Diamond Blackfan Anemia (DBA) is a rare inherited bone marrow failure syndrome characterized by red cell aplasia, congenital anomalies, and a predisposition to cancer. Compared to other inherited bone marrow failure syndromes including Fanconi anemia, Dyskeratosis Congenita and Shwachman–Diamond syndrome, DBA involves a specific block in erythropoiesis at the progenitor stage, resulting in the clinical hematologic phenotype of severe anemia [1-3].

More than 70% of patients with DBA have been shown to have a haploinsufficiency in one of 20 different ribosomal protein genes in both the small and large subunits (RPS and RPL) (Figure 1). Among these mutations, six genes (RPS19, RPL5, RPS26, RPL11, RPL35a, and RPS24) account for the majority of cases reported to date, with RPS19 being the most commonly mutated and accounting for 25% of cases [3,4]. While these mutations lead to a reduction in total ribosome levels which would be expected to have more global consequences, the most significant effects appear to be within the erythroid lineage. Previous studies have shown that the ribosomal insufficiency induces ribosomal stress leading to stabilization of the p53 protein and subsequent apoptosis that appears to specifically affect the erythroid progenitors when compared to other hematopoietic cell lineages [5,6]. Furthermore, it has been demonstrated that the global reduction in ribosome levels observed in DBA affects the translation of a subset of transcripts, notably GATA1, which plays an essential role in erythroid development. Indeed, mutations in GATA1 have been identified in patients with DBA [7]. However, the precise molecular mechanisms by which ribosome insufficiency in DBA specifically affects erythropoiesis is still an ongoing area of investigation.

Corticosteroid therapy remains the standard of care for DBA with the majority (80%) of patients showing an initial response with an increase in hemoglobin. However, some patients do not respond, become refractory to the therapy and/or develop serious side effects [8]. In these patients, chronic red cell transfusion is used to manage the anemia. Since the iron released from the transfused blood is not recycled in most hematologic conditions including DBA, the iron ultimately accumulates in tissues and chelation therapy is required to avoid the consequences from iron overload. Hematopoietic stem cell transplantation is currently the only curative option, but it can involve significant morbidity and is generally limited to those with human leukocyte antigens (HLA)-matched donors [9].

Given the limited treatment options, there are several novel drugs that are being tested in clinical trials for patients with DBA, including sotatercept (ClinicalTrials.gov Identifier: NCT01464164) and trifluoperazine (ClinicalTrials.gov Identifier: NCT03966053). Sotatercept is a human fusion protein containing the extracellular domain of a receptor for activin type IIA [10] and trifluoperazine is a calmodulin inhibitor. Both drugs have been shown to improve erythropoiesis in models of DBA [11,12]. Based on the more recent identification of other pathways and potential therapies in DBA, we undertook a study to look at other potential targets.

While investigating the signaling pathways regulating the pathogenesis of DBA, we found that Nemo-like Kinase (NLK) was hyperactivated in erythroid progenitors in cellular models of DBA as well as in bone marrow progenitor cells from patients [13]. NLK is an evolutionarily conserved atypical proline-directed serine/threonine mitogen-activated protein kinase. NLK plays a critical role in the regulation of diverse pathways, including the Wnt/β-catenin and Notch signaling pathways [14], and is involved in the development and the progression of neurodegenerative diseases [15]. NLK-deficient mice showed developmental and functional defects in the lungs, heart and skeleton [16-18]. Furthermore, it has been shown that NLK activity is required to regulate the function and differentiation of immune cells including CD8+ thymocytes and regulatory T-cells [19,20], in addition to regulating neurodegenerative diseases such as Huntington’s disease and spinocerebellar ataxias [21]. One of the regulation mechanisms of NLK expression involves a microRNA (miRNA), and a number of miRNAs, including miRNA-181 [22] reduces the expression level of NLK protein by binding to the 3’UTR of NLK mRNA. Our subsequent studies demonstrated that metformin, an oral antidiabetic drug, and ginsenoside Rb1, the active component of ginseng, suppressed NLK expression and improved erythropoiesis in models of DBA through induction of miRNA26a and miRNA208, respectively [23,24].

When activated, NLK forms a homo-dimer and auto-phosphorylates [25]. The molecular mechanisms of action vary in different cellular contexts and several NLK-activating kinases have been identified. Wnt-1 signal causes the nuclear entry of transforming growth factor-β (TGF-β)-activated kinase (TAK1), which phosphorylates homeodomain-interacting protein kinase 2 (HIPK2). Phosphorylated HIPK2 subsequently binds NLK and promotes autophosphorylation of NLK. Together with HIPK2, NLK binds c-Myb, resulting in the phosphorylation, ubiquitination and proteasome-dependent degradation of c-Myb [26]. c-Myb is a master regulatory transcription factor which regulates hematopoietic stem cell proliferation and differentiation. It has been shown to be downregulated in erythroid progenitors from patients with RPS19-mutated DBA [27]. Our data confirmed that the NLK-dependent phosphorylation of c-Myb was responsible for the observed proteasomal degradation in RPS19-insufficienct cells. It is also reported that p38 MAPK activates NLK by directly interacting with NLK in other cell types [28], although suppression of p38 did not improve erythropoiesis in DBA models [13].

We also found that Raptor, another substrate of NLK, was highly phosphorylated by hyper-activated NLK in RPS19-insufficiency [13]. Raptor is a subunit of the mechanistic target of rapamycin complex 1 (mTORC1) which regulates many fundamental cellular processes, such as protein translation, autophagy, and metabolism [29]. The lysosomally localized Rag guanosine triphosphatases (GTPases) recruit mTORC1 to lysosomes through interaction with Raptor, and mTORC1 is then activated at the outer lysosomal membrane by the Ras homolog enriched in brain (Rheb) GTPases embedded in the membrane [30]. Under stress conditions, NLK phosphorylates Raptor to inhibit the lysosomal localization of mTORC1 and thereby suppresses mTORC1 activation [31]. Our data demonstrated that Raptor was phosphorylated in an NLK-dependent manner and that the co-localization of Raptor with the lysosome was disrupted in RPS19-insufficient cells [13]. Furthermore, it has been shown that the amino acid L-leucine, a known activator of mRNA translation, increases protein translation in erythroid progenitors in human, murine and zebrafish models of DBA through the hyperactivation of mTORC1 [32,33]. Amino acids are known to regulate the lysosomal translocation and activation of mTORC1 [34], and leucin promotes the activation of the Rag GTPases and their interaction with Raptor, leading to the recruitment of mTORC1 to the lysosome membrane to stimulate mTORC1 activation [35,36]. Clinical trials of L-leucine for patients with DBA have been published with promising results at the suboptimal doses [37,38].

A summary of the upstream kinases activating NLK and downstream targets affected by NLK relevant to DBA are summarized in Figure 2.

Figure 2: The upstream kinases activating NLK and downstream targets affected by NLK relevant to DBA. c-Myb protein, a master regulatory transcription factor for hematopoiesis, is phosphorylated and ubiquitinated for proteasomal degradation via the pathway involving TAK1, HIPK2 and NLK [25,26]. Under stress conditions, NLK phosphorylates Raptor to disrupt its interaction with the Rag GTPase, which inhibits the translocation of mTORC1 to lysosomal where mTORC1 is activated by the Rheb GTPase [31]. GTPase: Guanosine Triphosphatase; HIPK2: Homeodomain-Interacting Protein Kinase 2; mTORC1: The mechanistic Target of Rapamycin Complex 1; NLK: Nemo-Like Kinase; Rheb: Ras homolog enriched in brain; TAK1: Transforming growth factor β-Activated Kinase.

TGF-β is a critical member of a large superfamily of secreted growth and differentiation factors. With respect to hematopoiesis, TGF-β plays an essential role in controlling a wide array of biological processes from homeostasis of the immune system to quiescence and self-renewal of HSCs [39]. Small molecule inhibitors of the TGF-β pathway have been shown to enhance the proliferation of hematopoietic progenitors from patients with DBA [40] and to improve erythropoiesis in a mouse model of myelodysplastic syndrome [41]. These findings suggest that the TGF-β pathway may be involved in the pathobiology of ineffective hematopoiesis in these diseases, but the underlying mechanism has not been established. Our data indicated that only a subset of TGF-β type I receptor inhibitors including SD208 [42], SB431542 [43] and galunisertib [44], improved erythropoiesis in models of DBA and that the improvement was not through direct inhibition of TGF-β receptors but rather by the inhibition of NLK activation, suggesting that NLK is a key factor in the pathobiology of erythroid defects in DBA [13].

We demonstrated that chemical inhibition of NLK activity with small molecules targeting NLK increased erythroid expansion in human and murine models of DBA, including those with RPS19 and RPL11 insufficiency [13]. Given that DBA is associated with mutations in both ribosomal subunits, pharmacologically targeting NLK as a common therapeutic target, could be a more feasible treatment option for patients in contrast to gene therapy or gene editing with autologous stem cell rescue. Although specific chemical compounds that inhibit NLK activity have not been developed to date, we did find that a number of clinically approved compounds inhibited NLK activity as an off-target effect. Among those compounds, off-target effects of a subset of TGF-β inhibitors showed up to 40% recovery of erythroid expansion due to inhibition of NLK activity in models of DBA when compared to control [13]. We hypothesize this will be sufficient to increase hemoglobin levels and minimize chronic red cell transfusions in patients with DBA.

In addition to small molecule inhibitors for NLK activity, chemical compounds that downregulate NLK expression though miRNA-mediated suppression, such as metformin and ginsenoside Rb1, may provide a clinical benefit to patients with DBA, particularly in combination with other drugs, with a higher specificity and fewer off-target effects [23,24]. These compounds would be expected to impact erythroid expansion more significantly by miRNA-mediated suppression and kinase inhibition of NLK together.

Beyond c-Myb and Raptor phosphorylation, the exact role of NLK activation during erythropoiesis remains to be explored more carefully. Our data demonstrated that NLK was transiently and modestly activated during early hematopoietic differentiation, and that NLK expression in non-erythroid lineages decreased as cells differentiate. Also, small molecule NLK inhibitors did not affect differentiation or proliferation in non-erythroid lineages, either in ribosome-competent or insufficient cells [13]. In RPS19-insufficient human HSPCs, genetic silencing of NLK by siRNA improved erythroid expansion to 46.7% of ribosome-competent controls, but the rescue was not complete [13]. For future studies, the effects of NLK on distinct stages of erythropoiesis in both normal and pathogenic models needs to be studied in more detail. This will help to ultimately define the optimal clinical trial for patients with DBA, including dose, timing, and duration for potential NLK inhibitors in the disease.

Diamond Blackfan anemia was the first disease associated with defects in ribosome biology. While our understanding has grown significantly over the years, there are still many areas to explore both from a pathophysiologic and therapeutic angle. In this review, we highlighted some of the ongoing work in the field and presented some of our work highlighting that potential role of NLK in DBA. We believe that NLK contributes to the erythroid defect in DBA and that the targeting of NLK has the potential to be a common therapeutic approach for patients with DBA, regardless of their mutational status. Although pharmacologically targeting NLK does not completely rescue the erythroid defect in ribosome-insufficient models of DBA in our pre-clinical studies, the improvement is potentially sufficient to provide significant clinical benefits to patients especially in combination with other drugs that are used and/or being investigated.

This study was supported by NIH T32 training grant (DK098132) (A.S and M.W.), Department of Defense (BM180024) (K.M.S.), California Institute of Regenerative Medicine (DISC2-12475) (K.M.S.), and National Heart Lung and Blood Institute (R01HL144436-01) (A.N.).

No conflicts of interests to declare.