Abstract
The mucosal lining (endometrium) of the human uterus undergoes cyclical processes of regeneration, differentiation and shedding as part of the menstrual cycle. Endometrial regeneration also follows parturition, almost complete resection and in post-menopausal women taking estrogen replacement therapy. In non-menstruating species, there are cycles of endometrial growth and apoptosis rather than physical shedding. The concept that endometrial stem/progenitor cells are responsible for the remarkable regenerative capacity of endometrium was proposed many years ago. However, attempts to isolate, characterize and locate endometrial stem cells have only been undertaken in the last few years as experimental approaches to identify adult stem/progenitor cells in other tissues have been developed. Adult stem cells are defined by their functional properties rather than by marker expression. Evidence for the existence of adult stem/progenitor cells in human and mouse endometrium is now emerging because functional stem cell assays are being applied to uterine cells and tissues. These fundamental studies on endometrial stem/progenitor cells will provide new insights into the pathophysiology of various gynaecological disorders associated with abnormal endometrial proliferation, including endometrial cancer, endometrial hyperplasia, endometriosis and adenomyosis.
Adult stem cells
Stem cells are rare undifferentiated cells present in many adult tissues and organs. Their rarity and lack of distinguishing morphological features and specific markers make it difficult to identify their location in tissues (Bongso and Richards, 2004; Shostak, 2006). Rather, adult stem cells are defined by their functional properties: high proliferative potential, substantial self-renewal capacity and ability to differentiate into at least one type of mature functional progeny (Potten and Loeffler, 1990; Morrison et al., 1997; Weissman, 2002; Eckfeldt et al., 2005) (Figure 1).
Hierarchy of stem cell differentiation. Stem cells undergo asymmetric cell divisions, which enable them to self-renew and replace themselves or differentiate to give rise to committed progenitors. These proliferate and give rise to more differentiated transit amplifying (TA) cells, which rapidly proliferate and finally differentiate to produce many terminally differentiated functional cells with no capacity for proliferation. The possible relationship of colonies initiated by human endometrial epithelial and stromal cells to the hierarchical model is shown. We postulate that the large colonies are initiated by putative stem/progenitor cells and the small colonies initiated by putative TA cells. Reproduced with permission from Chan et al. (2004).
Self-renewal or the ability to produce identical daughter stem cells is required to maintain the stem cell pool in tissues. Asymmetric cell division is one mechanism for producing an identical daughter cell and a more differentiated daughter (Morrison et al., 1997). However, stem cells also undergo symmetric divisions either producing daughter stem cells or transit amplifying (TA) progenitors.
Differentiation is defined as a change in cell phenotype because of expression of genes associated with cellular function rather than cell division (Potten and Loeffler, 1990; Bach et al., 2000). Stem cells exhibit a wide range of differentiation potential. The zygote is totipotential, producing all cell types in the embryo and extraembryonic tissues (Gage, 2000; Eckfeldt et al., 2005). Human embryonic stem cells (hESC) are pluripotential stem cells found in the inner cell mass of the blastocyst, which differentiate into all cell types of the three embryonic layers: ectoderm, mesoderm and endoderm, as well as trophectoderm (Gage, 2000; Xu et al., 2002; Bongso and Richards, 2004; Trounson, 2006). The differentiation potential of these cells becomes increasingly restricted as embryonic development proceeds (Eckfeldt et al., 2005). Arising from their progeny are the multipotential somatic stem cells found in adult tissues that differentiate into the component cells of the tissue in which they reside. An example is the haemopoietic stem cell (HSC) (Weissman, 2000). Progenitors or tissue-specific stem cells are committed to a particular differentiation pathway and have limited ability to self-renew (Fuchs and Segre, 2000; Lakshmipathy and Verfaillie, 2005; McCulloch and Till, 2005).
TA cells have properties intermediate between stem cells and end-stage differentiated cells, with limited proliferative potential and inability to self-renew, but they undergo several rounds of cell division progressively acquiring differentiation markers as part of the cellular amplification process producing numerous terminally differentiated cells (Potten and Loeffler, 1990; Fuchs et al., 2004) (Figure 1).
The stem cell niche
Stem cells are regulated by a specific physiological microenvironment termed the niche (Schofield, 1978). Our current knowledge on the structure, function and operation of the germ stem cell and adult stem cell niches have been derived from elegant studies in model organisms, in particular the Drosophila melanogaster ovary (Xie and Spradling, 2000) and testis (Yamashita et al., 2005) and Caenorhabditas elegans gonad, where the location of each cell is well characterized (Ohlstein et al., 2004; Li and Xie, 2005). Far less is known about mammalian adult stem cell niches that are anatomically more complex and difficult to study. A common feature of the Drosophila and mammalian stem cell niche is the precise location of the stem cell in close relationship to one or more surrounding differentiated tissue or niche cells that together with the extracellular matrix and various secreted molecules provides a microenvironment to regulate key adult stem cell functions (Fuchs et al., 2004; Li and Xie, 2005). Signals from this niche microenvironment impinge on intrinsic adult stem cell signalling to regulate stem cell proliferation and cell fate decisions (Eckfeldt et al., 2005; Moore and Lemischka, 2006).
Adult stem cell niches vary in their cellular composition, structure and location in different tissues. They have been identified and well characterized for epidermal stem cells in the hair follicle bulge and interfollicular regions of epidermis (Morris et al., 2004; Tumbar et al., 2004; Ito et al., 2005), for epithelial stem cells in the intestinal crypts (Booth and Potten, 2000; Sancho et al., 2004), the periosteum and blood vessels for HSC (Nilsson and Simmons, 2004; Kiel et al., 2005) and neural stem cells in the subventricular and subgranular zones of the central nervous system (Doetsch, 2003). Niche cells have been identified as osteoblasts lining trabecular bone in the HSC niche (Calvi et al., 2003; Zhang et al., 2003), endothelial cells in the neural stem cell (Doetsch, 2003) and HSC niches (Kiel et al., 2005; Wilson and Trumpp, 2006) and subepithelial mesenchymal cells in the intestinal stem cell niche (Mills and Gordon, 2001; He et al., 2004). Adhesion molecules and niche cells anchor the adult stem cell during periods of stem cell inactivity and manage the asymmetric stem cell divisions for their controlled release out of the stem cell niche (Fuchs et al., 2004; Li and Xie, 2005). Adhesion molecules identified in stem cell niches include cadherins mediating adhesive interactions between stem and niche cell(s) and integrins, which interact with the extracellular matrix (Zhang et al., 2003; Fuchs et al., 2004; Wilson and Trumpp, 2006). Niche cells maintain adult stem cells in a dormant state (G0) through signalling pathways inhibitory for growth and differentiation, often involving transforming growth factor-β (TGF-β) and bone morphogenetic protein (BMP) family members (Li and Xie, 2005).
One of the key functions of niche cells is to sense the need for tissue replacement and communicate proliferative and differentiation signals to resident stem cells (Moore and Lemischka, 2006). Signalling pathways identified in Drosophila germ stem cell niches appear to be conserved across species from germ stem cell to adult stem cell niches (Eckfeldt et al., 2005) and include Wnt/β-catenin, BMP (Eckfeldt et al., 2005; Li and Xie, 2005; Moore and Lemischka, 2006), Notch and Hedgehog pathways. Growth factors such as fibroblast growth factor-2 (FGF-2), insulin-like growth factor and vascular endothelial growth factor have roles in certain niches.
Adult stem cell plasticity
An area of considerable controversy is the concept of adult stem cell plasticity. A substantial body of literature suggests that adult stem cell fates are not limited to the tissue in which they reside or within embryonic germ layer boundaries (Blau et al., 2001; Raff, 2003; Wagers and Weissman, 2004; Eckfeldt et al., 2005). Transdifferentiation describes the conversion of cells of one tissue lineage into a different lineage, with concomitant loss of original tissue-specific markers and function, and acquisition of markers and function of the new cell type, without an intervening cell division (Tosh and Slack, 2002; Wagers and Weissman, 2004). It involves nuclear reprogramming and represents a form of metaplasia or alteration of key developmental genes (Tosh and Slack, 2002; Pomerantz and Blau, 2004). This plasticity of adult stem cells results from changes in the extracellular environment and occurs in the setting of tissue damage (Blau et al., 2001; Tosh and Slack, 2002; Quesenberry et al., 2005).
Bone marrow stem cells are known to traffic via the blood stream and appear to be incorporated into damaged tissues, changing into skeletal muscle cells, neurons and glia, hepatocytes, endothelial cells, myocardial cells and epithelial cells of gut, skin and lung, whereas neuronal cells have produced blood cells and skeletal muscle cells (reviewed in (Blau et al., 2001; Raff, 2003). Plasticity has also been detected in tissues of gender-mismatched bone marrow-transplant recipients by the co-expression of newly acquired tissue-specific antigens in Y chromosome containing cells or by tracking genetically tagged [e.g. green fluorescent protein (GFP)] transplanted cells. In the clinical setting, chimerism has been detected in the liver, gut and endometrium of bone marrow-transplant recipients (Kleeberger et al., 2002; Okamoto et al., 2002; Taylor, 2004), and in solid organ allografts, circulating recipient cells have colonized donor hearts and kidneys (Lagaaij et al., 2001; Quaini et al., 2002; reviewed in Korbling and Estrov, 2003). These findings suggest that adult stem cells are more ESC-like than originally thought and have important clinical implications for therapeutics (Raff, 2003).
The concept of stem cell plasticity has been refuted because it is a rare event (Wagers and Weissman, 2004). Alternative explanations include the transplantation of multiple stem cell types, especially if the source is bone marrow or skeletal muscle, cell fusion, de-differentiation after in vitro culture or the presence of truly pluripotent cells residing in the adult (Eckfeldt et al., 2005; Lakshmipathy and Verfaillie, 2005). Greater rigor in demonstrating robust and persistent engraftment of stem cells in regenerating tissues as well as their functional capacity has been advocated, as many studies have relied solely on morphology and immunostaining (Wagers and Weissman, 2004; Lakshmipathy and Verfaillie, 2005).
Role of adult stem cells in tissue homeostasis, repair and regeneration
Adult stem cells play a key role in tissue homeostasis, providing replacement cells in regenerating tissues lost by apoptosis or injury (Li and Xie, 2005; Snyder and Loring, 2005). Adult stem cell function is highly regulated to ensure an appropriate balance in stem cell replacement and provision of sufficient differentiated mature cells for tissue and organ function, without production of tumours (Moore and Lemischka, 2006; Shostak, 2006). A balance between self-renewal and differentiation is imperative and appears to be regulated by the stem cell niche.
It has become apparent that adult stem cells not only reside and function in highly regenerative tissues like the bone marrow, intestine and epidermis where they produce a steady supply of differentiated cells to maintain their respective tissues but are also found in tissues of low cell turnover, such as neural, liver, prostate and pancreas. In these tissues, adult stem cells function in maintaining tissue homeostasis by replenishing functional tissue cells lost by apoptosis (Tsujimura et al., 2002; Dor and Melton, 2004; Clarke and Smith, 2005). Following injury, normally quiescent adult stem cells undergo cell division producing TA cells that rapidly proliferate, to repair the lost tissue with sufficient numbers of functional end-stage cells. In some cases, such as liver and pancreas, fully mature cells have the capacity to revert to a proliferative phenotype to effect tissue replacement (Dor et al., 2004) in a poorly understood process. It appears that mature cells may de-differentiate into stem-like cells with a concomitant change in transcriptional profile or fully mature cells have latent stem cell capacities (Guasch and Fuchs, 2005).
Identification of adult stem cells and stem cell assays
Identification and characterization of adult stem cells is a major challenge because of the paucity of these cells and the lack of defining cell surface markers that enable their prospective isolation. Stem cell assays have been developed for identifying the hierarchies of HSC, which are easier to analyse than other adult stem cells because of their non-adherent growth, circulation in the blood stream, albeit in low numbers, and the identification of specific surface markers. Although adaptation of stem cell assays to assess functions of adult stem cells in solid tissues is rapidly developing, it is proving a difficult task (Kaur et al., 2004). There is a real challenge to develop predictive surrogate assays for tissue stem cell activity in cell populations removed from their natural microenvironment (Kaur et al., 2004).
Stem cell markers
There are no specific phenotypic markers of adult stem cells. A major research effort to identify specific markers of adult stem cells from several tissues has recently been undertaken (Simmons and Torok-Storb, 1991; Li et al., 1998; Uchida et al., 2000; Shi and Gronthos, 2003; Welm et al., 2003; Richardson et al., 2004; Xin et al., 2005; Shackleton et al., 2006; Stingl et al., 2006). The goal is to prospectively isolate pure stem cell populations for characterization, identification of stem cell location in situ and eventual use in the clinic as cell-based therapies. It is important to note that expression of a stem cell marker does not necessarily imply stem cell activity, as all the currently known adult stem cell markers are also expressed on mature cells (e.g. CD34 is a HSC and mature endothelial cell marker). Before a population of cells is designated stem cells on the basis of phenotypic marker expression, it is necessary to validate stem cell function using at least one surrogate stem cell assay (Kaur et al., 2004).
Assays of stem cell activity
In vitro assays examining the functions of adult stem cells include clonogenicity, proliferative potential, self-renewal and differentiation, whereas in vivo assays include self-renewal and tissue reconstitution. Clonogenicity is the ability of single cells to initiate colonies of cells when seeded at extremely low seeding densities or by limiting dilution. These colony-forming assays have been extensively used for characterizing HSC and their progenitors and provide a read-out for assessing potential stem cell markers when first characterizing candidate adult stem cells (Pellegrini et al., 1999; Gronthos et al., 2000; Stingl et al., 2001; Chan et al., 2004; van Os et al., 2004).
Proliferative potential is assessed by determining the total number of progeny or population doublings from a single cell or population of putative stem cells by serial passage until senescence (Li et al., 1998; Pellegrini et al., 1999; Stingl et al., 2001; Gronthos et al., 2003; Reynolds and Rietze, 2005). Serial cloning assays measure self-renewal of putative stem cells in vitro (Loeffler and Roeder, 2004) and serial transplantation in vivo (Kordon and Smith, 1998; Shackleton et al., 2006). These assays rely on the ability of the initial colony-forming cell/unit (CFU) to undergo a self-renewing division during colony formation or on transplantation and that the daughter cell has the same colony-initiating capacity on replating at limiting dilution or on re-transplantation (Dontu et al., 2003; van Os et al., 2004). Adult stem cells should self-renew multiple times and hence undergo multiple rounds of serial cloning. These techniques identify adult stem cells and progenitor cells, the difference being the extent to which they can undergo self-renewal (Reynolds and Rietze, 2005).
The differentiation potential of candidate stem cells is evaluated after culturing the cells in differentiation-induction media or transplanting them and analysing the tissue formed for expression of phenotypic differentiation markers or expression of tissue-specific transcription factors. Mesenchymal stem cells (MSCs) isolated from bone marrow, adipose tissue and dental pulp have been characterized for their capacity to differentiate into many mesenchymal lineages including adipocytic, smooth muscle, chondrocytic and osteoblastic lineages (Pittenger et al., 1999; Gronthos et al., 2000; Zuk et al., 2002).
The gold standard assay of adult stem cell activity involves in vivo reconstitution of the tissue from which the putative stem cell population was derived (Kaur et al., 2004; van Os et al., 2004; Joseph and Morrison, 2005). These techniques require the use of congenic strains of mice or immunocompromised mice, (e.g. NOD/SCID) as hosts for the engraftment of putative stem cell populations (Greiner et al., 1998). These assays are well established for murine stem cell populations and use sophisticated genetic approaches to distinguish the transplanted stem cell populations and their progeny from host cells (van Os et al., 2004; Burger et al., 2005; Xin et al., 2005; Shackleton et al., 2006). Other approaches include tagging cells with membrane intercalating dyes such as PKH26. Specific antibodies are used to detect species differences when transplanting human populations or when using congenic mouse strains. The value of these approaches is that transplanted cells and their derivatives can be identified within the architecture of the newly derived tissue (Booth et al., 1999; Gronthos et al., 2003). Competitive repopulation of stem cell and support cell populations enables quantification, but resident stem cells must first be depleted in the host using ablation regimes (Kaur et al., 2004; Shackleton et al., 2006).
Putative stem cell populations are often transplanted into ectopic sites, such as the kidney capsule or subcutaneous tissue (Booth et al., 1999). Although these sites do not recapitulate the tissue microenvironment of the stem cell niche or provide inductive cues, they do provide a rich vascular supply and contain the transplanted cells in a confined region (Xin et al., 2005).
Another in vivo approach for identification of adult stem cells in their stem cell niche is to capitalize on their quiescence (Braun and Watt, 2004). The label-retaining cell (LRC) approach uses a DNA synthesis label such as bromodeoxyuridine (BrdU) followed by a chase period to discriminate between adult stem cells, TA cells and their more differentiated progeny based on their relative frequencies of undergoing cell division. Infrequently, cycling stem cells retain the label for prolonged chase period (weeks to months), whereas the label is rapidly diluted in more mature, proliferating TA cells. LRC have been detected in epidermis, mammary gland and prostate (Morris and Potten, 1994; Tsujimura et al., 2002; Welm et al., 2002; Braun et al., 2003) and their stem cell properties demonstrated (Morris and Potten, 1994). More recently several elegant studies using conditional activation of fluorescently tagged transgenes under the control of tissue-specific promoters have been used to identify and characterize LRC in epidermis (Morris et al., 2004; Tumbar et al., 2004; Guasch and Fuchs, 2005).
Evidence for adult stem/progenitor cells in the uterus
Structure of the human uterus
The human endometrium of the uterus comprises the endometrial mucosal lining which is a highly regenerative tissue situated on the thick muscular myometrium (Figure 2). The endometrial–myometrial junction is irregular with no submucosal tissue to separate endometrial glandular tissue from the underlying smooth muscle of the myometrium (Uduwela et al., 2000). The endometrium and subendometrial myometrium originate from the Müllerian ducts during embryonic life, whereas the outer myometrial layer develops in fetal life and has a non-Müllerian origin (Noe et al., 1999; Uduwela et al., 2000). The endometrium is structurally and functionally divided into two major zones, the upper functionalis which contains glands extending from the surface epithelium loosely held together by supportive stroma and the lower basalis, consisting of basal region of the glands, dense stroma and lymphoid aggregates (Padykula, 1991; Okulicz et al., 1997; Uduwela et al., 2000; Spencer et al., 2005) (Figure 2). Both functionalis and basalis have been further subdivided into two morphologically distinct layers (Bartelmez, 1933; McLennan and Rydell, 1965; Kaiserman-Abramof and Padykula, 1989; Padykula, 1991), although others consider the distinctions between the four layers less obvious and prefer the concept of polarized microenvironments (Tabibzadeh, 1991a,b; Uduwela et al., 2000). The cellular components of human endometrium comprise luminal and glandular epithelium, stromal fibroblasts, vascular cells and leukocytes.
Full thickness human endometrium stained with epithelial marker epithelial cell adhesion molecule (Ep-CAM) demonstrating the functionalis and basalis layers (dotted lines). Note the branching glands near the endometrial–myometrial junction and the direct apposition of endometrial glands on the myometrium. Adapted from Chan RW (2006) Identification of human and mouse endometrial stem/progenitor cells. PhD Thesis, Monash University, Melbourne, Australia.
Indirect evidence for uterine stem/progenitor cells
Proliferation studies
The human endometrium has remarkable regenerative capacity, growing from 0.5–1 mm following menstruation to 5–7 mm in thickness each menstrual cycle (McLennan and Rydell, 1965) and is characterized by cyclic processes of cellular proliferation, differentiation and shedding. The concept that endometrial regeneration is mediated by stem cells located in basalis endometrium, rather than the functionalis or myometrium, was postulated many years ago (Prianishnikov, 1978; Padykula et al., 1984; Padykula, 1991). Evidence from early kinetic studies of endometrial cell proliferation shows zonal differences that predict an orderly replacement of endometrial epithelial and stromal cells from rarely proliferating putative stem/progenitors residing in the basalis near the endometrial–myometrial junction whose progeny are the rapidly proliferating TA cells observed in the functionalis (Ferenczy et al., 1979; Conti et al., 1984; Padykula et al., 1989; Okulicz et al., 1997). More recently, a disjunction between the proliferative index of endometrial glands in the basalis and functionalis has been observed in both the proliferative and the secretory stages of artificially cycled macaques using phosphorylated histone H3 as a proliferation marker (Brenner et al., 2003).
Endometrial gland methylation patterns
Endometrial epithelial stem cell kinetics has also been investigated by examining epigenetic errors encoded in methylation patterns of individual glands in human endometrium (Kim et al., 2005). This approach depends on endogenous DNA sequences becoming polymorphic as epigenetic variants arise during cell division when methylation at CpG sites alters within a particular gene. In a cyclically, remodelling tissue like human endometrium, a persistent polymorphism indicates heritable epigenetic variants in the stem cells, as those variants or mutations that occur in TA or mature cells are lost during shedding. Thus, the total number of stem cell divisions may be inferred from the numbers of somatic errors accumulated within individual glands (Ro and Rannala, 2001). Methylation sites of the CSX gene, a silent gene in human endometrium, were examined to ensure that any changes in methylation patterns would have no functional consequences but rather arose by random processes associated with ageing (Kim et al., 2005). The extent of methylation of endometrial glands increased with age until menopause, after which it remained relatively constant, indicating that the number of epigenetic errors was a reflection of the mitotic activity of endometrial stem cells (Kim et al., 2005).
Mathematical modelling of the data was more consistent with the concept that an individual gland contains a stem cell niche with an unknown number of long-lived stem cells rather than a single immortal stem cell that always divides asymmetrically. It would appear that symmetric and asymmetric cell divisions occurred in a stochastic manner to maintain a constant number of stem cells in the endometrial gland niche (Kim et al., 2005). Further evidence indicated that a reservoir of stem cells remained in atrophic endometrium, as there was no reduction in gland diversity after menopause. These random replication errors that accumulate over time provide a record of endometrial stem cell replication history, which may be useful for investigating a role for putative endometrial stem/progenitor cells in proliferative disorders of the endometrium.
Evidence from clinical observations
Further indirect evidence for endometrial stem/progenitor cells comes from primate studies and clinical practice, where endometrium completely regenerated and supported pregnancy after surgical resection of almost all endometrial tissue (Hartman, 1944; Wood and Rogers, 1993). In another clinical situation, pockets of endometrial tissue are observed to regenerate in a few women treated with electrosurgical ablation for menorrhagia (Tresserra et al., 1999), and pregnancies have been reported (Abbott and Garry, 2002). Clinical evidence suggesting the presence of adult stem cells in human endometrium comes from observations that it has the propensity to undergo ossification, often after termination of pregnancy, and although the calcified tissue is not of fetal origin, it is usually associated with chronic inflammation and trauma (Biervliet et al., 2004). These latter conditions are known to promote the incorporation of MSCs into regenerating tissues (van Os et al., 2004). In addition, tissues such as smooth muscle, bone and cartilage can also be found in endometrium (Bird and Willis, 1965; Roth and Taylor, 1966; Mazur and Kraus, 1980). MSCs have the capacity to differentiate into smooth muscle, fat, bone and cartilage in vivo and in vitro (Pittenger et al., 1999; Bianco and Robey, 2000; Gronthos et al., 2000; Short et al., 2003), and together, these observations suggest that under certain circumstances resident endometrial or bone marrow-derived multipotent MSCs may undergo an inappropriate differentiation.
Monoclonality of endometrial glands
More recent evidence indicates that endometrial glands are monoclonal in origin, suggesting that they arise from a single progenitor or stem cell. In almost half of histologically normal proliferative endometrial samples, rare glands have been observed that fail to express phosphatase and tensin homologue deleted on chromosome 10 (PTEN) protein (PTEN null glands) because of a mutation and/or deletion in the PTEN gene (Mutter et al., 2000). These PTEN-mutant glandular clones persist in the basalis region between menstrual cycles to regenerate their respective glands in the functional layer in subsequent cycles. PTEN null glands are increased in the endometrium of women in conditions of unopposed estrogen, particularly endometrial hyperplasia, a monoclonal epithelial proliferative disorder (Mutter, 2000; Mutter et al., 2001).
In a separate study, monoclonality was detected in carefully dissected individual endometrial glands using a PCR-based assay for non-random X chromosome inactivation of the androgen receptor gene (Tanaka et al., 2003a). Furthermore, adjacent glands up to 1 mm apart shared clonality indicating that well-circumscribed regions of endometrium were derived from the same precursor cell, suggesting that several glands share the same stem cell. This raises questions on the precise locality of candidate human epithelial stem/progenitor cells.
Evidence for uterine stem/progenitor cells from functional studies
Cell-cloning studies
Despite the likelihood that the amazing regenerative potential of human endometrium is mediated via resident stem/progenitor cells, it was only very recently that the first evidence based on functional assays was published (Chan et al., 2004; Gargett, 2004, 2006; Schwab et al., 2005). It was hypothesized that both epithelial and stromal or mesenchymal stem cells would be found in human endometrium, as a substantial stromal component supports both the glands and the surface epithelium (Gargett, 2004, 2006). Using purified single cell suspensions obtained from hysterectomy tissues, it was demonstrated that 0.22 ± 0.07% of endometrial epithelial cells and 1.25 ± 0.18% of stromal cells formed individual colonies of >50 cells/colony within 15 days when seeded at clonal density (Chan et al., 2004). These CFUs produced two types of colony for both epithelial and stromal cells. The large colonies containing >4000 cells were rare and comprised small, densely packed cells with high nuclear : cytoplasmic ratio that were postulated to be initiated by candidate endometrial stem/progenitor cells (Table I, Figure 1). The more common, small colonies were composed of large, loosely arranged cells with low nuclear : cytoplasmic ratio initiated by more mature CFU, probably TA cells (Table I, Figure 1). Further work has shown that the percentage of clonogenic epithelial and stromal cells in human endometrium does not vary significantly across the menstrual cycle (Schwab et al., 2005).
Cloning efficiency of human endometrial epithelial and stromal cells
| Clones | Clonogenicity (%) | ||
|---|---|---|---|
| Epithelial | Stromal | ||
| Large | 0.08 ± 0.03 | 0.02 ± 0.01** | |
| Small | 0.14 ± 0.04 | 1.23 ± 0.18** | |
| Total | 0.22 ± 0.07* | 1.25 ± 0.18* | |
| Clones | Clonogenicity (%) | ||
|---|---|---|---|
| Epithelial | Stromal | ||
| Large | 0.08 ± 0.03 | 0.02 ± 0.01** | |
| Small | 0.14 ± 0.04 | 1.23 ± 0.18** | |
| Total | 0.22 ± 0.07* | 1.25 ± 0.18* | |
Data are mean ± SEM for n = 16 epithelial and 13 stromal samples.
Data from Chan et al. (2004).
P = 0.0001; **P = 0.0002.
Cloning efficiency of human endometrial epithelial and stromal cells
| Clones | Clonogenicity (%) | ||
|---|---|---|---|
| Epithelial | Stromal | ||
| Large | 0.08 ± 0.03 | 0.02 ± 0.01** | |
| Small | 0.14 ± 0.04 | 1.23 ± 0.18** | |
| Total | 0.22 ± 0.07* | 1.25 ± 0.18* | |
| Clones | Clonogenicity (%) | ||
|---|---|---|---|
| Epithelial | Stromal | ||
| Large | 0.08 ± 0.03 | 0.02 ± 0.01** | |
| Small | 0.14 ± 0.04 | 1.23 ± 0.18** | |
| Total | 0.22 ± 0.07* | 1.25 ± 0.18* | |
Data are mean ± SEM for n = 16 epithelial and 13 stromal samples.
Data from Chan et al. (2004).
P = 0.0001; **P = 0.0002.
This contrasts with a cloning study on endometrial stromal cells derived from Pipelle biopsy tissues at various stages across the menstrual cycle, which demonstrated very high levels of cloning ranging from 45% in the early proliferative stage to 27% in the late proliferative stage using a limiting dilution assay and plating at 0.2 cells/well (Loughney and Redfern, 1995). However, in this study the average size of the clones after 14 days was only 12 cells. It would appear that the putative stem/progenitor cells may have been overlooked in this study because a total of 25 000 wells would be required to isolate one large stromal CFU. It also explains the inability to detect a relationship between proliferative potential and original colony size (Loughney and Redfern, 1995). In another limiting dilution cloning study, 13% of endometrial cells were clonogenic in a single sample that had been in continuous culture for 6 months (Tanaka et al., 2003b). The resulting clones exhibited high proliferative potential as they continued to proliferate for another 24 months in culture. It would appear that the culture conditions favoured self-renewing cell divisions of putative endometrial stem/progenitor cells present in the original sample.
Growth factor requirements for clonogenic endometrial cells
Several growth factors required for epithelial and stromal cell colony formation have been identified. In serum-free medium, clonogenic epithelial cells require either epidermal growth factor (EGF) or TGF-α or platelet-derived growth factor-BB (PDGF-BB) and fibroblast feeder layers to establish clonal growth (Chan et al., 2004; Schwab et al., 2005), indicating the importance of stromal–epithelial cell interactions in the endometrial epithelial stem/progenitor cell niche. It is likely that endometrial epithelial CFUs express EGF receptors, whereas PDGF-BB proliferative effects are probably mediated by PDGF receptor-β on the fibroblast feeder cells (Gargett, 2006). Clonogenic stromal cells also required EGF, TGF-α or PDGF-BB but were also clonogenic in FGF2-containing serum-free medium (Chan et al., 2004; Schwab et al., 2005), indicating that stromal CFUs express EGF receptors, PDGF receptor-β and FGF receptors. Whether combinations of these growth factors further enhance growth of clonogenic endometrial cells is not known.
Endometrial colony phenotypes
Cell type-specific markers were used to characterize cellular phenotype of the endometrial colonies and to identify differentiating progeny of the original CFU. Although small epithelial clones expressed epithelial cell adhesion molecule (Ep-CAM) and cytokeratin, the large colonies showed weak expression of these epithelial markers (Chan et al., 2004). However, they expressed α6-integrin (CD49f), also expressed on the basal membrane of endometrial epithelium in tissue, but not the stromal cell marker, collagen I. Both large and small stromal clones expressed CD90, a fibroblast marker (Chan et al., 2004). A significant proportion of stromal clones contained α-smooth muscle actin-expressing cells indicative of myofibroblast or smooth muscle cell differentiation. As it has been hypothesized that endometrial stem cells reside in the basalis, it is not surprising that stromal CFUs have differentiated into myofibroblast or smooth muscle cell lineages as there are considerable numbers of myofibroblasts in the basalis (Fujii et al., 1989; Uduwela et al., 2000), and inner myometrium develops from undifferentiated Müllerian duct mesenchyme (Bird and Willis, 1965; Konishi et al., 1984).
Stem cell activity of human endometrial clones
Current studies examining stem cell attributes of the rare epithelial and stromal CFU initiating large colonies have demonstrated their high proliferative potential as they undergo 30–32 population doublings before senescence or transformation (Gargett et al., 2005). Endometrial CFU also undergo self-renewing cell divisions in vitro as demonstrated by their serial cloning ability. This high proliferative potential of endometrial stromal cells has been noted earlier in kinetic growth studies of serially passaged bulk cultures (as opposed to CFU) where 50% of specimens underwent more than 24 population doublings, with several between 60 and 100 (Holinka and Gurpide, 1987). It is likely that several stromal stem/progenitor cells present in these cultures were responsible for this enormous proliferative capacity. Large secondary human endometrial stromal clones exhibit multilineage differentiation and similar to bone marrow and adipose tissue MSC differentiating into four mesenchymal lineages; adipocytes, smooth muscle cells, chondrocytes and osteoblasts, in vitro (Gargett et al., 2005). In contrast, epithelial and stromal cells initiating small colonies failed to serially clone and underwent limited proliferation, indicating that they are more differentiated endometrial cells with limited proliferative potential and self-renewal capacity (Gargett et al., 2005). These data suggest that the rare endometrial epithelial and stromal CFU initiating large colonies have characteristic properties of epithelial progenitor cells and MSC-like cells, respectively, that are probably responsible for the remarkable, cyclical and regenerative capacity of human endometrium (Gargett, 2006). Their presence in inactive and perimenopausal endometrium (Schwab et al., 2005) also explains the ability of atrophic endometrium to regenerate and is consistent with gland diversity demonstrated by epigenetic analysis of ageing endometrium (Kim et al., 2005). These initial findings lay the groundwork for further studies to characterize stem cell properties in vivo and for the discovery of specific endometrial stem/progenitor cell markers.
Markers of endometrial stem/progenitors cells
Currently, there are no known markers for endometrial epithelial stem/progenitor cells that distinguish them from their mature progeny. However, sorting endometrial stromal cells by magnetic beads or fluorescence-activated cell sorter (FACS) and examining the sorted populations for clonogenic activity have now excluded several potential markers. These include Stro-1 (Simmons and Torok-Storb, 1991), a bone marrow MSC marker, which did not enrich for endometrial stromal CFU and CD133 (Shmelkov et al., 2005), a HSC marker (Schwab and Gargett, unpublished observations). However, co-expression of two perivascular cell markers does give a significant 10-fold enrichment of multipotent, clonogenic stromal cells from human endometrium (Schwab and Gargett, 2006). This rare population of candidate endometrial MSC-like cells, postulated to reside near blood vessels in the basalis (Figure 3A), still need to be tested for stem cell properties in vivo. At this stage there has been no progress in identifying markers recognizing endometrial epithelial progenitor cells.
Schematic showing the possible location of putative endometrial stem/progenitor cells in (A) human and (B) mouse endometrium. In human endometrium, it is predicted that candidate epithelial stem/progenitors will be located in the basalis in the base of the glands and stromal stem/progenitors near blood vessels. In mouse endometrium, the location of epithelial and stromal label-retaining cells (LRCs) (candidate stem/progenitor cells) that have the capacity to rapidly proliferate during estrogen-stimulated growth of regressed endometrium is shown in the luminal epithelium and mainly near blood vessels at the endometrial–myometrial junction, respectively. Reprinted with permission from Gargett (in press).
There have been several expression studies examining endometrial tissues using antibodies to stem cell markers. Bearing in mind the caveats outlined previously, these studies may assist in determining markers worthy of examination for prospective isolation and subsequent testing in stem cell assays. Many of the classical stem cell markers have been examined in human endometrium for other purposes, but few have examined the basalis making them less informative. The expression of HSC markers, CD34, c-kit/CD117 and the survival marker, bcl-2, has been examined in hysterectomy tissues. Basically, this study found immunostaining for all three markers in basalis glands and stroma, with CD34 specific for basalis stroma (Cho et al., 2004). Flow cytometric analysis of fresh and cultured decidual stromal cells showed that some expressed Stro-1 and CD34, although their co-expression was not determined (Garcia-Pacheco et al., 2001). Whether any of these markers have value in identifying endometrial epithelial or stromal stem/progenitor cells is yet to be determined.
Recently, it was discovered that rare cells in the endometrial stroma of ∼44% of women (11/25) immunostain for Oct-4, a pluripotency marker and transcription factor of hESC (Matthai et al., 2006). Although the frequency of Oct-4+ endometrial stromal cells was not reported, nor their location in the basalis or functionalis documented and their cellular phenotype not determined, their presence suggests the possible presence of adult stem cells in human endometrium. Even though Oct-4 cannot be used to prospectively isolate endometrial MSC-like cells, as it is not a surface marker, its expression in human endometrial stromal clones and putative endometrial MSC-like populations should be examined.
Source of endometrial stem/progenitor cells
Remnant fetal stem cells
The embryonic female reproductive tract has its origins in the intermediate mesoderm, which begins to form soon after gastrulation. As this embryonic tissue proliferates, it is thought that some cells undergo mesenchymal to epithelial transition to give rise to the coelomic epithelium that later invaginates to form the paramesonephric or Müllerian ducts (Kobayashi and Behringer, 2003). These ducts comprise surface epithelium and underlying urogenital ridge mesenchyme. During fetal life, the glands commence developing as the undifferentiated uterine surface epithelium invaginates into the underlying mesenchyme, and smooth muscle differentiation of the mesenchyme commences to form the inner myometrium (Spencer et al., 2005).
A few fetal epithelial and MSCs are thought to remain in the adult endometrium and contribute to tissue replacement during its cyclic regeneration (Snyder and Loring, 2005). Whether there is an ultimate uterine stem cell that has capacity to replace all endometrial and myometrial cells, including epithelial, stromal, vascular and smooth muscle or whether there are separate epithelial and MSCs is not currently known. The different phenotypes, growth factor dependence and frequency of clonogenic endometrial epithelial and stromal cells suggest that there are at least two endometrial progenitor cells. However, this does not exclude the possibility of an unidentified, more primitive precursor in human endometrium. Sophisticated studies examining the relative reconstitution capacity of endometrial cells separated using lineage-specific and differentiation stage markers that have yet to be elucidated are required to distinguish these possibilities.
Circulating stem cells from the bone marrow
Another possible source of endometrial stem cells is the bone marrow. Increasingly, it is being recognized that bone marrow stem cells circulate, albeit in low numbers, and populate various organs (Blau et al., 2001). Significant chimerism, ranging from 0.2 to 52%, was detected using RT–PCR and immunohistochemistry in endometrial glands and stroma of four women who received single antigen HLA-mismatched bone marrow transplants, suggesting that bone marrow stem cells contributed to endometrial regeneration in a setting of cellular turnover and inflammatory stimuli (Taylor, 2004). It appears that the level of chimerism increased with time elapsed since transplantation, although the number of cases was small. The high level of chimerism may also be related to the original degree of endometrial damage and to loss of endogenous endometrial stem/progenitor cell populations resulting from pre-transplant conditioning or ongoing graft-versus-host disease involving the endometrium. The degree of endometrial chimerism observed is similar to that in other organs of bone marrow-transplant recipients (Korbling and Estrov, 2003).
The donor-derived cells were found in focal areas of glands and stroma suggesting local proliferation of the incorporated cells (Taylor, 2004). This observation together with the similar percentage of donor-derived epithelial and stromal cells in each patient suggests that there may be a single endometrial stem cell responsible for the production of both glands and stroma. Although most gland profiles observed were exclusively of the donor or host type, there was some chimerism within individual glands suggesting that not all were monoclonal (Taylor, 2004), which is consistent with the epigenetic error data (Kim et al., 2005) but contrasts with monoclonality described for non-transplanted women (Tanaka et al., 2003a). Polyclonal glands containing only a few donor cells could result from fusion of bone marrow-derived cells with endometrial cells, but this possibility was excluded after examining tissues stained with a DNA-complexing dye (Taylor, 2004). The endometrial epithelial and some stromal cells did not express CD45, a leukocyte marker, which distinguished them from donor endometrial leukocytes, although these would have been detected in the RT–PCR products.
These interesting observations rely solely on marker expression, but they raise many questions requiring further research to determine whether transdifferentiation of bone marrow stem cells into functional endometrial cells occurs. Bone marrow contains at least three different stem cells, HSC, MSC and endothelial progenitors, and all circulate. Which bone marrow stem cells or myeloid cells contribute to endometrial regeneration needs to be determined. Whether bone marrow-derived cells regularly engraft the endometrium during each menstrual cycle under normal physiological conditions or only occurs at the time of the bone marrow transplant or subsequently on the resumption of endometrial cycling is not known. Many mature bone marrow-derived cells traverse the endometrium on a regular basis. Whether the local tissue damage associated with menstruation is sufficient to attract bone marrow stem cells into the endometrium for permanent residence remains to be determined. Whether the bone marrow-derived stem cells incorporate into endometrial tissue in the basalis, functionalis or both regions is unknown. The studies required to answer some of these questions are not possible in humans and will rely on animal studies using transgenic reporter mice.
Evidence for adult stem/progenitor cells in mouse endometrium
The mouse is a well-established animal model for investigating endometrial function, although mice do not menstruate. Mouse endometrium undergoes cycles of cellular proliferation and apoptosis during its 4-day estrous cycle and can be induced to undergo substantial regeneration on administration of estrogen following ovariectomy (Martin and Finn, 1970; Martin et al., 1973). Mouse and human endometrium have a similar structure, although defined functionalis or basalis layers are not present in the mouse. Another similarity is the monoclonality of endometrial glands, as demonstrated in chimeric mice (Lipschutz et al., 1999), providing indirect evidence suggesting that single putative epithelial stem/progenitor cells are responsible for production of individual glands. One advantage of using mouse endometrium to investigate a role for endometrial stem/progenitor cells in endometrial growth is that glandular development is a post-natal event (Brody and Cunha, 1989; Spencer et al., 2005). The power of transgenic animals in dissecting molecular pathways also makes the mouse an attractive model.
LRCs in mouse endometrium
A reliable approach to identify adult stem cells and their location in the stem cell niche in vivo without the need for stem cell markers is the LRC technique. Important technical considerations in establishing the LRC technique in mouse endometrium included determining the optimal window for the initial labelling of most endometrial cells with BrdU, dosage of BrdU, length of chase before LRC analysis, defining LRC in immunostained sections, as a reduced intensity is observed after 1–2 cell divisions before BrdU becomes undetectable (3–4 divisions) (Lavker and Sun, 2000) and use of confocal microscopy to detect true co-localization of phenotypic markers and BrdU in individual cells. Using this approach, ∼3% of uterine epithelial and 6% of stromal cells were identified as LRC in C57BL/6J mice labelled on post-natal days 3–6 when the undifferentiated, luminal epithelium commences invagination into the mesenchyme to form the uterine glands (Brody and Cunha, 1989) and chased for 4–12 weeks (Chan and Gargett, 2006).
Endometrial epithelial LRC
Endometrial epithelial LRCs were observed as well as separated cells in the luminal epithelium rather than in clusters and were not present in glands, except for the occasional LRC in the neck of a gland (Figure 4A). This suggests that epithelial stem-like cells in luminal epithelium are responsible for the growth of glands in the adult mouse as well as during development. It is also likely that luminal epithelial LRCs have an important role in regenerating luminal epithelium, which undergoes substantial proliferation and apoptosis during the estrous cycle (Dharma et al., 2001; Mendoza-Rodriguez et al., 2002). Epithelial LRC may be located in inter-sites and may re-epithelialize the endometrium after parturition, although this still needs investigation. Although it is expected that glandular cells regenerate from luminal epithelium, they may have capacity to self-duplicate like pancreatic β cells (Dor et al., 2004) obviating the need for glandular epithelial stem/progenitors.
Confocal microscopy images showing (A, B) epithelial label-retaining cells (LRCs) and (C, D) stromal LRC in mouse endometrium. Post-natal BrdU-labelled mouse endometrium double immunostained for BrdU (red) and ER-α (green) showing (A) lack of co-expression in a single epithelial LRC in the luminal epithelium at 8-week chase (blue arrows) but ER-α expression in neighbouring mature epithelial cells and (C) co-localization of BrdU and ER-α in a stromal LRC beneath the surface epithelium at 12-week chase (white arrows). Endometrium from post-natal BrdU-labelled, 4- (B) and 8-week chased (D), ovariectomized mice double immunostained for BrdU (red) and proliferation marker, Ki-67 (green) to visualize estrogen-stimulated proliferation of epithelial LRC [(B) blue arrows] and of stromal LRC (D) at the endometrial–myometrial junction (white arrows). The x/z and y/z planes are shown on the far right and underneath the merged pictures demonstrating true co-localization of the two markers within individual whole nuclei. Dotted line indicates endometrial–myometrial junction. le, luminal epithelium; s, stroma; myo, myometrium. Scale bars: (A) 16 µm (B–D, 40 µm). Adapted from Chan and Gargett (2006) and Gargett (in press) with permission from AlphaMedPress and Taylor and Francis Medical Publishing.
Endometrial stromal LRC
Endometrial stromal LRC were observed adjacent to luminal and glandular epithelium (Figure 4C) and near the endometrial–myometrial junction (Figure 4D). Epithelial and stromal LRC did not express CD45, indicating that they were neither leukocytes nor derived from bone marrow cells (Chan and Gargett, 2006). Approximately 40% of stromal LRC were perivascular cells, either co-expressing α-smooth muscle actin or were in close association with CD31+ endothelial cells (Chan and Gargett, 2006) similar to bone marrow (Shi and Gronthos, 2003; Short et al., 2003) and human endometrial mesenchymal stem/progenitor cells (Figure 3). In a similar study using CD1 mice, 2–5% of stromal LRC were detected between 8- and 15-week chase, but no epithelial LRC was observed (Cervello et al., 2006). These stromal LRCs were also located near the endometrial–myometrial junction. One-third of these stromal LRCs co-expressed c-kit (CD117), the stem cell factor receptor found on HSC, and ∼10% expressed Oct-4, the pluripotent stem cell marker (Cervello et al., 2006).
Estrogen receptors in endometrial LRC
The estrogen receptor-1 (Esr1 or ERα) was differentially expressed in endometrial epithelial and stromal LRC (Chan and Gargett, 2006). Epithelial LRC did not express ERα, although they were adjacent to ERα-expressing luminal epithelium (Figure 4A), but 16% of stromal LRC expressed nuclear ERα (Figure 4C) (Chan and Gargett, 2006). Despite the lack of ERα expression, all epithelial LRC in BrdU-labelled, pulse-chased, ovariectomized mice underwent proliferation within 8 h of 17β-estradiol administration, as detected by co-localization of BrdU and the proliferation marker, Ki-67 (Figure 4B). In contrast, only some stromal LRC proliferated 8 h after 17β-estradiol, and these were mainly located near the endometrial–myometrial junction (Figure 4D). These data indicate that epithelial and some stromal LRC are recruited into cell cycle by estrogen, suggesting that they have the capacity to act as stem/progenitor cells and proliferate during physiological endometrial growth. The purpose and function of stromal LRC that fail to proliferate in response to estrogen stimulation are currently unknown, but they are unlikely to be endometrial mesenchymal stem/progenitor cells.
Endometrial epithelium is known to respond indirectly to estrogen via cell–cell interaction involving ERα-expressing stromal cells (Cunha et al., 1983; Cooke et al., 1997), which also appears to be recapitulated for epithelial LRC. The exact nature of the neighbouring stromal niche cells instructing epithelial LRC to proliferate is currently unknown (Figure 5). In contrast, estrogen appears to have a direct proliferative effect on some stromal LRC, and the role neighbouring vascular cells play in regulating stromal LRC function is currently unknown. These initial studies provide new insights and define a new approach to investigate the role of putative endometrial stem/progenitor cells in regulating epithelial and stromal growth in mouse endometrium.
Schematic of the putative endometrial (A) epithelial stem cell niche showing an ERα− epithelial stem cell (shaded cell) receiving indirect proliferative signals from surrounding ERα expressing mesenchymal niche cell. Subsequent signalling between mesenchymal niche cell and epithelial stem cell may be mediated via estrogen-induced release of epidermal growth factor (EGF) and/or transforming growth factor-α (TGF-α) from the niche cell to interact with stem/progenitor cell EGF receptors. (B) Mesenchymal stem niche in a perivascular location, indicating the possible pericyte and/or perivascular nature of the mesenchymal stem cells (MSC)-like cell (*), some of which express ERα (black nuclei). Surrounding ERα+ perivascular cells or endothelial cells may act as niche cells to regulate MSC-like cell proliferation through production of PDGF-BB, EGF, TGF-α or FGF2. Not all MSC-like cells respond to estrogen during endometrial regeneration. Endometrial MSC-like cells could be responsible for estrogen-induced growth of stromal tissue (perivascular MSC-like cells) and blood vessels (pericyte MSC-like cells).
Endometrial stem/progenitor cells in gynaecological disease
Several gynaecological conditions are associated with abnormal endometrial proliferation, and it is possible that putative endometrial stem/progenitor cells may play a role in the pathophysiology of diseases such as endometriosis, endometrial hyperplasia, endometrial cancer and adenomyosis (Gargett, 2004, 2006). Alterations in the number, function, regulation and location of epithelial and/or stromal endometrial stem/progenitor cells may be responsible for any one of these endometrial diseases. Furthermore, clonogenic epithelial and stromal cells are present in non-cycling and perimenopausal endometrium (Schwab et al., 2005) and may be responsible for regenerating endometrium in women who are given estrogen replacement therapy.
Cancer stem cells and endometrial cancer
There is increasing interest amongst cancer biologists and oncologists in the concept that cancers contain small numbers of cells with stem cell properties that are responsible for their growth and metastasis (Pardal et al., 2003; Miller et al., 2005). Emerging evidence indicates that there is a similar cellular hierarchy within tumours as in normal tissues (Figure 1), with adult stem cells and cancer stem cells having similar properties (Clarke and Fuller, 2006). Cancer stem cells have a long lifespan, enormous proliferative potential and self-renewal capacity enabling them to maintain and expand the cancer cell population, although they themselves are quiescent and rarely proliferate (Clarke and Fuller, 2006; Polyak and Hahn, 2006). They also produce abnormally differentiated cancer cell progeny with limited proliferative potential that form the bulk of the cancer (Reya et al., 2001). Many features of carcinoma can be explained by the stem cell concept, including clonal origin and heterogeneity of tumours, some associated with TA cells or progenitors, the mesenchymal influence on cancer behaviour, the local formation of precancerous lesions and the plasticity of tumour cells (Miller et al., 2005). Only a small proportion of the tumour actually comprises cancer stem cells, ∼0.02–1%. Thus, cancer stem cells act as precursor cells that produce the proliferating, more differentiated cancer cells killed by chemotherapy or radiation. Cancer stem cells differ from normal tissue stem cells in that their proliferation is no longer controlled by the neighbouring cells of the stem cell niche (Pardal et al., 2003; Polyak and Hahn, 2006). It is likely that cancer stem cells arise from adult stem cells that have accumulated multiple genetic and epigenetic changes over a period of time, giving them selective proliferative advantage and allowing their clonal expansion and succession in the stem cell niche (Miller et al., 2005). Thus, cancer stem cells may be the key tumour cells involved in the initiation, progression, metastasis and recurrence of tumours after treatment.
A key role of cancer stem cells has been established in acute myeloid leukaemia, breast cancer and glioblastoma (Pardal et al., 2003). Key genes involved in the self-renewal pathways that regulate adult stem cells, such as Wnt/β-catenin, sonic hedgehog and PTEN tumour suppressor gene are associated with a range of cancers (Beachy et al., 2004; Miller et al., 2005; Woodward et al., 2005). Microsatellite instability or mutations in the PTEN gene are known to be involved in endometrial hyperplasia and endometrioid endometrial carcinoma, but whether endometrial cancer stem cells are involved is not currently known. Furthermore, the Wnt-β-catenin signalling pathway is involved in endometrial carcinoma and endometrial stromal sarcomas (Latta and Chapman, 2002; Moreno-Bueno et al., 2002). It is currently unknown whether cancer stem cells have a role in endometrial cancers or in endometrial hyperplasia, but this is an area open for future research. Surgery is the main treatment for endometrial cancer (Amant et al., 2005); however, possible future therapies targeting cancer stem cells, perhaps blocking molecules regulating asymmetric cell division, may improve the prospects for those women with advanced or high grade, poorly differentiated type-2 tumours.
Endometriosis and endometrial stem/progenitor cells
Endometriosis, defined as the growth of endometrium outside the uterine cavity, is a common gynaecological disorder affecting 6–10% of women (Giudice and Kao, 2004). It is a major clinical problem causing inflammation, pain and infertility. Despite its common occurrence, little is known about its pathogenesis (Gazvani and Templeton, 2002; D’Hooghe et al., 2003). The most commonly held theory on the aetiology of endometriosis is that viable endometrial cells reach the peritoneal cavity through retrograde menstruation (Giudice and Kao, 2004). However, it is well known that menstrual debris is present in the peritoneal cavity of 90% of menstruating women, suggesting that only some endometrial cells from some women are capable of establishing endometriotic implants. Several theories have been suggested, including abnormal endometrium, genetic factors, altered peritoneal environment, reduced immune surveillance and increased angiogenic capacity (Starzinski-Powitz et al., 2001; Vinatier et al., 2001; Donnez et al., 2002; Giudice and Kao, 2004). It is possible that in 6–10% of women who develop endometriosis, endometrial stem/progenitor cells are inappropriately shed during menstruation and reach the peritoneal cavity where they adhere and establish endometriotic implants (Starzinski-Powitz et al., 2001; Gargett, 2004, 2006). Although long-term endometriotic lesions may develop from endometrial stem/progenitor cells, those that resolve may have been established by more mature TA cells.
The monoclonality of some endometriotic lesions (Jimbo et al., 1997; Tamura et al., 1998; Wu et al., 2003), the invasiveness of E-cadherin− N-cadherin+ endometriotic cells (Zeitvogel et al., 2001) and the increasing evidence that endometriosis can develop into ovarian clear cell and endometrioid carcinomas (Van Gorp et al., 2004) are consistent with the concept that endometriosis could have a stem cell origin. This is further supported by the demonstration of clonogenic cells in a long-term culture derived from a sample of endometriosis tissue (Tanaka et al., 2003b). If bone marrow stem cells have the capacity to seed the endometrium and transdifferentiate into functional endometrial cells, it is possible that they may have the capacity to behave similarly outside the endometrial environment, particularly in sites other than the peritoneal cavity (Taylor, 2004). Alternatively, it has been postulated that some forms of endometriosis may arise from remnant fetal Müllerian cells, which have characteristic stem cell properties of high proliferative potential, multipotency and self-renewal. Clearly, the role of endometrial stem/progenitor cells or bone marrow stem cells in the development of endometriosis will require an extensive research effort.
Adenomyosis and endometrial stem/progenitor cells
Adenomyosis, a condition affecting 1% of women, results from basal endometrium undergoing extensive invasion of the myometrium and is associated with smooth muscle hyperplasia; it is also considered to arise from fetal Müllerian cells (Ferenczy, 1998). It is possible that endometrial stem/progenitor cells or their niche cells demonstrate abnormal behaviour in adenomyosis, and these putative stem cells have an abnormally orientated niche such that their differentiating progeny move towards the myometrium rather than functionalis, producing pockets of endometrial tissue deep within the myometrium. Alterations in the putative endometrial stem cell niche, particularly in the niche cells regulating stem cell fate decisions may result in excessive smooth muscle differentiation of putative endometrial stem/progenitor cells producing the observed myometrial hyperplasia. Much research is required to establish whether there is a role for endometrial stem/progenitor cells in the pathogenesis of adenomyosis.
Conclusions and future directions for endometrial stem/progenitor cell research
Endometrial stem cell research is in its infancy, but major advances have already been made to identify rare populations of epithelial and stromal cells with progenitor activity in human and mouse endometrium that are probably responsible for its remarkable regenerative capacity. Whether there is a single more primitive endometrial stem cell that produces all endometrial and myometrial cell types is yet unknown, and whether the endometrial stromal progenitor has full MSC activity in vivo is yet to be determined. Much still needs to be done to fully identify and characterize endometrial epithelial and mesenchymal stem/progenitor cells. There is a need to identify specific markers of both cell types to allow their prospective isolation for molecular characterization and determining their location in tissue. Animal xenotransplantation models are needed to examine key stem cell properties of self‐renewal and tissue reconstitution in vivo. Whether the bone marrow is a source of endometrial stem/progenitor cells needs confirmation under physiological conditions, and whether bone marrow-derived cells are major or minor contributors to endometrial regeneration is yet to be determined. The putative endometrial stem cell niches for human epithelial and stromal stem/progenitors need further characterization. How estrogen and progesterone interact with endometrial stem/progenitor cells or their neighbouring niche cells needs clarification. The signalling pathways that regulate endometrial stem/progenitor cell activity have yet to be investigated, and candidate pathways involving the Wnt, hedgehog, BMP pathways and the HOX genes should be investigated as these molecules or pathways have already been detected in endometrium or have important roles during endometrial development.
As endometrial stem/progenitors cells become better characterized, their role in gynaecological disorders associated with abnormal endometrial proliferation can be assessed. This will not only increase the understanding of the pathogenesis of endometrial cancer, endometrial hyperplasia, endometriosis and adenomyosis but also has the potential to change the way these diseases may be treated in the future, particularly as therapeutic agents that target key stem cell functions become available.
Acknowledgements
The author thanks the National Health and Medical Research Council (ID 284344), the Royal Australian and New Zealand College of Obstetricians and Gynaecologists and the Australian and New Zealand Charitable Trusts for financial support, Rachel Chan, Kjiana Schwab and Rachel Zillwood for their contributions to these studies and Nancy Taylor and Nicki Sam for collection of human tissues.
References
Author notes
1Department of Obstetrics and Gynaecology, Monash University and 2Centre for Women’s Health Research, Monash Institute of Medical Research, Melbourne, Victoria, Australia
