STAT proteins: From normal control of cellular events to tumorigenesis

Abstract

Signal transducers and activators of transcription (STAT) proteins comprise a family of transcription factors latent in the cytoplasm that participate in normal cellular events, such as differentiation, proliferation, cell survival, apoptosis, and angiogenesis following cytokine, growth factor, and hormone signaling. STATs are activated by tyrosine phosphorylation, which is normally a transient and tightly regulates process. Nevertheless, several constitutively activated STATs have been observed in a wide number of human cancer cell lines and primary tumors, including blood malignancies and solid neoplasias. STATs can be divided into two groups according to their specific functions. One is made up of STAT2, STAT4, and STAT6, which are activated by a small number of cytokines and play a distinct role in the development of T-cells and in IFNγ signaling. The other group includes STAT1, STAT3, and STAT5, activated in different tissues by means of a series of ligands and involved in IFN signaling, development of the mammary gland, response to GH, and embriogenesis. This latter group of STATS plays an important role in controlling cell-cycle progression and apoptosis and thus contributes to oncogenesis. Although an increased expression of STAT1 has been observed in many human neoplasias, this molecule can be considered a potential tumor suppressor, since it plays an important role in growth arrest and in promoting apoptosis. On the other hand, STAT3 and 5 are considered as oncogenes, since they bring about the activation of cyclin D1, c-Myc, and bcl-xl expression, and are involved in promoting cell-cycle progression, cellular transformation, and in preventing apoptosis. J. Cell. Physiol. 197: 157–168, 2003© 2003 Wiley-Liss, Inc.

Signal transducers and activators of transcription (STAT) proteins comprise a family of transcription factors latent in the cytoplasm that consists of seven different members: STAT1, 2, 3, 4, 5A, 5B, and 6 (Darnell, 1997).

These transcription factors are activated by a series of extracellular signaling proteins such as cytokine, growth factors, and hormones that bind to specific cell-surface receptors. The resulting signal transduction pathways permit them to play different roles in normal physiological cell processes, such as differentiation, proliferation, apoptosis, and angiogenesis (Horvath, 2000).

However, aberrant activation of STAT-signaling gives rise to different pathological events, for example, cell transformation and oncogenesis. Although they neither directly regulate cell-cycle checkpoints nor contribute to the repair of DNA damage, they take part in tumorigenesis by means of the deregulation of the signal pathways in which they are implicated (Bowman et al., 2000).

STATs can be divided into two groups according to their specific functions. One is made up of STAT2, STAT4, and STAT6, which are activated by a small number of cytokines and play a distinct role in the development of T-cells and in IFNγ signaling. The other group includes STAT1, STAT3, and STAT5, activated in different tissues by means of a series of ligands, and involved in IFN signaling, development of mammary gland and response to GH, and embriogenesis, respectively. This latter group of STATS plays an important role in controlling cell-cycle progression and apoptosis and thus contributes to oncogenesis (Bromberg, 2002).

In this review we summarize the mechanisms of STATs activation, the signal pathways deriving from them and the mechanisms influencing their activity. Furthermore, we analyze the direct involvement of these proteins in oncogenesis, and discuss on their role as possible molecular targets for the development of new anticancer treatments.

ACTIVATION MECHANISMS OF STAT PROTEINS

STAT proteins become activated by means of a variety of soluble factors, such as cytokines, growth factors, and hormones (Fig. 1).

Figure 1. Activation mechanisms of STAT proteins: IFNα and IFNγ signaling (a) and other transductional pathways (b).

Cytokines comprise a large number of factors regulating cell growth, development, and immune responses and can be divided into two main groups. Type I cytokines include interleukins (IL2-7, IL9, IL11-13, and IL15), colony-stimulating factors (CSFs), neutrophic factors, and hormones; they bind to transmembrane receptor proteins containing four conserved cys residues, an extracellular trp-ser-X-trp-ser motive, and variable intracellular domains. Type II cytokines include the interferons (IFNs) and IL10; they bind to receptors with four cys residues, but without extracellular domains. Some receptors include only one subunit which homodimerizes at the moment of binding with the ligand, while others are made up of two, or, less frequently, three different oligomerizing subunits (Chen et al., 1999). Cytokine receptors, unlike those of the tyrosine kinase (TK) family, have no intrinsic enzymatic activity, but determine STAT phosphorylation in tyrosine by means of the proteins belonging to the Janus tyrosine kinase family (JAK) which they are constitutively associated with (Takeda and Akira, 2000).

The JAK family of receptor-associated tyrosine kinases consists of four members: JAK1, JAK2, JAK3, and TYK2, which are activated by receptors dimerization or oligomerization in a partly specific manner, in fact different cytokines can activate the same JAKs (Schindler, 2002). Nevertheless, JAKs may be activated through the influence of other kinases. Once it has become activated, the JAK kinase protein causes tyrosine phosphorylation of the receptor cytoplasmic tails to provide docking sites for the recruitment of molecules recognizing phosphotyrosine via their phosphotyrosine binding domain (PTB) or SRC-homology-2 (SH2) domain (Briscole et al., 1996). These molecules bound to the receptor are subsequently phosphorylated, thus becoming activated. The STATs have an SH2 domain through which they make contact with the receptor, dimerize and move into the nucleus (Bromberg and Darnell, 2000) (Fig. 1a).

STATs proteins are activated also by growth factor receptors, such as epidermal growth factor (EGF), hepatocyte growth factor (HGF), platened-derived growth factor (PDGF), and colony-stimulating factor-1 (CSF-1) receptors, which all possess an intrinsic tyrosine kinase activity (Catlett-Falcone et al., 1999) (Fig. 1b). These receptors may activate STAT proteins either indirectly, by means of JAK kinase proteins, or directly, as in the case of STAT1 activation by PDGF or EGF receptor (Fu and Zhang, 1993), which may phosphorylate STAT1 in vitro (Quelle et al., 1995; Choudhury et al., 1998).

Moreover, many receptors can activate the same STAT substrate by phosphorylating the tyrosine residue and may involve different JAKs. STATs activation specificity within the different cell lines depends on the specific interactions existing between the STAT proteins and their respective receptors (Briscole et al., 1996).

STAT proteins may also be activated by the non-receptor tyrosine kinases SRC and ABL (Fig. 1b).

STAT3 protein has been found associated in vivo and in vitro with SRC and this later phosphorylates STAT3 in vitro (Danial et al., 1995). Mammal cells transformed by oncogenic SRC show constitutively active STAT3 and negative-dominant forms of STAT3 block the transforming ability of SRC, demonstrating a close correlation between STAT3 activation and the oncogenic transformation of SRC (Turkson et al., 1998). Moreover, recent studies have shown that the constitutive activation of STAT3 in human breast cancer cells correlates with EGF receptors family kinase signaling and also with aberrant JAK and c-SRC activity (Garcia et al., 1997). In addition to v-SRC, many other transforming tyrosine kinases, such as v-Eyk, v-Ros, v-Fps, Etk/BMX, and Lck, constitutively activate STAT3 in the context of oncogenesis (Bowman et al., 2000).

Another example of tumorigenic stimuli known to activate STAT proteins is v-Abl (Briscole et al., 1996). v-Abl may constitutively activate STAT3 and STAT5, with direct involvement of the JAK kinases, whereas the fusion protein, BCR-Abl, may activate them in the absence of constitutive JAK activation, showing that the presence of the JAK kinases is not always essential for STAT activation by Abl (Bowman et al., 2000).

Besides the cytokine and growth factor receptors, several studies have also shown that the seven transmembrane pass (serpentine) receptors, such as those for angiotensin II, serotonin, and α-melatonin stimulating hormone (α-MSH), are able to activate STATs. This activation, however, appears to be independent of the G-proteins, which normally mediate the signal of such receptors (Williams, 1999) (Fig. 1b).

STRUCTURE OF STATS

The human STATs genes have been identified in three chromosomal clusters: STAT 1 and STAT 4 map on chromosome 2, STAT 2 and STAT 6 on chromosome 12, and STAT 3 and STAT 5a and 5b on chromosome 17 (Ihle, 2001). Despite functional differences of individual STAT proteins, crystallographic studies and protein sequence comparisons of STAT1, STAT3, and STAT4 revealed common STAT structural features (Becker et al., 1998; Chen et al., 1998).

STAT proteins are made up of about 850 amino acids, except for STAT 2 and STAT6, which have amino acids between 750 and 800 and range in size from 90–115 kDa. All STATs share several conserved domains, which are imperative for each specific function (Takeda and Akira, 2000) (Fig. 2).

Figure 2. Structure of STAT proteins with the most important conserved domains involved in specific functions. Y and S are phosphorylation sites in tyrosine (Y) and in serine (S).

DNA-binding domain, sited in the region between amino acids 320 and 475, is structurally very similar to the immunoglobin-like DNA-binding domain. Moreover, a DNA-binding fold between residues 320 and 490 contains several β-sheets that are folded similarly to those found in the DNA-binding domains of the transcription factors NF-kB or p53 (Levy and Darnell, 2002). This binding domain determines the DNA-binding specificity for each STAT protein (Horvath, 2000).

Domain SH2, sited in the region between the amino acid residues 600 and 700, is required for the recruitment of STATs to phosphorylated receptors and for the reciprocal SH2-phosphotyrosine interactions between monomeric STATs to form dimers (Shuai, 1999). STAT-binding to the receptor occurs through the interaction of the STAT SH2 domain with the phosphorylated tyrosine present in the receptor docking site. The differences in the STAT SH2 domain bring about selectivity of the STAT protein-binding to the different cytokine receptors (Becker et al., 1998).

The critical tyrosine residue (Tyr 701 for STAT1, Tyr 690 for STAT2, Tyr 705 for STAT3, Tyr 693 for STAT4, Tyr 694 for STAT5, and Tyr 641 for STAT6) required for SH-phosphotyrosine interaction and thus STAT activation is located near the SH2 domains. This tyrosine is rapidly phosphorylated by the active JAK determining STAT dimerization, which occurs by means of the binding of the SH2 domain of one molecule with the domain containing the phosphotyrosine of another molecule (Chatterjee-Kishore et al., 2000). The resulting dimers are thus stabilized by bivalent bonds (Chen et al., 1998). STAT2 is the only one that does not act as a homodimer, forming instead a complex with STAT1 and p48. As a response to several cytokines, the heterodimers STAT1-2 and STAT1-3 are formed, while no heterodimers with STAT 4, STAT 5a/b, and STAT6 have been identified, with the exception of STAT5a, which heterodimerizes with STAT5b (Horvath, 2000).

The phosphorylation of STATs into tyrosine residue is essential not only for dimerization, but also for the concomitant translocation of the dimers into the nucleus. Binding of STAT1 to importin-α5, one of the subunits of the nucleocytoplasmic transport machinery, has been described, and recent studies indicated that L407, K410, and K413 in STAT1 are critical residues in the nuclear import of tyrosine-phosphorylated STAT1 (McBride et al., 2002). Moreover, a region essential for nuclear import has been identified in STAT5b (Williams, 2000).

An alfa-helical linker domain (residues 488–576) bridges the DNA-binding and SH2 (Chen et al., 1998).

The transcriptionalactivation domain (TAD) at the C-terminal region of STATs between residues 661 and 851 is involved in communication with transcriptional complexes. A conserved serine in this STAT domain (apart from STAT 2 and STAT 6, which have no such serine), is a phosphorylation site and regulates STAT transcriptional activity. STAT 1 and 3 with this altered serine have a transcriptional capacity reduced by 20% (Imada and Leonard, 2000). One of the kinases responsible for the phosphorylation of this serine belongs to the MAP kinases family (ERKs and p38), which emphasizes the important “cross-talk” occurring between the two transductional pathways (David et al., 1995; Goh et al., 1999). Therefore, STATs can be phosphorylated in great many serine/threonine residues, which may modulate DNA-binding and/or transcriptional activity (Briscole et al., 1996).

The NH2-terminal region of STAT is highly conserved, provides protein–protein interaction sites and is required for the dimer–dimer interactions to form tetramers or oligomers STAT molecules. The tetramerization of STATs contributes to stabilize the STAT-DNA binding by means of the interaction with tandemly arranged low-affinity binding sites, thus increasing transcriptional activity (John et al., 1999).

The region of STATs between residues 130 and 315 consists of a four-stranded helical coiled coil domain. This domain associates with a number of potentially important regulatory modifiers, including IRF-9 and StIP1 (Collum et al., 2000; Horvath, 2000).

IFN-SIGNALING

The main features of the IFN-system, which have also provided the transduction pathway model for other cytokines, might be oversimplified as followss.

(a) Transduction pathways for the IFNγ signal: The class II cytokine receptor for IFNγ consists of two dissociated polypeptide chains without any ligand. The α-receptor subunit is mostly concerned with cytokine-binding, while the β is needed for signal transduction, with a secondary role in ligand interaction. The different receptor subunits are associated with inactive tyrosine-kinase JAKs. The α-chain is associated with JAK1 and the β subunit with JAK2. The binding between IFNγ and its receptor induces the dimerization of the two subunits α and β and the activation of the JAKs associated with them. Each dimerization juxtaposes the JAKS associated with each receptor chain, bringing about their activation (JAK1 and JAK2) by means of mutual phosphorylation. Nevertheless, several other kinases may have an effect on JAK activation (Stark et al., 1998; Chatterjee-Kishore et al., 2000).

Once activated, JAKs phosphorylate the α-receptor subunit within a specific tyrosine residue and generate the STAT docking site. In the case of the signal transduction triggered off by the interaction of IFNγ with its receptor, the member of the STAT family involved is STAT1. In fact, STAT1 binds with the α-subunit by means of the interaction of its SH2 domain with the phosphorylated tyrosine residue present in the receptor (Darnell, 1997).

(b) While associated with the α-receptor subunits, the STAT1s are phosphorylated by the JAKs within a conserved tyrosine residue (tyrosine 701). At this point, the STATs dissociate from the receptor subunit and dimerize. Each STAT1 dimer is further phosphorylated, this time in serine, which makes it an active transcription factor. The dimer translocates into the nucleus and induces the expression of target genes, which contain a regulatory sequence termed GAS (IFNγ activation site) in the promoter region (Ramana et al., 2000) (Fig. 1a).

Following IFNγ stimulation, a homodimer of STAT1β is also produced, and this, unlike the normal molecule (STAT1α), is lacking in the last 38 amino acid residues (Horvath, 2000). This homodimer does not apparently induce transcription, suggesting that sites necessary for the interaction with the RNA polymerase complex or transcription coactivators might be present in the lacking region.

(c) Transduction pathways for the IFNα signal: The class II cytokine receptor for IFNα consists of two dissociated polypeptide chains without ligand. These two chains are IFNAR1 and IFNAR2. The different receptor subunits are associated with inactive tyrosine-kinase JAKs. The IFNAR1 chain is associated with Tyk2 and the IFNAR2 chain with JAK1.

The binding between IFNα and its receptor induces the dimerization of the two subunits and the activation of the JAKs associated with them. Following binding with IFNα, the two receptor chains become associated, leading to the cross-phosphorylation of Tyk2 and of JAK1. Once activated, JAKs phosphorylate the IFNAR2 receptor subunit, thus generating the docking site for STAT2 and phosphorylate STAT2 in order to favor STAT1 recruitment. In fact, IFNAR2 forms a complex with STAT2 and STAT1. Nevertheless, IFNAR2 phosphorylation only generates the docking site for STAT2; the subsequent phosphorylation of STAT2 in tyrosine 690, once again mediated by JAKs, will favor binding with STAT1. The binding of STAT1 to IFNAR2 depends on STAT2, but not vice versa (Durbin et al., 1996; Meraz et al., 1996).

(d) Once it has been recruited into the receptor by STAT2, STAT1 is itself then phosphorylated by the JAKs in tyrosine 701, thus permitting the release of the heterodimer STAT1/STAT2, which associates with the p48 nuclear factor and forms the ISGF3 factor (IFN-stimulated gene factor) (Horvath et al., 1996). This complex activates specific target genes within the nucleus recognizing promoter sequences called IFN-stimulated response elements (IRSE) (Li et al., 1998) (Fig. 1a).

The two variants of STAT1, STAT1α, and STAT1β are frequently interchangeable with regard to ISGF3 activity in response to IFNα stimulation (Horvath, 2000). Furthermore, the absence of STAT2 is accompanied by a reduced tyrosine-phosphorylation and STAT1 activation, confirming the predominant role of STAT2 in the recruitment and activation of the complex. In response to IFNα, as an alternative, STAT1 may form heterodimers with STAT3 or else homodimers STAT1 and STAT2, which however lead to the activation of GAS elements (Chatterjee-Kishore, 2000b).

NEGATIVE REGULATION OF STATS

The duration of STATs activation is a temporary process, thus within hours the activating signals decay and the STATs are exported back to the cytoplam. Nevertheless, the mechanisms by which STAT signals decay are poorly understood, although some important regulators of this aspect of the pathway have been defined. These include members of the SOCS family of counter-regulatory proteins (Kile et al., 2001; Yamada et al., 2003), as well as, phosphatases and enzymes that mediate covalent modification of STAT proteins (Kisseleva et al., 2002). Such modifications include not only ubiquitination, which targets some STATs for proteolysis (Wang et al., 2000), but also serine phosphorylation (Decker and Kovarik, 2000), and potentially acetylation, arginine methylation, and SUMO-ylation (Sachdev et al., 2001), all potential mechanisms for controlling the duration and consequences of STAT signaling.

Negative nuclear regulators

The STATs dephosphorylation by nuclear tyrosine phosphatase occurs in the nucleus and is an important signal for export back to the cytoplam. There is evidence that TC45, a nuclear tyrosine phosphatase, is a relevant STAT phosphatase for STAT1 and STAT3 (Ibarra-Sanchez et al., 2000). In addition, it has been reported that cells lacking of this enzyme retain tyrosine-phosphorilated STAT1 for much longer than normal cells, and overexpression of TC45 leads to dephosphorilation of STAT5 (Aoki and Matsuda, 2002; ten Hoeve et al., 2002).

However, TC45 has also implicated in regulating cytoplasmic dephosphorilation of JAK1 and JAK3 (Simononic et al., 2002).

Recently, the negative activity on STAT protein of a group of nuclear proteins termed protein that inhibit activated STATs (PIAS) has been discovered. Studies in cultured mammalian cells indicated that PIAS1 and PIAS3 interact only with tyrosine-phosphorilated STAT1 and STAT3, respectively (Chung et al., 1997; Shuai, 2000). In addition, naturally occurring short forms of STATs can potentially act as dominant-negative proteins by occupying DNA as non-functional protein or by binding to wild-type STATs (Yu et al., 2000).

Negative cytoplasmatic regulators

Besides the aforementioned TC45, other phosphatases, such as SH2-containig phosphatase-1 (SH1), SH2, and protein-tyrosine-phosphatase-1B (PTP1B) have been implicated as cytoplasmic regulators of JAKs or STATs phosphorylation (You et al., 1999; Aoki and Matsuda, 2002).

The activity of STAT proteins is also regulated by the inhibitors of the suppressors of the cytokine signal (SOCS) family, responsible for modulating the JAK-STAT pathway by acting on the JAK kinases. These cytokine-induced proteins SOC are also known as JABs (JAK-binding proteins) or cytokine-inducible SH2 (CIS) proteins or STAT-induced STAT inhibitors (SSI) (Shuai, 1999; Dell'Albani et al., 2003).

The SOCS proteins are recruited to active receptor complexes to cause inhibition, and can also cause protein turnover of the receptor through a process of proteolytic degradation ubiquitine-proteasome mediated (Krebs and Hilton, 2001).

Moreover, some of the SOCS proteins are transcriptionally regulated by STATs themselves, suggesting that STATs can negatively regulate their own phosphorilation state (Starr and Hilton, 1999).

PHYSIOLOGICAL ROLE OF STATS

STAT1 is activated both by IFNα/β and by IFNγ, which play an important role in the activation of the macrophages and in the defense responses to pathogenic agents, it has been seen that STAT1 deficient mice are much more sensitive towards viral and microbic infections (Durbin et al., 1996).

STAT2 is the only member of STATs family, which does not bind GAS elements as homodimers. Homozygous deletions of STAT2 are accompanied by embryonic mortality suggesting an important role in normal development (Bowman et al., 2000).

It appears that IL-12 and IL-4 specifically activate STAT4 and STAT6, respectively (Hou et al., 1994; Thierfelder et al., 1996). IL-12 is implicated in the development of type I T-helper cells and also in the production of IFNγ. Very little is known about the transcriptional activity of STAT4, although it seems clear that it interacts with c-Jun, giving rise to transcriptional activation starting from specific promoters. STAT6 is directly responsible for IL-4 induction, leading to the differentiation of the type II T-helper cells. Moreover, STAT6 induces the transcriptional activation of IgEs. STAT4 and STAT6 are therefore involved in maintaining the equilibrium of the TH1 and TH2 responses; alterations in these proteins cause severe immune disorders (autoimmune diseases for TH1 and allergic diseases for TH2) (Takeda and Akira, 2000).

STAT5s are activated by a series of cytokines, which include prolactin, growth hormone, erythropoietin (Epo), thrombopoietin (Tpo), granulocyte-macrophage colony-stimulating factor (GM-CSF), IL-2, IL-3, IL-5, IL-7, IL-9, and IL-15, and are involved in a great many basic functions regarding cell growth regulation (Takeda and Akira, 2000). STAT5a is activated by prolactin and is responsible for the development of the mammary gland and lactogenesis (Liu et al., 1997). STAT5b is activated by the growth hormone and regulates sexual dimorphism (Udy et al., 1997). Stat5a/b knockout female mice are infertile because of altered corpus luteum development, suggesting that the function of the STAT5 proteins may be redundant (Teglund et al., 1998). STAT5b is also involved in the development of NK cells after stimulation with IL-15; Stat5a/b deficient mice present cytotoxic activity alterations in these cells, although it is more evident in mice lacking STAT5b (Imada et al., 1998).

STAT5 might induce c-myc transcription during IL-2-induced lymphocyte proliferation (Lord et al., 2000) and regulate the expression of cyclic D1 in hemopoietic cells, thus intervening in the growth control of these cells (Matsumara et al., 1999).

It is difficult to define the exact function of STAT3s in knockout mice, since their embryos die early during embryogenesis before reaching the gastrulation stage (Takeda et al., 1997). STAT3 might be extremely important during early embryogenesis for the development of the visceral endoderm (Akira, 2000).

Furthermore, the tissue-specific loss of STAT3 has made it possible to investigate its different functions (induction of cell-proliferation, apoptosis, survival) in the various cell lineages. STAT3 might be critical for the inactivation of the macrophages and the inhibition of inflammatory cytokines in the T-helper cell response.

Since IL-10 induces the anti-inflammatory response of the macrophages and of the neutrophils by activating STAT3, STAT3 mutant mice do not respond to IL-10 (Takeda et al., 1997).

In vitro studies have been suggested that STAT3 plays an important role in migration of skin epidermic cells, and is essential for skin remodeling (for example, during the hair cycle and wound-healing processes) (Sano et al., 1999).

Furthermore, STAT3 is involved in the involution of the post-lactating mammary gland (Chapman et al., 1999); this is an apoptotic process involving the epithelial cells and resulting from an increased level of the insulin-like growth factor binding protein-5 (IGFBP5). STAT3-mutant mice do not present cell apoptosis linked to IGFBP5, which indicates that STAT3 plays a role in controlling the level of this protein (Horvath, 2000; Imada and Leonard, 2000; Takeda and Akira, 2000; Ihle, 2001). These results suggest that both STAT3 and STAT5a are activated during the development of the mammary gland, but whereas STAT5a is inhibited during involution, STAT3 is activated at the beginning.

Since, compared to the other members of their family, STAT1, STAT5, and even more, STAT3, play a great many roles in several cell contexts, they would seem to have more importance than the others in the pathogenesis of different tumors, as we will see below (Levy and Lee, 2002).

STATS IN TUMORIGENESIS

Constitutive activation of several STATs has been observed in a wide number of human cell lines and primary tumors, including blood malignancies (leukemia, lymphoma, multiple myeloma) and solid neoplasias (head and neck, brain, breast, lung, pancreas, prostate cancers) (Catlett-Falcone et al., 1999; Bowman et al., 2000; Coffer et al., 2000; Huang et al., 2000; Lin et al., 2000; Song and Grandis, 2000; Garcia et al., 2001; Bromberg, 2002; Schaefer et al., 2002) (Table 1).

Table 1. Main features of STATS and their constitutive activation in primary tumors

Family member	Human chromosomal localization	Cytokine activating	Phosphorilations site	Physiological functions	Human primary tumors and tumors cell lines in which STATS are activated
STAT1	2q12-33	IFNs, IL-6	TYR 701, SER 727	Response to types I and II IFNs, IL-2, IL-6, EGF, PDGF, GM-CSF	Breast cancer, head and neck cancer, lung cancer, brain tumors
STAT2	12q13-14,1	IFNs	TYR 689	Response to type I IFN
STAT3	17q11,2-22	IL-6	TYR 705, SER 727	Embriogenesis, skin remodelling, involution of the post-lactating mammmary gland	Breast cancer, head and neck cancer, prostate cancer, melanoma, ovarian cancer, lung cancer, brain tumors, pancreatic cancer, renal carcinoma, acute lymphoblastic leukemia, erythroleukemia
STAT4	2q12-33	IL-12, IFN	TYR 722, SER 722	Response to IL-12 for the development of type I T-helper cells
STAT5a	17q11,2-22	PRL, GH, EPO	TYR 694, SER 726	Development of the mammary gland and lactogenesis	Chronic myelogenous leukemia, acute myeloid leukemia, acute lymphoblastic leukemia, erythroleukemia
STAT5b	17q11,2-22	PRL, GH, EPO	TYR 694, SER 731	Response to the GH, regulation of sexual dimorphism and development of NK cells	Chronic myelogenous leukemia, acute myeloid leukemia, acute lymphoblastic leukemia
STAT6	12q13-14,1	IL-4, IL-13	TYR 641	Response to IL-4 and IL-13 for the development of type II T-helper cells	Breast cancer

Constitutively activated STAT3 and STAT5 have been shown to possess transforming properties, and to be strongly associated with tumor development and progression (Bowman et al., 2000). Many studies suggested that STAT proteins could participate in oncogenesis through upregulation of genes encoding apoptosis inhibitors (Mcl-1, Bcl-x), cell-cycle regulators (cyclins D1/D2, c-Myc), and inducers of angiogenesis (VEGF) (Sinibaldi et al., 2000; Martino et al., 2001).

STAT1 may be considered as a tumor suppressor, since its activation is associated with growth arrest (Durbin et al., 1996).

Furthermore, the deregulation of the activity of STAT proteins in the immunitary system might contribute to tumor development by bringing about immune surveillance loss (Bromberg, 2002).

STAT5a/b is constitutively activated in a great many neoplasias, particularly in leukemias and lymphomas. Its constitutive activation may also be triggered off by the expression of fusion proteins causing persistent PTK activation, such as JAK2, PDGF-R, or ABL (Shuai et al., 1996; Lin et al., 2000).

However, leukemias induced by BCR-ABL and v-BCR does not require the presence of STAT5 in mice, since probably the latter does not have a critical role in the normal development of the hematopoietic system, although it does have some importance in the proliferative control of myeloid-line cells (Schwaller et al., 2000). In fact, both STAT5 and STAT3 have proved to be active in acute myeloid leukemia (Spiekerman et al., 2002). The expression of the oncogenic fusion protein Tel-JAK2 seems to lead rapidly to acute lymphocytic leukemia; in fact, activated STAT5a/b also causes the proliferation of peripheral T-cells (Carron et al., 2000).

Since STAT5 regulates the genic transcription of cyclins D1/D2 (Bowman et al., 2001) and of c-myc in some cell types (Lord et al., 2000), the constitutive activation of STAT5a/b, probably promotes tumorigenesis by deregulating the cyclin complexes D/CDK4-6, which control progression from the G1 to the S-phase of the cell cycle (Bowman et al., 2000), and by deregulation of c-myc-dependent cell growth.

Furthermore, the constitutive activation of STAT5a/b brings about the activation of antiapoptotic signals (Buettner et al., 2002). In response to various cytokines (IL-3 or EPO), STAT5 activation induces Bcl-xl expression, thus determining cell survival. (Dumon et al., 1999; Silva et al., 1999).

Mice that are nullizygous for STAT5a and STAT5b, as a result of the reduced expression of Bcl-xl, present fetal anemia resulting from the intense apoptototic activity of the erythroid cell precursors (Socolovsky et al., 1999). STAT5 also regulates Mcl-l in CML cells (Decker and Kovarik, 1999). In addition, constitutive mutant STAT5s promote the expression of bcl-x and pim-1, which suggests that STAT5 may contribute to factor-independent growth and survival (Nosaka et al., 1999). STAT5 is also implicated in Bcl-x induction following transformation induced by BCR-ABL. In fact, the use of negative dominants of STAT5a/b may lead to a reduction of Bcl-x levels, thus suppressing cell growth (de Groot et al., 1999).

STAT3 has been found to be constitutively activated in a large number of solid tumors and in cell lines transformed with v-Abl and v-SRC (Garcia et al., 1997; Catlett-Falcone et al., 1999; Coffer et al., 2000; Song and Grandis, 2000). So far, no STAT3 mutations inducing constitutive activation and, therefore, cell transformation have been identified. Nevertheless, dimerizable engineered mutant forms called STAT3C, which require no phosphorylation in tyrosine by a TK in order to become active, are able to migrate to the nucleus, guide transcription, and induce cell transformation, so that STAT3 assumes the role of a true protoncogene (Bromberg et al., 1999).

STAT3 activation depends mainly on a deregulation of the TK receptors or of their associated JAKs. In acute lymphocytic leukemias, in fact, aberrant JAK activation due to the expression of the oncogenic fusion protein TEL-JAK2 might also induce simultaneously with STAT5 and STAT1 the constitutive activation of STAT3, which promotes factor-dependent proliferation (Scwaller et al., 1998). Moreover, the introduction of wild-type STAT3 leads to an increase in the transforming potential of v-SRC, while the use of negative-dominant mutants of STAT3 reduces it, suggesting that STAT3 is required by v-SRC for cell transformation (Turkson et al., 1998). In addition, it has been demonstrated that STAT3 is activated by v-SRC in many cell lines (derived from gall bladder carcinomas or from breast epithelial cells) (Murakami et al., 1998; Smith and Crompton, 1998) also together with cooperation from JAK1 (Zhang et al., 2000).

On the contrary, the coexpression of dominant negatives of STAT3 and of the oncoprotein Ras, which appears not to induce STAT3, does not arrest Ras-induced transformation. These data suggest that STAT-signaling is only one of the pathways required for cell transformation induced by the oncogenic TK (Garcia et al., 1997).

Apart from v-SRC, there exist several other transforming protein kinases which can induce STAT activation during oncogenesis, such as v-Sis, v-Fps, v-Ros, v-Eyc, the insulin-like growth factor 1 (IGF-1) receptor, Etk/BMX, and Lck (Bowman et al., 2000).

In breast tumors, constitutive activation of STAT3 is associated with the induction of the expression and/or the activity of the EGF receptor family kinases (HER1/erbB-1, HER2/neu) and of SRC, due mainly to an aberrant expression of EGF-related ligands. STAT3 negative dominant expression in these cells brings about cell growth arrest and apoptosis (Garcia et al., 2001). Furthermore, once again in breast carcinoma cells, the coexpression of synergistically activated c-SRC and STAT3 induces the activation of HGF transcription which confers increased survival and growth during progression and metastasis (Hung and Elliott, 2001).

In acute myelogenous leukemia (AML) and in gastrointestinal stromal cell tumors (GISTs), cell transformation, associated with mutations activating the c-kit receptor, is blocked by the expression of negative-dominant STAT3 (Ning et al., 2001), while in thyroid tumors the transformation induced by the aberrant regulation of the Ret receptor tyrosine kinase is associated with the phosphorylation and activation of STAT3 (Schuringa et al., 2001).

As with STAT5, the constitutive activation of STAT3 brings about cell proliferation on the one hand and the activation of antiapoptotic signals on the other. STAT3 plays an important role in the G1-S of the cell-cycle transition, since it upregulates cyclins D (D1, 2, 3) and A, cdc25A, and downregulates p21 and p27 (Fukada et al., 1998).

STAT3 activation in cells transformed by v-SRC results in the expression of cyclin D1 promoter-based reporter, which is blocked in the presence of STAT3 dominant-negative protein (de Groot et al., 2000). Mutant forms of STAT3C, however, result in overexpressed cyclin D1 expression and are associated by cell transformation (Bromberg et al., 1999). Moreover, STAT3 activation, induced by v-SRC, results in the induction of p21/waf1 and growth arrest, whereas the coexpression of STAT3 dominant-negative protein blocks this activation (Sinibaldi et al., 2000). The overexpression of cyclin D and p21, therefore, appears to play an important part in the proliferation of the SRC-transformed cells. The oncogenic potential of STAT3 also derives from the fact that it is able to promote c-myc expression both during SRC transformation and following PDGF-induced mitogenesis (Bowman et al., 2001).

Several studies have shown that STAT3 regulates the expression of some antiapoptotic proteins, such as Bcl-xl and Mcl-1 during oncogenesis. The disruption of STAT3-signaling in multiple myeloma cells or in head and neck squamous cell carcinoma is, in fact, associated by a downregulation of Bcl-xl and by the induction of apoptosis (Catlett-Falcone et al., 1999; Buettner et al., 2002). Mcl-1 expression might depend on STAT3 in multiple myelomas and in some forms of leukemia (Epling-Burnette et al., 2001; Buettner et al., 2002).

STAT3 is also involved in tumoral progression through the activation of proangiogenic factors, such as VEGF (Niu et al., 2002).

STAT1 is considered a tumor suppressor, since it inhibits growth and acts as a proapoptotic factor.

It is required for the response signals to interferons, but STAT1 knockout mice are not lacking in all the IFN responses, since several of these do not depend on STAT proteins. Such mice do not develop spontaneous tumors, but are highly susceptible to develop cancer inducted by chemical carcinogens. Since STAT1 and p53 null mice show a more frequent and rapid tumor development, it is clear that STAT1 is implicated in mechanisms of tumor surveillance by the host and that the loss of STAT1-signaling enhances oncogenesis (Kaplan et al., 1998).

The involvement of STAT1 in growth arrest and apoptosis in many cell types may be explained by its capacity to induce caspase and p21/waf1 expression (Chin et al., 1996; Kumar et al., 1997).

The promoter region of p21 contains, in fact, two binding sites for STAT1 and one for STAT1 and STAT3 (Chin et al., 1996).

Normally, high p21 expression is associated with cell growth arrest, but a similar increase has also been observed in many human neoplasias. This contradiction has been explained by Bowman et al. (2000) with the fact that p21 is also responsible for the correct assemblage of the D1/CDK cyclin complex, and thus its increase may be necessary for cell-cycle progression.

In breast carcinomas, p21 activation by STAT1 may also involve BRCA1, which is the most important susceptibility gene in hereditary breast carcinoma. In fact, BRCA1 and STAT1 contribute synergically to p21/waf1 transcription by means of an interaction between the BRCA1 aminoacids 502–802 and the STAT1 TAD domain. This domain contains serine 727, whose phosphorylation is crucial for the transcriptional activity of STAT1 and which is in fact directly implicated in the recruitment of STAT1 transcriptional coactivators; its mutation causes defective STAT1-BRCA1 binding (Ouchi et al., 2000).

Furthermore, it has recently been shown that STAT1 may respond to IFNs and to PDGF by reducing the expression of the gene c-myc, thus confirming its important role as a cell growth inhibitor (Ramana et al., 2000). The non-phosphorylated form of STAT1 is also able to regulate genic expression; it seems, in fact, to repress hsp70 and the antiapoptosis factor bcl-xl and induce basal low molecular mass polypeptide-2 (LMP2) expression (Chatterjee-Kishore et al., 2000b).

Although STAT1 and STAT3 use quite different strategies for controlling cell growth and apoptosis, in a great many human neoplasias the constitutive activation of both of them can be observed. Nevertheless, their involvement appears different in specific cell types, for example, in squamous epithelial cells, aberrant stimulation of the epidermic growth factor receptor (EGFR) leads to the constitutive activation both of STAT1 and of STAT3, resulting, however, in very different effects. The use of direct antisense oligonucleotides against the initial translation site of STAT3 or of its negative-dominant mutants brings about cell growth inhibition, while the targeting of STAT1 with antisense or negative-dominant constructs has no effect on proliferation, which suggests that mediated autocrine growth of the EGFR and its ligands in such cells depends on STAT3 activation, but not on that of STAT1 (Grandis et al., 1998).

A recent study shows that in several breast cell lines, STAT6 may also be involved in oncogenesis, like STAT1, STAT3, and STAT5. The inhibition of cell growth and induction of apoptosis, reported by other authors in human breast cancer cells after treatment with IL-4 (Gooch et al., 1998) seems to be mediated by STAT6 activation (Gooch et al., 2002). Further studies will clarify the role of STAT6 in oncogenesis, which might well be similar to that of STAT1, since this too is an antigrowth and proapoptosis factor.

ANTICANCER THERAPY

The STAT proteins are involved in normal cellular processes such as embrionic development, regulation of cell differentiation, growth, apoptosis, and angiogenesis. Nevertheless, many studies have been indicated that the aberrant activation of STATs can be implicated in initiation and progression of human cancer, and suggested that STATs and STAT-signaling could represent effective molecular targets for the development of new anticancer treatments.

STAT-signaling inhibition does not only block cellular growth and induce apoptosis, but may also increase tumoral response to chemotherapy and conventional radiotherapy; it may well be that several types of common treatments fail because of the presence of constantly activated STAT proteins which inhibit apoptosis.

Molecular therapy may act at different levels of the STAT-signaling pathway.

New therapeutic targets might be the antagonists of cytokines or growth factors that bring about STAT signaling activation. These molecules are similar in structure to the physiological ligand and have more affinity with the receptor, but have no intrinsic capacity of activating STAT-signaling. An IL6 antagonist, Sant7, for example, blocks the constitutive activation of STAT3 in myeloma cells and is able to inhibit tumor growth (Dalton and Jove, 1999).

Furthermore, therapeutic antibodies can be used, which neutralize specific epitopes on receptors inhibiting interaction with their physiological ligand. For example, the antibodies anti-EGF-R (C225) or anti-HER2 (Herceptin) are now being used in clinical trials (Baselga, 2000a,b).

Since STAT activation depends mainly on the action of cellular protein kinases, tyrosine-kinases, such as JAKS, SRC, BCR-ABL, and serine-kinases, such as p38 and JNK, may also be considered as possible targets. The block of tyrosine-kinase activity, in fact, inhibits cell growth and induces apoptosis in hemopoietic cancers and in breast and in prostate tumor cells. For example, the EGF-R tyrosine-kinase inhibitor (PD0169414) has been seen to have an antitumoral activity in many types of neoplasias (Vincent et al., 2000). Furthermore, the tyrphostin AG490, a JAK2 inhibitor, induces cellular growth arrest and apoptosis in leukemias and in different tumor cells by blocking the STAT3 signal and preventing bcl-XL induction; pharmacological inhibition of the serine-kinase p38 activity also blocks induced v-Src transformation (Meydan et al., 1996).

Another way in which new antitumoral treatments can be effective is by the direct regulation of the action of STAT proteins. One of these methods involves the inhibition of dimerization, which is essential for nuclear translocation and transcriptional activity. When tyrosine, which is involved in dimerization, is replaced by phenylalanine, the mutant STATs are consequently unable to form dimers and to bind with DNA. Dimerization inhibition can be brought about by the use of artificial compounds, which have a high affinity for the STAT monomer, in order to generate compounds that are more stable than STAT-STAT homodimers (Bromberg et al., 1998). Recently, small phosphotyrosyl-peptides have been used; these bind at the STAT3 SH2 domain, thus blocking its phosphorylation, dimerization, DNA-binding, and gene regulation and, as a result, inhibiting induced v-Src cell transformation (Turkson et al., 2001). STAT coactivators play an extremely important role in guaranteeing DNA binding and the transcriptional activation carried out by the STATs. Therefore the use of pseudocoactivators with high STAT affinity but which compete with natural molecules might block the STAT-dependent biological effects.

At present, little is known about the precise mechanisms leading to activated STAT nuclear translocation, future studies in this field might also provide clearer indications regarding the use of several compounds for preventing this transfer and the consequent effect on the nucleus.

There are several other possible therapeutic approaches, which directly act on STATs by means of the use of antisense oligomers blocking the expression of specific STAT mRNA transcripts and of negative-dominant STAT proteins, so that they fail to carry out the most important STAT-signaling functions. One of these is the use of antisense oligomers against STAT3 in order to inhibit cell growth in head and neck squamous carcinoma cell lines and in prostate cancer cells (Grandis et al., 1998; Epling-Burnette et al., 2001). In ovarian tumor cells, the use of the form STAT3β, which lacks the C-terminal end and therefore the normal STAT3 transcriptional ability, prevents the expression of genes such as bcl-XL and cicline D1, thus determining a block of cell proliferation and consequent tumoral progression (Huang et al., 2000).

Finally, the identification of molecules, which imitate the action of proteins inhibiting the JAK-STAT pathway (PTP, SOCS, PIAS), thus blocking STAT action, might open new perspectives in the field of anticancer therapy.

CONCLUSION

The ever-increasing involvement of STAT proteins in human oncogenesis is opening up new possibilities regarding the study and the identification of new molecular targets for the development of future cancer therapy. Although they have no direct action on tumor development, the fact that they are at the center of tyrosine-kinase signaling means that STAT proteins play an important role in promoting cell-cycle growth and in inhibiting apoptosis, thus contributing on oncogenic progression.

STATs1, 3, and 5 might be considered as molecular markers for early detection of certain types of cancers and also prognostic indicators for determining tumoral aggressiveness and the response to various treatments. Molecular therapy leading to STAT-signaling inhibition may increase tumoral response to chemotherapy and conventional radiotherapy; since several types of common treatments fail because of the presence of constantly-activated STAT proteins which activate cell proliferation and inhibit apoptosis.

Future studies are needed in order to clarify the role of STAT proteins, of their coactivators and of the factors controlled by them, thus leading to further diagnostic as well as prognostic applications.