The regulation of JAKs in cytokine signaling and its breakdown in disease

Contents lists available at ScienceDirect Cytokine journal homepage: www.elsevier.com/locate/cytokine Review article The regulation of JAKs in cytokine signaling and its breakdown in disease Henrik M. Hammarén a,⁎ , Anniina T. Virtanen a , Juuli Raivola a , Olli Silvennoinen a,b,c a Faculty of Medicine and Life Sciences, University of Tampere, Arvo Ylpön katu 34, FI-33014 Tampere, Finland b Clinical Hematology, Department of Internal Medicine, Tampere University Hospital, Medisiinarinkatu 3, FI-33520 Tampere, Finland c Fimlab Laboratories, Pirkanmaa Hospital District, Tampere, Finland ARTICLE INFO Keywords: Janus kinase (JAK) Cytokine signaling Myeloproliferative neoplasm (MPN) Cancer Mutation ABSTRACT The JAK–STAT signal transduction pathway is responsible for mediating signals of over fifty cytokines, growth factors and hormones. Signaling through the JAK–STAT pathway is regulated on multiple levels, including intramolecular regulation by the JAK pseudokinase domain, and intermolecular regulation by a host of regulatory proteins. The advent of accessible genomic tools have provided a wealth of information on disease-associated mutations in the JAK–STAT pathway and its regulatory components. The vast number of these mutations in diseases ranging from immunodeficiencies and obesity to many cancers highlight the importance of correct regulation of JAK–STAT signaling for biological processes such as hematopoiesis, regulation of the immune system, metabolism, and growth. Simultaneously, JAK inhibitors are gaining traction in clinical use, both for treatment of diseases driven by JAK mutations, and for a host of inflammatory disorders, in which proinflammatory cytokine signaling through the JAK–STAT pathway is an integral part of pathogenesis. The elucidation of molecular mechanisms in the pathogenesis of complex diseases has also, however, brought the limitations of our current understanding on the regulation of cytokine signaling to the foreground. Indeed, deeper understanding of these regulatory mechanisms are a prerequisite for the development of the next generation of pharmacological modulators of the JAK–STAT pathway. In this review we discuss the current state of knowledge of the intraand intermolecular regulation of the JAK–STAT pathway, with a focus on diseases arising from disruptions in the regulatory apparatus. 1. Introduction Hematopoietic cytokines play a critical role in orchestrating fundamental processes such as the immune response and hematopoiesis, but also cell differentiation and growth. Thereby it is not surprising that the central cytokine signaling pathway through Janus kinases (JAKs) and signal transducer and activator of transcription (STAT) transcription factors has been shown to either cause or participate in https://doi.org/10.1016/j.cyto.2018.03.041 Received 30 November 2017; Received in revised form 29 March 2018; Accepted 30 March 2018 ⁎ Corresponding author. E-mail address: henrik.hammaren@uta.fi(H.M. Hammarén). Abbreviations: AL, Activation loop; ALL, Acute lymphoblastic leukemia; AML, Acute myeloid leukemia; AMKL, Acute megakaryoblastic leukemia; βc, Beta common; B-ALL, B-cell acute lymphoblastic leukemia; BCR, B-cell receptor; BSF-3, B-cell stimulating factor-3; CALR, Calreticulin; CLC, Cardiotrophin-like cytokine; CLCF1, Cardiotrophin-like cytokine factor 1; CLF, Colony-stimulating factor; CML, Chronic myeloid leukemia; CNTF, Ciliary neurotrophic factor; CNTFRα, Ciliary neurotrophic factor receptor αsubunit; CRLF2, Cytokine receptor-like factor 2; CSF3R, Granulocyte colony-stimulating factor receptor; CT-1, Cardiotrophin 1; del, Deletion; DS-AMKL, Down syndrome AMKL; EBI3, Epstein-Barr virus induced gene 3; EGFR, Epidermal growth factor; EPO(R), Erythropoietin (receptor); ET, Essential thrombocythemia; ETV6, Ets variant 6 (TEL); ETP-ALL, Early T-cell precursor acute lymphoblastic leukemia/ lymphoma; FERM, Band 4.1 protein, ezrin, radixin and moesin; fs, Frame-shift; γc, Gamma common; GCSFR, Granulocyte colony-stimulating factor receptor; GH(R), Growth hormone (receptor); GLMR, GP130-like monocyte receptor; GM-CSF-Rα, Granulocyte-macrophage colony-stimulating factor receptor α; GOF, Gain-of-function; gp130, Glycoprotein 130; HSE, Herpes simplex encephalitis; IFN, Interferon; IFNAR1, Interferon alpha/beta receptor 1; IFNGR1, Interferon gamma receptor 1; IFNLR1, Interferon lambda receptor 1; IL, Interleukin; IMF, Idiopathic myelofibrosis; JAK, Janus kinase; JH1/2, JAK homology 1/2; JMML, Juvenile myelomonocytic leukemia; LEPR, Leptin receptor; LIF, Leukemia inhibitory factor; LEPR, Leptin receptor; LIFRβ, Leukemia inhibitory factor receptor β; LOF, Loss-of-function; mda7, Melanoma differentiation associated gene-7; MDS, Myelodysplastic syndromes; MPL, Myeloproliferative leukemia protein, a.k.a. TPOR; MPN, Myeloproliferative neoplasms; MS, Multiple sclerosis; NF1-MPNSTs, Neurofibromatosis type 1-associated malignant peripheral nerve sheath tumors; NF-E2, Nuclear factor, erythroid 2; NKTCL, NK/T-cell lymphoma; NNT-1, Novel neurothrophin-1; NSCLC, Non-small cell lung cancer; OPR, OB-receptor; OSMRβ, Oncostatin M receptor β; PCM1, Pericentriolar Material 1; PH, Pleckstrin homology; PIAS, Protein inhibitor of activated STAT; PID, Primary immune deficiency; PMF, Primary myelofibrosis; PRL(R), Prolactin (receptor); PTK, Protein tyrosine kinase; PTP, Protein tyrosine phosphatase; PTPRC/T, Protein tyrosine phosphatase receptor type C/T; PTPBL, Basophil-like PTP; PV, Polycythemia vera; RA, Rheumatoid arthritis; RPN1, Ribophorin 1; RTK, Receptor tyrosine kinase; RUNX1, Runt-related transcription factor 1; SCID, Severe combined immunodeficiency; SH2, Src homology 2; SHP2, Protein tyrosine phosphatase, non-receptor type 11; SLE, Systemic lupus erythematosus; SOCS, Suppressors if cytokine signaling; SPAG9, Sperm associated antigen 9; SSBP2, Single-stranded DNA-binding protein 2; STAT, Signal transducer and activator of transcription; STRN3, Striatin 3; T-ALL, T-cell acute lymphoblastic leukemia; TCCR, T-cell cytokine receptor; TCPTP, T-cell protein tyrosine phosphatase; TEL, Ets variant 6 (ETV6); TM, Transmembrane; T-PLL, T-cell prolymphocytic leukemia; TPO(R), Thrombopoietin (receptor); TSLP(R), Thymic stromal lymphopoietin (receptor); TYK2, Tyrosine kinase 2 Cytokine 118 (2019) 48–63 Available online 21 May 2018 1043-4666/ © 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/). T tumorigenesis. The first demonstration of this were the identification of rare oncogenic fusions involving the JAK kinase domain, e.g., the TEL/ ETV6-JAK2 translocation resulting in leukemia driven by constitutively dimeric and active JAK2 kinase domain [1]. Subsequently, several other JAK2 fusions have been identified [2] as well as JAK2 gene amplifications in lymphomas [3] and triple-negative breast cancer [4]. Nevertheless, somatic JAK driver mutations are most commonly found in hematological malignancies, where a single oncogenic JAK2 mutation, JAK2 V617F, underlies more than half of all classical myeloproliferative neoplasm (MPN) cases (reviewed in [5,6]). JAK mutations are also common in leukemia, where, e.g., somatic JAK1 mutations are found in 10–20% of T-cell acute lymphoblastic leukemia (TALL) and JAK2 mutations in ∼20% of down syndrome-associated BALL [5]. Recently, (epi)genetic changes in JAKs, STATs, as well as regulatory components of the pathway have also been increasingly linked to other Table 1 Type I and II cytokines, their receptor chain configurations, as well as the corresponding JAKs and STATs are shown. JAKs and STATs critical for signaling identified with the highest confidence are shown in bold, associated JAKs and STATs for which the data is weaker are shown in parentheses. Cytokines that signal through the same receptor-JAK–STAT configuration are separated by commas. Synonyms are separated by a slash. In case of three or more often-used synonyms and for additional explanations, see footnotes. Cytokine Receptor chain(s) JAKs STATs References TYPE II CYTOKINE RECEPTORS IFN family IFN-I (typeI) * IFNAR1 IFNAR2 JAK1, TYK2 STAT1, STAT2, STAT3, STAT4 (STAT5, STAT6) [154,155] IFN-γ(typeII) IFNGR1 IFNGR2 JAK1, JAK2 STAT1 [150,154,155] IL-28a, IL-28b, IL29 ‡ IL-28R/ IFNLR1 IL-10RβJAK1, TYK2 STAT1, STAT2, STAT3, STAT5 [154,156–158] IL-10 IL-10RαIL-10RβJAK1, TYK2 STAT3, STAT1 [154–156] IL-19 IL-20RαIL-20RβJAK1, JAK2 STAT3, STAT1 [150,154,156] IL-20, IL-24/mda7 IL-20Rαor IL22R IL-20RβJAK1, JAK2 STAT3, STAT1 [150,151,154,156] IL-22/IL-TIF † IL-22R IL-10RβJAK1, TYK2 STAT3, STAT1, (STAT5) [150,154] IL-26/AK155 IL-20RαIL-10RβJAK1, TYK2 STAT3, STAT1 [151,154,156] TYPE I CYTOKINE RECEPTORS gp130 family IL-6 IL-6Rαgp130 JAK1, JAK2, TYK2 STAT3, STAT1 [90,154,155] IL-11 IL-11Rαgp130 JAK1, JAK2, TYK2 STAT3, STAT1 [90,154,155] LIF LIFRβgp130 JAK1, JAK2, TYK2 STAT3, STAT1 [90,154,155,159] CNTF CNTFRαLIFRβgp130 JAK1, (JAK2, TYK2) STAT3, (STAT1) [90,154,155,160] CLCF1 § , NP CNTFRαLIFRβgp130 JAK1, (JAK2) STAT3, STAT1 [90,154,160] CT-1 CNTFRαLIFRβgp130 JAK1, (JAK2, TYK2) STAT3 [90,154,155] OSM OSMRβor LIFRβ gp130 JAK1, (JAK2, TYK2) STAT3, STAT1 [90,154,155,159] IL-31 IL-31Rα/ GLMR OSMRβJAK1, (JAK2) STAT3, STAT5, STAT1 [154,161] G-CSF GCSFR/ CSF3R JAK1, (JAK2) STAT3 [90,154,155] Leptin LEPR/OBR JAK2 STAT3 [154] IL-12 (p35+p40) IL-12Rβ2 IL-12Rβ1TYK2, JAK2 STAT4 [153–155] IL-23 (p19+p40) IL-23R IL-12Rβ1TYK2, JAK2 STAT3, STAT4, STAT1 [153,154] IL-27 (p28+EBI3) || IL-27Rα ¶ gp130 JAK1, JAK2, TYK2 STAT1, STAT3, STAT4, (STAT5) [90,153,154] IL-35 (p35+EBI3) # IL-12Rβ2 gp130 JAK1, JAK2 STAT1, STAT4 [153,154] γ c family IL-2 IL-2RαIL-2Rβγ c JAK1, JAK3, (JAK2) STAT5, (STAT3) [90,154,155,162] IL-4 IL-4Rαγ c JAK1, JAK3 STAT6 [90,154,155,162] IL-7 IL-7Rαγ c JAK1, JAK3 STAT5, (STAT3) [90,154,155,162] IL-9 IL-9Rαγ c JAK1, JAK3 STAT5, STAT3 [90,154,155,162] IL-15 IL-15RαIL-2Rβγ c JAK1, JAK3 STAT5, (STAT3) [90,154,162] IL-21 IL-21R γ c JAK1, JAK3 STAT3, STAT5, (STAT1) [90,154,162] TSLP IL-7RαTSLPR/ CRLF2 JAK1, JAK2 STAT1, STAT3, STAT4, STAT5, STAT6 [154,162,163] IL-13 IL-4RαIL-13R JAK1, JAK2, TYK2 STAT6, (STAT3) [154,155,162] IL-3/β c IL-3 IL-3Rαβ c (gp140) JAK2, JAK1 STAT5, STAT3 [90,154,155,164] IL-5 IL-5Rαβ c (gp140) JAK2 STAT5, STAT1, STAT3 [90,154,155] GM-CSF GM-CSF-Rαβ c (gp140) JAK2 STAT5 [90,154,155] Single chain EPO EPOR JAK2 STAT5 GH GHR JAK2 STAT5, (STAT3) PRL PRLR JAK2 STAT5 TPO TPOR/MPL JAK2 STAT5 * In humans this family consists of 12 IFN-αs, IFN-ωand Limitin (a.k.a. IFN-ζ). † IL-22 also has a soluble receptor IL-22BP, which probably works as an agonist in vivo [150,151]. ‡ Interleukins 28 and 29 are also called Type III IFNs, or IFN-λs, as follows: IL-29/IFN-λ1, IL-28a/IFN-λ2, and IL-28b/IFN-λ3. § a.k.a. CLC, CLF, NNT-1, BSF-3; CLCF1 is secreted either with sCNTFR or CRLF1. || IL-27 p28 subunit is also called IL-30, which might signal through IL-6Rα[152]. ¶ a.k.a. WSX-1, TCCR. # IL-35 has also been reported to signal through IL-12Rβ2 or gp130 homodimers [153]. H.M. Hammarén et al. Cytokine 118 (2019) 48–63 49 cancers. Examples include hepatitis B-associated hepatocellular carcinoma (JAK1 mutations in ∼9% of patients), large granular lymphocytic leukemia (STAT3 mutations in 40% of patients), prostate cancer (amplification of STAT5A/B locus) and non-small cell lung cancer (NSCLC, hypermethylation of suppressor of cytokine signaling (SOCS) 3 promoter) (reviewed in [7]). Furthermore, several cytokines, particularly interleukin (IL-)6 mediate critical signals for the growth of solid tumors, and inhibition of JAK kinases by small-molecular weight inhibitors has been shown to efficiently abrogate tumor formation or restore sensitivity to other protein tyrosine kinase (PTK) inhibitors in xenograft models [8,9]. Currently, clinical JAK inhibition is focused on the treatment of MPNs [10] and rheumatoid arthritis (RA) [11], and clinical trials with JAK inhibitors are ongoing also in other autoimmune and inflammatory diseases [12]. Mounting evidence of the involvement of the JAK–STAT pathway in cancer, however, has already led to clinical trials with JAK inhibitors in, e.g., pancreatic cancer [13], advanced solid tumors (ClinicalTrials identifier: NCT02646748), NSCLC (NCT02917993), and triple-negative breast cancer (NCT02876302). The majority of clinical JAK mutations concentrate in the pseudokinase domain of the protein and highlight the importance of understanding the molecular mechanisms of normal and pathogenic JAK signaling. Given the prevalence of JAK mutations as well as the broad involvement of JAK-mediated signaling in disease, this understanding is a vital prerequisite for developing better future therapies with potential for wide-ranging implications. In this review, we discuss the current views on the molecular basis of regulation of JAKs, as well as mechanisms of JAK deregulation in disease. 2. Jaks and the JAK–STAT pathway JAKs are intracellular, non-receptor PTKs, but constitutively associated with their cognate receptors [14–16]. There are four JAKs in mammals, birds, and fish (JAK1–3, TYK2), with JAK1, JAK2 and TYK2 being ubiquitously expressed, while expression of JAK3 being mostly restricted to cells of hematopoietic origin [17]. JAKs are multi-domain proteins consisting of an N-terminal FERM-domain, a SH2-like domain, a so-called pseudokinase domain (JAK homology 2, JH2) and the catalytically active, signaling PTK domain (JH1). The FERM and SH2-like domains mediate the interaction of JAKs with their receptors [18,19], as well as regulating JAK kinase activity by as-of-yet unknown mechanisms [20,21]. Recent crystal structures have revealed that the FERM and SH2 domains form a structurally tightly-linked continuous unit with the receptor peptide running over, and intimately contacting both domains [22–25]. The most defining JAK domain is probably the pseudokinase domain, JH2, which constitutes the second kinase domain, and thus the second face of its two-faced namesake ancient Roman god Janus. JH2 strongly resembles eukaryotic protein kinases, but shows a distinct pattern of deviations from active kinases including lacking the catalytic aspartate, which is replaced by an asparagine in all JAK JH2s, thus gaining them the classification of ‘pseudokinase’[26–28].JH2 has critical regulatory functions, and is a veritable hotspot for clinically relevant mutations, including driver mutations underlying hematopoietic malignancies (JAK2 mutations), leukemia and lymphomas (all JAKs), and cancer (JAK1, JAK3), discussed in more detail below [5]. Lastly, JAKs consist of a C-terminal active PTK domain, JH1 , which is closely related to the kinase domain of receptor tyrosine kinases (RTKs), like epidermal growth factor receptor (EGFR), and seems to be evolutionarily distinct from JH2 [29]. JAKs mediate signaling of around sixty different cytokines, hormones, and growth factors ranging from regulators of the immune system and hematopoiesis, like interleukins (IL), interferons (IFN), erythropoietin (EPO), and thrombopoietin (TPO), to regulators of development and metabolism, like prolactin (PRL) and growth hormone (GH) (see Table 1)[30]. Signaling through the JAK–STAT pathway (Fig. 1A) is initiated by binding of a cytokine to the extracellular EPO pSTAT5 pSTAT5 Nucleus Cytoplasm Cell membrane Proliferation, survival, differentiation JAK2 pSTAT5 pSTAT5 N C SH2 JH2 JH1FERM Y119 Y201 Y206 Y221 Y317 Y372 Y373 S523 Y570 Y637 Y813 Y868 Y966 Y972 Y913 Y1007 Y1008 JH2 JH1 T875N D873N K539L pY570 pS523 R683S/G V617F D873 R683 F594 F595 F617 F537 K607 H606 H608 T875 A B C DE SOCS, SH2Bs PTPs PIAS MAPK/ ERK PI3K JAK2 JAK2 basal stimulation Fig. 1. The JAK–STAT signaling pathway and its regulation. A) Overview of the JAK–STAT signaling pathway using erythropoietin (EPO) signaling as an example. The most important intermolecular regulators are depicted along with their point of action along the signaling pathway. B) Characterized phosphorylation sites along JAK2. Activating phosphorylation sites are shown in green and inhibiting in red. The effect of Y206 phosphorylation on kinase activity is uncertain. See also Table 3. C) The JH2–JH1 inhibitory interaction using JAK2 as an example [44]. Sites of known clinical activating JAK2 mutations are shown with red spheres (αcarbon), and most important mutations are highlighted. The two inhibitory phosphorylation sites, pS523 and pY570, are encircled. D) Close-up of the D873–R683 interaction, along with surrounding residues, all of which are known sites of activating mutations (Table 2). E) The phenylalanine stack around F617 in JAK2 V617F crucial for V617F-induced JAK2 activation. H.M. Hammarén et al. Cytokine 118 (2019) 48–63 50 portions of its cytokine receptor chain(s). This induces dimerization, oligomerization, or a conformational change of the receptor complex [31], which activates associated JAKs inducing trans-autophosphorylation of the activation loop (AL) of JAK JH1s [32] and a subsequent increase in catalytic activity [33]. The molecular details of this activation mechanism are still unknown with multiple competing models currently under investigation (see below) [34]. Activated JAKs next phosphorylate specific tyrosines on the receptor chains, which serve as docking sites for SH2 domain-containing signaling molecules like STATs [35]. Receptor-bound STATs are next phosphorylated by JAKs on a specific tyrosine in the C-terminal tail, enabling SH2-mediated dimerization of STATs, and subsequent translocation into the nucleus [35–37], where they act as transcription factors with far-reaching effects on regulation of transcription and epigenetics (reviewed in [38]). This simple signal transduction pathway of cytokine receptors, associated JAKs, downstream STATs, along with key protein regulators is evolutionarily conserved throughout bilateria [39]. 3. Regulation of JAK activity Activity of the JAK–STAT pathway is tightly regulated on multiple levels to ensure suppression of signaling in the absence of cytokine stimulation, but also to allow rapid, transient activation upon stimulation (Fig. 1). Regulation relies primarily on control of JH1 tyrosine kinase activity by intraand intermolecular mechanisms, as well as on controlling the availability of parts of the signaling machinery. Many of these levels of regulation are initiated or modulated by phosphorylation of regulatory residues on JAKs, receptors, and STATs 3.1. Intramolecular regulation of JAK activity The most basic level of regulation of JAK tyrosine kinase activity is by the AL of JH1. Isolated JH1, when produced using recombinant protein technology, is constitutively active and able to autophosphorylate its AL in solution leading to full activation of the domain’s enzymatic activity [33,40,41]. In PTKs, this phosphorylation is usually necessary to relieve inhibition caused by the insertion of the unphosphorylated AL into the active site of the domain [42]. However, no crystal structures of JH1 currently exist of this presumed inactive form [43], and recent molecular dynamics data have suggested, that activation of JH1 via phosphorylation of its AL is not (only) due to vacation of the active site, but rather, due to destabilization of the autoinhibitory JH2–JH1 interaction [44]. 3.2. Inhibition and activation of JH1 by JH2 Early domain deletion experiments indicated that both JH2 and JH1 were needed for JAK-mediated signaling [45–48]. Subsequently, work on JAK2 and JAK3 found that loss of JH2 was also associated with increased basal JAK activity (i.e. in the absence of cytokine stimulation) [47,48]. Furthermore, adding JH2 to isolated JH1 in cells or in recombinant constructs has been shown to significantly decrease kinase activity of JH1 [40,41,47,49]. Thus, JH2 is a regulatory domain with two roles: firstly, inhibition of the kinase activity of JH1 in the absence of stimulation and, secondly, mediation of the stimulatory signal from the cytokine receptor to JH1 [48]. Both of these functions are strikingly apparent in the effects of various known clinical and experimental JAK mutations, as JH2 harbors both gain-of-function and loss-of-function mutations [46,50] (see Table 2). Recently, virtually identical structures of a JH2–JH1 interaction were proposed by long time scale molecular dynamics simulations for JAK2 [44] and a crystal structure for TYK2 [41], which provide explanation for JH2-mediated inhibition of JH1 (Fig. 1C). In the interaction, JH2 binds to the hinge-side of JH1 in a front-to-back orientation, loosely analogous to the interaction seen in the autoinhibited structure of Src kinase, in which SH2-SH3 domains bind to the hingeside of the kinase domain [51]. In the JH2–JH1 interaction, JH1 activity is probably inhibited by an opening of the catalytic site and conformational restriction of the JH1 lobes preventing the conformational dynamics of JH1 needed for the kinase reaction. The simulation-derived, autoinhibitory JAK2 JH2–JH1 interaction includes two previously known, JAK2-specific inhibitory phosphorylation sites: JAK2 S523 and Y570 [52–55] (Fig. 1B, Table 3), phosphorylation of both of which is expected to strengthen the interaction (Fig. 1C) [44]. 3.3. Activation of JAKs by mutation The inhibitory JH2–JH1 interaction interface harbors the majority of known clinical and experimental activating JAK mutations, and disruption of the inhibitory interaction thus provides an explanation for the hyperactive phenotype. Striking examples are mutations in JAK2 JH2 β7-β8 and αC-β4, and JH1 β2-β3 loops, which contain JAK2 R683 and D873 that form an ionic interaction over the JAK2 JH2–JH1 interface (Fig. 1D). Mutations in these loops (including mutations of R683 and D873 themselves) are expected to weaken the JH2–JH1 interaction by disrupting the R683–D873 link. Similarly, these residues are also known mutation sites in other JAKs (Table 2), indicating that the identified mode is likely to be conserved across all four JAKs. The multitude of known JAK2 exon 12 mutations (Table 2), on the other hand, fall into the SH2-JH2 linker region, which probably makes extensive contacts along the JH2–JH1 interface [44], and has previously been shown to be important in suppression and regulation of basal activity [19,56].Similarly, mutations such as JAK2 V617 are expected to disrupt the conformation of the SH2-JH2 linker, which is probably (at least part of) the activation mechanism of, e.g., JAK2 V617F. However, it is currently unclear, whether the high activating potency of, e.g., JAK2 V617F is explained by SH2-JH2 linker-mediated disruption of the JH2–JH1 interaction alone, or whether a distinct activating intraor intermolecular interaction is involved. Indeed, analysis of recombinant JAK2 fragments has shown that introduction of V617F to JH2-JH1 constructs only increases catalytic activity by ∼3fold [40], suggesting that another mechanism beyond disruption of the JH2–JH1 might be needed to explain the abnormally high activity of JAK2 V617F in vivo. CoIP experiments with full-length JAK2 have suggested that addition of the V617F mutation leads to a JAK2–JAK2 interaction in vitro, which cannot be detected with wild-type JAK2 [57]. The high density of activating mutations in the N lobe of JH2 as well as the SH2-JH2 linker has also been suggested to be circumstantial evidence for an activating interaction, probably involving these regions [58]. Transformation by practically all activating JAK2 mutations (including JAK2 V617F) requires the presence of cytokine receptors [59–61], which could be due to requirements for an activating transinteraction. Alternatively, the receptor could function as a simple scaffold allowing co-localization of activated JAKs with their substrate (s). Data showing that correct orientation of the receptor chains are not needed for activation by JAK2 V617F [62] support the latter model. Analysis of experimental mutations capable of inhibiting activation by clinical disease mutations (e.g., JAK2 V617F) has identified regions in JH2 required for mutational activation. These regions include the SH2-JH2 linker [63], a functional ATP-binding pocket in JH2 (at least in JAK1, JAK2 [64], and JAK3 (Raivola, Hammarén, Silvennoinen et al, manuscript in preparation)), as well as JH2 αC[62,65,66]. Indeed, molecular dynamics simulations of JAK2 JH2 suggest that V617F induces stabilization of JH2 αC, which can be reversed by addition of the inhibiting αC mutation JAK2 F595A thus breaking the phenylalanine stack around F617 (Fig. 1E) [67]. Similarly, loss of JH2 ATP-binding is likely to suppress mutational activation by destabilizing αC[64]. Interestingly, these mutations have been reported to not strongly inhibit cytokine-dependent JAK activation, suggesting that activation by cytokine could be mechanistically distinct from activation by mutations like JAK2 V617F [57,62,64–66]. Whether this is due to a distinct H.M. Hammarén et al. Cytokine 118 (2019) 48–63 51 Table 2 Clinical mutations in JAKs. Data in the table were expanded from similar compilations in [41,165–167]. Kinase activity –measured as basal JAK activation loop phosphorylation (e.g., JAK2 Y1007/1008), STAT phosphorylation, or transcriptional activity in reporter assay: +/−increase/decrease in activity, respectively. An increase in activity (+) refers to basal (unstimulated) activity (e.g., ligand-independent activation by JAK2 V617F); for decreased activity (−) signaling upon stimulation is included (e.g., cytokine-irresponsive JAK3 SCID mutations); NE –no appreciable effect. fs –frame-shift. Due to the increased pace of genetic analysis of patient samples, experimentally uncharacterized mutations are only shown, if of special interest. Exon # JAK domain Mutation Kinase activity Effect Associated disease Reference JAK1 3 FERM I62V B-ALL and T-ALL [168] 3 FERM S71C ETP-ALL [169] 4 FERM D82A Might lower receptor association (shown together with JAK1 Y81A) Gynecologic tumors [167,170] 5 FERM K142fs LOF by frame-shift Gynecologic tumors [167] 6 FERM K204M B-ALL [167] 8 FERM N339fs LOF by frame-shift Gynecologic tumors [167] 8 FERM R360W T-ALL [168] 8 FERM I377K ETP-ALL [169] 9 Linker V427M Pediatric T-ALL [171] 9 Linker N451S + Increased pSTAT3/5 Hepatocellular carcinoma (xenograft) [172] 9 Linker P430fs LOF by frame-shift Gynecologic tumors [167] 10 SH2 W467 * Breast Cancer (Triple Neg) [173] 10 SH2 T478S + Increased pSTAT1/3/5 AML [41] 10 SH2 E483D + Increased pSTAT3/5 Hepatocellular carcinoma (xenograft) [172] 11 SH2 G511D Gynecologic tumors [167] 11 SH2 S512L Overexpression of ALL (and JAK)-related genes T-ALL [168] 12 Linker R577Q Gynecologic tumors [167] 13 JH2 T593M Gynecologic tumors [167] 13 JH2 D604Y T-ALL [174] 13 JH2 V623A + Increased pSTAT1/3/5 AML [41] 13 JH2 L624-629W ALL [175] 13 JH2 I631R/G/I ALL [176] 14 JH2 A634D + B-ALL and T-ALL [168] 14 JH2 E637K Breast Cancer (HER2+) [173] 14 JH2 Q644H Hepatocellular carcinoma [123] 14 JH2 V645F + Constitutive pJAK1/STAT5 Hepatocellular carcinoma [123] 14 JH2 S646F/P + Increased pJAK1/STAT5 ALL [175,177] 14 JH2 H647Y Invasive ductal carcinoma [178] 14 JH2 K648N T-ALL [179] 14 JH2 V651M T-ALL [41] 14 JH2 Y652H + Constitutive STAT5 activity T-ALL [41,177] 14 JH2 L653F + Childhood ALL [168] 14 JH2 C657R Gynecologic tumors [167] 14 JH2 V658F/L + Constitutive STAT5 activity ALL [175,177,178,180] 15 JH2 E668Q + Pediatric T-ALL [171] 15 JH2 S703I + High basal pSTAT3/5 ALL & hepatocellular carcinoma [172,176,177] 16 JH2 R724H/Q/S + Increased STAT1 activity B-ALL and T-ALL [168] 16 JH2 S729C + Increased pSTAT3/5 Hepatocellular carcinoma (xenograft) [172] 16 JH2 P733L −Immunosuppression [181] 16 JH2 F734L T-ALL [177] 17 JH2 T782M Adenocarcinoma [178] 17 JH2 L783F + Increased STAT5 activity T-ALL [177,178] 17 JH2 L799P Gynecologic tumors [167] 18 Linker P832S −Immunosuppression [181] 19 Linker K860fs −LOF by frame-shift Gynecologic tumors [167] 19 Linker G871E SKN tumour cell line [182] 19 JH1 R879S/C/H + Increased STAT1 activity T-ALL [168] 19 JH1 G882E Loss of JH1 ATP-binding SKN tumour cell line [182] 20 JH1 T901G + T-ALL [171] 20 JH1 K908T Pediatric T-ALL [171] 20 JH1 K924fs LOF by frame-shift Gynecologic tumors [167] 22 JH1 G990_splice Gynecologic tumors [167] 23 JH1 E1051Q/P Gynecologic tumors [167] 24 JH1 A1086S + Increased pSTAT3/5 Hepatocellular carcinoma (xenograft) [172] 24 JH1 R1113H Gynecologic tumors [167] JAK2 3 FERM E61K Putative primary erythrocytosis [183] 4 FERM T108A NE Slightly EPO hypersensitive Found as germline mutation in V617Fpositive PV patient [184] 6 FERM E177V Putative primary erythrocytosis [183] 7 FERM G276A Putative primary erythrocytosis [183] 8 FERM R340Q PV [185] 9 Linker L393V NE Might be weakly EPO hypersensitive Found as germline mutation in V617Fpositive PV patient [184] SH2 12 Linker T514M MPNs [186] (continued on next page) H.M. Hammarén et al. Cytokine 118 (2019) 48–63 52 Table 2 (continued) Exon # JAK domain Mutation Kinase activity Effect Associated disease Reference 12 Linker N533I/Y PV (together with K539L) [166] 12 Linker M353I + AMKL [19] 12 JH2 K539L + MPNs [187] 12 JH2 I540T PV [41] 12 JH2 538–547 (Various deletions and insertions) MPNs [41,188] 12 JH2 D544G PV [41] 12 JH2 L545S PV [41] 12 JH2 F547L PV [41] 13 JH2 F556L MPNs [166] 13 JH2 R564Q + Hereditary ET [189] 13 JH2 V567A MPNs [166] 13 JH2 G571S Unknown (non-affected parent of MPN patient has this variant) [183] 13 JH2 H587N MPNs [166] 13 JH2 S591L MPNs [166] 14 JH2 H606Q MPNs [166] 14 JH2 K607N + AML [166] 14 JH2 H608Y MPNs [166] 14 JH2 L611S + ALL [166] 14 JH2 V617F + MPNs [116–119] 14 JH2 V617I (+) Cytokine hyperresponsiveness Hereditary thrombocythemia [190,191] 14 JH2 C618R + MPNs [192] 14 JH2 D620E MPS [193] 15 JH2 L624P MPNs [166] 15 JH2 I645V MPNs [166] 16 JH2 I682F + ALL [194] 16 JH2 R683S/G + MPNs [194,195] 17 JH2 S755R Hereditary thrombocythemia (together with R938Q in cis) [196] 19 Linker Y813D IMF [41] 19 Linker E846D (+) Increases JH1 activity, when EPOR is present, EPO hypersensitivity Germ-line mutation found in erythrocytosis and megakaryocytic atypia [197] 20 JH1 R867Q + ALL, hereditary thrombocythemia [175,196] 20 JH1 D873N + ALL [194] 20 JH1 T875N + AMKL [41] 21 JH1 P933R + ALL [194] 21 JH1 R938Q + Hereditary thrombocythemia (together with S755R in cis) [196] 24 JH1 R1063H (+) Increases JH1 activity, when EPOR is present, EPO hypersensitivity Germ-line mutation found in erythrocytosis and megakaryocytic atypia [197] 25 JH1 N1108S PV [41] JAK3 2 FERM M1V SCID [165] 2 FERM G36fs −LOF by frame-shift SCID [198] 2 FERM A58P/del −Lack of JAK3 mRNA/protein SCID [165,198] 3 FERM G62S + Does not transform BaF3, kinase activity higher than WT AML [199] 3 FERM I87T + Increased basal pJAK3/pSTAT5 DS-TMD [198,200] 3 FERM L98/99A Weakens receptor binding SCID [201] 3 FERM Y100C −Weakens receptor binding SCID [201,202] 4 FERM P132T/A + P132A oncogenic in mouse xenotransplants AMKL [203,204] 5 FERM P151R SCID, DS-TMD, AMLK [165,205] 5 FERM L156P + No significant increase in pSTAT5 in 293T cells, but does transform BaF3 cells ATLL [206] 5 FERM D169E −No IL-2 responsive pSTAT5 SCID [24] 5 FERM R172Q + No significant increase in pSTAT5 in 293T cells, but does transform BaF3 cells ATLL [206] 5 FERM E183G + ATLL [206] 6 FERM R223H Small cell carcinoma/neuroendocrine prostate cancer (NePC) [207] 6 FERM R272H Nontransforming passenger T-ALL [208] 6 FERM Q283H Nontransforming passenger T-ALL [209] 8 FERM T391fs LOF by frame-shift SCID [198] 9 SH2 R403H Nontransforming passenger T-ALL [208,210] 10 SH2 R445X −SCID [211] 11 SH2 E481G −SCID [212] 11 Linker K482S SCID [165] Linker fsX483 LOF by frame-shift SCID [165] 11 Linker del482-596 −SCID [212] 11 Linker Q501H + AMKL [41,200] 11 Linker Q507P T-PLL [213] 11 Linker M511I + Slight kinase activity increase, transforms Ba/F3 T-PLL, AML, JMML, NKTCL [208,213] 12 JH2 C565X −SCID [108,198] (continued on next page) H.M. Hammarén et al. Cytokine 118 (2019) 48–63 53 Table 2 (continued) Exon # JAK domain Mutation Kinase activity Effect Associated disease Reference 13 JH2 A572V/T + Increased pSTAT3/5; features of megakaryoblastic leukemia and transforms murine lymphoid cells in vivo AMKL [41,208] 13 JH2 A573V + T-ALL-like disease in mice DS-ALL,DS AMKL, NTCL [41,208,214–216] 13 JH2 M576L Adult non-DS AMKL [217] 13 JH2 del586-592 −No pSTAT5, but constitutive pJAK3 SCID [49] 13 JH2 R582W −Leads to two products, insufficient for signal transduction SCID [165] 13 JH2 H583Y + NTCL [198,216] 13 JH2 G589S −SCID [198] 13 JH2 G589D + NTCL [216] 13 JH2 delV590-S596 SCID [165] 13 JH2 del592 AMKL/TMD [205] 13 JH2 A593R/T + DS-AMKL [41,215] 15 JH2 R657Q(/H/W) + High basal pSTAT5, causes B-cell leukemia in mice T-PLL, AML, JMML, NKTCL [41,213] 15 JH2 R651W SCID [218] 15 JH2 V674A/F + Increased STAT5 activation, T-ALL-like disease in mice T-ALL [213,219,220] 15 JH2 V678L/M T-ALL [208,213] 16 JH2 P689S SCID [218] 16 JH2 R694K SCID [218] 16 JH2 E698X SCID [165] 16 JH2 V715I Breast cancer [178] 16 JH2 V722I + AMKL, SCID,NKTCL [203,211] 16 JH2 K733fs LOF by frame-shift SCID [221] 17 JH2 Del734-784 SCID [165] 17 JH2 C759R −No pSTAT5, but constitutive pJAK3 SCID [49] 17 JH2 V765D JMML [222] 17 JH2 Q766X SCID [165] 17 Linker R771X SCID [165] 18 Linker S789P + Increased kinase activity, but only weakly transforms BaF3 Childhood ALL [194,215] 18 JH1 Y824D T-PLL [223] 19 JH1 L857Q/P + Receptor independent, L857Q causes severe thymus hyperplasia in mice T-ALL/T-PLL/JMML [208,220] 19 JH1 Q865E Nontransforming passenger T-ALL [209] 19 JH1 fsX844 LOF by frame-shift SCID [165] 20 JH1 Y904X SCID [224] 20 JH1 P906S + ALL [220] 20 JH1 L910S −SCID [218] 20 JH1 R918C + Increased kinase activity, but only weakly transforms BaF3 AML [199,215] 20 JH1 R925S Nontransforming passenger T-ALL [208] 20 JH1 Y929X SCID [198] 21 JH1 E958K + T-ALL, JMML [209,220,222,225] 21 JH1 G987fs −LOF by frame-shift SCID [198] 21 JH1 Q988P T-ALL [226] 22 JH1 L1017M + Increased kinase activity, but only weakly transforms BaF3 CML [199] 22 JH1 Y1023X −SCID [218] 23 JH1 1024fs −LOF by frame-shift SCID [165] 24 JH1 E1106G Nontransforming passenger T-ALL [208] TYK2 3 FERM G36D T-ALL (MOLT-16 cell line) [134] 3 FERM S47N T-ALL (MOLT-16 cell line) [134] 3 FERM A53T NE PBMCs show decreased CXCL10 response Increased MS risk, found in herpes simplex encephalitis (HSE) patient [227–229] 8 FERM V362F NE Increased SLE and SSc risk [230] 8 FERM G363S NE AML [228] 9 FERM R425H T-ALL (MOLT-16 cell line) [134] Linker SH2 Linker 15 JH2 I684S −Impairs TYK2 kinase activity, but still enables signaling through JAK1/TYK2 or JAK2/TYK2 Protect against RA and Autoimmunity. Also found in T-ALL. [134,231,232] 15 JH2 R703W NE AML [228] 16 JH2 V731I + ALL (RPMI-8402 cell line) [134] 16 JH2 P760L + ALL (germline) [135] 16 JH2 G761V + ALL (germline) [135] Linker 20 JH1 M926V NE Germline mutation in ALL, but probably not deleterious [135] 20 JH1 A928V NE Protect against RA and Autoimmunity [228,231] 20 JH1 E957D + ALL (MOLT-4 cell line) [134,135] 22 JH1 A1016S NE AML [228] (continued on next page) H.M. Hammarén et al. Cytokine 118 (2019) 48–63 54 activating interaction employed by, e.g., JAK2 V617F, is still unknown. 3.4. Phosphorylation as mediator of regulation Much of the regulation of JAK–STAT signaling is controlled by phosphorylation (Fig. 1B). In the basal state, JAK2 (for which there is the most data available) is phosphorylated only on S523 (potentially by JH2 itself [54,68]), which strengthens the JH2–JH1 interaction described above [44]. Upon activation of receptor by cytokine, JAK2 is activated by (trans-)autophosphorylation not only on the JH1 AL (Y1007/Y1008), but also on multiple other residues (Y637, Y868, Y966, Y972), which are needed for full activation of kinase activity by as-of-yet unknown mechanisms (Table 3)[69]. Termination of signaling is subsequently initiated by phosphorylation of numerous inhibitory residues along JAK2 (Fig. 1B, Table 3), which may cause dissociation of JAK2 from its receptors (Y119 [70]), strengthen the JH2–JH1 inhibitory interaction (Y570, Fig. 2C [44,68,69]), or act as binding sites for regulatory proteins. Likewise, dephosphorylation of both JAK and receptor activating tyrosines is an important part of regulation of signaling, and is induced by multiple protein tyrosine phosphatases (PTP), including SH2 domain-containing phosphatases 1 and 2 (SHP1, SHP2), PTP1B, T-cellPTP (TCPTP), Receptor-type PTPs C (PTPRC/CD45) and T (PTPRT), as well as basophil-like PTP (PTPBL) (reviewed in [71]). Notably, SHP2, which is a known oncoprotein and increasingly interesting drug target itself [72], has been reported to be both an activator and suppressor of JAK–STAT signaling depending on the context [71]. Other phosphatases associated with regulation of JAK–STAT signaling function as classical negative regulators of signaling in termination or basal suppression of signaling (reviewed in [34,71]). Most recorded phosphorylation events on JAK2 have been shown to be dependent on JAK2 JH1 kinase activity (or AL phosphorylation), but the exact sequence of phosphorylation events or their mechanism of action during receptormediated activation is yet to be fully elucidated. Potential differences in phosphorylation-mediated regulation in different cytokine receptor–JAK contexts also remain mostly unknown. 3.5. Intermolecular regulation of JAK activity Most known intermolecular modes of JAK–STAT signaling regulation are involved in termination of signaling after cytokine stimulation, and are thus initiated by phosphorylation during cytokine-mediated Table 2 (continued) Exon # JAK domain Mutation Kinase activity Effect Associated disease Reference 22 JH1 R1027H + ALL (MOLT-16 and CCRF-CEM cell lines) [134] 23 JH1 P1104A −Impairs TYK2 kinase activity, but still enables signaling through JAK1/TYK2 or JAK2/TYK2 Decreased susceptibility to RA and Autoimmunity; Found in NF1-MPNSTs [228,231–233] Table 3 Regulatory phosphorylation sites, their effect on basal JAK2 activity (measured as pJAK2 (pY1007/Y1008) or a downstream measure like STAT phosphorylation (pSTAT) or transcriptional reporter activity), the probable phosphorylating kinase, and assumed effects of phosphorylation (mostly based on mutagenesis experiments) are shown. Cytokines, which have been experimentally tested for inducing specific phosphorylation, are shown in parenthesis. Sites for which no functional follow-up data has been published, have been omitted for clarity (see, e.g. [234,235]). NE –no effect. Autophos. –autophosphorylation by full-length JAK2. ? –no data available. Domain JAK2 Residue Kinase activity Phosphorylated upon Phosphorylated by Effect of phosphorylation/mutation of site References FERM Y119 −Stimulation (EPO) Probably JH1 * Phosphorylation likely to mimic Y119E, which induces dissociation from EPOR, GHR, PRLR, but not IFNGR2 [70] FERM Y201 + § ? Autophos. (in vitro) Y201F inhibits Ang II-mediated JAK2-signalling § [69,237] FERM Y206 NE ? ? Y206F has no appreciable effect on EPO signaling [69,237] FERM Y221 + Stimulation (GH, IL-3) Autophos. (in vitro and in cells) Y221F decreases basal pJAK2 [52,53] FERM Y317 −Stimulation (EPO) Probably mostly JH1 * Y317F causes ligand-independent pJAK2 [69] FERM Y372 + ? ? Y372F decreases basal and stimulated (IFN-γ, EGF) pJAK2, pSTAT1 and JAK2-STAT1 interaction [69,238] FERM Y373 + ? ? Y373F decreases basal pJAK2 [238] SH2 Linker S523 −Constitutive JH2 (in vitro) S523A slightly increases basal pJAK2 [54,55,68] JH2 Y570 −Stimulation (GH, EPO, IL3) Autophos. (in vitro), JH2 (in vitro) Y570F increases basal pJAK2, pSTAT3 and prolongs EPOinduced activity. pY570 likely to strengthen autoinhibitory JH2–JH1 interaction [41,44,52,53] JH2 Y637 + Stimulation (EPO) Probably mostly JH1 * Y637F lessens and shortens JAK2 activation upon EPO stimulation and partially inhibits V617F [69] Linker Y813 + Stimulation (GH) Autophos. (in vitro) Y813F reduces SH2-Bβinduced pJAK2 and pSTAT5. pY813 putative binding site for activator protein SH2-Bβ [239] JH1 Y868 + Stimulation (GH) JH1 (in vitro) Y868F decreases basal and GH-induced pJAK2, pSTAT3/5 [240] JH1 Y913 −Upon and after stimulation (EPO) Probably JH1 * Y913F increases EPO-induced pJAK2/pSTAT5. Y913E removes EPO-induced JAK2 activation [241] JH1 Y966 + Stimulation (GH) JH1 (in vitro) Y966F decreases basal and GH-induced pJAK2, pSTAT3/5 [240] JH1 Y972 + Stimulation (GH) JH1 (in vitro) Y972F decreases basal and GH-induced pJAK2, pSTAT3/5 [240] JH1 Y1007 + Stimulation JH1 Y1007F removes basal and cytokine-induced JAK2 activation and downstream signaling. pY1007 likely to weaken JH2–JH1 interaction [32,33,44] JH1 Y1008 NE Stimulation JH1 Y1008F has little effect on JAK2 activation [32] * Dependent on JAK2 Y1007/Y1008 phosphorylation and/or JH1 kinase activity. § Y201F has no effect on EPO-induced pJAK2, but inhibits Ang II-mediated pJAK2, pSTAT1/3. Phosphorylation probably enables binding of SHP2 and subsequent binding to AT1 receptor. Y201F has also been shown to inhibit activation by JAK2 V617F [236]. H.M. Hammarén et al. Cytokine 118 (2019) 48–63 55 activation. Mechanisms of termination involve (de)phosphorylation, production of inhibitory proteins, and removal of signaling complexes through internalization and lysosomal or ubiquitination-mediated proteasomal degradation of receptors [5,34]. One family of direct inhibitory proteins are suppressors of cytokine signaling (SOCS1–7 and CIS), transcription of which is upregulated by activated STATs, thus forming a negative feedback loop [73]. The mechanism for SOCS-mediated inhibition has been revealed for SOCS3, which binds to both JH1 and the associated cytokine receptor, leading to inhibition of JH1 activity by direct interaction with the JH1 AL, in addition to initiation of ubiquitination of JAKs [74,75]. Active JAK–STAT signaling also induces transcription of SH2B family proteins (consisting of SH2B1/SH2-B, SH2B2/APS, and SH2B3/ LNK), which each consist of a dimerization domain (DD), Pleckstrin homology (PH) domain and a C-terminal SH2 domain. SH2-B and APS bind to phosphorylated receptor or JAK tyrosines (e.g. JAK2 pY613 and pY813) via their SH2 domain [76–79], and are able to either activate or inhibit JAK activation, potentially by acting as dimerization scaffolds for JAKs [80]. The activating role of SH2-B has been most comprehensively demonstrated in leptin–JAK2 signaling [78,81], and mutations or deletions in SH2B1 have been linked to severe obesity (see Table 4). Interestingly, SH2-B has also been reported to bind to EPOR and putatively act as a negative regulator of EPOR signaling [82]. The third SH2B family member, LNK is a well-defined inhibitor of JAK2, as well as potentially JAK3 (reviewed in [79]). How LNK negatively regulates JAK activity is not clear, but it is likely to involve a direct interaction with the LNK N-terminal domains and JAK kinase domains, and/or ubiquitination via recruitment of E3 ligases [79]. JAK–STAT signaling is also regulated at the level of STATs by protein inhibitor of activated STATs (PIAS) proteins, which are one of only few known small ubiquitin-like modifier (SUMO) E3 ligases (see [83] for a review). As their name suggests, PIAS proteins (consisting of PIAS1, PIAS3, PIASx, and PIASy) were first identified as negative regulators of STATs [84,85], but they have since been shown to also be involved in regulation of multiple other cellular processes [86]. Following cytokine stimulation, PIAS proteins have been shown to inhibit STAT signaling activity through inhibition of their DNA-binding activity by directly interacting with STATs and and/or by inducing STAT SUMOylation (reviewed in [87]). 3.6. Activation of JAKs by cytokines –The role of the cytokine receptor Despite long-continuing research into the mechanisms of JAK activation by mutation and intermolecular regulation by various negative regulator proteins, the molecular mechanisms of JAK activation on cytokine receptors following cytokine stimulation have mostly remained elusive. One complicating factor is the sheer number of different JAK-associated receptors (Table 1) and their variable architectures. These range from ‘short receptors’(e.g., GHR, EPOR, γ c family, IFNGR) where ligand-binding domains are proximal to the membrane, and ‘tall receptors’(e.g., gp130, TCCR, LIFRβ) with ligand-binding domains up to 10 nm removed from the membrane [88–91]. Numerous crystal structures of receptor extracellular domains have been generated showing cytokine-bound and unbound states [90], but how these changes are translated into intracellular activation of JAKs is not known. The simplest mode of cytokine receptor activation is an induced dimer/oligomer model, where ligand binding induces dimerization/ oligomerization of receptor chains and associated kinases, which are subsequently activated by transphosphorylation due to proximity. This mode is widely seen in RTKs [92,93], with exceptions like insulin receptor kinase (IRK), which seems to exist as preformed dimers rearranged by ligand binding [94]. Studies with engineered artificial chimeric receptors have shown that simple dimerization is sufficient to activate JAKs [95] suggesting an induced-dimer activation model. Nevertheless, preformed dimers have been reported for many JAKassociated receptors, including EPOR [96], GHR [97], gp130 [98], LEPR [99], as well as IL-2Rβ/IL-9Rαand γ c [100]. However, some of these data may be confounded by the use of artificial overexpression systems and/or dimerization-prone fusion proteins. A current lack of methods to simultaneously measure receptor oligomerization and activation states also leaves it unclear, how many of the observed preformed dimers are actually actively signaling and thus contributing to basal signaling activity. Still, studies with forced EPOR or TPOR dimers have indicated that correct orientation of receptor chains is needed for activation of wild-type JAK2, thus potentially enabling non-signaling preformed dimers in vivo [101–103]. Additionally, FRET-studies with forced preformed GHR dimers suggested that the intracellular portions of GHR would move apart during stimulation, leading the study authors to hypothesize that JH2-mediated inhibition of JH1 could occur in trans at least for JAK2 [104]. How, or whether this architecture would apply to heteromeric JAK configurations, is unclear. Despite these earlier findings, recent single-molecule imaging studies have shown a distinct lack of preformed dimers for EPOR at least, and approaches using engineered protein ligands have suggested that activation of EPOR–JAK2 is mainly determined by the distance between receptor chains [105], thus again arguing for a ligand-induced dimer model [91]. Further research using less invasive methodology with (at least close to) native expression levels of receptors is thus sorely needed to settle the question of receptor-mediated activation of JAKs. 4. Failure of JAK regulation leads to disease Failure of JAK–STAT pathway regulation is a common cause of myeloid and lymphoid disorders but also plays a role in multiple cancers (Table 4)[6,7,106]. The first JAK disease-link was discovered in 1995 when JAK3 loss-of-function mutations were reported causative of severe combined immune deficiency (SCID) [107,108]. Shortly later, inhibition of acute lymphoblastic leukemia (ALL) with JAK-inhibitor AG-490 lead to the discovery of constitutive JAK2 activity in patientderived ALL cells [109]. Subsequently constitutively active JAK2 fusion protein TEL/ETV6-JAK2 was identified in early pre-B cell ALL and Tcell childhood ALL [1,110]. Oncogenic JAK2 rearrangements to multiple fusion gene partners (TEL, PCM1, BCR, SSPB2, STRN3, RPN1, NF-E2, RUNX1, SEC31A, SPAG9) have since been identified in ALL, acute myeloid leukemia (AML), atypical chronic myeloid leukemia (CML), MPN and/or Hodgkin lymphoma [2,110–114], and similar oncogenic fusions have also been reported for TYK2 [106]. The chimeras comprise of JH1, with or without (pieces of) JH2, fused to dimerization domain(s) of the fusion partner, leading to constitutive tyrosine kinase activity presumably via dimerization/oligomerization [1,112,115]. These translocations are rare, however: a cytogenetic study of 24 262 unique patients, for example, revealed these kinds of JAK2 abnormalities in only 0.06% of hematopoietic neoplasms [113]. In contrast, somatic JAK2 V617F is highly prevalent in MPN and causes approximately 95% of polycythemia vera (PV) and 50–60% essential thrombocythemia (ET) and primary myelofibrosis (PMF) cases [116–119]. Allelic burden probably plays a role in defining the clinical phenotype of JAK2 V617F as the burden increases from ET over PV to PMF [120]. Furthermore, the pSTAT profiles discriminate among the MPN diseases and are independent of JAK2 V617F: high pSTAT3 and pSTAT5 are typical in PV, high pSTAT3 and low pSTAT5 in ET, and low pSTAT3 and pSTAT5 in PMF [121], but the mechanism behind these differences have remained largely unknown. PV, ET and PMF can evolve into AML and therefore it is not surprising that approximately 5% of AML patients are also V617F-positive [122]. Numerous gain-offunction mutations identified in JAK1, JAK2 and JAK3 (Table 2) are drivers in ∼20% of T-ALL cases (mainly JAK1, also JAK2, JAK3) and to lesser extent in B-ALL (JAK1, JAK2) or hepatocellular adenoma (JAK1) [5,123]. V617F-negative MPN often show mutations in other components or H.M. Hammarén et al. Cytokine 118 (2019) 48–63 56 regulation of Jak2 by its auto-phosphorylation at tyrosine 913 via the epo signaling pathway, Cell. Signal. 20 (11) (2008) 1995–2001. [242] B.A. Croker, H. Kiu, S.E. Nicholson, SOCS regulation of the JAK/STAT signalling pathway, Semin. Cell Dev. Biol. 19 (4) (2008) 414–422. [243] E.G. Bochukova, N. Huang, J. Keogh, E. Henning, C. Purmann, K. Blaszczyk, et al., Large, rare chromosomal deletions associated with severe early-onset obesity, Nature 463 (7281) (2010) 666–670. [244] R. Walters, S. Jacquemont, A. Valsesia, A. De Smith, D. Martinet, J. Andersson, et al., A new highly penetrant form of obesity due to deletions on chromosome 16p11. 2, Nature 463 (7281) (2010) 671–675. [245] A. Volckmar, F. Bolze, I. Jarick, N. Knoll, A. Scherag, T. Reinehr, et al., Mutation screen in the GWAS derived obesity gene SH2B1 including functional analyses of detected variants, BMC Med. Genomics 5 (1) (2012) 65. [246] A. de la Chapelle, A.L. Traskelin, E. Juvonen, Truncated erythropoietin receptor causes dominantly inherited benign human erythrocytosis, Proc. Natl. Acad. Sci. U. S. A. 90 (10) (1993) 4495–4499. [247] I. Iacobucci, Y. Li, K.G. Roberts, S.M. Dobson, J.C. Kim, D. Payne-Turner, et al., Truncating erythropoietin receptor rearrangements in acute lymphoblastic leukemia, Cancer Cell 29 (2) (2016) 186–200. [248] M. Kleppe, I. Lahortiga, T. El Chaar, K. De Keersmaecker, N. Mentens, C. Graux, et al., Deletion of the protein tyrosine phosphatase gene PTPN2 in T-cell acute lymphoblastic leukemia, Nat. Genet. 42 (6) (2010) 530–535. [249] M. Porcu, M. Kleppe, V. Gianfelici, E. Geerdens, K. De Keersmaecker, M. Tartaglia, et al., Mutation of the receptor tyrosine phosphatase PTPRC (CD45) in T-cell acute lymphoblastic leukemia, Blood 119 (19) (2012) 4476–4479. [250] M. Tartaglia, E.L. Mehler, R. Goldberg, G. Zampino, H.G. Brunner, H. Kremer, et al., Mutations in PTPN11, encoding the protein tyrosine phosphatase SHP-2, cause noonan syndrome, Nat. Genet. 29 (4) (2001). [251] M. Bentires-Alj, J.G. Paez, F.S. David, H. Keilhack, B. Halmos, K. Naoki, et al., Activating mutations of the noonan syndrome-associated SHP2/PTPN11 gene in human solid tumors and adult acute myelogenous leukemia, Cancer Res. 64 (24) (2004) 8816–8820. [252] J. Ding, H. Komatsu, A. Wakita, M. Kato-Uranishi, M. Ito, A. Satoh, et al., Familial essential thrombocythemia associated with a dominant-positive activating mutation of the c-MPL gene, which encodes for the receptor for thrombopoietin, Blood 103 (11) (2004) 4198–4200. [253] A.D. Pardanani, R.L. Levine, T. Lasho, Y. Pikman, R.A. Mesa, M. Wadleigh, et al., MPL515 mutations in myeloproliferative and other myeloid disorders: A study of 1182 patients, Blood 108 (10) (2006) 3472–3476. [254] P.P. Zenatti, D. Ribeiro, W. Li, L. Zuurbier, M.C. Silva, M. Paganin, et al., Oncogenic IL7R gain-of-function mutations in childhood T-cell acute lymphoblastic leukemia, Nat. Genet. 43 (10) (2011) 932–939. [255] S. Degryse, J. Cools, JAK kinase inhibitors for the treatment of acute lymphoblastic leukemia, J. Hematol. Oncol. 8 (1) (2015) 91. [256] M. Cazzola, R. Kralovics, From janus kinase 2 to calreticulin: The clinically relevant genomic landscape of myeloproliferative neoplasms, Blood 123 (24) (2014) 3714–3719. [257] M. Araki, N. Komatsu, Novel molecular mechanism of cellular transformation by a mutant molecular chaperone in myeloproliferative neoplasms, Cancer Sci. 108 (10) (2017) 1907–1912. H.M. Hammarén et al. Cytokine 118 (2019) 48–63 63