Multiple myeloma–associated DIS3 gene is essential for hematopoiesis, but loss of DIS3 is insufficient for myelomagenesis Free

TO THE EDITOR:

Multiple myeloma (MM) is a plasma cell neoplasm that remains incurable despite the introduction of novel agents. Therefore, it is necessary to decode the molecular mechanisms of myelomagenesis. Recent advances in next-generation sequencing technologies have highlighted the significance of known events that drive MM, and have also identified novel genetic alterations in MM.¹ Importantly, DIS3 mutations have been identified in ∼10% of patients with MM, and loss of heterozygosity at chromosome 13q that leads to deletion of 1 allele of DIS3 has been observed in ∼40% of patients with MM.^1,2 Furthermore, DIS3 mutations/deletions are associated with poor prognosis in MM.² DIS3 is a member of the RNase II/R family. It is a catalytic subunit of RNA exosome that mediates degradation and processing of various RNAs.³ Recent studies have shown that Dis3 is required for immunoglobulin gene rearrangements, indicating its role in B cells.^4,5 However, the roles of DIS3 in hematopoiesis and myelomagenesis remain incompletely understood.

To explore the functions of Dis3 in hematopoiesis, we first generated C57BL/6 mice carrying the floxed allele of Dis3 by introducing a pair of loxP sequences into introns 2 and 4 of the Dis3 gene in fertilized eggs using a clustered regularly interspaced short palindrome repeats (CRISPR)-Cas9 system (supplemental Figure 1A). We confirmed insertion of loxP sites by polymerase chain reaction genotyping and Sanger sequencing (supplemental Figure 1B-C). We then set out to generate hematopoietic cell–specific Dis3 knockout (KO) mice by crossing Dis3^fl/fl mice and Mx-Cre mice. To avoid possible effects of Dis3 deficiency on nonhematopoietic cells, we transplanted bone marrow (BM) cells from Dis3^fl/fl and Dis3^fl/fl;Mx-Cre mice into lethally irradiated CD45.1⁺ recipient mice. Four weeks after transplantation, we deleted Dis3 by inducing Cre recombinase through an intraperitoneal injection of polyinosinic-polycytidylic acid (pIpC). We hereafter referred to the recipient mice recovered with the BM cells from Dis3^fl/fl and Dis3^fl/fl; Mx-Cre mice as wild type (WT) and Dis3^Δ/Δ_Mx mice, respectively (Figure 1A). We confirmed generation of the Dis3 KO allele by genomic polymerase chain reaction (Figure 1B) and validated reduced Dis3 messenger RNA expression in the BM cells of Dis3^Δ/Δ_Mx mice (Figure 1C). Of note, loss of Dis3 in hematopoietic cells resulted in severe pancytopenia at 2 weeks after Dis3 deletion (Figure 1D). Consistently, Dis3^Δ/Δ_Mx mice exhibited markedly reduced BM cell counts with relative preservation of mature lymphocytes (Figure 1E-F). Detailed flow cytometric analysis of the BM cells revealed lower frequency of Lineage⁻Sca-1⁺c-Kit⁺ (LSK) cells, hematopoietic stem cells (HSCs: CD150⁺CD48⁻CD135⁻ LSK), and multipotent progenitors (MPPs) in Dis3^Δ/Δ_Mx mice than in WT mice (Figure 1G). The frequency of the 3 MPP populations (MPP^{Megakaryocyte (Mk)/Erythroid (E)}, MPP^{Granulocyte (G)/Monocyte (M)}, and MPP^{Lymphoid (Ly)}) that retain multilineage potential despite undergoing specific lineage priming⁶ was also significantly lower in Dis3^Δ/Δ_Mx mice than in WT mice (Figure 1G). Dis3^Δ/Δ_Mx mice also exhibited a significant decrease in the frequency of common myeloid progenitor (CMP), granulocyte/macrophage progenitor (GMP), and megakaryocyte/erythroid progenitor (MEP) cells, compared with that in WT mice (Figure 1H). In contrast, the frequency of common lymphoid progenitor (CLP) cells was comparable between WT and Dis3^Δ/Δ_Mx mice (Figure 1H), whereas the absolute number of CLP cells was significantly reduced in Dis3^Δ/Δ_Mx mice compared with in WT mice (Figure 1I). These results indicate that Dis3 is essential for normal hematopoiesis.

To investigate the molecular mechanisms of hematopoietic failure after Dis3 depletion, we isolated 100 HSCs from WT and Dis3^Δ/Δ_Mx mice and performed random displacement amplification sequencing analysis. We isolated HSCs on day 4 after the mice were administered 2 doses of pIpC on days 0 and 2, because the frequency of HSCs in Dis3^Δ/Δ_Mx mice markedly reduced within a week after pIpC injection. Cell surface expression of Sca-1 is perturbed immediately after pIpC treatment.⁷ Therefore, we used the L⁻ESLAM (Lineage⁻EPCR⁺CD150⁺CD48⁻) definition for HSC sorting (supplemental Figure 2A),⁸ and validated proper isolation of HSCs by confirming the expression of HSC signature genes (supplemental Figure 2B).⁹ Under this condition, Cre recombination was insufficient, but Dis3 expression was significantly reduced in the HSCs of Dis3^Δ/Δ_Mx mice (∼50% reduction compared with that in WT mice; supplemental Figure 2B). Consistent with the previous finding,^4,5 we observed accumulation of various noncoding RNAs in Dis3-deficient cells (supplemental Figure 2C). A total of 535 genes were upregulated, whereas 304 genes were downregulated in the HSCs of Dis3^Δ/Δ_Mx mice compared with in HSCs from WT mice (cutoff of >1.5-fold change with false discovery rate of <0.05; Figure 2A).¹⁰ Pathway analysis revealed that genes associated with DNA damage response were enriched in upregulated genes (Figure 2B; supplemental Figure 2D). To further explore the significance of gene expression changes, we performed gene set enrichment analysis. Consistent with the previous reports that Dis3 is involved in B-cell differentiation,^4,5 Dis3-deficient cells significantly increased the expression of genes that were upregulated in Lin⁻ cells compared with that in B cells (Figure 2C). Interestingly, Dis3-deficient cells also increased expression of advanced MM signature genes (Figure 2C). Furthermore, the genes related to TP53 pathway and DNA damage response were induced in Dis3-deficient cells, confirming the result of pathway analysis (Figure 2C-D). Indeed, the BM cells isolated from Dis3^Δ/Δ_Mx mice had DNA damage, as evidenced by the elevated γH2AX (Figure 2E), in agreement with the molecular phenotype of Dis3-deficient cancer cell lines.¹¹ Furthermore, we found an increase of cleaved caspase-3– and annexin V–positive cells in the BM of Dis3^Δ/Δ_Mx mice (Figure 2E-F). These results indicate that loss of Dis3 induces DNA damage, leading to apoptosis of hematopoietic cells.

Finally, we examined whether loss of Dis3 in late B cells drives MM. To do so, we crossed Dis3^fl/wt or Dis3^fl/fl mice with Cγ1-Cre mice in which Cre recombinase is activated in the germinal center B cells on immunization.¹² We immunized 8- to 12-week-old Dis3^wt/wt;Cγ1-Cre, Dis3^fl/wt;Cγ1-Cre, and Dis3^fl/fl;Cγ1-Cre mice with NP-chicken γ globulin (NP-CGG), and rechallenged with NP-CGG, 4 weeks after primary immunization (supplemental Figure 3A). We confirmed generation of KO allele in the BM and spleen cells isolated from Dis3^fl/wt;Cγ1-Cre, and Dis3^fl/fl;Cγ1-Cre mice (supplemental Figure 3B-C). In the clinical samples of patients with MM, DIS3 mutations significantly coexist with t(14;16).¹ Thus, we also generated Dis3^fl/wt;Cγ1-Cre;Eμ c-Maf TG and Dis3^fl/fl;Cγ1-Cre;Eμ c-Maf TG mice by using t(14;16) model mice (Eμ c-Maf TG mice).¹³ We hereafter referred to Dis3^wt/wt;Cγ1-Cre, Dis3^fl/wt;Cγ1-Cre, Dis3^fl/fl;Cγ1-Cre, Dis3^fl/wt;Cγ1-Cre;Eμ c-Maf TG, and Dis3^fl/fl;Cγ1-Cre;Eμ c-Maf TG mice immunized with NP-CGG as WT, Dis3^Δ/+_Cγ1, Dis3^Δ/Δ_Cγ1, Dis3^Δ/+_Cγ1c-Maf, and Dis3^Δ/Δ_Cγ1c-Maf mice, respectively (supplemental Figure 3A). During a long-term observation period, Dis3^Δ/+_Cγ1c-Maf and Dis3^Δ/Δ_Cγ1c-Maf mice exhibited B-cell lymphoma (1 of 23 and 1 of 6 mice, respectively), consistent with a previous report that showed that a fraction of Eμ c-Maf TG mice developed B-cell lymphoma after a long latency.¹³ However, the compound mice that we generated did not exhibit an increase in the frequency of plasma cells in the BM (supplemental Figure 3D-E), and showed similar survival compared with that of WT mice (supplemental Figure 3F), indicating that these mice did not develop MM.

In conclusion, Dis3 prevents accumulation of DNA damage in hematopoietic cells, thereby supporting hematopoiesis. Furthermore, loss of Dis3 alone, and in combination with c-MAF transgene in germinal center B cells, do not exhibit plasma cell neoplasm in mice. Although recent studies have informed on the pathological significance of DIS3 deficiency in MM using in vitro experiments,^11,14 it has not been determined in vivo. Our study suggests that additional oncogenic events along with DIS3 deficiency are required for myelomagenesis. Future studies await to determine the role of DIS3 in the development of MM.

Acknowledgments: The authors thank Satoru Takahashi for providing Eμ c-Maf TG mice, and Yutaka Okuno for advice.

This work was supported by grants from the Japan Society for the Promotion of Science KAKENHI (20K08734 and JP 16H06276 [AdAMS], to H.O.), the Shinnihon Foundation of Advanced Medical Treatment Research (H.O.), the Friends of Leukemia Research Fund (H.O.), the SGH Foundation (H.O.), and the Research Clusters of Tokushima University (T. Harada and H.O.).

Contribution: H.O. and Y.O. designed and performed the experiments, analyzed the data, and wrote the manuscript; S.K. performed HSC sorting; K.E., A.H., and S.U. performed random displacement amplification sequencing analysis; T.Y.-N. and J.B. assisted with bone marrow transplantation; T. Masuda, Y.K., T. Harada, M.N., T. Minami, and T. Hideshima analyzed the data; K.A. generated the Dis3-floxed mice; and G.S. designed experiments, provided resources, and edited the manuscript.

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

Correspondence: Hiroto Ohguchi, Division of Disease Epigenetics, Institute of Resource Development and Analysis, Kumamoto University, 2-2-1 Honjo, Chuo-ku, Kumamoto 860-0811, Japan; email: [email protected].