Enhancing human cord blood hematopoietic stem cell engraftment by targeting nuclear hormone receptors

Abstract

Purpose of review

Allogeneic hematopoietic cell transplantation (HCT) is a life-saving therapy for hematological and non-hematological diseases. Cord blood (CB) is a source of transplantable hematopoietic stem cells (HSCs), but limited numbers of HSCs in single CB units, which may cause delayed neutrophil, platelet, and immune cell reconstitution, is a major problem for efficient transplantation. Ex vivo expansion and enhanced homing of CB HSC may overcome this disadvantage and improve its long-term engraftment. Here, we discuss the role of nuclear hormone receptors (NRs) signaling in human CB HSC engraftment.

Recent findings

Antagonizing Retinoid acid receptor (RAR) signaling promotes human HSC expansion and increases myeloid cell production. CB CD34⁺ cells expanded by SR1 promotes efficient myeloid recovery after transplantation compared with control groups, and leads to successful engraftment. Short-term treatment of glucocorticoids enhances homing and long-term engraftment of human HSCs and HPCs in NSG mice. PPARγ antagonism expands human HSCs and HPCs by preventing differentiation and enhancing glucose metabolism. These findings demonstrate that NR signaling components might be promising targets for improving human CB HCT.

Summary

Better understanding of molecular mechanisms underlying human HSC expansion and homing mediated by NR signaling pathways will facilitate enhanced HCT efficacy.

INTRODUCTION

Homeostasis of hematopoiesis is maintained by intrinsic and extrinsic signals by hematopoietic stem (HSC) and progenitor (HPC) cells, rare immature cell populations residing in microenviromental niches [1–3]. HSC self-renew, or differentiate through HPC into all mature blood cells throughout [4]. Human HSC are major contributors for HCT [5,6]. Once infused into peripheral blood of recipients, HSC and HPC migrate and engraft in bone marrow (BM) niches, where a new hematopoietic system builds up [7,8]. Three major sources of HSC for HCT are: BM, peripheral blood and CB [9,10]. Compared with the other sources, CB has advantages, including ready availability, less strict HLA matching requirements, lower incidence of graft vs. host disease and virus infection, and low probability of relapse among patients with minimal residual disease [11–13]. However, delayed hematopoietic recovery and immune reconstitution due to limited numbers of HSCs are major limiting factors preventing the wider use of HCT based cell therapy. Progress to improve CB HSC engraftment and reconstitution includes mitigating effects of Extra Physiologic Oxygen Shock/Stress (EPHOSS) during cell collection [14], ex vivo expansion of HSC [15–17*] and enhancing HSC homing efficiency [18].

Nuclear hormone receptors (NR) are transcription factors which regulate expression of target genes by binding to specific response elements in the transcription units [19,20]. NR, consisting of 48 members in humans, are activated by steroid hormones such as glucocorticoids and estrogen, or non-steroid hormones such as retinoic acid, vitamin D, and thyroid hormones. NR members share a similar structure, composed of a highly variable A/B transactivation domain, a conserved DNA binding domain (C domain), a nuclear localization signal containing D domain and a relatively conserved ligand binding E domain [19,20]. Upon ligand binding, the conformation of NR changes and a repressor complex such as NCoR and SMRT corepressor disassociate from NR, followed by recruitment of transactivation complexes with histone acetyltransferase (HAT) activity and other transcription machinery, thus facilitating transcription of target genes[21]. NR are involved in physiological processes including: proliferation [22], differentiation [23,24], development [25] and metabolism [26,27]. To treat diseases caused by dysfunction of NR, synthetic agonists and antagonists of NR were developed to modulate its activity [28]. These diversified synthetic ligands help explore novel biological functions of NR in humans.

Here, we summarize recent progress regarding enhancing human HSC and HPC engraftment by modulating NR signaling.

RETINOID ACID SIGNALING AND HUMAN HSC EXPANSION

Retinoid acid receptor (RAR) serves as the nuclear receptor for all-trans retinoid acid (ATRA), the biologically active form of retinol (vitamin A) [29]. There are three different subtypes of RAR (α, β, γ), and each is alternatively spliced into at least two distinct isoforms [30]. Although RAR is highly conserved between mouse and human, functioning of RAR signaling in murine and human HSC homeostasis is quite different. RARα knockout mice show no abnormal phenotype of HSCs and HPCs, while HSC numbers are significantly reduced in RARγ knockout mouse BM [31]. BM HPC numbers are increased in RARγ knockout mice compared with wild type animals, suggesting that RARγ might regulate differentiation of their HSCs. Meanwhile, ATRA ex vivo treatment enhances repopulating capability of murine BM Lineage⁻Sca-1⁺c-Kit⁺ (LSK) cells, effects dependent on RARγ [31]. Physiological ATRA is derived from retinol, which is not synthesized by animals but is obtained from food (vegetables, fruits, animal livers) [32], so HSC homeostasis can be affected by availability of retinol in food. ATRA treatment protects dormant HSCs from stress induced activation by restricting protein synthesis and reactive oxygen species (ROS) generation [33**]. Numbers of dormant HSCs in retinol free diet fed mice, in which the retinoid acid reservoir is depleted, is significantly reduced [33**].

RAR usually forms a heterodimer with retinoid X receptor (RXR) [34]. RXRα signaling negatively regulates granulocyte development from HSCs and HPCs in humans and mice [35], while in vivo hematopoiesis in RXRα knock out mice is normal [36]. Since there are two other isoforms of RXR (RXRβ, RXRγ) [30], normal hematopoiesis of RXRα knock out mice might be due to functional redundancy of different isoforms. It would be interesting to explore roles of RXR signaling in HSC self-renewal and maintenance using distinct RXR isoform double or triple knock out mice, since little is known of RXR in mouse and human HSC homeostasis.

Although activation of RAR signaling positively regulates development and function of mouse embryonic and adult HSC [31, 37, 38], human CB HSC self-renewal appears negatively modulated by the RAR pathway. Aldehyde dehydrogenase (ALDH) is a key enzyme in the RA synthesis pathway in vivo, and is characterized as a marker of HSCs and other tissue stem cells [39–41]. There are many ALDH isoforms. ALDH1 is considered the major isoform in HSCs and HPCs [42,43]. Treatment of human BM and CB CD34⁺CD38⁻ cells with diethylaminobenzaldehyde (DEAB), an ALDH inhibitor, enhances expansion and maintenance of HSCs and HPCs, as well as the in vivo frequency of SCID-repopulating cells (SRCs) in immune-compromised NOD/SCID mice [43]. Enhanced expansion of HSCs by DEAB can be reversed by RA induced differentiation [43]. Although HSCs themselves have the capability to convert retinol to RA [44], it would be interesting to determine whether inhibition of ALDH in HSCs or HPCs leads to down-regulation of RA signaling, and thus stimulates human HSC expansion and maintenance. BM stromal cells secrete CYP26 in the serum to inactivate and prevent RA signaling, and thus regulate human HSC self-renewal [45], suggesting that modulating RA signaling by targeting microenviromental niches represents a promising way to enhance ex vivo human HSC expansion. Treatment of CD34⁺CD38⁻ cells with LG1506, a selective RXR modulator, enhances human HSC and HPC expansion and long-term engraftment in NOD/SCID mice [46]. Mechanistically, once binding with LG1506, the RXR-RAR complex recruits silencing mediator of retinoic acid and thyroid hormone receptor (SMRT) corepressors to promoters of target genes such as CD38, thus inhibiting differentiation of HSCs [46]. Recently, it was found that RARα antagonist increases myeloid cell production and promotes expansion of human HSCs, while RARγ antagonist had no effect due to different sensitivity to ATRA [47]. RXR forms heterodimers with other NRs, including liver X receptor (LXR), thyroid hormone receptor (THR), vitamin D receptor (VDR), peroxisome proliferator-activated receptors (PPARs). Future studies should characterize roles of these RXR heterodimers in human HSC maintenance and engraftment.

GLUCOCORTICOID SIGNALING IN HOMING AND ENGRAFTMENT OF HUMAN HSCS

Glucocorticoid (GC) receptor (GR) is a member of the steroid nuclear hormone receptor subfamily, which also includes mineralocorticoid receptor (MR), estrogen receptor (ER), androgen receptor (AR), and progestogen receptor (PR) [19]. Unliganded GR resides in the cytoplasm and is complexed with HSP90 and HSP70 [48]. Upon binding with GC, GR translocates into the nucleus and binds to GC response elements (GREs) in transcription units of target genes [49]. Co-activators such as SRC-1and CBP/p300 are recruited by activated GR, thus facilitating transcription of target genes [49]. GC/GR signaling is involved in many physiological processes, including carbohydrate metabolism, energy homeostasis, development and immune function [48]. GC levels are controlled by hypothalamic-pituitary-adrenal (HPA) axis in vivo [48]. In mouse and rat, the major endogenous GC is corticosterone, while in primates including humans the endogenous active GC is cortisol (or hydrocortisone) [50]. There are many synthetic GCs, such as dexamethasone, prednisone and fluticasone, which have been widely used as FDA approved drugs for anti-inflammation, allergy or asthma [48,51].

GCs regulate erythropoiesis. Co-culture with GCs in vitro promotes proliferation of erythroid progenitors [52]; in vivo, GR signaling is essential for stress-induced expansion of erythroid progenitors [52]. GR synergizes with PPARα promoting self-renewal of erythroid progenitor cells [53]. GC/GR signaling also plays pivotal roles in HSC homeostasis, including HSC production, HSC mobilization and HSC homing. During embryonic zebrafish hematopoiesis, serotonin regulates formation of HSCs and HPCs via HPA axis-mediated GR activation [54*]. Upon developmental hypoxic stress, HIF-1α is activated and induces neuronal serotonin synthesis to regulate cortisol-dependent production of HSCs and HPCs [54*]. Whether HPA-GR mediated HSC and HPC production is conserved in adults, needs further study. The central nervous system not only regulates HSC formation, but also regulates HSC mobilization via the HPA axis. HPA-GR signaling induced by muscarinic receptor type-1 (Chrm1) in hypothalamus is essential for G-CSF induced HSC and HPC mobilization [55*]. Mechanistically, physiological concentrations of GCs promote migration of HSCs and HPCs by modulating actin polymerization and microtubule assembly [55*]. However, administration of high dose GCs significantly suppresses egress of HSCs and HPCs mobilized by G-CSF [55*]. It is unknown if this GCs dependent mobilization of G-CSF is conserved in humans. Interestingly, GR is highly expressed in human CB CD34⁺ HSCs and HPCs, suggesting that GC/GR signaling might play a role in human HSC physiological function [56**]. In a compound screen to search for chemical candidates which enhance human HSC and HPC homing, dexamethasone was identified as a significant enhancer of CXCR4 expression and SDF1/CXCR4 axis-mediated chemotaxis [56**]. Short-term pretreatment of CB CD34⁺ cells with other synthetic GCs including Flonase (fluticasone propionate), cortisol (hydrocortisone), and Medrol (methylprednisolone) promoted CXCR4 expression and chemotaxis of HSCs and HPCs [56**]. Flonase pretreated CB CD34⁺ cells manifested significantly enhanced SRCs, and enhanced human cell chimerism in primary and secondary recipient NSG mice, demonstrating effects of GCs on long-term repopulating, self-renewing HSC in CB cells [56**]. Upon GCs treatment, cytoplasmic GR in CB HSCs and HPCs was activated. Activated GR translocated into the nucleus and bound to GREs in the human CXCR4 transcription unit (Fig. 1). SRC1/p300 histone acetyltransferase complex was recruited to promote histone H4 acetylation associated with CXCR4, thus facilitating transcription and expression of CXCR4 in CB HSCs and HPCs. Unexpectedly, CBP was not involved in this process although in most case p300 and CBP function as a complex, further suggesting cell type and context specific effects of GC signaling. There are two isoforms of CXCR4 (CXCR4-A and CXCR4-B) in human cells due to alternative splicing. The promoters of CXCR4-A and CXCR4-B are totally different. The GRE locates in the promoter of CXCR4-A and the first intron of CXCR4-B. Since mRNA levels of CXCR4-A and CXCR4-B are significantly upregulated (our unpublished data), it would be interesting to know if there are functional differences between these two human CXCR4 isoforms in migration and homing of HSCs and HPCs. There is no GRE throughout the mouse Cxcr4 transcription unit, and Cxcr4 expression and SDF-1 induced chemotaxis of mouse BM HSCs and HPCs remain unchanged upon GC treatment [56**]. Whether GC pretreatment of mouse BM HSCs and HPCs enhances their homing and engraftment by mechanisms independent of the SDF-1/CXCR4 axis needs investigation.

FIGURE 1.

Glucocorticoid (GC) pretreatment enhances homing and engraftment of human CB HSCs/HPCs by transcriptionally increasing CXCR4 expression. GC activates glucocorticoid receptor in the cytoplasm, and GC-GR complex translocates into the nucleus and binds as a dimer to the GC response element in CXCR4 transcription unit. SRC1-p300 histone acetyltransferase complex is recruited to enhance histone H4 (H4K5 and H4K16) acetylation, thus facilitating transcription of CXCR4 in CB HSCs/HPCs [56**].

ANTAGONIZING PPARγ SIGNALING EXPANDS HUMAN HSCS

The peroxisome proliferator-activated receptors (PPARs) family consists of three subtypes, PPARα, δ/β and γ. They are ligand-dependent transcription factors regulating genes involved in energy metabolism [57]. PPARα is highly expressed in liver and other tissues/cells with a high fatty acid oxidation (FAO) rate, and mainly regulates lipid and lipoprotein metabolism [57]. Of interest, PPARα activation promotes expansion and self-renewal of erythroid progenitors [53]. Whether this process is dependent on FAO remains unknown. PPARδ is mainly involved in FAO in skeletal and cardiac muscles [57]. PPARδ dependent FAO regulates asymmetric division of mouse HSC, and thus is essential for HSC maintenance [58]. PPARγ is mainly expressed, but not restricted, in adipose tissue, and mediates transcription of genes involved in adipogenesis, lipid metabolism and glucose homeostasis [59].

In a compound screen, we found that PPARγ antagonist GW9662 significantly promotes ex vivo expansion of human CB HSCs and HPCs in RPMI1640 medium containing 10% fetal bovine serum (FBS) and cytokines (TPO, SCF, FLT-3L) [60**]. PPARγ is highly expressed in human CB HSCs, and knockdown of PPARγ in CB CD34⁺ cells enhances ex vivo expansion of HSCs, suggesting that PPARγ signaling functions as a negative regulator of human HSC self-renewal [60**]. Mechanistically, antagonizing PPARγ down-regulates differentiation associated genes, including CD38, CD1d, HIC1, FAM20C, DUSP4, DHRS3 and ALDH1A2, and a negative regulator of glycolysis, FBP1[60**]. It has been shown that inhibition of ALDH1A2 significantly promotes expansion and maintenance of CB HSCs [43], which demonstrates that antagonizing PPARγ expansion of CB HSCs may act in part by preventing RA-induced differentiation. Importantly, PPARγ antagonism promotes glycolysis of CB HSCs and HPCs, while it has no effect on mitochondria metabolism [60**]. Loss of function of FBP1 was sufficient to expand CB HSCs and HPCs which was dependent on glucose metabolism. Antagonizing PPARγ-FBP1 signaling had no effect on expansion and glucose metabolism of CB HSCs and HPCs when CD34⁺ cells were cultured in StemSpan™ Serum-Free Expansion Medium II (SFEM II) (our unpublished data). Interestingly, expression of PPARγ and FBP1 is greatly downregulated, while glycolysis is significantly enhanced when CD34⁺ cells are ex vivo cultured in SFEM II compared with 10% FBS containing RPMI1640 regular medium (our unpublished data), suggesting that PPARγ-FBP1 signaling is repressed by unknown ingredients in SFEM II. These findings indicate that the PPARγ-FBP1 axis might function as a metabolic switch for expansion and self-renewal of CB HSCs and HPCs (Fig. 2).

FIGURE 2.

PPARγ functions as a negative regulator of human CB HSC maintenance and expansion. During ex vivo culture in regular medium (RPMI1640+10% FBS+Cytokines), PPARγ signaling is activated. Activation of PPARγ increases expression of differentiation associated genes including CD38, CD1d, HIC1, FAM20C, DUSP4, DHRS3 and ALDH1A2; and a negative regulator of glucose metabolism FBP1. HSCs/HPCs will not be well maintained and expanded due to increased differentiation and repressed glucose metabolism. Antagonizing PPARγ signaling enhances maintenance and expansion of human HSCs/HPCs by preventing differentiation and switching on FBP1-repressed glycolysis [60**].

Inhibition of FBP1 enhances glucose metabolism in many types of cancer cells, while overexpression of FBP1 suppresses growth of cancer cells by compromising glycolysis [61,62]. It is unknown if leukemia cell growth is dependent on glucose metabolism activated by suppressing FBP1, and if inhibition of FBP1 impairs mitochondria metabolism in leukemia. PPARγ agonist glitazones decreases expression of STAT5 target genes, HIF2α and CITED2, and disrupts quiescence and stemness of chronic myeloid leukaemia (CML) stem cells LSCs [63]. Further studies are needed to determine if PPARγ activation impairs glucose metabolism by increasing FBP1 expression in CML LSCs. In primates infected by human and simian immunodeficiency viruses, the multipotent potential of their HPCs is disrupted, which is dependent on repression of STAT5 by ectopic activation of PPARγ [64]. Whether STAT5 signaling is involved in expansion of CB HSCs and HPCs by antagonizing PPARγ needs study. Knockdown of PPARγ in Fancd2 deficient murine HSCs significantly enhances their repopulating capability in recipient mice, while loss of function of PPARγ in wild type murine HSCs has no significant change on their engraftment [65*]. PPARγ antagonism has no effect on expansion of mouse BM HSCs and HPCs, and doesn’t downregulate expression of differentiation associated genes and Fbp1 in mouse BM lineage negative cells [60**]. PPARγ and its co-activators are upregulated in human FANCD2-deficient lymphoblastic cell lines and primary Fanconi patient samples [65*]. Future studies should determine the possible role of FBP1-mediated glucose metabolism and mitochondria metabolism in FANCD2-deficient HSCs.

CONCLUSION

Transcriptional regulation is one key intrinsic mechanism underlying human HSC expansion and self-renewal. As important members of transcription factors, NRs including RAR, GR and PPARγ play significant roles in expansion, self-renewal, homing and engraftment of human CB HSCs. Sufficient HSC numbers and efficient HSC homing are pivotal for successful clinical CB HCT, especially in adults. The more we know about molecular mechanisms of CB HSC self-renewal and homing, the better efficacy we can achieve in CB HCT clinical trials. Overexpression of Musashi-2 (MSI2), encoding an RNA binding protein, significantly promotes expansion of CB HSCs by suppressing aryl hydrocarbon receptor (AhR) signaling [66*]. Simultaneous transplantation of an SR1 (AhR antagonist) expanded CB-derived CD34⁺ cell population and an unmanipulated CB unit significantly promoted faster recovery of neutrophils and platelets after infusion [17*]. It would be of interest to determine the efficacy of combining SR1 with NR ligands, especially the homing enhancer Flonase in CB HCT. Recently, transplantable human HSCs have been successfully reprogrammed from iPS and endothelial cells [67,68**]. Can engraftment of these reprogrammed HSCs be enhanced by modulating NR signaling? Since many NR ligands are FDA approved drugs, NRs represent promising targets for enhancing engraftment of human CB HCT.

KEY POINTS.

• Antagonizing RAR signaling increases myeloid cell production and promotes human HSC expansion and engraftment by inhibiting differentiation.

• Attenuating AhR signaling enhances human HSC expansion, and SR1 expanded HSCs and HPCs show significant faster recovery of neutrophils and platelets in clinical study of HCT.

• Glucocorticoid pretreatment of human HSCs and HPCs enhances their homing and engraftment capability in NSG mice.

• PPARγ antagonism promotes self-renewal and engraftment of human HSCs and HPCs by switching on FBP1-repressed glucose metabolism and preventing differentiation. PPARγ is ectopically activated in primary fanconi patient samples, and PPARγ signaling might be an effective target in therapy of fanconi patients.