Integrating physiological regulation with stem cell and tissue homeostasis

Summary

Stem cells are uniquely able to self-renew, to undergo multilineage differentiation, and to persist throughout life in a number of tissues. Stem cells are regulated by a combination of shared and tissue-specific mechanisms and are distinguished from restricted progenitors by differences in transcriptional and epigenetic regulation. Emerging evidence suggests that other aspects of cellular physiology, including mitosis, signal transduction, and metabolic regulation also differ between stem cells and their progeny. These differences may allow stem cells to be regulated independently of differentiated cells in response to circadian rhythms, changes in metabolism, diet, exercise, mating, aging, infection, and disease. This allows stem cells to sustain homeostasis or to remodel relevant tissues in response to physiological change. Stem cells are therefore not only regulated by short-range signals that maintain homeostasis within their tissue of origin, but also by long-range signals that integrate stem cell function with systemic physiology.

INTRODUCTION

Stem cells are critical for the development and maintenance of tissues. The zygote gives rise to pluripotent cells in the embryo then these cells give rise to multipotent, tissue-specific stem cells that complete the process of organogenesis during fetal development. In a number of tissues, including the nervous and hematopoietic systems, tissue-specific stem cells persist throughout life to regenerate cells that are lost to turnover, injury, and disease. Self-renewing divisions, in which stem cells divide to make more stem cells, allow stem cell pools to be expand during fetal development and then to persist throughout adult life. The capacity to remain undifferentiated and to self-renew throughout life distinguishes stem cells from other cells.

Stem cells are required for the maintenance and function of a number of adult tissues. In the central nervous system (CNS), stem cells persist throughout life in the forebrain lateral ventricle subventricular zone as well as in the subgranular zone of the hippocampal dentate gyrus (Zhao et al., 2008). Stem cells in both regions of the adult brain give rise to new interneurons that regulate the ability to discriminate new odors or certain forms of spatial learning and memory, respectively (Alonso et al., 2006; Gheusi et al., 2000; Zhang et al., 2008). Hematopoietic stem cells (HSCs) give rise to blood and immune system cells throughout life and HSC depletion leads to immunocompromisation and hematopoietic failure (Park et al., 2003; van der Lugt et al., 1994). Stem cells also persist throughout life in numerous other tissues, including the intestinal epithelium (Barker et al., 2007).

Stem cells differ from restricted progenitors as a consequence of both intrinsic and extrinsic regulation. Stem cells often depend upon transcriptional and epigenetic regulators that are not required by restricted progenitors or differentiated cells in the same tissues (He et al., 2009). The environment also regulates stem cell function as specialized niches regulate stem cell maintenance throughout life using strategies that are often shared across species and tissues (Fuller and Spradling, 2007; Morrison and Spradling, 2008; Scadden, 2006).

Physiological homeostasis requires regulation at many levels, from the molecular level within cells, to the cellular level within tissues, to the systemic level within organisms. Much of systemic homeostasis in organisms is regulated by differentiated cells (e.g. pancreatic β cells that sense changes in glucose and secrete insulin, neurons that sense environmental inputs and modulate physiological and behavioral responses, etc.). Stem cells contribute to homeostasis partly by generating and regenerating appropriate numbers of differentiated cells. However, stem cell function itself must also be modulated in response to physiological changes to remodel tissues to keep pace with changing physiological demands (Drummond-Barbosa and Spradling, 2001; Hsu and Drummond-Barbosa, 2009; McLeod et al., 2010; Pardal et al., 2007).

Data increasingly suggest that many aspects of cellular physiology differ between stem cells and their progeny. At least some aspects of metabolic regulation differ between stem cells and restricted progenitors. This is interesting because most of what we know about metabolic pathways comes from studies of cell lines and non-dividing differentiated cells (such as liver and muscle). As a result, it remains unclear whether most aspects of metabolism are regulated similarly in all dividing somatic cells or whether different kinds of dividing somatic cells employ different metabolic mechanisms. If systemic physiological homeostasis depends upon the concerted regulation of stem cell function in multiple tissues, then stem cells may have distinct metabolic mechanisms that allow them to respond to these physiological changes.

In this review we will discuss mechanisms by which stem cells respond to physiological changes such as feeding, circadian rhythms, exercise, and mating. One of the key challenges for the next 10 years will be to understand how stem cell regulation is integrated with the physiology of whole organisms to maintain systemic homeostasis.

The importance of executive control by transcriptional networks

Embryonic stem (ES) cells are derived from the inner cell mass of the blastocyst prior to implantation. They are pluripotent and have indefinite self-renewal potential. These features of ES cells are regulated by a unique transcriptional network involving Oct4, Sox2, and Nanog (Jaenisch and Young, 2008). These transcription factors form a core auto-regulatory network that maintains pluripotency by inducing genes that promote self-renewal and by repressing genes that drive lineage restriction. Other epigenetic (Jaenisch and Young, 2008), transcriptional (Dejosez et al., 2008), and signaling (Ying et al., 2008) regulators collaborate with this network to sustain the pluripotent state. While the cell cycle (reviewed in (He et al., 2009)) and some aspects of metabolism (Wang et al., 2009) are also regulated differently in pluripotent stem cells as compared to other cells, it remains unclear how pervasive the differences in cellular physiology are relative to other cells.

The fact that differentiated adult cells can be reprogrammed to a pluripotent state by ectopically expressing a small number of pluripotency-associated transcription factors argues that all differences between pluripotent stem cells and other cells can be determined by this transcriptional network. Pluripotency can be induced in differentiated mouse and human cells by expressing Oct4 along with various combinations of other transcription factors (Park et al., 2008; Stadtfeld and Hochedlinger, 2010; Takahashi et al., 2007; Takahashi and Yamanaka, 2006; Yu et al., 2007). These induced pluripotent stem (iPS) cells resemble ES cells in terms of gene expression, cell cycle regulation, teratoma formation, and metabolic regulation (Prigione et al., 2010; Stadtfeld and Hochedlinger, 2010). Importantly, mouse iPS cells have the ability to generate a viable adult mouse upon injection into blastocysts (Boland et al., 2009; Zhao et al., 2009). This means that all of the aspects of cellular physiology that are necessary for pluripotent cells to differentiate into normal specialized cells can be induced by these transcription factors.

On the other hand, recent studies have identified epigenetic aberrations in iPS cells that indicate these cells are often not fully reprogrammed to a normal pluripotent state (Kim et al., 2010; Lister et al., 2011). This raises the question of whether some differences in cellular physiology, or at least epigenetic state, are regulated independent of the transcriptional network and whether these differences might stabilize the pluripotent state.

Stem cell self-renewal is different from restricted progenitor proliferation

Tissue-specific stem cells depend on transcription factors that regulate stem cell self-renewal but not restricted progenitor proliferation. The Sox17 transcription factor is required for the maintenance of fetal and neonatal HSCs but is not expressed by the vast majority of restricted progenitors in the hematopoietic system (Kim et al., 2007). Sox17 is not expressed by neural stem cells but other Sox family transcription factors likely perform similar functions in neural stem cells. Sox2 and Sox9 are required by CNS stem cells during fetal development as well as in the adult brain (Avilion et al., 2003; Favaro et al., 2009; Graham et al., 2003; Scott et al., 2010). Sox10 is required to maintain neural crest stem cells during peripheral nervous system (PNS) development but is not required by the restricted neuronal progenitors that arise from these cells (Kim et al., 2003). Different Sox family members are therefore required to maintain undifferentiated stem cells in different tissues during fetal development.

Prdm family transcription factors are also required by stem cells in multiple tissues. Prdm14 is required by primordial germ cells and stabilizes ES pluripotency (Chia et al., 2010; Yamaji et al., 2008). Prdm1/Blimp1 is required for primordial germ cells and progenitors in the sebaceous gland (Horsley et al., 2006; Ohinata et al., 2005). Prdm16 is required by stem cells in the hematopoietic and nervous systems but not by most restricted progenitors in the same tissues (Chuikov et al., 2010). Key transcription factors, or transcription factor families, therefore promote stem cell self-renewal in multiple tissues without necessarily promoting the proliferation of restricted progenitors in the same tissues (Figure 1A).

Figure 1. Some mechanisms promote stem cell self-renewal in multiple tissues but do not promote the proliferation of all progenitors.

A. A schematic depiction of stem cells from two different tissues giving rise to their respective restricted progenitors and differentiated cells. The yellow nucleus denotes that some transcriptional and epigenetic mechanisms that cell-intrinsically promote stem cell maintenance are conserved across tissues; however these mechanisms are not necessarily required by restricted progenitors or differentiated cells from the same tissues. Examples of this include Bmi-1, Sox17, Prdm16, and Hmga2 (see text for details and references). B. Stem cells also depend upon extrinsic short-range signals from the niche (yellow cells) for their maintenance. Niches in different tissues and different species often employ similar strategies, and even similar secreted factors, to promote stem cell maintenance. Stem cells often depend upon Wnt, Notch, and/or BMP ligands secreted by the niche for their maintenance. Exposure to this short-range signal from the niche distinguishes stem cells (yellow nucleus) from progenitors fated to differentiate and the size of the niche determines the number of stem cells in the tissue. Cells displaced from the niche by cell division or competition are fated to differentiate (Fuller and Spradling, 2007; Morrison and Spradling, 2008). C. Stem cells are also extrinsically regulated by long-range signals that reflect nutritional status, oxygen level, hormonal status, or other physiological changes.

Transcription factors collaborate with epigenetic regulators to maintain undifferentiated stem cells. The polycomb family chromatin regulator, Bmi-1, is required for the maintenance of postnatal stem cells in multiple tissues, including the hematopoietic and nervous systems, but not for the proliferation of most restricted progenitors in the same tissues (Lessard and Sauvageau, 2003; Molofsky et al., 2003; Park et al., 2003). The trithorax protein Mll is required for the maintenance of HSCs but not for the proliferation of restricted myeloid and lymphoid progenitors (Jude et al., 2007; McMahon et al., 2007). Mll is also required for neurogenesis by CNS stem cells but not for gliogenesis (Lim et al., 2009). Differences between stem cell self-renewal and restricted progenitor proliferation are not absolute as some restricted progenitors, such as lymphoid progenitors and cerebellar granule precursor cells, also depend on Bmi-1 for their proliferation (Leung et al., 2004; van der Lugt et al., 1994). Nonetheless, these transcriptional and epigenetic mechanisms do not generically regulate the proliferation of all cells, even when the mechanisms are widely conserved among stem cells in multiple tissues.

Cell cycle regulation also distinguishes stem cells from restricted progenitors in the same tissues. In some adult tissues, the stem cells are quiescent most of the time while most restricted progenitors divide more frequently. A good example is the hematopoietic system, where only a few percent of HSCs are in cycle at any one time (Kiel et al., 2007) and a subset of HSCs divide only once every few months (Foudi et al., 2009; Wilson et al., 2008). Although most restricted hematopoietic progenitors divide much more frequently, there are some restricted hematopoietic progenitors, including lymphoid progenitors (Pelayo et al., 2006) that can reversibly enter and exit the cell cycle over long periods of time, much like HSCs. As a consequence, bromo-deoxyuridine label retention is not a sensitive or specific marker of HSCs (Kiel et al., 2007), but can be used in concert with other HSC markers to identify a slowly-dividing subset of HSCs (Foudi et al., 2009; Wilson et al., 2008). There is also evidence that some adult neural stem cells (Doetsch et al., 1999; Morshead et al., 1994; Pastrana et al., 2009) and hair follicle stem cells (Blanpain et al., 2004; Cotsarelis et al., 1990; Tumbar et al., 2004) are quiescent much of the time. However, quiescence is not a defining feature of stem cells as stem cells in each of these tissues divide rapidly during fetal development (Lechler and Fuchs, 2005; Morrison et al., 1995; Takahashi et al., 1995) and can be reversibly recruited into cycle, such as after tissue injury (Doetsch et al., 1999; Harrison and Lerner, 1991; Kobielak et al., 2007; Lugert et al., 2010). Moreover, stem cells in the intestinal epithelium divide every day (Barker et al., 2007), demonstrating that even facultative quiescence is not an obligate feature of adult stem cells.

Stem cells and restricted progenitors can also differ in terms of cell cycle control. While neural stem cells are regulated by the cyclin-dependent kinase inhibitor, p21^Cip1 (Kippin et al., 2005), another family member, p27^Kip1, regulates restricted progenitor proliferation (Cheng et al., 2000; Doetsch et al., 2002). Other cell cycle regulators and tumor suppressors consolidate the transition of stem cells into transit amplifying progenitors by negatively regulating self-renewal. Deletion of p16^Ink4a, p19^Arf, and p53 dramatically expands HSC frequency by restoring long-term self-renewal potential to progenitors that normally only transiently self-renew (Akala et al., 2008). These tumor suppressors also limit the reprogramming of fibroblasts into iPS cells (Banito et al., 2009; Hanna et al., 2009; Hong et al., 2009; Kawamura et al., 2009; Li et al., 2009; Marion et al., 2009; Utikal et al., 2009). Tumor suppressors that negatively regulate cell cycle progression thus inhibit the acquisition of stem cell identity, perhaps by negatively regulating self-renewal.

Access to niche signals can distinguish stem cells from restricted progenitors

Many stem cells reside in specialized microenvironments, called niches, which promote stem cell maintenance and regulate stem cell function (Morrison and Spradling, 2008). One of the best characterized niches is the Drosophila testis in which spermatogonial stem cells reside at the apical tip of testis, anchored to hub cells through DE-cadherin and β-catenin/armadillo-mediated adherens junctions (Figure 1B) (Fuller and Spradling, 2007). In addition to anchoring stem cells within the niche, hub cells secrete short-range signals (Unpaired, a ligand that activates JAK/Stat signaling, and Decapentaplegic, a BMP homolog) that promote stem cell maintenance. Spermatogonial stem cells divide asymmetrically, oriented by the axis created by mother and daughter centrosomes, such that one daughter cell remains undifferentiated within the niche and the other daughter cell is displaced from the niche and fated to differentiate (Figure 1B) (Yamashita et al., 2007). Short-range niche signals can therefore determine the size of the stem cell pool (based on the space available in the niche) and which cells are fated to differentiate (based on whether they are displaced from the niche) (Figure 1B).

The C. elegans germline niche is conceptually similar in that spatially restricted Notch ligands expressed by the distal tip cell at the end of the gonad are required for the maintenance of undifferentiated stem cells. Cells displaced from the distal tip are fated to differentiate (Kimble and Crittenden, 2007). Unlike the Drosophila germline there is no evidence that C. elegans germline stem cells undergo asymmetric divisions. Stem cells thus undergo both asymmetric and symmetric divisions within their niches, depending on tissue and developmental context (reviewed in (Morrison and Kimble, 2006)).

Mammalian tissues also have specialized niches that secrete short-range factors that promote stem cell maintenance (Morrison and Spradling, 2008). As in the niches characterized in Drosophila and C. elegans, Notch ligands, BMPs, and Wnt proteins have been implicated in the regulation of stem cell maintenance in multiple mammalian tissues including in the CNS (Doetsch, 2003) and in hair follicles (Blanpain and Fuchs, 2006). These factors are presumed to be locally-secreted by supporting cells that create the niches, though the identities of these supporting cells are not yet well characterized in most mammalian tissues.

Stem cells are also extrinsically regulated by long-range signals, including an evolutionarily conserved role for insulin pathway regulation (Figure 1C). Circulating insulin-like peptide is required for the maintenance of Drosophila germline stem cells and intestinal stem cells, and quantitative changes in nutritional status lead to changes in stem cell function as a result of changing insulin-like peptide levels (LaFever and Drummond-Barbosa, 2005; McLeod et al., 2010). Mammalian stem cells are also positively regulated by insulin signaling as fetal forebrain stem cells adjacent to the lateral ventricle are regulated by IGF2 in cerebral spinal fluid (Lehtinen et al., 2011). Nonetheless, additional work will be required to determine whether mammalian stem cells are regulated by systemic nutritional status.

Aging has similar effects on stem cells in multiple tissues

Aging is associated with reduced regenerative capacity and stem cell function in multiple tissues including the CNS (Figure 2B) (Maslov et al., 2004). Stem cell function decreases with age in many tissues in an evolutionarily conserved manner. Fly spermatogonial stem cell function declines during aging as a consequence of both cell-intrinsic (Cheng et al., 2008) and niche changes (Boyle et al., 2007). In aging mammalian tissues, stem cells exhibit reduced self-renewal potential and accumulation of damage to DNA, mitochondria, and other macromolecules (Rossi et al., 2008; Sharpless and DePinho, 2007).

Figure 2. Stem cell self-renewal mechanisms change with age.

A. A pathway of tumor suppressors (red) and proto-oncogenes (green) that control stem cell self-renewal and that change in expression with age (Nishino et al., 2008). Arrows indicate whether expression increases or decreases with age in stem cells. B. The balance between proto-oncogenic signals and gate-keeping tumor suppressor signals changes with age within stem cells (Levi and Morrison, 2009). Proto-oncogenic signals dominate during fetal development when stem cells divide rapidly, and some gate-keeping tumor suppressor mechanisms that are critical during adulthood are not competent to negatively regulate division in fetal stem cells. Proto-oncogenes and gate-keeping tumor suppressors come into balance in young adult stem cells, which are often quiescent but remain able to enter cycle to regenerate tissues after injury. Gate-keeping tumor suppressors, such as p16^Ink4a, predominate in aging stem cells, which exhibit less regenerative potential. C. The changing balance between proto-oncogenes and tumor suppressors allows stem cells to change their properties throughout life in a way that mirrors changing tissue demands. Tissue growth and regenerative capacity decline over time while cancer incidence increases. The increasing tumor suppressor expression during aging may attenuate the increasing cancer incidence in aging tissues at the cost of reducing tissue regenerative capacity. Increasing expression of the gate-keeping tumor suppressors p16^Ink4a and p19^Arf increases mouse lifespan through mechanisms beyond reducing cancer incidence (Matheu et al., 2009). This suggests it is advantageous to negatively regulate the proliferation of some cells in aging tissues, despite the consequent reduction in stem cell function and tissue regeneration. D. Un-repaired DNA damage (red crosses) accumulates with age to a greater extent in stem cells as compared to restricted progenitors (Rossi et al., 2007). Cell-extrinsic factors also modulate stem cell aging. Muscle satellite cell function declines with age as these cells are exposed to a decreased level of Notch ligands and increased levels of Wnt and TGFβ (Brack et al., 2007; Carlson et al., 2008; Conboy et al., 2003; Conboy et al., 2005; Liu et al., 2007).

The declines in stem cell function during aging are also associated with increasing tumor suppressor expression (Figure 2C). The p16^Ink4a cyclin-dependent kinase inhibitor, a negative regulator of cell cycle progression that sometimes causes cellular senescence, is generally not detectable in young adult tissues but expression increases during aging (Krishnamurthy et al., 2004). This increase in p16^Ink4a expression contributes to the age-related decline in stem cell function in the hematopoietic and nervous systems as well as the decline in β cell proliferation in the pancreas. Deficiency for p16^Ink4a partially rescues the age-related declines in stem cell frequency, mitotic activity, and neurogenesis in the forebrain (Molofsky et al., 2006) as well as increasing HSC frequency (Janzen et al., 2006) and increasing the regenerative capacity of pancreatic β cells (Krishnamurthy et al., 2006). These results suggest that aging stem cells actively shut themselves down in multiple tissues by inducing the expression of tumor suppressors (Figure 2). This increase in tumor suppressor expression presumably reduces cancer incidence in aging tissues as p16^Ink4a-deficient mice exhibit a much higher cancer incidence during aging (Sharpless et al., 2001).

A pathway controlled by let-7 microRNAs changes with age to allow increased p16^Ink4a expression in stem cells (Figure 2A). In mouse neural stem cells, let-7b expression increases with age, reducing the expression of the Hmga2 transcriptional regulator, and increasing the expression of the JunB, p16^Ink4a, and p19^Arf tumor suppressors (Nishino et al., 2008). Hmga2 promotes stem cell self-renewal in fetal and young adult tissues by negatively regulating the expression of p16^Ink4a and p19^Arf, allowing these genes to be expressed in aging tissues as Hmga2 expression is extinguished by increasing let-7b expression. This pathway is likely to regulate age-related changes in stem cell function in multiple tissues without generically regulating the proliferation of all cells as Hmga2 is not required to promote the proliferation of restricted neuronal or glial progenitors (Nishino et al., 2008).

These results demonstrate that stem cell self-renewal is regulated by networks that balance proto-oncogenes, like Bmi-1 and Hmga2, with gate-keeping tumor suppressors, like p16^Ink4a and p19^Arf (Pardal et al., 2005). The way stem cells balance these competing signals changes with age (Figure 2C) (Levi and Morrison, 2009). Proto-oncogenic signals dominate during fetal development when stem cells divide rapidly to form tissues and when there is little risk of cancer. The proto-oncogenic signals come into balance with gate-keeping tumor suppressors during young adulthood, when most stem cells are quiescent most of the time, and when tumor suppression is required to avoid cancer. During aging, gate-keeping tumor suppressor expression increases. This reduces stem cell function and tissue regenerative capacity, as well as presumably reducing cancer incidence. By undergoing temporal changes in the balance between proto-oncogenic and tumor suppressor signals, stem cells favor growth over tumor suppression during fetal development and tumor suppression over tissue regeneration in aging tissues (Figure 2C).

This might also explain why mutation spectrum differs in childhood versus adult cancers (Downing and Shannon, 2002). If self-renewal pathways change with age, then different mutations may be competent to promote neoplastic proliferation in cells from different age patients.

Failure to repair DNA damage leads to phenotypes that resemble premature aging, including a premature decline in stem cell function (Blanpain et al., 2011; Rossi et al., 2008). Deficiencies in DNA repair proteins significantly reduce stem cell function, particularly under stressful conditions (Ito et al., 2004; Nijnik et al., 2007; Rossi et al., 2007). HSCs and primitive hematopoietic progenitors accumulate DNA lesions during aging, marked by γH2AX foci (Figure 2D) (Rossi et al., 2007). DNA damage may accumulate in HSCs because quiescent HSCs have enhanced survival mechanisms compared to differentiated progenitors, and rely on error-prone non-homologous end joining to repair DNA double strand breaks (Mohrin et al., 2010). The reliance upon non-homologous end joining to repair DNA double strand breaks is also observed in epidermal stem cells (Sotiropoulou et al., 2010) but not in all stem cells (Blanpain et al., 2011). Stem cells thus share mechanisms to suppress the accumulation of DNA damage. While experimental elimination of DNA repair mechanisms leads to a premature depletion of stem cells, an open question is the extent to which DNA damage in stem cells affects the properties of these cells during physiological aging.

Systemic environmental and metabolic changes also contribute to the aging of stem cells (Figure 2D). Aging reduces the regenerative capacity of muscle satellite cells through increases in the levels of Wnt and TGF-β and a decrease in the expression of Notch ligands (Brack et al., 2007; Carlson et al., 2008; Conboy et al., 2003; Conboy et al., 2005; Liu et al., 2007). Declines in mitochondrial function are also observed during aging (Balaban et al., 2005; Chan, 2006) and can be precipitated by premature declines in telomere length as a consequence of telomerase deficiency (Sahin et al., 2011). Given that defects in mitochondrial function can yield phenotypes that resemble premature aging (Balaban et al., 2005; Chan, 2006), these results suggest that defects in energy metabolism are one mediator of the effects of telomere attrition on aging.

Telomere maintenance is also critical for chromosome stability and stem cell maintenance. Stem cells express telomerase to attenuate the decline in telomere length with age or upon tissue regeneration (Morrison et al., 1996; Vaziri et al., 1994). Telomerase deficiency leads to reduced stem cell self-renewal, stem cell depletion, and to defects in the regeneration of proliferative tissues (Allsopp et al., 2003; Ferron et al., 2004; Jaskelioff et al., 2011; Lee et al., 1998). In telomerase deficient mice, these defects are only observed beginning in the third generation after loss of telomeres or upon serial transplantation of HSCs; however, inbred mice have much longer telomeres and shorter lifespans than humans. It therefore remains uncertain whether telomere length is limiting for stem cell function or tissue regeneration in the context of normal human aging. Surprisingly, reactivation of telomerase can elongate telomeres, rescuing epithelial stem cell function (Flores et al., 2005) as well as neural stem cell function, neurogenesis, and olfactory function in telomerase deficient mice (Jaskelioff et al., 2011). The regulation of telomerase activity is thus critical for the maintenance of stem cell function and tissue regenerative capacity.

Energy metabolism in stem cells

Stem cells must dynamically reprogram their cellular metabolism in response to changes in cell cycle status, which can occur during normal development or after injury. The ways in which they change their metabolism are not yet understood but presumably involve activation of nutrient uptake and consumption, and changes in the biosynthetic pathways that support survival and proliferation (DeBerardinis et al., 2008; Vander Heiden et al., 2009). Disruption of the mechanisms that regulate these metabolic pathways can lead to profound defects in stem cells without necessarily having the same effects on restricted progenitors and differentiated cells (Gan et al., 2010; Gurumurthy et al., 2010; Nakada et al., 2010). This suggests that some metabolic pathways are regulated differently in stem cells as compared to their progeny.

Stem cells and their differentiated progeny have distinct metabolic profiles. Cultured ES cells rely on glycolysis for ATP production but upregulate mitochondrial oxidative metabolism as they differentiate (Facucho-Oliveira and St John, 2009). Pluripotent cells in the inner cell mass also rely on glycolysis and upregulate oxidative metabolism during development (Facucho-Oliveira and St John, 2009). Highly proliferative, pluripotent cells therefore rely upon glycolysis for ATP production in vitro and in vivo, perhaps because glycolysis also yields substrates for anabolic biosynthetic pathways that proliferating cells depend upon (Vander Heiden et al., 2009) Adult HSCs have reduced concentrations of ATP and fewer mitochondria than differentiated cells and have been suggested to rely on glycolysis to generate ATP even though these cells are mainly quiescent (Inoue et al., 2010; Kim et al., 1998; Simsek et al., 2010). This raises the possibility that many undifferentiated stem cells preferentially rely upon glycolysis irrespective of whether they are highly proliferative or quiescent.

Stem cells must coordinate energy metabolism with cell division. The PI-3kinase pathway is activated in response to various growth factors and promotes cell growth and proliferation, partly by activating Akt and mTORC1 (Figure 3) (Engelman et al., 2006). The Pten tumor suppressor negatively regulates PI-3kinase pathway signaling and Pten deficiency increases cell growth and proliferation. Deletion of Pten increases the self-renewal of ES cells as well as in vivo neurogenesis and in vitro self-renewal by CNS stem cells (Gregorian et al., 2009; Groszer et al., 2006; Groszer et al., 2001). In contrast, conditional Pten deletion from adult HSCs drives HSCs into cycle but quickly leads to their depletion by activating a tumor suppressor response marked by increased p16^Ink4a and p53 expression (Lee et al., 2010; Yilmaz et al., 2006; Zhang et al., 2006). Deletion of TSC also leads to HSC depletion, partly by increasing mitochondrial mass and oxidative stress (Chen et al., 2008; Gan et al., 2008).

Figure 3. Signaling pathways that regulate energy and oxygen metabolism in stem cells.

Various growth factors activate PI-3kinase through their receptors, leading to activation of Akt (Engelman et al., 2006). Akt activation leads to mTORC1 activation, which promotes protein translation and lipid synthesis. The Lkb1-AMPK pathway becomes activated upon energy stress and inhibits the mTORC1 pathway by activating TSC (Shackelford and Shaw, 2009). Akt also negatively regulates FoxO transcription factor function. FoxO transcription factors promote the expression of enzymes that detoxify ROS (Salih and Brunet, 2008). Other transcriptional/epigenetic regulators, such as Bmi1 and Prdm16, also regulate oxidative stress and mitochondrial function. HIF transcription factors are stabilized in hypoxia by the inhibition of von Hippel Lindau (VHL)-mediated degradation. HIF promotes glucose uptake and glycolysis to adapt to hypoxia. Red ovals indicate tumor suppressors, green ovals indicate proto-oncogenes, and yellow ovals indicate proteins whose role in tumorigenesis is not clear.

The Lkb1-AMPK kinases are key regulators of cellular metabolism that coordinate cellular proliferation with energy metabolism by suppressing proliferation when the ATP to AMP ratio is low. Energy stress prompts AMPK signaling to activate catabolic pathways such as mitochondrial fatty acid oxidation while inhibiting anabolic pathways such as mTORC1-mediated protein synthesis (Figure 3) (Shackelford and Shaw, 2009). Lkb1 is a tumor suppressor that is mutated in Peutz-Jeghers syndrome patients (Hemminki et al., 1998; Jenne et al., 1998). Lkb1 deficiency increases the proliferation of many tissues (Contreras et al., 2008; Gurumurthy et al., 2008; Hezel et al., 2008; Pearson et al., 2008) and immortalizes mouse embryonic fibroblasts (Bardeesy et al., 2002). These data suggest that the primary function of Lkb1 in many adult tissues is to negatively regulate cell division, preventing tissue overgrowth. However, conditional deletion of Lkb1 from hematopoietic cells leads to a cell-autonomous defect in HSCs that rapidly increases proliferation and cell death (Gan et al., 2010; Gurumurthy et al., 2010; Nakada et al., 2010). HSCs depend more acutely on Lkb1 for cell cycle regulation and survival as compared to other hematopoietic cells. Lkb1 also has different effects on signaling pathways and on mitochondrial function within HSCs as compared to restricted progenitors (Nakada et al., 2010). This demonstrates that even key metabolic regulators have different functions in different kinds of dividing somatic cells.

The Lkb1 pathway regulates chromosome stability in HSCs in addition to energy metabolism. Lkb1-deficient HSCs exhibit supernumerary centrosomes and become aneuploid, whereas myeloid restricted progenitors appear to divide normally in the absence of Lkb1 (Nakada et al., 2010). AMPK-deficient HSCs do not become aneuploid, indicating that Lkb1 regulates mitosis in HSCs through AMPK-independent mechanisms. Lkb1 and AMPK homologs in Drosophila also regulate chromosome stability in neuroblasts, suggesting that Lkb1 is an evolutionary conserved regulator of mitosis in some cell types (Bonaccorsi et al., 2007; Lee et al., 2007). Therefore, regulation of mitotic processes including chromosome segregation differs between stem cells and some other progenitors.

Stem cells are sensitive to oxidative stress

Stem cells are particularly sensitive to the toxic effects of oxidative damage and are equipped with protective mechanisms that appear to be less active in some other progenitors. FoxO transcription factors regulate stem cell maintenance by regulating the expression of genes involved in cell cycle, apoptosis, oxidative stress, and energy metabolism (Figure 3) (Salih and Brunet, 2008). Deletion of FoxO1, 3, and 4 in CNS stem cells or in mouse HSCs leads to increased levels of reactive oxygen species (ROS) and to stem cell depletion (Paik et al., 2009; Tothova et al., 2007). Treating FoxO-deficient mice with the antioxidant, N-acetyl-L-cysteine, partially rescues these stem cell defects. FoxO3 appears to be the most important FoxO for stem cell function as deletion of FoxO3 alone also depletes CNS stem cells and HSCs (Miyamoto et al., 2007; Renault et al., 2009; Yalcin et al., 2008). In contrast to HSCs, FoxO-deficient restricted myeloid progenitors do not exhibit increased ROS levels (Tothova et al., 2007). This suggests that stem cells depend more upon FoxO transcription factors than certain downstream progenitors.

Prdm16 is another transcription factor that promotes stem cell maintenance in multiple tissues, at least partly by regulating oxidative stress (Figure 3). Prdm16 is a zinc finger protein that was originally identified as part of a chromosomal translocation in some human acute myeloid leukemias (Morishita, 2007). Consistent with this, over-expression of the Prdm16 proto-oncogene can immortalize myeloid cells (Nishikata et al., 2003); however, the physiological role of Prdm16 is to regulate stem cell function in multiple tissues. Prdm16 is necessary for the development of brown fat cells (Seale et al., 2008) as well as for the maintenance of stem cell activity in the nervous and hematopoietic systems (Chuikov et al., 2010). The depletion of neural stem cells is at least partially due to increased oxidative stress as the depletion can be partially rescued by treatment with N-acetyl-L-cysteine. Prdm16 appears to regulate mitochondrial function and to prevent the accumulation of ROS, though the mechanisms by which this occurs remain unknown.

The polycomb protein Bmi-1 promotes stem cell maintenance by negatively regulating p16^Ink4a and p19^Arf expression (Bruggeman et al., 2005; Jacobs et al., 1999; Molofsky et al., 2005; Oguro et al., 2006) and likely by regulating mitochondrial function and oxidative stress as well (Figure 3) (Liu et al., 2009). Cells from Bmi1-deficient mice have reduced mitochondrial oxygen consumption, reduced mitochondrial oxidative capacity, reduced ATP levels, and elevated ROS levels that appear to cause DNA damage (Liu et al., 2009). Treating Bmi1-deficient mice with N-acetyl-L-cysteine partially rescues the depletion of thymocytes, though it has not yet been tested whether this also rescues stem cell function. The observation that Bmi-1 regulates tumor suppressor expression and mitochondrial function suggests that key self-renewal mechanisms integrate energy metabolism with cell cycle control in a manner analogous to PI-3kinase pathway regulation by Pten, AMPK, and Lkb1.

Although elevated ROS levels are toxic to stem cells, physiological levels of ROS are required for certain stem cell functions. Consistent with the role of Akt in negatively regulating FoxO function (Salih and Brunet, 2008), deletion of Akt1 and Akt2 decreases ROS levels and attenuates the proliferation and differentiation of HSCs (Juntilla et al., 2010). Neural stem/progenitor cell proliferation and differentiation are also regulated by ROS (Le Belle et al., 2011; Prozorovski et al., 2008; Smith et al., 2000).

Stem cells, oxygen, and hypoxia

Changes in stem cell function are involved in the adaptation to declining oxygen availability such as occur with increasing altitude or cardiopulmonary disease. Neuron-like glomus cells in the carotid body mediate these responses by sensing oxygen levels in the blood and inducing hyperventilation during hypoxemia. Exposure of mice to hypoxia induces the proliferation of glia-like stem cells that remodel the carotid body in response to hypoxia to increase the number of glomus cells (Pardal et al., 2007). Hypoxia also increases erythropoiesis by inducing erythropoietin expression in the kidney and liver (Semenza, 2009). Hypoxia increases the total number and proliferation of HSCs and multipotent progenitors (Li et al., 2011). It is possible that this involves indirect effects of hypoxia on cell death or cell turn over. Alternatively, since most HSCs localize close to blood vessels (Kiel et al., 2005; Mendez-Ferrer et al., 2010) it is possible that their niche senses changes in oxygen levels. Since other stem cells, including some neural stem cells (Mirzadeh et al., 2008; Shen et al., 2008), also reside in perivascular microenvironments, it is conceivable that stem cells in multiple tissues are directly influenced by oxygen levels (Figure 4). Regardless of the mechanisms, multiple tissues are remodeled in response to hypoxia, partly due to changes in stem/progenitor cell function.

Figure 4. Integrating stem cell function with systemic physiology.

Physiological changes experienced by animals cause changes in stem cell function that lead to adaptive changes in tissue growth and remodeling. Stem cell function is modulated by long-range extrinsic signals that reflect changes in nutrition (A), circadian regulation (B), hormones (C, D), and oxygen tension (E). These signals sometimes act directly on stem cells and sometimes act on niche cells to indirectly modulate stem cell function (see text for details and references). The CNS is integral to the coordination of many physiological changes through neural activity (B) and the secretion of hormones (D). A. Stem cell function is modulated by insulin and other signals that reflect nutritional status. B. Stem cells are regulated by circadian rhythms generated centrally in the suprachiasmatic nucleus (SCN) of the brain and conveyed through the sympathetic nervous system (SNS). C, D. Sexual maturation, mating, and pregnancy can influence stem cell function by changing the expression of hormones from the gonads (C) and the pituitary gland (D). E. Physiological changes also modulate stem cell function by altering oxygen levels and systemic levels of cytokines and growth factors in the blood. Dashed arrows indicate factors secreted from the indicated organ that act on stem cells and their niches in multiple target tissues (green and orange).

It has been hypothesized that most stem cells reside in hypoxic niches that enable them to suppress oxidative damage by relying upon glycolysis rather than mitochondrial oxidative phosphorylation (Mohyeldin et al., 2010; Parmar et al., 2007; Simsek et al., 2010); however, this has not yet been tested in most tissues or in most developmental contexts. Hypoxic microenvironments may not protect stem cells from oxidative stress because hypoxia, paradoxically, can lead to the generation of elevated ROS levels (Brunelle et al., 2005; Guzy and Schumacker, 2006). Nonetheless, evidence suggests that many bone marrow HSCs and at least some neural stem cells in adult mice reside in hypoxic environments. This may appear superficially inconsistent with the idea that HSCs often reside perivascularly; however, HSCs reside adjacent to sinusoidal blood vessels in hematopoietic tissues (Kiel et al., 2005). Sinusoids are a specialized form of vasculature found only in hematopoietic tissues. Sinusoids carry slow veinous circulation that is not designed to transport oxygen around the body as much as to provide specialized vasculature through which hematopoietic cells can intravasate into circulation. Thus, the peri-sinusoidal environment in the bone marrow may be relatively hypoxic.

Stem cell maintenance also depends upon mechanisms that regulate adaptation to lower oxygen tensions. Hypoxia-inducible factor 1α (HIF1α) is a transcription factor that is stabilized in response to hypoxic stress, activating the expression of genes that promote non-oxidative carbon metabolism and ATP synthesis, such as those involved in glucose import and glycolysis (Figure 3) (Majmundar et al., 2010). Deletion of HIF1α from neural stem cells depletes neurogenic progenitors in the subgranular zone of the dentate gyrus (Mazumdar et al., 2010). HIF1α deletion also leads to a progressive decline in HSC function during bone marrow transplantation or aging (Takubo et al., 2010). Deletion of von Hippel Lindau, which encodes a ubiquitin ligase involved in the degradation of HIF1α, also leads to HSC defects even though this increases HIF1α levels (Takubo et al., 2010). This suggests that HIF1α levels must be tightly regulated. Stem cells likely depend on a variety of mechanisms to maintain homeostasis in the face of hypoxia or changes in oxygen tension.

Changes in nutrition affect stem cell function in many tissues

Caloric restriction increases longevity and reduces age-related disease in an evolutionarily conserved manner (Bishop and Guarente, 2007), partly by influencing the function of stem and progenitor cells. Caloric restriction in rodents enhances neurogenesis in the dentate gyrus by promoting the survival of newborn neurons and astrocytes (Bondolfi et al., 2004; Lee et al., 2002) and potentially by increasing progenitor proliferation (Kumar et al., 2009). In the hematopoietic system, short-lived mouse strains exhibit a decline in HSC frequency and function during aging while long-lived mouse strains do not (de Haan et al., 1997). Caloric restriction attenuates the age-related decline in HSC frequency in at least one short-lived mouse strain (Ertl et al., 2008). Feeding adult Drosophila a low nutrient diet alleviates the age-related reduction in the number and proliferation of male germline stem cells (Mair et al., 2010). Caloric restriction can therefore attenuate the reduction in stem cell function during aging in multiple tissues and species.

Nutritional changes can alter the expression of systemic factors that regulate stem cells (Figure 4). Protein starvation in Drosophila leads to a reversible loss of male germline stem cells and intestinal stem cells due to reduced expression of insulin-like peptides, possibly by insulin-producing cells in the brain (McLeod et al., 2010). Expression of constitutively active insulin receptor is able to suppress the starvation-induced loss of germline stem cells, suggesting that insulin directly regulates germline stem cell maintenance. This allows stem cell function in multiple tissues to be modulated by changes in nutritional status.

Changes in the nutritional status of the organism can also indirectly affect stem cell function by modulating the environment (Figure 4). Reduced insulin signaling after protein starvation reduces the proliferation of Drosophila female germline stem cells by acting directly on these cells and by reducing the capacity of the niche to maintain these cells (Hsu and Drummond-Barbosa, 2009; LaFever and Drummond-Barbosa, 2005).

Whereas decreased nutrition can reduce stem cell function, increased nutrition can increase stem cell function. Upon feeding, fat cells in Drosophila activate TOR signaling and secrete a fat-body-derived signal that regulates insulin-like peptide secretion by a subpopulation of nutritionally regulated glial cells. This insulin-like peptide activates neuroblast proliferation through PI-3kinase/TOR signaling (Chell and Brand, 2010; Sousa-Nunes et al., 2011). Additional work will be required to assess whether mammalian stem cells are also acutely regulated by changes nutritional status.

The stem cell response to injury and disease

Regeneration in many adult tissues involves the activation of stem cells to enter cycle and to increase the generation of differentiated cells. Loss of hematopoietic cells by cytotoxicity (Harrison and Lerner, 1991) or bleeding (Cheshier et al., 2007) leads to HSC expansion, mobilization from the bone marrow, and extramedullary hematopoiesis in the liver and spleen. Stroke and excitotoxic injuries induce cell death in the brain, but stem cells appear more resistant to these stresses and initiate a wound healing response that increases neural progenitor proliferation and neurogenesis (Parent, 2003; Romanko et al., 2004). Neural stem cells in the forebrain subventricular zone migrate to the site of injury and generate new neurons (Arvidsson et al., 2002; Parent et al., 2002; Yamashita et al., 2006). The physiological significance of this CNS injury response is uncertain as most of these new neurons are short-lived, fail to incorporate into neural circuits, and appear to contribute little to functional recovery (Zhao et al., 2008). Nonetheless, these responses illustrate the existence of mechanisms across tissues that activate stem cells in response to injury.

Inflammation modulates stem cell function in response to infection or injury. Bacterial and viral infections induce interferons, driving adult HSCs into cycle and expanding HSC numbers (Baldridge et al., 2010; Essers et al., 2009; Sato et al., 2009). This response must be highly regulated since chronic activation in many contexts leads to HSC depletion (Baldridge et al., 2010; Essers et al., 2009; Sato et al., 2009). Inflammation also inhibits neurogenesis and neural stem cell function in vivo (Ekdahl et al., 2003; Li et al., 2010; Monje et al., 2003). Pharmacological anti-inflammatory agents restore dentate gyrus neurogenesis after inflammation induced by irradiation (Monje et al., 2003). Microglial cells mediate the effect of inflammation on neurogenesis (Butovsky et al., 2006). Inflammatory signals can likely have both local and systemic effects on stem cell function, and much more study will be required to fully understand the influence of inflammation on stem cell function.

Circadian regulation of stem cell function

Circadian rhythms regulate many aspects of metabolism and physiology, including stem cell function (Figure 4). Circadian rhythms are evolutionarily conserved cyclical variations in gene expression and function that are approximately a day in length and are observed in cells from bacteria to mammals (Takahashi et al., 2008). The vertebrate CNS controls circadian rhythms throughout the body with oscillations of a master clock located in the suprachiasmatic nucleus of the hypothalamus (Figure 4). This master clock is entrained by light received by the retina, generating a transcriptional autoregulatory loop composed of the transcriptional activators Clock and Bmal1, and their target genes and feedback inhibitors Period1-3 (Per) and Cryptochrome1-2 (Cry) (Bass and Takahashi, 2010). Circadian rhythms regulate the expression of genes involved in protein turnover, mitochondrial respiration, lipid and glucose metabolism (Panda et al., 2002; Rutter et al., 2002), and are proposed to allow temporal orchestration of metabolic processes to maximize the utilization of nutrients (Tu and McKnight, 2006).

The circadian regulation of stem cells has been most extensively studied in the hematopoietic system (Figure 4). Circadian oscillations affect DNA synthesis and the frequency of colony-forming hematopoietic progenitors in mice and humans (Mendez-Ferrer et al., 2009), the ability of sublethally irradiated mice to engraft with transplanted bone marrow cells (D’Hondt et al., 2004), and the susceptibility of bone marrow to chemotherapy (Levi et al., 1988). All of these phenomena may reflect the influence of circadian regulation on the timing of cell division by hematopoietic cells, as this has been observed in a number of tissues (Mendez-Ferrer et al., 2009; Takahashi et al., 2008). Circadian rhythms also regulate neurogenesis in the hippocampus of multiple species, with increased proliferation at a specific circadian phase depending on the species (Goergen et al., 2002; Guzman-Marin et al., 2007).

HSCs and other progenitors are regularly mobilized from the bone marrow, into circulation, then back into hematopoietic tissues (Wright et al., 2001) and this process is subject to circadian regulation. In mice, the sympathetic nervous system regulates the oscillating expression of the chemokine Cxcl12, and its receptor Cxcr4, in the bone marrow such that Cxcl12 signaling is low during the inactive (light) phase of the cycle, allowing mobilization of hematopoietic progenitors into the blood (Katayama et al., 2006; Lucas et al., 2008; Mendez-Ferrer et al., 2008). This effect is also observed in humans, although the human diurnal cycle is inverted related to the mouse nocturnal cycle (Lucas et al., 2008). The physiological significance of this mobilization is not clear.

Exercise, mating, and pregnancy

Exercise, sex hormones, mating, and pregnancy all have effects on stem cell function (Figure 4). Exercise increases the number of neural stem cells and enhances cognitive parameters in mice and humans including learning and memory (Hillman et al., 2008). Exercise affects many aspects of energy metabolism, leading to the consumption of stored nutrients as well as hypoxia when exercise becomes anaerobic. In adult mice of all ages, voluntary running stimulates cell proliferation and neurogenesis in the hippocampus (van Praag et al., 1999; van Praag et al., 2005). Neurogenesis and neural stem cell frequency in the forebrain subventricular zone are increased by exercise as well (Blackmore et al., 2009).

The neurogenic response to exercise is likely to be mediated by multiple systemic factors. Growth hormone and insulin-like growth factor 1 (IGF1) expression are activated in rodents upon exercise (Carro et al., 2000; Eliakim et al., 1997). Growth hormone receptor-deficient mice did not show an increase of neurogenesis after voluntary exercise (Blackmore et al., 2009). Much of the activity of growth hormone is exerted through the production of IGF1, which is largely produced in the liver (Jones and Clemmons, 1995). Exercise stimulates the uptake of blood-borne IGF1 by specific groups of neurons involved in adaptive responses to exercise, and subcutaneous administration of IGF1 is sufficient to induce neurogenesis in the dentate gyrus (Carro et al., 2000). Antibody inhibition of IGF1 blocks the neurogenic and proliferative effects of exercise in the dentate gyrus (Trejo et al., 2001). These data suggest a systemic response to exercise that influences the behavior of neural stem cells and potentially stem cells in other tissues.

Courtship and pregnancy stimulate sex-specific changes in hormones that influence neurogenesis by mechanisms that differ from those induced by exercise. The estrus cycle and pregnancy in female mice are characterized by distinct patterns of gonadal hormones (Figure 4). During pregnancy, neurogenesis increases in concert with serum prolactin level and this has been proposed to be important for the recognition and rearing of offspring (Shingo et al., 2003). Prolactin is sufficient to induce neurogenesis and may act directly on neural stem cells (Shingo et al., 2003). In addition to pregnancy, exposure of female mice to male pheromones also induces neurogenesis (Mak et al., 2007). Pheromones from dominant males stimulate neurogenesis in the forebrain subventricular zone and in the dentate gyrus by inducing prolactin and luteinizing hormone, respectively. Neurogenesis stimulated by male pheromones affects mating preference and consequently the success of offspring (Mak et al., 2007). Neurogenesis during pregnancy and courtship may therefore induce adaptive behavioral changes.

There are additional physiological demands on pregnant and mating animals beyond neural adaptation. Mammalian mammary tissue is acutely sensitive to hormonal regulation that leads to profound changes in morphology and function. Estrogen and progesterone promote the expansion of mammary stem cells and mammary gland morphogenesis (Figure 4) (Asselin-Labat et al., 2010; Joshi et al., 2010). These steroid hormones induce expression of Wnt and RANK ligands by luminal cells, which then act on mammary stem cells to induce proliferation, resulting in more than 10-fold expansion of mammary stem cell frequency during pregnancy and the estrus cycle (Asselin-Labat et al., 2010; Joshi et al., 2010). The regular expansion and contraction of the mammary stem cell pool during each menstrual cycle provides a potential explanation for why breast cancer risk increases with the number of menstrual cycles in humans (Clemons and Goss, 2001). The hematopoietic system also undergoes considerable alterations during pregnancy, increasing erythropoiesis and extramedullary hematopoiesis (Fowler and Nash, 1968). Stem cells in multiple tissues are therefore likely to respond to global physiological cues that remodel tissues in response to pregnancy and sex hormones.

Perspective

Stem cells are regulated by diverse physiological cues that integrate stem cell function and tissue remodeling with physiological demands. Stem cell function is modulated by circadian rhythms, changes in metabolism, diet, exercise, mating, aging, infection, and disease. It is likely that these physiological changes have systemic effects on stem cells in multiple tissues.

Diverse transcriptional, metabolic, cell cycle, and signaling mechanisms regulate stem cell function without generically regulating the function of all dividing cells. Many factors critical for stem cell maintenance regulate energy metabolism and oxidative stress. The concerted regulation of energy metabolism and stem cell function may allow stem cell function to be closely matched to nutritional status. Understanding the key differences between stem cells and other progenitors should provide important insights into how tissue homeostasis is maintained throughout life and how regeneration might be enhanced by therapies that modulate stem cell metabolism.

Understanding these mechanisms could also improve the treatment of cancer. Proto-oncogenes and tumor suppressors likely evolved to regulate stem cell function and tissue homeostasis but cancer cells hijack these mechanisms to enable neoplastic proliferation. Proto-oncogenic pathways such as the PI-3kinase pathway are frequently over-activated in cancer, activating autonomous nutrient uptake, factor-independent growth and survival, increasing glycolysis and anabolic pathways (DeBerardinis et al., 2008). Collectively, this promotes aerobic glycolysis, also called the Warburg effect, in which cancer cells consume glucose by glycolysis without further activating oxidative metabolism (Warburg, 1956). An improved understanding of the mechanisms that regulate stem cell physiology would not only improve our understanding tissue homeostasis, but would likely yield new therapeutic strategies for cancer.