JAK2 signaling is important in both physiology and pathophysiology, as it plays prominent roles in embryonic development and haemopoiesis, as well as in cancer and inflammation (Khwaja,2006; Vijayakrishnan et al.,2011). JAK2, a member of the JAK family of protein tyrosine kinases (PTKs), is a crucial intracellular mediator of cytokine or hormone signaling, and ubiquitously expressed in virtually every cell type (Lai and Johnson,2010). In 1992, Harpur et al. (1992) cloned and published the cDNA sequence of JAK2 gene encoding a full-length protein of about 130 kDa. The predicted amino acid sequence of JAK2 contained the JH1 (JAK homology 1) and JH2 domains adjacently on the carboxyl half of the protein, and the JH3, JH4, JH5, JH6, and JH7 domains on the amino half (Wilks et al.,1991; Sandberg et al.,2004). The JH1 and JH2 domains correspond to kinase and pseudokinase domains, respectively, which are also the basis for the name of the JAK family, being named after Janus, the Roman god of gates and doors with double faces (Wilks et al.,1991; Sandberg et al.,2004).
Of the JAKs, the molecular events underlying JAK2 activation have been best studied. JAK2 is located in the cytoplasm, and binds to the membrane proximal proline-rich region of ligand-dependent receptors (Heim,1996; He and Zhang,2010). STAT3 and STAT5, conventionally thought to exist in the cytoplasm, are proved to be the main effector molecules of JAK2 (Godeny and Sayeski,2007; Starsíchová et al.,2010; Brooks and Waters,2010). A growing body of evidence demonstrates that dysregulation of this canonical JAK2 signaling is involved in carcinogenesis (Hedvat et al.,2009; Behera et al.,2010; He and Zhang,2010; Kim et al.,2011; Sun et al.,2011). Moreover, recent studies in our laboratory also showed that deviant JAK2 activity exists in gastric carcinoma tissue and cell lines (Ding et al.,2010). The functions ascribed to JAK2 have become more diverse since its discovery as a putative tyrosine kinase mutated in myeloproliferative cancers (James,2008; Kiss et al.,2010; Nielsen et al.,2011). However, the current studies show that there are non-canonical JAK-STAT pathways in mammalian cells. Videlicet, unphosphorylated STAT proteins are found in nucleus, and constantly shuttle between cytoplasm and nucleus (Vinkemeier,2004; Reich and Liu,2006). Even more exciting is the landmark discovery that JAK2 goes into the nucleus and directly interacts with nucleoprotein independent of STATs protein (Nilsson et al.,2006,2010; Dawson et al.,2009; Helmer et al.,2010).
JAK2 resides in the cytoplasm and nucleus, and its subcellular distribution is under strict control at the transcriptional level, as well as post-translationally by phosphorylation, acetylation, etc. (Shi et al.,2006; Valentino and Pierre,2006; Dawson et al.,2009; Rinaldi et al.,2010). Cytoplasmic JAK2 has a well-known role as a positive regulator of transcription of STAT target genes, such as Bcl-2, Bcl-xL, surviving, and cyclin D1 (Kim et al.,2011). Cytoplasmic JAK2 accomplishes regulation of gene expression through activation of the STAT family of transcription factors. Despite this, the identities of downstream target genes of the JAK2-STAT signaling are largely unknown. Constitutive activation of JAK2, caused by an acquired somatic mutation V617F, can phosphate and activate phosphatidylinositol 3-kinase (PI3K)/Akt and mitogen-activated protein kinase (MAPK) signaling proteins in addition to STAT proteins (Levine et al.,2007). Furthermore, the canonical JAK2-STAT signaling has been shown to crosstalk with other signaling pathways, such as nuclear factor-κB (NF-κB) (Digicaylioglu and Lipton,2001), Notch (Kamakura et al.,2004), epidermal growth factor receptor (EGFR) (Colomiere et al.,2009). It is becoming clear that the function of nuclear JAK2 is not the same as that of cytoplasmic JAK2. Accumulated evidence indicates that nuclear JAK2 plays an important role in epigenetic modification and regulation of nucleoprotein in addition to regulation of cell cycle progression, apoptosis and genetic instability (Ito et al.,2004; Dawson et al.,2009; Nilsson et al.,2006,2010; Rinaldi et al.,2010). However, the function of nuclear JAK2 is still being elucidated.
In the past two decades it has been established that JAK2 is linked to an incredibly diverse set of key cellular functions including cell cycle progression, apoptosis, proliferation, differentiation, survival, genetic instability, and alteration of heterochromatin. Thus, the emerging link between JAK2 and many diseases, especially cancer, has made it the focus of intense study. However, it is only in the past few years that evidence for the presence and function of JAK2 in the nucleus has become convincing. The existence of nuclear JAK2 pathways raises several important questions. Why are they needed? Are they truly independent of STATs? what downstream target genes they regulate? What does JAK2 do in the nucleus to effect cellular proliferation and carcinogenesis? Below we examine these questions and discuss on both form and function of nuclear JAK2 pathways as well as their impact on carcinogenesis and cancer treatment.
Canonical JAK2 Pathways
JAK2 is a nonreceptor tyrosine kinase that acts in numerous cellular signal transduction systems (Wallace and Sayeski,2006; Ma and Sayeski,2007). Activation of the JAK2 signaling can be initiated by binding of ligands, such as growth hormone (Argetsinger et al.,1993), interferon γ (Silvennoinen et al.,1993), erythropoietin (EPO) (Witthuhn et al.,1993), IL-6 (Parganas et al.,1998; Hong et al.,2007; Starsíchová et al.,2010), prolactin (PRL) (Sakamoto et al.,2007; Brooks and Waters,2010; Sakamoto et al.,2010), leptin (Frühbeck,2006; Robertson et al.,2010), or osteopontin (Behera et al.,2010), to their transmembrane receptors, respectively. In the classical pattern of JAK2–STATs signaling (Fig. 1), JAK2 is located in the cytoplasm and binds to a proline-rich region in the proximal membrane domain of various receptors. Activation of the pathway is initiated by binding of ligands (such as a cytokine) to transmembrane receptors (Rawlings et al.,2004; Li,2008). Upon binding of the ligand to its receptor, the receptors undergo a conformational change which brings two JAK2 molecules in close proximity to dimerize and trans- or autophosphorylate at conserved tandem tyrosines 1007, and the phosphorylation of JAK2 ultimately results in ubiquitination and eventual degradation of phospho-JAK2 (Feng et al.,1997; Aaronson and Horvath,2002; Rawlings et al.,2004; Li,2008). An phosphorylated JAK2 then phosphorylates specific tyrosine residues on the cytoplasmic tails of the receptors, offering docking sites for SH2 domain-containing proteins (Ferrand et al.,2005; Godeny and Sayeski,2007), such as the STATs, which subsequently dimerize, translocate to the nucleus and bind specific gene promoters that regulate proliferation and differentiation (Rawlings et al.,2004; Li,2008). Thus, JAK2 manages to conduct signals from the cell surface to the nucleus through tyrosine phosphorylation signaling cascades. However, activated JAK2 does not seem to exhibit specificity for a particular STAT, as different STATs might be phosphorylated by other tyrosine kinases under different conditions (Nieborowska-Skorska et al.,1999; Jatiani et al.,2010).
Previous studies have demonstrated that the function of canonical JAK2 in carcinogenesis is mediated through cytoplasmic phosphorylation of various STAT proteins (esp. STAT3 or STAT5), which are subsequently translocated into the nucleus to up-regulate STAT target genes (Fig. 1) (Aaronson and Horvath,2002; Rawling et al.,2004; Godeny and Sayeski,2007; Navarro et al.,2009; Yun et al.,2010). Among the genes known to be regulated by STAT proteins are those encoding cell-survival factors, such as the B-cell lymphoma 2 (Bcl-2) family of proteins (Zhao et al.,2008), those involved in cell proliferation, such as cyclin D1 and c-Myc (Catalano et al.,2009), and those implicated in angiogenesis or metastasis, such as vascular endothelial growth factor (VEGF) (Kupferman et al.,2009). Conceivably, upregulation of these genes that are important for cell proliferation and/or survival would promote cancer formation. Thus, it has been assumed that increased expression of JAK2/STATs encourages tumorigenesis (Fig. 1). Major questions remain concerning how the JAK2 cascade functions to control specific gene expression patterns, and how the cascades are regulated.
Aberrant activation of JAK2-STATs signaling has been shown in multiple solid tumors and leukemia (Yoshikawa et al.,2001; Tong et al.,2008; Meier et al.,2009). While many of the mechanisms involving JAK2-STATs activation during cancer progression are incompletely understood, striking data have shown that JAK2 and STAT3 were constitutively activated in human hepatocellular carcinoma (HCC) cells (Yoshikawa et al.,2001; Schneller et al.,2011). Additionally, overexpression of BRCA1 enhanced the ability of JAK2 to activate STAT3 in human prostate cancer cells (Gao et al.,2001). Interestingly, aberrant activation of STAT3 in tumors is associated with tumor escape from immune attack (Kusmartsev and Gabrilovich,2006). It is worth noting that the constitutive activation of JAK2 in primary mediastinal large B-cell lymphoma with constitutive STAT6 activation is not related to an overexpression at the protein level, but is due to delayed protein degradation of JAK2 (Guiter et al.,2004). Therefore, when studying the function of JAK2, in addition to the protein levels, we should detect the phosphorylation and degradation of JAK2 in cancer cells. The canonical JAK2 signaling also plays an important role in the regulation of other cancer cellular functions: for example, the predominant role of JAK2 in many types of cancer is tumorigenic via regulation of apoptosis (Yun et al.,2010), detachment-induced apoptosis-anoikis (Bretland et al.,2001), or autophagy (Yoon et al.,2010).
Nuclear Localization of JAK2
Early studies of JAK2 localization, using immunohistochemical staining techniques, found it in the cytoplasm; and these focused primarily on overexpressed protein in tumorigenic cell lines and tissues (Knoops et al.,2008; Gorantla et al.,2010). However, some JAK2-regulated genes, such as the haematopoietic oncogene LIM domain only 2 (LMO2), identified in previous studies did not contain a putative STAT binding site, suggesting that STAT proteins are not likely to mediate all of the signaling activities of JAK2 (Dawson et al.,2009; Sattleret al., 2009). This suggests that JAK2 can act via STAT independent mechanisms, and non-STAT JAK2 signaling may exist in mammalian cells. The hypothesis of a primary nuclear JAK2 signaling pathway is supported by data about constitutive expression of JAK2 found in the cell nucleus (Sorenson and Stout,1995; Lobie et al.,1996; Ram and Waxman,1997; Wallace et al.,2004; Nilsson et al.,2006; Dawson et al.,2009). A number of reports have highlighted the presence of active JAK2 in the nucleus of different cell lines and tissues as following:
Early as in 1995, Sorenson and Stout first observed the immunohistochemical localization of JAK2 in the nucleus of nearly all pancreatic islets isolated from neonatal, adult, and day 14 pregnant rats (Sorenson and Stout,1995). Subsequent publications demonstrated that constitutive nuclear localization of JAK2 has been observed in islet-derived INS-1 cells (Stout et al.,1997) and primary rat β-cells (Heitmeier et al.,1999). Moreover, the GH (growth hormone) receptor-associated JAK2 was found in the nucleus in cultured liver cells and rat liver in vivo (Ram and Waxman,1997), and STAT3, but not STAT5b, associated with nuclear JAK2 in rat liver cells (Chatterjee-Kishore et al.,2000). At the time, many researchers felt that the presence of JAK2 in the nucleus may be merely an artifact, because all the data had relied solely on immunohistochemical techniques; however, we can clearly infer that nuclear JAK2 does exist by finding that there is a direct interaction between JAK2 and nucleoproteins in the nucleus (Dawson et al.,2009). This study showed that JAK2 has a previously unrecognized nuclear pool in haematopoietic cells because JAK2 existed in the nuclei of cell lines with a V617F mutation. Similarly, a significant proportion of nuclear JAK2 was present in various cells independent of the JAK2 mutation status (Dawson et al.,2009). Since JAK2 is expressed constitutively in the nucleus along with a number of receptors of hormones, cytokines and growth factors that undergo nuclear translocation, the realization of a primary nuclear JAK2 signaling pathway is most likely true (Ito et al.,2004; Krolewski,2005). However, Moulin et al. (2003) found that after GH stimulation, JAK2 was clearly located in the cytoplasm but not in the nucleus by transfecting a tagged version of JAK2 expression vector in CHO cells. These studies suggest that subcellular localization of JAK2 may be cell type- and context-dependent.
The regulation of nuclear JAK2 is not well-understood, but very interesting in-vitro studies showed that growth hormone and its receptor undergo nuclear translocation to tyrosine phosphorylate nuclear localized JAK2, and exogenous GH can tyrosine phosphorylate nuclear JAK2 without affecting the subcellular location of JAK2 in cells (Lobie et al.,1996; Ram amd Waxman,1997). Early studies showed that nuclear JAK2 interacts with STAT5b by phosphorylating protein tyrosine phosphatase SHP-1 and enables the dephosphorylated STAT5b back to the cytoplasm (Ram and Waxman,1997). Similarly, Murray et al. (2005) found that decreased expression of the nuclear JAK2 tyrosine kinase and SHP-1 was associated with accumulation of nuclear phospho-STAT5 in vitamin A-deficient liver, and this nuclear retention of phospho-STAT5 is induced by decreased expression of SHP-1, which impairs dephosphylation of STAT5 and diminished nuclear localization of JAK2 in response to GH stimulation. The endogenous JAK2 was further so clearly detected in both cytosol and nuclear fractions in COS-7 cells (Lee and Duhé,2006). However, JAK2 activity and the phosphorylated form within the cell is not indispensable for nuclear translocation of GH; on the other hand, the nuclear shifting of both GH and its receptor are unrelated with JAK2 (Graichen et al.,2003; Mertaniet al.,2003). These studies suggest that JAK2 may have a special procedure of entering the nucleus, but the underlying molecular mechanism is poorly understood.
Accumulated evidence has indicated a nuclear localization of JAK2, but the cellular physiological functions of it has not been clear. Clearly, future investigations will identify additional functions of JAK2 that depend on its subcellular localization. On these basis, nuclear localization of phospho-JAK2 has been recently found in human brain tumors using a specific monoclonal antibody against JAK2 phosphorylated on tyrosine 1007 and 1008 (Kondyli et al.,2010). Meanwhile, confocal immunofluorescence imaging and Western blot analysis also corroborate a noteworthy pool of phosphonuclear JAK2 on tyrosine 1007 and 1008 in HRE-H9 cells, and JAK2 kinase inhibitors abolish this nuclear cluster of active JAK2 (Helmer et al.,2010). Interestingly, Rinaldi et al., (Rinaldi et al.,2010) found the V617F mutation leads to a nuclear accumulation of JAK2 in leukemia cells, which can be reversed by the JAK2 inhibitor AG490, suggesting the regulation of phosphorylation JAK2 could be one of the modifications that control nuclear translocation of JAK2, similarly to what happens to the STATs (Adach et al.,2009); however, so far it is undiscovered what exact signals are required for the nuclear translocation of normal and mutated JAK2. Given that recent data showed that LMO2 is directly regulated by nuclear JAK2 (Dawson et al.,2009), and AG490 is able to suppress cell growth of human leukemia cells by reverting nuclear JAK2 and normalizing LMO2 levels (Rinaldi et al.,2010), blocking of nuclear translocation of JAK2 may be a new treatment strategy for leukemia patients with JAK2 mutation (Rinaldi et al.,2010).
Although the STATs that are translocated into the nucleus are already phosphorylated by activated JAK2 in the cytoplasm (Godeny and Sayeski,2007; Li,2008), it remains to be determined whether JAK2 kinase can phosphorylate STAT proteins in nuclei. It is worth mentioning that GH-activated nuclear JAK2 binds specifically to STAT3 which is constitutively present in nuclei in the nontyrosine-phosphorylated state in rat liver cells, so nuclear inactive STAT3 is probably a substrate of GH-dependent nuclear JAK2 (Ram and Waxman,1997). Even so, the nuclear localization of JAK2 indicates that JAK2 may have a special function during cell signal transduction because JAK2 cannot bind to the membrane receptors as long as it is restricted to the nucleus. Indeed, nuclear JAK2 is believed to be kinetically activated to mediate signaling pathway within the nucleus to stabilize histone phosphorylation and nuclear factors (Nilsson et al.,2006; Dawson et al.,2009). Furthermore, the mechanism of this nuclear pathway may be similar to the cytoplasmic one, because the relevant receptors and ligands that are upstream of canonical JAK2 signaling are present in the nucleus (Ito et al.,2004). Interestingly, the earlier research by Lobie et al. (1996) has demonstrated that nuclear JAK2 could directly regulate GH-mediated signaling molecules in the nucleus.
Nuclear JAK2 Pathways
The concept that nuclear JAK2 plays a critical role in various cell physiological processes has been advanced in several recently published studies and editorials (Nilsson et al.,2006,2010; Dawson et al.,2009; Helmer et al.,2010; Rinaldi et al.,2010). Considering that STATs are not likely to mediate all of the JAK2 signaling activities, we speculate that JAK2 might phosphorylate other substrates in addition to STATs. Recent studies confirmed that active JAK2 goes into the nucleus, and directly phosphorylates histone H3Y41, NF1-C2, and RUSH-1α in the nucleus of mammary cells independent of STATs protein (Fig. 2) (Nilsson et al.,2006; Dawson et al.,2009; Sattler and Griffin,2009; Helmer et al.,2010). Newly discovered pathways mediated by nuclear JAK2 are described in detail as following
A work recently published in Nature confirmed that JAK2 has previously undiscovered functions in the nucleus to phosphorylate Tyr 41 (Y41) on histone H3 directly (Dawson et al.,2009). They found that nuclear JAK2 phosphorylates histone H3 at tyrosine 41 (H3Y41), and phospho-H3Y41 releases heterochromatin protein 1α (HP1α), from chromatin, then resulting in transcription of LMO2 which is originally repressed by HP1α (Fig. 2) (Dawson et al.,2009). Typically, HP1 is an important component of heterochromatin and plays an indispensable part in heterochromatin-mediated gene silencing (Grewal and Elgin,2002). Recently, it has been proven to serve as a tumor suppressor protein (Braig et al.,2005). In addition, Dawson et al. used the JAK2 inhibitor TG101209 to inhibit JAK2 activity of human leukaemic cells and observed that both the expression of LMO2 and the phosphorylation of H3Y41 decreased, while the binding of HP1α at the Y41 site increased (Dawson et al.,2009; Sattler and Griffin,2009). These studies uncover a direct relationship between JAK2 and LMO2, two oncogenic genes responsible for normal and leukaemic haematopoiesis (Nam et al.,2008). Interestingly, HP1α has been shown to interact with the oncogene c-Myc, one of STAT5-targeting genes (Norwood et al.,2004; Knoepfler,2007). What's more, occupation of genes by STAT5 overlap nearly one third of the cases with phosphorylated H3Y41 (Dawson et al.,2009; Sattler and Griffin,2009). Therefore, we suppose that any factor that interferes with H3Y41 phosphorylation may lead to heterochromatin-induced gene silencing and resistance to the oncogenic role of JAK2 activation.
However, it is still unknown whether the canonical JAK2 pathway and JAK2-H3Y41-HP1α pathway take effect respectively or cooperate mutually in gene regulation, since without a recognized STAT5 binding site does not absolutely exclude the binding of STAT5 to JAK2-regulated genes in the nucleus (He and Zhang,2010). Therefore, it is urgent to detect both STAT5 and HP1α binding at these JAK2-regulated genes by chromatin immunoprecipitation (ChIP) or other analyses to determine how important it is for HP1α to silence the specific oncogenes in normal and cancer cells.
The polypeptide hormone prolactin (PRL) binds to its receptor, which belongs to the cytokine receptor superfamily (Langenheim and Chen,2009; Sakamoto et al.,2010), and specifically activates JAK2, but not JAK1, JAK3, or Tyk2, then increases the subsequent phosphorylation of STAT5a and STAT5b in the cytoplasm, and then phosphorylated STAT5 dimerizes, translocates to the nucleus, and binds to its response elements (García-Martínez et al.,2010; Johnson et al.,2010). However, Nilsson et al. (Nilsson et al.,2006) found that prolactin-induced JAK2 activation is present in the nuclei of the mammary epithelial cells, and the JAK2 tyrosine phosphorylates NF1-C2 instead of STAT5 in the nuclei (Fig. 2), suggesting there is a non-STAT-including pathway by which the prolactin/JAK2 functions in the nuclei of the mammary epithelial cells. Even more interesting is that nuclear JAK2 could prevent NF1-C2 from proteasomal degradation, which regulates the proportion of active nuclear transcription factor; otherwise, NF1-C2 will be rapidly degraded in the absence of nuclear JAK2 (Nilsson et al.,2006,2010).
An earlier study done by Johansson and colleagues showed that NF1-C2 activates the tumor suppressor gene p53 transcription in mouse mammary glands by depending on binding of NF1-C2, not STAT5, to the p53 promoter (Johansson et al.,2003). In addition, they found that NF1-C2 inhibition suppresses p53 transcription with the use of NF1-C2 oligonucleotide decoys (Johansson et al.,2003). In the follow-up study, they further found that the increase in NF1-C2 protein levels caused by prolactin/JAK2 is regulated at the post-transcriptional level in the mouse mammary epithelial cells (Johansson et al.,2005; Nilsson et al.,2006). Collectively, these data suggest an interaction between prolactin/JAK2 and p53 in the mammary gland, through NF1-C2.
RUSH-1, the ortholog to human SWI/SNF-related matix-associated actin- dependent regulator of chromatin subfamily A member 3 (SMARCA3), is a nuclear effector of prolactin signals (Hewetson et al.,2002). RUSH-1 symbol recognizes two alternatively spliced transcription factors: a full-length progesterone-dependent α isoform, and a truncated oestrogen-dependent β isoform (Hayward-Lester et al.,1996). In vitro studies showed that resultant proteins have the same DNA-binding domains, nuclear localization signals, and RING (really interesting new gene) finger motifs, but RUSH-1α has seven conserved DNA-dependent ATPase domains in contrast with RUSH-1β, which is truncated after the first four (Sheridan et al.,1995; Hewetson and Chilton,2003; Chilton and Hewetson,2008).
A direct physical interaction between nuclear JAK2 and RUSH-1α in nucleus was proved by co-immunoprecipitation of the two nucleoproteins (Hewetson et al.,2002). Moreover, they found that RUSH contains no binding sites for STATs, and there was no physical affiliation between RUSH and all STAT family members (Hewetson et al.,2004). In addition, the current model of RUSH phosphorylation by nuclear active JAK2 is deduced from the previous finding that RUSH's DNA-binding ability is mediated by tyrosine phosphorylation (Hewetson et al.,2004). Thus it prompted speculation about a JAK2/RUSH pathway, which may be a substitute for the canonical JAK2/STAT signaling. Helmer et al. (2010) further confirmed that nuclear active JAK2 affects transcription of the RUSH gene by directly phosphorylating a progesterone-dependent pool of RUSH in HRE-H9 cells . It's encouraging that the recent follow-up studies confirmed a functional JAK2/RUSH pathway. Therefore, these findings strongly suggest that active nuclear JAK2 could directly phosphorylate RUSH-1α in the nucleus of mammary cells independent of STATs protein (Fig. 2). However, it is necessary to explore how RUSH-1α regulates JAK2-target genes in tumorous and nontumorous cells.
Crosstalk between Nuclear JAK2 and Other Pathways
Another important topic for future studies is to define the mechanisms of crosstalk between nuclear JAK2 and other pathways. Understanding the complexity of a signaling pathway depends on an elucidation of the underlying signaling regulatory networks, at the cellular and intercellular levels and in their temporal dimension (Caron et al.,2010; Guo et al.,2011). Recent studies demonstrated that crosstalk between nuclear JAK2 and other signaling pathways or factors, such as NF-κB, Notch and Hedgehog, can occur, and JAK2 may regulate transcription mediated by these other signaling pathways and factors in nucleus. The identification of nuclear JAK2 in regulation of transcription, translation, and epigenetic modification by cross-talking with other signaling pathways and factors will provide a brand-new vision of JAK2 biology.
Interestingly, a molecular link between JAK2 and NF-κB has also been identified in lymphoma cells (Joos et al.,2003; Savage et al.,2003). A previous study showed that JAK2 and NF-κB signaling are interrelated in erythropoietin-associated neuroprotective actions (Digicaylioglu and Lipton,2001). In addition, Funakoshi-Tago et al. (Funakoshi-Tago et al.,2011) reported that JAK2 is an important signal transducer in IL-33-activated NF-κB signaling. Furthermore, Kai et al. (2010) found that JAK2 signaling could induce the inactivation of IκBα (inhibitor of NF-κB α) under IFN-γ stimulation.
Another transcription factor c-Rel is proved to be a member of the NF-κB transcription factor family that generally functions in the nucleus (Gerondakis et al.,2006). In contrast, RUSH physically interactions with Egr-1 and c-Rel were biochemically confirmed by Hewetson and Chilton (Hewetson and Chilton,2008). Early growth response protein 1 (Egr-1), one nuclear phosphoprotein, has been known to interact with tumor suppressor p53 (Liu et al.,2001). Egr-1 can interact physically and functionally with NF-κB in vascular smooth muscle cells and vascular endothelial cells (Hasan and Schafer,2008; Lv et al.,2009). Chernyavsky and colleagues demonstrated that the SLURP-1 (secreted mammalian Ly-6/urokinase plasminogen activator receptor-related protein-1)-induced elevation of NF-κB is mediated by alpha7-nicotinic acetylcholine receptor and involves activation of JAK2 (Chernyavsky et al.,2010). In their study, inhibition of JAK2 activity by AG-490 significantly reduced the rSLURP-1-induced upregulation of the NF-κB gene expression in the esophageal squamous cell line Het-1A (Chernyavsky et al.,2010). These studies indicate that nuclear JAK2-RUSH can crosstalk with NF-κB pathways. Given the fact that both JAK2 and NF-κB often play a crucial role in carcinogenesis in the same tumor, such as breast cancer (Zhou et al.,2011), the exact mechanism of crosstalk between the two signaling pathways remains to be studied in depth in future research, especially in the field of cancer research.
IL-6 signals via a heterodimeric IL-6R/gp130 receptor complex, whose engagement triggers the activation of JAK kinases and the downstream effectors STAT proteins (Kishimoto,2005). Since Notch pathway is a critical downstream target of IL-6 (Sansone et al.,2007), JAK2 might functionally and biochemically interact with Notch signaling. Notch signaling has been reported to promote STAT3 activation, and its effectors Hes1 and Hes5 have been found to associate directly with JAK2 (Kamakura et al.,2004). Evidence for interaction between the JAK and Notch pathways has also been provided by work from Drosophila (Josten et al.,2004), and genetic screens in Drosophila have identified additional potential modifiers of the JAK pathway (Bach et al.,2003). Moreover, Nie et al. (Nie et al.,2010) found that Notch signaling transcriptionally activates the gene encoding ankyrin-repeat SOCS box-containing protein 2 (Asb2), while Asb2 binds JAK2 directly. In addition, they also found that the Notch intracellular domain (NICD), an active form of Notch, could induced ubiquitination and degradation of JAK2 in BaF3 cells (Nie et al.,2010).
Recent studies, including ours, show that NICD, Hes1, and Hes5 are mostly located in cellular nucleus to start transcription reactions upon activation of Notch (Duan et al.,2006; Yao et al.,2007; Bansal et al.,2009; Arnett et al.,2010), and NICD must accumulate in the nucleus to induce transformation of RKE cells (Jeffries and Capobianco,2000). These findings suggest that activation of Notch signaling occurs mainly in the nucleus. Interestingly, a direct interaction between Notch with histone H3 was found in the nucleus (Liefke et al.,2010; Di Stefano et al.,2011). Furthermore, Notch signaling has been confirmed to play a crucial role in the initiation and progression of LMO2-induced T-ALL (T-cell Acute Lymphoblastic Leukemia) (Weng et al.,2004; Curtis and McCormack,2010), while LMO2 has been proven to be one downstream target gene of JAK2-H3Y41-HP1α pathway (Dawson et al.,2009). Therefore, JAK2 might interact with Hes1, Hes5, NICD, or other components of Notch pathway via nucleoproteins, such as histone H3. These studies suggest that a crosstalk between nuclear JAK2-H3Y41-HP1α and Notch signaling occurs in cancer, although this hypothesis needs further validation by subsequent studies.
Previous studies described a novel signaling pathway by which prolactin, via nuclear localized JAK2, activates the transcription factor NF1-C2 (Nilsson et al.,2006), and prolactin can also take part in the regulation of cell development via Hedgehog signaling (Sbrogna et al.,2003). A more recent study reported that NF1-C2 binds to the promoter of the forkhead box F1 (FoxF1), an epithelial-to-mesenchymal transition (EMT) inducer and tumorigenic gene (Nilsson et al.,2010). Also, they found that, as a result of this binding, NF1-C2 curbs FoxF1 expression, and ultimately suppresses cell motility, invasion, EMT, and tumor progression (Nilsson et al.,2010). Moreover, FoxF1 is mediated by Hedgehog signaling to regulate the interactions between mesenchymal cells and epithelial cells in lung development (Shannon and Hyatt,2004; Saito et al.,2010). This research demonstrated that a crosstalk between nuclear JAK2-NF1-C2 and Hedgehog pathways occurs in tumor and nontumor cells.
Role of Nuclear JAK2 in Carcinogenesis
The function of nuclear JAK2 has become a hot topic recently. Because the mechanism for regulating the activation of JAK2 is complicated, analyses of the role of JAK2 in the nucleus may elucidate novel functions. Therefore, further functional studies on human tumors are necessary to elucidate the exact molecular mechanisms of JAK2 nuclear translocation and its effects on carcinogenesis in addition to various physiological processes.
The role of cytoplasmic JAK2 has been studied very extensively in the field of cancer research. In hematologic neoplasms such as chronic myelomonocytic leukemia and acute myeloid leukemia JAK2 is constitutively activated because of a mutation (Levine et al.,2005; Baxter et al.,2005; Tefferi et al.,2005; Bina et al.,2010), while aberrant JAK activity due to overexpression of wild-type JAK2 is also associated with a number of solid tumors, including gastric carcinoma, hepatocellular carcinoma, and head and neck squamous cell carcinoma (Lee et al.,2006; Fuke et al.,2007; Hedvat et al.,2009; Kupferman et al.,2009). In different experimental models, tumorigenesis is associated with increased activity of JAK2, and functional ablation of JAK2 protects against the onset of various tumors (Sakamoto et al.,2009; He and Zhang,2010; Agrawal et al.,2011). Studies in our laboratory showed that wild-type JAK2 is significantly upregulated in primary gastric cancers, and down-regulation of JAK2 significantly suppresses the proliferation of gastric cancer cells, suggesting that JAK2 may play a key role in gastric carcinogenesis (Ding et al.,2010). However, the role of nuclear JAK2 in carcinogenesis of solid tumor, such as gastric cancer, remains poorly characterized. These reports addressed the question of the role of nuclear JAK2 in carcinogenesis.
Nuclear JAK2 as an oncogene
Until recently, nuclear JAK2 signaling had always been associated with an oncogenic role during tumor formation (Rinaldi et al.,2010). Its function as an oncogene is best documented in the haematopoietic compartment in which aberrant nuclear JAK2 signaling leads to myeloproliferative neoplasia (Rinaldi et al.,2010). Inhibition of the JAK2 tyrosine kinase induced apoptosis and inhibited survival in CML (chronic myelogenous leukemia) cell lines with either JAK2 short interfering RNA or JAK2 kinase inhibitor AG490 (Samanta et al.,2006).
Mutations in pathways, such as IκB, Wnt, and ErbB, often lead to tumorigenesis, as is also true for JAK2 (Rumi et al.,2010). Recently, a novel somatic point mutation V617F in JAK2 was identified in the majority of polycythemia vera patients as well as other myeloproliferative disorders including essential thrombocythemia, and primary myelofibrosis. This very mutation locates in the pseudokinase domain, JH2, of JAK2, and leads to deregulated kinase activity and constitutive activation of JAK2, which can promote tumor development (Dusa et al.,2010; Ayad and Nafea,2011).
Meanwhile, Rinaldi et al. (Rinaldi et al.,2010) reported that nuclear JAK2 is present only in JAK2V617F-positive myeloproliferative neoplasia patients and JAK2V617F-positive cells, but not in patients with wild type JAK2, indicating JAK2V617F mutation may affect nucleo-cytoplasmic localization of JAK2. In addition, their work also showed that nuclear JAK2 up-regulates expression of LMO2 in those cells with JAK2V617F mutation, and the selective JAK2 inhibitor AG490 plays a role in anti-tumor by normalizing LMO2 levels in V617F positive K562 cells and restoring the cytoplasmic localization of JAK2 (Rinaldi et al.,2010). These findings strongly suggest that AG490 could shut off the pathway to prevent JAK2 molecules from getting into the nucleus, leaving the mutated JAK2 predominantly in the cytoplasm. However, most current JAK2 inhibitors are multikinase inhibitors, which can have an impact on other signaling pathways (Quintás-Cardama et al.,2011). The question of the specificity of AG490 to inhibit JAK2 needs to be verified (Gu et al.,2001; Wilks,2008). Thus, it's worthy of further exploration to support the hypothesis that an alteration of JAK2 activity could interfere with the nuclear localization of JAK2, using more highly specific JAK2 inhibitors or RNA interference targeting of JAK2. It will also be very interesting to verify whether AG490, other more highly specific JAK2 inhibitors or JAK2 siRNA could regulate or interfere with the nuclear localization of JAK2 in other types of tumor cells.
Recent work has described another exciting role for nuclear JAK2: excessive activation of JAK globally destabilizes heterochromatin and disturbs heterochromatin-induced gene silencing, which is essential for JAK overactivation-induced cell overproliferation and tumorigenesis (Shi et al.,2006). In any case, it is now clear that histones are a dynamic component of chromatin and not simply inert DNA-packing material in nuclei (Burgess and Zhang,2010). Reversible acetylation of nucleosomal histone tails, an epigenetic modification involved in the regulation of gene expression, is balanced by the opposing activities of histone acetyltransferases (HATs) and histone deacetylases (HDACs) (Shukla et al.,2008; Biancotto et al.,2010). Acetylation of histones provides a more open chromatin structure which correlates with gene activation, while histone deacetylation mediates transcriptional repression (Mai et al.,2005). For instance, methylation of histone H3 at Lys9 (H3mK9), catalyzed by methyltransferase Su(var)3–9 homologs, is associated with heterochromatin formation and heterochromatic gene silencing (Jenuwein and Allis,2001). H3mK9 provides binding sites for HP1, which in turn recruits Su(var)3–9 and associated proteins, leading to heterochromatin assembly and spreading (Lachner et al.,2001; Grewal and Elgin,2002). The levels of HP1 and the heterochromatic gene silencing machinery determine the outcome of a Drosophila JAK kinase, hopscotchTumorous-lethal (hopTum-l) tumorigenicity without affecting JAK/STAT signaling per se (Shi et al.,2006). Induction of heterochromatin formation mediated by Suv39h1 (a Su(var)3–9 ortholog in mammal) can function as a tumor-suppressive mechanism that restricts cancer development (Braig et al.,2005). These results have obvious implications for the development of potential anticancer therapies because many chemotherapeutic agents act through affecting heterochromatin and genetic stability.
Consistently, JAK loss of function enhances heterochromatic gene silencing, whereas overexpression of HP1 suppresses oncogenic JAK-induced tumors (Shi et al.,2006,2008). Constitutive activation of nuclear JAK2 can lead to oncogenic activation and genomic instability through direct phosphorylation of H3Y41 and displacement of the transcriptional repressor HP1α from heterochromatin, resulting in transcription of genes repressed by HP1α, such as the haematopoietic oncogene LMO2, which did not accord with the property of a STAT5-regulated gene (Dawson et al.,2009). Therefore, HP1 levels determine the effect of JAK tumorigenicity without affecting JAK/STAT signaling (Shi et al.,2006). Further studies are needed to explore whether there are other oncogenes whose transcription is activated due to release of HP1α from chromatin by active nuclear JAK2 independent of STATs.
Conversely, high expression of endogenous and/or exogenous JAK2 in leukemic cell lines and haematopoietic cells is usually accompanied by higher levels of H3Y41 phosphorylation (Dawson et al.,2009; He and Zhang,2010). This new finding corroborates that a JAK2 inhibitor and the histone demethylase JMJD2C knockdown have a synergistic effect on inducing lymphoma cell apoptosis by decreasing H3Y41 phosphorylation and increasing H3mK9 (Rui et al.,2010). Both in vitro and in vivo studies showed that JAK2 appears to be the only kinase responsible for H3Y41 phosphorylation (Dawson et al.,2009; He and Zhang,2010). However, a study recently published in Nature Cell Biology by Griffiths et al. (Griffiths et al.,2011) demonstrated that H3Y41 is not only phosphorylated by JAK2 in embryonic stem (ES) cells, but also by other members of the JAK family, such as JAK1, suggesting that multiple JAK family members can phosphorylate H3Y41 and changes in H3Y41 phosphorylation should be studied in diseases involving aberrant activation of JAKs.
The mutations in the genes encoding heterochromatin protein 1 (HP1) enhance tumorigenesis induced by JAK, while the overactivation of JAK can disrupt HP1-mediated heterochromatic gene silencing, allowing expression of genes that are repressed under normal circumstances (Shi et al.,2006). In addition, JAK inactivation enhances heterochromatic gene silencing, whereas overexpression of HP1 suppresses oncogenic JAK-induced cancers (Shi et al.,2006). However, Bártová et al. (2005) found that HDAC inhibitors (HDACi) trichostatin A (TSA) and sodium butyrate (NaBt) not only decrease the protein levels of HP1, but also reposition HP1 proteins more internally in the nucleus, which makes HP1 more close to interchromatin compartments. The HDACi-induced repositioning of HP1 may be involved in the mechanism of tumor suppression, which may partly explain the controversial issue here. However, whether HDACi causes dynamic reorganization of chromatin leading to JAK2 gene silencing needs to be confirmed.
Zhu et al. (2009) recently reported that the HDACi valproic acid (VPA) has potent and very highly selective inhibition on STAT3 tyrosine phosphorylation in normal and tumor cell lines that constitutively expressed STAT3 . In addition to this, the HDACi panobinostat appears to induce inactivation of STAT3 by increasing its binding to the serine/threonine protein phosphatase 2A (PP2A), which dephosphorylates phospho-STAT3 (Cotto et al.,2010). Since there is a direct interaction between JAK2 with histones in the nucleus (Dawson et al.,2009; Sattler and Griffin,2009), and STAT3 and STAT5 are the preferred downstream targets of phosphorylated JAK2 (Ihle and Gilliland,2007), we speculate HDACi could directly affect the actions of JAK2 in the nucleus.
Nuclear JAK2 as a tumor suppressor.
However, more recent data show that nuclear JAK2 can also act as a tumor suppressor by stabilizing NF1-C2 in the mammary gland. Forced expression of NF1-C2 abolishes tumorigenicity in nude mice, and suppresses the invasive phenotype, such as epithelial-to-mesenchymal transition (EMT), motility, cell–cell adhesion, tumorigenicity, of breast cancer cells (Nilsson et al.,2006,2010). Additionally, this study showed that active nuclear JAK2 suppresses tumorigenesis and EMT by repressing Forkhead box F1 (Nilsson et al.,2010). Thus, JAK2 in the nucleus of mammary epithelial cells appeared to be required for tumor-suppressing ability, and the activation of nuclear JAK2 could result in decreased tumor growth, explaining another aspect of JAK2 tumor-suppressor function.
There is no doubt that apoptosis is unequivocally antioncogenic in character (Li et al.,2010; Liu et al.,2011). It was suggested that JAK2 was capable of inducing expression of c-Myc, p53 and Bcl-2, which are actively involved in the promotion of cell apoptosis, thereby facilitating tumor cell death. STAT5 acts as a survival factor preventing apoptosis of terminally differentiated epithelial cells (Eilon and Barash,2011), while NF1-C2 might have opposing effects by stimulating p53 gene expression in the mouse mammary gland (Johansson et al.,2003). Thus decrease in nuclear JAK2/NF1-C2 correlates with a lower level of p53 mRNA (Concin et al.,2003). Hence, nuclear JAK2/NF1-C2 signaling participates in the tight control of p53 in some types of tumor cells, such as breast cancer cells (Nilsson et al.,2006). The function of nuclear JAK2 and NF1-C2 in p53 regulation is most probably to keep sufficient levels of latent p53 mRNA to ensure rapid accumulation of p53 protein in response to cellular stress (Nilsson et al.,2006). Thus, nuclear JAK2 might act as a tumor suppressor in collaboration with p53 via regulation of NF1-C2, and then affect apoptosis in some tumor cells.
The two faces of nuclear JAK2 in cancer.
It is not surprising to find that an oncogene acts as tumor suppressor in some cancers, while tumor suppressor acts as an oncogene in some other cancers (Radtke and Raj,2003; Lewis-Tuffin et al.,2010), such is the case with nuclear JAK2. An interesting aspect of nuclear JAK2 is its apparently opposite functions in tumor development, because it can act as an oncogene or tumor suppressor. It is possible that the outcome of nuclear JAK2 signaling is also dependent on its natural cellular context. However, the question of what this exact cellular contexts is remains difficult to answer. It has led to the belief that, in some normal, quiescent tissues or cells, JAK2 is localized primarily in the nucleus (Nilsson et al.,2006,2010); Whereas, in some neoplastic tissues or cells, cytoplasmic JAK2 predominates (Lee et al.,2006; Fuke et al.,2007; Hedvat et al.,2009; Kupferman et al.,2009). Therefore, the equilibrium between nuclear and cytoplasmic JAK2 might be important in the regulation of apoptosis and cell cycle; however, whether this is a general mechanism remains to be examined.
Remarkably, various cellular events can be triggered by the same signaling pathway when its cellular context is different. Indeed, JAK2 signaling has the potential to affect the various biological processes, including cell cycle progression, apoptosis and differentiation in both normal and abnormal cellular development or tumor formation (Ferrajoli et al.,2006; Lanning and Carter-Su,2006). For example, nuclear JAK2 could affect apoptosis and cell cycle via NF1-C2 in different ways. Stable NF1-C2 protein (NF1-C2S) expressing HC11 cells showed a markedly decreased proportion of S-G2 compared with control cells; however, no difference was observed in the sub-G1 fraction in HC11 cells with and without expression of NF1-C2S, suggesting that stable NF1-C2 protein does not induce apoptosis in HC11 cells (Nilsson et al.,2010). However, nuclear JAK2 can participate in the regulation of apoptosis via NF1-C2 rather than growth and differentiation of the mammary gland, which may partly explain why apoptosis induced by nuclear JAK2 is both cell type and context dependent (Nilsson et al.,2006,2010).
Nuclear JAK2 phosphorylates histone H3, NF1-C2 and RUSH-1α in the nucleus of different cell types independent of STATs protein (Nilsson et al.,2006; Dawson et al.,2009; Helmer et al.,2010), whereas cytoplasmic JAK2 modulates the expression of oncogenes, such as c-Myc and BCL-xL, dependent of STATs protein (Xie et al.,2002; Gozgit et al.,2008). These results suggest that JAK2, which is specifically localized in the nuclear or cytoplasmic compartment, regulates special signaling events. Since nuclear JAK2 could interfere with processes of carcinogenesis not only through its catalytic activity (Rinaldi et al.,2010), but also through its interaction with nucleoproteins involved in cellular transformation (Nilsson et al.,2006; Dawson et al.,2009; Helmer et al.,2010), we must also conduct studies of its downstream target genes to understand the two faces of nuclear JAK2 in cancer.
JAK2 has been confirmed to be involved in the initiation of prolactin-induced mammary cancer (Sakamoto et al.,2010). As to the biological effects of prolactin, previous studies showed that prolactin/JAK2 promotes tumor growth (Rouet et al.,2010; Faouzi et al.,2010) or suppresses invasiveness by upholding a differentiated state (Nouhi et al.,2006; Sultan et al.,2008). More importantly, Nilsson et al., (Nilsson et al.,2006,2010) further found that the prolactin/nuclear JAK2 pathway is a major factor determining NF1-C2 levels in breast cancer by biopsy techniques. Nuclear NF1-C2 levels positively correlate with patient prognosis and survival, indicating that nuclear JAK2 may act as a tumor suppressor in breast cancer. Nuclear JAK2 can also promote human brain tumor development if it is aberrantly activated or phosphorylated (Kondyli et al.,2010), so the contradictory outcomes occur when different cell types and different background levels of endogenous JAK2 are used. Therefore, it is more likely that JAK2 has dual functions, in other words, JAK2 is a tumor suppressor and oncogene, depending on its localization, cell type and context, which is very similar to prolactin (Fernandez et al.,2010). Future research to determine the underlying mechanism by which an extracellular signal, such as prolactin, induces JAK2 nuclear translocation and identify other downstream targets of nuclear JAK2 signaling in different types of cancer cells, will have a far-reaching impact on the understanding of the two faces of nuclear JAK2.
Carcinogenic effects of JAK2 are best demonstrated in the haematopoietic system disorders in which aberrant nuclear levels of JAK2 are major reasons for the formation of myeloproliferative neoplasms (Hellström-Lindberg,2010; Vannucchi,2010). More recent data show that nuclear JAK2 can also act as a tumor suppressor in the mammary gland (Nilsson et al.,2006,2010). This bifurcation of effects is further complicated by the alterations in subcellular localization of JAK2 that have been observed. A change in intracellular localization of JAK2 thus seems to be involved in neoplastic transformation and might modulate its activity. In megakaryocytes expressing wild-type or a cytoplasm-localized mutant of JAK2, the levels of phosphorylation of STAT5 and Akt increase (Grimwade et al.,2009). This indicates that these downstream targets of JAK2 activity are regulated by cytoplasmic JAK2. By contrast, expression of JAK2V617F which localizes to the nucleus results in an increase in LMO2 in K562 and in JAK2V617F-positive CD34+ cells, but not in those carrying wild-type JAK2 (Rinaldi et al.,2010). Since the JAK2V617F mutation results in the constitutive activation or dysregulation of JAK2 (Levine et al.,2005; Baxter et al.,2005; Tefferi et al.,2005; Bina et al.,2010), another interesting but unexplored area of research involving nuclear JAK2, is how its activation level or the V617F mutation affects its function.
Collectively, these apparently conflicting functions of nuclear JAK2 strongly suggest that the different outcomes of its activation are dependent upon both cell type and its context. Moreover, we can conclude that the biological reaction of normal or tumor cells to JAK2 depends on the interaction between STAT-dependent and STAT-independent pathways. In many instances, pro-survival signaling pathways, as well as pro-apoptotic signaling pathways, are activated at the same time. Therefore, to clarify the complicated interactive relationship between nuclear JAK2 and other factors or pathways will help to unravel the mystery of the two faces of nuclear JAK2 in cancer.