Expression of a homodimeric type I is required for JAK2V617F-mediated transformation

Xiaohui Lu*†, Ross Levine†‡, Wei Tong*, Gerlinde Wernig‡, Yana Pikman‡, Sara Zarnegar*, D. Gary Gilliland‡§¶, and Harvey Lodish*ʈ

*Whitehead Institute for Biomedical Research and the Department of Biology, Massachusetts Institute of Technology, Cambridge, MA 02142; and ‡Division of Hematology, Department of Medicine, and §Howard Hughes Medical Institute, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA 02115

Contributed by Harvey Lodish, November 9, 2005 A recurrent somatic activating mutation in the nonreceptor ty- says revealed that JAK2V617F transforms certain lines of hema- rosine kinase JAK2 (JAK2V617F) occurs in the majority of patients topoietic cells that express the receptor (EpoR) to with the myeloproliferative disorders vera, essential growth factor independence and induces constitutive phosphory- thrombocythemia, myelofibrosis with myeloid metaplasia, and, lation of the STAT5 transcription factor (9, 11). The specificity of less commonly, chronic myelomonocytic leukemia. We do not the JAK2V617F mutation in clonal disorders of the myeloid, but not understand the basis for the specificity of the JAK2V617F mutation lymphoid, lineage and the pleiotropic phenotype of JAK2V617F- in clonal disorders of the myeloid, but not lymphoid, lineage, nor associated myeloproliferative disorders are not fully understood. has the basis for the pleiotropic phenotype of JAK2V617F-associ- Here, we show that JAK2V617F functions differently from ated myeloproliferative disorders been delineated. However, the constitutively activated kinases, such as TEL-JAK2. In IL-3- presence of the identical mutation in patients with related, but dependent hematopoietic cell lines, the ability of JAK2V617F to clinicopathologically distinct, myeloid disorders suggests that in- induce cytokine-independent activation of the JAK2 and STAT5 teractions between the JAK2V617F kinase and other signaling pathways and transformation to cytokine independence requires molecules may influence the phenotype of hematopoietic progen- the coexpression of homodimeric Type I cytokine receptors, itors expressing JAK2V617F. Here, we show that coexpression of such as EpoR, receptor (TpoR), or granulocyte the JAK2V617F mutant kinase with a homodimeric Type I cytokine colony-stimulating-factor receptor (GCSFR). Further, a mutant receptor, the erythropoietin receptor (EpoR), the thrombopoietin EpoR that fails to bind JAK2 also fails to support JAK2V617F- receptor, or the granulocyte colony-stimulating-factor receptor, is mediated transformation and constitutive STAT5 activation. An necessary for transformation of hematopoietic cells to growth- EpoR mutant that lacks cytosolic tyrosines and is deficient in factor independence and for hormone-independent activation of Epo-induced STAT5 activation supports cytokine-independent JAK-STAT signaling. Furthermore, EpoR mutations that impair activation of JAK2V617F but is defective in supporting erythropoietin-mediated JAK2 or STAT5 activation also impair JAK2V617F-mediated STAT5 activation and cytokine- transformation mediated by the JAK2V617F kinase, indicating that independent growth. Our results suggest that, unlike other JAK2V617F requires a scaffold for its transform- constitutively activating kinase mutants previously identified, ing and signaling activities. Our results reveal the molecular basis JAK2V617F depends on cognate cytokine receptors for both its for the prevalence of JAK2V617F in diseases of myeloid lineage activation and the activation of downstream signal-transduction cells that express these Type I cytokine receptors but not in . This finding provides mechanistic insights into the lymphoid lineage cells that do not. prevalence of JAK2V617F in myeloproliferative diseases like essential thrombocythemia, polycythemia vera, and myeloid erythropoietin receptor ͉ myeloproliferative diseases ͉ STAT5 activation metaplasia and reveals the origins of dysregulation in JAK2V617F-mediated cell growth. onstitutive activation of tyrosine kinases by chromosomal Ctranslocation (1), interstitial deletion (2), internal tandem Materials and Methods duplication (3), or amino acid substitution (4) are frequent Expression Vectors. The murine JAK2 cDNA was cloned into the pathogenetic events in hematopoietic malignancies. These con- retroviral vectors MSCV-Neo and MSCV-GFP. The JAK2V617F stitutively activated kinases are disengaged from normal regu- mutation was generated by using site-directed mutagenesis latory mechanisms and activate signal-transduction pathways, (QuikChange-XL, Stratagene) and confirmed by full-length including the STAT, RAS͞MAPK, and PI3K͞AKT pathways DNA sequencing. that confer proliferative and survival advantage to hematopoi- etic progenitors. For example, TEL-JAK2 is a chimeric tyrosine Cell Culture. The 293T cells were grown in Dulbecco’s modified kinase that is expressed in hematopoietic malignancies as a Eagle’s medium with 10% FCS. Transient cotransfection of consequence of the t (9, 12)(p24;p13) translocation (5, 6). JAK2 293T cells and generation of retroviral supernatant were is normally tightly associated with the cytosolic domain of a performed as described in ref. 14. Parental Ba͞F3 cells, Ba͞F3 cytokine receptor and, after cytokine binding to the cognate receptor, becomes activated by transphosphorylation of a ty- rosine residue in the activation loop. In contrast, TEL-JAK2 is Conflict of interest statement: No conflicts declared. localized to the cytosol and activates STAT5 without a require- Abbreviations: Epo, erythropoietin; EpoR, erythropoietin receptor; GCSFR, granulocyte colony-stimulating-factor receptor; PrlR, ; TpoR, thrombopoietin ment for cytokine stimulation or binding to a cytokine receptor receptor. (7). Indeed, the TEL-JAK2 fusion lacks the conserved JH3 to †X.L. and R.L. contributed equally to this work. JH7 domains that are required for JAK2 association with ¶To whom correspondence may be addressed at: Brigham and Women’s Hospital, 1 cytokine receptors (8). Blackfan Circle, Room 5120, Harvard Medical School, Boston, MA 02115. E-mail: ggilliland@ We and others recently reported an association between a rics.bwh.harvard.edu. somatic point mutation of the JAK2 tyrosine kinase (JAK2V617F) ʈTo whom correspondence may be addressed at: 9 Cambridge Center, Cambridge, and the myeloproliferative diseases polycythemia vera, essential MA 02142. E-mail: [email protected]. thrombocythemia, and myeloid metaplasia (9–13). Functional as- © 2005 by The National Academy of Sciences of the USA

18962–18967 ͉ PNAS ͉ December 27, 2005 ͉ vol. 102 ͉ no. 52 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0509714102 Downloaded by guest on October 2, 2021 Fig. 1. A homodimeric is required for JAK2V617F-mediated transformation of Ba͞F3 cells. (a) Parental Ba͞F3 cells stably expressing wild-type JAK2 (Ba͞F3 WT-JAK2, open circles), JAK2V617F (Ba͞F3 JAK2V617F, open triangles), or TEL-JAK2 fusion cDNA (Ba͞F3 TEL-JAK2, open squares) were cultured in RPMI medium 1640͞10% FCS in the absence of IL-3, demonstrating cytokine-independent growth for Ba͞F3 cells expressing TEL-JAK2 but not wild-type JAK2 or JAK2V617F. Each data point represents the average of a duplicate, if not specified otherwise. (b)Ba͞F3 EpoR cells stably expressing wild-type JAK2 (Ba͞F3 EpoR WT-JAK2, filled circles), JAK2V617F (Ba͞F3 EpoR JAK2V617F, filled triangles), or TEL-JAK2 fusion cDNA (Ba͞F3 EpoR TEL-JAK2, filled squares) were cultured in RPMI medium 1640͞10% FCS in the absence of IL-3, demonstrating cytokine-independent growth for Ba͞F3 EpoR cells expressing either TEL-JAK2 or JAK2V617F but not wild-type JAK2. (c)Ba͞F3 cells stably coexpressing TpoR and JAK2V617F (Ba͞F3 TpoR JAK2V617F, filled triangles), JAK2V617F alone (Ba͞F3 JAK2V617F, open triangles), wild-type JAK2 alone (Ba͞F3 WT-JAK2, open circles), or TpoR together with wild-type JAK2 (Ba͞F3 TpoR WT-JAK2, filled circles) were cultured in RPMI medium 1640͞10% FCS in the absence of IL-3, demonstrating IL-3-independent growth for Ba͞F3 cells stably coexpressing TpoR and JAK2V617F but not cells expressing wild-type JAK2 or JAK2V617F alone. (d)Ba͞F3 cells stably coexpressing GCSFR and JAK2V617F (Ba͞F3 GCSFR JAK2V617F, filled triangles), expressing only JAK2V617F (Ba͞F3 JAK2V617F, open triangles), only wild-type JAK2 (Ba͞F3 WT-JAK2, open circles), or GCSFR together with wild-type JAK2 (Ba͞F3 GCSFR WT-JAK2, filled circles) were cultured in RPMI medium 1640͞10% FCS in the absence of IL-3, demonstrating IL-3-independent growth only for Ba͞F3 cells stably coexpressing JAK2V617F and GCSFR.

cells expressing the murine EpoR (BaF3 EpoR) (15), and Results Ba͞F3 cells expressing murine TpoR (Ba͞F3 TpoR) (14) were Expression of a Homodimeric Type I Cytokine Receptor Is Required for grown in RPMI medium 1640 containing 10% FCS and 10% JAK2V617F-Mediated Transformation of Ba͞F3 Cells. We demon- WEHI-3B cell supernatant as a source of IL-3 (WEHI media). strated that expression of TEL-JAK2 transforms the murine These cells were transduced with retroviral supernatant con- IL-3-dependent hematopoietic cell line Ba͞F3 to factor- taining either the MSCV-JAK2-Neo or MSCV-JAK2V617F- independent growth (7) (Fig. 1a). These properties are similar to Neo vectors, respectively, and then selected in G418 (1 mg͞ those reported by our group and others for a number of mutant ml). Ba͞F3 cells expressing human GCSFR (Ba͞F3 GCSFR) tyrosine kinases associated with hematopoietic malignancy that (16) and parental Ba͞F3 cells were transduced with retroviral each confer factor-independent growth to Ba͞F3 cells, including CELL BIOLOGY supernatant containing either the MSCV-JAK2-GFP or BCR-ABL, TEL-PDGFRB, HIP1-PDGFRB, H4-PDGFRB, MSCV-JAK2V617F-GFP vectors, respectively, and then TEL-ABL, ZNF198-FGFR1, FIP1L1-PDGFRA, NPM-ALK, sorted by flow cytometry to isolate GFP-positive cells. To FLT3-ITD, FLT3 activation-loop mutations, and KIT activation- assess for factor-independent growth, cells were washed three loop mutations (2, 17–26). In striking contrast, we observed that times in RPMI medium 1640 and cultured in the absence of expression of JAK2V617F in Ba͞F3 cells did not confer IL-3- Epo and IL-3 for 7 days. The number of viable cells was independent growth (Fig. 1a). We next tested Ba͞F3 cells that determined by trypan blue exclusion. express an ectopic EpoR, called Ba͞F3 EpoR cells. Importantly, JAK2V617F, but not wild-type JAK2, transformed Ba͞F3 EpoR Immunoprecipitation and Western Blotting. The following antibod- cells to cytokine-independent growth (Fig. 1b). ies were used for immunoprecipitation and Western blot Cell-surface EpoR is a dimer in the absence of hormone (27). analysis: anti-JAK2 (polyclonal antibody, Upstate Biotechnol- Epo binding reorients the receptor extracellular and membrane- ogy, Lake Placid, NY), anti-STAT5 (c-17, Santa Cruz Bio- spanning domains, such that the two attached JAK2 kinases are technology), anti-phosphotyrosine (4G10) (Upstate Biotech- brought into apposition, and one phosphorylates the critical nology), anti-phospho-STAT5 (Cell Signaling Technology, tyrosine in the activation loop of the other kinase. Among Type Beverly, MA), and anti-EpoR (m-20, Santa Cruz Biotechnol- I cytokine receptors, TpoR, GCSFR, and prolactin receptor ogy) antibodies. (PrlR) are most closely related to EpoR, in that each forms a

Lu et al. PNAS ͉ December 27, 2005 ͉ vol. 102 ͉ no. 52 ͉ 18963 Downloaded by guest on October 2, 2021 JAK2V617F to a dimerized cytokine receptor is essential for cytokine-independent autotransphosphorylation that leads to ki- nase activation. Indeed, Fig. 2a shows that expression of JAK2V617F in Ba͞F3 cells expressing EpoR, TpoR, or GCSFR results in a high level of JAK2 tyrosine phosphorylation in the absence of cytokine stimu- lation. In contrast, only background levels of JAK2 phosphorylation were observed after expression of JAK2V617F in parental Ba͞F3 cells or after expression of wild-type JAK2 in parental Ba͞F3 cells or Ba͞F3 cells expressing the EpoR, TpoR, or GCSFR. After cytokine binding to Type I cytokine receptor and the activation of JAK2, JAK2 phosphorylates multiple tyrosines in the receptor cytosolic domain; these tyrosines serve as ‘‘docking sites’’ for the binding and subsequent tyrosine phosphorylation of mul- tiple signal-transduction proteins, including STAT5. As expected, tyrosine phosphorylation of STAT5 in Ba͞F3 cells expressing EpoR, TpoR, or GCSFR and overexpressing wild-type JAK2 cannot be detected in the absence of cytokine stimulation. In contrast, there are high levels of tyrosine-phosphorylated STAT5 in Ba͞F3 EpoR, Ba͞F3 TpoR, and Ba͞F3 GCSFR cells expressing JAK2V617F in the absence of cytokine, whereas no tyrosine- phosphorylated STAT5 is seen in parental Ba͞F3 cells expressing Fig. 2. JAK2V617F constitutively activates the JAK2͞STAT5 pathway only JAK2V617F (Fig. 2b). Thus, cytokine-independent phosphoryla- when a homodimeric type I cytokine receptor is coexpressed. Twenty-five ͞ ͞ ͞ ͞ ͞ tion of both JAK2 and STAT5 in Ba F3 cells expressing million parental Ba F3, Ba F3 EpoR, Ba F3 TpoR, or Ba F3 GCSFR cells express- JAK2V617F, like cytokine-independent cell proliferation, requires ing either wild-type JAK2 or JAK2V617F were depleted of for 4 h. Cells were lysed and immunoprecipitated with antibodies specific for JAK2 (a) coexpression of a Type I cytokine receptor. or STAT5 (b) and then analyzed by Western blots using anti-phosphotyrosine antibody 4G10 (a Upper) or anti-phospho-STAT5 antibody (b Upper). Total EpoR Association Is Essential for JAK2V617F-Mediated Transforma- JAK2 and STAT5 in the immunoprecipitation pellets were detected with tion. JAK2 binds nascent EpoR and facilitates its surface ex- antibodies against JAK2 and STAT5 as control (a and b Lower). Arrows in a pression. W282, a highly conserved amino acid in the EpoR Upper indicate enhanced JAK2 phosphorylation in Ba͞F3 EpoR, Ba͞F3 TpoR, cytosolic domain, is important for this functional interaction, and Ba͞F3 GCSFR cells expressing JAK2V617F but not in cells expressing because the W282A mutation disrupts JAK2-enhanced EpoR wild-type JAK2 or in parental Ba͞F3 cells expressing JAK2V617F mutant; ͞ surface expression (8). The W282R mutation in the EpoR also Arrows in b indicate constitutively phosphorylated STAT5 in Ba F3 EpoR, abolishes productive JAK2 binding and Epo-dependent JAK2 Ba͞F3 TpoR, and Ba͞F3 GCSFR cells expressing JAK2V617F but not in cells ͞ activation (29, 30). Compared with Ba͞F3 cells expressing expressing wild-type JAK2 or in parental Ba F3 cells expressing JAK2V617F ͞ mutant. wild-type EpoR, Ba F3 cells expressing EpoRW282R exhibit no Epo-induced JAK2 phosphorylation, STAT5 phosphorylation, or EpoR phosphorylation (Fig. 3b–d, lanes 2 and 4). In contrast specifically oriented receptor homodimer after hormone bind- to Ba͞F3 EpoR or 32D EpoR cells coexpressing JAK2V617F, ing, activates the associated JAK2, and, in turn, activates the Ba͞F3 or 32D cells coexpressing EpoRW282R and JAK2V617F STAT5 transcription factor (28). Fig. 1 c and d shows that Ba͞F3 do not undergo factor-independent proliferation (Fig. 4; and see cells expressing TpoR or GCSFR are also transformed by Fig. 8, which is published as supporting information on the PNAS JAK2V617F into cytokine-independent growth. In contrast, web site), and there is no cytokine-independent tyrosine phos- parental Ba͞F3 cells expressing only JAK2V617F, Ba͞F3 TpoR, phorylation of either JAK2V617F or STAT5, (Fig. 5, lane 6). or Ba͞F3 GCSFR cells expressing wild-type JAK2 remain Thus, a functional interaction between JAK2 and EpoR is untransformed. We made similar observations with Ba͞F3 cells required for both Epo-dependent JAK2 activation and Epo- expressing the PrlR and also with the IL-3-dependent 32D cell independent activation of JAK2V617F. line expressing the EpoR, EpoR-TpoR, or EpoR-PrlR chimeras containing the extracellular domain of the EpoR and the EpoR Provides Scaffolds for the Activation of STAT5 and for the membrane-spanning and cytosolic domains of the TpoR or PrlR, Maximal Cell Growth Mediated by JAK2V617F. The EpoR mutant EpoRF8 has all eight of its intracellular tyrosine residues respectively (see Figs. 6 and 7, which are published as supporting mutated to phenylalanine (31), and, like EpoRW282R, it cannot information on the PNAS web site). support Epo-dependent growth when expressed in Ba͞F3 cells (Fig. 3a). When expressed in Ba͞F3 cells, EpoRF8 supports Constitutive Activation of JAK2V167F and Downstream STAT5 Require normal Epo-dependent activation of JAK2, but, as expected, the Presence of a Homodimeric Type I Cytokine Receptor. The above there is no Epo-dependent tyrosine phosphorylation of the findings suggest that, in contrast to all other known activating alleles EpoR and greatly reduced STAT5 phosphorylation (Fig. 3) (31). of tyrosine kinases associated with hematopoietic malignancies, (The low level of STAT5 phosphorylation may result from transformation to cytokine-independence by JAK2V617F requires binding of STAT5 to phosphotyrosines in JAK2 or may depend its binding to the cytosolic domain of a cytokine receptor. Cytokine- on other kinases that are activated by Epo stimulation.) Acti- receptor-triggered JAK2 transphosphorylation of the critical ty- vation of other Epo-dependent signaling pathways, such as the rosine residue in the activation loop requires the physical proximity PI-3 kinase͞AKT pathway is also greatly reduced in cells ex- of two JH1 kinase domains. In TEL-JAK2, the pointed domain of pressing EpoRF8, because these, too, depend on docking of SH2 the ets transcription factor TEL directly dimerizes the JH1 domain domains of signaling proteins to phosphotyrosines in the EpoR and induces constitutive kinase activation (5, 7). Unlike TEL-JAK2, cytosolic domain (32). Figs. 4 and 8 show that Ba͞F3 or 32D cells cytosolic JAK2V617F should not be able to form JAK2 ho- coexpressing EpoRF8 and JAK2V617F undergo cytokine- modimers. Because JAK2V617F retains domains JH3 to JH7 that independent proliferation at a much lower rate than do Ba͞F3 mediate receptor association (8), we hypothesize that binding of or 32D cells expressing wild-type EpoR and JAK2V617F, sug-

18964 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0509714102 Lu et al. Downloaded by guest on October 2, 2021 Fig. 3. EpoRW282R and EpoRF8 are defective in supporting EpoR signaling in the presence of Epo. (a) Parental Ba͞F3 cells (filled squares) and Ba͞F3 cells expressing either wild-type EpoR (filled triangles), EpoRW282R (filled diamonds), or EpoRF8 (filled circles) were grown in RPMI medium 1640͞10%FCS medium supplemented with 0.1 units͞ml Epo. Cell numbers were counted at days 2, 4, and 6. Only Ba͞F3 cells expressing wild-type EpoR exhibit sustained growth under this condition. (b) To measure EpoR phosphorylation, Ba͞F3 cells expressing wild-type EpoR, EpoRW282R, and EpoRF8 were grown in RPMI medium 1640 containing IL-3 then starved of cytokine for 4 h and then treated or not with 10 units͞ml Epo for 10 min. Lysates were subjected to immunoprecipitation with anti-phospho-tyrosine antibody PY99, followed by Western blotting using an anti-EpoR antibody. EpoR phosphorylation cannot be detected in Ba͞F3 cells expressing EpoRW282R or EpoRF8. (c and d)Ba͞F3 cells expressing wild-type EpoR, EpoRW282R, and EpoRF8 were deprived of IL-3 and Epo for 4 h before stimulation with 10 units͞ml Epo for 10 min. Equal numbers of stimulated and unstimulated cells were lysed, and JAK2 and STAT5 proteins were immunopre- cipitated with respective antibodies. Phosphorylation of JAK2 was detected by anti-phosphotyrosine antibody 4G10 (c), and phosphorylation of STAT5 was detected by anti-phospho-STAT5 antibody (d).

gesting that the phosphotyrosine scaffold of EpoR plays a major orientation that the activation loop of one JAK2 molecule can role in transforming Ba͞F3 cells by JAK2V617F. be phosphorylated by the other JAK2. Once these JAK2 The supposition that tyrosine residues in the EpoR cytosolic become activated by reciprocal transphosphorylation, they domain are essential for maximal cytokine-independent signaling phosphorylate several tyrosine residues in the EpoR intracel- by JAK2V617F is supported by the experiment shown in Fig. 5 lular domain, creating docking sites for SH2 domains of several (lanes 7 and 8). In Ba͞F3 EpoRF8 cells coexpressing JAK2V617F, signal-transduction proteins, such as STAT5, Shc, Grb2, and PI cytokine-independent tyrosine phosphorylation of the JAK2V617F 3Ј kinase (35). These events lead to the activation of multiple kinase occurs normally. However, there is a very low level of intracellular signaling pathways and specific inductions tyrosine phosphorylation of STAT5, presumably because essential and repressions, resulting in the survival, proliferation, and tyrosines in the EpoR cytosolic domain are missing. differentiation of erythroid progenitors (35).

Discussion Signal Transduction by JAK2V617F Requires a Cytokine Receptor as a JAK2 Plays a Central Role in the Signal Transduction by Homodimeric Scaffold. In one current structural model of JAK2, the activation Type I Cytokine Receptors Such as the EpoR. Without an intrinsic loop of JH1 kinase domain is positioned such that substrate kinase domain, cytokine receptors rely on members of the JAK cannot bind. The JH2 pseudokinase domain inhibits the activa- family of nonreceptor tyrosine kinases, such as JAK2, for their tion of the adjacent JH1 kinase domain by prohibiting the JH1 signaling. JAK2 is comprised of seven JAK homology (JH) activation loop from flipping out of the substrate-binding pocket, domains, denoted JH1–JH7. Segments JH3–JH7, containing a so that the crucial tyrosine residues in the JAK2 JH1-activation

FERM and SH2 domain, are essential for its binding to the loop cannot be reciprocally phosphorylated by adjacent JH1 CELL BIOLOGY membrane-proximal segment of the EpoR cytosolic domain (8). kinases, and JAK2 remains inactive (36). V617 in the JH2 Conserved regions in the EpoR cytosolic domain, including domain is thought to be located on the JH1–JH2 interface, W282, are important for this interaction (30). The C-terminal directly interacting with the activation loop of JH1 kinase JH1 domain of JAK2 is a catalytically active tyrosine kinase, domain. whereas the adjacent JH2 pseudokinase domain suppresses the The V617F mutation is thought to disrupt inhibition of JH1 by basal kinase activity of the JH1 domain (33). JH2, such that the activation loop of JH1 becomes extended and Many Type I cytokine receptors, such as EpoR, exist as then phosphorylated by an adjacent JAK2V617F kinase. For this homodimers on the cell membrane (27), and we presume that, to happen, two JAK2 molecules must be physically close to each in the absence of Epo, the appended JAK2 kinases are held in other. Unlike TEL-JAK2, which is a dimer and constitutively the ‘‘off’’ or nonphosphorylated state. Epo binding apparently active, JAK2V617F is a monomeric cytosolic and is changes the orientation of the two receptor extracellular inactive in the absence of cognate cytokine receptors, as we domains, possibly causing a rotation of the receptor trans- show here. membrane domains that is transmitted to the membrane- Our most important result is that coexpression of a ho- proximal cytosolic domains where JAK2 is bound (34). modimeric Type I cytokine receptor is essential for cell trans- Through unknown mechanisms, this conformational change formation by the mutant JAK2V617F; EpoR, PrlR, TpoR, or displaces the JH2 inhibition of the JH1 kinase domain and GCSFR can suffice. The receptor cytosolic domain presumably juxtaposes the two appended JAK2 molecules in such an binds JAK2V617F via its N-terminal FERM and SH2 domains

Lu et al. PNAS ͉ December 27, 2005 ͉ vol. 102 ͉ no. 52 ͉ 18965 Downloaded by guest on October 2, 2021 Fig. 4. EpoRW282R and EpoRF8 are defective in supporting cytokine- independent growth of Ba͞F3 cells expressing JAK2V617F. (a)Ba͞F3 cells stably expressing wild-type JAK2 and wild-type EpoR (open triangles), EpoRW282R (open diamonds), or EpoRF8 (open circles) were cultured in RPMI Fig. 5. EpoRW282R and EpoRF8 are defective in activating JAK2 or STAT5 in ͞ ͞ ͞ medium 1640͞10% FCS in the absence of IL-3; no cell growth was observed. (b) Ba F3 cells expressing JAK2V617F. Equal numbers of parental Ba F3, Ba F3 ͞ ͞ Ba͞F3 cells stably expressing JAK2V617F, and the wild-type EpoR cells (filled EpoR, Ba F3 EpoRW282R, or Ba F3 EpoRF8 cells stably expressing either ͞ triangles), EpoRW282R (filled diamonds), or EpoRF8 (filled circles) were cul- wild-type JAK2 or JAK2V617F were cultured in RPMI medium 1640 10% FCS tured in RPMI medium 1640͞10% FCS in the absence of IL-3; maximum growth in the presence of IL-3 and then deprived of cytokines for 4 h. Cells were lysed was observed only in cells coexpressing JAK2V617F and wild-type EpoR. Cells and immunoprecipitated with antibodies specific for JAK2 (a) or STAT5 (b), coexpressing EpoR F8 and JAK2V617F grew at a considerably slower rate, and phosphorylation of JAK2 and STAT5 in these cells was analyzed by whereas no growth was observed in cells coexpressing EpoRW282R and Western blot using anti-phosphotyrosine antibody 4G10 (a Upper) or anti- JAK2V617F. Each data point represents the average of two independent phospho-STAT5 antibody (b Upper). Total JAK2 and STAT5 in the immuno- experiments. precipitates were detected by Western blots using antibodies against JAK2 and STAT5 (a and b Lower). (a) Enhanced JAK2 phosphorylation is observed in Ba͞F3 EpoR and Ba͞F3 EpoRF8 cells expressing JAK2V617F (lanes 4 and 8) but ͞ (8). When JAK2V617F proteins bind to each of the EpoR not in cells expressing wild-type JAK2 (lanes 1, 3, 5, and 7) or Ba F3 EpoR W282 cells expressing JAK2V617F (lane 6), suggesting that EpoRW282R cannot subunit in a receptor homodimer, the weakened JH2 inhibition support ligand-independent JAK2 activation in cells expressing JAK2V617F. of the JH1 kinase domain allows one JAK2 kinase domain to (b) In cells expressing wild-type JAK2, phosphorylation of STAT5 is barely phosphorylate the critical tyrosine on the other, even in the detectable (lanes 1, 3, 5, and 7). In cells expressing JAK2V617F, a high level of absence of cytokine binding. This result is supported by our phosphorylated STAT5 is detected only in Ba͞F3 EpoR cells expressing finding that EpoRW282R is defective in both Epo-dependent JAK2V617F (lane 4) but not in parental Ba͞F3 or Ba͞F3 EpoRW282R cells (lanes activation of wild-type JAK2 and in cytokine-independent acti- 2 and 6). A low level of STAT5 phosphorylation is detected in Ba͞F3 EpoRF8 vation of JAK2V617F. cells expressing JAK2V617F mutant (lane 8), correlating with the reduced rate Cytosolic tyrosines in the EpoR cytosolic domain are dis- of proliferation without cytokines (Fig. 4). pensable for the tethering of JAK2 to the EpoR and Epo- dependent activation of JAK2 (Fig. 3b); however, cytosolic important therapeutic strategy in JAK2V617F-positive myelo- tyrosines are crucial for the maximal cell growth and activation proliferative disorders. of downstream signaling molecules induced by Epo. EpoRF8, Epo binding triggers conformational changes in the EpoR lacking all cytosolic tyrosines, shows much reduced cytokine- intracellular domains. This structural change may allow an independent growth conferred by JAK2V617F, indicating that appended JAK2V617F to assume a more favorable orientation EpoR not only serves as an anchorage for constitutive activa- for JAK2V617F transphosphorylation and͞or JAK2-mediated tion of JAK2V617F, but EpoR’s ability to recruit downstream EpoR and͞or STAT5 phosphorylation, which may result in the signaling molecules also plays a crucial role in the transforming observed Epo hypersensitivity (8). capability of JAK2V617F. Cytosolic tyrosines in the EpoR We do not know whether receptors or other proteins besides cytosolic domain provide a scaffold that is essential for homodimeric Type I cytokine receptors can bind and activate activation of downstream signal-transduction proteins, espe- JAK2V617F. The IL-3 receptor, a heterodimeric Type I recep- cially STAT5, by EpoR-attached JAK2V617F, as evidenced by tor, is endogenously expressed in our parental Ba͞F3 and 32D our finding that EpoRF8 does not fully activate STAT5. These cells, because these grow in the presence of IL-3. However, this data also suggest that activation of STAT5 signaling is critical level of expression of the IL-3 receptor is insufficient to support for JAK2V617F-mediated cytokine-independent proliferation JAK2V617F-mediated transformation. It is important to assess and that inhibition of this signaling pathway may represent an other cytokine receptors for their ability to engage the trans-

18966 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0509714102 Lu et al. Downloaded by guest on October 2, 2021 forming potential of JAK2V617F. Our observation that the cells that express these Type I cytokine receptors but not in ability of JAK2V617F to activate STAT5 and transform IL-3- lymphoid lineage cells that do not. dependent cells requires coexpression of homodimeric Type I cytokine receptors may provide insights into phenotypic pleiot- We thank Dr. Dan Link (Washington University, St. Louis) for kindly ropy of diseases mediated by this allele and into the natural providing the Ba͞F3 GCSFR cell line. This work was supported, in part, selection of a single activating allele of JAK2V617F. Further- by National Institutes of Health Grants P01 HL 32262 (to H.L.) and more, the ability of the EpoR, TpoR, and GCSFR to bind and DK50654 and CA66996 (to D.G.G.), the Leukemia and Lymphoma activate JAK2V617F provides a molecular basis for the preva- Society (D.G.G.), the Doris Duke Charitable Foundation (D.G.G.), the lence of JAK2V617F in polycythemia vera, essential thrombo- Howard Hughes Medical Institute (D.G.G.), and a research grant from cythemia, and myeloid metaplasia, diseases of myeloid-lineage Amgen, Inc. (to H.L.).

1. Bartram, C. R., de Klein, A., Hagemeijer, A., van Agthoven, T., Geurts van 18. Carroll, M., Tomasson, M. H., Barker, G. F., Golub, T. R. & Gilliland, D. G. Kessel, A., Bootsma, D., Grosveld, G., Ferguson-Smith, M. A., Davies, T., (1996) Proc. Natl. Acad. Sci. USA 93, 14845–14850. Stone, M., et al. (1983) Nature 306, 277–280. 19. Ross, T. S., Bernard, O. A., Berger, R. & Gilliland, D. G. (1998) Blood 91, 2. Cools, J., DeAngelo, D. J., Gotlib, J., Stover, E. H., Legare, R. D., Cortes, J., 4419–4426. Kutok, J., Clark, J., Galinsky, I., Griffin, J. D., et al. (2003) N. Engl. J. Med. 348, 20. Schwaller, J., Anastasiadou, E., Cain, D., Kutok, J., Wojiski, S., Williams, I. R., 1201–1214. LaStarza, R., Crescenzi, B., Sternberg, D. W., Andreasson, P., et al. (2001) 3. Nakao, M., Yokota, S., Iwai, T., Kaneko, H., Horiike, S., Kashima, K., Sonoda, Blood 97, 3910–3918. Y., Fujimoto, T. & Misawa, S. (1996) Leukemia 10, 1911–1918. 21. Golub, T. R., Goga, A., Barker, G. F., Afar, D. E., McLaughlin, J., Bohlander, 4. Yamamoto, Y., Kiyoi, H., Nakano, Y., Suzuki, R., Kodera, Y., Miyawaki, S., S. K., Rowley, J. D., Witte, O. N. & Gilliland, D. G. (1996) Mol. Cell. Biol. 16, Asou, N., Kuriyama, K., Yagasaki, F., Shimazaki, C., et al. (2001) Blood 97, 4107–4116. 2434–2439. 22. Smedley, D., Demiroglu, A., Abdul-Rauf, M., Heath, C., Cooper, C., Shipley, 5. Lacronique, V., Boureux, A., Valle, V. D., Poirel, H., Quang, C. T., J. & Cross, N. C. (1999) Neoplasia 1, 349–355. Mauchauffe, M., Berthou, C., Lessard, M., Berger, R., Ghysdael, J. & Bernard, 23. Bai, R. Y., Dieter, P., Peschel, C., Morris, S. W. & Duyster, J. (1998) Mol. Cell. O. A. (1997) Science 278, 1309–1312. Biol. 18, 6951–6961. 6. Peeters, P., Raynaud, S. D., Cools, J., Wlodarska, I., Grosgeorge, J., Philip, P., 24. Hayakawa, F., Towatari, M., Kiyoi, H., Tanimoto, M., Kitamura, T., Saito, H. Monpoux, F., Van Rompaey, L., Baens, M., Van den Berghe, H. & Marynen, & Naoe, T. (2000) Oncogene 19, 624–631. Blood 90, P. (1997) 2535–2540. 25. Grundler, R., Thiede, C., Miething, C., Steudel, C., Peschel, C. & Duyster, J. 7. Schwaller, J., Frantsve, J., Aster, J., Williams, I. R., Tomasson, M. H., Ross, (2003) Blood 102, 646–651. T. S., Peeters, P., Van Rompaey, L., Van Etten, R. A., Ilaria, R., Jr., et al. (1998) 26. Kitayama, H., Kanakura, Y., Furitsu, T., Tsujimura, T., Oritani, K., Ikeda, H., EMBO J. 17, 5321–5333. Sugahara, H., Mitsui, H., Kanayama, Y., Kitamura, Y., et al. (1995) Blood 85, 8. Huang, L. J., Constantinescu, S. N. & Lodish, H. F. (2001) Mol. Cell 8, 790–798. 1327–1338. 27. Constantinescu, S. N., Keren, T., Socolovsky, M., Nam, H., Henis, Y. I. & 9. Levine, R. L., Wadleigh, M., Cools, J., Ebert, B. L., Wernig, G., Huntly, B. J., Lodish, H. F. (2001) Proc. Natl. Acad. Sci. USA 98, 4379–4384. Boggon, T. J., Wlodarska, I., Clark, J. J., Moore, S., et al. (2005) Cancer Cell 28. Watowich, S. S., Wu, H., Socolovsky, M., Klingmuller, U., Constantinescu, S. N. 7, 387–397. 12, 10. Baxter, E. J., Scott, L. M., Campbell, P. J., East, C., Fourouclas, N., Swanton, & Lodish, H. F. (1996) Annu. Rev. Cell Dev. Biol. 91–128. S., Vassiliou, G. S., Bench, A. J., Boyd, E. M., Curtin, N., et al.(2005) Lancet 29. Miura, O., Cleveland, J. L. & Ihle, J. N. (1993) Mol. Cell. Biol. 13, 1788–1795. 365, 1054–1061. 30. Miura, Y., Miura, O., Ihle, J. N. & Aoki, N. (1994) J. Biol. Chem. 269, 11. James, C., Ugo, V., Le Couedic, J. P., Staerk, J., Delhommeau, F., Lacout, C., 29962–29969. Garcon, L., Raslova, H., Berger, R., Bennaceur-Griscelli, A., et al. (2005) 31. Klingmuller, U., Bergelson, S., Hsiao, J. G. & Lodish, H. F. (1996) Proc. Natl. Nature 434, 1144–1148. Acad. Sci. USA 93, 8324–8328. 12. Kralovics, R., Passamonti, F., Buser, A. S., Teo, S. S., Tiedt, R., Passweg, J. R., 32. Klingmuller, U., Wu, H., Hsiao, J. G., Toker, A., Duckworth, B. C., Cantley, Tichelli, A., Cazzola, M. & Skoda, R. C. (2005) N. Engl. J. Med. 352, 1779–1790. L. C. & Lodish, H. F. (1997) Proc. Natl. Acad. Sci. USA 94, 3016–3021. 13. Levine, R. L., Loriaux, M., Huntly, B. J., Loh, M., Beran, M., Stoffregen, E., 33. Saharinen, P., Takaluoma, K. & Silvennoinen, O. (2000) Mol. Cell. Biol. 20, Berger, R., Clark, J. J., Willis, S. G., Nguyen, K., et al. (2005) Blood 106, 3387–3395. 3377–3379. 34. Seubert, N., Royer, Y., Staerk, J., Kubatzky, K. F., Moucadel, V., Krishnaku- 14. Tong, W. & Lodish, H. F. (2004) J. Exp. Med. 200, 569–580. mar, S., Smith, S. O. & Constantinescu, S. N. (2003) Mol. Cell 12, 1239–1250. 15. D’Andrea, A. D., Yoshimura, A., Youssoufian, H., Zon, L. I., Koo, J. W. & 35. Wojchowski, D. M., Gregory, R. C., Miller, C. P., Pandit, A. K. & Pircher, T. J. Lodish, H. F. (1991) Mol. Cell. Biol. 11, 1980–1987. (1999) Exp. Cell Res. 253, 143–156. 16. Jacob, J., Haug, J. S., Raptis, S. & Link, D. C. (1998) Blood 92, 353–361. 36. Lindauer, K., Loerting, T., Liedl, K. R. & Kroemer, R. T. (2001) Protein Eng. 17. Daley, G. Q. & Baltimore, D. (1988) Proc. Natl. Acad. Sci. USA 85, 9312–9316. 14, 27–37. CELL BIOLOGY

Lu et al. PNAS ͉ December 27, 2005 ͉ vol. 102 ͉ no. 52 ͉ 18967 Downloaded by guest on October 2, 2021