cap cells of terminal buds may be derived from PI-MECs and their progeny. This conclusion was confirmed by demonstrating that the LacZ-positive cells in these structures could be doubly stained for mammary cell lineage markers for myoepithelium (smooth muscle actin, Fig. 3), estrogen receptor alpha (ER-α), or progesterone receptor (PR). Thus, PI-MECs are not only self-renewing, but they are pluripotent as well, giving rise to progeny that differentiate along all the epithelial cell lineages of the mammary gland.

Fig. 3.  This composite shows LacZ-negative acini stained for SMA in the upper left panel and LacZ-positive acini in the remaining three panels. The arrows indicate the myoepithelial cells demonstrated by positive SMA-staining. The LacZ-positive cells appear as dark gray in this grayscale figure. Bars equal 5 mM.

WAP-TGF-β1 expression aborts self-renewal of PI-MECs in transplants

The reproductive capacity of the mammary epithelial stem cell is reduced coincident with the number of symmetric divisions it must perform. In a study using WAP-TGF-β1 transgenic mice, it was observed that mammary epithelial stem cells were prematurely aged due to ectopic expression of TGF-β1 under the regulation of the WAP gene promoter. To assess whether TGF-β1 expression in PI-MECs abolishes their capacity to self-renew, mammary epithelia from WAP-TGF-β1/WAP-Cre/Rosa-LacZ triple transgenic mice were transplanted into wild-type recipients. It is important to note that the percentage of labeled cells in the triple transgenic glands after a single parity was indistinguishable from that observed in WAP-Cre/Rosa-LacZ double transgenic controls. As expected, mammary tissue implants and dispersed cells from the triple transgenic females, after either a single pregnancy or multiple gestation cycles, failed to produce full lobular development in full-term pregnant hosts. Perhaps more importantly, LacZ-positive cells were not observed in the ducts in these transplant outgrowths either in nulliparous or early pregnant hosts. LacZ-expressing cells did appear in the transplant population and were present in the lobular structures during late pregnancy in these transplants (after 15 days to parturition). In summary, the results of these studies demonstrate that the PI-MECs that develop during pregnancy and survive subsequent tissue remodeling in the absence of lactation in WAP-TGF-β1 females were incapable or severely limited in their ability to self-renew in transplants and could not contribute to ductal development in subsequent transplant outgrowths. Therefore, self-renewal (expansion outside of a stem cell niche) and proliferation competence (asymmetric divisions within a niche) appear to be properties independently affected by autocrine TGF-β1 expression in the PI-MECs.

By definition, the self-renewal of stem cells occurs by two different processes. In asymmetric divisions, the most common activity of stem cells residing in a niche, the stem cell is preserved and one daughter becomes committed to a particular cell fate. Alternatively, a stem cell may divide symmetrically and expand to produce two or more stem cell daughters that retain stem cell properties. This latter form of self-renewal is essential for expansion of the stem cell population during allometric growth of the tissue (i.e., during ductal growth and expansion in the postpubertal female or when the mammary epithelial implant is growing in the transplanted mammary fat pad). The negative effect of TGF-β1 on the expansive self-renewal of PI-MECs supports our earlier observation regarding protection from mouse mammary tumor virus (MMTV)-induced mammary tumorigenesis in WAP-TGF-β1 transgenic females. This might suggest that the cellular targets for MMTV-mediated neoplastic transformation are PI-MECs because multiple pregnancies accelerate MMTV-induced oncogenesis.

PI-MECs and mammary tumorigenesis

Pregnancy has a dual effect on human breast cancer (protection or promotion), depending on the age of an individual, the period after a pregnancy and the genetic predisposition. In genetically engineered strains that are highly susceptible to mammary tumorigenesis and exhibit accelerated tumor development in postpartum or parous females, one might expect that PI-MECs serve as targets for neoplastic transformation. The unique growth properties of PI-MECs (i.e., responsiveness to pregnancy hormones, survival during involution, and ability to self renew) make this epithelial subtype a potential target for pregnancy-associated tumorigenesis. Transgenic mice expressing the wild-type Her2/neu (ErbB2) oncogene under transcriptional regulation of the MMTV-LTR seem to be suitable for studying the involvement of PI-MECs in pregnancy-associated mammary tumorigenesis since this animal model exhibits a relatively long latency of tumorigenesis (T50 of 205 days). Using this animal model, we demonstrated that (a) multiparous females consistently exhibited accelerated tumorigenesis compared with their nulliparous littermate controls in a mixed genetic background and (b) PI-MECs were, indeed, primary targets of neoplastic transformation in this model. The de novo generation and amplification of a large number of hormone-responsive and apoptosis-resistant epithelial cells (i.e., PI-MECs) during the first and subsequent reproductive cycles might, therefore, account for the significantly increased cancer susceptibility of parous MMTV-neu transgenic females.

To further substantiate that PI-MECs are primary targets for neoplastic transformation in MMTV-neu transgenic mice, we eliminated or greatly impaired the growth of PI-MECs by deleting the Tsg101 gene in cells that transiently activated WAP-Cre (i.e., females that carry two transgenes, MMTV-neu and WAP-Cre, in a homozygous Tsg101 conditional knockout background). The complete deletion of Tsg101 can serve as an excellent negative “selection marker” for WAP-Cre expressing cells since this gene is indispensable for the survival of normal, immortalized, and fully transformed cells. In multiparous MMTV-neu females, impaired genesis or elimination of PI-MECs resulted in a significantly reduced tumor onset, suggesting that restraining the growth and survival of differentiating alveolar cells during pregnancy (and therefore PI-MECs in parous mice) eliminates the cellular basis for transformation in this model.

Some PI-MECs are asymmetrically dividing long-label retaining cells

It was proposed over 30 years ago that somatic stem cells avoid accumulation of genetic errors resulting from DNA synthesis prior to dividing by selectively retaining their template DNA strands and passing the newly synthesized strands to their committed daughters. Therefore, somatic epithelial stem/progenitor cells labeled by DNA analogs during their inception will become long-label retaining epithelial cells (LREC). Label retention has long been considered to be a characteristic of somatic stem cells and this propensity to retain DNA label has been explained by postulating that somatic stem cells seldom divide and are mainly proliferatively quiescent. Recent studies have, however, shown that in multiple tissues long-label-retaining cells are actively dividing and asymmetrically retain their labeled template DNA strands while passing the newly synthesized DNA to their differentiating progeny. Interestingly, PI-MECs that have proliferated extensively in transplants give rise to LacZ-positive progeny, which retain the original DNA label for long periods and when pulsed with a second alternative DNA label prove to be actively traversing the cell cycle and thus become doubly labeled incorporating the second label into new DNA strands. Subsequent to a short chase period the second label is transferred along with the new DNA strands to LacZ-positive progeny (Fig. 4). This evidence demonstrates that during self-renewal PI-MECs produce progeny (in addition to luminal and myoepithelial offspring) that behave as asymmetrically dividing stem/progenitor cells responsible for the steady-state maintenance of the diverse LacZ-positive mammary epithelial population in the resulting outgrowth.

Fig. 4.  PI-MEC long-label-retaining (3H-thymidine) with autoradiographic grains were doubly labeled with 5BrdU (darker nuclei) in transplant labeled with thymidine 7 weeks earlier following a 2-day pulse with 5BrdU and produced 5BrdU-labeled-only (arrows) daughters after a 6-day chase.

Stem cell biology in the mammary gland

Stem cells have the ability to produce progeny that are different from themselves and to self-renew. The presence of mammary stem cells in mice was proven in the 1950s when it was demonstrated that any portion of the mouse mammary gland, regardless of age or parity status, could regenerate a functional gland when transplanted into a fat pad cleared of endogenous epithelium in prepubescent female recipients. This model has become a powerful tool in the study of mammary stem cells as both tissue fragments and dispersed cells can regenerate an entire mammary gland. This is especially important because before any cell can be considered a stem cell, it must fulfill the following functional criteria, regardless of its biochemistry. It must self-renew, differentiate into at least one other cell type and participate significantly in the generation and maintenance of its tissue.

The study of human mammary stem cells has been hampered until recently by the lack of in vivo models. The generation of immunocompromised animals that minimizes tissue rejection has advanced study immensely. More recently Kuperwaser and colleagues have developed a new in vivo model in which they “humanize” a mouse mammary fat pad by transplanting human fibroblasts into the cleared murine fat pad. The human fibroblasts are allowed to establish prior to the addition of human mammary epithelial cells allowing for the epithelium to engraft in a more natural environment.

In the mouse mammary gland three distinct classes of mammary progenitor cells have been identified: multilineage progenitors, ductal-limited progenitor cells, and lobule-limited progenitor cells. These classes are distinguished by their functional properties evident during pregnancy. Multilineage progenitor cells have the capacity to generate any epithelial subtype; ductal-only progenitor cells do not develop secretory lobules during subsequent pregnancies following transplantation while lobule-limited progenitor cells develop functional secretory alveoli during pregnancy but do not generate a complex ductal network.

Stem cell isolation in the hematopoietic system has been at the forefront for years. Efficient flow cytometric techniques for stem cell isolation have allowed characterization of these cells. Using techniques based on the hematopoietic system, four types of mammary progenitor cells have been identified in humans based on cell surface marker expression phenotypes: multilineage, myoepithelial-restricted, and two classes of luminal-restricted. A number of different candidate markers have been investigated including CD10/CALLA, ESA (epithelial specific antigen), MUC-1 and the cytokeratins 8, 9, 18, and 19. Myoepithelial-restricted cells are isolated based on an EpCAMlowMUC1CD49f+phenotype. Luminal-restricted progenitors are isolated based on expression of MUC1 and either K19+K14 or K19K14 phenotypes.

A valuable in vitro tool for studying mammary stem cells is the newly developed “mammosphere” culture system where anchorage-independent cells generate clonal floating sphere structures. Mammospheres formed under these conditions possess multilineage progenitors that self-renew and begin to differentiate to form luminal and myoepithelial cells. We have used this system to demonstrate that lobule-limited PI-MECs self-renew in culture. Mammospheres can be transplanted and recapitulate an entire gland. A small number of cells within mammospheres also segregate BrdU-labeled DNA suggesting asymmetric cell divisions (Fig. 5). Mammosphere cultures that were exposed to differentiating conditions, such as activation of the Notch or Hedgehog pathways, showed an increase in secondary mammospheres formed after dissociation as well as an increase in branching morphogenesis and cellular proliferation. This indicates that both Notch and Hedgehog are involved in the self-renewal and differentiation pathways in mammary stem cells. These molecular pathways, along with the Wnt pathway, are involved in the maintenance of the hematopoietic stem cell hierarchy.

Fig. 5.  A mammosphere after 10 days in anchorage-independent culture conditions. The arrow shows a BrdU labeled nucleus (see Color Plates).

Two different methods have been used to demonstrate that a single cell can regenerate a mouse mammary gland by acting as a multilineage mammary stem cell. Kordon and Smith showed first in 1998 through serial transplantation studies that a single retrovirally tagged mammary cell formed a mammary gland comprised of ductal and luminal components of both luminal and myoepithleial lineages. Later in 2006, two groups published their findings where they demonstrated that by transplanting a single, visually confirmed cell that was fluorescently sorted on the basis of being CD24+ Sca-1lowCD49fhighCD29high and not expressing CD45, Ter119, or CD31 they were able to generate a mammary gland. CD24 acts as a P-selectin ligand and may modulate integrin function while CD29 is β1-integrin and CD49f is α6-integrin. CD29 and CD49f are markers for basal epithelia suggesting that the basal epithelial compartment contains the mammary stem cells. Furthermore, only luminal epithelial cells are ERα+, which have almost no fat pad outgrowth potential.

Cancer stem cells

Stem cells are believed to be the targets of tumorigenesis in the mammary gland. Most DNA mutations rely on replication and cellular division. It is hypothesized that stem cells protect themselves against such mutations through asymmetric division. Since the majority of breast cancers are heterogeneous this argues that malignant transformation originates in cells that can generate progeny of different lineages.

Cancer stem cells do not have to arise from normal somatic stem cells. Another theory of mammary tumor heterogeneity is that the different types of mammary cells (luminal, basal, myoepithelial) vary in degrees of susceptibility to genetic mutations and transformation. Transit amplifying cells or differentiated cells may generate cancer stem cells if a genetic mutation reactivates a self-renewal pathway, although there is as of yet no direct evidence of this occurring.

Approximately 15–21% of human breast cancers present a basal phenotype characterized by the lack of expression of ER, PR, and erbB2. These “triple negative” breast cancers have a poor prognosis. Compared with patients having other forms of breast cancer, these patients are more likely to have a recurrence within 5 years and the mortality rates are correspondingly higher. Triple negative tumors are associated with BRCA1 and usually occur in younger women. BRCA1 (breast cancer 1) mutations are synonymous with basal breast cancer. It has been hypothesized that BRCA1 acts as a stem cell regulator and loss of BRCA1 in a conditional knockout model demonstrates a proliferation defect during pregnancy. BRCA1 tumors often have mutated p53 and Trp53, key cell cycle regulators.

In 2003, Al-Hajj and colleagues isolated tumor-initiating cells as determined by xenotransplatation into immunocompromised NOD/SCID mice. The tumor-initiating cells isolated from eight pleural effusions and one primary tumor had an EpCam+ CD44+ CD24−/low phenotype. They found that the frequency of tumor-initiating cells within that subpopulation of the tumor cells was <1%. This observation agrees with other human tumor types including colon, brain, and pancreatic cancers. However, clinical studies have confirmed through both breast cancer cell lines and breast tumors that the CD44+/CD24 phenotype is not associated with patient outcome or metastasis. In light of these and other studies the question that has arisen is as follows: are CD44+/CD24 cells better at adapting to the mouse mammary environment or are they really more tumorigenic in humans?

The human breast stem cell niche in both normal and tumor tissues has been characterized as CD44+CD24 and PROCR+ (a basal marker) but CD10 (a myoepithelial marker). Gene expression profile analysis showed that CD44+CD24 PROCR+ CD10 cells from normal tissue are more similar to CD44+CD24 PROCR+ CD10 tumor cells than to normal CD24+ cells.

Mammosphere initiating cells from the MCF-7 and MDA-MB-231 cell lines that are CD44+CD24low/− have an increased radio resistance. CD133+ glioma stem cells display similar resistance due to improved DNA damage repair as compared to nonstem cells. “Differentiation therapy” of BMP4 can eliminate glioma stem cells, and other molecular therapies have shown promise in eliminating leukemia stem cells. Better understanding of mammary stem cells, both normal and tumor-initiating, will help to devise new therapies that will allow elimination of tumor-initiating cells through molecular targeting.

Pregnancy has a dual effect on human breast cancer (protection or promotion), depending on the age of an individual, the period after a pregnancy, and the genetic predisposition. In genetically engineered strains of mice that are highly susceptible to mammary tumorigenesis and exhibit accelerated tumor development in postpartum or multiparous females, one might expect that PI-MECs serve as targets for neoplastic transformation. PI-MECs represent reporter gene marked lobule-limited mammary epithelial stem/progenitor cells that persist following pregnancy, lactation, and involution. These pluripotent cells exist in nulliparous glands but do not express the reporter gene, which is activated by Cre-lox recombination upon expression of Cre from the whey acidic protein (WAP) promoter, hence parity-identified. The unique growth properties of PI-MECs (i.e., responsiveness to pregnancy hormones, survival during involution, and ability to self-renew) make this epithelial subtype a potential target for pregnancy-associated tumorigenesis. Transgenic mice expressing the wild-type neu oncogene under the transcriptional regulation of the MMTV-LTR are suitable for studying the involvement of PI-MECs in pregnancy-associated mammary tumorigenesis since this animal model exhibits a relatively long latency of tumorigenesis (T50 of 205 days). MMTV-neu mice generate ERα-negative lesions that exhibit pathological features similar to a subset of human breast cancers. More importantly, the overexpression of neu/erbB2/HER2 has been observed in a significant subset of pregnancy-associated breast cancers in humans. Using this animal model, Henry et al., demonstrated that (a) multiparous females consistently exhibited accelerated tumorigenesis compared with their nulliparous littermate controls in a mixed genetic background and (b) PI-MECs were, indeed, primary targets of neoplastic transformation in this model. Interestingly, the significantly fewer lesions that arose in nulliparous controls originated from hormone-responsive cells that transiently activated WAP-Cre (i.e., an epithelial subpopulation that represents only 1–4% of all epithelial cells in the virgin gland) during estrus. The de novo generation and amplification of a large number of hormone-responsive and apoptosis-resistant epithelial cells (i.e., PI-MECs) during the first and subsequent reproductive cycles might, therefore, account for the significantly increased cancer susceptibility of parous MMTV-neu transgenic females.

We have hypothesized that if PI-MECs are the targets for MMTV-neu induced tumors then neu-transformed PI-MECs might represent a subpopulation of tumor-initiating stem cells responsible for maintenance and expansion of the MMTV-neu-induced tumorigenic population. We mixed increasing dilutions of MMTV-neu tumor cells collected from WAP-Cre/Rosa26R/MMTV-neu females with normal wild-type mammary epithelial cells from FVB/N mice. These mixtures were inoculated into the epithelium-divested mammary fat pads of immunocompromised female hosts. Our results demonstrate that some MMTV-neu-initiated tumor cells are capable of interacting with the normal epithelium contributing to normal mammary gland growth and regeneration, and in the process produce both luminal and myoepithelial epithelial progeny.

Influence of the microenvironment on stem cell biology and cancer

Stem and progenitor cells in most adult tissues reside in specialized, highly regulated microenvironments called stem cell niches. In general terms, niches are made up of signaling cells, characteristic of extracellular matrix (ECM), soluble mediators, and the stem cell (Fig. 6). The niche interprets a myriad of signals and controls whether stem cells remain quiescent, expand via symmetric division, or self-renew while contributing to the pool of progenitor cells by dividing asymmetrically. A constant dialogue made up of physical and molecular interactions between stem cells and the “support” cells of the niche shelters stem cells from differentiation or apoptotic stimuli that might deplete their stores. Homeostasis is maintained by the niche by balancing the need for additional differentiated progeny with the long-term protection against overproduction, which if not properly controlled may lead to cancer. One might even argue that adult or somatic stem cells have limited function outside their niche and that the justification for studying the niche emerges from its ability to impose function on its resident stem cells.

Fig. 6.  Schematic of a generic somatic stem cell niche showing the various elements that participate in regulating the fate of the resident stem cell. The distance from the ultimate stem cell position may determine the relative strength of each of these signals and thus contribute to the differentiation of the daughter cell. ECM extracellular matrix.

The idea of a niche as a specialized microenvironment sheltering stem cells was first proposed by Schofield almost 30 years ago in the context of mammalian hematology. However, most of the early evidence supporting the niche as a physical entity came from invertebrate models, like the gonads of Drosophila and C. elegans. In both cases the germ cells reside at the distal tip of a tapered organ and depend on the interaction with surrounding somatic cells to maintain their undifferentiated state. The idea that heterologous cell types compose the niche has now been well documented and drives much of our search for similar cell-based niche components in mammalian tissues. Whether normal stem cell function requires multiple cell types has recently been brought into question. Two reports showed that a posterior midgut population of cells in Drosophila could renew and differentiate into enterocytes and enteroendocrine cells in a clonal fashion. These cells do not appear to be in contact with a heterologous cell type but do sit on the basement membrane. In addition, two papers showed the successful transplantation of single mammary stem/progenitor cells that generated full ductal outgrowths. These data suggest that the extracellular matrix and other noncellular constituents might be sufficient to regulate stem cell function in some tissues.

In the murine mammary gland a pool of stem cells produces a branching bilayered epithelial tissue with an inner luminal epithelial layer and an outer myoepithelial layer bounded by a basement membrane surrounded by a fatty stroma of adipocytes with vasculature, nerves, and lymph vessels. What is special about the mammary niche is its ability to generate these structures by coordinating local stimuli with systemic hormonal stimuli to generate new mammary structures. Three groups have shown that putative mammary stem cells are ERα-negative so it would follow that a subset of the support cells within or nearby the niche would be estrogen and/or progesterone responsive. However, because the location of niches in situ has not yet been identified, formal proof remains to be provided. In a recent review on how hormones influence the mammary niche, Brisken and Duss proposed that the pubertal niche and the adult niche are overlapping entities in terms of what cells compose them. A hormone receptor positive sensor cell and the stem cell are identical in both but any additional niche cells that are recruited in response to estrogens versus progesterones are different. They further speculated about whether there is a “special” subpopulation of receptor positive sensor cells that directly control the niche or whether their inductive effect is indirect and mediated through a sensed depletion of mammary progenitors. These hints should allow for a better stage-specific characterization of the mammary niche.

Despite knowing the combination of stem/progenitor cell surface markers outlined in the previous section, the physical location of the mammary niche remains elusive. Several reports have suggested that the cap region of the terminal end buds (TEBs) in mice is one possible location of the stem cell niche during estrogen-dependent ductal outgrowth. During ductal elongation the cap cells are able to directly interact with the stroma because the basal lamina is degraded locally suggesting that some stromal cells participate in this process. Bolstering this idea is the observation that wild-type epithelium does not grow out in an EGFR-deficient fat pad but EGFR-deficient epithelium does reconstitute a wild-type fat pad. This result coupled with the fact that EGFR mRNA is enriched in the stem cell compartment and is an important downstream effector of ERα signaling via paracrine amphiregulin argues that the pubertal niche has a stromal component. However, since the TEBs and cap cells no longer exist after the completion of the ductal tree, the cap region is an unlikely location for the adult stem cell niche.

Proliferation and differentiation (secondary side branching) resulting from the stimulation of the adult mammary niche are driven primarily by the cyclic release of progesterone during the estrous cycle. Unlike estrogen and its receptor, progesterone can stimulate proliferation of both PR-positive and PR-negative luminal epithelial cells. At least part of progesterone’s paracrine effects are thought to be mediated through the receptor activator of NF-κB ligand (RANKL) and Wnt-4. The Wnt signal is believed to act directly on the stem cell within the niche and trigger an asymmetric division followed by the differentiating daughter’s expansion induced by RANKL.

In general terms of the location of the adult niche, some older data suggested that the mammary ducts of mice and rats may contain stem cells. In virgin mice, putative stem cells were found in ducts rather than alveoli, and these were shown to regenerate ductal epithelium as well as to form new alveolar buds. Sca-1 was proposed as a marker of mammary stem cells in mice when its expression was observed in ducts as well as in invading TEBs. Earlier work from our laboratory corroborated this idea by showing that rudimentary ducts from postgestational mice, when transplanted into cleared fat pads of TGF-β1 transgenic mice, grew out and retained the capacity to reactivate lobular structures at late pregnancy. In addition, ductal niches appear to respond specifically to the MMTV-c-myc transgene by amplifying the stem cell compartment, which supported the previous findings in humans that an entire TDLU represents the progeny from a single early ductal progenitor. Recent evidence employing microdissection to collect organoids from reduction mammoplasties demonstrated that the terminal ducts are one of the major sites of stem cells in adults. Cells from these isolated structures produced self-renewing mammospheres, multilineage colonies in 2D culture and TDLU-like structures in 3D ECM cultures, while cells from the lobules did not display any of these capacities. A further subdivision of these human ductal cells revealed a subpopulation that double stain positive for both K14 and K19. K14 typically marks myoepithelial cells and K19 is expressed mostly on luminal cells. The expression of both lineage markers in the same cell suggests that it might identify a subpopulation of stem or progenitor cells in the human breast. It has even been proposed that K19 functions as a “switch keratin,” enabling a cell to transition from one type of cytoskeleton to the other. The same K14+/K19+ cells also express chondroitin sulfate, K6a, K15, and SSEA-4, which overlap nicely with stem cell markers from other tissues, including brain, hair follicle, and prostate. This set of putative stem cell markers coincides with the EpCam+/CD49f+ population discussed earlier. Taken together, these data suggest the likely location of the mammary stem cell niche to be most closely associated with the ducts. However, the data published to date do not rule out the possibility of other niches and it is unknown whether the niche in the postpubertal virgin gland remains the same throughout adulthood or after pregnancy. The fact that the parity-induced mammary epithelial cells (PI-MECs) produced luminal and myoepithelial lineages in ductal-limited and lobule-limited outgrowths when employed in limiting dilution transplantation experiments using parous mammary tissue from WapCreRosa26stopβGal mice and the fact that PI-MECs survive involution suggest that new niches might be established during the extensive proliferation of pregnancy. This same population of cells was later shown to be enriched for CD24+/CD49f+ cells corroborating their stem/progenitor status.

Unlike other tissues, where the niche has been associated with a particular anatomical structure, the fact that any portion of the mammary gland can be transplanted and generate a complete outgrowth capable of full lactation indicates that the mammary niche resides at regular intervals throughout the gland. Studies from other animal model systems have shown that adult stem cells are generally focal in their distribution and not necessarily colocalized with the majority of the transiently amplifying cells. For example, stem cells have been shown to reside in the hair follicle bulge and the proximal ducts of the prostate, both separate from the main site of proliferation in each tissue. In an attempt to ascertain how many niches (or stem cells) are responsible for generating a mouse mammary gland, several groups have tried to estimate stem cell frequency by serial dilution transplantation experiments. The estimates range from 1 mammary repopulating unit (MRU) in 200 dissociated cells to 1 in 5,000 cells.

Many human stem cell studies are forced to use ex vivo culture of putatively enriched stem cell or progenitor populations before transplantation into animal hosts. Several of these studies, upon first glance, appear to contain contradictory results about the repopulating ability of certain marked stem or progenitor populations. Confusion in these reports revolves around the Sca-1 status, inclusion in the side population, and whether endothelial cells were excluded from study. Most of the discrepancies in these papers can be explained if one considers the different culture conditions used. It has been suggested that any manipulation of mammary epithelial cells in culture selects for immature progenitors or transit amplifying cells with repopulating activity, but not stem cells. This might suggest that bona fide stem cells will not exist outside their niche; that stem cells are defined by their place in the niche. This idea originated in tissues where downstream progenitors (immediate daughters of the ultimate stem cell) can occupy vacated niches and reacquire stem cell traits. For example, the melanocyte stem cell niche in mice is located in the base of the hair bulb in the transient portion of the hair follicle. Stem cells occupy the very bottom of the follicle and as they divide their progeny ascend and line the sides of the follicle, becoming increasingly differentiated with increasing distance from the niche. When the stem cell is selectively ablated experimentally, committed progenitors traversed the interstitial space, took up residence in the vacated niche, and began to function like stem cells. Similarly, when endogenous skeletal muscle satellite cells were killed by irradiation, transplanted myoblasts derived from explanted, quiescent satellite cells could fuse with the existing muscle fiber and regenerate it after successive damage with snake venom. This same mechanism was also suggested when glands resulting from transplantation of epithelial cell into cleared fat pads underwent multiple rounds of pregnancy.

The dominance of the niche over the stem cell’s autonomous phenotype has also been demonstrated in several reports involving cells crossing lineage “boundaries” to regenerate “foreign” tissues. Mice that received bone marrow transplants from syngeneic donors, using transgenes as markers for self and nonself, revealed that bone marrow-derived cells could participate in repair of muscle damage. In three of the five studies, bone marrow-derived cells resided in the satellite cell position on muscle fibers after irradiation and were able to respond appropriately to damage cues brought on by the physiological response to exercise, but the conversion from bone marrow stem (progenitor) cell to muscle cell was not complete. Interestingly, myoblasts derived from satellite cells can easily and completely undergo myogenesis in culture after serum withdrawl but marrow-derived cells could only be pushed toward the myoblast fate by culturing them in myoblast-conditioned media or by coculturing them with a myoblast cell line. These experiments might suggest that the muscle microenvironment is important in maintaining the ability of nonmuscle-derived cells to exhibit myogenic characteristics. In addition, recent work from our lab has shown that the mammary microenvironment is dominant over testicular stem cells. Germinal stem cells from WapCreRosaLacZ mouse testis enriched for CD49f (α6-integrin) expression, when mixed with mammary epithelial cells, were able to self-renew and give rise to cells that expressed either myoepithelial or luminal cell markers and more importantly, showed evidence of milk production. Taken together, these experiments demonstrate that the lineage of adult tissue-specific stem cells can be redirected if given the right microenvironment. Therefore, consideration of the stem cell niche requires a reversal of the reductionist approach often adopted by molecular biologists. Indeed, it demands the reassembly and integration of the reductionist data at the level of the tissue or organ since at this level, structure is function.



comments powered by Disqus







Copyright (c) 2013. All rights reserved.