AML: Acute Myeloid Leukemia; ATP: Adenosine Triphosphate; CDR: Commonly Deleted Region; CK1: Casein Kinase 1; EPO: Erythropoietin; ERK: Extracellular Signal-Regulated Kinase; ESA: Erythropoiesis-Stimulating Agent; G-CSF: Granulocyte Colony-Stimulating Factor; HSC: Hematopoietic Stem Cell; HSP: Heat Shock Protein; JAK2: Janus Kinase 2; MDS: Myelodysplastic Syndromes; NADH: Nicotinamide Adenine Dinucleotide Hydrogen; OXPHOS: Oxidative Phosphorylation; RBC: Red Blood Cell; STAT5: Signal Transducer and Activator of Transcription 5; TCA: Tricarboxylic Acid

Myelodysplastic syndromes (MDS) are acquired neoplastic myeloid proliferations in the bone marrow, where hematopoietic stem cells (HSCs) are impacted. MDS are characterized by clonal proliferation of HSCs, recurrent genetic abnormalities, progressive cytopenia, increased risk of acute myeloid leukemia (AML), and myelodysplasia [1]. The symptoms of MDS patients normally include fatigue, shortness of breath, pallor, unusual bruising or bleeding, petechiae, anemia, and frequent infections [2,3]. The diagnosis of MDS require at least 10% of the cells in one hematopoietic lineage to be dysplastic, specifically, a megakaryocyte lineage needs to be at least 10% micromegakaryocytes or at least 40% dysmegakarypoiesis [4]. Treatment for MDS includes supportive care, drug therapy, and stem cell transplantation. Supportive care may include transfusion therapy, erythropoiesis-stimulating agents (ESAs), and antibiotic therapy. Drug therapy can slow the progression to AML which includes lenalidomide, immunosuppressive therapy, azacitidine and decitabine, and chemotherapy [5].

Approximately 10-15% of MDS cases acquire an interstitial deletion on chromosome 5q, known as del(5q) [6,7]. Del(5q) is among the most common cytogenetic aberrations in MDS and defines a unique MDS sub-category [8]. Two distinct commonly deleted regions (CDR), proximal and distal CDR, have been identified on the del(5q) region in MDS patients. The proximal CDR at 5q31.2-5q31.3 is associated with higher-risk MDS with higher rate of progression to AML. The distal CDR is a 1.5 Mb deletion encompassing 5q32–5q33.2 and present in the classical 5q- syndrome with better prognosis (Figure 1) [9-11]. The precise number of genes located in each CDR is not fully characterized. Some studies reported that the proximal CDR contains ~30 genes [12] and distal CDRs contains ~ 41 genes [13]. The analysis based on the copy number variations reveals ~ 405 genes on 5q including ~ 41 in proximal CDR and ~ 55 in distal CDR [11].

A number of genes in CDRs have been implicated in essential for hemopoiesis and pathogenesis of del(5q) MDS through haploinsufficiency (Table 1) [14]. For example, the RPS14 gene encodes for ribosomal protein S14, a structural protein of the 40S ribosomal subunit with important functions in ribosomal biogenesis [15]. Pellagatti et al. reported haploinsufficiency of the ribosomal gene RPS14 in HSCs in patients with the 5q− syndrome, and also indicated its implication in the etiology of congenital anemia like Diamond-Blackfan anemia [16]. Ebert et al. showed that suppressing RPS14 expression in normal cord blood CD34+ cells resulted in a marked lack of erythroid maturation, and that overexpression of RPS14 in del(5q) CD34+ cells restored it [17]. This groups also reported that haploinsufficiency for ribosomal protein genes including RSP14 causes selective activation of p53 in human erythroid progenitor cells [18]. The CSNK1A1 gene encodes casein kinase 1 alpha (CK1α), the smallest isoform of the CK1 kinase family. The heterozygous deletion of this gene in myeloid lineage results in red blood cell (RBC) proliferation and expansion in mice, indicating CSNK1A1 plays a central role in the biology of del(5q) MDS and is a promising therapeutic target [19]. Mutations in CSNK1A1 E98 have been reported in ~ 18% and D140 in ~ 7% del(5q) MDS patients, which are associated with poor prognosis and decreased life expectancy. Liu et al. reported that E98K and D140A mutants have reduced ability to promote phosphorylation of β-catenin, resulting in enhanced Wnt signaling. These mutations linked functional changes may promote expansion of abnormal myeloid progenitors in del(5q) MDS [20]. Genes with important functions in hematopoiesis and implication in pathogenesis of del(5q) MDS are summarized in Table 1. Most del(5q) patients have deletions that span both CDRs, therefore, simultaneous deletion of genes on both CDRs may cooperate to contribute to altered hematopoiesis seen in patients.

Anemia is the most common clinical manifestation of MDS patients and present in the majority of MDS cases. Anemia in MDS is not only a transient manifestation, but also a disease similar to the situation of congenital anemias. The pathophysiologic causes of anemia include the dysplasia of erythopoietic progenitors and the abnormal HSCs losing their ability to mature into RBCs [21]. It has been demonstrated that the formation of lipid raft microdomains, assembling erythropoietin (EPO) receptor, STAT5, JAK2, and Lyn kinases, are key regulators of EPO signaling and erythroid maturation [22]. The plasma membrane of hematopoietic cells contains sphingolipid and cholesterol enriched microdomains called lipid rafts, which serve as cluster signaling intermediates to create focused signaling platforms that facilitate receptor-induced activation of signal transduction molecules. It has been observed that MDS erythroid progenitors display a significantly diminished lipid raft assembly and smaller raft aggregates, suggesting disruption of raft integrity may be responsible for the impaired EPO signaling in MDS [23] Although MDS dysplastic erythroid progenitor cells may express EPO receptor at a normal density, in vitro studies showed that the signaling pathway in response to EPO stimulation is defective, such as lack of STAT5 phosphorylation [24]. In addition, Frisan et al. reported that defective ERK1/2 phosphorylation upon EPO stimulation signaling is altered in MDS progenitors [25]. For the clinical treatment of anemia in del(5q) MDS patients, multiple strategies have been used including blood transfusion, iron chelation therapy, erythropoiesis-stimulating agents (ESAs), and granulocyte colony-stimulating factor (G-CSF) [5,22].

 The HSPA9 gene encodes the protein mortalin, and its location is mapped to 5q31.2, a region located in the proximal CDR of the del(5q) MDS region [26]. The HSPA9/mortalin is a highly conserved heat-shock chaperone belonging to the heat shock protein 70 (HSP70) family. It is predominantly presented in the mitochondria, but also found in other sub-cellular compartments including plasma membrane, endoplasmic reticulum, and cytosol [27]. HSPA9/mortalin is critical in regulating a variety of cell physiological functions such as response to cell stress, control of cell proliferation, and inhibition/prevention of apoptosis [28]. Our laboratory has been studying the role of HSPA9 gene in haematopoiesis and erythroid maturation for years. We observed that Hspa9 homozygous deletion mice are embryonic lethal, while Hspa9 heterozygous deletion mice have normal basal hematopoiesis, but display altered B-cell lymphopoiesis [29]. This is consistent with our later reports showing that haploinsufficiency of multiple del(5q) genes, including Hspa9, also induce B-cell abnormalities in mice [12]. In addition to using mouse models, we have identified that HSPA9 knockdown induces apoptosis in human CD34+ hematopoietic progenitor cells, which is likely a TP53-dependent process, suggesting that reduced levels of HSPA9 may contribute to p53/TP53 activation and increased apoptosis observed in del(5q) MDS (Figure 2) [30]. More recently, we reported that HSPA9/mortalin inhibition disrupts erythroid maturation through a TP53-dependent mechanism in human CD34+ hematopoietic progenitor cells [31]. This study indicates that insufficiency of specific del(5q) MDS-associated gene could disrupt erythroid maturation, providing insights into ineffective erythropoiesis and anemia observed in patients.

The precise relationship between erythroid maturation, TP53, and HSPA9 remains unclear. For future studies, there are at least four directions that we can pursue to elucidate the mechanisms. First, we plan to investigate the energy metabolism during the erythroid maturation process. Energy metabolism includes two main pathways: glycolysis and mitochondrial oxidative phosphorylation (OXPHOS). Glycolysis takes place in cytosol and converts glucose into pyruvate and to produce ATP and NADH. Oxidative phosphorylation takes place in mitochondria and uses energy released through the TCA cycle to produce ATP with higher efficiency than glycolysis [32]. In the human blood system, HSCs and progenitor cells possess the ability of both differentiation (into downstream mature blood cells) and self-renewal (proliferation without differentiation). It was suggested that a balance between glycolysis and OXPHOS metabolism could guide the choice between differentiation and self-renewal in stem cell fate [33]. Glycolysis was shown to be characteristic of self-renewing stem and progenitor cells, while there is a switch from glycolytic metabolism toward OXPHOS during cell differentiation [33,34]. Specifically, Richard et al. reported that erythroid differentiation requires a peak of energy consumption associated with glycolytic metabolism rearrangements, in which glycolysis is less involved in differentiating erythroid cells compared to OXPHOS [35]. The relationship between HSPA9/mortalin and glycolytic metabolism has been studied by other groups. Echtenkamp et al. showed that pharmacological inhibition of mortalin by JG-98, an allosteric inhibitor of HSP70, could resensitize castration-resistant prostate cancer to androgen deprivation drugs by suppressing aerobic respiration. JG-98’s primary effect is to inhibit mitochondrial translation, leading to disruption of electron transport chain activity [36]. Ferguson et al. found that JG-98 could increase the potency against myeloma cells resistant to proteasome inhibitors. JG-98 could localize to mitochondria, inhibit mortalin, and deplete mitochondrial ribosome proteins [37]. Our published data showed that HSPA9/mortalin inhibition disrupts erythroid maturation through a TP53-dependent mechanism [30,31]. We expect that that the TP53-dependent regulation of erythroid maturation by HSPA9/mortalin is through glycolic metabolism modification, more specifically, inhibiting glycolysis and enhancing OXPHOS (Figure 1). Second, Chen et al. defined the stages during erythroid differentiation based on dynamic changes in the expression of red cell membrane proteins such as maturation such as Band3, CD44, CD71, spectrin, and beta-actin [38]. Thus, it is worth to investigate whether HSPA9 is involved in the process of red cell membrane development, which is critical for erythroid maturation. Third, we also showed that HSPA9 knockdown inhibits Stat5 activation in vitro in the presence of IL-7 [29]. Stat5a and 5b homozygous deletion mice showed ineffective erythropoiesis due to decreased survival of early erythroblasts [39]. Hence, HSPA9 could possibly regulate erythroid maturation through STAT5 signaling. Fourth, Ganguli et al. reported that p53 inhibits glucocorticoid-induced proliferation of erythroid progenitors, suggesting that TP53 may inhibit erythropoiesis which is consistent with our hypothesized model in Figure 1 [40]. Interestingly, human HSPA9 gene and glucocorticoid receptor gene (known as NR3C1) are both located on chromosome 5q31 [41,42]. Thus, HSPA9/mortalin may interact with glucocorticoid receptors, which mediates erythropoiesis through TP53 regulation. In addition, other studies showed that glucocorticoid receptor-Stat5 interaction in hepatocytes controls body size and maturation-related gene expression in mice [43,44]. Therefore, it is also worth to explore whether STAT5 signaling could potentially be triggered by HSPA9 and glucocorticoid in erythropoiesis.

Research reported in this publication was supported by the National Institute of General Medical Sciences of the National Institutes of Health under Award Number 2U54GM104942-08 (Tuoen Liu). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. This study was also supported in part by NIH grant P20GM103434 to the West Virginia IDeA Network of Biomedical Research Excellence. This study was also supported by West Virginia School of Osteopathic Medicine intramural fund (Tuoen Liu).

Alfredo Fallorina and Tuoen Liu wrote, edited, and approved the manuscript.

The authors declare no conflict of interest.