**4. Selection of genetically modified HSCs in vivo: Positive selection**

Ongoing studies of the mechanisms controlling HSC self-maintenance and commitment continue to identify novel factors that bring about HSC expansion in vivo when over-expressed. A less than exhaustive set of these factors is listed in the Table 2. Arguably, the most extensively studied gene with such properties is the homeobox transcription factor HoxB4. Forced expression of HoxB4 in murine HSCs induces remarkable ex vivo and in vivo cell expansion without compromising their differentiation or inducing leukemic transformation (Sauvageau et al., 1995, Antonchuk et al., 2002). Similar effects were obtained using recombinant TAT-HOXB4 protein (Krosl et al., 2003). In some reports, HoxB4 and negative selection marker MGMT were used together to further increase percentage of modified HSCs (Chinnasamy et al., 2005). However, attempts to use HoxB4 for positive selection of HSCs in larger animals were much less successful, with a major expansion of short-term repopulating cells only (Zhang X et al., 2006). Besides, a significant number of leukemia occurrences apparently related to unregulated expression of HoxB4 were observed in these animals (Zhang X et al., 2008).

cycle but incidentally failed to maintain their stem cell status (Bunting et al., 1999). This situation seems to have been ameliorated after introduction of transduction protocols that rely on the use of serum-free media that lack factors inducing SC differentiation (Mostoslavsky et al., 2005) and novel growth factors that better preserve cell stemness (Zhang C et al., 2008). It remains yet to see whether these improvements are sufficient to

As current efficiency of transduction of human LTR HSCs with viral vectors appears to be quite low and there are no clinically proven protocols for expansion of these cells ex vivo, the most promising solution at hand to this problem is an in vivo selection of modified cells after their transduction and re-transplantation back to a patient. Conceptually, one might distinguish negative and positive in vivo selection strategies. The first one can be defined as a strategy that is aimed at elimination of stem and progenitor cells that do not bear integrated functional constructs. Positive selection implies a strategy that does not target the construct-negative stem cells but rather provides selective survival and growth advantage to the cells that bear the inserted construct. The negative selection gains presently much of attention and seems to be better poised for a clinical advancement in the near future. Arguably, the most promising and advanced variant of negative selection is based on the use of O6-MGMT as a selection marker and various alkylating compounds as selection agents (Davis et al., 2000, Ragg et al., 2000). Using this approach and multiple rounds of selection in vivo, overall peripheral blood chimerism has been driven in mice and larger animal models to levels higher than 75%. However, the clinical applicability of this technique is as yet unclear, as recent experiments performed by two research teams with non-human primates using MGMT-mediated selection produced rather conflicting results. One team demonstrated successful implementation of this strategy in monkeys, although with selection efficiencies and chimerism rates highly variable between individual animals (Beard et al., 2010), whereas another team reported a rather

significantly increase the efficiency of HSC gene therapy in clinical settings.

**3. Selection of genetically modified HSCs in vivo: Negative selection** 

negligible increase in chimerism rates upon selection in vivo (Larochelle et al., 2009). Various implementations of negative selection strategy are listed in the Table 1.

**4. Selection of genetically modified HSCs in vivo: Positive selection** 

expression of HoxB4 were observed in these animals (Zhang X et al., 2008).

Ongoing studies of the mechanisms controlling HSC self-maintenance and commitment continue to identify novel factors that bring about HSC expansion in vivo when over-expressed. A less than exhaustive set of these factors is listed in the Table 2. Arguably, the most extensively studied gene with such properties is the homeobox transcription factor HoxB4. Forced expression of HoxB4 in murine HSCs induces remarkable ex vivo and in vivo cell expansion without compromising their differentiation or inducing leukemic transformation (Sauvageau et al., 1995, Antonchuk et al., 2002). Similar effects were obtained using recombinant TAT-HOXB4 protein (Krosl et al., 2003). In some reports, HoxB4 and negative selection marker MGMT were used together to further increase percentage of modified HSCs (Chinnasamy et al., 2005). However, attempts to use HoxB4 for positive selection of HSCs in larger animals were much less successful, with a major expansion of short-term repopulating cells only (Zhang X et al., 2006). Besides, a significant number of leukemia occurrences apparently related to unregulated


Table 1. Strategies for negative selection of genetically modified HSC

Some other members of the HOX family, either alone or fused with specific cellular partners, are also able to induce expansion of hematopoietic progenitors in mice. Of particular importance is a fusion gene NUP98-HoxA10, which has a remarkable ability of multi-log expansion of murine repopulating cells ex vivo, exceeding that of HoxB4 (Ohta et al., 2007; Watts et al., 2011).

Recently, the powerful effect of overexpression of early acting transcription factor SALL4 on ex vivo expansion of human hematopoietic cells capable of long-term repopulation of NOD/SCID mice was demonstrated (Aguila et al., 2011). Significant ex vivo expansion could be also achieved using recombinant TAT-SALL4B protein.

There are at least a dozen of other genes that, when overexpressed, induce significant expansion of HSCs in mice in vivo. One of the most interesting groups of such factors are epigenetic regulators. Of particular interest is Bmi1, a member of Polycomb group, which is involved in regulation of mantenance of various adult stem cell types. Inactivation of Bmi1

Gene Therapy of Hematopoietic and Immune Systems: Current State and Perspectives 445

Gene Observed effects References

strongly increased in vivo repopulation. Zhu et al., 2005

pool and develop a myeloproliferative disease. Mori et al., 2008

long-term NOD/SCID repopulating cells. Aguila et al., 2011

increased numbers of HSCs. Liang et al., 2007

Antonchuk et al., 2002; Sauvageau et

Ohta et al., 2007; Watts et al., 2011

Iwama et al., 2004; Rizo et al., 2008

Kamminga et al.,

Cavazzana-Calvo et al., 2010; Ikeda et al., 2011; Wang et

Métais & Dunbar, 2008; Sellers et al.,

Du et al., 2005; Ott

Guo et al., 2010; Ooi et al., 2010

Bersenev et al.,

Mullican et al.,

Heath et al., 2004; Zhang P et al. 2004;

2008

2007

2006

2011

al., 2010

2010

et al., 2006

al., 1995

Overexpression of HoxB4 induces significant ex vivo and in vivo expansion of murine long-term

Enforced expression of NUP98-HOXA10 fusion protein results in significant expansion of murine repopulating cells ex vivo exceeding that of HoxB4.

Enforced expression of Bmi1 leads to striking ex vivo expansion of multipotential progenitors and marked augmentation of HSC repopulating

Jhdm1b Same as above. Konuma et al.,

competitive repopulation and serial transplantation.

Frequently found in the vicinity of integrated constructs in gene therapy trials; HMGA2 expressing cells have growth advantage in

Forced expression of miR-125a was capable of increasing the number of HSCs cells several-fold. Overexpression of miR-125b enhances HSC function, as judged by serial transplantation.

Mice with inactivation of both Nor-1 and Nur77 have abnormal expansion of HSCs and myeloid progenitors

C/EBP alpha-deficient HSCs are hyperproliferative and have enhanced competitive repopulating

and develop lethal acute myeloid leukemia.

repopulating HSCs.

capacity in vivo.

repopulating HSCs.

NF-Ya Murine HSCs overexpressing NF-Ya demonstrate

Ezh2 Overexpression prevents exhaustion of long-term

Jab1 Mice with Jab1 overexpression have expanded HSC

Evi-1 Frequently found in the vicinity of integrated constructs in gene therapy trials.

PRDM16 Frequently found in the vicinity of integrated constructs in gene therapy trials.

Sall4 Enforced expression results in ex vivo expansion of

Lnk Mice with Lnk inactivation have an expanded HSC pool with enhanced self-renewal.

Latexin Mouse strains expressing lower latexin levels have

activity.

Table 2. Genes affecting in vivo expansion of HSCs

HOXB4

NUP98- HOXA10

Bmi1

Fbxl10/

HMGA2

MicroRNAs miR-125a, miR-

Nur77/NR4A1

125b

& Nor-1/NR4A3

C/EBP

leads to defect in HSC self-renewal (Park et al., 2003), whereas its enforced expression results a striking ex vivo expansion of multipotential progenitors and marked augmentation of HSC repopulating capacity in vivo (Iwama et al., 2004). In addition, enforced expression of Bmi1 in human CD34-positive cells leads to the ex vivo expansion of NOD/SCID repopulating cells (Rizo et al., 2008). Another Polycomb group gene that potentially could be used for positive selection is Ezh2; upon overexpression, it prevents HSC exhaustion (Kamminga et al., 2006). Forced expression of yet another epigenetic regulator, histone demethylase Fbxl10/Jhdm1b in HSCs abolishes exhaustion of the LTR HSCs following serial transplantation. This property of Ezh2 and Fbxl10/Jhdm1b makes them especially appropriate for schemes combining positive and negative selection since the latter one places very significant stress on hematopoietic system.

Another group of genes that might be used for positive selection are those that are frequently activated in predominant hematopoietic cell clones arising after retro- or lentiviral transduction, and are likely therefore to act as factors inducing in vivo expansion of these clones. The most prominent among such genes are MDS1/Evi-1 (Sellers et al., 2010; Métais & Dunbar, 2008), PRDM16 (Du et al., 2005; Ott et al., 2006) HMGA2 (Wang et al., 2010; Cavazzana-Calvo et al., 2010) and LMO2 (McCormack et al., 2003; McCormack et al., 2010). As a note of caution, forced expression of these genes may produce undesired effects; for example, expression of Evi-1 was reported to be associated with chromosomal instability (Stein et al., 2010).

In addition to protein factors, micro RNAs also have effect on HSC function and population size. In particular, miR-125a and miR-125b were shown to increase number of HSCs in vivo or enhance their repopulation capacity (Guo et al., 2010; Ooi et al., 2010).

Having focused on genes that expand stem cell population, one should not overlook another group of genes that exert an opposite effect, namely negative influence on HSC pool size. Thanks to RNA interference technology, suppression of gene expression in various cell types nowadays is nearly as simple as overexpression. If gene knockout or knockdown results in expansion of stem cell population, this property may potentially be used for positive selection. Among genes of interest in this respect are C/EBP alpha, Lnk and Nur77, to name a few. C/EBP alpha-deficient hematopoietic stem cells (HSCs) are hyperproliferative, have increased expression of Bmi-1 and enhanced competitive repopulating activity (Zhang et al. 2004; Heath et al., 2004). Inactivation of Lnk, inhibitory adaptor protein, leads to an expanded HSC pool with enhanced self-renewal (Bersenev et al., 2008). Mice with inactivation of both Nor-1 and Nur77 have abnormal expansion of HSCs and myeloid progenitors and develop lethal acute myeloid leukemia (AML).

Regardless of what gene is being used for positive selection, it is clear that its constitutive expression would eliminate one or more of the negative growth controls imposed on HSCs by organism, and thus increase risks of neoplastic transformation. Therefore, any clinically acceptable protocol for gene therapy using positive selection of transduced HSCs should be based on transient, tightly regulated gene expression. Given that positive selection, if correctly implemented, promises to provide significant advantages over negative selection schemes, further research into creation of robustly regulated expression systems for positive selection in HSCs seem to be fully warranted.

leads to defect in HSC self-renewal (Park et al., 2003), whereas its enforced expression results a striking ex vivo expansion of multipotential progenitors and marked augmentation of HSC repopulating capacity in vivo (Iwama et al., 2004). In addition, enforced expression of Bmi1 in human CD34-positive cells leads to the ex vivo expansion of NOD/SCID repopulating cells (Rizo et al., 2008). Another Polycomb group gene that potentially could be used for positive selection is Ezh2; upon overexpression, it prevents HSC exhaustion (Kamminga et al., 2006). Forced expression of yet another epigenetic regulator, histone demethylase Fbxl10/Jhdm1b in HSCs abolishes exhaustion of the LTR HSCs following serial transplantation. This property of Ezh2 and Fbxl10/Jhdm1b makes them especially appropriate for schemes combining positive and negative selection since the latter one

Another group of genes that might be used for positive selection are those that are frequently activated in predominant hematopoietic cell clones arising after retro- or lentiviral transduction, and are likely therefore to act as factors inducing in vivo expansion of these clones. The most prominent among such genes are MDS1/Evi-1 (Sellers et al., 2010; Métais & Dunbar, 2008), PRDM16 (Du et al., 2005; Ott et al., 2006) HMGA2 (Wang et al., 2010; Cavazzana-Calvo et al., 2010) and LMO2 (McCormack et al., 2003; McCormack et al., 2010). As a note of caution, forced expression of these genes may produce undesired effects; for example, expression of Evi-1 was reported to be associated with chromosomal instability

In addition to protein factors, micro RNAs also have effect on HSC function and population size. In particular, miR-125a and miR-125b were shown to increase number of HSCs in vivo

Having focused on genes that expand stem cell population, one should not overlook another group of genes that exert an opposite effect, namely negative influence on HSC pool size. Thanks to RNA interference technology, suppression of gene expression in various cell types nowadays is nearly as simple as overexpression. If gene knockout or knockdown results in expansion of stem cell population, this property may potentially be used for positive selection. Among genes of interest in this respect are C/EBP alpha, Lnk and Nur77, to name a few. C/EBP alpha-deficient hematopoietic stem cells (HSCs) are hyperproliferative, have increased expression of Bmi-1 and enhanced competitive repopulating activity (Zhang et al. 2004; Heath et al., 2004). Inactivation of Lnk, inhibitory adaptor protein, leads to an expanded HSC pool with enhanced self-renewal (Bersenev et al., 2008). Mice with inactivation of both Nor-1 and Nur77 have abnormal expansion of HSCs and myeloid progenitors and develop lethal acute myeloid leukemia

Regardless of what gene is being used for positive selection, it is clear that its constitutive expression would eliminate one or more of the negative growth controls imposed on HSCs by organism, and thus increase risks of neoplastic transformation. Therefore, any clinically acceptable protocol for gene therapy using positive selection of transduced HSCs should be based on transient, tightly regulated gene expression. Given that positive selection, if correctly implemented, promises to provide significant advantages over negative selection schemes, further research into creation of robustly regulated expression systems for positive

or enhance their repopulation capacity (Guo et al., 2010; Ooi et al., 2010).

places very significant stress on hematopoietic system.

(Stein et al., 2010).

(AML).

selection in HSCs seem to be fully warranted.


Table 2. Genes affecting in vivo expansion of HSCs

Gene Therapy of Hematopoietic and Immune Systems: Current State and Perspectives 447

Yet another line of evidence indicates that activation of beta-catenin in the niche components rather than in HSCs may produce support of LTR cells ex vivo (Nemeth et al., 2009). Currently, there is little doubt that Wnt signaling plays important role in HSC biology, but the issue is apparently more complex than was implied by initial publications

Other embryonic signaling pathways also might be exploited in HSC culture. Morphogens of the hedgehog family, namely Sonic and Indian hedgehogs, are able to support ex vivo expansion of human NOD/SCID repopulating cells (Bhardwaj et al., 2001; Kobune et al., 2004), despite the fact that in vivo Hedgehog signaling seems to not be necessary for adult murine hematopoietic stem cell function (Hofmann et al., 2009). BMP4, a member of BMP superfamily, is a critical component of the hematopoietic niche that regulates both HSC number and function (Goldman et al., 2009), and is able to expand NOD/SCID-repopulating

In addition to the use of secreted proteins to for ex vivo HSC culture, one apparent trend of the last years is the application of low-molecular weight chemicals, in particular agonists or inhibitors of particular intracellular signaling pathways, for ex vivo culture. Thus, specific inhibitor of p38 kinase induces self-renewal and ex vivo expansion of HSCs as shown by the in vitro cobblestone area forming cell assay and serial transplantation (Wang et al., 2011). GSK-3β inhibitors, which stimulate Wnt signaling, were shown to promote engraftment of cultured HSCs (Ko et al., 2011; Trowbridge et al., 2006). Of significant clinical interest is the finding that ex vivo treatment with stabilized prostaglandin E2 enhances frequency of both hematopoietic progenitors and long-term repopulating HSCs present as analyzed by competitive transplantation (North et al., 2007). According to other data, only the short-term

The initial studies demonstrating substantial degree of expansion of HSCs ex vivo relied the use of stromal cells as feeder layers (Moore et al., 1997). Based on the substantial progress in identification of HSC niches in bone marrow, there is currently a revival of interest in development of protocols for co-culture of HSC with stromal cell layers (Chou & Lodish, 2010; De Toni et al., 2011). These stromal cells produce a range of factors that significantly improve the maintenance and expansion of HSCs in culture, most likely by mimicking more or less successfully niche conditions. Very prominent components of the HSC niche are cell surface proteins, in particular cell adhesion molecules. The importance of cell-cell interactions was highlighted by the study by Wagner et al., 2007, indicating that maintenance of primitive hematopoietic progenitors by stromal lines is associated with expression of cell adhesion proteins rather than with secretory profiles of these lines. In particular, N-cadherin was shown to be an important component of the osteoblastic HSC niche (Zhang et al., 2003). However, importance of N-cadherin for HSC-niche interactions was later questioned (Kiel et al., 2007), thus rising substantial controversy. In an elegant in vitro study Lutolf et al. (2009) have shown that N-cadherin, as well as Wnt3a, are the only proteins among those tested that were capable of supporting self-renewal divisions of HSCs in vitro. N-cadherin expression was also shown to be important for maintenance of long-term repopulating cells in culture (Hosokawa et al., 2010). Ability of stromal cell line FMS/PA6-P to support primitive murine hematopoietic cells was found to depend critically on N-CAM expression (Wang et al., 2005). Yet another cell adhesion protein, namely mKirre, plays a prominent role in hematopoietic supportive capacity

repopulating HSCs are expanded by this treatment, though (Frisch et al., 2009).

and remains highly controversial.

cells in culture (Hutton et al., 2006).

of OP9 stromal cells (Ueno et al., 2003).
