Loss of Runx1 perturbs adult hematopoiesis and is associated with a myeloproliferative phenotype

Abstract

Homozygous loss of function of Runx1 (Runt-related transcription factor 1 gene) during murine development results in an embryonic lethal phenotype characterized by a complete lack of definitive hematopoiesis. In light of recent reports of disparate requirements for hematopoietic transcription factors during development as opposed to adult hematopoiesis, we used a conditional gene-targeting strategy to effect the loss of Runx1 function in adult mice. In contrast with the critical role of Runx1 during development, Runx1 was not essential for hematopoiesis in the adult hematopoietic compartment, though a number of significant hematopoietic abnormalities were observed. Runx1 excision had lineage-specific effects on B- and T-cell maturation and pronounced inhibition of common lymphocyte progenitor production. Runx1 excision also resulted in inefficient platelet production. Of note, Runx1-deficient mice developed a mild myeloproliferative phenotype characterized by an increase in peripheral blood neutrophils, an increase in myeloid progenitor populations, and extramedullary hematopoiesis composed of maturing myeloid and erythroid elements. These findings indicate that Runx1 deficiency has markedly different consequences during development compared with adult hematopoiesis, and they provide insight into the phenotypic manifestations of Runx1 deficiency in hematopoietic malignancies.

Figure 1.

Inducible Runx1 excision in adult mice induces thrombocytopenia. (A) Schematic representation of Runx1 gene-targeting strategy used to flank exon 4 with LoxP-targeting sites. B indicates BamHI; S, SspI; E4, Runt domain exon 4. (filled boxes) 5′ and 3′ targeting probes. (dashed arrows) BamHI-digested genomic DNA fragments for wild-type (7.3-kb), Floxed (2.8-kb), and excised (2.5-kb) DNA detected with excision (Δ) probe (open boxes) located 3′ to the distal LoxP site. (B) Southern blot analysis of BamHI-digested genomic DNA from bone marrow [B] and spleen [S] cells of representative mice killed at 14, 35, and 91 days after pIpC injection. The dose of pIpC is indicated. The blots are probed with Δ probe, which detects the Floxed (2.8-kb) and excised (2.5-kb) Runx1 alleles. Numbers indicate percentage excision. (C) Mean ± SD of total WBC counts, total red blood cell (RBC) counts, total platelet counts, percentage lymphocytes, and percentage neutrophils in PB of pIpC-treated Runx1^F/F—Tg(Mx1-Cre) (red symbols) and Runx1^F/F (black symbols) mice. Mice were bled 12 days before pIpC injection and 14, 21, 28, 35, 49, 91, 119, and 154 days after pIpC injection. The number of animals evaluated for each genotype at the respective time points is N (Runx1^F/F—Tg(Mx1-Cre) = 14, 14, 12, 8, 10, 9, 8, 7 and 6; N (Runx1^F/F) = 7, 5, 7, 7, 5, 7, 6, 6, and 6.

Figure 2.

Stem cells from pIpC-treated Runx1^F/F—Tg(Mx1-Cre) mice are reduced in competitive repopulation ability. (A) Ethidium bromide (EtBr)–stained 3% agarose gel of 3-primer PCR of Runx1 loci from sorted HSCs derived from fresh BM 17 weeks after pIpC (Table 1) and unfractionated BM and PB. 1 = Runx1^F/F; 2 = Runx^F/F—Tg(Mx1-Cre). Mr = 1-kb ladder, sizes indicated (bp). Control samples include tail DNA from Runx1^+/+, Runx1^F/+, and Runx1^F/F mice and from Runx1^Δ/Δ, Runx1^F/Δ, and Runx1^+/Δ embryos. Runx1^Δ alleles were generated by mating Runx1^F/F mice to Tg(EIIa-Cre) mice and intercrossing Runx1^F/Δ mice. (B) Composite showing gating strategy and CD45.1/CD45.2 staining of PB from representative mice that underwent transplantation with Runx1^F/F or Runx1^F/F—Tg(Mx1-Cre) and competitor marrow 8 weeks after transplantation. Green indicates lymphocytic (R2), and red indicates granulocytic (R3) fractions based on scatter gating (SSC, side scatter; FSC, forward scatter). Gate assignments were confirmed by back-gating of control stainings with Gr1⁺–, Mac1⁺–, CD3⁺–, T-cell receptor β (TCRβ⁺)–, CD19⁺–, or B220⁺–stained PB (not shown). The percentage contribution to the granulocytic and lymphocytic lineages of PB from each marrow isotype (CD45.1/CD45.2) was determined. (C) Plotted is the mean (±SD) percentage contribution of donor-derived CD45.2⁺ cells in PB from mice that underwent transplantation with Runx1^F/F—Tg(Mx1-Cre) (red symbols; n = 6) or Runx1^F/F (black symbols; n = 6) and competitor bone marrow to total WBCs, granulocytes, and lymphocytes of recipient mice at 4, 8, and 19 weeks after transplantation. Two of 8 mice receiving Runx1^F/F—Tg(Mx1-Cre) marrow that failed to show contribution of CD45.2⁺ marrow to PB of recipient mice at any time point are excluded. (D) Nineteen weeks after transplantation, EtBr-stained 3% agarose gel of 3-primer PCR of Runx1 loci of BM from mice that underwent competitive transplantation. (E) EtBr-stained 3% agarose gel of 3-primer PCR of Runx1 loci from BM of mice that underwent transplantation without competitor marrow.

Figure 3.

Lymphoid development is inhibited in Runx1-excised mice. (A) Representative B220/CD19 and B220/CD43 staining of BM from Runx1^F/F and Runx1^F/F—Tg(Mx1-Cre) mice. CD43 staining is shown for cells gated as IgM^–NKK1.1^–Lin^–. Numbers indicate percentage of cells gated in each respective quadrant. (B) EtBr-stained 3% agarose gels with 3-primer PCR products of Runx1 loci in PB, spleen (Sp), BM, and thymocytes (Th) of mice 154 days after pIpC. Genotypes are as indicated. Control samples are as in Figure 2. The primers do not amplify a product from the Runx1^rd allele. (C) Shown are CD4 and CD8 staining from representative animals of genotypes, Runx1 ^F/+, Runx1^F/rd, and Runx1^F/rd—Tg(Mx1-Cre). Numbers indicate percentage of cells gated in each respective quadrant. (D) Shown are CD25 and CD44 staining of CD45.2⁺CD4^–CD8^– gated thymocytes from representative animals, as in panel C. (E) Composite of high-speed flow cytometric analysis of BM from representative Runx1^F/F and Runx1^F/F—Tg(Mx1-Cre) mice. Numbers indicate percentages of total BM for indicated populations. Gated populations indicated include IL-7Rα⁺Lin^– (R1), IL-7Rα⁺Lin⁺ (R2) (which includes immature B cells), and CLP (R3), as indicated in Table 1. PI indicates propidium iodide.

Figure 4.

Megakaryocyte maturation is inefficient in pIpC-treated Runx1-excised mice. BM from representative (A) Runx1^F/F and (B) Runx1^F/F—Tg(Mx1-Cre) mice 143 days after pIpC injection shows an absence of normal megakaryocytes in Runx1-excised mice (hematoxylin and eosin [H&E] staining; original magnification, × 600). (C) BM from representative Runx1^F/F—Tg(Mx1-Cre) recipient mice 98 days after transplantation (H&E; original magnification, × 600). (A-C) Yellow arrows indicate representative megakaryocytes. Red and green arrowheads indicate representative erythroid and myeloid elements, respectively. Sections demonstrate a notable absence of megakaryocytes and an increased ratio of maturing myeloid-to-erythroid forms in Runx1-excised versus nonexcised marrows. (D) Histograms depicting number of cultured Runx1^F/F and Runx1^F/F—Tg (Mx1-Cre) bone marrow cells stained with propidium iodide to show ploidy. Colors indicate CD41 gates: green, CD41⁺; red, CD41^–. (E, left) Plotted are mean ± SD numbers of acetylcholinesterase-positive colonies per 5 × 10⁴ BM cells plated from pIpC-treated Runx1^F/F (□) and Runx1^F/F—Tg(Mx1-Cre) (▪) mice. Results shown are for 3 experiments performed in quadruplicate. 1 and 3, fresh marrow; 2, previously frozen marrow. (Right) Plotted is the average fold increase for Runx1^F/F—Tg(Mx1-Cre) (▪) relative to Runx1^F/F (□) in the 3 experiments (P ≤ .034). Representative megakaryocyte colonies from (F) Runx1^F/F and (G) Runx1^F/F—Tg(Mx1-Cre) mice show acetylcholinesterase staining (brown) of megakaryocytic cells (original magnification, × 100).

Figure 5.

Mature and progenitor myeloid cell populations are expanded in pIpC-treated Runx1-excised mice. (A) Plotted are the mean ± SD of total (P < .05), myeloid (P < .05), BFU-E (P = .5), and mixed-lineage colonies (P = <.05) per 2 × 10⁴ BM cells plated in vitro from pIpC-treated Runx1^F/F (□) and Runx1^F/F—Tg(Mx1-Cre) (▪) mice. Results are the average of 3 experiments performed in duplicate. (B) Representative cytospin from mixed-lineage colony derived from Runx1^F/F—Tg(Mx1-Cre) BM (Wright-Giemsa staining; original magnification, × 600). Red arrowheads indicate representative erythroid elements. Green arrowheads denote myeloid and monocytic forms. (C) Representative EtBr-stained 3% agarose gel of Runx1 PCR products from single GM or mixed-lineage colonies derived from Runx1^F/F—Tg(Mx1-Cre) (lanes 1-20) and Runx1^F/F (lanes 21-25) mice. Of 70 Runx1^F/F—Tg(Mx1-Cre) colonies evaluated, 9 (12.8%) failed to show any amplification, 5 of 61 (8.2%) indicated partial excision, and 56 of 61 (91.8%) were completely excised. Tail and embryo genomic DNA controls are as indicated. (D) Gr1 and Mac1 staining of BM from representative CXB6F1 mice 154 days after pIpC induction. Numbers indicate percentages of cells gated in each respective quadrant.

Figure 6.

Runx1-excised mice display myeloid expansion in spleen and liver. Spleen (A-B) and liver (C-D) sections from representative (A,C) Runx1^F/F and (B,D) Runx1^F/F—Tg(Mx1-Cre) mice 143 days after pIpC (H&E; original magnification, ×100). Insets' original magnification, × 600. (E-F) Spleen sections from representative mice 14 weeks after transplantation with (E) Runx1^F/F and (F) Runx1^F/F—Tg(Mx1-Cre) marrow (H&E; original magnification, × 100). Insets' original magnification, ×600. (A-F) Yellow arrows indicate representative lymphocytes. Red and green arrowheads indicate representative erythroid and myeloid elements, respectively. Sections demonstrate the presence of extramedullary hematopoiesis in the liver and splenic red pulp in Runx1-excised mice, absent in nonexcised control animals. (G) Gr1/Mac1 and CD45.2/Ter119 staining of splenocytes of indicated genotype from representative primary mice or mice that underwent transplantation. Numbers indicate percentage of cells gated in each respective quadrant.