Targeting the IL-6/JAK/STAT3 signalling axis in cancer

Abstract

The IL-6/JAK/STAT3 pathway is aberrantly hyperactivated in many types of cancer, and such hyperactivation is generally associated with a poor clinical prognosis. In the tumour microenvironment, IL-6/JAK/STAT3 signalling acts to drive the proliferation, survival, invasiveness, and metastasis of tumour cells, while strongly suppressing the antitumour immune response. Thus, treatments that target the IL-6/JAK/STAT3 pathway in patients with cancer are poised to provide therapeutic benefit by directly inhibiting tumour cell growth and by stimulating antitumour immunity. Agents targeting IL-6, the IL-6 receptor, or JAKs have already received FDA approval for the treatment of inflammatory conditions or myeloproliferative neoplasms and for the management of certain adverse effects of chimeric antigen receptor T cells, and are being further evaluated in patients with haematopoietic malignancies and in those with solid tumours. Novel inhibitors of the IL-6/JAK/STAT3 pathway, including STAT3-selective inhibitors, are currently in development. Herein, we review the role of IL-6/JAK/STAT3 signalling in the tumour microenvironment and the status of preclinical and clinical investigations of agents targeting this pathway. We also discuss the potential of combining IL-6/JAK/STAT3 inhibitors with currently approved therapeutic agents directed against immune-checkpoint inhibitors.

The IL-6/JAK/STAT3 pathway has a key role in the growth and development of many human cancers. Elevated levels of IL-6 are observed in chronic inflammatory conditions, such as rheumatoid arthritis and inflammatory bowel disease, and in a large number of patients with haematopoietic malignancies or solid tumours¹. In the pathogenesis of cancer, elevated levels of IL-6 stimulate hyperactivation of JAK/STAT3 signalling, which is often associated with poor patient outcomes^2–5. Furthermore, the genes encoding JAK enzymes, particularly JAK2, are frequently mutated in myeloproliferative neoplasms, leading to constitutive activation of JAK/STAT3 signalling. Hyperactivation of STAT3 signalling occurs in the majority of human cancers and also correlates with a poor prognosis. STAT3 hyperactivation in tumour cells can occur as a result of elevated IL-6 levels in the serum and/or in the tumour microenvironment, owing to signals from other growth factors and/or their receptors, activation by non-receptor tyrosine kinases (such as SRC and BCR–ABL1), or loss-of-function mutations affecting negative regulators of STAT3. These negative regulators include members of the protein inhibitor of activated STAT (PIAS) and suppressor of cytokine signalling (SOCS) families as well as several cellular phosphatases (tyrosine-protein phosphatase non-receptor type 6 (SHP1; also known as PTPN6), tyrosine-protein phosphatase non-receptor type 11 (SHP2), dual specificity protein phosphatase 22 (DUSP22), receptor-type tyrosine-protein phosphatase-δ (PTPRD), receptor-type tyrosine-protein phosphatase T (PTPRT), tyrosine-protein phosphatase non-receptor type 2 (PTPN2) and tyrosine-protein phosphatase non-receptor type 1 (PTPN1))^6–11. Aberrant expression of microRNAs (miRNAs) that regulate STAT3 expression can also contribute to elevated STAT3 activity in tumours.

IL-6 is produced by multiple cell types located within the tumour microenvironment, including tumour-infiltrating immune cells, stromal cells, and the tumour cells themselves^1,12–15. IL-6 acts directly on tumour cells to induce the expression of STAT3 target genes, which encode proteins that then drive tumour proliferation (such as cyclin D1) and/or survival (such as BCL2-like protein 1 (BCL-x_L)). The ability of STAT3 to promote IL6 gene expression then results in a feedforward autocrine feedback loop¹⁶. STAT3 also induces the expression of factors that promote angiogenesis, such as VEGF; invasiveness and/or metastasis, such as matrix metalloproteinases (MMPs); and immunosuppression, such as IL-10 and TGFβ (in addition to VEGF and IL-6)^14,17,18.

In addition to direct effects on tumour cells, IL-6 and JAK/STAT3 signalling can have a profound effect on tumour-infiltrating immune cells. STAT3 is often hyperactivated in tumour-infiltrating immune cells and exerts negative regulatory effects on neutrophils, natural killer (NK) cells, effector T cells, and dendritic cells (DCs), suggesting that STAT3 activation in immune cells likely leads to downmodulation of antitumour immunity^19–29. At the same time, STAT3 positively regulates regulatory T (T_reg) cells and myeloid-derived suppressor cell (MDSC) populations^17,19. Collectively, these effects contribute to a highly immunosuppressive tumour microenvironment.

The understanding that IL-6/JAK/STAT3 signalling promotes tumour growth and progression while severely hindering antitumour immunity has stimulated the search for clinical agents that can effectively inhibit this pathway. Siltuximab and tocilizumab are antibodies that target IL-6 and the IL-6 receptor-α (subsequently referred to as IL-6R), respectively, and have been approved by the FDA for the treatment of multicentric Castleman disease (siltuximab), arthritis (tocilizumab), and chimeric antigen receptor (CAR) T cell-induced cytokine-release syndrome (tocilizumab). Similarly, tofacitinib is a small-molecule tyrosine kinase inhibitor that primarily targets JAK1 and JAK3 and has been approved by the FDA for the treatment of arthritis, whereas ruxolitinib is a small-molecule inhibitor of JAK1 and JAK2 and is approved for use in patients with myelofibrosis or polycythaemia vera. Clinical evaluations of these agents in patients with haematopoietic or solid tumours are currently ongoing. Moreover, a large number of novel IL-6, IL-6R, JAK, and STAT3 inhibitors are currently the subject of preclinical and/or clinical investigations. In this Review, we summarize our current understanding of the role of IL-6/JAK/STAT3 signalling in cancer and in antitumour immunity, and the progress being made towards the development of clinical agents targeting this vital signalling pathway. Perspective is offered on the prospect of combining IL-6/JAK/STAT3 inhibitors with antibodies targeting the immune-checkpoint proteins programmed cell death protein 1 (PD-1), programmed cell death 1 ligand 1 (PD-L1), and cytotoxic T lymphocyte protein 4 (CTLA-4).

The IL-6/JAK/STAT3 signalling pathway

IL-6

Chronic inflammation promotes the development and progression of tumours. IL-6 can be expressed at high levels in the tumour microenvironment and is a major mediator of inflammation. In addition, IL-6 can directly stimulate the proliferation, survival, and invasiveness of tumour cells. IL-6 also induces the production of pro-inflammatory and angiogenesis-promoting factors, including IL-1β, IL-8, C-C motif chemokine (CCL)2, CCL3, CCL5, GM-CSF, and VEGF, which act in an autocrine and/or paracrine fashion on immune and non-immune cells within the tumour microenvironment³⁰. Together, these effects underscore the important role that IL-6 has in a variety of different cancers as well as the prognostic value of circulating IL-6 levels in patients with this disease.

The IL-6 protein is 21–28 kDa in size, depending on the extent of glycosylation. A viral form of IL-6, encoded by human herpesvirus 8, has approximately 25% homology with human IL-6 (REFS ^31,32). IL-6 was first identified as a factor capable of promoting B cell development and regulating the acute-phase immune response^33,34. Loss of IL-6 signalling reduces the effectiveness of both innate and adaptive immune responses to invading microorganisms and parasites³⁵. Notably, IL-6-deficient mice are resistant to both antigen-induced and collagen-induced arthritis and to multicentric Castleman disease^36–38. Elevated levels of IL-6 are seen in patients with arthritis or Castleman disease, and this observation has stimulated the successful clinical application of inhibitors of IL-6 signalling in the treatment of patients with these conditions^39,40.

IL-6 is produced in the tumour by infiltrating immune cells, by the tumour cells themselves, and by stromal cells. Thus, tumour-associated macrophages, granulocytes, and fibroblasts, as well as cancer cells, are all primary sources of IL-6 (REFS ^1,12–14). Adipocytes, T cells, and MDSCs also contribute to the elevated levels of IL-6 seen in tumours^1,14,15. Nuclear factor-κB (NF-κB) is a key transcription factor that drives the expression of IL-6 (REF. ⁴¹). Notably, hyperactivation of NF-κB is commonly observed in many human cancers. Hyperactivation of STAT3 in tumour cells also induces the production of IL-6, thus generating a positive-feedback loop¹⁶.

The IL-6R

IL-6 signalling is mediated by two different pathways: the classic signalling pathway and the trans-signalling pathway (FIG. 1). The classic signalling pathway involves binding of IL-6 to IL-6R on the cell surface and the subsequent interaction of this complex with the membrane-spanning protein IL-6 receptor subunit-β (gp130; also known as IL-6Rβ) to initiate intracellular signalling. In the trans-signalling pathway, IL-6 binds to a secreted form of the IL-6R (sIL-6R), followed by interaction of the IL-6–sIL-6R complex with gp130. Each pathway regulates distinct biological effects of IL-6. Classic signalling is particularly important for the acute-phase immunological response, haematopoiesis, and central homeostatic processes³⁵. Trans-signalling has a key role in the tumour microenvironment, acting to control the recruitment of leukocytes and the inflammatory activation of tumour-associated stromal cells^35,42.

Fig. 1. IL-6 signalling pathways.

a | In the classic IL-6 signalling pathway, IL-6 binds to the membrane-bound IL-6 receptor-α (subsequently referred to as IL-6R), thus inducing formation of a heterohexameric complex consisting of two molecules each of IL-6, IL-6R, and the IL-6 receptor subunit-β (gp130). Formation of this complex results in activation of the JAK/STAT3 signalling pathway, leading to the transcription of STAT3 target genes. The IL-6/IL-6R/gp130 complex can also activate the PI3K/AKT/mTOR (mechanistic target of rapamycin) and RAS/RAF/MEK/ERK pathways (not pictured). b | In the IL-6 trans-signalling pathway, soluble IL-6R (sIL-6R) binds to IL-6. sIL-6R can be generated by alternative splicing of IL6R mRNA or cleavage of membrane-bound IL-6R by disintegrin and metalloproteinase domain-containing protein 170(ADAM10) or ADAM17. When IL-6 binds with sIL-6R, this complex is then able to bind to and induce the dimerization of gp130, leading to the activation of downstream signalling pathways (as described above for classic IL-6 signalling pathways). gp130 is ubiquitously expressed, although the IL-6R is expressed in only a limited number of cell types. Trans-signalling via sIL-6R enables IL-6 to act on cells with limited or absent IL-6R expression. IL-6 trans-signalling can be negatively regulated by soluble gp130 (sgp130), which is generated by alternative splicing. This molecule competes with membrane-bound gp130 for binding to the IL-6–sIL-6R complex, thereby inhibiting IL-6 trans-signalling but not the classic IL-6 signalling pathway.

During classic signalling, IL-6 binds to the IL-6R, which is a single membrane-spanning protein of 80 kDa with a short cytoplasmic domain that lacks signalling capacity. Expression of IL-6R is largely restricted to specific subsets of leukocytes, megakaryocytes, hepatocytes, and certain barrier epithelial cells, thus limiting classic IL-6 signalling to these cell types^43,44. Intracellular signalling is initiated by the formation of the heterohexameric complex, consisting of IL-6, IL-6R, and gp130, followed by the recruitment of cellular signalling proteins, including JAKs and STAT3 (REF. ⁴⁵)(FIG. 1). Unlike IL-6R, gp130 is expressed on most cell types and is a shared co-receptor for other cytokines and growth factors, including IL-11, IL-27, LIF, oncostatin M, CNTF, CT1 (also known as CTF1), neuropoietin, humanin, and cardiotrophin-like cytokine⁶.

The discovery of trans-signalling was aided by the detection of sIL-6R in serum and urine samples^46,47. The findings of subsequent studies demonstrated that sIL-6R are generated by proteolytic cleavage of the IL-6R by the disintegrin and metalloproteinase domain-containing protein 17 (ADAM17) or ADAM10 proteases or by alternative splicing^48,49. The IL-6–sIL-6R complex retains the capacity to bind and activate signalling via gp130 (FIG. 1). Importantly, because gp130 is ubiquitously expressed, cells that do not express the IL-6R can respond to IL-6 through the trans-signalling pathway. Thus, shedding of sIL-6R by tumour-infiltrating neutrophils, monocytes, and T cells provides an opportunity for IL-6 to activate signalling in tumour and stromal cells with low levels of IL-6R expression and in those that do not express IL-6R³⁵.

Alternative splicing can generate four different secreted forms of gp130 (sgp130)^35,50. Secreted sgp130 proteins bind to the IL-6–sIL-6R complex and selectively inhibit trans-signalling⁵¹ (FIG. 1). This effect seems to be specific to IL-6 signalling as sgp130 has not been reported to affect signalling by other cytokines that activate signalling pathways involving gp130 (REF. ⁵²). The inhibitory activity of sgp130 has provided the basis for the development of agents and strategies designed to selectively inhibit IL-6-induced trans-signalling in patients.

JAKs

IL-6 signalling, via either the classic or trans-signalling pathways, involves the engagement of gp130, which leads to activation of gp130-associated JAKs. Four mammalian JAKs have been identified, JAK1, JAK2, JAK3, and TYK2, all of which are expressed in human cells, share considerable structural similarity, and have molecular masses ranging from 120 to 140 kDa. JAK1, JAK2, and TYK2 are widely expressed, while expression of JAK3 is largely restricted to cells of a haematopoietic origin⁵³. The carboxyl-terminal regions of JAKs contain a JAK homology (JH)1 domain, in which the tyrosine kinase domain is located. The JH1 domain is preceded by a pseudokinase domain (JH2), which interacts with the JH1 domain to restrain kinase activity in unstimulated cells⁶. Engagement of gp130 by the IL-6–IL-6R complex results in selective activation of JAK1, JAK2, and/or TYK2 via associations of these enzymes with membrane proximal domains (Box domains 1 and 2) in the gp130 protein⁵⁴ (FIG. 2). These interactions disrupt JH2-mediated inhibition of JH1 kinase activity, resulting in reciprocal transphosphorylation and full activation of JAKs. The activated JAK enzymes subsequently phosphorylate multiple tyrosine residues in the cytoplasmic region of gp130, which then serve as docking sites for proteins initiating the PI3K/AKT, RAS/RAF/MEK/MAPK, and JAK/STAT3 signalling pathways.

Fig. 2. Signalling downstream of the IL-6 receptor.

JAK proteins bind to the Box 1 and Box 2 domains in the intracellular portion of IL-6 receptor subunit-β (gp130). This leads to JAK-mediated phosphorylation of gp130 at several tyrosine residues, including four C-terminal residues that serve as docking sites for STAT3. Once bound to gp130, STAT3 is phosphorylated by JAKs at tyrosine 705, leading to STAT3 dimerization and nuclear translocation, followed by STAT3-mediated transcription of target genes. The figure shows signalling initiated by classic signalling. Trans-signalling pathways initiate downstream signalling in the same fashion. Tyrosine phosphorylation of STAT3 can also be induced by other oncogenic proteins, including SRC and BCR–ABL1. IL-6/JAK/STAT3 signalling is negatively regulated by a number of mechanisms. Suppressor of cytokine signalling (SOCS) 1 and SOCS3 bind to and inhibit the kinase activity of JAKs. SOCS3 is a STAT3 target gene; following transcription, SOCS3 then acts as a component of a negative-feedback loop that maintains tight regulation of this pathway. The following phosphatases also have a role in the negative regulation of this pathway: tyrosine-protein phosphatase non-receptor type 6 (SHP1; also known as PTPN6); tyrosine-protein phosphatase non-receptor type 11 (SHP2); dual specificity protein phosphatase 22 (DUSP22); receptor-type tyrosine-protein phosphatase-δ (PTPRD); receptor-type tyrosine-protein phosphatase T (PTPRT); tyrosine-protein phosphatase non-receptor type 1 (PTPN1); tyrosine-protein phosphatase non-receptor type 2 (PTPN2). Protein inhibitor of activated STAT3 (PIAS3), an E3 SUMO protein ligase, as well as PDZ and LIM domain protein 2 (PDLIM2), an ubiquitin E3 ligase, are additional endogenous proteins that inhibit STAT3 by mediating STAT3 degradation. The expression of PIAS3 and PDLIM2 can be inhibited by oncogenic microRNAs (miRNAs): miR-18a targets PIAS3, whereas miR-221 and miR-222 target PDLIM2. Another miRNA, miR-551b-3p, promotes STAT3 gene expression. Other miRNAs act to negatively regulate the IL-6/JAK/STAT3 pathway: expression of IL6R is inhibited by miR-218 and miR-34a, while STAT3 expression is inhibited by miR-17-5p, miR-20a, and miR-124.

STAT3

The JAK/STAT3 signalling pathway has a prominent role in mediating many of the effects of IL-6 on tumour cell proliferation, survival, invasion, and metastasis as well as suppressive effects on antitumour immunity. Among the seven members of the STAT protein family (STATs 1–4, 5A, 5B, and 6), STAT3 is the most strongly associated with the promotion of tumour growth and immunosuppression^17,55 and is the only family member whose genetic deletion results in embryonic lethality^56,57. In the inactive state, STAT3 exists as a monomer located in the cytoplasm. Following JAK-mediated tyrosine phosphorylation of growth factor receptors such as gp130, the SRC homology domain 2 (SH2) domain of STAT3 recognizes and binds to these phosphotyrosine docking sites, placing STAT3 within close proximity of active JAK enzymes, which subsequently phosphorylate STAT3 at Tyr705 (FIG. 2). Phosphorylation of Tyr705 results in SH2-domain-mediated, head-to-tail dimerization of the STAT3 protein and translocation of the STAT3 dimer to the nucleus via an importin-α–importin-β1-dependent mechanism⁵⁸. In the nucleus, STAT3 binds to consensus response elements in the promoters of target genes, thus inducing the transcription of a broad panel of genes encoding regulators of cellular proliferation (such as cyclin D1 and MYC) and survival (such as BCL-x_L and survivin) as well as angiogenesis-promoting (such as VEGF) and immunosuppressive growth factors and cytokines (such as IL-6)^14,17. The ability to phosphorylate and/or activate STAT3 is not limited to JAKs, as SRC and BCR–ABL1 tyrosine kinases have also been shown to directly activate STAT3 (REFS ^59,60). Additional post-translational modifications that might affect the function and/or stability of STAT3 have been reported, including phosphorylation of Ser727, and ubiquitylation, sumoylation, acetylation, or methylation of other residues^61,62. Phosphorylation of Ser727, although less well studied than Tyr705 phosphorylation, can be mediated by several kinases, including JNK and other MAPKs and protein kinase C, and generally promotes STAT3 activity.

Stimulation of nonmalignant cells with IL-6 or other cytokines results in transient phosphorylation and/or activation of STAT3. In these cells, STAT3 activation can be controlled by at least three different classes of negative regulators: PIAS proteins, SOCS proteins (such as suppressor of cytokine signalling (SOCS) 1, 2, and 3), and several cellular phosphatases (SHP1, SHP2, DUSP22, PTPRD, PTPRT, and PTPN1 and 2)^8,56,63,64. As described in a subsequent section, inhibition or reduced expression of these negative regulators can lead to constitutive activation of STAT3, an effect that is often observed in patients with cancer.

Signalling via the IL-6/JAK/STAT3 pathway is also regulated via a complex interplay with cellular miRNAs. Several miRNAs have been shown to dampen IL-6/JAK/STAT3 signalling by reducing expression of the components of this pathway, either in tumour cells or in tumour-infiltrating immune cells (FIG. 2). For example, miR-17-5p, miR-20a, and miR-124 reduce STAT3 expression, and miR-34a and miR-218 suppress IL-6R expression^65–68. However, several miRNAs directly induce STAT3 upregulation (miR-551b-3p) or act to reduce expression of negative regulators of STAT3 (miR-18a, miR-221, and miR-222)^69–71. Regarding the latter activity, miRNA-18a negatively regulates expression of E3 SUMO protein ligase PIAS3, while miR-221 and miR-222 suppress expression of PDZ and LIM domain protein 2 (PDLIM2), an E3 ubiquitin ligase that promotes polyubiquitylation and subsequent proteasomal degradation of STAT3. Altered expression of miRNAs that regulate cellular levels of STAT3 is likely to have a role in the hyperactivation of STAT3 signalling in cancer.

IL-6, JAKs, and STAT3 in cancer

IL-6, IL-6R, and gp130

IL-6 acts to recruit immune cells within the tumour microenvironment, therefore stimulating the production of additional pro-inflammatory cytokines. Thus, IL-6 serves as a key link between chronic inflammation and tumour progression. Similar to patients with arthritis and those with Castleman disease, or following infection, elevated IL-6 levels are observed in patients with a variety of different forms of cancer¹. In particular, elevated circulating levels of IL-6 have been reported in patients with breast⁷², cervical⁷³, colorectal⁷⁴, oesophageal⁷⁵, head-and-neck^76,77, ovarian^78,79, pancreatic⁸⁰, prostate⁸¹, and renal⁸² cancers as well as in those with non-small-cell lung cancer (NSCLC)⁸³ or multiple myeloma⁵. Furthermore, circulating IL-6 levels are increased by surgery⁸⁴ and chemoradiation⁸⁵, and are reported to be increased in patients with recurrent tumours¹. Elevated serum IL-6 levels are also observed in patients with inflammatory bowel disease, and IL-6 levels generally correlate with tumour size, stage, and metastasis in patients with colorectal cancer⁸⁶. In those with breast cancer, the highest levels of IL-6 are detected at the leading edges of the tumours, and IL-6 levels have been shown to be closely correlated with advanced-stage disease, as indicated by the number of tumour-positive lymph nodes¹⁶. Importantly, circulating IL-6 levels have been shown to be prognostic indicators of survival^{73,76,78,79,83,86,87} as well as predictors of a response to therapy^75,85,88 in several different types of cancer.

Data from preclinical studies have confirmed that IL-6 signalling has important roles in tumour development. Findings from preclinical models and patient samples demonstrate that IL-6 promotes the development of breast^16,89, colorectal^90–92, lung (NSCLC)⁹³, pancreatic⁹⁴, and skin³⁰ cancers. In breast cancer, IL-6 induces Notch 3 activation, thus promoting self-renewal of tumour stem cells⁸⁹, and exogenous expression of IL-6 has been shown to promote breast cancer metastasis¹⁶. IL-6 signalling also stimulates epithelial-to-mesenchymal transition in breast⁹⁵ and head-and-neck⁹⁶ cancers. IL-6 signalling is also required for the survival of nonmalignant intestinal epithelial cells as well as the development of colitis-associated and sporadic forms of colorectal cancer^92,97.

To date, no clinically relevant genomic alterations in the genes encoding IL-6, IL-6R, or gp130 have been detected in the tumour types analysed by The Cancer Genome Atlas (TCGA)^98,99. However, activating mutations in gp130 have been shown to occur in approximately 60% of surgical specimens from patients with inflammatory hepatocellular adenomas¹⁰⁰. Additionally, a single-nucleotide polymorphism (-174G>C) in the promoter region of the IL6 gene has been shown to result in increased expression of this cytokine¹⁰¹. Epigenetic alterations have a prominent role in aberrant activation of the IL-6/IL-6R/JAK/STAT3 pathway in cancer, and changes in the expression and/or activation of transcription factors might have a prominent role in the elevated expression of IL-6 in cancer.

JAKs

JAKs mediate the activation of STAT3 in both nonmalignant and malignant cells exposed to IL-6. The anti-inflammatory effects of JAK inhibitors, which have been broadly investigated, are largely attributed to inhibition of STAT3 and/or STAT5 activation. Much of the interest in the role of JAKs in cancer and the clinical application of JAK inhibitors has focused on haematological malignancies and conditions involving chronic inflammation^6,102–106. For example, the JAK2^V617F mutation is found in >95% of patients with polycythaemia vera and results in a JAK2 enzyme that is constitutively active independent of cytokine signalling^{102,106–108}. JAK2^V617F is also observed in patients with essential thrombocythemia and in those with primary myelofibrosis. Additional JAK2 mutations occur in B cell acute lymphoblastic leukaemia (B-ALL), Down syndrome ALL, Hodgkin lymphoma, and B cell lymphoma^105,106. In paediatric T cell acute lymphoblastic leukaemia (T-ALL), a chromosomal translocation results in the constitutively active transcription factor ETV6 (TEL)–JAK2 fusion protein¹⁰⁹. Moreover, mutations leading to dysregulation of JAK1 activity have been reported in B-ALL, adult T-ALL, and essential thrombocythemia¹⁰⁶. Similarly, JAK3 mutations are also found in patients with B-ALL, adult T-ALL, and essential thrombocythemia¹⁰⁶. Mutations in the genes encoding JAK enzymes seem to be much less common in solid tumours. However, JAK1 is mutated in 3.8% and 11.5% of hepatocellular carcinomas and endometrial cancers, respectively, including the presence of activating mutations^98,99,110.

STAT3

Aberrantly elevated STAT3 activity has been estimated to occur in >70% of human cancers^111,112. Malignancies in which hyperactivation of STAT3 has been reported include acute myeloid leukaemia (AML), multiple myeloma, and solid tumours of the bladder, bone, breast, brain, cervix, colon, oesophagus, head-and-neck, kidney, liver, lung, ovary, pancreas, prostate, stomach, and uterus^113–125. The levels of phosphorylated and/or activated STAT3 have been shown to correlate with a poor clinical prognosis in several of these cancers. Exogenous expression of a constitutively active form of STAT3 confers anchorage-independent growth and tumorigenic capacity on fibroblasts, thus demonstrating the oncogenic activity of the STAT3 protein¹²⁶.

Preclinical studies aimed at modulating the expression and/or function of STAT3 have demonstrated important roles of STAT3 in the development and/or progression of multiple cancers, including bladder, colorectal, head-and-neck, lung, pancreatic, prostate, and skin cancers^127–133. STAT3 also promotes resistance to conventional chemotherapy and radiation therapy as well as to targeted therapies, such as cetuximab^134,135. Indeed, activation of STAT3 via a positive-feedback loop constitutes a primary mechanism of resistance in drug-treated, oncogene-addicted cells¹³⁶. These preclinical data suggest that agents and approaches that block the activity of STAT3 in tumour cells could have substantial additional value in preventing or reversing anticancer drug resistance.

The effects of STAT3 activation on the growth of tumour cells are mediated, in large part, by the STAT3-mediated induction of key target genes that regulate cellular proliferation and metabolism, suppression of apoptosis, and responses to hypoxia¹⁴. Activated STAT3 also induces the expression of VEGF and selected MMPs, which promote angiogenesis and invasiveness and/or metastasis, respectively¹⁸. In addition, STAT3 binds to the IL6 promoter, generating a positive-feedback loop, leading to increased IL-6 expression¹⁶. Both VEGF and IL-6 can also have immunosuppressive effects, which might facilitate immune evasion by tumour cells harbouring hyperactivated STAT3. The tumour-promoting effects of STAT3 might also be mediated by induction of miR-21 and miR-181b-1, which suppress the expression of the tumour suppressors PTEN and ubiquitin carboxyl-terminal hydrolase CYLD, respectively¹³⁷.

Emerging evidence indicates that STAT3 is also hyperactivated in tumour-infiltrating immune cells and might have profound effects on antitumour immunity^19,20. Conditional deletion of the Stat3 gene in murine haematopoietic cells has revealed several potent immunosuppressive effects of STAT3, including negative regulation of neutrophil and NK cell function, induction of PD-1 expression, inhibition of effector T cell function, inhibition of DC maturation and function, and expansion of T_reg cell and MDSC numbers in the tumour microenvironment^{19,21–29,138}. STAT3-mediated induction of immunosuppressive factors in the tumour microenvironment, including IL-6, IL-10, TGFβ, and VEGF, stimulates positive-feedback amplification of STAT3 activation in both tumour cells and tumour-infiltrating immune cells^{17,20,29,139,140}. The end result of this complex crosstalk between cancer and immune cells in the tumour microenvironment is increased tumour cell growth and survival, with diminished antitumour immunity. Collectively, these observations suggest that selectively targeting STAT3 in cancer could provide multiple benefits: inhibition of cell-autonomous effects on tumour cell growth and metastasis; inhibition of cell-autonomous immunosuppressive effects in infiltrating immune cells; and inhibition of immunosuppressive crosstalk between tumour cells and tumour-infiltrating immune cells.

STAT3 mutations are rare and primarily restricted to patients with haematological malignancies. Mutations in the exon encoding the STAT3 SH2 domain were originally identified by Koskela et al.¹⁴¹ in 31 of a cohort of 77 patients with large granular lymphocytic (LGL) leukaemia. Additional activating mutations in the SH2 domain have been reported in NK and T cell lymphomas¹⁴². Activating mutations outside of the SH2 domain have also been identified by Andersson et al.¹⁴³ in patients with LGL leukaemia.

Hyperactivation of STAT3 in tumours can occur through a variety of mechanisms. Elevated expression of IL-6 and increased stimulation of IL-6R commonly result in hyperactivation of JAK/STAT3 signalling in tumours. Autocrine stimulation of growth factor receptors, such as EGFR, can also lead to induction of STAT3 signalling. In certain malignancies, such as NSCLC, EGFR is overexpressed or mutated to a constitutively active form, and a similar situation can occur with JAK enzymes¹⁴⁴. Furthermore, many tumours harbour hyperactivation of SRC, which can promote STAT3 phosphorylation and/or activation¹⁴⁵. Thus, therapies targeting EGFR or SRC, in addition to those targeting components of the IL-6 pathway, could potentially be used to downmodulate STAT3 activation and signalling.

Alterations in proteins that negatively regulate STAT3 in nonmalignant cells can also contribute to aberrant activation of STAT3. Loss of SOCS1, and/or 3 expression, often owing to promoter hypermethylation, occurs in multiple forms of cancer, including tumours of the brain, cervix, colon, head and neck, liver, lung, ovary, pancreas, and prostate as well as in melanoma⁶. Inhibition or mutation of the SHP1 and SHP2 phosphatases, in addition to loss of expression of PIAS family members, has also been reported^6,7. Mutations in the genes encoding PTPRT and PTPRD have been found in 5.6% and 3.7%, respectively, of head-and-neck cancers, with promoter methylation contributing further to loss of expression^8–11.

Hyperactivation of STAT3 is primarily associated with the promotion of tumour growth, although a growing body of evidence indicates that activated STAT3 might have tumour-suppressive functions in certain settings. STAT3 is reported to be a negative regulator of the development and progression of KRAS-induced lung cancer in Apc-mutant mice^146–148. Levels of activated STAT3 have been positively correlated with a better prognosis in patients with nasopharyngeal carcinoma, colorectal carcinoma, or leiomyosarcoma^149–151. Thus, the eventual clinical application of STAT3 inhibitors requires careful consideration of the role of STAT3 activation in each specific cancer.

Targeting the IL-6/JAK/STAT3 pathway

Inhibitors of IL-6 and IL-6Rs

Three main approaches to inhibition of IL-6-mediated signalling at the ligand and/or receptor level are currently in use: directly targeting IL-6 with antibodies, such as siltuximab; targeting the IL-6R with antibodies, such as tocilizumab; and targeting the IL-6–sIL-6R complex using fusion proteins incorporating sgp130. Direct targeting of IL-6 and IL-6 receptors inhibits both classic and trans-signalling, while targeting the IL-6–sIL-6R complex with sgp130 fusion proteins selectively inhibits trans-signalling.

Siltuximab, a chimeric mouse–human antibody, is currently the most extensively developed clinical agent targeting IL-6 (FIG. 3; TABLE 1). Following positive results of several clinical trials, siltuximab was granted FDA approval in 2014 for the treatment of multicentric Castleman disease^152–154. The findings of preclinical studies indicate that siltuximab increases the activity of melphalan in in vitro models of multiple myeloma¹⁵⁵, although the addition of siltuximab to bortezomib-based or melphalan-based regimens did not improve overall response rates or progression-free survival rates in clinical trials^156–160. In preclinical models of solid tumours, siltuximab demonstrated antitumour efficacy against ovarian¹⁶¹, prostate¹⁶², and lung¹⁶³ cancers. Analyses of tumour material from phase I–II studies involving patients with prostate cancer have shown that siltuximab decreases levels of activated STAT3 and MAPKs and results in a serum PSA-defined response rate of 3.8%, with no significant improvements in patient outcomes^164–166. Stabilized disease was obtained in >50% of patients with metastatic renal cell carcinoma receiving siltuximab in a phase I–II clinical trial¹⁶⁷. However, no clinical activity was observed in phase I–II trials involving patients with advanced-stage cancer (including colorectal, head-and-neck, lung (NSCLC), ovarian, or pancreatic tumours)¹⁶⁸. Thus, while promising results have been obtained in preclinical studies, evidence demonstrating activity of siltuximab against solid tumours in clinical trials has been largely limited. These findings highlight the possibility that targeting IL-6 alone in unselected patient populations is unlikely to have a marked effect on the outcomes of patients with solid tumours. Overcoming this obstacle will require the development of effective combination therapies as well as the identification of reliable and robust biomarkers that are predictive of a response. Additional anti-IL-6 antibodies currently in preclinical development in cancer include sirukumab, olokizumab, clazakizumab, and MEDI5117 (REFS ^169–171).

Fig. 3. Inhibitors of the IL-6/JAK/STAT3 signalling pathway.

Various targeted agents that inhibit different nodes of the IL-6 signalling pathway have been developed. Siltuximab, sirukumab, olokizumab, clazakizumab, and MEDI5117 are anti-IL-6 monoclonal antibodies. Tocilizumab and sarilumab are monoclonal antibodies that target IL-6R. These antibodies inhibit both the classic and trans-signalling pathways. By contrast, the gp130–Fc fusion protein olamkicept inhibits IL-6 trans-signalling but not the classic signalling pathway. Tofacitinib, ruxolitinib, pacritinib, and AZD1480 are small-molecule tyrosine kinase inhibitors that target JAKs, preventing phosphorylation of STAT3. C188-9, OPB-31121, OPB-51602, and other Src homology domain 2 (SH2) domain inhibitors interfere with STAT3 dimerization. The STAT3 antisense oligonucleotide AZD9150 binds to and causes the destruction of STAT3 mRNA, thus decreasing STAT3 expression. The cyclic STAT3 decoy contains a nucleotide sequence derived from the promoter of the STAT3 target gene FOS. This decoy competitively inhibits STAT3 binding to genomic response elements in the promoter regions of target genes.

Table 1.

Anti-IL-6 or anti-IL-6-receptor antibodies in clinical trials

Inhibitor (type of agent)	Indication	Study phase	NCT identifier	Trial results	Refs
Siltuximab (anti-IL-6 mAb)	Multiple myeloma, B cell non-Hodgkin lymphoma, Castleman disease	I	NCT00412321	No DLTs observed; recommended dose for future studies determined	154
	Multiple myeloma (smoldering or indolent)	I	NCT01219010	10% M-protein response, 30% minor M-protein response; acceptable safety profile	NA
	Multiple myeloma	I	NCT01309412	Study terminated owing to safety concerns	156
	Multiple myeloma	I/II	NCT01531998	90.9% ORR (PR) in combination with RVD; MTD determined	160
	Multiple myeloma	II	NCT00401843	No increase in PFS or OS compared with bortezomib alone	157
	Multiple myeloma (high-risk smoldering)	II	NCT01484275	NA	NA
	Multiple myeloma	II	NCT00402181	No response to single-agent siltuximab; 17% ORR in combination with dexamethasone in patients with dexamethasone-refractory disease	158
	Multiple myeloma	II	NCT00911859	Addition of siltuximab to VMP did not improve the number of patients having a CR, PFS, or OS but did improve the number of patients with a VGPR	159
	Myelodysplastic syndromes	II	NCT01513317	Study terminated owing to a lack of efficacy	NA
	Prostate cancer	I	2047 SN:218/4.2 (Innsbruck Medical University)	No siltuximab-related adverse events observed	164
	Metastatic, hormone-refractory prostate cancer	I	NCT00401765	62.2% of patients had a PSA-defined response; 89.7% of patients discontinued treatment prior to completion of all 14 cycles	NA
	Metastatic, hormone-refractory prostate cancer	II	NCT00385827	Study terminated owing to a lack of efficacy; well tolerated in combination with MP	166
	Metastatic, hormone-refractory prostate cancer	II	NCT00433446	Minimal clinical activity despite evidence of a reduction in IL-6 levels (decrease in serum CRP levels)	165
	Solid tumours	I/II	NCT00841191	No clinical activity observed but well tolerated as monotherapy; recommended phase II dose determined	168
	Metastatic renal cell carcinoma	I/II	NCT00265135	SD in >50% of patients; no DLTs observed	167
Tocilizumab (anti-IL-6 receptor mAb)	B cell chronic lymphocytic leukaemia	I	NCT02336048	NA	NA
	Metastatic HER2+ breast cancer	I	NCT03135171	NA	NA
	Ovarian cancer	I/II	NCT01637532	Immunological changes consistent with decreased levels of immunosuppression; no DLTs observed	174
	Pancreatic cancer	II	NCT02767557	NA	NA

CR, complete response; CRP, C-reactive protein; DLT, dose-limiting toxicity; mAb, monoclonal antibody; MP, mitoxantrone plus prednisone; MTD, maximum-tolerated dose; NA, not available; ORR, overall response rate; OS, overall survival; PFS, progression-free survival; PR, partial response; PSA, prostate-specific antigen; RVD, lenalidomide plus bortezomib and dexamethasone; SD, stable disease; VGPR, very good partial response; VMP, bortezomib plus melphalan and prednisone.

Tocilizumab is a humanized monoclonal antibody that recognizes IL-6R and disrupts both classic and trans-signalling (FIG. 3; TABLE 1). Tocilizumab has been approved by the FDA for use in adult patients with rheumatoid arthritis and in patients with systemic juvenile idiopathic arthritis and for the management of cytokine-release syndrome in adult or paediatric patients with B-ALL receiving treatment with CAR T cells. The findings of preclinical studies suggest that tocilizumab is effective against ovarian¹⁷², pancreatic¹⁷³, and colitis-associated colorectal cancers⁹². The findings of a phase I clinical trial demonstrated the combination of tocilizumab with carboplatin and/or doxorubicin to be feasible and safe in patients with ovarian cancer¹⁷⁴. Early phase trials exploring the safety and effectiveness of tocilizumab in patients with B cell chronic lymphocytic leukaemia (B-CLL) and in those with breast or pancreatic cancers are currently ongoing (NCT02336048, NCT03135171, and NCT02767557). Another monoclonal antibody that targets the IL-6R, sarilumab, is currently in preclinical development.

Selective inhibition of trans-signalling might be of particular value in patients whose tumours have either very limited or no IL-6R expression. Trans-signalling can be selectively inhibited by proteins incorporating the sgp130 sequence, which bind with and inhibit the IL-6–IL-6R complex⁵¹. An sgp130-Fc fusion protein has been shown, in preclinical models, to inhibit the development and/or progression of KRAS-driven NSCLC, pancreatic cancer, and colitis-associated premalignant colorectal cancer^173,175,176. Olamkicept, an sgp130-Fc fusion protein, is currently in phase I–Ib testing in patients with rheumatoid arthritis and in those with inflammatory bowel diseases. In addition to activating STAT3, IL-6 can also activate STAT1 via gp130 (REFS ^177,178). STAT1 is known to have tumour-suppressive properties. Thus, targeted inhibition of IL-6 signalling in patients with cancer has the potential downside of reducing STAT1 activity in tumour cells.

JAK inhibitors

Tofacitinib, ruxolitinib, and pacritinib are currently the most extensively investigated clinical inhibitors of JAKs, with several other agents currently in preclinical development (FIG. 3). To date, the clinical application of JAK inhibitors has focused heavily on conditions involving chronic inflammation and myeloproliferative neoplasms, with less evaluation of these agents in patients with solid tumours.

Tofacitinib is an orally administered JAK inhibitor that selectively inhibits JAK1 and JAK3, with a lower affinity for JAK2 (REF. ¹⁷⁹). Tofacitinib has been approved by the FDA for the treatment of rheumatoid arthritis^180–183. Tofacitinib is also undergoing clinical evaluation as a treatment of other inflammatory conditions, including chronic plaque psoriasis¹⁸⁴, ulcerative colitis¹⁸⁵, and Crohn’s disease¹⁸⁶. Ruxolitinib is an orally administered JAK inhibitor with selectivity for JAK1 and JAK2 and has received FDA approval for the treatment of patients with intermediate-risk or high-risk myelofibrosis and for patients with hydroxyurea-resistant or intolerant polycythaemia vera¹⁰⁵. The FDA approval of ruxolitinib for myelofibrosis was based on data from the COMFORT-1 and COMFORT-2 trials, which compared the efficacy of ruxolitinib with that of placebo or the best available therapy^187–189. Ruxolitinib was found to induce a marked reduction in spleen volumes and total symptom scores in both trials, although myelosuppression and anaemia were frequent dose-limiting events. Given the importance of JAK2 for haematopoiesis¹⁹⁰, the myelosuppression associated with ruxolitinib is not surprising. Interestingly, the effectiveness of ruxolitinib in patients with myelofibrosis is not dependent on the presence of JAK2^V617F mutations¹⁹¹. The FDA approval of ruxolitinib as a treatment of polycythaemia vera was based on findings of the RESPONSE clinical trial, in which ruxolitinib had superior efficacy to that of the standard-of-care approach (including hydroxyurea, interferons or pegylated interferons, pipobroman, anagrelide, or immunomodulators such as lenalidomide or thalidomide)^192,193. Evaluations of the efficacy of ruxolitinib in patients with essential thrombocythemia in later-phase clinical trials are currently ongoing.

The dose-limiting myelosuppression that frequently accompanies treatment with ruxolitinib has stimulated the search for JAK inhibitors with a reduced risk of adverse events. Pacritinib is an orally administered, selective JAK2 and FLT3 inhibitor classified as lacking in myelosuppressive activity. Clinical evaluations of the efficacy of pacritinib in the PERSIST-1 trial in patients with myelofibrosis revealed favourable levels of activity compared with that of the best available therapy (ruxolitinib not included), even in patients with thrombocytopenia^194,195. However, accrual to the subsequent PERSIST-2 trial, which was designed to evaluate the efficacy of pacritinib in patients with myelofibrosis with thrombocytopenia, was temporarily halted by the FDA owing to the emergence of severe cardiac and haemorrhagic adverse events. This trial was terminated, but pacritinib is now being investigated in a phase II clinical trial in patients with myelofibrosis who hab been previously treated with ruxolitinib (NCT03165734).

Clinical investigation of the efficacy of JAK inhibitors in patients with solid tumours is currently limited. In preclinical studies, JAK inhibition has been shown to curtail the in vivo growth of a broad variety of solid tumours, including brain, breast, colorectal, gastric, head-and-neck, liver, lung, pancreatic, and ovarian cancers^14,196–200. Many of the early preclinical studies used the JAK1 and JAK2 inhibitor AZD1480. However, phase I testing in patients with solid tumours revealed a major risk of neurological adverse events following treatment with AZD1480, including anxiety, ataxia, behavioural changes, hallucinations, and memory loss, resulting in the discontinuation of further testing of this agent²⁰¹. Phase I evaluations of the safety and tolerability of ruxolitinib in children with recurrent and/or refractory solid tumours have revealed acceptable levels of tolerability and have enabled the identification of a recommended dose for further testing in phase II studies²⁰². Data from a phase II study involving adults with metastatic pancreatic cancer have shown that the combination of ruxolitinib and capecitabine is well tolerated and might lead to improved survival outcomes²⁰³. Additional early phase trials exploring the safety and effectiveness of ruxolitinib in patients with breast, colorectal, head-and-neck, lung, ovarian, pancreatic, or prostate cancers are currently ongoing (for example, NCT02066532, NCT03153982, NCT02713386, and NCT03274778).

STAT3 inhibitors

STAT3 has established tumour-promoting properties, is overexpressed and/or hyperactivated in the majority of human cancers, and is commonly associated with a poor prognosis; therefore, considerable effort has gone into identifying and developing STAT3 inhibitors that can be applied in the clinic. However, given the status of STAT3 as an intracellular transcription factor that, therefore, lacks enzymatic activity, it has often been considered an ‘undruggable’ target, and the development of possible inhibitors has been difficult²⁰⁴. Nonetheless, several compounds that inhibit either the function or expression of STAT3 have now reached clinical trials (FIG. 3; TABLE 2).

Table 2.

STAT3 inhibitors in clinical trials

Inhibitor	Indication	Study phase	NCT identifier	Trial results	Refs
AZD9150 (STAT3 antisense oligonucleotide)	DLBCL	I	NCT02549651	NA	NA
	Advanced-stage and/or metastatic hepatocellular carcinoma	I	NCT01839604	NA	NA
	Advanced-stage solid tumours, metastatic HNSCC	I/II	NCT02499328	NA	NA
	Advanced-stage cancers, DLBCL, advanced-stage lymphomas	I/II	NCT01563302	SD or PR in 44% of patients; durable PR in 2 of 6 patients with DLBCL; MTD determined	240
	Advanced-stage pancreatic cancer, NSCLC, and CRC	II	NCT02983578	NA	NA
C188-9 (STAT3 SH2 domain binder)	Advanced-stage cancers	I	NCT03195699	NA	NA
OPB-31121 (STAT3 SH2 domain binder)	Solid tumours	I	NCT00955812	No objective responses; MTD determined	225
	Advanced-stage solid tumours	I	NCT00657176	SD in 8 of 18 evaluable patients; MTD determined	226
	Hepatocellular carcinoma	I/II	NCT01406574	SD in 6 of 23 patients; authors described antitumour activity as insufficient, thus precluding further clinical development	227
OPB 51602 (STAT3 SH2 domain binder)	Multiple myeloma, NHL, AML, ALL, and CML	I	NCT01344876	No clear therapeutic response; MTD and recommended dose determined	229
	Nasopharyngeal carcinoma	I	NCT02058017	Study terminated owing to the emergence of intolerable lactic and metabolic acidosis	NA
	Advanced-stage cancers	I	NCT01423903	NA	NA
	Advanced-stage solid tumours	I	NCT01184807	PR in 2 of 37 evaluable patients (both with EGFR–TKI -refractory NSCLC); MTD and recommended phase II dose determined	228
STAT3 decoy oligonucleotide (STAT3 response element from FOS)	HNSCC	0	NCT00696176	Decreased STAT3 target gene expression in tumours following intratumoural injection; no toxicities reported	238

ALL, acute lymphoblastic leukaemia; AML, acute myeloid leukaemia; CML, chronic myeloid leukaemia; CRC, colorectal cancer; DLBCL, diffuse large B cell lymphoma; HNSCC, head and neck squamous cell carcinoma; MTD, maximum-tolerated dose; NA, not available; NHL, non-Hodgkin lymphoma; NSCLC, non-small-cell lung cancer; PR, partial response; SD, stable disease; SH2, SRC homology domain 2; STAT3, signal transducer and activator of transcription 3; TKI, tyrosine kinase inhibitor.

The first direct inhibitors of STAT3 to be developed were based on tyrosine-phosphorylated peptides (PY*LKTK) and peptidomimetics (ISS-610, PM-73G), which are capable of binding to the SH2 domain of STAT3, and disrupt STAT3 dimerization and DNA-binding activity^205–208. These agents have demonstrated pro-apoptotic and antitumour activity against cancer cells with STAT3 hyperactivation, although they are limited in potency, cellular uptake, stability, and potential immunogenicity and have primarily been used as research tools. A number of non-peptide SH2 domain inhibitors have also been identified and have been shown to inhibit the growth of cells and/or tumours with elevated levels of activated STAT3, including STA-21, LLL-3, STATTIC, WP1066, S3I-201, BP-1-102, STX-0119, and HJC0123 (REFS ^209–221). Of these compounds, BP1-102, STX-0119, and HJC0123 are orally bioavailable in preclinical models.

The nonpeptide STAT3–SH2 domain antagonists OPB-31121, OPB-51602, and C188-9 have all been evaluated in early phase clinical trials^222–230. Phase I studies exploring the safety and tolerability of orally administered OPB-31121 in patients with hepatocellular carcinoma or other advanced-stage solid tumours have been completed, and maximum-tolerated doses have been established^225–227. Evidence of antitumour activity was reported by Oh et al.²²⁶, although peripheral sensory neuropathy was observed in the study conducted by Okusaka et al.²²⁷. OPB-51602 can also be orally administered and has been evaluated in phase I studies involving patients with relapsed and/or refractory haematological malignancies and treatment-refractory solid tumours^228,229. Possible antitumour activity was seen in patients with NSCLC, although both studies revealed tolerability-related difficulties, including peripheral neuropathy and drug-induced pneumonitis. C188-9 is a high-affinity STAT3 inhibitor (K_d ~5 nM) that can also be orally administered. It inhibits the growth of radioresistant head-and-neck cancer xenografts and is currently being evaluated in a phase I study involving patients with advanced-stage solid tumours²³⁰.

An alternative method of inhibiting STAT3 function involves competitive inhibition of the interactions between STAT3 and promoter elements in target genes. A 15-bp double-stranded decoy oligonucleotide corresponding to the STAT3 response element in the FOS promoter has been shown to competitively inhibit STAT3 binding to DNA and to suppress the tumour growth of preclinical models of brain, breast, head-and-neck, lung, ovarian, and skin cancers as well as AML^{129,231–237}. A phase 0 study involving intratumoural injection of this linear STAT3 decoy oligonucleotide in patients with head-and-neck cancer has demonstrated downregulation of STAT3 target genes²³⁸. A cyclic version of the STAT3 decoy has subsequently been generated²³⁸. The cyclic STAT3 decoy has increased heat and nuclease resistance, antitumour activity against xenograft tumour models following intravenous administration, and no apparent toxicities when administered at high doses^238,239.

Inhibition of STAT3 expression using antisense oligonucleotides provides another distinctly different approach to inhibition of cellular STAT3 activity. AZD9150 is a second-generation STAT3 antisense oligonucleotide that is optimized from previous iterations by incorporating 2′–4′ constrained ethyl-modified residues²⁴⁰. Preclinical testing has revealed a lack of end-organ damage and other adverse events following AZD9150 administration²⁴¹. Notably, intravenous delivery of AZD9150 has been shown to inhibit the growth of lymphoma and NSCLC xenografts and to increase the sensitivity of cell-line models of neuroblastoma to cisplatin^240,242. Moreover, clinical evaluations of AZD9150 have revealed activity against treatment-refractory lymphoma and NSCLC, with a maximum-tolerated dose of 3 mg/kg (REF. ²⁴⁰). The major toxicities associated with AZD9150 in previous studies proved modest and primarily involved rapid induction of thrombocytopenia in two of the nine patients treated with 4 mg/kg doses. The generally favourable safety profile and preliminary evidence suggesting efficacy in patients support the further evaluation of AZD9150 in clinical trials. A further challenge facing the clinical implementation of both AZD9150 and the cyclic STAT3 decoy involves the delivery of nucleotide-based agents owing to the high molecular weight of both molecules.

Immunotherapy combinations

The targeted inhibition of immune checkpoints using monoclonal antibodies has led to dramatic improvements in the treatment of patients with advanced-stage cancer. Ipilimumab inhibits CTLA-4 activation, while pembrolizumab and nivolumab inhibit PD-1 signalling. Both CTLA-4 and PD-1 are inhibitory cell-surface receptors that act to restrain T cell-mediated immune responses. PD-L1 activates PD-1 and is commonly expressed on the surface of tumour cells or tumour-infiltrating immune cells. Collectively, these antibodies have received FDA approval for the treatment of a diverse range of cancers, including melanoma, Hodgkin lymphoma, bladder and head-and-neck cancers, and NSCLC. Given the early successes observed with immune-checkpoint inhibition, broader clinical application of these and similar antibodies targeting immune-checkpoint proteins is likely. Inhibition of IL-6/JAK/STAT3 signalling can also affect the tumour microenvironment and has implications for antitumour immunity; therefore, determining whether co-targeting of immune checkpoints and the IL-6/JAK/STAT3 signalling pathway might be beneficial is important. Early indications suggest that inhibition of IL-6/JAK/STAT3 signalling will be useful in combating the various adverse inflammatory effects resulting from treatment with immune-checkpoint inhibitors. Moreover, preclinical evidence is emerging that inhibition of IL-6/JAK/STAT3 signalling might augment the antitumour efficacy of immune-checkpoint inhibitors.

Treatment of patients with cancer with immune-checkpoint inhibitors can stimulate the production of IL-6 (REFS ^243–245). The elevation of serum IL-6 levels in these patients is typically manifested as inflammatory conditions, such as psoriasiform dermatitis, arthritis, and Crohn’s disease^{243,244,246,247}. Importantly, treatment with tocilizumab has been shown to resolve these conditions and to enable patients to continue to receive immune-checkpoint inhibitors^245–247.

The findings of numerous studies have shown that signalling via the IL-6/JAK/STAT3 pathway induces the expression of PD-1 and/or PD-L1 (REFS ^248–252). Inhibition of IL-6/JAK/STAT3 signalling downregulates PD-1 and/or PD-L1 expression and might, hypothetically, be expected to have one of two possible effects on the efficacy of immune-checkpoint inhibition. Targeting IL-6/JAK/STAT3 signalling might be expected to improve the effectiveness of pembrolizumab and nivolumab as a result of direct inhibitory effects on tumour cells as well as effects on immune cells in the tumour microenvironment. Alternatively, downregulation of PD-1 and PD-L1 by inhibitors of IL-6/JAK/STAT3 signalling could result in the attenuation of the activity of pembrolizumab and/or nivolumab by reducing the expression of the proteins targeted by these antibodies. Substantial preclinical and clinical research will be required to address this important issue, although preliminary studies in preclinical models suggest a clinical benefit from the combination of agents targeting the IL-6/JAK/STAT3 pathway with immune-checkpoint inhibition. Co-targeting of IL-6 and PD-L1 leads to enhanced inhibition of tumour growth in mouse models of both pancreatic ductal and hepatocellular carcinomas^253,254. Treatment with ruxolitinib has been shown to overcome resistance to anti-PD-1 antibodies in mice with pancreatic orthotopic tumours²⁵⁵. Clearly, further investigations involving these combination approaches are warranted.

Future directions

The IL-6/JAK/STAT3 pathway is hyperactivated in many patients with cancer, and the findings of numerous studies involving preclinical in vitro and in vivo models demonstrate that targeting individual nodes in this pathway can have antitumour effects. In addition to the importance of IL-6/JAK/STAT signalling in tumour cells, activation of this pathway has also been implicated in suppressing antitumour immune responses in the tumour microenvironment. Thus, therapies that target this pathway are likely to benefit patients with cancer, both by inhibition of tumour cell growth and through stimulation of antitumour immunity. Agents that target individual nodes, including IL-6, IL-6R, and JAKs, are all approved by the FDA for the treatment of inflammatory conditions or myeloproliferative neoplasms. Many of these agents are also currently under active investigation as treatments of other haematopoietic malignancies and solid tumours. Novel inhibitors of the IL-6/JAK/STAT3 pathway, including STAT3-selective agents, are also being developed, and early phase clinical trials are currently ongoing.

Predictive biomarkers, beyond pathway hyperactivation, are needed in order to rationally incorporate IL-6/JAK/STAT3-targeting agents into multimodality treatment plans, including combinations with chemotherapy, radiotherapy, and immune-checkpoint inhibitors. In addition, as the therapeutic armamentarium of these agents increases, comparative evaluations, which are currently lacking, will be needed in both preclinical and clinical settings. The search for biomarkers that predict a response should include the evaluation of miRNAs, which are currently not widely explored. Moreover, molecular targeting or exogenous expression of specific miRNAs might prove to be a fruitful means of suppressing signalling via the IL-6/JAK/STAT3 pathway. Interestingly, a mimic of miR-34, which represses the expression of IL-6R, has entered phase I evaluations in patients with solid tumours. However, the pursuit of miRNAs as targets or therapeutic agents will be challenged by the broad effects of miRNAs on protein expression. Thus, the effects of miRNA mimics or inhibitors are unlikely to be highly selective for IL-6/JAK/STAT3 signalling pathways.

The clinical use of IL-6/JAK/STAT3-targeting agents will benefit from a deeper understanding and consideration of the genomic profile of patients’ tumours. Information regarding the identity of oncogenic drivers will also be useful in guiding the development of personalized combination therapies. For example, mutations in the PIK3CA gene are common in a wide variety of cancers and lead to activation of the PI3K signalling pathway. The PI3K pathway promotes tumour growth independently of JAK–STAT3; therefore, inhibitors of JAK–STAT3 signalling are likely to be ineffective as monotherapies in tumours with PIK3CA-activating mutations. Instead, the presence of PIK3CA mutations in concert with STAT3 hyperactivation suggests that a therapeutic regimen comprising a JAK–STAT3 inhibitor, in combination with an inhibitor of the PI3K signalling pathway, will be an effective approach.

Conclusions

In summary, targeting the IL-6/JAK/STAT3 signalling axis, which has already been shown to be beneficial in the treatment of certain cancers in human patients, holds considerable promise for the suppression of tumour growth and the restoration of antitumour immunity. The identification of predictive biomarkers and the development of rational combination therapies based on the immune and genomic profiles of tumours is required in order to maximize the broad utility and efficacies of agents targeting the IL-6/JAK/STAT3 signalling pathway and their use in precision medicine for patients with cancer.

Key points.

• The IL-6/JAK/STAT3 signalling pathway is aberrantly hyperactivated in patients with chronic inflammatory conditions and in those with haematopoietic malignancies or solid tumours

• Multiple cell types in the tumour microenvironment produce IL-6, leading to activation of JAK/STAT3 signalling in both tumour cells and tumour-infiltrating immune cells, which can promote tumour-cell proliferation, survival, invasiveness, and metastasis

• STAT3 is hyperactivated in tumour-infiltrating immune cells and acts to negatively regulate neutrophils, natural killer cells, effector T cells, and dendritic cells while positively regulating populations of myeloid-derived suppressor cells and regulatory T cells

• Targeting components of the IL-6/JAK/STAT3 signalling pathway can inhibit tumour cell growth and relieve immunosuppression in the tumour microenvironment

• Inhibitors of IL-6, the IL-6 receptor, or JAKs have all received FDA approval for various malignancies, and other novel inhibitors of the IL-6/JAK/STAT3 signalling pathway are currently in clinical and/or preclinical development

• Investigations of the efficacy of IL-6/JAK/STAT3 inhibitors, in combination with immune-checkpoint inhibitors, are warranted