The STAT3 encoding gene is located on the human genome chromosome 17 (17q21.1) and the protein contains six classical functional segments of the STAT family: i) N-terminal domain (ND), which is able to stabilize the dimerized STAT3 and promote the formation of tetramers of two STAT3 dimers to make it more stable with DNA (12); ii) coiled-coil domain (CCD), which mediates STAT3 direct binding to the receptor and facilitates STAT3 phosphorylation on 705-tyrosine site (Y705) (13); iii) DNA binding domain (DBD), which, by recognizing the γ-interferon activating sequence (GAS), will direct STAT3 to the promoters of target genes, and initiate transcriptional activation of the target genes (14); iv) the linker region, the function of which unknown at present; v) Src homology 2 (SH2) domain, the most conserved part of STAT3, which shares the same core sequence 'GTFLLRFSS' with the SH2 domain of tyrosine kinase Src, and the phosphorylation and subsequent dimerization of SH2 domain and plays a critical role in the process of signal transduction (15); vi) C-terminal transcriptional activation domain (TAD), in which there are several important tyrosines/serines located or near the region, and the phosphorylation of specific residues are important for STAT3 function. For example, Y705 is located between SH2 and TAD, which is crucial for STAT3 activation and dimerization, and the 727-serine site (S727) is located in TAD, which is thought to enhance the transcriptional activity of STAT3 (16–18) (Fig. 1A).

STAT3 has four isoforms: STAT3α, STAT3β, STAT3γ and STAT3δ. Among the structures of STAT3 isoforms, STAT3α is the most common structure. It consists of the ND, CCD, DBD, Linker, SH2 and TAD domains, with a molecular weight of approximately 92 kDa, and is mainly associated with cell proliferation and transformation (19,20) (Fig. 1A). STAT3β originates from the alternative splicing of the STAT3 gene transcript, which results in a 55 amino acid deletion at the 3' end of the open reading frame of STAT3α with a molecular weight of approximately 83 kDa; STAT3β lacks the C-terminal transactivation domain and S727, and acts as a dominant transcriptional inhibitory factor that is particularly crucial for granulocyte colony-stimulating factor (G-CSF)-mediated cell differentiation (21–23) (Fig. 1B). STAT3γ, with molecular weight of 72 kDa, acts as a dominant negative form of STAT3α. It lacks the C-terminal transactivating portion, and due to controlled proteolysis, the corresponding residues of S727 and Y705 of STAT3α are absent in STAT3γ, which retains the SH2 domain and can be recruited to tyrosine-phosphorylated receptor proteins through the SH2 domain (Fig. 1C). STAT3γ is predominantly activated in terminal differentiated neutrophils, and is involved in the regulation of cell proliferation (19,24). STAT3δ, a putative isoform of STAT3 with a yet unknown structure, has been found to be expressed during the early stage of granulocytic differentiation (25).

STAT3 can be activated through a number of mechanisms, including through the Janus kinase (JAK)/STAT3, Ras/mitogen-activated protein kinase (MAPK) and non-receptor tyrosine kinase signaling pathways (26,27). STAT3 is negatively regulated by two types of regulating factors, including the suppressor of cytokine signaling (SOCS) and the protein inhibitor of activated STAT (PIAS), which regulate the active status of STAT3 through different mechanisms (28). These activation pathways and regulatory factors regulate STAT3 activity and function synergistically, and play essential roles in physiological and pathological processes (29–31) (Fig. 2). A description of these activation pathways and regulatory factors is provided as follows:

In the process of leukemia cell proliferation, STAT3 can promote the proliferation of leukemia cells through JAK/STAT3, Ras/Raf/MAPK, PI3K/AKT/mammalian target of rapamycin (mTOR) and other signaling pathways. For example, G-CSF stimulates the activation of STAT3α to promote the proliferation of AML cell lines (47), and protein kinase CK2 regulates the transcription of the Forkhead box O3 (FOXO3a) gene by activating STAT3 through JAK/STAT3 and PI3K/AKT/mTOR, and promotes the proliferation of leukemia stem cells (48). Moreover, STAT3 can also be activated via the JAK/STAT3 and Ras/Raf/MAPK pathways, which initiates the transcription of the Forkhead box M1 (FOXM1) gene and promotes the proliferation of the CML cell line, K562 (49), wherease the inhibition of STAT3 activity can effectively reduce the proliferation and survival of leukemia cells (50–52). For example, the STAT3 inhibitor, nifuroxazide, and the PI3K inhibitor, GDC-0032, alone or in combination, have been shown to effectively reduce the proliferation of human ALL CCRF-CEM cells (50). In CD34 antigen-positive AML (AML CD34⁺) cells, the siRNA-mediated knockdown of STAT3 has been shown to significantly inhibit the growth and survival of AML CD34⁺ cells (51). In addition, it has been demonstrated that the combined use of the survivin gene inhibitor, YM155, with the STAT3 inhibitor, S3I-201, significantly inhibit the proliferation of YM155-tolerant human T lymphocytic leukemia MT-2 cells (52). These studies collectively suggest that STAT3 overexpression can promote leukemia cell proliferation and that the inhibition of STAT3 activity can reduce leukemia cell survival and proliferation (Fig. 3).

Hematopoietic progenitor cells can be divided into myeloid and lymphoid progenitor cells. Myeloid progenitor cells have the potential to multi-differentiate into myeloid cells (erythrocytosis, granulocytosis, thrombocytosis, etc.), while lymphoid progenitor cells have the potential to differentiate into lymphatic subfamilies. Leukemia cells, due to differentiation and maturity disorders, are always accompanied by abnormal hyperplasia, and are stagnated at different stages of cell development. For example, acute leukemia (AL) cell differentiation is stagnated at the relatively earlier stages, and the majority of AL cells are progenitor cells and early immature cells; chronic leukemia (CL) cell differentiation is stagnated at the later stages, and the majority of CL cells are mature naive cells or mature cells. We hereby discuss the roles of STAT3 in the differentiation of leukemia cells as follows:

The occurrence of leukemia is closely related to the apoptotic disorder of leukemia cells. STAT3 plays a key role in leukemia cell apoptosis. In general, STAT3 activation can inhibit the apoptosis of leukemia cells by inducing anti-apoptotic gene expression. For example, internal tandem duplication (ITD) in the juxtamembrane in the kinase domain of Flt3 is a common genetic lesion in AML, and the stimulation of the Flt3-ITD AML cells, MV4-11, results in the elevation of p-STAT3 levels, which upregulates the expression of the anti-apoptotic gene, survivin, to protect AML cells from apoptosis (82). In the CML cell line, K562, the expression levels of the anti-apoptotic genes, Bcl-2, Bcl-xL, Mcl-1 and survivin and those of the inhibitor of apoptosis protein, c-IAP2, have been shown to be upregulated by STAT3 activation to protect K562 cells from apoptosis (83). In CLL cells, the expression levels of the anti-apoptotic genes, Bcl-2, Bcl-xL, cyclin D1 (CCND1) and ROR1, were upregulated following the activation of STAT3, and protected CLL cells from apoptosis (84) (Fig. 6A). The inhibition of STAT3 expression can upregulate the expression of pro-apoptotic genes and induces leukemia cell apoptosis. For example, in K562 cells, the STAT3-specific inhibitor, LLL-3, has been shown to induce leukemia cell apoptosis by activating the caspase-3-and -7-mediated pro-apoptotic pathway (85). In TEL-AML1 fused ALL, the STAT3 inhibitor, S3I-201, has been shown to induce leukemia cell apoptosis by upregulating the expression of the pro-apoptotic genes, caspase-3, p27 and p21 (86). The cytotoxic agent, bromo analogue (TBr), has been shown to enhance the expression of the pro-apoptotic genes caspase-3, -8 and -9, and to decrease the ratio of the anti-apoptotic genes, Bcl-2/Bax, by downregulating p-STAT3 and upregulating p-ERK to induce the apoptosis of HL-60 cells (87). Moreover, (R)-5-hydroxy-2-methylchroman-4-one (HMC), isolated from a novel endophytic fungus, can significantly reduce the level of p-STAT3, which activates the pro-apoptotic genes, Bax, Bid, caspase-3, -8 and -9 to induce leukemia cell apoptosis (88) (Fig. 6B). However, a previous study reported that the overexpression of STAT3 in CLL upregulates caspase-3 expression, and induces CLL cell apoptosis (89), suggesting that the different activation status of STAT3 may play differential roles in the regulation of leukemia cell apoptosis.

In summary, STAT3 is closely related to the development of leukemia. It plays essential roles in the development of leukemia by promoting the proliferation of leukemia cells (Fig. 7A), regulating the differentiation of granulocytes, monocytes/macrophages and dendritic cells and blocking the apoptosis of leukemia cells (Fig. 7B and C). The STAT3 expression and activation status may thus be an important target for leukemia treatment and prognosis assessment.

STAT3 plays an important role in the diagnosis of leukemia. For example, Xia et al (90) found that in patients newly diagnosed with AML, 13 out of 17 patients were found to exhibit the constitutive phosphorylation of STAT3 on Y705. Stevens et al (91) demonstrated that the STAT3 pathway was more sensitive to ligand stimulation in patients with relapsing AML. Benekli et al reported that constitutively active STAT3 was detected in ten of 36 patients newly diagnosed with AML (92), and in another study, almost 78% of the patients (21 of 27) expressed STAT3β protein (25).

Thus, collectively, it is suggested that STAT3 plays an important diagnostic role in both newly diagnosed and relapsing AML. T-cell large granular lymphocytic leukemia (T-LGLL) sustains the phosphorylation of STAT3, and mutations of the STAT3 gene have been identified as a recurrent genetic abnormality in T-LGLL (93–99).

STAT3 single nucleotide polymorphisms (SNPs) are also prospective candidates for use in predicting susceptibility to leukemia, which is associated with the individual sensitivity of leukemia and is important for leukemia diagnosis. For example, Zhong et al (100) investigated association between STAT3 SNPs and leukemia, and found that rs17886724 located on STAT3 intron 4 was closely related to the pathogenesis of leukemia; Lautner-Csorba et al (101) found that two SNPs of rs3816769 and rs12949918 in the STAT3 gene, were associated with a decreased risk of hyperdiploid ALL occurrence (≥50 chromosomes). All these studies indicate the significance of STAT3 in the diagnosis of leukemia.

STAT3 plays an important role in leukemia, and STAT3 intervention is currently a 'hot' topic of research, which may reveal novel targets and strategies for leukemia treatment (102,103). At present, different types of STAT3 inhibitors have been implemented for leukemia treatment, including STAT3 SH2 domain inhibitors, DNA-binding domain inhibitors, STAT3 gene expression oligonucleotide inhibitors and N-terminal domain inhibitors (104). For example, small molecule C188-9, a STAT3 inhibitor, has been shown to induce the apoptosis of multiple AML cell lines and primary cells (8). The STAT3 inhibitor, OPB-31121, has also been shown to strongly inhibit STAT3 and STAT5 phosphorylation, and to exert a significant anti-proliferative effect on human leukemia cells (105). In addition, a number of chemotherapy drugs can also target STAT3 to exert a therapeutic effect. For instance, decitabine has been shown to markedly inhibit the proliferation of HL-60 cells and to enhance the cytotoxicity of natural killer cells to HL-60 cells, which may be related to the STAT3 signaling pathway (106). Zoledronic acid and bortezomib can both inhibit the proliferation of K562 cells and induce apoptosis via the STAT3 signaling pathway (107,108). Chidamide has been shown to affect AML cell viability by inhibiting the JAK2/STAT3 signaling pathway (109). Although some inhibitors have been shown to be effective in the treatment of leukemia in vitro, few inhibitors of STAT3 have been approved for leukemia therapy in clinical practice (Table II).

Moreover, many natural extracts of traditional Chinese medicine have also been shownt o inhibit the activation of STAT3 to exert antitumor effects. For instance, the traditional Chinese medicine, cucurbitacin B, has been shown to inhibit STAT3 activation along with the Raf/MEK/ERK pathway and to promote the apoptosis of K562 cells (110). It has also been demonstrated that dehydrocostus lactone significantly suppresses the proliferation of K562 cells by inhibiting STAT3 phosphorylation and the expression of downstream target genes (111). Tanshinone IIA and cryptotanshinone have been shown to induce K562 cell apoptosis by modulating JAK/STAT3/5 and SHP1/2 signaling distinctively (112). Matrine suppresses the growth of human CLL cells via its inhibition of the IL-6/JAK/STAT3 signaling (113). On the whole, traditional Chinese medicine has great potential in the treatment of leukemia. It has also been reported that certain biological therapies can also be used to target STAT3 signaling pathways for leukemia treatment. Lentivirus-mediated STAT3 silencing may inhibit the expression of its downstream genes, c-myc, Bcl-xL and cyclin D1, to suppress the malignant biological behaviors, and STAT3 silencing also inhibits the leukemogenic potency of K562 cells in mice (114). Moreover, the suppression of the survivin gene induced by the BCR/ABL/JAK2/STAT3 pathway sensitizes imatinib-resistant CML cells to different cytotoxic drugs (115). Thus, the role of biological therapy in the treatment of leukemia cannot be ignored.

The roles of STAT3 in the prognosis of leukemia have drawn increasing attentions, but are still controversial. Recent studies have found that Mcl-1 gene expression and indoleamine 2,3-dioxygenase 1 (IDO1) activity can both be regulated by STAT3 which are associated with the poorer prognosis of leukemia (116,117). Bruserud et al (103) examined the association between STAT3 expression in bone marrow and the prognosis of patients with AML, and found that STAT3 expression and Y705 activation had an adverse prognostic impact on human AML. Moreover, Adamaki et al (118) discovered that 8 out of 14 childhood patients with ALL, in which STAT1 and STAT3 were expressed at lower levels on day 33, were relapse-free with a high survival rate, indicating that STAT3 gene upregulation is a poor predictor of the clinical prognosis of childhood ALL. Furthermore, Zhong et al (119) investigated the association between genomic polymorphisms of STAT3 and AML patient response to chemotherapy in a Chinese population, and found that there were strong associations between unfavorable cytogenetics, partial remission (and even no remission) and GG genotype frequency in rs9909659, which suggested that the STAT3 GG genotype in rs9909659 may confer an enhanced resistance to standard chemotherapy, and predict an unfavorable prognosis. Besides, it has also been reported that constitutive STAT3 activation is associated with the shortest disease-free survival in AML, and the subgroup of AML patients, which expressed STAT3β and showed constitutive activity of STAT3, experienced a particularly poor outcome (92).

By contrast, some studies have suggested that p-STAT3 also has a beneficial effect on the prognosis of leukemia. Redell et al (8,120) discovered that when the phosphorylated level of STAT3 on Y705 and S727 increased in AML upon cytokine stimulation, the disease-free and survival rates of patients were improved. Levidou et al (121) discovered that the immunoreactivity of tyrosine phosphorylated STAT3 was marginally associated with a prolonged overall survival. In addition, the upregulation of tyrosine phosphorylated STAT3 was associated with a longer time to first treatment, and the absence of p-STAT3 was associated with a shorter time to progression, which indicated that p-STAT3 also has a favorable prognostic effect on leukemia. Nevertheless, the upregulation of STAT3 is also associated with the favorable prognosis of AL (122). Collectively, the impact of STAT3 on the prognosis of leukemia remains controversial. Due to the defects of current clinical studies, including small sample sizes and so on, extensive clinical studies are warranted in order to further confirm the role of STAT3 in the evaluation of the prognosis of leukemia.