**2. Physiological pathways involved in the regulation of stem cells**

HSC fate decisions are supported by the orchestration of several pathways such as Wnt, Notch and Hedgehog pathways that critically balance cell cycling and quiescence, leading to proliferation and apoptosis, self-renewal or differentiation (Fig. 3). The ultimate decision is dependent on hundreds of inputs including concentrations of different growth factors, cytokines, hormones, oxygen levels that must be integrated to subsequently activate these different signal transduction cascades. Understanding their regulation might led to the effective and more spread out utilization of HSC in clinical settings. The most relevant aspects of these pathways are briefly resumed below.

Fig. 3. A schematic representation of signaling pathways collectively influencing stem cell fate.

#### **2.1 The hedgehog (Hh) signaling**

In adult tissues, Hh signaling is involved in the maintenance of stem cells, regeneration and tissue repair where it governs processes like cell proliferation, cell renewal and differentiation. The three Hh ligand homologues: Sonic Hh, Indian Hh, and Desert Hh bind interchangeably the two related twelve-pass membrane Patched (Ptc) receptors. They relieve the inhibition of smoothened (SMO), a serpentine receptor resembling G protein–coupled receptors allowing activation of a family of zinc-finger transcription factors called GLI and the modification of the expression Hh target genes (Kasper et al., 2009).

The development of HSC-based therapies however, is to some extent prevented by the scarce representation of HSC in the BM and their finite lifespan *ex vivo*. Increasing their utilisation needs enhancement of hematopoietic stem cells availability or *de novo* generation of HSC. This presumes i) the development of robust methods to efficiently control HSC regulatory processes; ii) the therapeutic *in vivo* or *in vitro* expansion of HSC number and iii)

HSC fate decisions are supported by the orchestration of several pathways such as Wnt, Notch and Hedgehog pathways that critically balance cell cycling and quiescence, leading to proliferation and apoptosis, self-renewal or differentiation (Fig. 3). The ultimate decision is dependent on hundreds of inputs including concentrations of different growth factors, cytokines, hormones, oxygen levels that must be integrated to subsequently activate these different signal transduction cascades. Understanding their regulation might led to the effective and more spread out utilization of HSC in clinical settings. The most relevant

Fig. 3. A schematic representation of signaling pathways collectively influencing stem cell

In adult tissues, Hh signaling is involved in the maintenance of stem cells, regeneration and tissue repair where it governs processes like cell proliferation, cell renewal and differentiation. The three Hh ligand homologues: Sonic Hh, Indian Hh, and Desert Hh bind interchangeably the two related twelve-pass membrane Patched (Ptc) receptors. They relieve the inhibition of smoothened (SMO), a serpentine receptor resembling G protein–coupled receptors allowing activation of a family of zinc-finger transcription factors called GLI and

the modification of the expression Hh target genes (Kasper et al., 2009).

the utilisation of optimized protocols to generate available HSC from ESC or IPSC.

**2. Physiological pathways involved in the regulation of stem cells** 

aspects of these pathways are briefly resumed below.

fate.

**2.1 The hedgehog (Hh) signaling** 

The role of Hh signaling in HSC is controversial. Bhardwaj et al provided evidence for a role of Hh signaling in HSC (Bhardwaj et al., 2001). In this study, suppression of Hh signaling inhibited proliferation of HSC and addition of soluble SHh induced expansion of hematopoietic repopulating cells (Bhardwaj et al., 2001). More recent reports confirmed that suppression of the Hh pathway leads to a severe defect in HSC functions (Merchant et al., 2010; Trowbridge et al., 2006) whereas others reported that this pathway can be dispensable for HSC biology (Gao et al., 2009; Hofmann et al., 2009). In *Ptc1*+/−mice, which have increased Hh activity, activation of the Hh signaling pathway induces expansion of primitive blood cells under homeostatic conditions. However, when HSC are challenged to regenerate the blood system, persistent Hh activation leads to HSC exhaustion (Trowbridge et al., 2006). Furthermore, Indian Hh gene transfer can confer enhanced hematopoietic support ability to BM stromal cells, suggesting that it is involved in the interaction between HSC and the stromal cells. This leads to an increase in proliferation and repopulating capacity of primitive hematopoietic cells (Kobune et al., 2004). These results suggest a role for Hh signaling in balancing homeostasis and regeneration *in vivo*. In contrast, other reports show that Hh signaling is dispensable for adult HSC functions (Gao et al., 2009; Hofmann et al., 2009). In these studies conditional deletion of SMO, the only non redundant component of the Hh cascade, or pharmacologic inhibition of Hh signaling have no apparent effect on adult hematopoietic, including peripheral blood count, number or cell cycle status of stem or progenitor cells, hematopoietic colony-forming potential or long-term repopulating activity in *in vivo* assays. In agreement with this notion, genome-wide transcriptome analysis revealed that silencing the Hh signaling does not significantly alter the HSC-specific gene expression ''signature.'' Taken together, these conflicting data suggest that Hh signaling may influence HSC through more complex networks such as cell-niche interactions.

#### **2.2 Fibroblast growth factor (FGF) signaling**

FGF belongs to a family of heparin-binding polypeptides that shows multiple functions, including effects on cell proliferation, differentiation and survival (Baird, 1994). Twenty-four members of the FGF family have been identified in human and mice. FGFs bind and activate their cognate FGFRs that are encoded by four genes (FGFR1– 4). This results in receptor dimerization, tyrosine kinase autophosphorylation, and recruitment of signaling complexes. The FGF signal transduction proceeds by one, or a combination, of three main pathways: Ras/mitogen-activated protein kinase (MAPK) signaling; planar cell polarity/calcium; phosphoinotitide-3-kinase (PI3K)/Akt (extensively reviewed by Bottcher and Niehrs, 2005). Both FGF-1 and FGF-2 support HSC expansion when unfractionated mouse BM cells are cultured in serum-free medium (de Haan et al., 2003; Yeoh et al., 2006). Crcareva et al*.* confirmed that FGF-1 stimulates ex-vivo expansion of HSC (Crcareva et al., 2005). Conditional derivatives of FGFR-1 have also been used to support short-term HSC expansion and long-term HSC survival (Weinreich et al., 2006). This factor seems to also support *ex vivo* expansion of murine and human HSC in combination with other cytokines, i.e stem cell factor [SCF], thrombopoietin [TPO], insulin-like growth factor-2 [IGF-2], and fibroblast growth factor-1 [FGF-1] (Zhang and Lodish, 2005). Moreover, a recent study showed that addition of SCF, TPO, and FGF-1 to a mesenchymal stem cells (MSC) culture stimulates proliferation, maintenance of primitive immunophenotype, and expansion of CFU-initiating cells. This supports the notion that expansion of HSC requires complex stimulation of different signal cascades activated by soluble growth factors as well as adhesion proteins (Walenda et al., 2011).

Searching for the Key to Expand Hematopoietic Stem Cells 211

recovery (16 days in patients receiving the expanded unit, compared to 26 days in patients of the concurrent cohort). Similarly, preliminary evaluation of time needed for platelet recovery compared favourably in those patient receiving the expanded cell product compared with those receiving non-manipulated cells (Dahlberg et al., 2011). In addition, comparable overall survival and graft-versus-host disease risk of patient receiving nonmanipulated cells was observed within the average follow-up of 354 days. The expanded cell population may also have retained long-term repopulating capacities as two patients display *in vivo* persistence of cultured donor cells. The lack of *in vivo* persistence in the remaining patients may either be due to loss of stem cell self-renewal capacity during *ex vivo* culture or to immune mediated rejection. Indeed, it has been well documented that in most of the patients who received two non-manipulated cord blood units for transplantation, only one contributes to persistent long-term engraftment. The mechanism responsible for this single donor dominance remains yet to be defined. Larger phase II/III studies are required to evaluate whether co-infusion of this expanded cell product decreases the occurrence of serious infection, improves survival, or affects duration of hospital stay

The TGF superfamily consist of a large collection of secreted proteins that regulate cell growth, differentiation, apoptosis, cellular homeostasis, and other functions in both the adult organism and the developing embryo. The more than 30 TGF family ligands are organized into three subgroups (reviewed in (Lyssiotis et al., 2011)). The TGF (which comprises SMAD and Activin/Nodal ligands), bone morphogenetic protein (BMP), and the growth differentiation factors (GDF). The TGF signaling leads to the phosphorylation of Smads by activated receptors resulting in their partnering with the common signaling transducer Smad4, and translocation to the nucleus. Once activated, Smads regulate diverse biological effects by partnering with transcription factors resulting in cell-state specific

A significant number of studies have demonstrated that TGF inhibits proliferation of both murine and human HSC *in vitro*. It was suggest that TGF induces quiescence in HSC since its neutralization was showed to release early hematopoietic progenitors cells from quiescence (Hatzfeld et al., 1991; Yamazaki et al., 2009). In agreement with studies performed *in vitro*, injection of TGF1 into the femoral artery of mice effectively inhibits proliferation of multipotent hematopoietic progenitors in the BM, establishing an inhibitory role of TGF1 also *in vivo* (Goey et al., 1989). Despite a key role *in vitro*, TGFdid not seem to provide the necessary signals that maintain quiescence and the stem cell pool *in vivo* 

To block the entire Smad signaling pathway, the Smad7 was overexpressed in murine HSC using a retroviral gene transfer approach. Forced expression of Smad7 significantly increased the self-renewal capacity of HSC *in vivo* (Blank et al., 2006). In a similar approach using human hematopoietic cells, overexpression of Smad7 resulted in a shift from lymphoid-dominant engraftment toward the myeloid lineage, and an increase of the myeloid-committed clonogenic progenitor frequency in NOD-SCID mice (Chadwick et al., 2005). Instead, Smad4-deficient HSC displayed a significantly reduced repopulative capacity

(Delaney et al., 2010).

(Larsson et al., 2005).

**2.4 The transforming growth factor beta (TGF) superfamily** 

modulation of transcription (Kaivo-Oja et al., 2003).

#### **2.3 Notch signaling**

The Notch pathway is also an evolutionarily conserved mechanism that plays a fundamental role in regulating cell-fate decisions (Bolos et al., 2007). Four types of Notch receptors (Notch 1-4) and five Notch ligands (Jagged 1 and 2, Delta 1, 3 and 4) have been identified in vertebrates. Notch ligands are single-pass transmembrane proteins consisting of multiple EGF-like repeats and a characteristic DSL (Delta, Serrate, and LAG-2) domain (see for review Ohishi et al., 2003; Shimizu et al., 2000). One characteristic of this signaling pathway is the dual role of Notch as both a transmembrane receptor and a transcription factor in a system where no second messengers are used (Matsuno et al., 1995). Notch can have opposite functions in different self-renewing organs indicating that the outcome of Notch activation depends to a great extent on the cell context and the specific growth factors present in the microenvironment. For example, activation of Notch1 by Delta ligands 1 and 4 is required for inducing T-cell and inhibiting B-cell differentiation whereas Notch2 activation by Jagged1, and possibly Delta1, acts on HSC (Han et al., 2002; Radtke et al., 1999; Varnum-Finney et al., 2011).

A role for Notch in hematopoietic was initially suggested by detection of the human Notch1 gene in CD34+ or lineage (Lin)- CD34+ hematopoietic cells (Milner et al., 1994). Transduction of murine HSC with a retrovirus expressing a constitutively active form of Notch1 induced the emergence of an immortalized pluripotent cytokine-dependent cell line capable of both myeloid and lymphoid repopulation *in vivo*, thereby demonstrating a role for Notch in HSC self-renewal (Varnum-Finney et al., 2000). Similar results were obtained using an immobilized form of the Notch ligand Delta-1 since incubation of murine HSC with immobilized Delta-1 and cytokines led to a several-log expansion of cells capable of shortterm *in vivo* reconstitution (Varnum-Finney et al., 2003).

In contrast to the murine studies, only a modest or no increase in the progenitor numbers was achieved by expressing activated Notch-1 in human CD34+ cord blood cells (Carlesso et al., 1999; Chadwick et al., 2007) or by incubation with Delta-1 (Jaleco et al., 2001), Delta-4 (Lauret et al., 2004) or Jagged-1 (Karanu et al., 2000; Karanu et al., 2001; Walker et al., 1999). This contrast with other reports showing that incubation of human cord blood cells with the immobilized Delta-1 combined with fibronectin fragments and cytokines induce a 100-fold increase in the number of CD34+ cells compare to controls (Ohishi et al., 2002) and a 16-fold increase in SCID Repopulating Cells (SRC) number compared to uncultured cells. *In vivo* transplanted cells persisted 9 weeks post-transplantation and in secondary recipients, suggesting the presence of both long-term and short-term repopulating cells following culture of human cord blood cells on Delta-1 ligand (Delaney et al., 2010). The SRC enhancement by relatively low density of immobilized ligand and the preference to promote differentiation toward the T-cell lineage at higher ligand density revealed important ligand dose-dependent effects of Notch signaling (Delaney et al., 2005).

The engineered Notch ligand approach for *ex vivo* expansion of human cord blood cells is now under clinical investigation (http://clinicaltrials.gov/ct2/show/record/ NCT00343798). In this phase 1 clinical trial, patients undergoing a myeloablative double cord blood transplantation are receiving one non-manipulated cord blood unit along with a second cord blood unit that has undergone Notch-mediated *ex vivo* expansion. These cells were safely infused and led to a significant reduction in the time needed for neutrophil

The Notch pathway is also an evolutionarily conserved mechanism that plays a fundamental role in regulating cell-fate decisions (Bolos et al., 2007). Four types of Notch receptors (Notch 1-4) and five Notch ligands (Jagged 1 and 2, Delta 1, 3 and 4) have been identified in vertebrates. Notch ligands are single-pass transmembrane proteins consisting of multiple EGF-like repeats and a characteristic DSL (Delta, Serrate, and LAG-2) domain (see for review Ohishi et al., 2003; Shimizu et al., 2000). One characteristic of this signaling pathway is the dual role of Notch as both a transmembrane receptor and a transcription factor in a system where no second messengers are used (Matsuno et al., 1995). Notch can have opposite functions in different self-renewing organs indicating that the outcome of Notch activation depends to a great extent on the cell context and the specific growth factors present in the microenvironment. For example, activation of Notch1 by Delta ligands 1 and 4 is required for inducing T-cell and inhibiting B-cell differentiation whereas Notch2 activation by Jagged1, and possibly Delta1, acts on HSC (Han et al., 2002; Radtke et al., 1999;

A role for Notch in hematopoietic was initially suggested by detection of the human Notch1

of murine HSC with a retrovirus expressing a constitutively active form of Notch1 induced the emergence of an immortalized pluripotent cytokine-dependent cell line capable of both myeloid and lymphoid repopulation *in vivo*, thereby demonstrating a role for Notch in HSC self-renewal (Varnum-Finney et al., 2000). Similar results were obtained using an immobilized form of the Notch ligand Delta-1 since incubation of murine HSC with immobilized Delta-1 and cytokines led to a several-log expansion of cells capable of short-

In contrast to the murine studies, only a modest or no increase in the progenitor numbers was achieved by expressing activated Notch-1 in human CD34+ cord blood cells (Carlesso et al., 1999; Chadwick et al., 2007) or by incubation with Delta-1 (Jaleco et al., 2001), Delta-4 (Lauret et al., 2004) or Jagged-1 (Karanu et al., 2000; Karanu et al., 2001; Walker et al., 1999). This contrast with other reports showing that incubation of human cord blood cells with the immobilized Delta-1 combined with fibronectin fragments and cytokines induce a 100-fold increase in the number of CD34+ cells compare to controls (Ohishi et al., 2002) and a 16-fold increase in SCID Repopulating Cells (SRC) number compared to uncultured cells. *In vivo* transplanted cells persisted 9 weeks post-transplantation and in secondary recipients, suggesting the presence of both long-term and short-term repopulating cells following culture of human cord blood cells on Delta-1 ligand (Delaney et al., 2010). The SRC enhancement by relatively low density of immobilized ligand and the preference to promote differentiation toward the T-cell lineage at higher ligand density revealed important ligand

The engineered Notch ligand approach for *ex vivo* expansion of human cord blood cells is now under clinical investigation (http://clinicaltrials.gov/ct2/show/record/ NCT00343798). In this phase 1 clinical trial, patients undergoing a myeloablative double cord blood transplantation are receiving one non-manipulated cord blood unit along with a second cord blood unit that has undergone Notch-mediated *ex vivo* expansion. These cells were safely infused and led to a significant reduction in the time needed for neutrophil

CD34+ hematopoietic cells (Milner et al., 1994). Transduction

**2.3 Notch signaling** 

Varnum-Finney et al., 2011).

gene in CD34+ or lineage (Lin)-

term *in vivo* reconstitution (Varnum-Finney et al., 2003).

dose-dependent effects of Notch signaling (Delaney et al., 2005).

recovery (16 days in patients receiving the expanded unit, compared to 26 days in patients of the concurrent cohort). Similarly, preliminary evaluation of time needed for platelet recovery compared favourably in those patient receiving the expanded cell product compared with those receiving non-manipulated cells (Dahlberg et al., 2011). In addition, comparable overall survival and graft-versus-host disease risk of patient receiving nonmanipulated cells was observed within the average follow-up of 354 days. The expanded cell population may also have retained long-term repopulating capacities as two patients display *in vivo* persistence of cultured donor cells. The lack of *in vivo* persistence in the remaining patients may either be due to loss of stem cell self-renewal capacity during *ex vivo* culture or to immune mediated rejection. Indeed, it has been well documented that in most of the patients who received two non-manipulated cord blood units for transplantation, only one contributes to persistent long-term engraftment. The mechanism responsible for this single donor dominance remains yet to be defined. Larger phase II/III studies are required to evaluate whether co-infusion of this expanded cell product decreases the occurrence of serious infection, improves survival, or affects duration of hospital stay (Delaney et al., 2010).

#### **2.4 The transforming growth factor beta (TGF) superfamily**

The TGF superfamily consist of a large collection of secreted proteins that regulate cell growth, differentiation, apoptosis, cellular homeostasis, and other functions in both the adult organism and the developing embryo. The more than 30 TGF family ligands are organized into three subgroups (reviewed in (Lyssiotis et al., 2011)). The TGF (which comprises SMAD and Activin/Nodal ligands), bone morphogenetic protein (BMP), and the growth differentiation factors (GDF). The TGF signaling leads to the phosphorylation of Smads by activated receptors resulting in their partnering with the common signaling transducer Smad4, and translocation to the nucleus. Once activated, Smads regulate diverse biological effects by partnering with transcription factors resulting in cell-state specific modulation of transcription (Kaivo-Oja et al., 2003).

A significant number of studies have demonstrated that TGF inhibits proliferation of both murine and human HSC *in vitro*. It was suggest that TGF induces quiescence in HSC since its neutralization was showed to release early hematopoietic progenitors cells from quiescence (Hatzfeld et al., 1991; Yamazaki et al., 2009). In agreement with studies performed *in vitro*, injection of TGF1 into the femoral artery of mice effectively inhibits proliferation of multipotent hematopoietic progenitors in the BM, establishing an inhibitory role of TGF1 also *in vivo* (Goey et al., 1989). Despite a key role *in vitro*, TGFdid not seem to provide the necessary signals that maintain quiescence and the stem cell pool *in vivo*  (Larsson et al., 2005).

To block the entire Smad signaling pathway, the Smad7 was overexpressed in murine HSC using a retroviral gene transfer approach. Forced expression of Smad7 significantly increased the self-renewal capacity of HSC *in vivo* (Blank et al., 2006). In a similar approach using human hematopoietic cells, overexpression of Smad7 resulted in a shift from lymphoid-dominant engraftment toward the myeloid lineage, and an increase of the myeloid-committed clonogenic progenitor frequency in NOD-SCID mice (Chadwick et al., 2005). Instead, Smad4-deficient HSC displayed a significantly reduced repopulative capacity

Searching for the Key to Expand Hematopoietic Stem Cells 213

The individual contribution of these pathways to the hematopoietic development of HSC have been extensively addressed (Cerdan and Bhatia, 2010). However, there are many potential intersections along them and therefore the impact of their collective contribution towards influencing the fate of HSC should be carefully considered. Some of these

Ducan et al. provide a model for how HSC may integrate multiple signals to maintain the stem cell state. They showed that although the proliferation and survival of HSC exposed to Wnt proteins seem unaffected when Notch signaling is impaired, their ability to remain undifferentiated is substantially altered (Duncan et al., 2005). These results demonstrated that the Notch pathway is imperative in maintaining HSC in an undifferentiated state. These findings do not preclude the possibility that a stronger Wnt signal, such as activated -catenin, may be able to overcome the consequences of loss of Notch signaling. Moreover, Wnt3a regulates the expression of established Notch target genes (Duncan et al., 2005) and the inhibition of GSK-3, a downstream target of Wnt signaling that affects HSC fate through mechanisms involving both Wnt and Notch target genes (Trowbridge et al., 2006). These findings suggest that these pathways could play a role in HSC self renewal using a common network of regulatory circuits with Wnt enhancing proliferation and survival, and Notch preventing differentiation (Blank et al.,

Furthermore, there is substantial evidence for the cross-talk between the Wnt signaling pathway and FGFs and TGF-b by means of the association between Smad4 and Hox proteins. *Homeobox* (*hox*) genes encode transcription factors that function as regulators of hematopoiesis and are frequently dysregulated in human leukemia, particularly acute myeloid leukemia (Kroon et al., 1998). Recently, Wang et al described a mechanism whereby TGF-β/BMP inhibited the BM transformation capacity of HoxA9 and HoxA9-Nup98 fusion protein through a Smad4-dependent mechanism. Accordingly, Smad4 was shown to interact directly with HoxA9 and Nup98-HoxA9 fusion protein, thus precluding their DNA binding capacity and subsequent transcriptional activity (Wang et al., 2006). Smad4 also seems to participate in other signaling cascades such as Wnt or Notch (Itoh et al., 2004; Labbe et al.,

These studies show the high interdependence between the different pathways, and the impact of their collective contribution on HSC self-renewal. This should be carefully

Epigenetic modifications, in addition to the intracellular pathways described in the previous section also play an essential role in regulating self-renewal, differentiation and tissue development. They induce gene expression regulation and can be grouped into three main categories: i) DNA methylation, ii) Histone modifications and iii) Nucleosome positioning. Recent studies suggest that epigenetic mechanisms contribute to establish the HSC unique

considered when trying to expand HSC for clinical purposes.

characteristics. The following is a description of some of these examples.

**2.7 Epigenetic control and HSC self-renewal** 

**2.6 Cross-talk between these pathways** 

intersection points are resumed below.

2008).

2000).

of primary and secondary recipients (Karlsson et al., 2007). Because overexpression of Smad7 versus deletion of Smad4 would be anticipated to yield similar hematopoietic phenotypes, it is conceivable that Smad4 functions as a positive regulator of self-renewal independently of its role as a central mediator of the canonical Smad pathway. In the context of adult hematopoiesis, a high concentration of BMP-4 was shown to promote maintenance of human cord blood cells *in vitro*, while lower concentration of BMP4, BMP2 and BMP7 induced proliferation and differentiation of HSC (Bhatia et al., 1999).

### **2.5 Wingless-type (Wnt) pathway**

Wnt proteins are secreted morphogens necessaries for basic developmental processes, such as cell-fate specification, progenitor-cell proliferation and the control of asymmetric cell division, in many different species and organs (Bejsovec, 2005; Moon et al., 2004). Wnt proteins bind to cell surface receptors of the Frizzled family which can translocate the signals to the nucleus and function as transcriptional activators through intracellular catenin. Different Wnt pathways are known but their clear separation and their independence remain controversial. There is one canonical pathway that acts on the stability of -catenin and interacts with T cell transcription factors in the nucleus. There are many non-canonical pathways like the PCP and Wnt/Calcium pathways. The most distinctive differences between the canonical and non-canonical pathways include the specific ligands activating each pathway, ß-Catenin, LRP5/6 co-receptor, and Dsh-DEP domain independence, respectively, and the ability of the non-canonical pathways to inhibit the canonical pathway. Ligands that activate the non-canonical pathways are Wnt4, Wnt5a, and Wnt11.

Recent evidence based on genetic models suggests that canonical Wnt signaling, regulates HSC self-renewal. Active -catenin promotes HSC proliferation and inhibits differentiation (Kirstetter et al., 2006; Scheller et al., 2006) whereas deficiency in -catenin inhibits HSC selfrenewal (Cobas et al., 2004; Luis et al., 2009; Zhao et al., 2007). Moreover, purified Wnt3a treatment of adult HSC increases self-renewal of murine HSC, as determined by *in vivo* reconstituting assays (Willert et al., 2003) and of human Lin-CD34+ cells as measured by immunophenotype and colony assays (Van Den Berg et al., 1998).

The role of the non-canonical pathways is not well defined, but surprisingly, their activation and consequently inhibition of the canonical pathway, appears also to be able to expand HSC. Murdoch et al. demonstrated that injecting mice with Wnt5a conditioned media prior to transplant of human umbilical cord blood cells increased engraftment more than 3-fold (Murdoch et al., 2003). Furthermore, culturing Lin-Sca-1+c-Kit+ (LSK) cells with recombinant murine Wnt5a resulted in an enhancement of hematopoietic reconstitution in a BM transplant assay. Wnt5a seems to activate the non-canonical signaling pathways leading to a 3.5- fold more HSC in G0 phase (Nemeth et al., 2007).

Overexpression of Wnt4 led to a modest increase in HSC frequency as measured by phenotype and limiting dilution transplant assays and Wnt4-/- mice showed decreased frequencies of HSC in BM. Similar to the results obtained using Wnt5a, overexpression of Wnt4 led to an increase in the percentage of HSC in G0 (Louis et al., 2008). Whether Wnt4 and Wnt5a inhibit the canonical pathway in a similar fashion remains to be elucidated. These results show the importance of a balanced regulation of these two overlapping Wnt signaling pathways.

#### **2.6 Cross-talk between these pathways**

212 Advances in Hematopoietic Stem Cell Research

of primary and secondary recipients (Karlsson et al., 2007). Because overexpression of Smad7 versus deletion of Smad4 would be anticipated to yield similar hematopoietic phenotypes, it is conceivable that Smad4 functions as a positive regulator of self-renewal independently of its role as a central mediator of the canonical Smad pathway. In the context of adult hematopoiesis, a high concentration of BMP-4 was shown to promote maintenance of human cord blood cells *in vitro*, while lower concentration of BMP4, BMP2

Wnt proteins are secreted morphogens necessaries for basic developmental processes, such as cell-fate specification, progenitor-cell proliferation and the control of asymmetric cell division, in many different species and organs (Bejsovec, 2005; Moon et al., 2004). Wnt proteins bind to cell surface receptors of the Frizzled family which can translocate the signals to the nucleus and function as transcriptional activators through intracellular catenin. Different Wnt pathways are known but their clear separation and their independence remain controversial. There is one canonical pathway that acts on the stability of -catenin and interacts with T cell transcription factors in the nucleus. There are many non-canonical pathways like the PCP and Wnt/Calcium pathways. The most distinctive differences between the canonical and non-canonical pathways include the specific ligands activating each pathway, ß-Catenin, LRP5/6 co-receptor, and Dsh-DEP domain independence, respectively, and the ability of the non-canonical pathways to inhibit the canonical pathway. Ligands that activate the non-canonical pathways are Wnt4, Wnt5a, and

Recent evidence based on genetic models suggests that canonical Wnt signaling, regulates HSC self-renewal. Active -catenin promotes HSC proliferation and inhibits differentiation (Kirstetter et al., 2006; Scheller et al., 2006) whereas deficiency in -catenin inhibits HSC selfrenewal (Cobas et al., 2004; Luis et al., 2009; Zhao et al., 2007). Moreover, purified Wnt3a treatment of adult HSC increases self-renewal of murine HSC, as determined by *in vivo* reconstituting assays (Willert et al., 2003) and of human Lin-CD34+ cells as measured by

The role of the non-canonical pathways is not well defined, but surprisingly, their activation and consequently inhibition of the canonical pathway, appears also to be able to expand HSC. Murdoch et al. demonstrated that injecting mice with Wnt5a conditioned media prior to transplant of human umbilical cord blood cells increased engraftment more than 3-fold

murine Wnt5a resulted in an enhancement of hematopoietic reconstitution in a BM transplant assay. Wnt5a seems to activate the non-canonical signaling pathways leading to a

Overexpression of Wnt4 led to a modest increase in HSC frequency as measured by phenotype and limiting dilution transplant assays and Wnt4-/- mice showed decreased frequencies of HSC in BM. Similar to the results obtained using Wnt5a, overexpression of Wnt4 led to an increase in the percentage of HSC in G0 (Louis et al., 2008). Whether Wnt4 and Wnt5a inhibit the canonical pathway in a similar fashion remains to be elucidated. These results show the

importance of a balanced regulation of these two overlapping Wnt signaling pathways.

Sca-1+c-Kit+ (LSK) cells with recombinant

immunophenotype and colony assays (Van Den Berg et al., 1998).

(Murdoch et al., 2003). Furthermore, culturing Lin-

3.5- fold more HSC in G0 phase (Nemeth et al., 2007).

and BMP7 induced proliferation and differentiation of HSC (Bhatia et al., 1999).

**2.5 Wingless-type (Wnt) pathway** 

Wnt11.

The individual contribution of these pathways to the hematopoietic development of HSC have been extensively addressed (Cerdan and Bhatia, 2010). However, there are many potential intersections along them and therefore the impact of their collective contribution towards influencing the fate of HSC should be carefully considered. Some of these intersection points are resumed below.

Ducan et al. provide a model for how HSC may integrate multiple signals to maintain the stem cell state. They showed that although the proliferation and survival of HSC exposed to Wnt proteins seem unaffected when Notch signaling is impaired, their ability to remain undifferentiated is substantially altered (Duncan et al., 2005). These results demonstrated that the Notch pathway is imperative in maintaining HSC in an undifferentiated state. These findings do not preclude the possibility that a stronger Wnt signal, such as activated -catenin, may be able to overcome the consequences of loss of Notch signaling. Moreover, Wnt3a regulates the expression of established Notch target genes (Duncan et al., 2005) and the inhibition of GSK-3, a downstream target of Wnt signaling that affects HSC fate through mechanisms involving both Wnt and Notch target genes (Trowbridge et al., 2006). These findings suggest that these pathways could play a role in HSC self renewal using a common network of regulatory circuits with Wnt enhancing proliferation and survival, and Notch preventing differentiation (Blank et al., 2008).

Furthermore, there is substantial evidence for the cross-talk between the Wnt signaling pathway and FGFs and TGF-b by means of the association between Smad4 and Hox proteins. *Homeobox* (*hox*) genes encode transcription factors that function as regulators of hematopoiesis and are frequently dysregulated in human leukemia, particularly acute myeloid leukemia (Kroon et al., 1998). Recently, Wang et al described a mechanism whereby TGF-β/BMP inhibited the BM transformation capacity of HoxA9 and HoxA9-Nup98 fusion protein through a Smad4-dependent mechanism. Accordingly, Smad4 was shown to interact directly with HoxA9 and Nup98-HoxA9 fusion protein, thus precluding their DNA binding capacity and subsequent transcriptional activity (Wang et al., 2006). Smad4 also seems to participate in other signaling cascades such as Wnt or Notch (Itoh et al., 2004; Labbe et al., 2000).

These studies show the high interdependence between the different pathways, and the impact of their collective contribution on HSC self-renewal. This should be carefully considered when trying to expand HSC for clinical purposes.

#### **2.7 Epigenetic control and HSC self-renewal**

Epigenetic modifications, in addition to the intracellular pathways described in the previous section also play an essential role in regulating self-renewal, differentiation and tissue development. They induce gene expression regulation and can be grouped into three main categories: i) DNA methylation, ii) Histone modifications and iii) Nucleosome positioning. Recent studies suggest that epigenetic mechanisms contribute to establish the HSC unique characteristics. The following is a description of some of these examples.

Searching for the Key to Expand Hematopoietic Stem Cells 215

Polycomb group (PcG) and Trithorax group (TrxG) proteins have emerged as key players in gene regulation and are thought to function coordinately to orchestrate DNA accessibility. These epigenetic regulators act antagonistically to either promote (TrxG) or repress (PcG) transcription through regulation of specific amino acid modifications in histones. It is not known how the PcG and TrxG proteins switch and balance between transcriptionally silenced heterochromatin (for example, enriched in histone H3 lysine 27 trimethylation, H3K27me3) and transcriptionally competent euchromatin (for example, enriched in histone H3 lysine 4 trimethylation, H3K4me3), respectively, during

In vertebrates, polycomb group proteins participate mainly in two complexes, Polycomb Repressive Complex (PRC) 1 and PRC2. Probably the best example of a chromatinassociated factor involved in self-renewal is BMI1, which is a component of PRC1. BMI1 is expressed in HSC and its expression decreases upon differentiation towards myeloid or erythroid cells, but is retained within the lymphoid compartments. Upon deletion of BMI1, no changes in the number of HSC in the fetal liver were observed, but in postnatal BMI1-/ mice, the number of HSC was markedly reduced. Targeted deletion of BMI1 in murine HSC impaired their competitive repopulation capacity (Park et al., 2003). *In vitro*, BMI1-/- HSC proliferated poorly and displayed an accelerated loss of multilineage differentiation potential and overexpression of BMI1 enhanced the self-renewal of HSC and enhanced their

Overexpression of BMI1 in cord blood CD34+ cells resulted in stem cell maintenance. After an *in vitro* culture period of 10 days, BMI1-overexpressing cells display a much better engraftment in NOD-SCID mice. Although the mechanisms involved remain to be elucidated, it was observed in single-cell assays that the percentage of CD34+/CD38-

undergoing apoptosis was reduced, whereas the percentage of quiescent HSC not undergoing cell cycle progression was increased upon BMI1 overexpression (Rizo et al., 2008). Lentiviral downmodulation of BMI1 in human cord blood CD34+ cells impaired longterm expansion, progenitor-forming capacity and stem cell frequencies, both in cytokinedriven liquid cultures and in BM stromal cocultures. This was associated with higher expression of p14ARF and p16INK4A and enhanced apoptosis, which coincided with increased levels of intracellular reactive oxygen species (ROS) and reduced FOXO3A

Another example of a chromatin-associated factor involved in self-renewal is the mixed lineage leukemia (MLL) protein, which encodes a trithorax-group chromatin regulator. Using *Mll*-deficient ESC to generate chimeras, Ernst et al*.* showed a cell-intrinsic requirement for MLL in the generation of lymphoid and myeloid populations in adult animals (Ernst et al., 2004). Moreover, MLL is often fused to the AF9 protein in leukemia and have been reported to impart leukaemia stem cell properties on committed hematopoietic progenitors. The leukemia stem cells generated can maintain the global identity of the progenitor from which they arose while activating a limited stem-cell- or self-renewal-associated programme (Krivtsov et al., 2006). Moreover, this MLL-AF9 fusion drives high-level expression of multiple *Hox* genes and can overcome Bmi1-

deficiency to establish leukemic stem cells (Smith et al., 2011).

HSC

development.

engraftment potential (Iwama et al., 2004).

expression (Rizo et al., 2008).

#### **2.7.1 Methylation of DNA**

The most widely studied epigenetic modification in humans is cytosine methylation. DNA methylation occurs almost exclusively in the context of CpG dinucleotides that tend to cluster in regions called CpG islands. A group of enzymes, the DNA methyltransferases (DNMTs) tightly regulate both the initiation and maintenance of these methyl marks. DNA methylation can inhibit gene expression by various mechanisms. Methylated DNA can promote the recruitment of methyl-CpG-binding domain proteins which in turn recruit histone-modifying and chromatin-remodeling complexes to methylated sites. DNA methylation can also directly inhibit transcription by precluding the recruitment of DNA binding proteins from their target sites. In contrast, unmethylated CpG islands generate a chromatin structure favorable for gene expression (Portela and Esteller, 2010).

Methylation is controlled by at least 3 DNMTs: DNMT3a and DNMT3b for *de novo* methylation and DNMT1 for methylation maintenance. Conditionally disruption of *Dnmt3a*, *Dnmt3b*, or both *Dnmt3a* and *Dnmt3b* (*Dnmt3a/Dnmt3b*) showed that Dnmt3a and Dnmt3b function as *de novo* DNA methyltransferases during differentiation of hematopoietic cells. Unexpectedly, *in vitro* colony assays showed that both myeloid and lymphoid lineage differentiation potentials were maintained in Dnmt3a-, Dnmt3b-, and Dnmt3a/Dnmt3bdeficient HSC. However, Dnmt3a/Dnmt3b-deficient HSC, but not Dnmt3a- or Dnmt3bdeficient HSC, were incapable of long-term reconstitution in transplantation assays, suggesting a role for DNA methylation by Dnmt3a and Dnmt3b in HSC self-renewal (Tadokoro et al., 2007).

Conditional disruption of Dnmt1 in the mouse hematopoietic system revealed defects in self-renewal, niche retention, and in the ability of cells to give rise to multilineage hematopoiesis. Loss of Dnmt1 had specific impact on myeloid progenitor cells, causing enhanced cell cycling and inappropriate expression of mature lineage genes (Trowbridge et al., 2009). Consistent with these results, Broske et *al.* showed that Dnmt1 is essential for HSC self-renewal but dispensable for homing, cell cycle control and suppression of apoptosis but also implicated Dnmt1 in lymphoid differentiation (Broske et al., 2009).

#### **2.7.2 Histone modifications and nucleosome positioning**

A nucleosome is a histone octamer composed by a histone H3-H4 tetramer and two H2A-H2B dimers, around which DNA, 147 base pairs in length, is wrapped in 1.75 superhelical turns. Nucleosomes are connected by the so-called linker DNA and the histone H1. Histones post-transcriptional modifications, including acetylation, methylation, phosphorylation, ubiquitination, SUMOylation and ADP-ribosylation, occur predominantly in histone tails. They have important roles in transcriptional regulation as they can provide either an ON or OFF signature which result in the tight regulation of gene expression but display also important roles in DNA repair, DNA replication, alternative splicing and chromosome condensation. Nucleosomes act as barriers to transcription. They block access of activators and transcription factors to their sites on DNA and inhibit the elongation of the transcripts. The packaging of DNA into nucleosomes appears to affect all stages of transcription, thereby regulating gene expression. Nucleosome positioning plays also an important role in shaping the methylation landscape (Portela and Esteller, 2010).

The most widely studied epigenetic modification in humans is cytosine methylation. DNA methylation occurs almost exclusively in the context of CpG dinucleotides that tend to cluster in regions called CpG islands. A group of enzymes, the DNA methyltransferases (DNMTs) tightly regulate both the initiation and maintenance of these methyl marks. DNA methylation can inhibit gene expression by various mechanisms. Methylated DNA can promote the recruitment of methyl-CpG-binding domain proteins which in turn recruit histone-modifying and chromatin-remodeling complexes to methylated sites. DNA methylation can also directly inhibit transcription by precluding the recruitment of DNA binding proteins from their target sites. In contrast, unmethylated CpG islands generate a

Methylation is controlled by at least 3 DNMTs: DNMT3a and DNMT3b for *de novo* methylation and DNMT1 for methylation maintenance. Conditionally disruption of *Dnmt3a*, *Dnmt3b*, or both *Dnmt3a* and *Dnmt3b* (*Dnmt3a/Dnmt3b*) showed that Dnmt3a and Dnmt3b function as *de novo* DNA methyltransferases during differentiation of hematopoietic cells. Unexpectedly, *in vitro* colony assays showed that both myeloid and lymphoid lineage differentiation potentials were maintained in Dnmt3a-, Dnmt3b-, and Dnmt3a/Dnmt3bdeficient HSC. However, Dnmt3a/Dnmt3b-deficient HSC, but not Dnmt3a- or Dnmt3bdeficient HSC, were incapable of long-term reconstitution in transplantation assays, suggesting a role for DNA methylation by Dnmt3a and Dnmt3b in HSC self-renewal

Conditional disruption of Dnmt1 in the mouse hematopoietic system revealed defects in self-renewal, niche retention, and in the ability of cells to give rise to multilineage hematopoiesis. Loss of Dnmt1 had specific impact on myeloid progenitor cells, causing enhanced cell cycling and inappropriate expression of mature lineage genes (Trowbridge et al., 2009). Consistent with these results, Broske et *al.* showed that Dnmt1 is essential for HSC self-renewal but dispensable for homing, cell cycle control and suppression of apoptosis but

A nucleosome is a histone octamer composed by a histone H3-H4 tetramer and two H2A-H2B dimers, around which DNA, 147 base pairs in length, is wrapped in 1.75 superhelical turns. Nucleosomes are connected by the so-called linker DNA and the histone H1. Histones post-transcriptional modifications, including acetylation, methylation, phosphorylation, ubiquitination, SUMOylation and ADP-ribosylation, occur predominantly in histone tails. They have important roles in transcriptional regulation as they can provide either an ON or OFF signature which result in the tight regulation of gene expression but display also important roles in DNA repair, DNA replication, alternative splicing and chromosome condensation. Nucleosomes act as barriers to transcription. They block access of activators and transcription factors to their sites on DNA and inhibit the elongation of the transcripts. The packaging of DNA into nucleosomes appears to affect all stages of transcription, thereby regulating gene expression. Nucleosome positioning plays also an important role in

chromatin structure favorable for gene expression (Portela and Esteller, 2010).

also implicated Dnmt1 in lymphoid differentiation (Broske et al., 2009).

**2.7.2 Histone modifications and nucleosome positioning** 

shaping the methylation landscape (Portela and Esteller, 2010).

**2.7.1 Methylation of DNA** 

(Tadokoro et al., 2007).

Polycomb group (PcG) and Trithorax group (TrxG) proteins have emerged as key players in gene regulation and are thought to function coordinately to orchestrate DNA accessibility. These epigenetic regulators act antagonistically to either promote (TrxG) or repress (PcG) transcription through regulation of specific amino acid modifications in histones. It is not known how the PcG and TrxG proteins switch and balance between transcriptionally silenced heterochromatin (for example, enriched in histone H3 lysine 27 trimethylation, H3K27me3) and transcriptionally competent euchromatin (for example, enriched in histone H3 lysine 4 trimethylation, H3K4me3), respectively, during development.

In vertebrates, polycomb group proteins participate mainly in two complexes, Polycomb Repressive Complex (PRC) 1 and PRC2. Probably the best example of a chromatinassociated factor involved in self-renewal is BMI1, which is a component of PRC1. BMI1 is expressed in HSC and its expression decreases upon differentiation towards myeloid or erythroid cells, but is retained within the lymphoid compartments. Upon deletion of BMI1, no changes in the number of HSC in the fetal liver were observed, but in postnatal BMI1-/ mice, the number of HSC was markedly reduced. Targeted deletion of BMI1 in murine HSC impaired their competitive repopulation capacity (Park et al., 2003). *In vitro*, BMI1-/- HSC proliferated poorly and displayed an accelerated loss of multilineage differentiation potential and overexpression of BMI1 enhanced the self-renewal of HSC and enhanced their engraftment potential (Iwama et al., 2004).

Overexpression of BMI1 in cord blood CD34+ cells resulted in stem cell maintenance. After an *in vitro* culture period of 10 days, BMI1-overexpressing cells display a much better engraftment in NOD-SCID mice. Although the mechanisms involved remain to be elucidated, it was observed in single-cell assays that the percentage of CD34+/CD38- HSC undergoing apoptosis was reduced, whereas the percentage of quiescent HSC not undergoing cell cycle progression was increased upon BMI1 overexpression (Rizo et al., 2008). Lentiviral downmodulation of BMI1 in human cord blood CD34+ cells impaired longterm expansion, progenitor-forming capacity and stem cell frequencies, both in cytokinedriven liquid cultures and in BM stromal cocultures. This was associated with higher expression of p14ARF and p16INK4A and enhanced apoptosis, which coincided with increased levels of intracellular reactive oxygen species (ROS) and reduced FOXO3A expression (Rizo et al., 2008).

Another example of a chromatin-associated factor involved in self-renewal is the mixed lineage leukemia (MLL) protein, which encodes a trithorax-group chromatin regulator. Using *Mll*-deficient ESC to generate chimeras, Ernst et al*.* showed a cell-intrinsic requirement for MLL in the generation of lymphoid and myeloid populations in adult animals (Ernst et al., 2004). Moreover, MLL is often fused to the AF9 protein in leukemia and have been reported to impart leukaemia stem cell properties on committed hematopoietic progenitors. The leukemia stem cells generated can maintain the global identity of the progenitor from which they arose while activating a limited stem-cell- or self-renewal-associated programme (Krivtsov et al., 2006). Moreover, this MLL-AF9 fusion drives high-level expression of multiple *Hox* genes and can overcome Bmi1 deficiency to establish leukemic stem cells (Smith et al., 2011).

Searching for the Key to Expand Hematopoietic Stem Cells 217

Soluble growth factors, such as ANGPLT5 and IGFBP2, produced by the endothelium may enhance HSC expansion *ex vivo* when used with conventional cytokines. Although the addition of ANGPLT5 and/or IGFBP2 to a 10 days-human CD133+ cord blood cells culture has no effect upon the total nucleated cells number *in vitro*, it significantly enhances *in vivo* repopulation of NOD-SCID mice 2 months post-transplantation as well as secondary transplantation (Zhang et al., 2008a). These results were confirmed recently using human cord blood CD34+CD133+ cells cultured for 10 days in the presence of IGFBP2 and ANGPLT5. Expanded cells were shown to be capable of long-term multi-lineage and multi-

Pleiotropin, which have mitogenic and angiogenic activities, has been found to be essential for maintenance of murine HSC. Mice transplanted with LSK CD34- cells treated with Ptn and a standard cocktail of cytokines showed 6-fold increase in HSC frequency compared to cells treated with cytokines alone. *In vivo,* systemic administration of Ptn was found to increase the number of BM LSK cells both in irradiated and nonirradiated mice, suggesting a role for this factor in the *in vivo* regeneration of HSC. Treatment of human cord blood Lin-CD34+CD38- cells with Ptn for 7 days induced a 4-fold increase in CFC content and a 3- or 7 fold improved engraftment at 4 or 7 weeks respectively in NOD-SCID mice compared with controls. This factor may activate the PI3-Kinase/AKT and Notch pathways by alleviating

The homeobox gene family member HoxB4 is the most investigated transcription factor for its potential to increase the self-renewal potential of HSC. HOXB4 belongs to a large family of transcription factors that share a highly conserved DNA-binding domain, the homeodomain. In mammals, there are 39 *Hox* genes grouped in four clusters referred to as A, B, C and D. In the hematopoietic system, 16 different *Hox* genes are transcribed during normal hematopoiesis. Primitive subpopulations express primarily genes of the A and B cluster (Giampaolo et al., 1995; Pineault et al., 2002; Sauvageau et al., 1994). Mice transplanted with marrow overexpressing HOXB4 resulted in a 47-fold increase of the competitive repopulating unit (CRU) numbers and did not develop leukemic transformation (Sauvageau et al., 1995). *HOXB4* overexpression in mouse HSC cultured for 14 days induced a primitive cell-specific growth advantage contrary to a progressive depletion of HSC usually observed under these conditions. Total cell growth (mostly mature cells) was enhanced by 2-fold, progenitors by 3-fold and HSC by 1000-fold in cells overexpressing

In humans, transient overexpression of HOXB4 in hematopoietic cord blood cells, did not increase proliferation of primitive progenitors, frequency of CFC, and LTC-ICs but induced an iincrease in myeloid differentiation (Brun et al., 2003). Other studies showed that

**3.1.2 Angiopoietin-like 5 (ANGPLT5) and insulin-like growth factor binding protein 2** 

site hematopoiesis in serial reconstitution in NSG mice (Drake et al., 2011).

activation of its receptor, RPTP-/ (Himburg et al., 2010).

**3.2 Transcription factors: The HOX- family** 

HOXB4 (Antonchuk et al., 2002).

**(IGFBP2)** 

**3.1.3 Pleiotropin (Ptn)** 

**3.2.1 HOXB4** 

The studies described in this section establish that epigenetic alterations can modulate the self-renewal process. Epigenetic state in stem cells can be stably heritable or can be erased (partly or completely) by cell division. These changes might facilitate the transition of a progenitor cell to a self-renewing stem cell, or might prompt a stem cell to differentiate, divide or lose its ability to self-renew.
