Requirement for JAK2 inErythropoietin-Induced Signalling Pathways

Cell. Signal. Vol. 9, No. 1, pp. 85–89, 1997 Copyright  1997 Elsevier Science Inc.

ISSN 0898-6568/97 $17.00 PII S0898-6568(96)00121-0

Requirement for JAK2 in Erythropoietin-Induced Signalling Pathways Thomas Bittorf,* Robert Jaster, Britta Lu¨dtke, Beatrice Kamper and Josef Brock Institute of Medical Biochemistry, Medical Faculty of the University Rostock, Schillingallee 70, PF100888, 18055 Rostock, Germany

ABSTRACT. Erythropoietin (EPO) exerts its activities by the induction of multiple signalling pathways through interaction with the erythropoietin receptor (EPOR). Previous studies have suggested that the Ras/MAP kinase as well as the JAK/STAT signalling cascades play significant roles in the induction of EPO-responsive genes. Here we show that, in HCD-57 erythroleukemic cells, both pathways are activated by EPO in a dosedependent manner with similar sensitivities and kinetics. The activation of signalling molecules is closely related to the proliferative status of the cells. Using an antisense strategy, we were able to show that the downregulation of the JAK2 protein level in HCD-57 cells results in a distinct reduction of the ability to induce not only STAT5 DNA-binding, but also MAP kinase activity. Our results thus provide evidence for a significant contribution of the cytosolic tyrosine kinase JAK2 to the EPO-induced activation of the Ras/MAP kinase cascade. Copyright  1997 Elsevier Science Inc. cell signal 9;1:85–89, 1997. KEY WORDS. Erythropoietin, Signalling pathways, MAP kinase, JAK/STAT

INTRODUCTION The proliferation, differentiation and functional activities of hematopoietic cells are regulated by growth factors or cytokines through binding to specific membrane receptors on their target cells [1]. Erythropoietin (EPO) is the main regulator of erythropoiesis and exerts its action by the activation of receptors on progenitor cells [2]. Although cytokine receptors do not have kinase domains, the ligands induce rapid tyrosine phosphorylation of cytosolic molecules as well as the receptors themselves [3]. Multiple signalling pathways are activated by the interaction of receptor phosphotyrosine residues with SH2 domains on certain signalling molecules. A number of such SH2-containing proteins associate with the erythropoietin receptor (EPOR) including Grb2 [4], SH-PTP1 [5], Shc [4] and PI 3-kinase [6]. The phosphorylation of receptor-associated molecules is thought to be carried out, at least in part, by the EPOR-associated tyrosine kinase JAK2 [7, 8], a member of the Janus Kinase family, or other unidentified cytosolic tyrosine kinases. However, there is still no experimental evidence for a direct phosphorylation of these target molecules. Along with the activation of JAKs, latent cytoplasmic transcription factors, * Author to whom all correspondence should be addressed. Abbreviations: EPO–erythropoietin; EPOR–erythropoietin receptor; MAP kinase–mitogen-activated protein kinase; STAT–signal transducer and activator of transcription; JAK–Janus kinase; SH-PTP1–src-homology domain containing protein tyrosine phosphatase; ECL–enhanced chemiluminescence; EMSA–electrophoretic mobility shift assay; BSA–bovine serum albumine; PBS–phosphate buffered saline; PI 3-kinase–phosphatidylinositol 3-kinase Received 22 May 1996; and accepted 1 July 1996

known as STATs (Signal Transducer and Activator of Transcription), are tyrosine phosphorylated and translocated into the nucleus where they bind to specific target sequences. EPO specifically induces the activation of STAT5 [9, 10], originally identified as a transcriptional factor which is activated in the lactating mammary gland [11]. The identification of other cytoplasmic kinases of the JAK family and their interaction with additional members of the STAT family suggests that a common pathway exists linking the activation of cytokine receptors directly to the regulation of gene transcription [12]. Another major signalling pathway which is activated by many cytokines is the well established and highly conserved Ras/MAP kinase pathway. We and others have previously shown that both signalling pathways are involved in the EPO-induced biological responses of erythroid cells [8–10, 13–15]. Recent studies indicate that the two separate cascades are linked at the level of MAP kinase [16, 17], which contributes to the activation of STATs during the interferon-stimulated transcription of early response genes. The potential role of JAKs for the activation of the Ras/MAP kinase pathway has not been addressed. It is also not known if the two signalling pathways differ in their sensitivity to cytokine stimulation. It was recently shown that the ability of EPOR mutants to transmit proliferative signals depends on the EPO concentration used [18]. In this study, we analysed the induction of characteristic signalling events in the presence of different EPO concentrations and employed an antisense strategy to elucidate the contribution of JAK2 for the activation of the Ras/MAP kinase pathway. Our results show, that in HCD-57 cells, both

86

signalling cascades are induced in a dose-dependent manner with a similar sensitivity and kinetics. The activation of signalling molecules is closely correlated to the proliferative status of the cells. Experiments using JAK2 antisense treated cells reveal that the depletion of the JAK2 protein level results in a significant reduction of the ability to induce STAT5 and MAP kinase activities. The data provide evidence for a contribution of JAK2 in the activation of the Ras/MAP kinase cascade emphasizing the importance of JAK2 in the signal transduction through the EPOR. MATERIALS AND METHODS Cell Culture The murine EPO-dependent erythroleukemia cell line HCD-57 [19] was grown in suspension culture in Iscove’s modified Dulbecco’s medium (IMDM, Gibco, Eggenstein, Germany) supplemented with 30% foetal calf serum, penicillin 100 mg/ ml, 2 3 1025 M b-mercaptoethanol, and 0.6 U/ml recombinant human EPO (Boehringer, Mannheim, Germany). The cells were resuspended three times a week at a density of 2 3 105 cells/ml. For short–term stimulation experiments, cells were cultured without EPO for 24 h and resuspended in IMDM at 1 3 106 cells/ml before the addition of EPO.

T. Bittorf et al.

Preparation of Nuclear Extracts and EMSAs Nuclear extracts were prepared essentially as described previously [20]. For electrophoretic mobility shift assays, nuclear proteins corresponding to 105 cells were incubated with 16 fmol double stranded oligonucleotide derived from the bovine b-casein promoter (59-AGA TTT CTA GGA ATT CAA ATC-39). The oligonucleotides were endlabelled with [g32P] ATP (Amersham, U.K.) by polynucleotide kinase. The shift assays were performed in a total volume of 20 ml in the following buffer: 10 mM Tris-HCl (pH 7.5), 50 mM potassium chloride, 0.1 mM EDTA, 1 mM dithiothreitol, 1 mg/ml BSA, 5% glycerol, 0.1% NP40, 1 mM Pefabloc (Boehringer Mannheim, Germany). The reactions, also containing 2 mg poly(dl-dC) (Boehringer Mannheim, Germany), were performed at room temperature for 30 min and initiated by the addition of nuclear extract. Supershift analysis was done by including either 0.5 mg of anti-STAT1 or anti-STAT5 antibody (Santa Cruz Biotechnology, CA, U.S.A.) at 48C for an additional 20 min incubation. Complexes were analysed by electrophoretic separation on a 6% polyacrylamide gel in 0.25 3 TBE buffer. Assay of SH-PTP1-Phosphatase Activity

Antisense Oligonucleotides to JAK2 HPLC-purified phosphorothioate-modified oligonucleotides were purchased from Biognostik (Go¨ttingen, Germany). The antisense JAK2 sequence (59-GCT TGT GAG AAA GC-39) corresponds to nucleotides 1902-1915 of murine JAK2 [7]. A random oligonucleotide (59-GTC CCT ATA CGA AC-39) was used as a control. Both oligonucleotides were applied at a concentration of 5 mM and replenished whenever cell cultures were diluted. Assay of Cell Proliferation Cell proliferation was analysed by [3H]thymidine incorporation. Cells (104-5 3 104) were pulsed with 0.6 mCi [3H]thymidine (83 Ci/mmol) for 4 h as described previously [20]. DNA was harvested onto glass fibre filters and assayed by liquid scintillation counting. Cell viability was assessed by their ability to exclude trypan blue. Immunoblotting Cell extracts were prepared by boiling 5 3 105 cells of each sample in 50 ml SDS sample buffer (50 mM Tris pH 6.8, 10% glycerol, 1.5% SDS, 4.2% b-mercaptoethanol, 0.01% bromophenolblue). Total protein was separated through SDS-polyacrylamide gels and transferred to nitrocellulose membranes (Hybond-ECL, Amersham, U.K.) as described previously [13]. Membranes were blocked with 1% BSA and incubated with anti-MAP kinase and JAK2 (Santa Cruz Biotechnology, CA, U.S.A.) antibodies. After a final incubation with a horseradish peroxidase-labelled antirabbit Ig antibody, blots were developed with ECL (Amersham, U.K.).

To assay SH-PTP1 activity, HCD-57 cells were stimulated by EPO as described and lysed in lysis buffer (20 mM TrisHCl, pH 7.5, 137 mM NaCl, 1% Triton X-100, 2 mM EDTA, 10% glycerol, 0.15 U/ml aprotinin) for 1 h. Lysates were precleared by centrifugation and supernatant aliquots corresponding to 2 3 106 cells were incubated with rabbit polyclonal SH-PTP1 antibody (Santa Cruz Biotechnology, CA, U.S.A.) prebound to Protein A-sepharose beads (0.5 mg antibody/5ml beads per sample) at 48C for 2 h. After washing the beads three times in lysis buffer, the beads were transferred to the phosphatase assay buffer (20 mM TrisHCl, pH 7.2, 0.1% b-mercaptoethanol, 1 mg/ml bovine serum albumine). The assay was performed by means of a nonradioactive phosphatase assay system supplied by Boehringer (Mannheim, Germany). Briefly, a biotin-labelled phosphopeptide was immobilized on a streptavidin-coated microtiter plate and incubated with the SH-PTP1 immunoprecipitates at 378C for 1 h. The reactions were terminated by the addition of sodium orthovanadate (100 mM). After washing the wells with PBS (pH 7.2), a peroxidase-labelled anti phosphotyrosine-antibody was added at 378C for 1 h. Finally, the fraction of unmetabolized phosphopeptide was determined by the addition of the peroxidase substrate ABTS (2,2-Azino-di-[3-ethylbenzthiazoline sulfonate]). Absorbance was measured after 3 min at 405 nm using an ELISA reader. RESULTS EPO Induces Transient STAT5, MAP Kinase and SH-PTP1 Activities in HCD-57 Cells To evaluate the stimulation of the JAK/STAT as well as the Ras/MAP kinase pathway in HCD-57 cells by EPO, EMSA

JAK2 in Erythropoietin-Induced Signalling Pathways

FIGURE 1. Kinetics of STAT5, MAP kinase and SH-PTP1 ac-

tivation by EPO in HCD-57 cells. Cells were challenged by EPO for the indicated times. (A) Nuclear extracts were prepared from 106 cells and analysed by EMSA using b-casein oligonucleotides as probe. Supershift analysis was done as described. Shifted and supershifted (ss) complexes are pointed out by arrows. (B) 105 cells of the same samples were lysed and subjected to immunoblotting with an anti-MAP kinase antibody. The positions of p42 MAP kinase and its slower migrating, activated form are pointed out by arrows. (C) 2 3 106 cells per sample were assayed for SH-PTP1 activity as described. Results are expressed as percentage of unstimulate control cells. High phosphatase activities correspond to a low absorbance. Data are means 6 S.E.M. of three independent experiments.

was employed for the analysis of STAT5 DNA-binding activity, using the STAT5 binding site derived from the bovine b-casein promoter as a probe and immunoblotting for the demonstration of MAP kinase activity changes. Fig. 1A shows the kinetics of STAT5 activation after EPO challenge. In the absence of EPO, no DNA/protein complexes were detectable. After hormonal stimulation, a rapid and transient induction of a specific DNA-binding protein was observed. The complex was supershifted by an anti STAT5antiserum but not by an antibody specific for STAT1, indicating that STAT5 is involved in the DNA/protein interaction. Both p42 and p44 MAP kinase are expressed in HCD-57 cells, but p42 MAP kinase at much higher levels than p44 MAP kinase (data not shown). During activation of p42 MAP kinase, a slower migrating form of the enzyme can be observed by immunoblotting. We have previously shown that the shifted band corresponds to the phosphorylated p42 MAP kinase [13]. The increasing phosphorylation can be correlated to an enhanced enzyme activity using in-gel kinase assays [14]. The data presented in Fig. 1B demonstrate that EPO induces a rapid phosphorylation of p42 MAP kinase displaying a kinetics which is very similar to the activation of STAT5. The tyrosine specific phosphatase

87

FIGURE 2. Activation of MAP kinase, STAT5 and induction of cellular growth by EPO is dose-dependent. (A) Cells were EPOdeprived for 24 h and then exposed for 10 min to EPO at the indicated concentrations. Lysates of 105 cells were subjected to Western analysis with anti-MAP kinase antibodies. (B) Nuclear extracts from the same cells were subjected to EMSA as described (C) 24 h after exposure to EPO cell proliferation was assayed by [3H]thymidine incorporation. Data are means 6 S.E.M. of three independent experiments.

SH-PTP1 was shown to associate with the EPOR and it was speculated that the induction of its activity terminates EPO-induced signalling events [5]. Here we show that the activity of SH-PTP1 is transiently induced by EPO (Fig. 1C). In contrast to the activities of MAP kinase and STAT5 peaking after about 15 min, the highest activity was found after 1 h. The Signalling Pathways Triggering the Activation of STAT5 and p42 MAP Kinase are Highly Sensitive to EPO and Respond in a Dose-Dependent Manner to Hormonal Stimulation The majority of studies on EPO-induced signalling has been conducted with non-erythroid cells transfected with the EPOR and using supraphysiological concentrations of the hormone. Therefore, we investigated the dose responsiveness of the JAK/STAT and Ras/MAP kinase signalling pathways using EPO concentrations ranging from physiological levels up to 500 mU/ml. The corresponding results are shown in Fig. 2 together with 3H-thymidine-incorporation values reflecting the proliferation status of the cells. The results indicate that HCD-57 cells are very sensitive to EPO and respond with the stimulation of MAP kinase and STAT5 activities even at low (10 mU/ml) physiological concentrations (Fig. 2A/B). The activation of both signalling molecules is dose-dependent and correlates well with the induction of proliferation as assessed by 3H-thymidineincorporation (Fig. 2C).

88

T. Bittorf et al.

FIGURE 4. Activation of STAT5 DNA-binding and MAP ki-

FIGURE 3. Reduction of JAK2 protein levels during JAK2 anti-

sense treatment of HCD-57 cells. Cells were exposed to a JAK2 antisense- or a random control oligonucleotide at a concentration of 5 mM for at least 72 h or left untreated. Lysates of 105 cells were subjected to Western analysis with an anti-JAK2 antibody. Equal protein loading was checked by probing the same blot with an anti-MAP kinase antibody.

JAK2 Antisense Oligonucleotide Treatment Results in a Decreased Level of STAT5 DNA-Binding Activity After EPO Challenge The involvement of JAK2 in the transduction of signals through the EPOR is well established [8, 23]. To analyse the role of JAK2 in the activation of downstream signalling molecules, we employed an antisense strategy. HCD-57 cells were treated with oligonucleotides for at least 3 d. The growth of the cells was not noticeably affected by the presence of antisense or control oligonucleotides, respectively (data not shown). The protein level was checked by comparing untreated cells and cells treated with the oligonucleotides using Western analysis (Fig. 3). We observed a significant reduction of the JAK2 protein level in antisensetreated cells. Reprobing the same blot with an anti-MAP kinase antibody confirmed equal loading (lower panel). Having established cell cultures displaying reduced levels of JAK2 we next investigated the consequences of these changes for the ability of the cells to respond to EPO by the stimulation of STAT5 DNA-binding activity. Fig. 4A shows the results of electrophoretic mobility shift assays using oligonucleotides with STAT5-binding sequences. It is obvious that the ability of JAK2 depleted cells to respond to low levels (50 mU/ml) of EPO with STAT5 activation is strongly reduced. At higher EPO levels (500 mU/ml) STAT5-activity is induced in JAK2 depleted cells, but still at a much lower level than in cells treated with control oligonucleotides or untreated cells, respectively. JAK2 is Involved in the Activation of the Ras/MAP Kinase Signalling Pathway Recently it was shown that MAP kinase activity is necessary for the regulation of the JAK/STAT signalling cascade

nase activity in JAK2-depleted HCD-57 cells. Following a 72 h oligonucleotide treatment (see Fig. 3) cells were again exposed to the oligonucleotides but EPO-deprived for 24 h and than exposed to EPO (50 or 500 mU/ml) as indicated. A. Nuclear extracts were prepared as described and subjected to EMSA using b-casein oligonucleotides. B. Lysates of 105 cells were subjected to Western analysis with anti-MAP kinase antibodies.

in interferon-stimulated cells [16]. As there is more and more experimental evidence for links between the two signalling pathways, we were interested in the role of JAK2 for the induction of MAP kinase activity. To address this question we used JAK2-depleted cells to compare their EPOinduced MAP kinase activation to cells treated with a control oligonucleotide and untreated cells, respectively. We were able to show that cells having a reduced JAK2 protein level are clearly diminished in their ability to activate MAP kinase after EPO challenge (Fig. 4B). Our results provide evidence that JAK2 is involved in both the activation of MAP kinase and STAT5 DNA-binding.

DISCUSSION Erythroid progenitor cells respond to the glycoprotein hormone EPO by the activation of multiple signalling molecules including the Janus Kinase JAK2 [8], the PI 3-kinase [6], and Shc [4]. JAK2 seems to play an important role in the propagation of signals as it was shown to be responsible for the activation of the recently identified transcription factor STAT5 [9, 10]. In addition, JAK2 may be involved in coupling the EPOR to signalling pathways like the Ras/ MAP kinase cascade [24]. In this study, we examined the time course and dosedependency of STAT5 and MAP kinase activation in the erythroleukemic, EPO-dependent HCD-57 cell line. Both molecules were previously shown to be involved in the realization of EPO-dependent biological functions [8–10, 13– 15]. It is tempting to speculate that the signalling pathways triggered by EPO may have different sensitivities to the hormone and the response of a given cell in vivo depends on the local EPO concentration and receptor occupancy, respectively. Our results demonstrate that at least in this cell line the activation of MAP kinase and STAT5 occurs simultaneously at the same EPO concentrations. The down-regulation of their activities goes along with a transient activation of the phosphatase HCP, which is thought to be involved

JAK2 in Erythropoietin-Induced Signalling Pathways

in terminating EPO-induced signalling, most probably by dephosphorylating JAK2 [5]. These data suggest that JAK2 may be responsible for coupling the activated EPOR to both signalling pathways. In contrast, JAK2 function is not essential for the EPO-induced activation of the ribosomal S6 kinase p70S6k, which is mediated through both PI 3-kinasedependent and independent pathways (R. Jaster et al., accompanying manuscript). To get further insight, we focused on the contribution of JAK2 for the induction of the Ras/MAP kinase pathway and other EPO-induced signalling events. To address this question, we employed an antisense strategy involving downregulation of JAK2 protein levels in HCD-57 cells. Our data indicate that JAK2-depleted cells are not only diminished in their ability to respond to EPO with enhanced STAT5 DNA-binding, but also activation of MAP kinase, suggesting a significant contribution of JAK2 in the induction of the Ras/MAP kinase pathway, too. It was already shown that a point mutation in a membrane proximal domain of the EPOR abolishes the association of JAK2 as well as Shc phosphorylation and MAP kinase activation [22]. The authors couldn’t find any correlations between the defective signalling events and the mitogenic activity of the cells and concluded that MAP kinase activation may not be required for proliferative signalling. The JAK2depleted cells used in this study are also not diminished in their proliferative capacity, but this may be due to the high EPO levels applied in the propagation of HCD-57 cells. It was recently shown that, using another erythroid cell line and receptor mutants lacking tyrosine phosphorylation sites, high EPO concentrations can compensate for the loss in proliferative capacity [18]. Our data support these observations because the significant reduction of MAP kinase and STAT5 activity in JAK2-depleted cells treated with physiological EPO doses is partially compensated at higher EPO levels (Fig. 4). Taken together, the results provide evidence for the simultaneous activation of the coexistent major signalling pathways, the Ras/MAP kinase and the JAK/STAT cascades and that the receptor-associated tyrosine kinase JAK2 contributes substantially to the induction of both pathways. It remains to be analysed whether JAK2 activates the Ras/ MAP kinase cascade by direct phosphorylation of Shc or interaction with other downstream targets.

89 This work was supported by grants from the Deutsche Forschungsgemeinschaft and the SANDOZ Foundation.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.

Metcalf D. (1989) Nature 339, 27–30. Krantz S. B. (1991) Blood 77, 419–434. Ihle J. N. (1995) Nature 377, 591–594. Damen J. E., Liu L., Cutler R. L. and Krystal G. (1993) Blood 82, 2296–2303. Klingmu¨ller U., Lorenz U., Cantley L. C., Neel B. G. and Lodish H. F. (1995) Cell 80, 729–738. Miura O., Nakamura N., Ihle J. N. and Aoki N. (1994) J. Biol. Chem. 269, 614–620. Harpur A. G., Andres A. C., Ziemiecki A., Aston R. R. and Wilks A. (1992) Oncogene 7, 1347–1353. Witthuhn B. A., Quelle F. W., Silvennoinen O., Yi T., Tang B., Miura O. and Ihle J. N. (1993) Cell 74, 227–236. Gouilleux F., Pallard C., Dusanter-Fourt I., Wakao H., Haldosen L.-A., Norstedt G., Levy D. and Groner B. (1993) EMBO J. 14, 2005–2013. Gouilleux F., Wakao H., Mundt M. and Groner B. (1994) EMBO J. 13, 4361–4369. Wakao H., Gouilleux F. and Groner B. (1994) EMBO J. 13, 2182–2191. Ihle J. N., Witthuhn B. A., Quelle F. W., Yamamoto K. Thierfelder W. E., Kreider B. and Silvennoinen O. (1994) TIBS 19, 222–227. Bittorf Th., Jaster R. and Brock J. (1994) Cell. Signal. 6, 305– 311. Tilbrook P. A., Bittorf Th., Busfield S. J., Chappell D. and Klinken S. P. (1996) J. Biol. Chem. 271, 3453–3459. Tilbrook P. A., Bittorf Th., Callus B., Busfield S. J., Ingley E. and Klinken S. P. K. (1996) Cell Growth Differ. 7, 511–520. David M., Petricoin III E., Benjamin C., Pine R., Weber M. J. and Larner A. C. (1995) Science 269, 1721–1723. Wen Z., Zhong Z. and Darnell J. E. Jr. (1995) Cell 82, 241– 250. Damen J. E., Wakao H., Miyajima A., Krosl J., Humphries R. K., Kutler R. and Krystal G. (1995) EMBO J. 14, 5557– 5568. Spivak J. L., Pham T., Isaaks M. and Hankins W. D. (1991) Blood 77, 1228–1233. Jaster R., Bittorf Th., Klinken S. P. and Brock J. (1996) Biochem. Pharmacol. 51, 1181–1185. Andrews N. C. and Faller D. V. (1991) Nucl. Acids. Res. 19, 2499. Miura Y., Miura O., Ihle J. N. and Aoki N. (1994) J. Biol. Chem. 269, 29962–29969. Miura O., Nakamura N., Quelle F. W., Witthuhn B. A., Ihle J. N. and Aoki N. (1994) Blood 84, 1501–1507. He T.-C., Jiang N., Zhuang H. and Wojchowski D. M. (1995) J. Biol. Chem. 270, 11055–11061.