Aberrant stem cell and developmental programs in pediatric leukemia

Hematopoiesis is a finely orchestrated process, whereby hematopoietic stem cells give rise to all mature blood cells. Crucially, they maintain the ability to self-renew and/or differentiate to replenish downstream progeny. This process starts at an embryonic stage and continues throughout the human lifespan. Blood cancers such as leukemia occur when normal hematopoiesis is disrupted, leading to uncontrolled proliferation and a block in differentiation of progenitors of a particular lineage (myeloid or lymphoid). Although normal stem cell programs are crucial for tissue homeostasis, these can be co-opted in many cancers, including leukemia. Myeloid or lymphoid leukemias often display stem cell-like properties that not only allow proliferation and survival of leukemic blasts but also enable them to escape treatments currently employed to treat patients. In addition, some leukemias, especially in children, have a fetal stem cell profile, which may reflect the developmental origins of the disease. Aberrant fetal stem cell programs necessary for leukemia maintenance are particularly attractive therapeutic targets. Understanding how hijacked stem cell programs lead to aberrant gene expression in place and time, and drive the biology of leukemia, will help us develop the best treatment strategies for patients.


Introduction
Stem cells perform a complex balancing act between self-renewal and differentiation throughout ontogeny.To perform these functions, stem cells proliferate rapidly and repair DNA damage.However, these "stemness" properties present a vulnerability as, if hijacked, they provide cancer cells with the pathways required for growth and survival (Taipale and Beachy, 2001;Hanahan and Weinberg, 2011).

Stem cell programs in leukemia
In leukemia, stem cell programs may be inappropriately reactivated or retained and/or co-opted from fetal development.This may be a consequence of some leukemias originating in utero, especially in children (Greaves, 2005).
In this review, we discuss stem cell programs that are aberrantly active in the wrong cellular context ("place") or stage of ontogeny ("time") in pediatric leukemia and their potential applications in developing targeted therapies (Figure 1).

Stem cell programs in the wrong place
Aberrant stem cell genes in leukemia Leukemic blasts may exhibit stem cell properties, conferring a more aggressive phenotype.Several stem cell genes are important for leukemia biology (Table 1), and some key examples are discussed below.

Homeobox genes
Homeobox genes form a group of transcription factors (TFs) with 235 functional genes in humans (Holland et al., 2007).

CDX2
CDX2 of the ParaHox family of homeobox genes (Brooke et al., 1998) is not normally expressed in hematopoietic cells and inhibits the hematopoietic potential of murine embryonic stem cells (ESCs) (McKinney-Freeman et al., 2008).

IRX genes
The Iroquois genes (IRX1-6) belong to the TALE group of homeobox genes.IRX1 is required for normal development of kidney and neural tissues (Alarcón et al., 2008;Freese et al., TABLE 1 (Continued) Summary of developmental/stem cell genes involved in the pathophysiology of pediatric leukemia.

Normal function Aberrant expression and function
Therapeutic potential

Key reference
Adhesion proteins Frontiers in Cell and Developmental Biology frontiersin.org2014) but not expressed in most hematopoietic precursors (Nagel et al., 2022).IRX3 and IRX5 are not expressed in hematopoiesis.

Aberrant stem cell gene expression in ontogeny
Fetal stem cell programs in pediatric leukemia Pediatric leukemias [and solid malignancies such as neuroblastoma (Molenaar et al., 2012)] may exhibit properties of fetal HSCs.This may represent a fetal cell of origin or indicate reactivation of fetal programs (Monk and Holding, 2001;Sharma et al., 2022;Solé et al., 2022).Either way, the inappropriate expression of fetal genes is important for cancer biology.Leukemic blasts, especially LSCs (Somervaille et al., 2009), show fetal-specific gene expression profiles (Wu et al., 2015;Helsmoortel et al., 2016b;Elcheva et al., 2020;Bai et al., 2021;Tran et al., 2021).Down syndrome (DS)-associated leukemias and juvenile myelomonocytic leukemia (JMML) are two examples resulting from perturbation of fetal HSPCs.
Children with DS have an increased risk of AML and ALL (Hasle et al., 2000).Mutations in the megakaryocyte-erythroid transcription factor GATA1 in fetal life lead to transient abnormal myelopoiesis (TAM) in the fetal/neonatal period (Roberts et al., 2013;Wagenblast et al., 2021).Additional mutations, most commonly in the cohesin complex genes (Labuhn et al., 2019), are required for myeloid leukemia of DS (Roberts and Izraeli, 2014).DS-ALL also likely stems from perturbed lymphopoiesis, which begins in utero, and is characterized by CRLF2/TSLPR overexpression in 50% and JAK2 mutations in 20% (Li et al., 2023).Key Hsa21 genes important for leukemogenesis are listed in Table 1.
Fetal oncogenes relevant to stem cell activity and implicated in pediatric leukemia act via a diverse range of mechanisms and are discussed below (Figure 1).

Fetal genes with transcription factor activity implicated in pediatric leukemia
SCL/TAL1 was originally identified as overexpressed in T-ALL (Ferrando et al., 2002).Ablation of the gene causes embryonic death (Porcher et al., 2017), but it is dispensable for adult HSCs (Mikkola et al., 2003).

Fetal genes important for posttranscriptional regulation
The bulk of post-transcriptional control is exerted by RNAbinding proteins (RBPs).Some RBPs such as LIN28B and insulinlike growth factor 2 mRNA-binding proteins (IGF2BP1/IGF2BP3) have a fetal expression pattern, a role in stem cell biology (Copley and Eaves, 2013;Degrauwe et al., 2016;Zhang et al., 2016) and pediatric leukemia.
A meta-analysis showed that 7.5% of pediatric leukemias express LIN28B (Helsmoortel et al., 2016b).Aberrant LIN28B expression defines a poor prognosis subgroup in JMML (Helsmoortel et al., 2016a), where H19, a fetal oncogene (Matouk et al., 2014), is stabilized in the presence of LIN28B (Helsmoortel et al., 2016b).AML in children <3 years has higher levels of LIN28B (and IGF2BP1/3) expression than in children >3 years (Bolouri et al., 2021).Although LIN28B has predominantly been reported to have a pro-leukemic role in AML (Zhou et al., 2017), one study on a murine KMT2A::MLLT3 AML model suggests that LIN28B abrogates perinatal leukemia development (Eldeeb et al., 2023).Given >50% of human neonatal leukemias are of myeloid lineage, these findings seem counterintuitive, although it is possible that neonatal AML arises from LIN28B negative progenitors.Given the role of LIN28B in fetal B-lymphopoiesis, it may also be important for ALL initiation or maintenance.

IGF2BP1 and IGF2BP3
IGF2BP1 and IGF2BP3 are important for fetal organogenesis and are expressed in FL HSCs, but not in adult HSCs (Wang D. et al., 2022).Induction of IGF2BP3 in adult HSCs induces a fetal-type output (Palanichamy et al., 2016;Wang et al., 2019).
IGF2BP1 and IGF2BP3 have been linked to leukemia, as well as solid malignancies, and are often co-expressed with LIN28B (Elcheva et al., 2020;Tran et al., 2021).The mechanism of action for IGF2BP3 in oncogenesis is segregation of mRNA transcripts from the cytoplasmic RNA-induced silencing complex, including the let-7 miRNA family.

Fetal genes important for epigenetic regulation
HMGA1 and 2 HMGA1 and HMGA2 are fetal oncogenes affecting epigenetic regulation.The HMGA family encodes proteins with AT hooks which interact with DNA to alter the chromatin architecture.These genes have much lower expression in adult tissues than in the fetal counterparts (Kumar et al., 2019;Roy et al., 2021), and HMGA1 can promote a pluripotent state (Shah et al., 2012).HMGA2 is expressed mainly in fetal HSC/MPP and influences both differentiation and proliferation of stem cells (Battista et al., 2003;Li et al., 2007;Copley et al., 2013), as well as promoting long-term in vivo reconstitution by cord blood CD34 + cells (Kumar et al., 2019).
Reactivation of HMGA1 and HMGA2 has been demonstrated in a wide range of malignancies (Huso and Resar, 2014;Mansoori et al., 2021) including leukemia (Efanov et al., 2014).HMGA1 expression has been linked to risk of relapse in pediatric B-ALL (Roy et al., 2013).In pediatric and adult AML, high expression of HMGA2 is linked to poor prognosis, and knockdown of the gene has induced differentiation in primary blasts (Marquis et al., 2018;Tan et al., 2018).HMGA2 induces T-ALL in a Eμ-HMGA2 transgenic mouse (Efanov et al., 2014).

microRNAs in leukemia
Aberrant expression of microRNAs (miRNAs) specific to fetal life and stem cell compartment (O'Connell et al., 2010) is implicated in pediatric leukemia (Grobbelaar and Ford, 2019;Gaur et al., 2020) (Table1).Pediatric AML can be risk-stratified by a 24-miRNA signature (Esperanza-Cebollada et al., 2023).Eight of these have target genes within the pLSC6 signature and includes let-7 miRNAs (known repressors of oncogenes), with lower let-7g/let-7i expression in high-risk AML.One of the pLSC6 genes (FAM30A) is an lncRNA.Signatures based on lncRNA differentiate pediatric leukemia subtypes, but do not inform prognosis (Buono et al., 2022).

Targeting aberrant stem cell programs in leukemia for therapy
The inappropriate expression of stem cell genes, while conferring survival advantage to leukemic cells, can also render them dependent on specific proteins or pathways, and thus vulnerable to targeted disruption.Fetal stem cell genes are the most attractive targets as they are not expressed in healthy postnatal tissues, ameliorating concerns about off-target effects.Genes expressed in leukemic and healthy postnatal stem cells present more of a challenge.However, excessive leukemic reliance on the aberrant pathway, the so-called "oncogenic addiction," can generate a therapeutic window, whereby leukemic cells can be killed while sparing normal stem cells.Potential targeting strategies are summarized in Table 1.Specific approaches relating to stem cell genes discussed in this review are explored below.

Small-molecule inhibitors
Many stem cell genes code for TFs or other DNA-binding proteins, considered "undruggable," owing to their intrinsically disordered nature.Recent improvements in screening methods have identified HMGA2-binding compounds, including the antimicrobials sumarin and ciclopirox (Huang et al., 2019;Su et al., 2020) and MEIS1/2 inhibitors (Turan et al., 2020).
An alternative approach employs small molecules that bind the minor groove of the TF cognate sequence.DNA binders of this type can inhibit HOXA9 (Depauw et al., 2019), with in vitro activity against HOXA9-dependent cells lines (Sonoda et al., 2021).Similarly, netropsin and trabectedin demonstrate antitumor activity in HMGA2+ neoplasia.Treatment with both drugs shows a synergistic anti-proliferative effect in infant ALL cell lines (Wu et al., 2015).Other approaches to target HMGA2 include targeting downstream pathways such as G2M transition (Moison et al., 2022) and PI3K/Akt/mTOR (Tan et al., 2016).
Direct inhibition of CDX2 has not yet been possible; however, the observation that PPARγ signaling restores KLF4 expression offers a potential therapeutic route to partially opposing CDX2 activity.PPARγ agonists are toxic to CDX2 overexpressing leukemia cell lines in vitro (Faber et al., 2013;Esmaeili et al., 2021).

Targeting RNA-binding proteins
Like DNA-binding proteins, small-molecule inhibition of RBPs is difficult, although recent high-throughput approaches have generated candidates (Wu, 2020).The most promising LIN28(A/ B) inhibitor is C1632, which targets LIN28B + cell lines both by disruption of LIN28B-let-7 interaction (Franses et al., 2020;Zhang Q. et al., 2023;Shahab et al., 2023) and in Ewing sarcoma by disruption of the interaction between EWS-FLI1 mRNA and LIN28B (Keskin et al., 2020).Other molecules such as LI71 bind the cold shock domain and have efficacy against LIN28B + cancer cell lines (Wang L. et al., 2018).
Another potential strategy to boost let-7 miRNA expression, thus inhibiting several oncogenes (Cinkornpumin et al., 2017), has yet to be applied in leukemia.

Immune effector cell therapy
Stem cell markers as targets for immunotherapy: The cell surface marker PROM1 is a highly attractive target for immunotherapy.A CD19/CD133 tandem CAR-T (Li et al., 2018) and CD19/ 133 bispecific CAR-iNKT (Ren et al., 2023) show efficacy in vivo against KMT2A-r cell lines.However, valid concerns about stem cell toxicity when targeting CD133 in patients have been raised (Bueno et al., 2019).Preclinical testing and early-phase trials using CD133-CAR-T in solid malignancies (Wang et al., 2018b;Dai et al., 2020;Vora et al., 2020) revealed no BM aplasia and only transient, reversible hematological toxicities.Longer-term follow-up and assessment will be required to confirm the safety of CD133 targeting.CAR-T directed against NG2 has shown promise in mobilizing leukemic blasts and rendering them more sensitive to chemotherapy in mouse models (Lopez-Millan et al., 2019).Anti-CD117 CAR-T therapy shows preclinical efficacy (Myburgh et al., 2020) but also eliminates healthy HSCs, necessitating novel approaches such as terminating "safety switches" (Magnani et al., 2023).Both CAR-T (Wang et al., 2018c) and CAR-NK (Mansour et al., 2023) cells have been used to target FLT3.Anti-CD123 CAR-T therapy has been used in earlystage clinical trials, appearing safe and potentially effective (Yao et al., 2019;Wermke et al., 2021).

Discussion
Stem cells possess unique properties allowing the expansion, selfrenewal, and differentiation required for tissue homeostasis.These programs are frequently co-opted by leukemias, where they provide growth and survival advantages.Although there is renewed interest in reprogramming adult HSCs to become more "fetal-like", the potential of fetal stem cell genes to also promote oncogenesis must be considered.
Understanding stem cell programs in leukemia, including oncofetal genes, is vital to disentangling the biology of leukemias, including treatment resistance/relapse, and identifying mechanisms vulnerable to novel targeted therapies.

FIGURE 1
FIGURE 1 Role of stem cell genes in leukemia cell survival.Potential therapeutic strategies targeting aberrant stem cell-associated pathways are shown in red.TF = transcription factor, ZF = zinc finger.

TABLE 1 (
Continued) Summary of developmental/stem cell genes involved in the pathophysiology of pediatric leukemia.