Effect of Nucleophosmin1 (Npm1) Haploinsufficiency on Hematopoietic Stem Cells (HSC)

Nucleophosmin1 (NPM1), located at 5q35.1, is frequently mutated, deleted, or translocated in a number of hematopoietic malignancies (1). Deletion of 5q encompassing the NPM1 locus occurs in approximately 40% of high risk MDS/AML with complex karyotypes (2), although the common 5q interstitial deletion that occurs in MDS does not include this region (3). Mice genetically engineered to harbor heterozygous loss of Npm1 (Npm1+/−) were shown to have a mild MDS-like phenotype characterized by dyserythropoiesis and dysplastic megakaryocytes with macrocytic anemia, but without significant cytopenias or alterations in lineage commitment. These mice also exhibited increased susceptibility to hematologic malignancies with age (4, 5). While these studies confirmed NPM1’s role as a tumor suppressor, they did not characterize a role for NPM1 in hematopoietic stem or progenitor cell function, a likely phenotype alteration given the propensity of NPM1 mutant mice to develop hematologic neoplasms. In order to better understand the effects of loss of a single NPM1 allele on hematopoiesis, we obtained Npm1 heterozygous deficient mice (#11744-UCD) from the Mutant Mouse Regional Resource Centre (UC Davis, CA). An embryonic stem cell clone with a single gene trap event in the Npm1 locus was used to generate the Npm1+/− line (genetic background 129/SvEvBrd X C57BL/6J).

Western blot analysis of cell lysates from total bone marrow cells demonstrated an approximate 50% reduction in Npm1 protein levels in hematopoietic cells in NPM1+/− mice as compared to wild type controls (Supplementary Fig 1). This reduction is similar to that reported in the previously characterized NPM1+/− mutant (4). Complete loss of Npm1 (Npm1−/−) resulted in embryonic lethality as reported previously (4, 6); however, Npm1+/− mice were viable, grossly normal, and born at expected Mendelian ratios (Npm1+/+:Npm1+/−:Npm1−/− ratios of 1:2.4:0; c² = 0.86, p > 0.3).

Next, we evaluated NPM1+/− mice for alterations in hematopoiesis. Complete blood counts (CBCs) did not differ significantly between age-matched Npm1+/+ and Npm1+/− mice (Supplementary Fig. 2). This result is in contrast to a previous report showing alterations in red blood cell size and anisocytosis (4). In addition, the number of bone marrow cells per femur did not differ between Npm1+/+ and Npm1+/− mice, regardless of age (data not shown). Flow cytometric analysis revealed no significant change in the absolute number or relative percentage of mature myeloid (Gr-1+, Mac1+) or lymphoid (B220+ or CD3+) cells (Supplementary Fig 3A, data not shown) in Npm1+/− mice as compared to wild type littermates (up to 2 yrs of age). Evaluation of erythroid cells using CD71 and Ter119 antibodies revealed that the percentage of mature Ter119⁺, CD71⁻ erythroid cells (calculated as a percentage of total bone marrow cells) was significantly decreased, as previously reported (Supplementary Fig 3B, p < 0.05 in Npm1+/− mice); however, in contrast to this prior report, the immature CD71+Ter119+ erythroid cell population was unaffected (p < 0.3).

As NPM1 loss-of-function is associated with the development of hematologic malignancies containing stem-cell like populations, we asked whether Npm1 haploinsufficiency affects HSC function (Figure 1). The number of HSCs in the bone marrow exhibited a small, but statistically significant increase (1.4-fold) in the percentage of HSCs (Lin-cKit+Sca-1+CD34−CD150+) within the Lin-cKit+Sca+ population of 12 Npm1+/− mice as compared to 13 wild type littermate controls (Figure 1B; p < 0.01). In addition, the absolute number of HSCs was significantly increased (1.8-fold) (Figure 1C; p < 0.04). The expansion in HSC numbers persisted as mice aged from 4 to 20 months. These data suggest that the level of Npm1 expression plays a role in regulating HSC numbers in the bone marrow

Figure 1. Quantitation of HSCs in Npm1+/− mice.

A. Representative FACS profile pre-gated on live, lineage lo/−, cKit+ and Sca1+ cells showing LT-HSC (CD34−, CD150+) in Npm1+/+ and Npm1+/− mice. (B) The frequency of HSC (Lin-cKit+Sca-1+CD34−CD150+) within the LKS (Lin-cKit+Sca-1+) population in Npm1+/+ (circle) and Npm1+/− (square) mice. (C) Absolute numbers of HSC (Lin-cKit+Sca-1+CD34−CD150+) in a single femur bone of Npm1+/+ (circle) and Npm1+/− (triangle) mice. Each bar represents the average value for each category.

To determine whether functional differences exist between Npm1+/+ and Npm1+/− HSC, we examined the ability of HSC to form colonies in vitro. Using double FACS-sorted HSC, the total number and frequency of methylcellulose colony forming unit-granulocyte/monocyte (CFU-GM) and colony forming unit-granulocyte/erythrocyte/monocyte/megakaryocyte (CFU-GEMM) were similar between Npm1+/+ (n=8) and Npm1+/− HSC (n=9) (Supplementary Fig. 4A). Similarly, clonal HSC liquid culture assays did not show differences in plating efficiency, cell number, or lineage composition when grown in myeloid-promoting conditions (data not shown). Together, these data indicate that Npm1 haploinsufficiency does not impair HSC survival, differentiation, or growth in vitro, an observation that is consistent with the normal peripheral blood cell counts in Npm1+/− mice. As the increased HSC cell number in Npm1+/− mutant mice may be due to Npm1 effects on HSC self-renewal, colonies from methylcellulose assay were harvested and serial replating assays were performed for four rounds with 10,000 cells, with each round of replating occurring after 10 days. No significant differences in the number of colonies were observed during serial replating of HSCs from Npm1+/+ (n=3) and Npm1+/− (n=3) mice (Supplementary Fig 4B)

As the increased numbers of HSC observed in NPM+/− mice may be due to increased growth or decreased cell death, we determined whether Npm1+/− HSC (Lin-cKit+Sca1+CD34−CD150+) exhibit increased cell cycling. Evaluation of HSC cell cycle status in Npm1+/+ (n=4) and Npm1+/− (n=7) mice using Ki67 and DAPI staining revealed no significant difference in their cell cycle distribution (Figure 2C, 2D). The majority of both Npm1+/+ and Npm1+/− HSC were in a quiescent state (Ki67−, DAPI−) and fewer than 15% cells were in the G1 phase (Ki67+, DAPI−). A representative cell cycle profile of Npm1+/+ and Npm1+/− HSC is shown in Figure 2B. In addition, there was no difference in HSC apoptotic rates, either by staining with annexin V/PI (data not shown) or by calculating the clonal HSC plating efficiency using methylcellulose or in vitro liquid culture systems (Supplementary Fig. 4)

Figure 2. Reduced Engraftment Potential of Npm1+/− HSC and their cell cycle status.

300 double sorted HSC (Lin-cKit+Sca1+CD34−CD150+) from Npm1+/+ and Npm1+/− mice were engrafted into lethally irradiated C57/B6 mice (CD45.2). The percentage of donor chimerism (CD45.1 and CD45.2) was studied in the peripheral blood (A) and the bone marrow (B) of the recipient mice 16 weeks after transplantation using flow cytometry. Each circle and square represents Npm1+/+ and Npm1+/− HSC recipient mice, respectively, and each bar represents the average for each category. (C) The cell cycle status of HSC (Lin-cKit+Sca1+CD34−CD150+) from Npm+/+ and Npm1+/− mice was evaluated by staining for Ki67 and DAPI. The bar shows G0 (Ki67−DAPI−), G1 (Ki67+DAPI−) and S/G2/M (Ki67+DAPI+) phases in HSC from four Npm1+/+ and seven Npm1+/− mice. (D) A representative cell cycle profile of HSC from one Npm1+/+ and one Npm1+/− mice.

To further characterize the function of Npm1+/− HSC, we performed competitive transplantation assays in which equal numbers of double-sorted HSC from Npm1+/+ and Npm1+/− mice (CD45.1+CD45.2+) were transplanted into lethally irradiated C57BL/6 (CD45.2) recipients together with competitor cells (CD45.2). Short term (2 months) as well as long term (4 months) competitive repopulating capability was studied by monitoring peripheral blood donor chimerism of recipient mice by flow cytometry. Mice transplanted with Npm1+/− HSC exhibited markedly lower engraftment than the recipients of Npm1+/+ HSC as measured by total chimerism (3-fold) (Fig. 2A, p < 0.03). The percentage total and myeloid chimerism in individual transplanted mice did not differ significantly at 2 and 4 months (data not shown). We also quantified donor chimerism in the bone marrow of the recipient mice 4 months after transplantation (Figure 2B). Similar to the results in the peripheral blood, the competitive repopulating ability of the HSC from Npm1+/− was significantly reduced (8.5-fold) (Fig. 2B, p < 0.02) in the bone marrow of the recipient mice. While the engraftment potential of Npm1+/− HSC was reduced, we did not observe any effects on the composition of donor-derived mature hematopoietic cells in the peripheral blood of recipient mice (Supplementary Fig 5). Thus, the level of Npm1 expression has a direct effect on HSC repopulating ability in vivo, but has no effect on hematopoietic fate commitment.

In summary, our NPM1+/− mouse model indicates that Npm1 plays a role in maintaining HSC number and in preserving the functional integrity of these cells in the context of competitive transplantation. However, it does not appear to play a role in regulating HSC differentiation. Npm1 haploinsufficiency did not result in significant alterations in mature blood cell numbers or significant dysplasia in our mice. These results are at some variance with a previous report in which Npm1 heterozygosity was associated with an MDS-like phenotype in older (6–10 months) mice (4). We cannot completely exclude the possibility that the different techniques used to generate these mutant mice may explain the differences in our results; however, both models are associated with homozygous null embryonic lethality as well as decreased NPM1 mRNA (data not shown) and protein transcription in Npm1 +/− mice. Our study did use mice in which Npm1 was functionally eliminated by a single gene trap event in the Npm1 locus between exon 7 and 8 (genetic background 129/SvEvBrd X C57BL/6J) while in the previous study exons 2–7 were replaced by GFP cassette introducing the stop codon after exon 2 (genetic background 129/Sv X C57BL/6). Minor genetic variability between the two different substrains used in these studies (129/Sv vs 129/SvEvBrd) might also influence their respective phenotypes (7). Although we have no definitive explanation for the difference in phenotypes, only our study directly assessed HSC function. We speculate that the increased number of HSC in NPM+/− mice could be an important early step in myeloid pathogenesis, as these HSC could provide a milieu in which additional genetic events might occur, thereby increasing the susceptibility to subsequent development of MDS or myeloid leukemia. Indeed, prior studies have demonstrated that transgenic Eμ-myc overexpression in the NPM1+/− mutant setting can facilitate AML development (4). However, these data also indicate that loss of NPM1 function alone is not sufficient to result in increased development of hematologic malignancies, which suggests that other cooperating lesions are likely necessary for disease pathogenesis. Further studies focused on understanding the interactions between NPM1 and other genes associated with MDS or AML progression will help to identify these cooperating events.