Runx3 deficiency results in myeloproliferative disorder in aged mice

The Runx genes belong to the Polyomavirus enhancer binding protein 2/core binding factor family of heterodimeric transcription factors that are important for development and are implicated in various human diseases.^1-3  Of 3 family genes, RUNX1 is a well-documented hematopoietic and leukemia factor. RUNX1 is abrogated by multiple types of genetic alterations in human hematological malignancies^3-5  and is critical for the generation and maintenance of hematopoietic stem cells (HSCs).^6-10 Runx1 regulates the expression of stemness- and niche-related factors, such as Bmi1, Cxcr4, and integrin α₂, deregulation of which in adult Runx1 knockout (KO) mice led to an expanded hematopoietic stem/progenitor cell (HSPC) compartment and subsequent stem cell exhaustion.^6-12  In addition, adult Runx1 KO mice show abnormal megakaryocytic differentiation and defective lymphoid development.^11-14  In contrast, the role of Runx3 in hematopoiesis remains elusive, despite the fact that Runx3 is expressed in various hematopoietic tissues in adult mice.¹⁵  We therefore sought to examine its effects on hematopoiesis in Runx3-deficient mice.

Runx3^fl/fl chimeric mice obtained from genetically manipulated 129P2/O1aHsd-derived E14 embryonic stem cells injected into C57BL/6 blastocysts¹⁶  were backcrossed onto C57BL/6 inbred mice for ≥3 generations before being used for subsequent crosses with Mx1-Cre⁺ mice. For induction of the Mx-Cre transgene, 4-week-old Runx3^fl/fl;Mx1-Cre⁺ mice were injected intraperitoneally with 600 µg of polyinosinic-polycytidylic acid (pIpC; Sigma-Aldrich) on 7 alternate days.

Mice were subcutaneously injected with 300 μg/kg/day murine granulocyte-colony stimulating factor (G-CSF; PeproTech) daily for 4 days. Twenty microliters of peripheral blood (PB) was obtained for the colony-forming unit culture (CFU-C) assay.

For complete information on hematological analyses, flow cytometric analyses, CFU-C assay, bone marrow transplantation (BMT) procedures, and quantitative real-time polymerase chain reaction analysis, see supplemental Methods on the Blood website.

Due to the neonatal lethality of conventional Runx3^−/− mice, conditional Runx3 KO mice were generated. To induce Cre expression in hematopoietic cells, Runx3^fl/fl;Mx1-Cre⁺ mice and Runx3^fl/fl;Mx1-Cre⁻ littermates were injected with pIpC (hitherto referred to as Runx3 KO and Runx3 wild-type [WT] mice, respectively). Deletion of the Runx3 locus was nearly complete in the BM of Runx3 KO mice (supplemental Figure 1A).

Young Runx3 KO mice (6-8 weeks old) did not display gross hematopoietic abnormalities (supplemental Figure 2), evidenced by normal white blood cell (WBC), hemoglobin (Hb), and platelet (Plt) counts. Consistent with these findings, flow cytometric analysis showed no significant changes in various lineage populations, namely the myeloid, lymphoid, and erythroid lineages. There were no significant differences in BM cellularity, spleen weight, and thymus weight between Runx3 KO and Runx3 WT mice.

The HSPC compartment, immunophenotypically defined as a c-Kit⁺Sca-1⁺Lin⁻ (KSL) population, in the BM of Runx3 KO mice exhibited a slight increase, although this was statistically not significant (P = .11, Student t test; supplemental Figure 2D). In addition, a trend toward an increase in short-term HSC frequency with a marginal decrease in long-term HSC frequency in Runx3 KO mice was observed (supplemental Figure 2E). Interestingly, a significant increase in CFU-C activity was observed in cells from Runx3 KO mice (Figure 1A-E). The increased CFU-C activity was further pronounced after the replating assay (Figure 1B-D). Notably, Runx3 KO cells do not replate for a longer period of time compared with the WT cells. These results suggest that that Runx3 KO mice have enhanced CFU-C forming capacity of HSPCs compared with their WT littermates.

HSPC analysis in the spleen of Runx3 KO mice revealed a marginal increase in the KSL and myeloid progenitor (c-Kit⁺Sca-1⁻Lin⁻) populations (Figure 1F-G). As increased HSPC numbers in the spleen is indicative of an enhanced mobilization ability, a mobilization assay was conducted using G-CSF, which is widely known to enhance the mobilization of HSPCs.¹⁷  The result showed that the number of CFU-Cs in the PB of Runx3 KO mice was significantly greater than that in Runx3 WT mice at all time points (Figure 1H). Therefore, Runx3 KO HSPCs appear to be easily mobilized into the periphery by G-CSF.

To explore whether the hematopoietic abnormalities will be enhanced over time, Runx3 KO mice were aged to 18 months old. Deletion of the Runx3 locus in BM cells from aged Runx3 KO mice remained almost complete (supplemental Figure 1B). There was a significant increase in WBC counts in aged Runx3 KO mice compared with WT controls, whereas Hb and Plt counts remained unchanged (Figure 2A). Notably, the leukocytosis was dominated by a population of myeloid cells, characterized by abundant granulocytic cells in the PB and an increase in Mac-1⁺ and Gr-1⁺ populations in the BM and spleen of aged Runx3 KO mice (Figure 2B-C and data not shown). Analysis of other hematopoietic lineages in the BM revealed a significant decrease in the erythroid lineage (Figure 2C). These changes were accompanied by significant increases in BM cellularity and spleen weight (Figure 2D-E). HSPC analysis in aged Runx3 KO mice showed an increase of KSL and myeloid progenitor populations in BM and spleen compared with aged Runx3 WT mice (Figure 2F-I). Although the frequency of KSL in BM remained unchanged, which is thought to be attributable to the decreased Lin⁻ fraction in Runx3 KO BM, the absolute number of KSL cells was significantly increased in the aged Runx3 KO mice (supplemental Figure 3A-C). Analysis of the proliferative ability of KSL cells by Ki-67 staining revealed a tendency toward more Ki-67⁺ cells in Runx3 KO mice, suggesting that the increased HSPCs in aged mice is probably due to enhanced proliferation in a subtle but sustained manner (supplemental Figure 3D). Together, these results suggest that the increased WBC count and increased BM cellularity were due to an expansion of KSL and myeloid progenitor compartments and myeloid cells. Of note, development of spontaneous leukemia was not observed in aged Runx3 KO mice.

The Runx family genes show a high degree of conservation in their amino acid sequences within the family and share the same capability to bind to their consensus DNA sequence. Therefore, the observation that Runx3 function was not fully unveiled in young Runx3 KO mice may be attributable to functional compensation among Runx family genes. However, Runx1 expression levels were not significantly altered in both young and aged Runx3 KO mice (supplemental Figure 3E). However, aged Runx3 KO mice partially but clearly phenocopied Runx1 KO mice, implying that Runx3 per se is also a critical factor for hematopoiesis. Although a clear increase of KSL fraction was observed in young Runx1 KO mice,⁶  a mild myeloid proliferation phenotype was only occasionally observed in aged Runx1 KO mice.^11,13  In contrast, all aged Runx3 KO mice examined showed a myeloproliferative status accompanied by a modest HSPC expansion. Consistent with the higher penetrance of MPD in Runx3 KO mice than in Runx1 KO mice, HSC exhaustion was not observed in the Runx3 KO mice, unlike Runx1 KO mice, as Runx3 KO BM cells transplanted into WT recipient mice were not outcompeted by WT competitor cells at different doses (test: competitor ratios of 1:1, 2:1, and 4:1) in BMT experiments. In fact, transplantation with Runx3 KO donor cells resulted in greater chimerisms and higher WBC counts in the recipient mice, as opposed to mice transplanted with Runx3 WT BM cells (Figure 2J; supplemental Figure 3F; data not shown), implicating that Runx1 and Runx3 seem to play distinct roles in regulating the HSC behavior. The BMT result also suggests that the increase of phenotypic HSCs in aged Runx3 KO mice corresponds to an increase of functional HSCs and that the phenotypes observed were due to cell autonomous mechanisms. Differentiation blocks in megakaryocytic and lymphoid lineages, which were observed in Runx1 KO mice, were not evident in Runx3 KO mice. Intriguingly, Runx3-deficient HSPCs showed enhanced HSPC mobilization suggested by hypersensitivity to G-CSF, like RUNX1 haploinsufficient status, which is known to cause familial Plt disorder with predisposition to acute myeloid leukemia (AML).¹⁸ 

Human data further implicate RUNX3 in leukemogenesis. Deletion in chromosome 1p36, containing the RUNX3 locus, is frequently observed in human leukemia cells.^19-21  Methylation of the RUNX3 promoter in AML has also been reported,²²  and RUNX3 expression is down-regulated in AML with t(8;21) or inv(16).^23,24  Collectively, our work represents the first direct evidence for the pivotal role of Runx3 in hematopoiesis. Further analysis of Runx3 would provide us with deeper insight into leukemogenesis and may lead to future drug development.

The authors thank K. Rajewshky for Mx-Cre Tg mice and members of the Biological Resource Center, Biopolis, and MD2 Vivarium, NUS, for mouse husbandry.

This work was supported by the A*STAR (Agency of Science, Technology and Research), Biomedical Research Council, National Medical Research Council, Singapore National Research Foundation, and the Ministry of Education under the Research Center of Excellence Programme.

Contribution: C.Q.W. performed experiments, analyzed data, and wrote the manuscript; L.M. performed experiments; M.S., Y.I., and I.T. provided research tools; M.O. and V.T. designed research and wrote the manuscript; and C.Q.W., L.M., M.S., Y.I., I.T., M.O., and V.T. reviewed and approved the manuscript.

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

Correspondence: Motomi Osato, Cancer Science Institute of Singapore, National University of Singapore, 14 Medical Dr, Singapore 117599; e-mail: csimo@nus.edu.sg; and Vinay Tergaonkar, Institute of Molecular and Cell Biology, 61 Biopolis Dr, Singapore 138673; e-mail: vinayt@imcb.a-star.edu.sg.