Maintaining tissue homeostasis: dynamic control of somatic stem cell activity

Homeostatic proliferation

Signaling events regulating homeostatic proliferation in epithelial tissues of mammals have been extensively studied, yet the cell-specific requirements for individual signaling events remain unclear. In the intestine, for example, Wnt signaling, Notch signaling and BMP signaling all promote proliferation, but also influence differentiation in the crypt (Crosnier et al., 2006; van der Flier and Clevers, 2009). Whether these effects are primarily a consequence of influencing stem cell proliferation or transit amplifying cell proliferation, however, remains unclear.

In flies, the availability of lineage-tracing techniques for lineages derived from homozygous mutant stem cells has allowed characterizing the signaling requirements for homeostatic proliferation of ISCs in detail, and has led to the emerging concept that signaling mechanisms ensuring homeostatic proliferation and signaling events inducing proliferation in response to injury are distinct.

Drosophila ISCs are mostly slow- or non-proliferating in young, unchallenged intestines, but become highly proliferative after an environmental challenge or tissue injury. The existence of a ‘quiescent’ state for ISCs has been under some debate: lineage tracing in ISCs had initially led to the impression that ISCs are continuously dividing cells, since once induced, clones appear to grow linearly, then enter a steady state where production of new cells seems to be in balance with the turnover rate of ECs (Ohlstein and Spradling, 2006). This interpretation is complicated, however, by mitotic and S-phase labeling experiments, which indicate long periods of very limited ISC proliferation in young, healthy guts. BrdU incorporation, for example, is observed in only 5-10 % of all ISCs in a 48 hour window (Hochmuth et al., 2011). A low basal level of tissue renewal has been revealed, on the other hand, by separate lineage tracing studies. In these, a ‘Flp-out’ strategy was used to induce heritable GFP expression in all ISCs and their progeny in adulthood, demonstrating that the whole tissue is turned over in about 2 weeks in females and in over a month in males (Jiang et al., 2009). This corresponds to total tissue turnover of about 4 times in females and twice in males over the whole lifespan of the animal.

This basal homeostatic proliferation and self-renewal of ISCs requires the activity of several growth factor signaling pathways (Figure 2). Using the MARCM method (Lee and Luo, 1999)(Textbox 1) to induce homozygosity for mutations in the EGF Receptor and the Insulin Receptor, it was shown that the growth factor response pathways activated by these receptors are essential for ISC proliferation under unstressed conditions (resulting in mutant clones consisting of mostly single stem cells) (Biteau and Jasper, 2011; Biteau et al., 2010; Jiang et al., 2011; Xu et al., 2011). Downstream mediators of these pathways include the Insulin Receptor Substrate, PI3Kinase, Akt, Ras and ERK and all these molecules have also been shown to be essential for ISC proliferation. Consistent with a general permissive role for these signaling pathways, activated ERK (dpERK) can be detected in all ISCs under normal conditions (Biteau and Jasper, 2011; Jiang et al., 2011; Xu et al., 2011). Interestingly, constitutive activation of EGFR/InR signaling components can also increase ISC proliferation rates, indicating that the level of RTK signaling activity modulates the proliferative state of ISCs (Biteau and Jasper, 2011; Jiang et al., 2011; Xu et al., 2011).

In addition, the MAPK p38 is required for ISC proliferation under unstressed conditions (Park et al., 2009), and it was suggested that it acts as a mediator of the effect of the PDGF/VEGF-like receptor signaling pathway (composed of Pvf ligands and the PvR receptor) on ISC proliferation (Choi et al., 2008; Park et al., 2009). It has not yet been established, however, whether Pvf and PVR are also required for homeostatic proliferation.

Interestingly, ligands for the EGFR and the InR pathways are dynamically controlled in response to environmental challenges. Nutritional state can significantly affect insulin-like peptide (Dilp) expression, while oxidative stress or DNA damage results in repression of dilp expression (Geminard et al., 2009; Karpac et al., 2011; Slaidina et al., 2009; Wang et al., 2005). This regulation may thus allow adjusting ISC proliferation to systemic nutrient and stress levels (Amcheslavsky et al., 2009; McLeod et al., 2010). Nutrition also influences the activity of the TSC/Tor signaling pathway, and excessive Tor activation seems to have deleterious consequences for ISC activity (Amcheslavsky et al., 2011). Infection with pathogenic bacteria, on the other hand, induces EGF-like ligand expression in the gut (Buchon et al., 2010; Buchon et al., 2009b; Jiang et al., 2011; Xu et al., 2011). These ligands are secreted by both epithelial cells (including ISCs and ECs), as well as by the surrounding visceral muscle, displaying a remarkable redundancy and indicating that ISC maintenance and proliferation may be coordinated across the entire organ by ligands derived from multiple regions. Insulin and EGF-like growth factors thus serve as permissive signals for ISC proliferation, while also contributing to the proliferative response to stress.

Activation of stem cell proliferation by injury and stress

Recent studies in mammals support the idea that the mechanisms regulating SC proliferation in response to tissue injury are, at least partially, distinct from those essential for homeostatic regeneration. For example, the Hippo and Focal Adhesion Kinase pathways were recently shown to be required in the intestinal crypt for acute tissue regeneration in response to DSS or gamma irradiation, respectively, while being dispensible for homeostatic proliferation (Ashton et al., 2010; Cai et al., 2010). A comprehensive and detailed analysis of stress-induced SC proliferation is thus warranted and promises to provide important insight into the maintenance of tissue homeostasis under challenging conditions.

In Drosophila, upon stress or injury, ISCs respond by dramatically increasing their proliferative activity to replenish the epithelium with functional differentiated cells. This activation occurs in response to infection (Apidianakis et al., 2009; Buchon et al., 2009a; Chatterjee and Ip, 2009; Cronin et al., 2009; Jiang et al., 2009), oxidative stress (Biteau et al., 2008; Buchon et al., 2009a; Choi et al., 2008), DNA damage (Amcheslavsky et al., 2009) and heat stress, as well as other factors that damage the epithelium (such as DSS) and that cause EC apoptosis (Amcheslavsky et al., 2009; Jiang et al., 2009). Stress-induced activation of ISCs is essential for flies to survive these stresses (Buchon et al., 2009a; Cronin et al., 2009; Jiang et al., 2009).

A comprehensive picture of the signals mediating this proliferative response is emerging (Figure 2): ISCs respond to intrinsic challenges, such as oxidative stress through the Jun-N-terminal Kinase (JNK) pathway (Biteau et al., 2008; Choi et al., 2008), to local cues through the JAK/Stat signaling pathway (Buchon et al., 2009a; Cronin et al., 2009; Jiang et al., 2009), and to changes in tissue integrity through the hippo/yorkie pathway (Karpowicz et al., 2010; Ren et al., 2010; Shaw et al., 2010; Staley and Irvine, 2010). These inductive pathways are characterized by the fact that their activation is not required for ISC proliferation under homeostatic conditions, but is required and sufficient to increase ISC proliferative activity in response to stress. They are integrated in multiple ways: JAK/Stat signaling is activated in ISCs by locally derived interleukin 6 - like cytokines named Unpaired 1-3 (Upd1-3). These cytokines are secreted from damaged and dying ECs in response to infection (Buchon et al., 2009a; Cronin et al., 2009; Jiang et al., 2009). Artificial induction of JNK in ECs is sufficient to induce Upd expression (Jiang et al., 2009), potentially through the activation of the Yorkie transcription factor,(Staley and Irvine, 2010). However, JNK is not required in ECs for the proliferative response upon infection (Buchon et al., 2009a; Jiang et al., 2009). Yorkie also induces Upd expression in ISCs themselves, activating proliferation in an autocrine manner (Karpowicz et al., 2010). In addition, JAK/Stat signaling acts indirectly to promote ISC proliferation by increasing the expression of the EGF-like ligand Vein in the visceral muscle (Buchon et al., 2010). Through secreted Upd cytokines, injured or dying ECs thus initiate and promote regenerative activity in the epithelium by directly activating cell cycle progression in ISCs, while at the same time triggering increased growth factor secretion from the muscle. In addition to its potential role in dying ECs, the JNK pathway also autonomously activates ISC proliferation by phosphorylating the AP-1 transcription factor Fos (Biteau and Jasper, 2011). Fos is phosphorylated by JNK and the EGFR responsive MAPK kinase ERK on distinct sites, thus integrating both permissive and inductive signals (Ciapponi et al., 2001). Accordingly, Fos is required for both homeostatic ISC proliferation and for stress-induced activation of ISCs.

Redox state and stem cell function

An important question that remains unanswered in most stem cells lineages, is how exactly ISC proliferation is increased by stress signals. Two non-exclusive mechanisms can be envisioned (Figure 3A): Quiescent stem cells that do not divide under homeostatic conditions may be activated, leading to an overall increase in proliferating cells within the stem cell population. Alternatively, the cell cycle of actively cycling stem cells might be shortened, increasing the proliferation rate within an identical stem cell pool. Further studies of the proliferative response in the fly intestinal epithelum are required to differentiate between these possibilities, and are expected to generate testable hypotheses for the regulation of mammalian stem cell systems.

In this context, one recent study supports the first possibility, namely the existence of quiescent cells within the ISC pool, and the stress-induced transition of these cells towards an actively dividing state (Hochmuth et al., 2011). Strikingly, this transition appears to be regulated primarily by changes in the intracellular redox state of ISCs, highlighting the emerging critical role of redox regulation in stem cell biology. A low intracellular concentration of reactive oxygen species (ROS) is increasingly recognized as a critical condition for stemness, self-renewal and pluripotency in both mammals and in Drosophila.

In mice, elevated ROS levels result in reduced regenerative potential and self-renewal in a wide range of stem cell populations, including neuronal and glial progenitors and HSCs (Diehn et al., 2009; Ito et al., 2004; Le Belle et al., 2011; Liu et al., 2009; Miyamoto et al., 2007; Miyamoto et al., 2008; Smith et al., 2000; Tothova and Gilliland, 2007; Tothova et al., 2007; Tsatmali et al., 2005). The reason and consequence of the correlation between redox state and stem cell behavior remains enigmatic, but recent findings in flies have begun unraveling signaling mechanisms mediating this effect. Increased ROS concentration primes hematopoietic progenitors of the larval lymphgland for differentiation (Owusu-Ansah and Banerjee, 2009) and promotes ISC proliferation in the adult gut (Biteau et al., 2008; Buchon et al., 2009a; Choi et al., 2008; Hochmuth et al., 2011). In the intestine, high levels of ROS are produced by ECs to control commensal and pathogenic bacterial populations, and increased ROS production is a likely cause of the increased ISC proliferation observed in aging animals (Biteau et al., 2008; Buchon et al., 2009a; Choi et al., 2008; Ha et al., 2009; Ha et al., 2005; Hochmuth et al., 2011). A central regulator of the intracellular redox state in vertebrates and invertebrates is Nrf2, a member of the ‘cap-and-collar’ (Cnc) family of transcription factors. By influencing the stem cell redox state, the Drosophila homologue of Nrf2, CncC, and its negative regulator Keap1 control ISC proliferation rates, and this regulation is required to limit ISC hyperproliferation and intestinal degeneration in aging flies (Hochmuth et al., 2011). Interestingly, Keap1 mutant mice show significant hyperkeratosis of the esophageal epithelium, suggesting that Keap1/Nrf2 also influences cell differentiation and proliferation in mammalian barrier epithelia (Wakabayashi et al., 2003). The Keap1/CncC regulatory module thus emerges as central in the control of ISC proliferation and intestinal regeneration.

How the changes in redox state induced by Nrf2 activity might be translated into specific stem cell responses remains unclear. The control of hematopoietic progenitor differentiation by high ROS levels suggest an interesting role for the JNK pathway, however, as it acts to induce Foxo activity and promote Polycomb complex downregulation in this system (Owusu-Ansah and Banerjee, 2009). In the intestine, JNK activation induces proliferation of ISCs, but Foxo activation suppresses proliferation (Biteau et al., 2008; Biteau et al., 2010). While a similar JNK response thus might play a role downstream of Nrf2 in ISCs, the mechanistic details of this response are likely to be substantially distinct from the hematopoietic system.

Proliferative plasticity is required for tissue maintenance and longevity

It is of interest that all the signaling pathways described above have been implicated in tumor formation in vertebrates (Radtke and Clevers, 2005). For example, a role of AP-1 transcription factors and JNK signaling in the regulation of intestinal stem cell proliferation and intestinal cancer has been reported in mice: JNK can induce cell proliferation in the intestinal crypt and increases tumor incidence and growth in an inflammation-induced colon cancer model (Sancho et al., 2009). It remains unclear, however, if this proliferative response is initiated in Lgr5+ or Bmi1+ stem cells, or rather in the transit amplifying cell population. The Fos binding partner c-Jun may mediate these effects, as indicated by the fact that knock-in of non-phosphorylatable c-Jun mutants into the c-Jun locus reduces tumorigenesis in the intestine caused by loss of APC (Nateri et al., 2005). Further illustrating the conserved function of Fos in the regulation of stem cell function, increased Fos activity has also been found to be sufficient to promote hematopoietic stem cell self-renewal in mice (Deneault et al., 2009). In the fly, ISC-specific activation of Wnt signaling, by mutating APC or expressing an active form of b-catenin or wingless itself, is also sufficient to induce the formation of tumor-like stem cell clusters (Cordero et al., 2009; Lee et al., 2009; Lin et al., 2008). Whether Wnt signaling also interacts with JNK and Jun/Fos in Drosophila ISCs remains to be tested, yet regulation of stem cell proliferation by JNK signaling and AP-1 appears to be a broadly conserved phenomenon.

The precise role of stem cell specific stress signaling in the development of tumors remains unclear, however, and the detailed analysis of stem cell activity in the Drosophila intestine is providing important new insight into the requirements for dynamic stem cell control to prevent loss of epithelial homeostasis. The emerging model highlights the critical role for negative feedback regulation of inductive signaling to ensure an appropriate but limited burst of ISC proliferation following injury. Thus, when stimuli are removed, ISC proliferation is rapidly decreased after the initial induction of proliferation (Buchon et al., 2010; Jiang et al., 2009). Chronic and excessive activation of JNK signaling in the ISC lineage, in turn, causes intestinal dysplasia (Biteau et al., 2008), while over-expression of tumorigenic RasV12 predisposes ISCs to tumor formation when exposed to a stimulus (Apidianakis et al., 2009). It is likely that the down-regulation of proliferative activity is the consequence of the induction of multiple negative regulators of inductive signaling pathways, such as Socs36E (inhibitor of JAK/Stat signaling), Puckered (inhibitory phosphatase of JNK) or Argos (negative regulator of EGFR) (Biteau et al., 2008; Buchon et al., 2009a; Buchon et al., 2009b; Jiang et al., 2009). However, this hypothesis remains to be validated, and additional mechanisms have to be tested. For example, reactivation of the Nrf2 pathway may contribute to the return of ISCs to a non-proliferative state. Strikingly, a recent study in mouse muscle satellite cells demonstrate the essential role of an inhibitor of EGFR signaling in the reversible quiescence of this stem cell popuation, demonstrating the potential conservation of the mechanisms proposed in the Drosophila intestine (Shea et al., 2010).

The importance of the transient aspect of the proliferative response is particularly highlighted by the deleterious consequences of constant ISC proliferation in aging flies. In older animals, an increase in bacterial load, and subsequent oxidative stress, in the intestine results in chronic activation of the regenerative response, including JNK, Upd/JAK/Stat and Pvf/p38 activities (Biteau et al., 2008; Buchon et al., 2009a; Choi et al., 2008; Park et al., 2009). This leads to sustained ISC proliferation and loss of tissue homeostasis, as non-functional supernumerary cells accumulate in the epithelium. This deregulation impairs metabolic homeostasis and limits Drosophila longevity (Biteau et al., 2010).

Maintenance of stem cell populations

The exact mechanisms ensuring maintenance of an appropriate stem cell population are currently under intense investigation. An elegant recent study has suggested that the stem cell population in the mouse crypt is maintained by population asymmetry (termed ‘stochastic model’), rather than fixed intrinsic asymmetry of stem cell divisions (‘hierarchical model’) (Figure 3C)(Snippert et al., 2010). Snippert et al used multi-color lineage tracing to distinguish between the stochastic and hierarchical model in the maintenance of crypt lineages by Lgr5⁺ stem cells. Since mosaically labeled crypts become mostly ‘monochrome’ over a period of 2 months, the authors infer that this tissue is maintained by population asymmetry rather than by fixed asymmetry of the stem cell division (in which case each individually induced lineage is expected to be stable over time, preserving mosaicism). It remained unclear, however, if the selective survival of individual lineages might reflect an unknown stem cell hierarchy in the crypt, in which a single stem cell is responsible for the long-term maintenance and replacement of all other ‘secondary’ stem cells in this tissue. This second possible interpretation of Snippert et al’s results has become more likely due to a new study, in which a hierarchy between Lgr5+ and Bmi1+ stem cells in the crypt has been identified (Tian et al., 2011). Strikingly, complete ablation of the Lgr5+ cell population in the intestine does not affect the integrity and homeostasis of the intestinal epithelium, while Bmi1+ stem cells can give rise to all structures of the crypt, including Lgr5+ stem cells (Quyn et al., 2010; Tian et al., 2011). One interpretation of Snippert et al’s results is therefore that crypts drift toward clonality over time because the Lgr5+ stem cell population is being replenished from individual Bmi1+ stem cells. Furthermore, the spindle orientation of dividing stem cells in the crypt indicates a predominance of asymmetric divisions, supporting a hierarchical model, as in a stochastic model divisions are expected to be to a large extent symmetric (Quyn et al., 2010).

The simpler organization of the Drosophila ISC lineage has allowed characterizing maintenance strategies in this system more definitively, but important questions remain. ISCs and their function are maintained throughout the life of the organism, as illustrated by the fact that clone numbers observed in individual guts are in effect stable until late in life (i.e. for 4-6 weeks; (Lin et al., 2008; Ohlstein and Spradling, 2006). The mechanisms ensuring ISC self-renewal are starting to be characterized, but it remains unclear if ISCs that have been specified during development simply survive throughout life of the animal, or if there are active replenishment processes in which ISCs are regenerated through symmetric divisions of older cells (Figure 3B). ISC nests display an intrinsic asymmetry, as the ISC is basally localized and closely contacts the basement membrane and the visceral muscle, while the EB resides more apically. Furthermore, these cells can be clearly distinguished by the asymmetric distribution of the Notch ligand Delta (Dl), which is detected exclusively in the ISC, and of Notch signaling reporters, which are invariably active in EBs but not ISCs. The outcome of ISC divisions thus seems to be consistently asymmetric in young, healthy guts. The asymmetry of the ISC division itself has been debated (Hou, 2010; Simons and Clevers, 2011), however, as no asymmetric localization of intracellular signaling components prior to division has been described in ISCs to date (a hallmark of asymmetric divisions observed for example in Drosophila neuroblasts (Knoblich, 2010)). Supporting the model of an initially symmetric cell division, vesicles containing the Dl protein are equally segregated during ISC cell division (but Dl expression is rapidly inhibited in the EB), and rare nests containing two delta-positive cells can be observed (Ohlstein and Spradling, 2007).

On the other hand, similar to what is observed in the mouse and human intestinal stem cells (Quyn et al., 2010), the mitotic plane in dividing ISCs is oriented at an invariable angle with respect to the basement membrane, with the prospective EB located apically (Ohlstein and Spradling, 2007). This observation might indicate an imposed orientation in the division of ISCs that is derived from local cues and guarantees that the cell located basally retains stem cell identity, by preventing N activation. To date, such signals have not been described. Interestingly, however, the visceral muscle that surrounds the intestinal epithelium does secrete cytokines (the Wnt-like molecule wingless) and growth factors (the EGF ligand vein) that may be acting in a concentration gradient on ISCs (Biteau and Jasper, 2011; Buchon et al., 2010; Jiang et al., 2011; Lin et al., 2008; Xu et al., 2011), thus potentially providing cues for the positioning of the division plane. Indeed, EGFR and Frizzled mutant clones were shown to be eliminated from the intestinal epithelium, suggesting that ISCs lacking these receptors lose functionality and/or identity (Lin et al., 2008; Xu et al., 2011). However, an effect of these mutations on cell survival cannot be ruled out. Furthermore, while knocking down Vein in the muscle by RNAi limits the proliferative response of ISCs to stress (Biteau and Jasper, 2011; Buchon et al., 2010; Jiang et al., 2011; Lin et al., 2008; Xu et al., 2011), it remains to be clearly established whether muscle-derived ligands are critical for long-term maintenance of ISCs, or whether they might serve a supporting role while ligands derived from EBs, EEs, or ECs contribute to the ISC ‘niche’. Interestingly, Paneth cells, which constitute the stem cell niche in the mouse intestinal crypt, express the ligands Wnt3, Wnt11 and EGF and are essential to promote stem cell function in vitro and in vivo (Sato et al., 2011). Some of the support mechanisms by which the niche influences stem cell maintenance may thus be conserved between fly and mouse intestines.

It is possible that stem cell division in the Drosophila intestinal epithelium can switch from a consistently asymmetric outcome to symmetric division in response to certain challenges. Supporting this model, recent reports indicate a potential increase in ISC numbers after starvation and refeeding (McLeod et al., 2010), but no direct evidence identifying completely symmetric divisions in the posterior midgut has been reported. In addition, infection, oxidative stress and aging result in the formation of ectopic Dl expressing cells in the posterior midgut. Under these exceptional hyperproliferative conditions, the ISC population might thus resort to symmetric divisions to adjust ISC numbers to tissue requirements. It is important to note, however, that under these conditions ISC daughter cells are also formed the are polyploid and express the ISC/EB marker esg in combination with Dl, while also showing Notch activation. These cells are mis-differentiated and accumulate on the basement membrane, disrupting the epithelium and complicating the analysis of stem cell population size under such conditions (Biteau et al., 2008). A careful analysis using lineage tracing approaches that individually label both stem and daughter cells will be required to confirm a possible switch from asymmetric to symmetric divisions in ISCs (Griffin et al., 2009; Yu et al., 2009).

Self renewal and Differentiation

As suggested by the differential activation of Notch signaling between ISC and EBs, this pathway regulates the balance between self-renewal/stem cell identity and differentiation, both in mammals and flies. In mouse, the effect of Notch activation on stem and progenitor cell populations is highly context dependent (for review see (Liu et al., 2010a)). In hematopoietic, neuronal, and intestinal lineages, Notch is required for self-renewal and promotes proliferation (Aguirre et al., 2010; Duncan et al., 2005; Fre et al., 2005; Hitoshi et al., 2002; Imayoshi et al., 2010; van Es et al., 2005), whereas activation of the Notch pathway induces differentiation of SC populations in the epidermis and the mammary epithelium (Blanpain et al., 2006; Bouras et al., 2008). In depth characterization of the exact mechanisms by which Notch regulates stem cell identity and differentiation, as well as its interactions with other signaling pathways, will be required to fully understand the control of stem cell self-renewal and differentiation. Studies in flies, where Notch promotes differentiation and regulates cell fate decision of ISC daughter cells, have already shed new light on these processes, providing new insight also into the regulation of mammalian stem cell systems: Notch signaling was recently shown to regulate cell-fate decisions in the regenerating airway epithelium in a manner reminiscent of the Drosophila ISC lineage (Rock et al., 2011). Dl/N signaling in the fly ISC lineage promotes differentiation of EBs into either EEs or ECs. The cell fate decision between ECs and EEs seems to be regulated by the intensity of the Dl signal, i.e. high levels of N activity in EBs result in EC differentiation, while the absence of N activation promotes EE differentiation (Ohlstein and Spradling, 2007). This regulation is conserved in vertebrate airway epithelia (Rock et al., 2011). Similar to fly ISCs, basal stem cells (BCs) express the Notch ligands Dll1 and Jagged2, activating N in their daughter cells. N promotes BC differentiation, and it has been proposed that high N activation promotes differentiation into the secretory fate, while lower activation of N results in differentiation into ciliated cells (Rock et al., 2011). Both in flies and in mice it remains unclear, however, how the intensity of Notch activation in daughter cells, and thus their lineage decision, is regulated.

Other studies have provided important insight into the mechanisms controlling ISC daughter cell differentiation: Maintaining the close proximity of ISCs and EBs through ECadherin-mediated cell adhesion is crucial for successful EB differentiation (Maeda et al., 2008). Dl-mediated N activation in EBs increases the activity of the Suppressor of Hairless (Su(H)) transcription factor, presumably by replacing the Hairless transcriptional repressor from Enhancer of Split (E(spl)) complex promoters with the Notch intracellular domain (N^ICD) (Bardin et al., 2010). Differentiation of EBs into the EC lineage also requires an active JAK/Stat signaling pathway (Beebe et al., 2010; Jiang et al., 2009). However, the exact role of this pathway in the control of ISC proliferation, maintenance and EB cell fate decision, as well as its integration with the Notch signaling pathway remain uncertain due to conflicting reports (Beebe et al., 2010; Jiang et al., 2009; Lin et al., 2010; Liu et al., 2010b).