DYRK1a mediates BAFF-induced noncanonical NF-κB activation to promote autoimmunity and B-cell leukemogenesis

B cells play a central role in mediating humoral immune responses and, when deregulated, contribute to the development of autoimmune diseases and B-cell malignancies.^1-3 Early stages of B-cell development occur in the bone marrow, where B-cell progenitor cells progress through pro-B, pre-B, and immature B-cell stages defined by ordered rearrangement and expression of B-cell receptor (BCR) genes.⁴ Immature B cells migrate to the spleen as transitional 1 (T1) B cells, which undergo a process of maturation to become T2 immature cells and then mature B cells under the regulation of signals from the BCR and B-cell activating factor (BAFF) receptor (BAFFR).⁵ BAFFR is particularly important for the survival of T2 and mature B cells.⁶ Defect in BAFF expression or BAFFR signaling is associated with humoral immune deficiencies, whereas aberrant BAFF production contributes to the pathogenesis of autoimmune diseases.⁷ 

A major signaling event stimulated by the BAFFR is activation of the noncanonical nuclear factor (NF)-κB (ncNF-κB) pathway, which is critical for peripheral B-cell development and survival.⁸ This pathway is based on proteasomal processing of the NF-κB2 precursor protein p100, which generates mature NF-κB2 and p52 and causes nuclear translocation of p100-sequestered NF-κB members, p52 and RelB.^9,10 A central component of the ncNF-κB pathway is NF-κB–inducing kinase (NIK), encoded by the gene Map3k14, which along with its downstream kinase, IKKα, induces phosphorylation-dependent processing of p100.^9,11 The mechanism of ncNF-κB signaling centers on the regulation of NIK degradation. Under steady-state conditions, NIK is targeted for ubiquitin-dependent degradation by an E3 ubiquitin ligase complex, composed of cIAP, TRAF2, and TRAF3 (hereafter called cIAP-TRAF complex), which serves as a mechanism to prevent signal-independent ncNF-κB activation.^10,12-15 In response to signals from a subset of tumor necrosis factor receptor superfamily members, which in B cells include BAFFR and CD40, NIK is liberated from this E3 ligase complex, and the accumulation of stabilized NIK triggers ncNF-κB activation.¹⁰ Genetic deficiencies in ncNF-κB signaling components can cause immunodeficiencies, whereas deregulated activation of this pathway is associated with autoimmune diseases and B-cell malignancies.^10,16-18 However, to date, the molecular mechanism underlying signal-induced NIK stabilization is still poorly defined.

As an approach to elucidate the mechanism underlying the regulation of cIAP-TRAF E3 ligase complex, we performed an unbiased screening to identify novel factors physically associating with TRAF3. We identified a novel TRAF3-binding protein, dual-specificity tyrosine phosphorylation-regulated kinase 1a (DYRK1a). DYRK1a gene locus amplification has been associated with human B-cell acute lymphoblastic leukemia (B-ALL), and DYRK1a is important for leukemogenesis in a mouse model of B-ALL.^19,20 Notably, although B-ALL cells are of pre–B-cell origin, they are characterized by a high level of BAFFR expression.²¹ We show in the present study that DYRK1a is a kinase that responds to BAFFR and CD40 stimulation and mediates ncNF-κB activation and peripheral B-cell survival. Our data suggest that DYRK1a promotes NIK stabilization by phosphorylating TRAF3. These findings establish DYRK1a as a novel mediator of ncNF-κB activation and provide new insight into peripheral B-cell development and B-ALL pathogenesis.

M12 cells were transduced with a retroviral vector encoding biotin identification (BioID) TRAF3 fusion protein with an N-terminal Myc epitope tag (pCLXSN(GFP)-myc-BioID-TRAF3) or the control pCLXSN-myc-BioID vector. The cells were cultured in biotin-supplemented medium for 24 h and then subjected to BioID screening, as described previously.²² TRAF3-interacting proteins were identified using mass spectrometry at the Mass Spectrometry Proteomics Core of Baylor College of Medicine (Houston, TX).

B cells were purified from the splenocytes by anti–B220-conjugated magnetic beads (Miltenyl Biotec). Purified B cells in replicate wells of 96-well plates (2 × 10⁵ cells per well) were stimulated at 37°C with CD40L (500 ng/mL) or BAFF (200 ng/mL). After stimulation for 40 h, B cells were pulse-labeled for 6 h with [³H] thymidine and then collected for thymidine-incorporation assays. B-cell survival was analyzed by SYTOX Blue dead cell stain (Invitrogen) after 48-h culture with indicated condition. APC-annexin V and propidium iodide (BD Pharmingen) were used to quantify the apoptotic cell population.

Cell suspensions from various organs or cell lines were stained as indicated and acquired using an LSRFortessa flow cytometer (BD Biosciences), as described previously.²³ The data were analyzed using FlowJo software. Gating strategy is shown in supplemental Figure 1 (available on the Blood Web site).

The BM12 model of lupus-like autoimmune disease was induced and analyzed as described.²² 

Statistical analysis was performed using GraphPad Prism software. Significant changes between 2 groups were analyzed with an unpaired, 2-tailed Student t test. Two-way analysis of variance (ANOVA) with Bonferroni multiple-comparison test were used for in vitro cell survival assay. Kaplan-Meier analyses were performed, and the log-rank Mantel-Cox test was used to determine statistical difference between the survival curves of two groups. Values of P < .05 were considered significant, and the level of significance was indicated as *P < .05, **P < .01, and ***P < .001. All data are presented as mean ± standard deviation (SD).

To characterize novel regulators and functions of the ncNF-κB signaling pathway, we screened for TRAF3-binding proteins by proximity-dependent BioID, a technique that allows identification of signaling partners–based in vivo protein association.²⁴ We constructed retroviral vectors encoding either a Myc-BioID control or Myc-BioID–TRAF3 fusion protein and stably infected a B-cell line, M12 (supplemental Figure 2A-B). After biotin treatment of the cells, we performed streptavidin bead pull-down to isolate the biotinylated proteins (supplemental Figure 2C-D). Mass spectrometry analysis identified several known TRAF3-binding proteins, such as TBK1, CYLD, TRAF5, as well as novel TRAF3-interactors, including DYRK1a (supplemental Table 1).

Co-immunoprecipitation (CoIP) assays confirmed the strong interaction between TRAF3 and DYRK1a (Figure 1B,D) and revealed the essential role of the C-terminal TRAF domain (TRAF-C) and a polyhistidine motif of DYRK1a for this molecular interaction (Figure 1A-E). The interaction between TRAF3 and DYRK1a was also readily detected under endogenous conditions in B cells, and this interaction was constitutive and not further enhanced upon BAFF stimulation (Figure 1F). These results identify DYRK1a as a novel factor physically interacting with TRAF3.

TRAF3 controls B-cell survival and maturation through regulating ncNF-κB activation.^25-27 We therefore examined the role of DYRK1a in regulating B-cell development and maturation by generating DYRK1a B-cell–conditional knockout (Dyrk1a^BKO) mice. We crossed the Dyrk1a-flox mice with Cd19-Cre or Mb1-Cre mice (supplemental Figure 3); the Cd19-Cre mice are known to express the Cre recombinase predominantly in peripheral B cells, whereas the Mb1-Cre mice strongly express the Cre recombinase in early stages of bone marrow B cells.²⁸ In line with a role of DYRK1a in pre–B-cell development,²⁹ deletion of DYRK1a using the Mb1-Cre reduced the number of bone marrow B cells from the pre-B stage (supplemental Figure 4A-B). Interestingly, a more striking defect was seen with the mature B cells (supplemental Fig. 4A-B). Because mature B cells are generated in the spleen, we examined the role of DYRK1a in regulating peripheral B cells by using the Dyrk1a^BKO mice generated with Cd19-Cre. Consistent with the poor expression of Cd19-Cre in early stages of bone marrow B cells, deletion of DYRK1a using this Cre did not significantly reduce the number of pro-B, pre-B, or immature B cells but led to a reduction in the mature B-cell population (supplemental Figure 4C-D). We used this BKO model for the following studies except as otherwise indicated. Compared with wild-type (WT) control mice, the Dyrk1a^BKO mice had a decreased frequency and absolute number of B cells in peripheral lymphoid organs (Figure 2A). The DYRK1a deficiency predominantly reduced the number of mature B cells, although it also moderately reduced the number of immature B cells (Figure 2B).

Immature B cells in the spleen include transitional (T) 1 and T2 populations.⁶ The B-cell–specific DYRK1a deficiency caused a reduction in the T2, but not T1, population (Figure 2C), suggesting the requirement of DYRK1a for the maintenance of T2, but not T1, stage of immature B cells. Splenic mature B cells include follicular and marginal zone B-cell populations.⁶ Both the follicular and the marginal zone B cells were reduced in the Dyrk1a^BKO mice (Figure 2D). These results demonstrate a crucial role of DYRK1a in the generation or maintenance of mature B cells and T2 immature B cells. Of note, these phenotypes of the Dyrk1a^BKO mice were reminiscent of mutant mice defective in ncNF-κB signaling.⁶ 

B cells are a major mediator of systemic autoimmune diseases, including systemic lupus erythematosus (SLE), and manipulation of B cells has been explored as a therapeutic approach. Because of the requirement of DYRK1a in peripheral B-cell maintenance, we examined the role of DYRK1a in autoimmunity using a mouse model of SLE.³⁰ This model involves adoptive transfer of T cells from a C57BL/6 variant mouse strain, BM12, harboring a 3-amino acid substitution in the MHCII molecule to the WT or Dyrk1a^BKO mice (in C57BL/6 background). The BM12 donor T cells become activated in the recipient via alloreactivity and then activate self-reactive recipient B cells to cause SLE-like symptoms. Upon BM12 T-cell transfer, the WT mice developed SLE-like symptoms, including enlargement of the spleen, B-cell hyperplasia associated with increased plasma cells, autoantibody accumulation, and deposition of immunoglobulin (Ig) G/complement immune complexes to the kidney glomeruli (Figure 2E-I). Importantly, these phenotypes were ameliorated in the Dyrk1a^BKO recipient mice (Figure 2E-I). In a control experiment, transplantation of the WT and Dyrk1a^BKO mice with C57BL/6 T cells did not cause plasma cell increase or autoantibody production (supplemental Fig. 5). These results suggest that DYRK1a is critical for the induction of SLE-like autoimmune symptoms.

The results described above raised the question of whether DYRK1a might regulate B-cell survival or expansion induced by ncNF-κB inducers. As expected, wild-type B cells underwent massive cell death when cultured ex vivo, which was inhibited by the ncNF-κB inducers, BAFF and CD40L (Figure 3A). Importantly, DYRK1a deficiency impaired BAFF- and CD40L-induced B-cell survival and proliferation (Figure 3B-C).

Remarkably, the DYRK1a deficiency attenuated BAFF-induced ncNF-κB activation, as revealed by the impaired loss of cytoplasmic p100 and induction of nuclear p52 and RelB (Figure 3D). A central step in ncNF-κB signaling is stabilization of NIK, which is normally caused by degradation of its inhibitory protein TRAF3.³¹ Interestingly, the DYRK1a deficiency did not inhibit TRAF3 degradation. Nevertheless, the DYRK1a deficiency impaired induction of NIK protein accumulation (Figure 3E-F) without inhibiting Map3k14 mRNA expression (Figure 3G). These results, along with the DYRK1a-TRAF3 interaction, suggested a role for DYRK1a in regulating ncNF-κB signaling.

Because DYRK1a regulates signal-induced, but not steady-state, ncNF-κB activation and B-cell survival, we questioned whether this kinase responds to the ncNF-κB signals. In this regard, DYRK1a activation involves its autophosphorylation at a conserved tyrosine residue in the activation loop.³² Importantly, while the basal level of DYRK1a phosphorylation was low in B cells, it was potently stimulated by BAFF and CD40L (Figure 3H). These results establish DYRK1a as a kinase that facilitates the induction of ncNF-κB signaling and B-cell survival by BAFFR and CD40.

To determine the role of the ncNF-κB pathway in DYRK1a-mediated B-cell regulation, we used the R26Stop^FLMap3k14 transgenic mice carrying a Map3k14 transgene under the control of a loxP-flanked STOP cassette.³³ By crossing these mice with the Dyrk1a^BKO mice, we generated mutant mice with B-cell–specific NIK overexpression and DYRK1a deletion (Dyrk1a^BKOMap3k14^Btg). Remarkably, while the Dyrk1a^BKO mice had reduced frequency and number of splenic B-cell populations, this phenotype was reversed in the Dyrk1a^BKOMap3k14^Btg mice (Figure 4A-B). Consistently, NIK expression rescued the activation of ncNF-κB in DYRK1a-deficient B cells and restored the survival ability of these mutant B cells (Figure 4C-E). Compared with the WT mice, the Map3k14^Btg mice had increased frequency and number of B cells (supplemental Figure 6A-B), which was in line with a previous study.³³ Importantly, DYRK1a deficiency in the Map3k14^Btg mice (Dyrk1a^BKOMap3k14^Btg) did not cause a reduction in B-cell populations (supplemental Figure 6A-B), suggesting that ectopic NIK expression rescues the B-cell defect of DYRK1a deficiency. Compared with WT control B cells, the Map3k14^Btg B cells also displayed increased survival, which was only moderately reduced upon DYRK1a deletion (Map3k14^BtgDyrk1a^BKO) (supplemental Figure 6C-D). Consistently, NIK ectopic expression led to strong ncNF-κB activation (nuclear p52 expression) in both the Map3k14^Btg and Map3k14^BtgDyrk1a^BKO B cells (supplemental Figure 6E), suggesting that NIK functions downstream of DYRK1a. Thus, DYRK1a mediates peripheral B-cell survival and maturation via facilitating activation of the ncNF-κB pathway.

Because DYRK1a is a kinase interacting with TRAF3, we examined whether DYRK1a might phosphorylate TRAF3 by mass spectrometric analysis of TRAF3 co-expressed in 293 with WT DYRK1a or a kinase-dead DYRK1a mutant, K188R (supplemental Figure 7A). We identified 3 major phosphorylation sites, S9, T29, and S75, in TRAF3, among which T29 was specifically induced by the WT DYRK1a (supplemental Figure 7B-C; supplemental Table 2). T29 is located in the N-terminal region of human TRAF3, and this site is conserved as either a T or S across vertebrate species (supplemental Figure 7D). T29 and S29 were also identified as phosphorylation sites of human and mouse TRAF3, respectively, in public phosphorylation database (supplemental Figure 7E). In vitro kinase assays confirmed that purified DYRK1A could phosphorylate the N-terminal region (amino acid 1-165) of both human and mouse TRAF3 (Figure 5A-B). Furthermore, mutation of T29 in human TRAF3 (1-165) and S29 in mouse TRAF3 (1-165) abolished TRAF3 phosphorylation by DYRK1a (Figure 5A-B). Using the in vitro kinase assay, we further demonstrated that the catalytic activity of endogenous DYRK1a was stimulated in B cells by BAFF and CD40L (Figure 5C).

To determine the functional significance of DYRK1a-mediated TRAF3 phosphorylation, we generated bone marrow chimeric mice by adoptively transferring the lymphocyte-deficient Rag1-KO mice with TRAF3-deficient bone marrow cells reconstituted with either WT TRAF3 or TRAF3 T29A. Spleen B cells from the TRAF3 T29A chimeric mice had reduced processing of p100 and nuclear translocation of p52 and RelB compared with the B cells of WT TRAF3 chimeric mice (Figure 5D). Consistently, the TRAF3 T29A chimeric mice displayed a reduction in the percentage and absolute number of peripheral B cells, most strikingly mature follicular B cells (Figure 5E-F). Together, these results suggest that DYRK1a mediates ncNF-κB activation and peripheral B-cell development via phosphorylation of TRAF3.

DYRK1a has been implicated in human leukemia and recently shown to be important for a mouse model of B-ALL.^19,20 Our analysis of a human pediatric ALL database revealed DYRK1a amplification or mRNA overexpression in 10.83% of total cases (203 samples from 154 patients) (supplemental Figure 8). Notably, a hallmark of B-ALL cells is high-level expression of BAFFR despite their pre–B-cell origin.²¹ Because BAFFR is normally expressed in mature and T2 transitional B cells but not pre-B cells,⁶ its expression in malignant pre-B cells suggests the involvement of BAFFR signaling in B-ALL development. In support of the previous study, our flow cytometric assays revealed BAFFR expression in several human B-ALL cell lines, including SEMK2, TANOUE, CALL4, MTUZ5, and BALL1 (Figure 6A). Consistent with the lack of BAFFR expression in Burkitt lymphoma,³⁴ BAFFR was not detected in the Burkitt lymphoma cell line DAUDI (Figure 6A). Using the BALL1 cell line, we demonstrated that as seen in normal B cells, DYRK1a and TRAF3 physically interacted in B-ALL cells (supplemental Figure 9). Furthermore, the BALL1 cells responded to BAFF for activation of ncNF-κB, as revealed by cytoplastic p100 degradation and nuclear p52 induction (Figure 6B). The BAFF-induced ncNF-κB activation was impaired upon DYRK1A knockdown in BALL1 cells (Figure 6B). BAFF-induced, although not steady-state, BALL1 cell survival was compromised by the DYRK1A knockdown (Figure 6C). Remarkably, ectopic NIK expression efficiently rescued the survival of DYRK1A-knockdown BALL1 cells (Figure 6C), suggesting a DYRK1a-NIK signaling axis in mediating B-ALL cell survival.

We next examined whether pharmacological inhibition of DYRK1a interferes with B-ALL cell survival. Interestingly, BAFF-induced BALL1 cell survival was largely impaired in the presence of a selective DYRK1a inhibitor, EHT1610,³⁵ causing a drastic increase in cell death (Figure 6D). Treatment of BALL1 cells with EHT1610 also promoted BALL1 cell death under steady-state conditions, albeit less strikingly than under BAFF-induced conditions (Figure 6D). Importantly, ectopic expression of NIK in BALL1 cells rendered these B-ALL cells largely resistant to EHT1610-mediated cell death (Figure 6D). Consistently, EHT1610 potently inhibited BAFF-induced p100 processing, which was rescued by ectopic NIK expression. These results further suggest that DYRK1a promotes B-ALL cell survival via the ncNF-κB signaling pathway.

To further determine the role of the DYRK1a-NIK signaling axis in B-ALL, we used an animal model³⁶ by transplanting lethally irradiated B6 mice with WT or Dyrk1a^BKO bone marrow cells infected with retroviruses expressing the oncogene BCR-ABL and GFP marker gene (supplemental Figure 10A). In this case, we generated the Dyrk1a^BKO mice by crossing Dyrk1a-flox mice with Mb1-Cre mice known to strongly express the Cre recombinase in early stages of bone marrow B cells.²⁸ Mice transplanted with BCR-ABL–transduced WT bone marrow cells developed B-ALL, characterized by early lethality and the presence of a high frequency of GFP⁺ leukemic B cells in the blood, bone marrow, and various other organs, but these B-ALL phenotypes were severely attenuated in the recipient mice of Dyrk1a^BKO bone marrow cells (supplemental Figure 10B-D). Like the human B-ALL cells, these mouse leukemic cells displayed pre–B-cell surface markers (CD19⁺, IgM⁻, and BP-1⁺) and expressed high levels of BAFFR (Figure 7A; supplemental Figure 10E). Interestingly, freshly isolated WT B-ALL cells displayed potent ncNF-κB activation, which was largely impaired in the Dyrk1a^BKO B-ALL cells (Figure 7B). Consistently, the Dyrk1a^BKO B-ALL cells had increased apoptosis and decreased proliferative ability compared with the WT B-ALL cells (Figure 7C-D). These findings suggest that DYRK1a-mediated B-ALL development may involve activation of the ncNF-κB pathway.

To further determine the role of the ncNF-κB pathway in DYRK1a-mediated B-ALL induction, we tested whether transgenic expression of NIK in the Dyrk1a^BKO bone marrow B cells is able to rescue their defect in leukemogenesis. For these studies, we performed B-ALL using BCR-ABL–transduced bone marrow cells derived from the wild-type, Dyrk1a^BKO, and Dyrk1a^BKOMap3k14^Btg mice. While mice transplanted with the Dyrk1a^BKO bone marrow cells displayed improved survival, this phenotype was completely reversed in mice transplanted with the Dyrk1a^BKOMap3k14^Btg bone marrow cells (Figure 7E). To assess the role of the DYRK1a-NIK signaling axis in the survival and expansion of the B-ALL cells, we performed secondary B-ALL by injecting primary B-ALL cells to lethally irradiated B6 mice. As expected,³⁶ induction of secondary B-ALL occurred rapidly, and all mice receiving the WT primary leukemic cells died within 16 days (Figure 7F). Moreover, mice receiving the Dyrk1a^BKO primary leukemic cells had profoundly improved survival and reduced leukemic cell frequency (Figure 7F-G). Importantly, ectopic NIK expression in the Dyrk1a^BKO primary leukemic cells completely restored their ability to efficiently induce secondary B-ALL (Figure 7F-G). Collectively, these findings suggest a critical role for the DYRK1a-NIK signaling axis in regulating B-ALL development.

To assess the molecular mechanism by which DYRK1a-NIK axis regulates the survival of B cells and B-ALL cells, we analyzed the expression of a panel of anti-apoptotic and pro-apoptotic genes from the Bcl2 family by quantitative reverse transcription polymerase chain reaction (qRT-PCR). We identified two anti-apoptotic genes, Bcl2 and Bcl2l1 (Bcl-XL), which was induced by BAFF in splenic B cells in a Dyrk1a-dependent manner (supplemental Figure 11A). Bcl2 and Bcl2l1 were also highly expressed in B-ALL cells (compared with normal bone marrow cells), and the expression of these survival genes was significantly reduced in the Dyrk1a^BKO B-ALL cells (Figure 7H). The protein level expression of Bcl-2 and Bcl-XL was also substantially reduced in the Dyrk1a^BKO B-ALL cells (Figure 7I). Moreover, pharmacological inhibition of DYRK1a also attenuated BAFF-stimulated expression of Bcl-2 and Bcl-XL in the human B-ALL cell line BALL1, which could be rescued by ectopic NIK expression (Figure 7J). By analyzing a human pediatric ALL dataset, we found that the expression level of Bcl-XL, as well as of DYRK1A, is inversely correlated with patient survival (supplemental Figure 11B). These results provide mechanistic insight into the function of DYRK1a-NIK signaling axis in B-ALL cell survival.

Data presented in this paper establish DYRK1a as a kinase that responds to the ncNF-κB inducers BAFF and CD40L and mediates peripheral B-cell survival and maturation. DYRK1a deficiency causes a profound reduction in the mature and T2 immature B-cell populations, which are known to rely on BAFF for survival. We identified DYRK1a as a TRAF3-binding protein required for BAFF- and CD40L-induced ncNF-κB activation and B-cell survival.

It is generally thought that TRAF3 functions as an adaptor recruiting NIK to the cIAP-TRAF E3 ligase complex for ubiquitination and proteolysis.^14,15 However, how the receptor signals might trigger modifications of the cIAP-TRAF components leading to NIK stabilization has remained unclear. Our data suggest that the kinase DYRK1a responds to BAFF and CD40L signals and mediates phosphorylation of TRAF3 at S-29. This residue is conserved across the species and located adjacently to the RING domain of TRAF3 known to be essential for its function in the regulation of ncNF-κB signaling.³⁷ It is thus possible that the DYRK1a-mediated TRAF3 phosphorylation may interfere with its function, thereby facilitating NIK accumulation and ncNF-κB activation.

A previous study suggests that DYRK1a deletion in hematopoietic progenitor cells using the interferon-inducible Mx1-Cre causes a defect in lymphopoiesis, leading to reduced frequency of pre-B cells.²⁹ In support of this previous work, we found that DYRK1a deletion using Mb1-Cre, which efficiently expresses the Cre recombinase in early stages of bone marrow B cells,²⁸ results in reduced frequency of pre-B cells. However, this early-stage B-cell defect was not seen in Dyrk1a^BKO mice generated using the Cd19-Cre, likely because of the predominant Cre recombinase expression in peripheral, but not bone marrow, B cells.²⁸ Using the Cd19-Cre Dyrk1a^BKO mice, we demonstrated a crucial role of DYRK1a in peripheral B-cell maturation and maintenance. Our data suggest that DYRK1a also plays a role in regulating autoimmune B-cell activation and antibody production in the BM12 model of SLE. The requirement of DYRK1a in T-cell–mediated autoimmune B-cell activation is in line with our finding that DYRK1a also mediates ncNF-κB activation by CD40, a crucial receptor involved in T-cell–dependent B-cell activation.³⁸ 

B-ALL represents the most common form of childhood cancer.³⁹ Despite the substantial advancement in B-ALL treatment, it fails in up to 20% of patients, emphasizing the importance for characterizing additional drug targets. Human DYRK1a gene locus amplification is associated with Down syndrome and B-ALL, and a role for DYRK1a in regulating B-ALL has also been suggested in a mouse model study.^20,40,41 Our analysis of a human pediatric ALL database revealed DYRK1a amplification or mRNA overexpression in 10.83% of total cases. Integration of DYRK1a inhibitor into the treatment of DYRK1a^hi ALL population may improve the outcomes. Better understanding of the mechanism of DYRK1a function will create additional therapeutic opportunities. Our present study suggests a crucial role of DYRK1a in mediating the BAFFR survival pathway. Notably, although normal B cells do not express BAFFR in their early stages of development, B-ALL cells are characterized by high levels of BAFFR expression despite pre–B-cell origin.²¹ We found that DYRK1a mediates BAFF-induced ncNF-κB activation and survival of human B-ALL cells. The survival defect of the DYRK1a-knockdown B-ALL cells could be rescued by ectopic expression of NIK. Furthermore, leukemia cells derived from the mouse B-ALL model also express BAFFR, and the defect of Dyrk1a^BKO mice in B-ALL induction could be rescued by transgenic expression of NIK. These findings suggest that mediating ncNF-κB signaling is an important mechanism by which DYRK1a regulates peripheral B-cell survival and B-ALL pathogenesis.

The authors thank Kamon Sanada for DYRK1a expression vectors, as well as the personnel from the flow cytometry, advanced microscopy, DNA analysis, and animal facilities at The University of Texas MD Anderson Cancer Center for their technical assistance.

This study was supported by grants from the National Institutes of Health, National Institute of General Medical Sciences (GM84459). The Baylor College of Medicine Mass Spectrometry Proteomics Core is supported by the Dan L. Duncan Comprehensive Cancer Center NIH Award (P30 CA125123) and CPRIT Core Facility Award (RP170005). T.G. was a visiting student supported by a scholarship from the China Scholarship Council with the grant number of 201906380080.

Contribution: Y.L. designed and performed the research, prepared the figures, and wrote the manuscript; X.X., Z.J., L.Z., J.-Y.Y., C.-J.K., T.G., and X.C. contributed experiments; A.J. and S.Y.J. performed MS and analyzed the MS data; N.B. and M.Y.K. contributed key reagents; and S.-C.S. supervised the work and wrote the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Shao-Cong Sun, The University of Texas MD Anderson Cancer Center, 7455 Fannin St, Unit 902, Houston, TX 77030; e-mail: ssun@mdanderson.org.