Metallothionein 1 mediates growth and survival of Dnmt3a;Npm1-mutant acute myeloid leukemia

Acute myeloid leukemia (AML) is an aggressive malignancy associated with poor outcomes.¹ Deregulation of transcriptional programs resulting in malignant cell self-renewal and impaired differentiation is a molecular hallmark of AML.² Thus, targeting transcriptional dependencies of AML is a promising approach to impair cell-intrinsic mechanisms sustaining this malignancy. Here, we investigated transcriptional patterns across cell states utilizing a sequentially inducible model of AML³ combining two of the most common and frequently co-mutated genes in cytogenetically normal human de novo AML; DNA methyltransferase 3A (DNMT3A) and nucleophosmin (NPM1).⁴

We performed single-cell RNA sequencing (RNA-seq) and cellHarmony analysis⁵ of primitive c-Kit⁺ cells enriched using magnetic-activated cell sorting from four primary Dnmt3a^R878H/+ Npm1^cA/+ (Dnmt3a;Npm1-mutant) AML spleen samples³ (Figure 1A) (GSE277963). All mouse studies were approved by The Jackson Laboratory’s Institutional Animal Care and Use Committee (IACUC). We integrated this data with a healthy hematopoietic cell atlas of 87 cell states spanning all major bone marrow (BM) lineages. This atlas includes rare and transitional cell states (Online Supplementary Figure S1A). As this atlas excluded predominant hematopoietic and stromal cell populations in the spleen, we included distinct cell states identified in spleen single-cell RNA-seq for the annotation of our datasets (Online Supplementary Figure S1B, C). The most primitive cell populations found in all Dnmt3a;Npm1-mutant AML samples were multi-lineage primed MultiLin-2 (ML-2) progenitors (Online Supplementary Figure S1D). Using differential gene expression analysis of Dnmt3a;Npm1-mutant AML progenitors compared to their normal BM counterparts, the transcript Mt1 was the top gene significantly increased in expression across all samples (Figure 1B; Online Supplementary Figure S2). We compared Mt1 expression in MultiLin-2 cells profiled from eight different mouse models harboring pre-leukemic mutations,⁶ wild-type (WT) BM controls, and Dnmt3a;Npm1-mutant AML mice from this study (2-4 biological replicates per genotype). We observed that increased expression of Mt1 was highly specific to Dnmt3a;Npm1-mutant AML and is not induced by individual mutations in Dnmt3a or Npm1 alone (Figure 1C, D).

Metallothionein 1 (Mt1) is a low molecular weight, cysteine-rich intracellular protein that can bind to both essential and toxic metals.⁷ To identify potential role(s) of Mt1 in Dnmt3a;Npm1-mutant AML, we performed knockout of Mt1 by electroporation of caspase 9 single-guide RNA (Cas9-sgRNA) complexes (RNP) into primary Dnmt3a;Npm1-mutant AML cells. We began by pooling three sgRNA targeting the Mt1 locus (Figure 2A). With this approach, >80% of cells were confirmed by sequencing to have an insertion/deletion (INDEL) at the targeted region of the Mt1 locus (Figure 2B). We performed differential expression analysis using RNA-seq from three independent replicates compared to control Dnmt3a;Npm1-mutant AML cells electroporated with Cas9 but no sgRNA (GSE262264). From this analysis, 256 genes were increased (ex. Cpa3, Ctla2a, Tpsb2, Cma1) and 89 genes decreased (ex. Mt1, Ccn2, Aebp1, Col5a2) with a significance cutoff of P<0.05 (Figure 2C). Gene set enrichment analysis revealed that knockout of Mt1 reduced the expression of cell cycling signatures and increased expression of genes signatures related to differentiation, pyroptosis, and oxidative stress (Figure 2D). Based on these data, we hypothesized that knockout of Mt1 would promote cell death and reduced cell cycling of primary Dnmt3a;Npm1-mutant AML cells. To test this, we evaluated cell proliferation and growth of Mt1 knockout Dnmt3a;Npm1-mutant AML cells relative to an expanded set of negative control conditions including no sgRNA, scrambled sgRNA, and sgRNA targeting of the non-essential Rosa26 locus (Figure 2E, F). Compared to these negative controls, knockout of Mt1 in primary Dnmt3a;Npm1-mutant AML cells resulted in decreased proliferation assessed by Ki67 staining (Figure 2G) and reduced growth in 5-day cultures (Figure 2H). To assess the rigor of sgRNA-mediated knockout of Mt1, we tested individual sgRNA which each achieved ~10-40% knockout efficiency (Figure 2I). These individual sgRNA were sufficient to impair growth of Dnmt3a;Npm1-mutant AML cells in 5-day cultures to a level comparable to knockout of Pcna (Figure 2J). We utilized Pcna as a positive control which is an essential component of the DNA replication machinery known to be a common cellular dependency.

To evaluate how knockout of Mt1 impacts leukemogenicity in vivo, sublethally irradiated recipient mice (CD45.1⁺) were transplanted with Mt1-knockout or control primary Dnmt3a;Npm1-mutant AML cells (CD45.2⁺) (Figure 3A). At 4 weeks post-transplant, we found significantly fewer AML (CD45.2⁺) cells detected in the recipients of Mt1-knockout AML compared to negative controls (Figure 3B). Mice receiving Mt1-knockout AML cells also survived significantly longer than mice transplanted with control AML cells (Figure 3C). In moribund recipient mice, the proportion of AML cells retaining Mt1 knockout was variable (Figure 3D), and survival of individual recipient mice positively correlated with the proportion of Mt1-knockout AML cells remaining (Figure 3E), demonstrating that knockout of Mt1 reduced the leukemogenicity of Dnmt3a;Npm1-mutant AML in vivo. In humans, there are eight functional MT1 isoforms and the transcript levels of MT1F, MT1G, MT1H and MT1X have been shown to be increased in human AML compared to normal BM samples in the ONCOMINE database.⁸ In the DNMT3A;NPM1-mutant human AML cell line, OCI-AML3, we observed significantly increased expression of the isoform MT1G compared to normal BM cells and the human DNMT3A-mutant NPM1-WT AML cell line OCI-AML2 (Figure 3F). We performed knockout of MT1G in OCI-AML2 and OCI-AML3 cells by electroporation of Cas9-RNP using a sgRNA targeting exon 1, achieving a similar knockout efficiency in both cell lines (Figure 3G). In OCI-AML2 cells, knockout of MT1G did not impact cell growth in culture unlike the positive control PCNA which significantly decreased viability (Figure 3H). In contrast, in OCI-AML3 cells, knockout of MT1G decreased cell counts in culture to a similar level as the positive control PCNA (Figure 3I). These data support that in human AML cells, similar to what we see in mouse Dnmt3a;Npm1-mutant AML cells, increased expression and dependency of MT1G occurs in DNMT3A;NPM1-mutant AML which is not induced by mutation in DNMT3A alone.

Figure 1.Mt1 is a top upregulated gene in mouse Dnmt3a;Npm1-mutant acute myeloid leukemia progenitor cells. (A) Experimental design for single-cell RNA sequencing (RNA-seq). Created in BioRender. Trowbridge, J. (2025) https://BioRender.com/nzoaehw. (B) Volcano plots of differential gene expression in Multi-Lin-2 cells from Dnmt3a;Npm1-mutant acute myeloid leukemia (AML) versus CellHarmony reference. (C) Violin plot of Mt1 expression in Multi-Lin-2 cells from 8 previously reported mouse models harboring pre-leukemic mutations,⁶ control wild-type (WT) bone marrow (BM) and Dnmt3a;Npm1-mutant AML mice from this study (2-4 biological replicates). Individual dots are single-cell log-transformed normalized expression values. (D) Dotplot of fold change in Mt1 expression (log2) in Multi-Lin-2 cells. The size and color of each dot corresponds to the -log10 P value of a Welch’s t test (two-sided, unequal variance assumption) comparing log-transformed normalized expression values in mutants versus WT control BM. AML: acute myeloid leukemia; KO: knockout; het: heterozygous; ITD: internal tandem duplication; hom: homozygous.

Figure 2.Knockout of Mt1 impairs cell growth of mouse Dnmt3a;Npm1-mutant acute myeloid leukemia cells. (A) Experimental design for caspase 9 single-guide RNA (Cas9-sgRNA)-mediated Mt1 knockout. Created in BioRender. Trowbridge, J. (2025) https://BioRender. com/hobbbyh. (B) Proportion of Dnmt3a;Npm1-mutant acute myeloid leukemia (AML) cells with insertions/deletions (INDEL) detected at the Mt1 locus. Each dot represents a biological replicate. (C) Volcano plot showing significantly differentially expressed genes in Mt1 knockout Dnmt3a;Npm1-mutant AML cells versus control Dnmt3a;Npm1-mutant AML cells. N=3 biological replicates per condition. Mt1 is denoted in red. (D) Normalized enrichment scores of gene signatures enriched in Mt1 knockout compared to control Dnmt3a;Npm1-mutant AML cells. (E, F) Proportion of AML cells with indels detected at the (E) Rosa26 or (F) Mt1 locus. Each dot represents a biological replicate. (G) Frequency of Ki67⁺ cells in control and Mt1 knockout Dnmt3a;Npm1-mutant AML cells. Dots represent individual biological replicates (N=3). **P<0.01; ***P<0.001 by one-way ANOVA with Tukey’s multiple comparisons test. (H) Fold change in viable cell count after 5 day culture of control and Mt1 knockout Dnmt3a;Npm1-mutant AML cells. Dots represent individual biological replicates (N=4-6). **P<0.01; ***P<0.001 by one-way ANOVA with Dunnett’s multiple comparisons test. (I) Proportion of AML cells with indels detected at the Mt1 locus. (J) Fold change in viable cell count after 5 day culture of control and Mt1 knockout Dnmt3a;Npm1-mutant AML cells. Dots represent individual biological replicates (N=3). *P<0.05, ****P<0.0001 by one-way ANOVA with Dunnett’s multiple comparisons test. -ve: positive; cntl: control; RNP: ribonucleoprotein complex.

Figure 3.Knockout of Mt1 reduces growth of mouse Dnmt3a;Npm1-mutant acute myeloid leukemia cells in vivo and the human DNMT3A;NPM1-mutant acute myeloid leukemia cell line OCI-AML3. (A) Experimental design for in vivo assays. Created in BioRender. Trowbridge, J. (2025) https://BioRender.com/viw75wf. (B) Frequency of control or Mt1 knockout Dnmt3a;Npm1-mutant acute myeloid leukemia (AML) (CD45.2⁺) cells in the peripheral blood of transplanted recipient mice at 4 weeks post-transplant. Each dot represents 1 recipient mouse (N=5). *P<0.05; **P<0.01 by one-way ANOVA with Fisher’s LSD. (C) Survival curve of mice transplanted with control and Mt1 knockout Dnmt3a;Npm1-mutant AML cells (N=5 per condition). **P<0.01 by log-rank (Mantel-Cox) test. (D) Proportion of AML cells at mouse harvest with insertions/deletions (INDEL) detected at the Mt1 locus. Each dot represents an individual mouse. (E) Correlation between mouse survival and Mt1 knockout detected at harvest for mice transplanted with Mt1 knockout Dnmt3a;Npm1-mutant AML cells. Each dot represents an individual mouse. (F) Relative expression of MT1G in OCI-AML2 and OCI-AML3 cells relative to wild-type BM CD34⁺ hematopoietic stem and progenitor cells. Dots represent biological replicates (N=3). ***P<0.001 by one-way ANOVA with Tukey’s multiple comparisons test. (G) Proportion of AML cells with INDEL detected at the Mt1G locus. (H, I) Viable cell counts after 3-day culture of (H) OCI-AML2 and (I) OCI-AML3 cells with MT1G knockout compared to controls. Dots represent independent replicate experiments (N=3). **P<0.01; ***P<0.001 by one-way ANOVA with Dunnett’s multiple comparisons test. +ve: positive; cntl: control; sgRNA: single-guide RNA; -ve: negative; RNP: ribonucleoprotein complex.

In mouse hematopoietic stem cells (HSC), the Mt1 locus is methylated by Dnmt3a as evidenced by Mt1 hypomethylation in Dnmt3a-knockout HSC using whole-genome bisulfite sequencing.^9,10 However, Mt1 is not increased in expression in Dnmt3a-knockout or mutant HSC,^9,10 indicating that cooperativity with an Npm1 mutation and transformation to AML are prerequisites for observing transcriptional upregulation of Mt1. Understanding the function of Mt1 in AML progenitor cells requires further study. It is known that metallothionein protects cells against radiation and chemotherapeutic agents by virtue of its free radical scavenging property.^11,12 In newly diagnosed pediatric AML, metallothionein expression is associated with increased risk of relapse.¹³ Our study expands on this literature by demonstrating that mouse and human Dnmt3a;Npm1-mutant AML is sensitive to loss of metallothionein expression. A current limitation is that the only mouse model available is a dual Mt1 and Mt2 germline knockout¹⁴ which is a barrier to ascertaining the role of Mt1 in normal HSC, and no small molecule inhibitors targeting MT1 or MT1G are available to our knowledge. Therefore, development of new tools and methodologies are needed to better understand the role of metal toxicity and oxidative stress in AML risk and relapse.

Footnotes

• Received February 21, 2025
• Accepted August 4, 2025

Correspondence

J.J. TROWBRIDGE - Jennifer.trowbridge@jax.org

Disclosures

RL is on the supervisory board of Qiagen, a co-founder/board member at Ajax, and is a scientific advisor to Mission Bio, Kurome, Anovia, Bakx, Syndax, Scorpion, Zentalis, Auron, Prelude, and C4 Therapeutics; for each of these entities he receives equity/ compensation. He has received research support from the Cure Breast Cance Foundation, Calico, Zentalis and Ajax, and has consulted for Jubilant, Goldman Sachs, Incyte, Astra Zeneca and Janssen. JJT has previously received research support from H3 Biomedicine, Inc., and patent royalties from Fate Therapeutics. All other authors have no conflicts of interest to disclose.

Contributions

Funding

This work was supported by U01AG077925 (to JJT and RLL), R01DK118072 (to JJT), R01AG069010 (to JJT), P30CA034196 (to JJT), RC2DK122376 (to HLG), R01CA284595 (to HLG), R01HL122661 (to HLG), R01CA253981 (to HLG), and the Edward P. Evans Foundation (to JJT). JJT was supported by a Leukemia and Lymphoma Society Scholar award. JJM was supported by a Leukemia & Lymphoma Society Career Development Program Fellow Award and The Jackson Laboratory Scholar Award.

Acknowledgments

The authors thank all members of the Trowbridge laboratory for experimental assistance, input, and critical discussions, particularly Maria Telpoukhovskaia, Griffin Nye and Xiurong Cai.