Activation of nemo-like kinase in diamond blackfan anemia suppresses early erythropoiesis by preventing mitochondrial biogenesis

Abstract

Diamond Blackfan Anemia (DBA) is a rare macrocytic red blood cell aplasia that usually presents within the first year of life. The vast majority of patients carry a mutation in one of approximately 20 genes that results in ribosomal insufficiency with the most significant clinical manifestations being anemia and a predisposition to cancers. Nemo-like Kinase (NLK) is hyperactivated in the erythroid progenitors of DBA patients and inhibition of this kinase improves erythropoiesis, but how NLK contributes to the pathogenesis of the disease is unknown. Here we report that activated NLK suppresses the critical upregulation of mitochondrial biogenesis required in early erythropoiesis. During normal erythropoiesis, mTORC1 facilitates the translational upregulation of Transcription factor A, mitochondrial (TFAM), and Prohibin 2 (PHB2) to increase mitochondrial biogenesis. In our models of DBA, active NLK phosphorylates the regulatory component of mTORC1, thereby suppressing mTORC1 activity and preventing mTORC1-mediated TFAM and PHB2 upregulation and subsequent mitochondrial biogenesis. Improvement of erythropoiesis that accompanies NLK inhibition is negated when TFAM and PHB2 upregulation is prevented. These data demonstrate that a significant contribution of NLK on the pathogenesis of DBA is through loss of mitochondrial biogenesis.

Keywords: anemia; erythropoiesis; hematopoiesis; mitochondrial biogenesis; pediatric diseases; ribosomopathy.

Figure 1

NLK activation suppresses mTORC1 activity. A, mononuclear cells from the bone marrow aspirates of one healthy donor and 3 DBA patients were differentiated in erythroid media for 5 days. Immunoprecipitated mTORC1 was subjected to in vitro kinase assay for phosphorylation of (i) S6K or (ii) 4E-BP1. B, cord blood CD34+ HSPCs were transduced with shRNA against RPS19 or control and non-targeting or siRNA against NLK and differentiated for 5 days in erythroid media and mTORC1 activity against (i) S6K and (ii) 4E-BP1 was assessed. C, (i) Human CD34+ HSPCs were transduced with shRNA against RPS19 or control and siRNA against NLK or NT. After 5 days of differentiation in the presence or absence of BI-D1870, cells were lysed and subjected to Western blot analysis. (ii) Sibling-matched Rpl11^+/Δ were fed tamoxifen or not for 14 days. CD71+ cells were isolated from blood and subjected to Western blot analysis. D, CD34+ HSPCs expressing control or shRPS19 were differentiated for the indicated days and (i) NLK and (ii) mTORC1 were immunoprecipitated. In vitro phosphorylation of Raptor by NLK, or 4E-BP1 by mTORC1, was analyzed. Data are represented as mean ± SD and significance is defined as p < 0.05 (n = 3). See also Fig. S1.

Figure 2

NLK activation suppresses the translation of a subset of TOP-containing transcripts.A, differentiating HSPCs expressing control or RPS19 were incubated with OPP for 24 h at indicated times, and fluorescence at 494/521 nm was determined by flow cytometry. B, control or shRPS19 and non-targeting or siNLK expressing HSPCs were differentiated for 5 days. Cell lysates were separated by ultracentrifugation on a sucrose gradient and the abundance of 34 ribosome insufficiency-sensitive mRNAs sequences in polysome fractions were quantitated by qRT-PCR and normalized to total cellular fractions. To extrapolate the rescue effect, the variance between control and shRPS19 samples was calculated and designated 100%. Values from samples expressing both shRPS19 and siNLK were compared with 0% representing the same value as shRPS19 and 100% representing the same value as the control. Transcripts were segregated based on the possession of a TOP sequence in the 5′UTR or not and compared by student t test. The NLK translational rescue value of individual TOP-containing (C) and transcripts not containing a TOP sequence (D) are plotted. Three individual repeats were performed in triplicate. Data are represented as mean ± SD of the triplicate means and significance is defined as p < 0.05 (n = 3). See also Figs. S2 and S3.

Figure 3

NLK activation impacts translation of mitochondrial biogenesis transcripts that are upregulated during erythroid differentiation.A, CD34+ HSPCs were transduced with shRPS19 with or without siNLK and differentiated for 5 days. Total cellular mRNA for 12 mitochondrial genes that are impacted by ribosome insufficiency was assessed by qRT-PCR. B, after differentiation, mRNA expression in polysome fractions was assessed by qRT-PCR after sucrose gradient ultracentrifugation. The effect of NLK knockdown on transcript expression in RPS19 insufficient samples compared to controls was calculated. Transcripts in red are rescued by >50% and in blue by >20%. C, NLK rescue values for mitochondrial transcripts were segregated into those containing TOP sequences in the 5′UTR or not. D, the translation of mitochondrial transcripts in ribosome insufficiency was plotted again the nucleotide length of the 5′UTR. CD34+ HSPCs were transduced with control or shRPS19 and differentiated for 2, 5, or 8 days. Total cellular mRNA (E) and polysome-associated mRNA (F) were determined as above. Three individual repeats were performed in triplicate. Data are represented as mean ± SD of the triplicate means and significance is defined as p < 0.05 (n = 3). See also Figs. S4 and S5.

Figure 4

Mitochondrial biogenesis is not upregulated in ribosome insufficiency due to NLK activity. A, cord blood CD34+ HSPCs were transduced with shRPS19 or control and differentiated for two or 5 days in the presence of erythroid media. At indicated times, mtDNA was assessed by qRT-PCR (i), mitochondrial mass (ii) was assessed by flow cytometry analysis of mitotracker green uptake, and intracellular ATP (iii) was determined by CellTiter-Glo Luminescent Assay. B, Lin-Kit+ HSPCs from Rpl11^+/+ or Rpl11^+/Δ mice were differentiated in erythroid media for two or 5 days and assessed for mtDNA, mitochondrial mass (ii) and intracellular ATP (iii). C, mononuclear cells from bone aspirates of a control and DBA patient were differentiated for 2 and 5 days prior to mtDNA (i), mitochondrial mass (ii) and intracellular ATP (iii) analysis. D, cord blood CD34+ HSPCs were transduced with shRNA against RPS19 or control and siRNA against NLK or non-targeting prior to differentiation for indicated times and assessment for mtDNA (i), mitochondrial mass (ii) and intracellular ATP (iii). E, human cord blood CD34+ HSPCs transduced with control or shRNA against RPS19 were differentiated in the presence or absence of SRT2104 and assessed for mtDNA (i), mitochondrial mass (ii) intracellular ATP (iii) and expansion of CD235+ erythrocytes (iv). Three individual repeats were performed in triplicate. Data are represented as mean ± SD of the triplicate means and significance is defined as p < 0.05 (n = 3). See also Figs. S6 and S7.

Figure 5

Mitophagy cooperates with translation defects in suppression of mitochondrial output downstream of NLK and mTORC1 in ribosome insufficiency. A, cord blood CD34+ HSPCs were transduced with control or shRNA against RPS19 with Keima-tagged TOMM20 (i) or ACTB (ii) alone, or TOMM20 (iii) and ACTB (iv), in the presence of torin, GSK3-IN-3, liensinine or siRNA against NLK. The ratio of free to lysosome-associated Keima was determined by comparing excitation at 488 and 561 nm respectively. B, cord blood CD34+ HSPCs transduced with control or shRNA against RPS19 were differentiated in the presence or absence of GSK3-IN-3 (upper), liensinine (lower), and assessed for mtDNA (far left), mitochondrial mass (middle left) intracellular ATP (middle right) and expansion of CD235+ erythrocytes (far right). C, cord blood CD34+ HSPCs were transduced with shRNA against RPS19 and either siRNA against NLK or control in the presence of Torin (upper), co-transduction with shRNA against TFAM and PHB2 (middle) or GSK3-IN-3 (lower). mtDNA (far left), mitochondrial mass (middle left), intracellular ATP (middle right), and expansion of CD235+ erythrocytes (far right) were assessed. Three individual repeats were performed in triplicate. Data are represented as mean ± SD of the triplicate means and significance is defined as p < 0.05 (n = 3 or 5). See also Fig. S8.

Figure 6

Proposed model of how NLK activation in ribosome insufficiency contributes to failed erythropoiesis in DBA. Proposed model of how NLK activation in ribosome insufficiency contributes to failed erythropoiesis in DBA. Activated NLK phosphorylates and suppresses mTORC1. The reduction in mTORC1 activity induces mitophagy and reduces translation of factors that induce mitochondrial biogenesis. Collectively, the mitochondrial output cannot be upregulated at day 5 and erythropoiesis is ineffective.