Targeting Janus Kinases and Signal Transducer and Activator of Transcription 3 To Treat Inflammation, Fibrosis, and Cancer: Rationale, Progress, and Caution

A. Canonical Janus Kinase/Signal Transducer and Activator of Transcription 3 Signaling

The Janus kinase (JAK)/signal transducer and activator of transcription (STAT) signal transduction pathway is an evolutionarily conserved pathway present in Drosophila melanogaster through Homo sapiens (Hou et al., 2002). This pathway is activated in response to many protein ligands, including cytokines, growth factors, interferons (IFNs), and peptide hormones, where it regulates a wide range of cellular processes, including cell growth, proliferation, differentiation, and apoptosis (Rawlings et al., 2004; O’Shea et al., 2013). Protein ligands bind to the extracellular domains of their receptors, which transmit signals into the cytoplasm through a series of conformational changes and post-translational modifications, notably tyrosine phosphorylation, leading to reprogramming of the targeted cells. Most cytokine receptors lack intrinsic kinase activity; consequently, central to their signaling is a family of protein tyrosine kinases known as JAK that are constitutively associated with the cytoplasmic region of the receptors and provide tyrosine kinase activity. The binding of cytokines to cognate receptors leads to a conformational change within the receptor complex that repositions membrane-proximal, receptor-bound JAKs into an active orientation, resulting in mutual transphosphorylation that increases their activity toward tyrosine sites within the receptor. Specific phosphotyrosine (pY)–peptide motifs then act as recruitment sites for specific STAT proteins, via their Src homology 2 (SH2) domains, leading to their being phosphorylated at key tyrosine residue within a loop domain located immediately C-terminal to the SH2 domain, followed by their SH2-to-SH2 homodimerization. These activated homodimers accumulate in the nucleus, where they bind to promotor regions of many genes and activate their transcription.

1. Janus Kinases

The human genome encodes four JAKs—JAK1, JAK2, JAK3, and tyrosine kinase 2 (TYK2)—that associate selectively (Fig. 1) with different receptors (Wilks, 1989; Firmbach-Kraft et al., 1990; Wilks et al., 1991; Harpur et al., 1992). Their essential role in developmental biology is underscored by the fact that deficiency in JAK1 and JAK2 is embryonically lethal due to neurologic defects and deficiencies in erythropoiesis, respectively, whereas deficiencies in JAK3 and TYK2 are associated with a variety of severe immunodeficiency syndromes in animal models and humans (Ghoreschi et al., 2009).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Schematic illustrating the complexity of cytokine signaling. Individual cytokines bind to more than one receptor complex, which associates with more than one JAK and activates one or more STAT proteins.

JAKs have a unique architecture (Fig. 2) that is distinguishable from other protein tyrosine kinases. Traditionally, JAK structure has been described based on its distinct regions of high homology consisting of seven JAK homology (JH) domains. Recent X-ray crystal structural studies have provided a clearer delineation of JAK structural architecture, with four distinct domains: 1) N-terminal band 4.1, ezrin, radixin, moesin (FERM) domain; 2) SH2 domain; 3) the catalytically active JH1 kinase domain located at the C-terminal end; and 4) the pseudo-kinase domain (JH2), located adjacent to it, which is enzymatically inactive but has a crucial regulatory role as a repressor of kinase activity (Lupardus et al., 2014). Not surprisingly, the majority of disease-causing mutations in JAKs are located in the JH2 domain. The N-terminal FERM and SH2 domains are tightly associated to form a single module that is required for the binding of JAKs to the cytoplasmic portion of cytokine receptors. The FERM–SH2 module has been shown to mediate specific association with two receptor motifs known as Box1 and Box2. Despite having structural data, it is still not clear how specificity for different groups of receptors is achieved (Yamaoka et al., 2004; Ferrao and Lupardus, 2017).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Schematic showing the seven sequence homology-based domains of JAK (JH1 through JH7) organized into four structural domains: FERM, SH2, pseudokinase, and kinase.

Many cytokines bind to receptors composed of heterologous polypeptide chains, and several of the polypeptide chains are common or shared. In addition, a single or combination of two to three JAKs selectively associates with each receptor (Lin et al., 1995; Baker et al., 2007; Kiu and Nicholson, 2012) (Fig. 2). The consequence is that a combination of different STATs can be activated by a single cytokine, resulting in a broad range of signaling outcomes.

2. Signal Transducer and Activator of Transcription Proteins

STAT3 is a member of a family of seven proteins known to play critical roles in cytokine and growth factor signaling (Darnell, 2002). More than 30 different cytokines in various tissue and cell systems have been implicated in the canonical signaling pathways of STAT3, involving nuclear translocation and activation of transcription of its target genes (Bharadwaj et al., 2016b). Similarly, multiple growth factors also signal via activation of STAT3 (Bharadwaj et al., 2016b). Latent STAT3 is predominantly localized in the cytoplasm as homotypic N-terminal domain (NTD)-to-NTD dimers. STAT3 activation is achieved when cytokines or peptide hormones bind to cell surface receptors, orchestrating a series of events, including phosphorylation at key sites, primarily tyrosine (Y) 705,which leads to activation or rearrangement of the dimer from NTD-to-NTD to SH2-to-SH2, nuclear localization, and transcriptional activation of specific genes by binding to respective DNA elements at their promoter sites.

STAT3 is structured into seven functional domains (Fig. 3): 1) the globular N-terminal domain, 2) coil-coiled domain, 3) DNA-binding domain, 4) linker domain, 5) SH2 domain, 6) loop domain, and 7) transactivation domain, all of which play a role in STAT3 function (Shao et al., 2004). The N-terminal and SH2 domains are essential for the dimerization that, respectively, stabilizes the inactive and active conformations of STAT3. The SH2 domain is also required for docking to tyrosine-phosphorylated receptors (Vogt et al., 2011). In addition to recognizing specific DNA elements, the DNA-binding domain, in conjunction with the coil-coiled domain, participates in binding to the nuclear trafficking apparatus with both domains containing nuclear localization signals. The loop domain contains Y705, which loops into the SH2 domain of its dimerizing partner. The transactivation domain (TAD) plays a role in the recruitment of coactivators with histone acetyltransferase activity to facilitate transcriptional activation. This domain is missing in the major alternative isoform of STAT3, STAT3β, due to alternative mRNA splicing that shifts the reading frame and replaces the 55-amino-acid residue TAD with seven distinct residues that confer prolonged nuclear retention to STAT3β compared with STAT3βΔ (Shao et al., 2001).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Schematic showing amino acid residue sites within STAT3 that undergo post-translational modification.

B. Noncanonical Signal Transducer and Activator of Transcription 3 Activities

In addition to its canonical transcriptional activity, STAT3 has activities that are independent of its transcriptional role.

A. Janus Kinase/Signal Transducer and Activator of Transcription 3 in Hematopoiesis and Immune Cell Function

The JAK/STAT3 signaling pathway is well characterized in hematopoiesis and immune cells, where it contributes to proliferation, differentiation, activation, and regulation of multiple types of hematopoietic and immune cells in both the normal and stress conditions.

B. Janus Kinase/Signal Transducer and Activator of Transcription 3 and Metabolism

A. Radiation

Early experiments showed that UV light treatment of mouse embryonic fibroblasts resulted in phosphorylation of S727 in STAT1 via p38 MAPK (Kovarik et al., 1999; Ramsauer et al., 2002). STAT3 also is activated by UV through DNA damage and reactive oxygen species (Shen et al., 2001; Aziz et al., 2007a; Barry et al., 2010; Bito et al., 2010b). Although activated STAT1 seems to move cells toward apoptosis (Shen et al., 2001), STAT3 activation seems to be protective (Shen et al., 2001; Chan et al., 2004b; Sano et al., 2005; Aziz et al., 2007a; Barry et al., 2010; Bito et al., 2010a,b). Acutely, PKCε-mediated pS-STAT3 (Aziz et al., 2007a,b) and pY-STAT3 seem to work together to transcribe genes that are UV-protective, whereas persistently elevated pY-STAT3 levels mediated by chronic UV exposure are a key step in UV-induced skin cancer (Chan et al., 2004a,b; Aziz et al., 2007a,b; Barry et al., 2010). Levels of pY-STAT3 also are increased in in vitro models, as well as in patients with ionizing radiation (IR)–resistant tumors, by IR at doses used for radiotherapy (≥2 Gy), contributing to selection of IR-resistant cells in many different cancers (Ho et al., 2010; Liu et al., 2013; Pan et al., 2013; Yan et al., 2013; Gao et al., 2014; Klein et al., 2014; Lau et al., 2015; Zang et al., 2017; Lu et al., 2018). Interestingly, very low doses of radiation, e.g., 0.1 Gy, seem to reduce triple-negative breast cancer progression by deactivating pY-STAT3 through pJAK1 inhibition (Kaushik et al., 2017), leading to reduced numbers of CD44⁺/CD24⁻ breast cancer cells, reduced ability of these cells to form mammospheres, and their reduced ability to self-renew and form metastasis (Kaushik et al., 2017). Recently, the importance of pS-STAT3 in radioresistance in glioblastoma multiforme also has been suggested (Ouédraogo et al., 2016).

B. Oxidative Stress

Reactive oxygen intermediates are produced when cells are subjected to oxidative stress. Although reactive oxygen intermediates are known to contribute to DNA damage and tumorigenesis, they also function as signal mediators in many normal cellular processes. Oxidative stress is implicated in the positive modulation of the activity of a number of protein tyrosine kinases, including those that activate STAT3. Peroxide treatment leads to phosphorylation of STAT3 (pY-STAT3) and its nuclear translocation in human lymphocytes (Carballo et al., 1999), and JAK2, STAT1, and STAT3 are activated by oxidized low-density lipoprotein (Maziere et al., 2001), suggesting a role of these proteins in cellular response to oxidative stress.

STAT3 itself can be subject to direct modification by oxidative stress, as 9 of the 14 cysteine residues within STAT3 have been reported to be redox sensitive. These modifications affect the transcriptional activity of STAT3 through inhibiting Y705 phosphorylation and/or DNA binding, as well as inducing formation of higher order complexes and interfering with the function of the TAD (Li and Shaw, 2004; Xie et al., 2009; Li et al., 2010a; Kurdi et al., 2012; Butturini et al., 2014; Sobotta et al., 2015). Similarly, thiol-targeting agents also inhibited IL-6–induced STAT3 activation by S-glutathionylation in HepG2 cells (Xie et al., 2009), human endothelial cells, neonatal rat cardiomyocytes, and adult mouse cardiac myocytes (Zgheib et al., 2012) presumably through a STAT3–peroxiredoxin-2 association (Sobotta et al., 2015) and/or steric hindrance. JAK2-mediated Y705 phosphorylation also was blocked (Butturini et al., 2014) by S-glutathionylation of C328 and C542 within the DNA binding domain and Linker domain of STAT3, respectively.

Peroxide treatment induced homodimerization of STAT3 formation through a disulfide bonding of cysteines within the amino terminus of STAT3 (Li and Shaw, 2004). Cysteine residues within the DNA binding domain and the transactivation domain appear to be involved in peroxide-induced STAT3 higher-order complex formation (Sobotta et al., 2015). ROS also may differentially affect STAT3 binding to various STAT3-binding elements, suggesting oxidative stress may alter the profile of genes activated by STAT3 (Li et al., 2010a). STAT3 activation in cardiac myocytes was impaired by glutathione depletion (Kurdi et al., 2012), which could be reversed by glutathione monoethyl ester, which is cleaved intracellularly to glutathione, as well as by the reducing agent, N-acetyl-cysteine.

S-nitrosylation is another redox-related post-translational modification that inhibits STAT3 activation. In microglial cells, endogenous nitric oxide (NO) produced by inducible NO synthase (iNOS) or by treatment with s-nitrosoglutathione led to s-nitrosylation of STAT3 on C259, which inhibited JAK2-STAT3–induced gene expression and cell proliferation (Kim et al., 2014).

Increasing oxygen concentration is a mainstay for therapy of lung failure, e.g., adult respiratory distress syndrome, but hyperoxia also may contribute to the pathogenesis of acute lung injury (Altemeier and Sinclair, 2007), which is characterized by noncardiogenic pulmonary edema, inflammation, and respiratory failure. Respiratory epithelial cell–specific STAT3-KO mice progressed more rapidly than STAT3 wild-type mice to lung injury after exposure to 95% oxygen (Hokuto et al., 2004), indicating a hyperoxia-protective role for STAT3. Levels of surfactant protein B, a lipid-associated protein essential for surfactant function that is STAT3 regulated (Yan et al., 2002; Yang et al., 2004), were reduced in the bronchial lavage fluid of the respiratory epithelial cell–specific STAT3-KO mice. Correspondingly, overexpressing of STAT3C (a constitutively active form of STAT3) in respiratory epithelial cells (Lian et al., 2005) improved animal survival following hyperoxia and protected the lung from inflammation and injury through reducing mRNA levels of matrix metalloproteases, e.g., matrix metalloproteinases (MMP)9 and MMP12, and levels of adhesion molecules, which facilitate migration of neutrophils into the lung. STAT3 also is essential for the protective effects of heme oxygenase-1 in oxidant-induced lung injury (Zhang et al., 2006). IL-11 and IL-6 pretreatment protected endothelial cells from H₂O₂-induced cell death reportedly through the activation of both MAPK and STAT3 signaling pathways (Waxman et al., 2003).

Many other toxins, e.g., asbestos, silica, cigarette smoke, airborne particulate matter, diesel exhaust, and ozone, cause lung injury through ROS generation (Mossman et al., 2006), which in turn induces production of chemokines or cytokines mediating inflammation. Cigarette smoke extract (CSE) induced IL-6 expression and increased levels of pY-STAT3 and STAT3-DNA binding in human bronchial epithelial cells (HBECs) (Liu, 2007), which protected these cells from apoptosis. Anti–IL-6 neutralizing antibody, as well as STAT3 small interfering RNA (siRNA), rendered HBECs sensitive to CSE-induced DNA damage, suggesting that STAT3 mediates HBEC survival in response to CSE-induced DNA damage (Liu, 2007).

C. Genotoxic Stress

The fact that STAT3 protects from apoptosis by DNA-damage agents, e.g., radiation (Shen et al., 2001; Chan et al., 2004b; Sano et al., 2005; Aziz et al., 2007a; Barry et al., 2010; Bito et al., 2010a,b) and ROS (Hokuto et al., 2004), suggests a direct protective role STAT3 from genotoxic stress. In fact, cells lacking STAT3 are less efficient in repairing damaged DNA due to reduced activity of ataxia-telangiectasia mutated (ATM)-Chk2 and ATM and RAD3‐related (ATR)-Chk1, two major pathways involved in sensing DNA damage (Barry et al., 2010). STAT3 transcriptionally regulates mediator of DNA damage checkpoint 1, the master regulator of the ATM-chk2 pathway (Barry et al., 2010). The epidermal growth factor receptor/STAT3 pathway upregulates the endonuclease, Eme1, to reduce DNA damage after topoisomerase I inhibition (Vigneron et al., 2008). JMJD2B, a hypoxia-inducible factor 1α–regulated master regulator of DNA repair genes, which maintains histone methylation balance important for the transcriptional activation of many oncogenes, acts through activating STAT3 (Chen et al., 2014).

Some of the DNA-protective functions of STAT3 now clearly are established to be dependent on pS-STAT3. Following DNA damage induced by topoisomerase I inhibition, STAT3 is phosphorylated by cdk5 at S727, but not Y705; pS-STAT3 transcriptionally upregulates Eme1, the endonuclease involved in DNA repair, which reduces DNA damage (Courapied et al., 2010).

D. Cell Senescence

Both normal and tumor cells undergo senescence in response to various insults that cause persistent DNA damage (Kuilman et al., 2010; Rodier and Campisi, 2011; Sikora et al., 2011; Ohtani and Hara, 2013); senescence is characterized by the inability to proliferate despite sufficient nutrients and mitogens while maintaining cell viability and metabolism (Kuilman et al., 2010; Rodier and Campisi, 2011; Sikora et al., 2011; Ohtani and Hara, 2013). There are three different types of senescence: replicative senescence, oncogene-induced senescence, and stress-induced senescence. Senescence is implicated in many age-related diseases, including renal failure, type 2 diabetes mellitus, idiopathic pulmonary fibrosis, cardiovascular disease, and age-related cachexia. There is a close relationship between elevated proinflammatory factors e.g., IL-1α, IL-6, TNF, and NF-κB, which tend to increase in tissues with senescence and age (McHugh and Gil, 2018).

JAK/STAT signaling has long been linked to senescence (Kojima et al., 2013). Induction of C/EBPβ and IL-6 was shown to be required for B-raf proto-oncogene serine/threonine-protein kinase-induced senescence (Kuilman et al., 2008). Also, embryonic lung fibroblast TIG3 cells, after 55 or more passages, expressed greater levels of IL-6, IL-6Rα, and pY-STAT3 compared with TIG3 cells at earlier passages (≤33). IL-6 plus soluble IL-6Rα treatment of young TIG3 cells (passage ≤33) resulted in growth arrest and senescence-associated-β-galactosidase activity, mimicking a p53/ROS driven DNA damage response. STAT3 was similarly described to be essential for both the early ROS increase, as well as the later senescence-associated-β-galactosidase activity, through STAT3-regulated soluble factor insulin-like growth factor binding protein 5 in both human fibroblasts (Udono et al., 2012) and vascular endothelial cells (Kim et al., 2007).

IL-6 and pY-STAT3 induced by chemotherapy promoted premature senescence in lymphoma cells in a paracrine manner, which maintained a minimal residual number of chemoresistant cells that could lead to relapse. Antibody against IL-6, gp130 siRNA, or dominant-negative STAT3 enhanced chemo-induced death in lymphoma (Gilbert and Hemann, 2010), as well as colon cancer (Yun et al., 2012), indicating that targeting IL-6/STAT3 signaling may be a viable strategy for bypassing senescence-induced chemoresistance. Interestingly, senescent breast cancer cells were better able to elicit antitumor immunity through cytotoxic natural killer cells and CD4⁺ T cells following treatment with STAT3 inhibitor (Tkach et al., 2012).

E. Autophagy

Autophagy is a process used by cells to disassemble unnecessary or dysfunctional organelles that are critical for cell survival and regulated by a network of signaling pathways, including STAT3 (Kroemer et al., 2010; Pietrocola et al., 2013). STAT3 transcriptionally activates genes of several anti-autophagy–related proteins e.g., BCL2, BCL2L1, and MCL1 (Fukada et al., 1996; Bromberg et al., 1999; Bowman et al., 2000). These proteins bind competitively to BECN1 and disrupt formation of BECN1/PIK3C3 complexes, which are essential for autophagy. Induction of autophagy occurs when these protein are competitively displaced by proapoptotic BCL-2 homology 3-containing proteins (Kroemer et al., 2010). Sorafenib, a small-molecule kinase insert domain receptor inhibitor, activates autophagy in multiple liver cancer cell lines through targeting STAT3 and reducing MCL1 expression (Tai et al., 2013). STAT3 also negatively regulates both PIK3C3 (Yamada et al., 2012) and BECN1 (Miao et al., 2014) through recruitment of HDAC3 on its promoter (Miao et al., 2014).

STAT3 also promotes autophagy, e.g., leukemia inhibitory factor/STAT3 and oncostatin M/STAT3 signaling promote autophagy in mammary gland cells by upregulating both the p55 and p50 isoforms of PI3K regulatory subunit 1, PIK3R1 (Pensa et al., 2014). Furthermore, STAT3 upregulates microRNAs that target autophagy-related genes, thus promoting autophagy, e.g., STAT3 upregulates MIR17HG (miR-17-92 cluster), members of which target autophagy-related genes ULK1, BECN1, and BCL2L11 (Wu et al., 2012; Chatterjee et al., 2014; Spaccarotella et al., 2014), through a highly conserved STAT3 binding site in the promoter (Brock et al., 2009).

In contrast to nuclear STAT3, cytoplasmic STAT3 inhibits autophagy by inhibiting eukaryotic translation initiation factor 2-α kinase 2 activity, by binding to its catalytic domain, thereby inhibiting its enzymatic activity, thereby preventing formation of phosphorylated eukaryotic translation initiation factor 2-α, a known autophagy activator (Shen et al., 2012). The loss of mtSTAT3 results in a significant increase in ROS (Szczepanek et al., 2012), which is implicated in the induction of autophagy, especially the selective autophagic degradation of mitochondria, or mitophagy (Scherz-Shouval and Elazar, 2007). Mitochondrial translocation of STAT3 also suppresses autophagy during ischemia by inhibiting cytochrome c, somatic release, and preventing opening of the mitochondrial permeability transition pore (Szczepanek et al., 2011).

F. Ischemia and Reperfusion Stress

Increased production of IL-6 and granulocyte cell–stimulating factor and activation of STAT3 was found to occur early in rat lungs following resuscitation from hemorrhagic shock, a whole-body ischemia–reperfusion injury (Hierholzer et al., 1998b,c; Meng et al., 2000, 2001). Examination of hemorrhagic shock in induced iNOS-KO mice or following administration of an iNOS inhibitor demonstrated that both NF-κB and STAT3 were activated in the lungs and livers in an iNOS-dependent manner (Hierholzer et al., 1998a), and that iNOS inhibition resulted in a marked reduction of lung and liver injury. In addition, administration of a NO scavenger also resulted in reduced lung injury and reduced activation of NF-κB and STAT3 after resuscitated hemorrhagic shock (Hierholzer et al., 2002), presumably through downregulation of acute inflammation (Hierholzer et al., 2002). Importantly, resuscitation from hemorrhagic shock also resulted in apoptosis of lung type I alveolar endothelial cells (Moran et al., 2009), liver hepatocytes (Moran et al., 2008), and cardiomyocytes (Thacker et al., 2013), which resulted in increased susceptibility to pneumonia (Thacker et al., 2014) and sepsis following intraperitoneal bacterial challenge (Arikan et al., 2006). Importantly, apoptosis of these critical parenchymal cells in the lung, liver, and heart and the increased susceptibility to pneumonia and bacterial peritonitis could be reversed by administration of IL-6 at the start of resuscitation; this action of IL-6 was mediated by activation of STAT3. Thus, STAT3 augmentation may be a viable strategy to prevent hemorrhagic shock–mediated pathologies, resulting in decrease in mortality, and its timing early in resuscitation may mitigate against it, which contributes to the deleterious acute inflammatory response that ensues following parenchymal cell apoptosis.

Ischemia–reperfusion (I/R) lung injury is a significant source of morbidity following cardio-pulmonary bypass and lung transplantation (Ng et al., 2006). Under ischemic conditions, activated neutrophils infiltrate the inflamed lung and produce ROS, leading to release of lysosomal enzymes, which contribute significantly to I/R lung injury (Ng et al., 2006). Ischemia followed by 15 minutes of reperfusion in the lung induces STAT3 and NF-κB nuclear translocation (Farivar et al., 2006). In studies of lung models of in vitro anoxia–reoxygenation and in vivo I/R, carbon monoxide attenuated endothelial cell apoptosis presumably through enhanced STAT3 activation via PI3K/Akt and p38 MAPK pathways (Zhang et al., 2003, 2005). IL-6 treatment also reduced endothelial disruption and neutrophil sequestration in lung and alveolar spaces, resulting in improved oxygenation after I/R (Zhang et al., 2016).

Leukemia inhibitory factor binding to its receptor and gp130 activates Akt, STAT3, and ERK1/2 and protects cardiomyocytes from oxidative and ischemic stress by acting on mitochondria and inducing gene transcription (Zouein et al., 2013b). JAK-induced pY-STAT3 and ERK1/2-activated pS-STAT3 seem to work together to increase transcription of antiapoptotic, antioxidative stress, and proangiogenic genes (Kurdi and Booz, 2007; Zouein et al., 2013a). Studies with cardiomyocyte-targeted STAT3-KO mice reveal that although cardiac STAT3 depletion did not affect infarct size, STAT3 is crucial for cardiac preconditioning (Smith et al., 2004). Following I/R injury (Hilfiker-Kleiner et al., 2004), cardiomyocyte-specific STAT3-KO mice had greater infarct size and cardiomyocyte apoptosis presumably due to increased mRNA levels of the proapoptotic and proautophagy gene, BNIP3, and decreased mRNA levels of the prosurvival gene, heat shock protein 70, as well as decreased mRNA encoding antiapoptotic proteins (Mcl-1, Bcl-xL, c-FLIP_L, and c-FLIP_S) and cardioprotective proteins (cyclo-oxygenase-2 and heme oxygenase-1) linked to cardiac preconditioning (Bolli et al., 2011). STAT3 in cardiomyocytes also helps heart remodeling during the subacute phase of myocardial infarction (MI); STAT3 deletion during days 11 to 24 after MI resulted in the following: 1) worsened cardiac function and increased mortality, 2) exacerbated cardiac fibrosis through upregulation of fibrosis-related genes, 3) increased death of cardiac myocytes, and 4) cardiac hypertrophy, which reduced capillary density in the border zone (Enomoto et al., 2015). However, persistently activated STAT3 within cardiomyocytes resulted in a worse outcome following MI due, in part, to increased inflammation (Hilfiker-Kleiner et al., 2010). Of note, pS-STAT3 optimizes mitochondrial respiration and limits ROS formation from complex I following cardiac ischemia (Szczepanek et al., 2012).

G. Hyperosmotic Stress

Hyperosmotic stress can also activate STATs. In the slime mold Dictyostelium, hyperosmotic stress led to increased levels of pY-STAT1 without any known involvement of JAK or MAPK (Araki et al., 2003). In mammalian cells, sorbitol-induced hyperosmolarity resulted in phosphorylation of JAK1, JAK2, and TYK2 and led to activation of STAT1 and STAT3, independent of gp130, leading to formation of STAT1/STAT3 heterodimer complexes and increased DNA binding (Gatsios et al., 1998). Hyperosmotic shock in COS-7 cells led to MKK6/p38 activation and pY-STAT1 (Bode et al., 1999). Prolonged elevated glucose could also signal as a cell stressor through multiple pathways, including hyperosmolarity (McGinn et al., 2003), protein kinase C (PKC) activation (Mandarino, 1992), and oxidative damage (Nishikawa et al., 2000), likely involving pS-STAT3, as well as pY-STAT3.

A. Strategies to Target Janus Kinases

All Jakinibs in use or undergoing clinical trials to this point in time are ATP analogs that target the kinase (JH1) domain (Fig. 1). Each of these small-molecule inhibitors has been shown to bind in the ATP-binding cleft located between the β-sheet–rich N-lobe and α-helix–rich C-lobe of the JH1 domain (Alicea-Velázquez and Boggon, 2011). Although the pseudokinase (JH2) domain and FERM domain play an important role in regulating JAK activity and provide a potentially rich source of sites for targeting JAKs, there are no reports of attempts to reduce the activity of JAKs through targeting of these domains.

Excessive JAK activity has been linked to numerous disorders, including inflammatory diseases such as psoriasis and inflammatory bowel disease; myeloproliferative diseases, e.g., polycythemia vera (PV); and hematologic malignancies, including myeloid leukemia and multiple myeloma. The development of Jakinibs has focused mainly on treatment of hematologic disorders and inflammatory diseases, as outlined below.

B. Strategies to Target Signal Transducer and Activator of Transcription 3

Several approaches to reduce STAT3 activity have been proposed and include targeting of specific receptors or receptor-associated kinases, especially JAKs, upstream of STAT3, in addition to strategies to directly interfere with STAT3 activity by preventing its recruitment to receptors, dimerization, nuclear localization, and DNA binding [see Bharadwaj et al. (2016b)]. Other approaches currently underway involve modulation of autoregulators of STAT3 activity, such as strategies to activate protein phosphatases, members of the protein inhibitors of activated STAT family, or suppressor of cytokine signaling (SOCS) protein members (Heppler and Frank, 2017; Durham et al., 2019).

Substantial effort has focused on identifying a small molecule capable of blocking key residues within the STAT3 SH2 domain (Fig. 3) responsible for the binding of STAT3 to pY-peptide ligands located within ligand-activated receptors and within the loop domain of STAT3 itself. Such a molecule would inhibit the ability of STAT3 to modulate gene transcription by preventing several key steps in STAT3 activation, including its recruitment to activated receptor complexes, its phosphorylation by receptor-intrinsic or receptor-associated tyrosine kinases, its dimerization, its nuclear accumulation, and its DNA binding. Several groups have successfully identified small-molecule inhibitors using structure-based virtual ligand screens, biophysical techniques, and cell-based phenotypic screening that have focused on the STAT3 SH2 domain, either its pY-peptide–binding domain or dimerization interface. STAT3 inhibitors identified by structure-based high-throughput virtual screening include STA-21, Stattic (Stat three-inhibitory compound), S3I-201/NSC74859, BP-1-102, OPB-31121, TTI-101 (C188-9), and their analogs. Each demonstrates the ability to inhibit the growth of one or more cancer cell lines owing in some instances to their inhibition of STAT3 binding to its pY-peptide ligands and blocking IL-6–induced STAT3 activation, nuclear translocation, and transcriptional gene activation (Zhang et al., 2013c). It should, however, be noted that not all the molecules identified as STAT3 inhibitors are competitive inhibitors. Some, such as Stattic and some members of the S3I-201 family of compounds, inhibit STAT3 by covalently reacting with residues (largely cysteines) at key locations within STAT3, as well as other off-target proteins (Heidelberger et al., 2013; Ball et al., 2016).

Studies investigating the impact of disease-causing mutations on STAT3 suggest that there are long-range communications between domains of STAT3, implying that STAT3 activity may be subject to allosteric modulation (Zhang et al., 2000; Mertens et al., 2015; Namanja et al., 2016). Thus, in addition to orthosteric inhibitors, emerging evidence points to allostery as a mechanism of STAT3 inhibition observed in a number of small-molecule inhibitors. For example, hydrogen–deuterium exchange experiments and other studies show that STAT3 inhibitors—SF-1-066, BP-1-102, and MM206—may, in addition to binding to the SH2 domain, affect the function of the SH2 domain by binding to and perturbing the structure of domains of STAT3 at a distance from the SH2 domain, suggesting an allosteric destabilizing effect on the SH2 domain (Resetca et al., 2014; Minus et al., 2015). These findings suggest the possible utility of developing allosteric inhibitors of STAT3. Of note, compounds acting allosterically may be more specific as they tend to bind to less conserved regions of target proteins; this could be of potential benefit when attempting to specifically target a protein such as STAT3 that has a SH2 domain that is structurally very similar to other members of the STAT protein family. However, it is important to note that the toxicity of some STAT3 inhibitory compounds, as well as their antitumor activity, may arise from their ability to cause marked changes in STAT3 structure, leading to its aggregation within critical organelles. For instance, it has been reported that STAT3 forms proteotoxic aggregates in mitochondria when tumor cells are treated with OPB-51602, STA21, or WP1066 (Genini et al., 2017).

A number of inhibitors have been developed that appear to directly target the DNA-binding domain of STAT3. Nucleic acid–based drugs, including decoy oligonucleotides and, perhaps, G-quartet oligonucleotides, have been identified that inhibit STAT3-mediated gene expression and inhibit growth of different cancer cell lines (Leong et al., 2003; Lewis et al., 2008; Sun et al., 2008; Shen et al., 2009; Zhang et al., 2013b; Liu et al., 2014; Sen et al., 2014). Platinum IV compounds—CPA-1, CPA-7, IS3 295, and platinum (IV) tetrachloride—also can inhibit STAT3 DNA-binding activity and inhibit cancer cell line growth (Turkson et al., 2004). Other small-molecule inhibitors besides platinum-based compounds also have been identified, including InS3-54 and HO-3867, which specifically target the DNA-binding domain of STAT3, inhibit its binding to DNA, and block STAT3-mediated gene activation (Rath et al., 2014).

Strategies targeting STAT3 mRNA have been employed to inhibit STAT3 protein expression and thereby all of its activities. Approaches based on siRNA and short hairpin RNA targeting of STAT3 activity have resulted in the development of a naked antisense oligonucleotide–based drug, AZD9150 (IONIS-STAT3-2.5_Rx or danvatirsen), which has entered Phase II clinical trials in cancer patients (see below).

C. Janus Kinase/Signal Transducer and Activator of Transcription 3 Signaling and Its Targeting in Asthma

Asthma has traditionally been associated with overactivation of T helper (Th)2 cells, infiltration of the lung with eosinophils (Wenzel, 2006; Fahy, 2009; Lin et al., 2013), and good responses to treatment with corticosteroids. However, higher morbidity and mortality are seen in approximately 10% of asthma patients who are refractory to corticosteroid treatment. These patients show a Th17 cell–driven phenotype and neutrophilic infiltration of the lung (American Thoracic Society, 2000; McKinley et al., 2008; Al-Ramli et al., 2009). STAT3 is known to regulate both Th2 cell recruitment during allergic inflammation and the production of Th17 cells and the expansion and recruitment of neutrophils (Chen et al., 2007; Ma et al., 2008; Milner et al., 2008). Targeting JAK/STAT3 signaling may be of benefit in the treatment of neutrophilic, steroid-resistant asthma. An early example of this approach involved the use of STAT1/3 decoy oligonucleotides in a mouse model of neutrophil-predominant asthma due to intranasal instillation of house dust mite (HDM) extract. Administration of STAT1/3 decoy oligonucleotides or the Jakinib, Tyr1A, was shown to mitigate airway inflammation and airway hyper-responsiveness in lungs of mice challenged with HDM (Simeone-Penney, 2008). Pharmacological blockade of STAT3 in HDM-challenged mice with TTI-101 (C188-9) inhibited allergen-induced STAT3 phosphorylation, resulting in concomitant reduction in airway inflammation and remodeling. These changes were accompanied by normalization of IL-4, IL-5, IL-13, and IL-17A cytokine levels and reduced expansion of Th2 cells and Th17 cells (Gavino et al., 2016). The data summarized above suggest that targeting JAK and STAT3 with small-molecule inhibitors may provide a superior anti-asthma treatment approach compared with standard therapies as, they would target both Th2- and Th17-associated asthmas (Choy et al., 2015).

D. Janus Kinase/Signal Transducer and Activator of Transcription Signaling and Its Targeting in Anaphylaxis

The Jakinib, ruxolitinib, was shown to reduce food allergy manifestations in mice through prevention of mast cell hyperplasia and inhibition of mast cell activation (Yamaki and Yoshino, 2014). Similar to findings in mut-Stat3 mice, Hox et al. (2016) found that hematocrit levels and local plasma transudate in response to the IgE/antigen anaphylaxis model and to histamine alone were significantly lower in TTI-101–treated mice compared with untreated mice. These data were supported by a decreased permeability induced by histamine and other vascular permeability–inducing agents in normal human umbilical vein endothelial cells treated in vitro with TTI-101. Together with the findings in the mut-Stat3 mice described above, these results suggest that STAT3 inhibition interferes with a common pathway regulating vascular permeability and raises the possibility of using JAK/STAT3 inhibition to prevent anaphylaxis in those suffering from severe allergies to peanuts or other allergens that are low-level contaminates of many foods.

E. Janus Kinase/Signal Transducer and Activator of Transcription 3 Signaling and Its Targeting in Inflammatory Bowel Disease

Ulcerative colitis (UC) and Crohn’s disease (CD) are the major forms of inflammatory bowel disease (IBD). Both are idiopathic chronic inflammatory diseases of the gastrointestinal (GI) tract that cause substantial morbidity, occasional mortality, and a 20- to 30-fold increased risk of developing colorectal cancer (CRC) (Kasembeli et al., 2018). UC and CD are distinct in the location of inflammation within the GI tract—the rectum and distal colon are typically involved in UC, whereas any part of the gut can be involved in CD—and the depth of tissue involvement (Kasembeli et al., 2018). Population-based, genome-wide–association studies followed by subsequent meta-analysis of population-based, genome-wide–association studies and immunochip data have identified more than 160 loci linked to IBD susceptibility (Jostins et al., 2012). Among these genes are included some related to intestinal mucosal immune responses, including STAT3 (Lees et al., 2011; Kasembeli et al., 2018). In addition, a number of cytokines found to be elevated in the plasma and tissue of patients with IBD are known to activate STAT3 and/or its downstream targets, including IL-1β, IL-2, IL-6, IL-12, IL-15, IL-21, IL-23, IL-17, IL-18, IL-10, IL-11, IL-22, IFN-α/β/γ and MMP9, TNF-α, and IFN-γ (Kasembeli et al., 2018). Increased levels of STAT3 protein have been detected in affected tissues in both UC and CD, which correlated with increased levels of pY-STAT3 and the histologic degree of inflammation (Mudter et al., 2005; Musso et al., 2005; Kasembeli et al., 2018); levels of pY-STAT3 were observed in both T cells and colonic epithelial cells. Studies undertaken to tease out the role played by CD4⁺ cells in IBD determined that activation of STAT3 in pathogenic T cells contributes to colitis by increasing their survival (Kasembeli et al., 2018). In contrast, genetic deletion of STAT3 in innate immune cells, such as myeloid cells (neutrophils and macrophages), and enterocytes resulted in chronic colitis (Takeda et al., 1999; Atreya et al., 2000; Yamamoto et al., 2000; Suzuki et al., 2001; Pickert et al., 2009).

Two isoforms of STAT3 (α, β), derived from one gene by alternative mRNA splicing, are expressed in most cells in a 4:1 ratio (α:β). STAT3α is proinflammatory and antiapoptotic, whereas STAT3β has opposing effects on STAT3α. Manifestations of dextran sulfate sodium–induced colitis were more severe in STAT3^Δβ/Δβ transgenic mice, which express only STAT3α (Maritano et al., 2004; Marino et al., 2014), indicating that a STAT3 activity globally contributes to dextran sulfate sodium–induced colitis.

Humanized anti–IL-6R monoclonal antibody (tocilizumab), anti–IL-6R monoclonal antibody, anti–IL-6 avimer antibody (C326), and small-molecule IL-6R antagonist (TJ301) have shown beneficial effects in humans with CD (Ito et al., 2004; Kasembeli et al., 2018). Anti-TNF antibodies that disrupt the interaction of TNF-α with its receptor, including infliximab (a mouse–human chimeric anti–TNF-α antibody), adalimumab (a human monoclonal anti–TNF-α antibody), and certolizumab pegol (recombinant humanized anti–TNF-α antibody), appear to be beneficial in CD patients (Targan et al., 1997; Hanauer et al., 2002, 2006; Colombel et al., 2007; Schreiber et al., 2007; Sandborn et al., 2011). Vidofludimus, an immunosuppressive drug that inhibits dihydroorotate dehydrogenase involved in pyrimidine nucleoside biosynthesis, inhibits the proliferation of T cells and B cells, and the generation of IL-17A, IL-17F, and IFN-γ has shown to have clinical benefit in an open-label uncontrolled entrance study of patients with IBD (Herrlinger et al., 2013). Ustekinumab, a human monoclonal antibody that binds to the p40 subunit of both IL-12 and IL-23, causing disruption of their binding to receptor, induced a clinical response within 4 to 6 weeks of start of treatment in two clinical trials involving CD patients (Sandborn et al., 2008; Bruce, 2016). Also, antibodies against IL-12 and IL-23 have been evaluated in early Phase II studies, and these agents have demonstrated efficacy in moderate-to-severe CD patients who had previously failed infliximab (Sandborn et al., 2008; Bruce, 2016).

A selective, small-molecule inhibitor of JAK3, Janex-1, has shown promise as a therapeutic option for UC treatment (Durant et al., 2010; Anderson et al., 2011; De Vries et al., 2017). Furthermore, Phase II and III clinical studies of tofacitinib, an oral JAK3 inhibitor, have been Food and Drug Administration approved as the only nonsteroidal oral treatment that induces remission in moderate-to-severe UC (Sandborn et al., 2012; De Vries et al., 2017). In addition, filgotinib and upadacitinib (ABT-494, AbbVie), which selectively target JAK1, have demonstrated benefit in Phase II clinical studies in CD patients (https://clinicaltrials.gov/ct2/show/NCT02365649; Ananthakrishnan, 2017; De Vries et al., 2017).

F. Janus Kinase/Signal Transducer and Activator of Transcription 3 Signaling and Its Targeting in Cachexia

Proinflammatory cytokines, such as TNF-α, IL-6, and IFN-γ, drive catabolic signaling associated with cachexia (Mantovani et al., 2000; Miyamoto et al., 2016; Porporato, 2016). The link between IL-6/STAT3 signaling and muscle and adipose tissue wasting is well established in mouse models of cachexia (White et al., 2011; Silva et al., 2015) and has been reinforced preliminarily in clinical studies. A Phase II trial of the JAK1/2 inhibitor, ruxolitinib, has been shown to increase body weight, in patients with primary myelofibrosis (PMF) and metastatic pancreatic adenocarcinoma (Jatoi, 2013; Cerchione et al., 2017). In patients with myelofibrosis, changes in body weight were shown to closely correlate to changes in plasma levels of leptin and erythropoietin, both of which have been linked to metabolic regulation through STAT3 signaling (Mesa et al., 2015; Dey et al., 2016).

Recent studies have highlighted the direct role of JAK/STAT3 signaling in protein catabolism through muscle-specific proteolytic pathways. STAT3 directly binds to the promoter of caspase-3 in muscle cells, increasing levels of caspase-3 protein; caspase-3 cleaves actinomyosin and myofibrillar proteins, which triggers protein degradation via the ubiquitin/proteasome system. Another proteolytic pathway activated by JAK/STAT3 signaling involves pY-STAT3–mediated transcriptional activation of C/EBPδ, which, in turn, upregulates myostatin, leading to increased expression of the muscle-specific E3 ligases, MAFbx and Atrogin-1, and protein degradation via the ubiquitin/proteasome system. Emerging evidence suggests that STAT3 may also exert its procachexic effects in collaboration with other signaling networks; for instance, its activation has been shown to modulate IFN-γ/TNF-α–induced muscle wasting via the iNOS/NO pathway (Ma et al., 2017).

Accelerated loss of white adipose tissue is a salient characteristic of cachexia. IL-6 signaling through STAT3 has been shown to control this process, as evidenced by inhibition of white adipose tissue lipolysis in mice treated with anti–IL-6R antibody (Babaei et al., 2018; Wang et al., 2018a). Additionally, STAT3 is thought to be involved in cachexia-related adipose tissue breakdown by modulating enzymes involved in lipolysis. Results from experiments with bovine adipocytes treated with leptin suggest that there may be a link among lipolysis, adipose triglyceride lipase induction, and STAT3 signaling (Koltes et al., 2017). These observations are further strengthened by the results showing that STAT3 KO in adipose tissues of mice leads to obesity due to hypertrophy of adipose tissue (Cernkovich et al., 2008; Xu et al., 2013).

Levels of pY-STAT3 are increased in skeletal muscle in mouse models of cachexia, including models of chronic kidney disease (Zhang et al., 2013a) and cancer (Silva et al., 2015). Inhibition of STAT3 with TTI-101 reversed muscle loss in both cachexia models by suppressing the expression of MAFbx/Atrogin-1 and MuRF1 and reducing caspase-3 activity (Silva et al., 2015). Recent work by Guo et al. (2017) corroborates the findings of Zhang et al. (2013a) by demonstrating that pantoprazole inhibits JAK2/STAT3 signaling, which leads to reduced expression of MuRF1 and Fbx32 proteins, thereby mitigating the progression of tumor-associated muscle atrophy and protein degradation.

G. Janus Kinase/Signal Transducer and Activator of Transcription 3 Signaling and Its Targeting in Fibrosis

STAT3 plays a major role in many of the cellular processes that drive the development and progression of fibrosis, including fibroblast and myofibroblast cell proliferation, cell survival, induction of angiogenesis, and cell plasticity (Kasembeli et al., 2018). Fibrotic tissue samples from patients often show elevated levels of pY-STAT3 (Ogata et al., 2006). STAT3 is known to act both independently and in conjunction with other signaling networks such as TGF-β1, which is a well-known master regulator of fibrosis (Chakraborty et al., 2017). In fact, there is ample direct and indirect evidence indicating significant crosstalk between TGF-β1 signaling and STAT3. Studies have shown that TGF-β1 signaling can induce STAT3-dependent fibrotic responses by phosphorylation and activation of JAK2 (Zhang et al., 2015).

Genetic and pharmacological inhibition of STAT3 signaling has been shown to suppress fibrosis in many animal models, including models of lung fibrosis, in which haploinsufficiency of STAT3 eliminated the increase in lung collagen (hydroxproline) content in bleomycin-treated mice (O’Donoghue et al., 2012). pY-STAT3 activates transcription programs of various cellular processes to drive the development and progression of fibrosis. These include the following: 1) production of extracellular matrix (ECM) by upregulating collagen type I expression; 2) transcriptional control of MMPs, as well as their cognate inhibitor proteins—tissue inhibitors of MMPs—that also are necessary for the maintenance of ECM (Sun et al., 2008; Wang et al., 2011b; Giannandrea and Parks, 2014; Matsui et al., 2017); 3) resistance to apoptosis of ECM-producing cells, which slows resolution of the wound-healing process due to myofibroblast persistence (Kuhn and McDonald, 1991; Katzenstein and Myers, 1998; Habiel and Hogaboam, 2014; Xu et al., 2014a; Milara et al., 2018); and 4) contribution to fibroblast plasticity—a defining hallmark of fibrosis (Lamouille et al., 2014; Chakraborty et al., 2017; Liu et al., 2017; Pedroza et al., 2017).

Results from preclinical and clinical studies have confirmed the essential role of JAK/STAT3 signaling in the pathogenesis of fibrosis and generated considerable interest directed at targeting this pathway to develop novel treatments. Indeed, several Jakinibs are undergoing clinical testing in patients with fibrosis (Kasembeli et al., 2018), whereas preclinical studies examining small-molecule STAT3 inhibitors have shown promising results. Both TTI-101 and S3I-201 have been tested in several mouse models of fibrosis, including bleomycin-induced lung and skin fibrosis (scleroderma) and genetically-engineered models of scleroderma, as well as a model mouse of interstitial kidney fibrosis induced by unilateral ureteral obstruction, in which inhibition of STAT3 phosphorylation by these compounds blocks the development of organ fibrosis by reducing expression levels of proteins that mediate fibrosis and impairing myofibroblast differentiation (Pang et al., 2010; Pedroza et al., 2016, 2017; Chakraborty et al., 2017). In addition, cucurbitacin-B attenuates carbon tetrachloride–induced liver fibrosis, which was accompanied by reduced levels of pY-STAT3, along with reduced levels of hydroxyproline, collagen-1α, α-smooth muscle actin, and TGF-β in liver tissue (Sallam et al., 2018). Others have shown that sorafenib, and its derivative SC-1, can ameliorate liver fibrosis through STAT3 inhibition in hepatic stellate cells. In the thioacetamide liver fibrosis mouse model, administration of either sorafenib or SC-1 resulted in activation of SHP-1, which reduced levels of pY-STAT3 (Su et al., 2015). We demonstrated that TTI-101 administration to hepatocyte-specific Pten KO (HepPten-) mice, a model of nonalcoholic steatohepatitis (NASH) progressing to cirrhosis, reduced liver injury and fibrosis (Jung et al., 2017). S3I-201 also has shown activity in mouse preclinical models of carbon tetrachloride–induced fibrosis (Wang et al., 2018b). Zerumbone, a cyclic sesquiterpene derived from rhizomes of wild ginger, inhibits STAT3 phosphorylation and decreases mRNA expression of fibrotic genes such as α-SMA and TGF-β1, but not TGF-β2, in alkali-injured corneas of mice. Treatment with zerumbone reduced corneal neovascularization and fibrosis, thus promoting wound healing (Kim et al., 2017a).

Myeloproliferative neoplasms (MPNs) are a group of Philadelphia chromosome-negative chronic hematologic disorders that include PV, essential thrombocythemia (ET), and PMF. Dysregulation of bone marrow proinflammatory cytokine production leads to chronically elevated levels of circulating inflammatory cytokines, which drives clonal evolution in MPNs and leads to increased deposition of reticulin and collagen-containing bone marrow stromal fibers (Tefferi et al., 2011). Bone marrow fibrosis is a shared feature of all MPNs, but is the defining histopathological feature of PMF, a major diagnostic criterion (Zahr et al., 2016), and has been associated with poor outcomes (Kuter et al., 2007; Byrne et al., 2018).

Soluble factors associated with megakaryocytes and platelets known to signal through STAT3 have been implicated in the modulation of stromal cells and bone marrow fibrosis (Agarwal et al., 2016; Zahr et al., 2016). In addition, the JAK2 gain-of-function mutation (V617F) has been suggested to play a prominent role in MPNs (Baxter et al., 2005; Levine et al., 2005), being present in in ∼95% of PV, 50% to 60% of ET, and ∼50% of all PMF patients and associated with increased risk of mortality and splenomegaly (Verstovsek, 2010). JAK2 signaling through STAT5, which is constitutively active in JAK2 (V617F) mutant-positive cells, seems to mediate PV, as deletion of either STAT5A or STAT5B genes in JAK2V617F mice prevented the development of PV (Walz et al., 2012).

Currently, the only curative treatment option for PMF is allogenic hematopoietic cell transplantation, but long-term survival is usually hampered by disease relapse and/or graft-versus-host disease, leaving the unmet need to develop targeted therapies for PMF, e.g., JAK/STAT pathway inhibition. Ruxolitinib, a JAK1/JAK2 inhibitor, has demonstrated promising activity in PMF patients, showing increased benefit across several end points, e.g., reduction in splenomegaly, improvement in disease-related symptoms, and increased overall survival, when compared with placebo or best available therapy (Harrison et al., 2012; Verstovsek et al., 2012; Vannucchi et al., 2015). Pacritinib (a JAK2/FLT3 inhibitor) was even more effective than ruxolitinib in reducing splenomegaly and constitutional symptoms (Mascarenhas et al., 2018). Mometinib (a Jak1/2 inhibitor) and Itanicibib (a JAK1 inhibitor) also showed varying degrees of efficacy (Byrne et al., 2018). There is no clear correlation between the presence of a JAK2 V617F mutation and bone marrow fibrosis in MPN. However, when compared with best available therapy, Ruxolitinib and other Jakinibs appeared to reduce the odds of worsening bone marrow fibrosis and greater reduction in spleen size (Kvasnicka et al., 2018), suggesting that long-term JAK inhibition may be disease modifying.

It is important to note, however, that whereas JAK-STAT inhibitor treatment leads to improvement in constitutional symptoms and reduced splenomegaly, they show only modest benefit in reduction of clonal burden (Deininger et al., 2015), suggesting that the bulk of the observed clinical benefit of these inhibitors may arise from modulation of JAK-STAT signaling not associated with the expanded mutant clone(s). In fact, inhibition of JAK signaling in both wild-type and mutant cells was shown to be necessary to achieve improved clinical benefit in MPNs (Kleppe et al., 2015), and abnormal cytokine production continues to persist in plasma of many PMF patients after JAK2 inhibition with ruxolitinib (Fisher et al., 2019). Also, there is a word of caution regarding the use of Jakinibs in PMF as they may increase the risk of secondary malignancy, e.g., lymphoma (Porpaczy et al., 2018).

The BRD4/BET/NF-κB signaling pathway also has been suggested to contribute to MPN, as the addition of BET inhibitors to Jakinibs has been shown to markedly reduce serum levels of inflammatory cytokines, accompanied by reduction in disease burden, and reversal of bone marrow fibrosis in preclinical models (Kleppe et al., 2018). In addition, simultaneous inhibition of STAT5 phosphorylation on tyrosine and serine residues had synergistic effects against JAK2V617F mutant-positive cell lines, suggesting that approaches that target both post-translational modifications of STAT5 may be of benefit to patients with PMF (Bartalucci et al., 2017).

H. Janus Kinase/Signal Transducer and Activator of Transcription 3 Signaling and Its Targeting in Metabolic Disorders

Hepatic steatosis is an accumulation of fat in hepatocytes that can arise as a complication of obesity, type 2 diabetes mellitus, and alcohol intoxication. Treatment of mice in a murine model of chronic-binge ethanol feeding with STAT3-inducing cytokines, e.g., IL-6/IL-22, ameliorates fat accumulation in hepatocytes (Wang et al., 2004a; Ki et al., 2010). In morbidly obese individuals with congenital leptin deficiency, leptin replacement dramatically decreases insulin resistance, steatosis, dyslipidemia, and hyperglycemia (Farooqi et al., 2002; Licinio et al., 2004). Moreover, hepatocyte-specific gp130- or STAT3-KO mice exhibit exacerbated steatosis induced by a choline-deficient, methionine-supplemented diet; alcohol-containing diet; or a high-fat diet (Kroy et al., 2010). The antisteatogenic effect of hepatic STAT3 is suggested to be through direct inhibition of lipogenic sterol-regulatory element-binding protein-1 (a master transcription factor that controls lipid synthesis), acetyl-coenzyme A carboxylase, and fatty acid synthase, and stimulation of fatty acid oxidation genes, adenosine monophosphate-activated protein kinase, and carnitine palmitoyltransferase 1 (Inoue et al., 2004; Ueki et al., 2004; Horiguchi et al., 2008; Kroy et al., 2010). Inhibiting SOCS3 also increases pY-STAT3 levels, which ameliorates hepatic steatosis by downregulating sterol-regulatory element-binding protein-1 (Ueki et al., 2004). Interestingly, most of these mice models recapitulated alcoholic liver disease (ALD), which makes one speculate (Miller et al., 2011) that, probably in ALD, activation of STAT3 reduces hepatic steatosis. In fact, a current paper, which analyzed liver inflammatory and immune responses at early stages of alcoholic liver disease in a cohort of alcohol-dependent patients undergoing a highly standardized alcohol withdrawal program, concluded that in early stages of ALD, deficient IL-6/STAT3 signaling, high TLR7, and type I IFN production were risk factors for developing severe liver damage and fibrosis in these patients (Stärkel et al., 2019).

Recently, it was shown that an oxidative hepatic environment in obesity inactivates the STAT1 and STAT3 phosphatase T cell protein tyrosine phosphatase (TCPTP) and increases STAT1 and STAT3 signaling (Grohmann et al., 2018). Genetic deletion of TCPTP in hepatocytes also promoted T cell recruitment, NASH, and fibrosis, as well as hepatocellular carcinoma (HCC) in obese C57BL/6 mice that normally do not develop NASH, fibrosis, or HCC (Grohmann et al., 2018). Attenuating STAT1 signaling prevented T cell recruitment, NASH, and fibrosis, but did not prevent HCC. In contrast, correcting STAT3 signaling prevented HCC without affecting NASH and fibrosis. However, other studies have clearly established a role of STAT3 in non-alcoholic fatty liver disease. Pathogenesis of non-alcoholic fatty liver disease proceeds through lipid accumulation in hepatocytes that triggers a series of cytotoxic events that subsequently induces numerous pathologic processes, including insulin resistance, leptin deficiency, oxidative stress, fat accumulation, and liver tissue inflammation (Lonardo et al., 2016). STAT3, Krueppel-like factor 6, PPARs and methylenetetrahydrofolate reductase, and various proinflammatory cytokines are involved in the above processes leading to NASH (Sookoian et al., 2008). Activated IL-6/STAT3 in non-alcoholic fatty liver disease is associated with increased insulin resistance (Min et al., 2015). In a recent study (Belloni et al., 2018) targeting a phospho-STAT3/miR21 axis using either the STAT3 inhibitor S3I-201 or metformin, reduced STAT3 activation and improved vesicular hepatic steatosis in an in vitro fatty dHepaRG model (hepatic differentiated HepaRG cells treated with sodium oleate, S3I-201) as well as an in vivo high-fat diet induced age-dependent liver steatosis model (metformin).

In a hepatocyte-specific Pten KO mouse model of NASH-mediated HCC, TTI-101 treatment blocked hepatocellular carcinoma tumor development, through reducing liver steatosis, inflammation, and bile ductular reactions, thereby improving all pathologic lesions of NASH. It also greatly reduced liver injury, as measured by serum aspartate aminotransferase and alanine transaminase levels, as well as liver fibrosis, as indicated above. The reason for TTI-101 being able to reduce symptoms of both NASH/fibrosis and HCC might stem from the fact that TTI-101 treatment of HepPten⁻ mice resulted in inhibition of both STAT1 and STAT3 pathways (Jung et al., 2017). Similarly, STAT3 inhibition in adipose tissue and, to a lesser extent, in hepatic tissues (Hua et al., 2018) was how known antilipemic agents (PPARγ agonists, gemfibrozil, and fenofibrate) attenuated obesity, high blood pressure, and insulin resistance in a PPARα-null mouse model of metabolic syndrome.

STAT3 also has been reported to play a definite detrimental role in development of insulin resistance. Prolonged IL-6 exposure leads to insulin resistance and glucose intolerance in human adipocytes, hepatocytes (Gurzov et al., 2016), and skeletal muscles. Blocking IL-6 in diet-induced obese mice represses hepatic inflammation via inhibition of the IL-6/JAK2/STAT3 pathway (Park et al., 2010), thereby increasing insulin sensitivity and resultantly decreasing diet-induced obesity (Priceman et al., 2013). TCPTP in the liver dephosphorylates the insulin receptor, JAK1, and STAT3 and negatively regulates both IFN regulatory factor/PI3K and STAT3 signaling pathways, thereby modulating expression of gluconeogenic genes and hepatic glucose output (Fukushima et al., 2010). Moreover, STAT3 silencing improves insulin sensitivity in IL-6–treated myotubes and prevents development of lipid-induced insulin resistance (Mashili et al., 2013; Gurzov et al., 2016). IL-6/STAT3 signaling in muscle during intense exercise, in contrast, plays a beneficial role in glucose homeostasis (Wunderlich et al., 2013). Thus, it seems, whereas transient IL-6/STAT3 seems to be helping maintain glucose homeostasis, chronic IL-6/STAT3 signaling impairs insulin action, leading to insulin resistance and inflammation-associated glucose intolerance, probably through an increase in basal SOCS3 levels in generally insulin-sensitive adipose tissue, liver, and skeletal muscles, by mechanism (Yang et al., 2012) similar to leptin/STAT3/SOCS3 axis in the CNS (Wunderlich et al., 2013).

IL-6/JAK/STAT3-induced SOCS3 impairs insulin action by binding to IRS-1 and IRS-2, resulting in their ubiquitination and degradation (Shi et al., 2004). Recent studies have reported that STAT3 is required for pancreatic β cell reprogramming and regeneration, although the downstream targets are unknown (Gurzov et al., 2016). Muscle-specific deletion of SOCS3 (Jorgensen et al., 2013) or STAT3-siRNA (Kim et al., 2017b) protects against the development of hyperinsulinemia and insulin resistance due to elevated IRS-1 and Akt phosphorylation that contribute to increased glucose uptake. Interestingly, hepatocyte-specific SOCS3–KO mice have improved insulin sensitivity resulting from increased pY-IRS-1, but develop systemic insulin resistance and obesity over a long period due to the hyperactivity of STAT3 and resultant chronic inflammation (Torisu et al., 2007), indicating a non-SOCS3–mediated mechanism of chronic STAT3-mediated insulin resistance.

Chronic inflammation has been known to accelerate insulin resistance and other metabolic abnormalities associated with obesity (Kalupahana et al., 2012; Wensveen et al., 2015). Obesity-induced inflammation also drives an influx of immune cells, including B cells and CD4⁺ and CD8⁺ T cells, leading to further release of proinflammatory cytokines, such as IL-6, from recruited inflammatory cells, such as macrophages (Pradhan et al., 2001; Hu et al., 2004; Wunderlich et al., 2013), which exacerbates unintended immune responses contributing to the pathology (Wensveen et al., 2015). Thus, the IL-6/JAK2/STAT3 signaling pathway is a valid target to treat NASH, non-alcoholic fatty liver disease, and insulin resistance, especially in the setting of obesity and type 2 diabetes.

I. Cancer