SUMMARY, DISCUSSION and FUTURE PERSPECTIVES
1. The signal transduction cascade involved in IL-6-induced STAT3 ser727 phosphorylation: kinetics and specificity
Although STAT tyrosine phosphorylation is the first critical event in ligand-induced STAT activation, which allows STAT dimerization, translocation the nucleus and binding of response elements in target gene promoters, STAT3 serine phosphorylation is a second event that critically regulates STAT activity (reviewed in [63]). Except for STAT2 and STAT6, all other STATs contain serine residues that become phosphorylated in a stimulus-regulated manner [1,61-63]. This second phosphorylation event provides the possibility to modulate and fine-tune STAT signal transduction and might introduce specificity in the effects that cytokines and growth factors have on cells.
Figure 1 describes a model for IL-6-induced STAT3 tyr705 and ser727 phosphorylation in HepG2 cells (Chapters 2 and 3). IL-6 initiates signaling by associating with its ligand-binding receptor (IL-6R), which allows dimerization of the gp130 receptor components.
The gp130 associated JAK kinases transphosphorylate tyrosine residues 767, 814, 905 and 915 that form docking sites for STAT3 once phosphorylated. Association of STAT3 with the gp130 receptor enables JAK-mediated STAT3 tyr705 phosphorylation. In HepG2 cells, this occurs with rather quick kinetics: within 2 min upon IL-6 stimulation maximal levels of STAT3 tyr705 phosphorylation were observed (Chapter 3). Once phosphorylated on tyr705, STAT3 dimerizes via reciprocal interactions between the SH2 domains that enable nuclear translocation. This fast nuclear import process was also demonstrated by immunofluorescence microscopy studies revealing that STAT3 cytoplasmic-nuclear translocation occurs within 5 min upon IL-6 stimulation (Fig.2, see also Chapter 3).
IL-6-induced STAT3 ser727 phosphorylation involves the sequential activation of Vav, Rac-1, MEKK, SEK-1/MKK-4 and PKCδ. The guanine nucleotide exchange factor Vav is associated with the membrane-distal region of the gp130 receptor and becomes phosphorylated on tyrosine residue(s) upon IL-6 stimulation (Chapter 2, [34]). Within 5 min, Vav associates with the small GTPase Rac-1, and probably regulates GDP-GTP exchange on Rac-1 [223], thus leaving it in the activated conformation. The kinetics of IL-6-induced Vav tyrosine phosphorylation correlate with the kinetics of dissociation of the
Vav-Rac-1 complex, suggesting that tyrosine phosphorylation of Vav is not a prerequisite for Vav-mediated Rac-1 activation but rather plays a role in the dissociation process, although mechanistic explanations for this model are still lacking. Upon activation, Rac-1 initiates a signal transduction cascade that is comprised of the MAP kinase kinase kinase MEKK-1, the MAP kinase kinase SEK-1/MKK-4 and PKCδ. IL-6 induces a transient activation of SEK-1/MKK-4 as determined by phosphorylation of the residue Thr223 within 5 min, which reached maximal levels at 10 min upon IL-6 stimulation (Chapter 2).
In unstimulated cells, SEK-1/MKK-4 is present as a complex with PKCδ, and upon IL-6 stimulation PKCδ dissociates from 1/MKK-4 within 10-15 min. Presumably, SEK-1/MKK-4 phosphorylates PKCδ on Thr505, which then translocates to the nucleus.
Figure 1. Proposed model for IL-6-induced STAT3 ser727 phosphorylation.
Phosphorylated PKCδ was only found in the nucleus reaching maximal phosphorylation levels at 5-10 min upon IL-6 stimulation, suggesting that PKCδ nuclear translocation is a rather quick event. At timepoint 15 min, PKCδ associates with STAT3 and phosphorylates STAT3 on ser727. STAT3 ser727 phosphorylation reaches maximal levels after 15 min.
Importantly, STAT3 ser727 phosphorylation was was only detected in the nuclear fractions and was absent from the cytoplasmic fractions. Taken together, these results suggest that IL-6-induced STAT3 ser727 phosphorylation is a nuclear event, particularly since STAT3 nuclear translocation occurs within 5 min upon IL-6 stimulation and precedes STAT3 ser727 phosphorylation.
In agreement with our data, Jain et al. have described that PKCδ is directly involved in IL-6-induced STAT3 ser727 phosphorylation [243]. In contrast, they report that STAT3-PKCδ associations occur mainly in the cytoplasm. In immunoprecipitation studies from IL-6-stimulated nuclear fractions of HepG2 cells, no nuclear PKCδ was detected, whereas we find a significant amount of PKCδ in total nuclear fractions of HepG2 cells (chapter 3).
Possibly, due to a lack of detectable immunoprecipitated PKCδ from nuclear fractions they were not able to observe a nuclear PKCδ-STAT3 association. Furthermore, they indicate that STAT3 tyr705 phosphorylation appears to be a prerequisite for PKC δ-mediated STAT3 ser727 phosphorylation, particularly since stimulation with PMA, which is a very potent activator of PKCδ and does not induce STAT3 tyr705 phosphorylation, does not result in association of PKCδ with STAT3 [243]. Since we find that practically all tyrosine phosphorylated STAT3 is present in the nucleus within 5 min upon IL-6 stimulation, these data suggest that IL-6-induced STAT3 ser727 phosphorylation is a nuclear event.
We can exclude the possibility that the ERK signal transduction cascade is involved in IL-6-induced STAT3 ser727 phosphorylation in HepG2 cells since inhibition at various levels of this signaling cascade did not interfere with IL-6-induced STAT3 ser727 phosphorylation (Chapter 2). Also, PI-3K or Src activity is not required since inhibition of the kinase activity of these molecules by treating cells with the chemical inhibitors wortmannin and PP2A, respectively did not reduce IL-6-induced STAT3 ser727 phosphorylation. Inhibition of p38 kinase activity by using the inhibitor SB203580 also did not interfere with IL-6-induced STAT3 ser727 phosphorylation, but rather enhanced both ser727 phosphorylation as well as STAT3 transactivation (Chapter 2). Since LPS and TNFα induce SOCS-3 expression via the p38 pathway [136], we speculated that a similar mechanism might be involved in IL-6 signaling as well. Inhibition of p38 kinase activity would then prevent the IL-6-induced upregulation of SOCS-3, which would allow an increase in STAT3-mediated gene transcription. However, IL-6 still induced SOCS-1 and SOCS-3 RNA in the presence of SB203580 to similar levels and with similar kinetics as compared to untreated cells (data not shown), indicating the p38 kinase activity is not required for the IL-6-induced upregulation of SOCS proteins. Possibly, p38 is required to activate a (nuclear) phosphatase in order to downregulate STAT3 signal transduction.
Further experiments are required to resolve this issue.
It is somewhat peculiar that the MAP kinase JNK is not involved in IL-6-induced STAT3 ser727 phosphorylation in HepG2 cells. In many cases, JNK-1 is the end-point kinase of the Rac-MEKK-SEK-1 signal transduction cascade. Although it has been demonstrated that JNK-1 is the kinase that phosphorylates STAT3 on ser727 in response to stress stimuli such as UV and Anisomycin [83], we can exclude the possibility that JNK-1 phosphorylates STAT3 in response to IL-6 based on the following observations: (i) JNK-1 is not activated in response to IL-6 stimulation (Chapter 2 and 3); (ii) JNK-1 is not nuclear-translocated in response to IL-6 (Chapter 2); (iii) no association between SEK-1/MKK-4 and JNK-1 was detected in mammalian two hybrid assays (Chapter 3); (iv) no IL-6-induced JNK-1 association with STAT3 was observed in immuno-precipitation
Figure 2. Kinetics of IL-6-induced STAT3 nuclear translocation. In unstimulated cells (0 min), STAT3 is distributed over the cytoplasm and the nucleus. Upon stimulation, STAT3 translocates to the nucleus within 5 min. After 60 min of IL-6 stimulation, STAT3 is relocated to the cytoplasm.
studies (data not shown). Thus, we conclude that JNK-1 is not involved in IL-6-induced STAT3 ser727 phosphorylation. PKCδ is the end-point kinase of the Rac-MEKK-SEK-1 signal transduction cascade and phosphorylates STAT3 on the ser727 residue in response to IL-6. It is plausible that JNK-1 and PKCδ are anchored in different signal transduction protein complexes and that these complexes are activated in a strictly ligand-dependent manner. Recently, two groups of proteins have been identified which might function as scaffold-proteins for the JNK-1 signal transduction cascade, the JNK-1 Interacting Proteins (JIP) and JNK/Stress-activated protein kinase-Associated Proteins (JSAP) [343,344]. These putative scaffold proteins interact with specific members of the JNK signal transduction cascade, including isoforms of the MAP kinase kinase kinases MEKK-1, -2, -3 and -4, the MAP kinase kinases SEK-1/MKK-4 and MKK-7, and the MAP kinases JNK-1, -2 and –3. It appears that these scaffold proteins only interact with a specific subset of isoforms of the JNK signal transduction cascade, and that these proteins selectively enhance the activation of signaling pathways by forming an anchor for the specific proteins. Thus, scaffold proteins will contribute to the specificity of numerous distinct signaling pathways in cells. The fact that these scaffold proteins can homo- or hetero-dimerize either via leucine-zipper or SH3 domains and contain multiple regulatory phosphorylation sites might introduce further specificity. It will be challenging to study whether scaffold proteins are also involved in IL-6-induced activation of the Vav-Rac-MEKK-SEK/MKK-4-PKCδ signal transduction cascade and determine their role in IL-6-induced STAT3 ser727 phosphorylation.
In HepG2 cells, IL-6-induced STAT3 signaling is transient. Within 60 min, both STAT3 tyr705 and ser727 phosphorylation is reduced to basal levels and STAT3 is re-entered into the cytoplasm. The kinetics of STAT3 dephosphorylation and cytoplasmic re-localization occur with similar kinetics as the upregulation of SOCS proteins. SOCS proteins downregulate STAT signal transduction by association with Jaks or the activated receptor complex, thereby preventing STAT tyrosine phosphorylation. SOCS-1 and SOCS-3 mRNA was first detected upon 30 min of IL-6 stimulation and prolonged for several hours (data not shown), suggesting that this negative feedback loop indeed downregulates STAT3 signal transduction. However, SOCS proteins will only inhibit STAT3 signaling at the receptor level and will prevent a re-activation of STAT3 when the IL-6-gp130 receptor complex is still in its active conformation. Thus, phosphatases must play an important role as well. Presumably, as is the case for STAT1 [115], nuclear tyrosine phosphatases will de-phosphorylate STAT3 tyrosine residues in the nucleus, which allows dissociation of the STAT3 dimer. STAT3 monomers will then be shuttled to the cytoplasm via a Nuclear Export Signal (NES)-mediated mechanism as has also been demonstrated for STAT1 [71].
In the cytoplasm, SOCS proteins will now prevent a re-activation of STAT3. Further experiments are required to identify the tyrosine phosphatase(s) involved, as well as the roles that potential serine phosphatases, receptor internalization and protein degradation might fulfill in the downregulation of IL-6-induced STAT3 signal transduction.
An important point concerns the ligand-dependent and cell-type specific conditions that determine which signal transduction cascade is utilized to phosphorylate STAT3 on ser727 (reviewed in [63]). For instance, IL-2 induces STAT3 ser727 phosphorylation via the MEK/ERK pathway in T lymphocytes, but in combination with IL-12 it utilizes the p38 pathway [88]. EGF induces ser727 phosphorylation via the MEK/ERK pathway in
NIH-3T3 cells [82], similar as insulin, OnM and LIF in adipocytes [345] and GM-CSF and G-CSF in human neutrophils [346]. Stress stimuli such as UV and TNFα utilize the JNK pathway to phosphorylate STAT3 on ser727 in COS-7 cells [83], whereas v-Src-induced STAT3 ser727 phosphorylation is mediated by both JNK-1 and p38 in v-Src-transformed NIH-3T3 cells [89]. In HepG2 cells, IL-6-induced STAT3 ser727 phosphorylation is independent of ERK activation (Chapter 2). Jain et al demonstrated that PKCδ directly phosphorylates STAT3 on ser727 [243], in agreement with our data. So, although IL-6 is capable of activating ERK-1 (Chapter 2), ERK is not involved in the IL-6-induced ser727 phosphorylation of STAT3, while other ligands such as EGF or insulin do utilize this signal transduction cascade in some cells. Similar, in response to stress stimuli, JNK-1 is the end-point kinase of the JNK signal transduction cascade, whereas in the case of IL-6, JNK-1 is not involved. The kinetics and the level of activation of the specific signal transduction cascades might account for these observations, as well as the involvement of scaffold proteins, which activation patterns might depend on the cytokine or growth factor that a cell is challenged to.