STAT5 activation correlates with erythropoietin receptor-mediated erythroid differentiation of an erythroleukemia cell line.

Interaction between erythropoietin (EPO) and its membrane receptor induces the proliferation and differentiation of erythroid progenitors. EPO has been shown to activate the JAK2-STAT5 pathway in various hematopoietic cell lines, although the physiological role of this pathway is unclear. We have previously shown that epidermal growth factor activates a chimeric receptor bearing the extracellular domain of the epidermal growth factor receptor linked to the cytoplasmic domain of the EPO receptor, resulting in proliferation of interleukin-3-dependent hematopoietic cells and erythroid differentiation (globin synthesis) of EPO-responsive erythroleukemia cells. In the present study, we introduced various deletion and tyrosine to phenylalanine substitution in the cytoplasmic domain of the chimeric receptor and expressed these mutant chimeras in an EPO-responsive erythroleukemia cell line, ELM-I-1. Mutant chimeric receptors retaining either Tyr343 or Tyr401 could activate STAT5, judged by tyrosine-phosphorylation of STAT5 and induction of CIS, a target gene of STAT5. These mutants were able to induce erythroid differentiation. However, a chimeric receptor containing both Y343F and Y401F mutations could not activate STAT5 nor induce erythroid differentiation. Thus, Tyr343 or Tyr401 of the EPO receptor are independently necessary for erythroid differentiation as well as STAT5 activation. Moreover, exogenous expression of dominant-negative STAT5 suppressed EPO-dependent erythroid differentiation. These findings suggest that STAT5 plays an important role in erythroid differentiation through the EPO receptor cytoplasmic domain.


plays an important role in erythroid differentiation through the EPO receptor cytoplasmic domain.
Erythropoietin (EPO) 1 is a glycoprotein hormone required for the survival, proliferation, and differentiation of committed erythroid progenitor cells (1). The EPO receptor (EPOR) belongs to the cytokine receptor superfamily, which includes receptors for other hematopoietic growth factors such as interleukins, colony-stimulating factors, and growth hormone (2). A novel subfamily of protein tyrosine kinases, known as Janus kinases (JAKs), has been shown to associate with cytokine receptors, including the EPOR, and to play an important role in cytokine-dependent cell proliferation and gene regulation (3). Activated JAKs in turn convert latent cytoplasmic transcription factors, known as STATs (signal transducers and activators of transcription), into their active forms by tyrosine phosphorylation. The tyrosine phosphorylated STATs form homo-or heterodimers and translocate into the nucleus, where they bind to their specific target sequences, and then regulate expression of target genes (4).
The STAT family presently includes six members, each of which functions in a specific cytokine system. STAT5, which was originally identified as mammary gland factor regulated by prolactin (5), is activated by multiple cytokines such as IL2, IL3, IL5, granulocyte-macrophage colony stimulating factor, EPO, and thrombopoietin (6 -11). Although IL2 and the other cytokines activate different JAKs, they all activate STAT5. The physiological function of STAT5 in hematopoietic cells, however, remains unclear. The role of STAT5 in cell proliferation is still controversial (9,(12)(13)(14)(15)(16), and little is known about its role in differentiation. However, because STAT3 has been shown to play a critical role in macrophage differentiation through gp130 (17)(18)(19), it would be interesting to determine whether STAT5 is involved in differentiation. We previously reported that exogenous expression of a chimeric receptor containing the extracellular domain of the EGF receptor (EGFR) and the cytoplasmic domain of the EPOR conferred EGF-dependent erythroid differentiation on EPO-responsive cells (20). Moreover, the membrane-proximal 127 amino acids are competent to induce differentiation. This region contains only one tyrosine residue (Tyr 343 ), but it is sufficient to activate STAT5 (12)(13)(14)(15). To clarify the role of STAT5 in the EPOR-mediated differentiation, we constructed various deletion and tyrosine substitution of chimeric EPOR mutants and introduced them into a highly EPO-responsive cell line, ELM-I-1. Tyrosine-mutant receptors that failed to activate STAT5 could not induce erythroid differentiation, suggesting that STAT5 positively regulates EPOR-mediated erythroid differentiation.

EXPERIMENTAL PROCEDURES
Plasmid Construction-The wild type EGFR-EPOR chimera (designated wt chimera in this study) and a truncated chimeric receptor (EGFR-EPORH, designated Ϫ108 in this study) consisting of the extracellular domain of the EGFR and the cytoplasmic domain of the EPOR have been described previously (20,21). These chimeric cDNAs were * This work was supported in part by grants from the Ministry of Science, Education and Culture of Japan, the Kato Memorial Foundation, the Japanese Foundation for Multidisciplinary Treatment of Cancer, the Haraguchi Memorial Foundation, the Uehara Memorial Foundation, the Kowa Life Science Foundation, and the Naito Memorial Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
To substitute phenylalanine for tyrosine in the Ϫ76 deletion (designated -76YY), the SphI/EcoRI fragment, which includes Tyr 343 and Tyr 401 was mutated by PCR using primers containing Tyr to Phe substitutions. The forward primer for Y343F was TGAGCATGCCCAG-GACACCTTCTTG, and the reverse primer for Y401F was TGAAT-TCAGGGGTCCAGGATGGTGAACTCA. PCR fragments were digested with SphI and EcoRI and then exchanged for the SphI/EcoRI fragment of Ϫ76YY. The resulting mutant containing Y343F and Y401F was designated Ϫ76FF. Similarly, the Ϫ76 deletion mutant containing Y343F alone was created and designated Ϫ76FY. The Ϫ76 deletion containing Y401F alone was designated Ϫ76YF. The Ϫ135 deletion containing Y343F (Ϫ135F) was created by using the reverse primer, AGAATTCACTTATCCAATACCAGAAGGTG. The integrity of the sequences was confirmed by sequencing. The resulting constructs are listed in Fig. 1.
Myc-tagged sheep wt and dominant-negative (dn) STAT5s were created by PCR. The dnSTAT5 lacking the 110 C-terminal amino acids was created according to the method of Mui et al. (16). PCR with the forward primer GGAGAATTCGGGCTGGATCCAGGCC and reverse primer CA-GACTAGTAGTACTTGGAGAAGAC (for dnSTAT5) or AACACTAGT-TCAGGAGAGCGAGCCT (for wtSTAT5) was performed, and the products were then subcloned into pCS-MTϩ (22) to add Myc tag (MEQKLISEEDLNE, six times repeated) at the N-terminal. The Hin-dIII/XbaI fragments containing Myc tag and cDNA were recloned into mammalian expression vector pcDNA3.
Cells, Transfection, and Differentiation Assay-Murine EPO-responsive erythroleukemia cell line ELM-I-1, which was established from x-ray-induced erythroblastic leukemia, was maintained in DMEM containing 10% fetal calf serum (23). The plasmids carrying chimeric receptors were introduced into ELM-I-1 cells by electroporation, as described previously (20). After selection with G418 (1 mg/ml), stable transformants expressing the chimeric receptors were identified by flow cytometer analysis using a monoclonal anti-EGFR extracellular domain antibody (Genzyme). Transformants expressing wt or dnSTAT5 were identified by immunoblotting with anti-Myc antibody (22). The transformants expressing chimeric receptors or dnSTAT5 were maintained in DMEM containing 0.5 mg/ml G418. For differentiation assay, cells were cultured in EGF (1 g/ml) or EPO (10 unit/ml) for 3 days. Hemoglobin-positive cells were scored after benzidine staining, and globin content was measured by immunoblotting, as described previously (20).

Construction of Chimeric Receptors and Their Expression in
ELM-I-1-Phosphorylated tyrosine residues of the EPOR have been shown to be important for recruiting signal transducing molecules containing the SH2 domain. For example, phosphatidylinositol 3-kinase binds to phosphorylated Tyr 479 (Y8 in Fig.  1)(24), a hematopoietic cell-specific tyrosine phosphatase to Tyr 429 (Y3) (25,26), and Syp, another phosphatase to Tyr 401 (Y2) (27). Although it has not been demonstrated that STAT5 directly binds to the phosphorylated tyrosine residues of the EPOR, Tyr 343 (Y1) and Tyr 401 (Y2) have been shown to be independently necessary for STAT5 activation. Thus, specific tyrosine residues of the EPOR are probably responsible for each signal transduction pathway. To clarify which tyrosine residues are important for EPOR-mediated erythroid differentiation, we constructed chimeric receptors (Fig. 1) bearing the extracellular domain and the transmembrane domain of the EGFR linked to full-length (wt) or several deletion mutants of the cytoplasmic domain of EPOR (Ϫ55, Ϫ76YY, Ϫ108, Ϫ135Y, and Ϫ181 in Fig. 1). Previous reports have indicated that deletion of the 135 C-terminal amino acids can support proliferation in Ba/F3 cells but that the Ϫ181 mutant cannot (21,28). Tyrosine substitution mutants were also created as shown in Fig. 1.
These chimeric receptors were introduced into erythroleukemia cell line ELM-I-1. Previously, we expressed wt and Ϫ108 chimeras in a Friend virus-transformed erythroleukemia cell line, TSA8, and found that these chimeras could induce erythroid differentiation (20). Whereas EPO induced globin synthesis in TSA8 cells, EPO alone was incapable of inducing more mature differentiation, such as hemoglobin synthesis or extrusion of the nucleus. Moreover, part of the endogenous EPOR may be activated by gp55 of the Friend virus (29). ELM-I-1, on the other hand, is retrovirus-free, and EPO induces not only globin but also hemoglobin synthesis (approximately 10 -30% of cells become benzidine-positive after cultivation in the presence of EPO (30)). Thus, EPO can induce greater terminal  and 479 (Y8)). The chimeric receptors are named according to the number of amino acids deleted from the C-terminal of the EPOR cytoplasmic domain. Substitutions of the Y1 or Y2 tyrosine residues by phenylalanine are indicated by F. The typical results of benzidine positivity in ELM-I-1 transformants are shown on the right. Benzidine staining was carried out after culturing the cells in 1 g/ml EGF or 10 unit/ml EPO for 3 days. Data from one representative clone of each transfection are shown. Benzidine positivity was calculated from about 1000 cells. Two or three independent clones for each construct were tested, and similar results were obtained. erythroid differentiation in ELM-I-1 than in TSA8. In particular, IL3 strongly potentiates EPO-induced erythroid differentiation (30). This is quite similar to natural bone marrow cell differentiation in the presence of EPO and IL3. In several artificial systems, e.g. Ba/F3 sublines expressing exogenous EPOR (31, 32), IL3 suppressed EPO-induced globin synthesis. In this sense, ELM-I-1 cells seem to be more suitable for the study of erythroid differentiation in vitro than other EPO-dependent hematopoietic cells. We therefore introduced chimeric receptors into this cell line.
Differentiation Potential of Chimeric Receptor Mutants-The cDNAs for the chimeric receptors subcloned into a mammalian expression vector were introduced into the ELM-I-1 cells. Expression of the hybrid receptors was examined by flow cytometry using anti-EGFR antibody (data not shown). Positive clones (2-3 clones per chimeric receptor) were cultured in the presence of EGF or EPO for 3 days and then stained with benzidine. Without EPO or EGF, the background of hemoglobin-positive cells was very low. Less than 0.6% of the cells were benzidine-positive in any transformants without EPO or EGF stimulation ( Figs. 1 and 2, Ϫ). Every clone expressing wt or mutant chimeric receptors was positive (5-48%) after culture in EPO, thus differentiation potential was retained after chimeric receptor expression. Representative staining is shown in Fig. 2, and typical staining results for each transformant are shown in Fig. 1 (right). The globin content of each clone cultured in EPO or EGF is also compared by immunoblotting (Fig.  3). As demonstrated in TSA8 cells (20), the Ϫ108 deletion possessed differentiation potential in ELM-I-1 cells. The Ϫ135 deletion was still capable of inducing EGF-dependent differentiation, but Ϫ181 could not (Figs. 1 and 3, Ϫ135Y and Ϫ181). Thus, the differentiation signal from the EPOR required membrane-proximal 90 amino acids.
The presence of Y1 in the cytoplasmic domain of the EPOR has been shown to be sufficient to activate STAT5 in the Ϫ108 deletion mutant (12). To determine whether tyrosine residues are important for EPOR-mediated differentiation, we created Tyr to Phe substitutions. The Ϫ135F mutant contains both deletion of the 135 C-terminal amino acids and the Y343F mutation. In contrast to the Ϫ135Y deletion mutant, Ϫ135F did not induce erythroid differentiation (Figs. 1 and. 3, Ϫ135F). Thus, Y1 is necessary for the differentiation signal when the 135 C-terminal amino acids have been deleted.
STAT5 has been shown to be fully activated through either Y1 or Y2 of the EPOR cytoplasmic domain (13, 15). Thus, we introduced Tyr to Phe substitution in the Ϫ76 deletion mutant containing only two tyrosine residues (Y1 and Y2) (Fig. 1,  Ϫ76YY). As shown in Figs. 2 and 3, the Ϫ76FY mutant, which contains Y343F substitution but retains Y2, was able to induce EGF-dependent erythroid differentiation. Similarly, Ϫ76YF, which contains Y401F but retains Y1, was also capable of inducing differentiation (Figs. 2 and 3, Ϫ76YF). However, Ϫ76FF, which contains double mutation of Y343F and Y401F, could not induce EGF-dependent differentiation (Figs. 2 and 3,  Ϫ76FF). Thus, one of the two tyrosine residues in the Ϫ76 deletion mutant is independently critical for erythroid differentiation.
JAK2 and STAT5 Activation by Tyrosine Mutants-To confirm that the Ϫ76FY and the Ϫ76YF mutant but not the Ϫ76FF mutant can activate STAT5 in ELM-I-1 cells, we measured STAT5 phosphorylation and CIS-induction through the mutant chimeric receptors. The Ϫ181, Ϫ76YY, Ϫ76YF, Ϫ76FY, and Ϫ76FF transformants were stimulated with EGF, and FIG. 4. Activation of STAT5 by chimeric receptor mutants. A, transformants (1 ϫ 10 7 /sample) were incubated in serum-free DMEM medium for 5 h at 37°C and then stimulated with or without 1 g/ml EGF at 25°C for 10 min. Cell extracts were immunoprecipitated (IP) with anti-JAK2 (␣JAK2) or anti-STAT5 (␣-STAT5), then resolved by SDS-polyacrylamide gel electrophoresis, and then immunoblotted (blot) with anti-phosphotyrosine (␣PY), ␣JAK2, or ␣STAT5 antibodies. B, transformants were cultured in 1 g/ml EGF or 10 unit/ml EPO for 2 h, and then total RNA (10 g/lane) was separated from the cells and hybridized with CIS and ␣-enolase probes. then JAK2 and STAT5 were immunoprecipitated. Their tyrosine phosphorylation was detected with anti-phosphotyrosine antibody. As shown in Fig. 4A, EGF stimulated JAK2 phosphorylation in every mutant chimera. JAK2 phosphorylation was observed even in the Ϫ181 mutant, probably because this mutant contains box1 and box2 regions, which are putative JAK binding sites (Fig. 1). Thus, activation of JAK2 does not require tyrosine residues in the EPOR cytoplasmic domain. In contrast, STAT5 was not phosphorylated in response to EGF in the Ϫ181 nor the Ϫ76FF mutant, whereas Ϫ76YY, Ϫ76YF, and Ϫ76FY were able to induce STAT5 phosphorylation.
CIS is an immediate early gene induced by multiple cytokines such as IL2, IL3, and EPO in various hematopoietic cells and a direct target of STAT5 (12,33). We prepared RNA from EGF-or EPO-stimulated cells expressing Ϫ76YY, Ϫ76YF, Ϫ76FY, and Ϫ76FF. Whereas EPO induced CIS expression in all transformants, EGF induced CIS in Ϫ76YY, Ϫ76YF, and Ϫ76FY mutant cells but not in Ϫ76FF mutant cells (Fig. 4B). These data are consistent with previous reports that Y1 or Y2 of the EPOR cytoplasmic domain is sufficient to activate STAT5.
Effect of Dominant-negative STAT5-To clarify the critical role of STAT5 in erythroid differentiation, we expressed wt and dn type STAT5 in ELM-I-1 cells. The dominant-negative form of sheep STAT5 was created by deleting the C-terminal transactivation domain as reported by Mui et al. (16). This construct has been shown to inhibit STAT5 activation and partially suppress cell proliferation. The Myc epitope tag was introduced into the N terminus to detect exogenous protein. Three independent clones from each transfection were tested for EPOinduced hemoglobin synthesis. As shown in Fig. 5, expression of wtSTAT5 did not affect EPO-induced benzidine positivity, whereas dnSTAT5 suppressed differentiation to about 5%. These data support that STAT5 plays a critical role in EPORmediated erythroid differentiation in ELM-I-1 cells. However, we could not rule out the possibility that dnSTAT5 may suppress the activation of other signaling molecules through Y1 and Y2, because dnSTAT5 may tightly bind to these two phosphorylated tyrosine residues. Extensive study is under way to determine whether STAT5 is directly involved in differentiation.
Our results contrast strongly with a recent report by Chretien et al. (34). They found that EPO-induced differentiation in a human leukemia cell line, TF-1, correlates with impaired STAT5 activation, although direct evidence that STAT5 sup-presses differentiation was not provided. This discrepancy is apparently caused by the difference in cell lines used. ELM-I-1 can grow in the absence of any cytokines, whereas TF-1 requires granulocyte-macrophage colony stimulating factor for growth. These two cells may be derived from different developmental stages of hematopoietic cells, and such differences may cause distinct STAT5 requirements for differentiation. However, the same as natural BFU-E cells, IL3 enhances EPOinduced erythroid differentiation of ELM-I-1 cells, although it does not induce erythroid-differentiation by itself. It is particularly important to examine the role of STAT5 in in vivo differentiation of natural erythroid progenitor cells. Studies are under way using bone marrow and fetal liver cells infected with retrovirus bearing chimeric receptors or dnSTAT5.