Expression of oncogenic K-ras from its endogenous promoter leads to a partial block of erythroid differentiation and hyperactivation of cytokine-dependent signaling pathways

Deregulated Ras signaling frequently occurs in human patients with acute myeloid leukemia, myelodysplastic syndromes, and myeloproliferative disorders (reviewed in Shannon¹ ). Dysregulated Ras signaling is mainly achieved through 2 mechanisms. First, Ras signaling is aberrantly activated by constitutive activation of upstream tyrosine kinases (eg, tel-PDGFRβ and BCR-ABL) and tyrosine phosphatases (eg, PTPN11) or inactivation of NF1, a Ras GTPase-activating protein (reviewed in Lauchle et al² ). Second, elevated Ras signaling results from oncogenic mutations in the N- and K-ras genes

In human patients with various hematopoietic malignancies, the red cell lineage is often affected, characterized by different defects in erythroid differentiation.³  We have been using mouse fetal liver erythroid progenitors as a model system to study the role of oncogenic Ras signaling in erythroid differentiation.^4,5  In this system, we can monitor erythroid differentiation step-by-step in vivo and in culture based on the expression of erythroid-specific TER119 and nonerythroid-specific CD71 (transferrin receptor) surface proteins. We can purify large amounts of erythroid progenitors and early erythroblasts from mouse fetal livers with approximately 75% to 85% purity. We can culture purified erythroid progenitors in vitro such that they undergo normal terminal proliferation and differentiation.

Using these tools, we previously showed that overexpression of oncogenic Ras leads to the constitutive activation of its downstream signaling pathways, including p44/42 MAPK, Akt, and Rlf, a severe block of terminal erythroid differentiation, and cytokine-independent growth of primary erythroid progenitors. However, these results might not mimic the conditions in human cancer patients carrying oncogenic Ras mutations in their endogenous loci. To better model human cancers, we used a tetracycline-inducible system to induce K-ras^G12D expression from its endogenous locus. Here, we show that expression of endogenous K-ras^G12D leads to a partial block of terminal erythroid differentiation and hyperactivation of cytokine-dependent signaling pathways.

All mouse lines were maintained in mixed 129sv/Jae and C57BL/6 genetic backgrounds. The embryos were generated from crossing R26-M2rtTA^/+;LSL K-ras^G12D/+ mice with LC-1/LC-1;R26R/R26R mice.^6–9  The pregnant mice were fed with water containing 0.2% doxycycline and 1% sucrose at E13.5. The embryos were harvested 42 to 48 hours later. Genotyping of the adult mice and embryos was performed as described.^6–9 

TER119-negative cells were purified from individual embryos as previously described.¹⁰  β-Galactosidase-neo (βgeo) activity in the purified cells was detected with a fluorogenic substrate fluorescein di β-D-galactopyranoside (FDG) (Invitrogen Molecular Probes, Carlsbad, CA) as described.¹¹ 

Fetal liver cells (FLCs) were isolated from E15.5 embryos and were simultaneously stained for CD71 and TER119 as previously described.⁴  Flow cytometry was carried out on a Becton Dickinson FACSCalibur machine (BD Biosciences, Franklin Lakes, NJ).

TER119-negative FLCs were purified from individual embryos and purified cells were divided into 4 aliquots. Cells were starved in Iscove modified Dulbecco's medium containing 0.5% bovine serum albumin for 30 minutes at 37°C and then stimulated with or without 0.3 to 3 U Epo/mL for 10 minutes at 37°C. Western blot analysis was performed essentially as described.¹⁰ 

In preliminary studies, we used different cre lines to express oncogenic K-ras from its endogenous promotor at different embryonic stages. Because germ line expression of endogenous K-ras^G12D results in early embryonic lethality,⁹  we first used the Mox2 cre line, in which cre expression starts at E5 and is restricted to epiblasts.¹²  The late onset and the exclusion of the placenta from cre expression has been reported to rescue the early embryonic lethality seen in Rb (retinoblastoma) germ line knockouts.¹³  However, embryos carrying both the LSL K-ras^G12D and the cre alleles died around E11.5 in mixed 129sv/Jae and C57BL/6 genetic backgrounds (data not shown). Because the first wave of definitive erythropoiesis in mouse fetal liver occurs at E12, the early lethality of these embryos prevented us from studying oncogenic K-ras signaling in primary erythroid progenitors.

To overcome this problem, we used a tetracycline-inducible system (Figure 1A). In this system, expression of a modified tetracycline-controlled transactivator (M2rtTA) is under the control of the Rosa26 promoter (R26-M2rtTA) line.^7,14,15  In the presence of doxycycline (Dox), M2rtTA binds to tet operators (TetO) and presumably results in cre expression throughout the embryos.⁸  To test the recombination efficiency of this inducible system in fetal liver erythroid progenitors (TER119-negative cells), we used a conditional β-galactosidase-neo (βgeo) reporter line (R26R line)⁶  (Figure 1A). The expression of βgeo was detected with its fluorogenic substrate FDG (fluorescein di-β-galactopyranoside). Mouse embryos carrying all 3 alleles (R26-M2rtTA, TetO-cre, and R26R) were generated. In the absence of Dox, the percentage of FDG-positive cells in TER119-negative population was, as expected, close to zero (Figure 1B). In contrast, greater than 90% of TER119-negative cells (predominantly primitive progenitor cells, including mature erythroid burst-forming units [BFU-Es] and erythroid colony-forming units [CFU-Es]) became FDG-positive on Dox induction for as short as 1 day. Moreover, the floxed K-ras allele was undetectable in the embryos by polymerase chain reaction after Dox treatment (data not shown). Thus, this inducible system can be used to efficiently remove a floxed stop cassette and induce gene expression in early erythroid cells.

Induction of oncogenic K-ras is achieved by feeding the pregnant mice with water containing Dox. We determined the optimal induction scheme to be starting induction on E13.5 and harvesting FLCs 2 days later (on E15.5). Thus, after 2 days of Dox induction, FLCs were isolated from individual embryos and analyzed by fluorescence-activated cell sorting (FACS). The percentage of TER119-negative cells (erythroid progenitors and early erythroid blasts) in embryos expressing oncogenic K-ras significantly increased compared with that in wild-type embryos (22.0% and 10.2%, respectively; P < 10⁻⁶) (Figure 2A-B), suggesting a partial block of terminal erythroid differentiation in vivo. Thus, these embryos showed anemia rather than erythrocytosis as a result of inefficient erythroid differentiation. This result is consistent with the inefficient erythropoiesis phenotype observed in adult mice expressing endogenous K-ras^G12D.¹⁶ 

We further purified TER119-negative cells from individual embryos and stimulated them with or without Epo (Figure 2C). In the absence of Epo stimulation, there was no constitutive activation of downstream signaling pathways in cells expressing oncogenic K-ras. When stimulated with increasing concentrations of Epo, wild-type cells showed increasing activation of the 3 downstream signaling pathways tested, Stat5, Akt, and p44/42 MAPK, and maximum activation was achieved at 3 U Epo/mL. In contrast, in cells expressing oncogenic K-ras signaling pathways were hyperactivated at low concentrations of Epo and maximum activation was observed at 1 U Epo/mL. These data indicate that oncogenic K-ras expressed from its endogenous locus does not lead to constitutive activation of downstream signaling pathways but does hyperactivate them on Epo stimulation.

The hyperactivation of cytokine-dependent signaling pathways by endogenous K-ras^G12D is in sharp contrast to the constitutive activation of Ras downstream signaling pathways seen with overexpressed oncogenic Ras.⁵  Consequently, expression of endogenous K-ras^G12D only leads to a mild block of terminal erythroid differentiation and mild hyperproliferation of primary erythroid progenitors in our in vitro culture system (data not shown). Importantly, endogenously expressed K-ras^G12D does not support cytokine-independent growth of these cells (data not shown) in contrast to overexpression of oncogenic Ras.⁴ 

The hyperactivation of cytokine-dependent signaling pathways by endogenous K-ras^G12D is consistent with previous observations that hematopoietic progenitors expressing endogenous K-ras^G12D are hypersensitive to cytokine stimulation in various colony assays.^16–18  To our knowledge, this is the first time that endogenous K-ras^G12D is shown to hyperactivate cytokine-dependent signaling pathways in primary hematopoietic progenitors.

We thank Dr Philippe Soriano for generously providing us with R26R mice and Dr Hermann Bujard for kindly providing us with LC-1 mice. We thank Drs Qiang Chang and Alec Gross for helpful discussion and critical comments on the manuscript. We also thank Stacey Sullivan and Tony Chavarria for their help with mice.

This work was supported by a grant from the National Institutes of Health (grant PO1 HL 32262) (H.F.L.) and a postdoctoral fellowship from the Leukemia and Lymphoma Society (J.Z.).

National Institutes of Health

Contribution: J.Z. experimental design and execution and writing manuscript; Y.L. experimental execution; C.B., D.A.T., R.J., and T.E.J. development of knock-in mice; H.F.L. experimental design and writing manuscript.

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

Correspondence: Harvey F. Lodish, Center for Cancer Research and Department of Biology, Massachusetts Institute of Technology, Cambridge, MA 02138; e-mail: lodish@wi.mit.edu.