Loss of Dnmt3a increases self-renewal and resistance to pegIFN-α in JAK2-V617F-positive myeloproliferative neoplasms

Abstract

Pegylated interferon alfa (pegIFN-α) can induce molecular remissions in patients with JAK2-V617F-positive myeloproliferative neoplasms (MPNs) by targeting long-term hematopoietic stem cells (LT-HSCs). Additional somatic mutations in genes regulating LT-HSC self-renewal, such as DNMT3A, have been reported to have poorer responses to pegIFN-α. We investigated whether DNMT3A loss leads to alterations in JAK2-V617F LT-HSC functions conferring resistance to pegIFN-α treatment in a mouse model of MPN and in hematopoietic progenitors from patients with MPN. Long-term treatment with pegIFN-α normalized blood parameters and reduced splenomegaly and JAK2-V617F chimerism in single-mutant JAK2-V617F (VF) mice. However, pegIFN-α in VF;Dnmt3aΔ/Δ (VF;DmΔ/Δ) mice worsened splenomegaly and failed to reduce JAK2-V617F chimerism. Furthermore, LT-HSCs from VF;DmΔ/Δ mice compared with VF were less prone to accumulate DNA damage and exit dormancy upon pegIFN-α treatment. RNA sequencing showed that IFN-α induced stronger upregulation of inflammatory pathways in LT-HSCs from VF;DmΔ/Δ than from VF mice, indicating that the resistance of VF;DmΔ/Δ LT-HSC was not due to failure in IFN-α signaling. Transplantations of bone marrow from pegIFN-α-treated VF;DmΔ/Δ mice gave rise to more aggressive disease in secondary and tertiary recipients. Liquid cultures of hematopoietic progenitors from patients with MPN with JAK2-V617F and DNMT3A mutation showed increased percentages of JAK2-V617F-positive colonies upon IFN-α exposure, whereas in patients with JAK2-V617F alone, the percentages of JAK2-V617F-positive colonies decreased or remained unchanged. PegIFN-α combined with 5-azacytidine only partially overcame resistance in VF;DmΔ/Δ mice. However, this combination strongly decreased the JAK2-mutant allele burden in mice carrying VF mutation only, showing potential to inflict substantial damage preferentially to the JAK2-mutant clone.

Graphical abstract

Figure 1.

Disease phenotype characterization of JAK2-V617F (VF) and JAK2-V617F;Dnmt3a^Δ/Δ (VF;Dm^Δ/Δ) mice. (A) Schematic drawing of induction with tamoxifen injections for 4 weeks (red box) and experimental procedures. (B) Analysis of Cre-mediated excision of Dnmt3a. Left panel shows representative gel from Bioanalyzer DNA chip, with bands corresponding to floxed (fl) and deleted (Δ) Dnmt3a alleles. Right panel shows quantification of fl and Δ Dnmt3a alleles (n = 4 per genotype). (C) Time course of nonfasting blood glucose levels (n = 5-7 mice per genotype). (D) Time course of peripheral blood counts (n = 5-7 mice per genotype). (E) Spleen weight at terminal workup after 16 weeks postinduction. (F) BM cellularity per 4 bones (n = 4-5 mice per genotype). (G) Frequencies of HSPCs in BM and spleen at terminal workup after 16 weeks of treatment (n = 4-5 mice per genotype). (H) Analysis of erythroid progenitor’s frequencies in BM and spleen at terminal analysis. (I) Quantification BM fibrosis and osteosclerosis (n = 4-5 mice per genotype). The degree of myelofibrosis was scored and assigned on a scale from MF-0 to MF-3. Osteosclerosis was scored as present or absent. All data are presented as mean ± standard error of the mean. Two-way analyses of variance (ANOVA) with subsequent Tukey (B-C) and Dunnett posttest (G; erythroid progenitors), 1-way ANOVA with subsequent Tukey posttest (D,E,G; Ter119⁺ cells, H) or unpaired t test with Welch correction (F) were used. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001.

Figure 2.

Differential effects of pegylated IFN-α on hematopoiesis in JAK2-V617F (VF) and JAK2-V617F;Dnmt3a^Δ/Δ (VF;Dm^Δ/Δ) mice. (A) Schematic drawing of the experimental setup for BM transplantations and treatment with pegIFN-α. (B) Time course of body weight (n = 24 mice per group). (C) Time course of blood counts and GFP chimerism of recipient mice (n = 24 mice per group). (D) Spleen weight at terminal workup after 16 weeks of treatment. (E) Frequency and GFP chimerism of VF;GFP or VF;Dm^Δ/Δ;GFP HSPCs in BM at terminal workup after 16 weeks of treatment (n = 15 mice per group). (F) Frequency and GFP chimerism of VF;GFP or VF;Dm^Δ/Δ;GFP HSPCs in spleen at terminal workup after 16 weeks of treatment (n = 15 mice per group). (G) Analysis of the proportions of CD41^hi and CD41^lo subpopulations within the mutant (GFP-positive) LT-HSCs. All data are presented as mean ± standard error of the mean. ANOVA with subsequent Tukey (C,E,F,G) posttest or 1-way ANOVA with subsequent Tukey (D) posttest were used. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001. CBC, complete blood count.

Figure 3.

Effects of IFN-α treatment on accumulation of DNA damage and ROS and on cell division in HSPCs. (A) Accumulation of ROS and presence of γH2AX in lin^–/Sca1⁺/Kit⁺ (LSK) cells from BM and spleen of recipient mice. GFP chimerism in LSK cells, gating strategy, and detection of DNA damage by anti-γH2AX antibody and ROS by staining with CellRox (left). Quantification of γH2AX and ROS in 9 mice per group (right). (B) Analysis of cell division (Ki67-positive cells) and DNA damage within the mutant (GFP-positive) LT-HSCs; LT-HSC chimerism and gating strategy (left); and the results obtained from 9 mice per group (right). (C) Analysis of ROS levels in GFP-positive LT-HSCs. These data were obtained from mice described in supplemental Figure 13. All data are presented as mean ± standard error of the mean. ANOVA with subsequent Tukey posttest was used. ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001.

Figure 4.

Transplantations of BM cells from primary recipients into secondary recipients. (A) Schematic drawing of the experimental setup for noncompetitive BM transplantations into secondary recipient mice (n = 6 mice per group). (B) Percentages of mice that showed engraftment defined as >1% GFP chimerism in Gr1⁺ cells in peripheral blood at 20 weeks after transplantation. (C) Spleen weight at terminal workup 20 weeks after transplantation. (D) Time course of blood counts and GFP chimerism in secondary recipients. (E) GFP chimerism in LT-HSCs at terminal workup 20 weeks after transplantation. (F) Schematic drawing of the experimental setup for competitive BM transplantations into secondary recipients at 1:50 dilution (n = 12 mice per group). (G-J) Same annotations as in panels B-E. ANOVA with subsequent Tukey (D,I) posttest or 1-way ANOVA with subsequent Tukey (E, H) posttest were used. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001.

Figure 5.

Effects of IFN-α on HSPCs from patients with MPN carrying JAK2-V617F and DNMT3A mutations. (A) Experimental setup for liquid cultures of single-cell FACS sorted HSPCs. (B) Growth efficiency of single-cell sorted HSPCs expressed as the percentages of wells in which colonies grew. (C) Genotyping of HSPC–derived colonies treated with IFN-α or vehicle. Stacked bars represent the percentages of colonies for each of the genotypes. Numbers above the bars indicate the numbers of colonies analyzed. (D) Relative change (%) in the numbers of JAK2-V617F-positive colonies. ANOVA with subsequent Sidak (D) posttest were used. FACS, fluorescence-activated cell sorting.

Figure 6.

Differential effects of IFN-α treatment on VF and VF;Dm^Δ/Δ LT-HSCs. (A) Experimental setup for single-cell RNA seq of FACS sorted LT-HSCs. (B) Expression of selected lineage and cycling genes in reduced dimension plots. TSNE, t-distributed stochastic neighbor embedding. (C) Clustering of cells is based on the gene expression in reduced dimension plot (D) Shifts in cell identity induced by IFN-α treatment in reduced dimension plots (left); and the relative cell abundance per cell type across genotypes and treatments (right). (E) Heat map analysis of IFN response genes expression (Reactome R-MMU-91353) in quiescent LT-HSCs. Normalized expression of genes in pseudobulk samples was plotted. (F) Gene set enrichment analysis of Hallmark gene sets comparing IFN-α–treated VF and VF;Dm^Δ/Δ quiescent LT-HSCs (qLT-HSCs). (G) Fold change in the normalized expression of Cxcl9 and Cxcl10 genes derived from number of reads (counts per million) in pseudobulk analysis. Expression in vehicle–treated VF cells was set to 1. (H) PROGENy analysis in quiescent LT-HSCs. ANOVA with subsequent Tukey posttest was used. ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001. FACS, fluorescence-activated cell sorting.

Figure 7.

Treatment with combination of Aza or At with pegIFN-α. (A) Schematic drawing of the experimental setup for BM transplantations and treatment. (B) Time course of changes in body weight (%); n = 6-8 per group. (C) Spleen weight at terminal workup after 12 weeks of treatment (n = 5-8 mice per group). (D) Liver weight at terminal workup (n = 5-8 mice per group). (E) Time course of blood counts and GFP chimerism of recipient mice (n = 6-8 mice per group). (F) Frequencies and GFP chimerism of HSCs (LT-HSCs) in BM and spleen at terminal workup after 12 weeks of treatment (n = 5-8 mice per group). ANOVA with subsequent Tukey posttest was used. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001.