Runx1 deficiency predisposes mice to T-lymphoblastic lymphoma

AML1ETO, the fusion gene resulting from t(8;21) in acute myeloid leukemia (AML) subtype M2, and CBFB-MYH11, the fusion gene resulting from inv(16) in AML subtype M4 Eo, are both capable of blocking hematopoiesis in mouse models.^1,2 Runx1^+/ETO and Cbfb^+/MYH11 heterozygous embryos die at embryonic day 12.5 (E12.5) to E13.5 from hemorrhages and defective hematopoiesis, similar to Runx1^-/- and Cbfb^-/- embryos.^3-5  In vitro studies have demonstrated that the proteins encoded by these 2 fusion genes are capable of repressing the expression of Runt-related transcription factor 1 (Runx1)/core binding factor β (Cbfβ) target genes.^6,7  These observations have led to the hypothesis that the fusion genes block hematopoietic differentiation and induce leukemia by inhibiting Runx/Cbfb activity in a dominant-negative manner. More recently, sporadic loss-of-function mutations, both monoallelic and biallelic, have been found in patients with AML.^8,9  In addition, the disease familial platelet disorder with propensity to develop AML (FPD/AML) has been linked to heterozygous mutations in RUNX1.^10,11  Together these studies have led to the hypothesis that inhibition of Runx1/Cbfb function is a critical step in the development of hematopoietic malignancies.

In this study, we used a mouse model to determine whether disruption of Runx1 expression predisposes mice to the development of hematopoietic malignancies. Chimeras generated with Runx1^lacZ/lacZ embryonic stem (ES) cells developed lymphoma at increased incidence after treatment with N-ethyl-N-nitrosourea (ENU) to induce additional genetic changes. This finding provides strong evidence that Runx1-deficient cells are susceptible to malignant transformation.

Runx1^lacZ/lacZ ES cells were generated from Runx1^+/lacZ ES cells¹²  (Runx1^tm2Spe/+) by growth in 4.0 mg/mL G418.¹³ Runx1^lacZ/lacZ (LZD) and wild-type 129 (TC1) ES cells were used to generate chimeric mice using standard protocols. One-month-old mice were given a single dose of ENU (100 mg/kg) by intraperitoneal injection.¹⁴ 

Liver sections were stained with an anti-CD3 antibody and a peroxidase staining kit (DAKO, Carpinteria, CA). Spleen sections were stained with a rabbit anti–β-galactosidase antibody and a diaminobenzidine (DAB) substrate kit for peroxidase (Vector Laboratories, Burlingame, CA).

Pictures of the stained tissue sections were obtained using a Nikon Eclipse E800 microscope equipped with a Plan-Apo 40 ×/0.95 objective lens (Nikon, Melville, NY) and a Photometrics CoolSnap HQ camera (Roper Scientific, Tucson, AZ). Pictures were acquired through IPLab 3.5.5 software (Scanalytics, Fairfax, VA) and were organized for publication with Adobe Photoshop 6.0.1 (Adobe, San Jose, CA).

Thymocytes were stained with phycoerythrin (PE)–conjugated anti-CD4 and Cy-chrome–conjugated anti-CD8 (BD Pharmingen, San Diego, CA) and analyzed using FACSCalibur (BD Biosciences, San Diego, CA).

Southern blot analysis was performed on DNA isolated from bone marrow cells, using a 400–base pair (bp) Runx1 intron 8 probe as described previously.¹² 

Thymocytes were stained with a PE-conjugated anti-CD4 and a fluorescein isothiocyanate (FITC)–conjugated anti-CD8 antibody (BD Pharmingen) and sorted with FACS Aria (Becton Dickinson, San Jose, CA). Bone marrow cells were isolated and sorted after staining with an FITC-conjugated anti–c-kit antibody and PE-conjugated antibodies against CD3, B220, Mac-1, Gr-1, and TER119 (BD Pharmingen). The DNA was extracted from the subpopulations. Polymerase chain reaction (PCR) was performed using 200 to 400 ng DNA and the following PCR conditions: 94°C for 4 minutes; 40 cycles of 94°C for 1 minute, 56°C for 1 minute, and 72°C for 2 minutes; 72°C for 4 minutes. The PCR primers used were LacZ forward (ACTGGCAGATGCACGGTTAC) and LacZ reverse (GTGGCAACATGGAAATCGCTG). Spleen sections effaced by lymphoma were treated with zylene and methanol to remove paraffin, and the genomic DNA was isolated using QIAamp DNA mini kit (QIAGEN, Valencia, CA). Primers for markers D7Mit44 and D11Mit188 were used for PCR genotyping to distinguish 129 from C57BL/6J cells.¹⁵ 

We generated chimeric mice using Runx1^lacZ/lacZ ES cells and monitored for leukemia development. None of the mice developed malignancy (Figure 1A). We then treated them with a single injection of 100 mg/kg ENU. Chimeras generated with wild-type ES cells were used as controls. Due to technical difficulties, only a relatively small number of chimeras (8 for the test and 9 for the control groups, respectively) became available for the study. Four of the Runx1^lacZ/lacZ chimeras developed lymphoblastic lymphoma within the first 12 months after ENU treatment. These 4 lymphomas were all derived from the ES cells (see below). Two mice died from unknown causes at 4 and 8 months after injection and were excluded from statistical calculation. During the same time period, one chimera generated with wild-type ES cells developed lymphoma. However, genotyping by PCR indicated that the lymphoma was derived from the C57BL/6J cells (data not shown). In addition, 2 other chimeras generated with wild-type ES cells died from unknown causes at 7 and 8 months after ENU injection. These 3 mice were excluded from the statistical analysis. Therefore, 4 of the 6 Runx1^lacZ/lacZ chimeras developed lymphoma of ES-cell origin, whereas none of the 6 wild-type ES-cell chimeras did (Figure 1A). This represents a significant increase (P = .03, by Fisher exact test, one-sided) in the incidence of lymphoma in the Runx1^lacZ/lacZ chimeras compared with wild-type chimeras.

Runx1^lacZ/lacZ chimeras that developed lymphoblastic lymphoma (LZD nos. 2, 3, 4, and 5) had enlarged thymuses that were completely effaced by the malignant cells (Figure 1B). There was neoplastic infiltration in the lungs, liver (Figure 1B), kidney, spleen (Figure 2A), and bone marrow (Figure 1B). Staining of liver sections with anti-CD3 antibody revealed that the lymphoma was of T-cell origin (Figure 1B). Flow cytometric analysis revealed coexpression of CD4 and CD8 (Figure 1C), a common phenotype for T-lymphoblastic lymphoma.¹⁶ 

To verify the ES-cell origin of the lymphoblastic lymphomas that developed in the Runx1^lacZ/lacZ chimeras, we stained spleen sections for β-galactosidase and performed Southern blot hybridization with DNA isolated from bone marrow cells. The spleens of Runx1^lacZ/lacZ chimeras that developed lymphoma were completely effaced by β-galactosidase–positive tumor cells (Figure 2A), indicating their ES-cell origin. In LZD nos. 2, 3, and 4, the bone marrow was completely effaced by lymphoma and only the targeted LacZ allele was detected by Southern blot hybridization (Figure 2B), confirming that the malignant cells were derived from ES cells. On the other hand, only the wild-type allele was detected in the bone marrow of LZD no. 6 (unaffected; Figure 2B). These results suggest that the 4 T-lymphoblastic lymphomas that developed within 7 months were of ES-cell origin.

Since there is an early block in hematopoiesis in Runx1^lacZ/- embryos,¹²  it is somewhat puzzling that there are hematopoietic Runx1^lacZ/lacZ cells in the adult chimeras that can give rise to lymphoma. It is likely that similar to Cbfb^+/MYH11 chimeras,¹⁴  some Runx1^lacZ/lacZ precursors are able to survive in the context of a chimeric animal. To address this possibility, we analyzed the contribution of Runx1^lacZ/lacZ ES cells to the hematopoietic compartments in untreated healthy Runx1^lacZ/lacZ chimeras. By genomic DNA PCR we were able to detect the presence of ES cell–derived cells in all populations from the thymus (Figure 2C). ES cell–derived cells were also detected by PCR in the c-kit⁺/lin^- and c-kit⁺/lin⁺ subpopulations in the bone marrow but were barely or not at all detectable in the c-kit^-/lin⁺ population (Figure 2D). However, the overall hematopoietic contribution from the ES cells must be small, since by Southern blot we can only detect the wild-type allele in LZD no. 6, an ENU-treated Runx1^lacZ/lacZ chimera that did not develop neoplasia (Figure 2B). In addition, there was no significant β-galactosidase staining of the spleen sections from LZD no. 6 (Figure 2A) and 2 untreated healthy Runx1^lacZ/lacZ chimeras (data not shown). The results suggest that Runx1^lacZ/lacZ ES cells gave rise to hematopoietic stem cells and early progenitors, but later stages of differentiation were blocked.

It is somewhat unexpected that T-lymphoblastic lymphoma was induced in the Runx1^lacZ/lacZ chimeras, since RUNX1 deficiency is more commonly associated with AML. However, there is one report of a translocation involving RUNX1 in a 12-year-old boy with T-cell acute lymphoblastic leukemia,¹⁷  suggesting that defects in RUNX1 expression can be associated with T-cell malignancies in humans as well as in mice. While our results may reflect some of the biologic differences between humans and mice, it is also possible that Runx1 deficiency is capable of inducing both myeloid and lymphoid malignancies, and myeloid leukemia would have been observed with a larger set of mice or higher dose of ENU. In one recent report, a transgenic mouse model was used to determine if expression of AML1-ETO can induce leukemia in mice.¹⁸  The AML1-ETO transgenic mice developed both AML and T-cell lymphoma (55% and 45%, respectively) after treatment of ENU at a relatively high dose (300 mg/kg), whereas 100% of the wild-type controls developed lymphoma.

Our results suggest that Runx1^lacZ/lacZ -derived cells are more susceptible than wild-type cells to developing lymphoblastic lymphoma after chemical mutagenesis. Although we cannot exclude the possibility that the Runx1-lacZ fusion gene exerts some novel activity, the phenotype of Runx1^lacZ/- embryos is identical to that of Runx1^-/- embryos with embryonic lethality between E12.5 and E13.5 from hemorrhages and absence of definitive hematopoiesis.¹²  Therefore, our results suggest that Runx1 deficiency can predispose to development of hematologic malignancies in mice. Additional studies in Runx1^-/- chimeras or Runx1 conditional knockout mice will be needed to confirm the findings reported here.

We would like to thank Shelley Hoogstraten-Miller and Eugene Elliott for assistance with veterinary issues and technical support with the ENU injections and Joan Bailey-Wilson for statistical analysis.