Oxidative Stress Impairs Multiple Regulatory Events to Drive Persistent Cytokine-Stimulated STAT3 Phosphorylation

CORE Metadata, citation and similar papers at core.ac.uk

Provided by Elsevier - Publisher Connector

Biochimica et Biophysica Acta 1843 (2014) 483–494

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

journal homepage: www.elsevier.com/locate/bbamcr

Oxidative stress impairs multiple regulatory events to drive persistent cytokine-stimulated STAT3 phosphorylation

Ivan H.W. Ng a,b, Yvonne Y.C. Yeap a, Lynette S.R. Ong a,1,DavidA.Jansb,⁎, Marie A. Bogoyevitch a,⁎⁎

a Department of Biochemistry and Molecular Biology, Bio21 Molecular Science and Biotechnology Institute, University of Melbourne, Victoria 3010, Australia b Department of Biochemistry and Molecular Biology, Monash University, Victoria 3800, Australia

article info abstract

Article history: Although cytokine-driven STAT3 phosphorylation and activation are often transient, persistent activation of Received 20 August 2013 STAT3 is a hallmark of a range of pathologies and underpins altered transcriptional responses. As triggers in Received in revised form 31 October 2013 disease frequently include combined increases in inﬂammatory cytokine and reactive oxygen species levels, Accepted 19 November 2013 we report here how oxidative stress impacts on cytokine-driven STAT3 signal transduction events. In the Available online 25 November 2013 model system of murine embryonic ﬁbroblasts (MEFs), combined treatment with the interleukin-6 family cytokine Leukemia Inhibitory Factor (LIF) and hydrogen peroxide (H O ) drove persistent STAT3 phosphorylation Keywords: 2 2 STAT3 whereas STAT3 phosphorylation increased only transiently in response to LIF alone and was not increased by Oxidative stress H2O2 alone. Surprisingly, increases in transcript levels of the direct STAT3 gene target SOCS3 were delayed during Cytokines the combined LIF + H2O2 treatment, leading us to probe the impact of oxidative stress on STAT3 regulatory

Nuclear transport events. Indeed, LIF + H2O2 prolonged JAK activation, delayed STAT3 nuclear localisation, and caused relocalisation of nuclear STAT3 phosphatase TC-PTP (TC45) to the cytoplasm. In exploring the nuclear import/ export pathways, we observed disruption of nuclear/cytoplasmic distributions of Ran and importin-α3 in cells

exposed to H2O2 and the resultant reduced nuclear trafficking of classical importin-α/β-dependent protein cargoes. CRM1-mediated nuclear export persisted despite the oxidative stress insult, with sustained STAT3 Y705 phosphorylation enhancing STAT3 nuclear residency. Our studies thus reveal for the first time the striking impact of oxidative stress to sustain STAT3 phosphorylation and nuclear retention following disruption of multiple regulatory events, with significant implications for STAT3 function. © 2013 Elsevier B.V. All rights reserved.

1. Introduction to downstream transcriptional changes, have been increasingly recognised as playing major roles in cancer and pathologies associated The proteins of Janus Kinase (JAK)/Signal Transducer and Activator with enhanced inflammation [1,2]. The STAT transcription factor pro- of Transcription (STAT) pathway, initially identified as important teins are largely inactive and cytoplasmic under normal conditions, direct signalling components linking cytokine receptor activation but upon JAK-stimulated tyrosine phosphorylation, translocate to the cell nucleus to participate directly in the expression of the STAT- fi Abbreviations: JAK, Janus kinase; STAT, Signal Transducer and Activator of speci c gene targets [3]. One member of the STAT family, STAT3, has Transcription; TC-PTP, T-cell protein tyrosine phosphatase; SOCS3, suppressor of cytokine attracted attention as a transcription factor mediating the gene expres- signalling 3; H2O2, hydrogen peroxide; LIF, leukemia inhibitory factor; OSM, oncostatin M; sion changes upon stimulation by the interleukin-6 family of cytokines MAPK, mitogen-activated protein kinase; MEF, murine embryonic fibroblast; FCS, foetal [4]. In this classical model of activation, STAT3 phosphorylation and calf serum; CLSM, confocal laser scanning microscopy; GFP, green fluorescent protein; nuclear localisation are usually transient, but persistent STAT3 phos- NES, nuclear export sequence; NLS, nuclear localisation sequence; Fn, nuclear fluorescence; Fc, cytoplasmic fluorescence; Fn/c, the nuclear to cytoplasmic ratio; SEM, standard phorylation and activation have frequently been linked with events of error of the mean cellular injury, inflammation and cancer [3,5,6]. ⁎ Correspondence to: D.A. Jans, Nuclear Signalling Laboratory, Department of Although the mechanisms leading to persistent STAT3 activation Biochemistry and Molecular Biology, Monash University, Clayton, VIC 3800, Australia. remain largely unclear, several regulatory events control the magnitude Tel.: +61 3 9901 9341; fax: +61 3 9902 9500. ⁎⁎ Correspondence to: M.A. Bogoyevitch, Department of Biochemistry and Molecular and extent of STAT3 activation under control conditions. These regula- Biology, Bio21 Molecular Science and Biotechnology Institute, University of Melbourne, tory events include the elevation of cytokine levels that promote recep- 30 Flemington Road, Parkville, VIC 3010, Australia. Tel.: +61 3 8344 2289; fax: +61 3 tor activation with downstream activation of tyrosine kinases [7], and 9348 1421. conversely the involvement of negative regulators such as the tyrosine E-mail addresses: [email protected] (D.A. Jans), [email protected] phosphatase T-cell protein tyrosine phosphatase (TC-PTP) [8–10] and (M.A. Bogoyevitch). 1 Present Address: Centre for Animal Biotechnology, School of Veterinary Science, the suppressor of cytokine signalling 3 (SOCS3) protein [11]. Notably, University of Melbourne, Parkville, VIC 3010, Australia. SOCS3 is itself a gene target of active STAT3 and its direct regulation

0167-4889/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.bbamcr.2013.11.015 484 I.H.W. Ng et al. / Biochimica et Biophysica Acta 1843 (2014) 483–494 by STAT3 provides an immediate negative feedback mechanism (Amersham Life Science). Proteins of interest were blotted using the attenuating signalling by the JAK/STAT3 pathway following cytokine following primary antibodies: anti-STAT3 N-terminus (BD Bioscience), activation [11]. anti-phospho-Y705 STAT3 or phospho-S727 STAT3 (Cell Signaling Increased intracellular reactive oxygen species levels, leading to Technology), anti-TC-PTP (R & D systems), anti-α-tubulin (Sigma) and oxidative stress, have also been linked to disease pathologies, often anti-c-Myc epitope (Santa Cruz Biotechnology). After incubation with promoting inflammatory processes and thus further elevating pro- horseradish peroxidase-linked secondary antibody (Thermo Scientific), inflammatory cytokine levels [12,13]. Whilst recent studies have immunoreactive proteins were visualized using an enhanced chemilu- demonstrated STAT3 activation by hydrogen peroxide (H2O2)asan minescence detection system (Thermo Scientific) and quantitation oxidative stress trigger [14–16], the impact of oxidative stress on the carried out using ImageJ 1.38 public domain software (National magnitude and kinetics of cytokine-stimulated STAT3 phosphorylation Institutes of Health, USA). and activation has not been evaluated. In this study, we focus on the consequences of oxidative stress on STAT3 activation by the interleukin-6 family cytokine Leukemia Inhibitory Factor (LIF). Our 2.3. Immunofluorescence, confocal laser scanning microscopy and image studies provide the first evidence that oxidative stress prolongs analysis cytokine-driven STAT3 activation by deregulating multiple regulatory events. Samples were prepared and analysed as described previously [32].Briefly, treated cells on coverslips were washed with cold 2. Materials and methods phosphate-buffered saline (PBS) before fixation using 4% [w/v] para- formaldehyde (15 min, 37 °C) and permeabilisation in 0.2% [v/v] 2.1. Cell culture, transfection and treatments Triton X-100/PBS (15 min, room temperature). Non-specific binding was blocked by incubation in 1% [w/v] bovine serum albumin/PBS Wild-type (WT) murine embryonic fibroblasts (MEFs), STAT3−/− (30 min, room temperature), then coverslips incubated with prima- MEFs, or HeLa cells were maintained in high glucose Dulbecco's modi- ry antibodies [STAT3 (C-20) (Santa Cruz), Ran (BD Bioscience), fied Eagle's medium (DMEM) containing L-glutamine (Gibco) supple- importin-α3/KPNA4 or importin-β1/NTF97 (Abcam), 1:200 dilution mented with 10% [v/v] foetal calf serum (FCS; DKSH Australia) and in sterile filtered 1% [w/v] bovine serum albumin/PBS] then washed 100 U/ml penicillin/streptomycin (Gibco) then changed to serum-free with PBS before incubation with Cy2/Cy3- (Millipore) or Alex488- media conditions 16 h before treatment with either 10 ng/ml murine coupled (Invitrogen) secondary antibodies. Nuclei were visualized recombinant LIF (Sigma), 1 mM H2O2 (Ajax Finechem), or a combina- using DAPI (Sigma, 1:15000 in PBS). Coverslips were mounted tion of 10 ng/ml LIF and 1 mM H2O2 (LIF + H2O2). For selected studies, (Biomeda Gel Mount, ProSciTech) onto glass slides and confocal 50 μM menadione was used to generate superoxide in place of oxidative laser scanning microscopy was performed using a Leica TCS SP2 stress initiated by H2O2. For protein kinase inhibitor studies, WT MEFs imaging system with a 100 × 1.35 NA objective. Quantitation of were pre-treated with a Mitogen-Activated Protein Kinase (MAPK) relative fluorescence intensities in digitised confocal images was pathway inhibitor (20 μM JNK Inhibitor VIII [17];20μMMEKinhibitor, carried out by measuring fluorescence in an area of the nucleus UO126 [18];20μM p38 inhibitor, SB203580 [19]), 0.5 μMJAKinhibitorI and cytoplasm of cells in 10 different fields (ImageJ 1.38 public [20] or 10 μM Src family kinase inhibitor SU6656 [21] for 1 h prior to LIF, domain software) from 3 independent experiments (n = 3). The

H2O2,orLIF+H2O2 treatment. Cells were pre-treated with the STAT3 values for the nuclear (Fn)andcytoplasmic(Fc) fluorescence, pathway inhibitor STATTIC [22] (20 μM) 30 min prior to treatment subsequent to subtraction of background fluorescence, enabled the with LIF, H2O2,orLIF+H2O2. All inhibitors were from Calbiochem. specific nuclear to cytoplasmic ratio (Fn/c) to be calculated [32]. For the delivery of anti-importin-α3 or anti-importin-β1 antibodies (Abcam) into intact cells, the Chariot peptide system was used according to the manufacturer's instruction (Active Motif) [23,24].Briefly, 2.4. RNA preparation and quantitative real-time PCR antibodies (1 μgin100μl PBS) were mixed with 3.5 μlChariotpeptide solution and 100 μldistilleddeionisedH2O, added to 150 μLofserum- Samples were prepared and analysed as described previously [32]. free culture medium, and then preincubated with cells (12 well dishes) Briefly, total RNA was extracted from MEFs using the Purelink RNA for 4 - 12 h prior to exposure to LIF. Transient transfection studies were mini-kit (Invitrogen) and reverse-transcribed to cDNA using the RT carried out with the expression of myc epitope-tagged phosphatase TC- High Capacity kit (Applied Biosystems) according to the manufacturer's PTP [25], or Green Fluorescent Protein (GFP)-tagged nuclear export protocols. Quantitative real-time TaqMan® PCR to determine the sequence (NES) and nuclear localisation sequence (NLS) constructs: regulation of SOCS3 transcript levels was performed using 50 ng cDNA GFP–Rev–NES [26], GFP-T-antigen (GFP-T-ag(114–135)) [27],GFP- in a 20 μl reaction containing TaqMan® Gene Expression Master pUL54(1145–1161) [28], GFP-VP3(74–121), GFP–VP3–NES mutant Mix and specific TaqMan® Gene Expression Assay (AssayID: SOCS3, [29,30] or a control GFP-only construct using Lipofectamine™ LTX Mm00545913_s1; Applied Biosystems). Amplification of cDNA was car- with Plus™ reagent and Opti-MEM according to the manufacturer's ried out in a 48-well Step One real-time PCR system (Applied Biosystems) instructions (Invitrogen) prior to further treatments. In control using the following PCR conditions: 2 min at 50 °C and 10 min at 95 °C, experiments, GFP–Rev–NES cells were additionally pre-treated with followed by 40 cycles of 15 s at 95 °C and 1 min at 60 °C. The data was 10 ng/ml Leptomycin B, a CRM1 inhibitor [31] for 60 min prior to normalised to GAPDH (AssayID: Mm99999915_g1) in the respective analysis. samples and data quantitation was carried out using the 2−ΔΔCT method and expressed relative to the control sample. RNA isolation and expres- 2.2. Cell lysate preparation and immunoblot analysis sion analysis were performed on 3 independent occasions.

Following treatment, cell lysates were prepared using RIPA buffer (50 mM Tris–HCl, pH 7.3, 150 mM NaCl, 0.1 mM EDTA, 1% [v/v] sodium 2.5. Statistical analysis deoxycholate, 1% [v/v] Triton X-100, 0.2% [w/v] NaF and 100 μM

Na3VO4) supplemented with complete protease inhibitor mix (Roche Statistical analysis was carried out using Graphpad Prism 5 software Diagnostic). Protein samples, diluted with 3× protein sample buffer, (Windows version 5.00, GraphPad Software) and the unpaired student were resolved by SDS-PAGE (10% [v/v] polyacrylamide gels) and T-test was used to compare data from control cells versus treated cells at the separated proteins were transferred onto PVDF membranes each corresponding time point. I.H.W. Ng et al. / Biochimica et Biophysica Acta 1843 (2014) 483–494 485

3. Results not the result of irreversible damage to the cell regulatory system during oxidative stress (Fig. 1E, and quantitative data in Supp. Fig. 1C).

3.1. Treatment with LIF and H2O2 in combination results in sustained STAT3 To determine the impact of oxidative stress on STAT3 transcriptional Y705 and S727 phosphorylation and delayed expression of SOCS3 activity, we used RT-PCR to evaluate the expression of SOCS3, a well- known STAT3 direct target gene [11]. In the presence of LIF alone, a To test the impact of oxidative stress on STAT3 signalling, we robust increase in SOCS3 was observed at 15 min (Fig. 2), correspond- exposed MEFs to 1 mM H2O2 or 10 ng/ml Leukemia Inhibitory Factor ing with the peak in STAT3 pY705 and pS727 (Fig. 1). H2O2 treatment (LIF), as well as a combination of both stimuli (LIF + H2O2), for up to alone did not induce SOCS3 expression, but LIF + H2O2 treatment 2 h. STAT3 activation was analysed by immunoblotting for phosphory- resulted in a delayed increase in SOCS3 expression. We conﬁrmed lation of STAT3 at Y705 (pY705 STAT3, required for dimerization and STAT3 inhibition by STATTIC (20 μM) [22] (Supp. Fig. 2), and demon- nuclear translocation [3]), and S727 (pS727 STAT3, linked to increased strated the requirement for STAT3 in SOCS3 expression by showing STAT3 transactivation [33]). Consistent with LIF-stimulated STAT3 acti- that pre-treatment with STATTIC blocked SOCS3 induction in vation, we observed rapid and transient increases in both pY705 and LIF + H2O2-treated WT MEFs, and that there was a lack of SOCS3 induc- −/− pS727 STAT3 following exposure to LIF (Fig. 1A). Exposure to H2O2 for tion in response to LIF/H2O2/LIF + H2O2 in STAT3 MEFs (Fig. 2). up to 2 h stimulated a detectable increase in pS727 STAT3 without an Taken together, these results indicate a delayed STAT3-dependent reg- accompanying increase in pY705 STAT3, but the combined treatment ulation of SOCS3 expression following the combined cytokine and of LIF + H2O2 resulted in increased and sustained pY705 STAT3 togeth- oxidative stress treatment despite persistent STAT3 phosphorylation. er with delayed pS727 STAT3 (Fig. 1A). Densitometric analysis and statistical evaluation of the quantitative data from three independent 3.2. JAKs and Src family kinases, but not MAPKs, contribute to STAT3 experiments conﬁrmed these observations of sustained pY705 STAT3 phosphorylation during combined LIF + H2O2 exposure and delayed pS727 STAT3 following LIF + H2O2 co-exposure (Fig. 1B). To ensure this sustained STAT3 Y705 phosphorylation was not We next evaluated whether upstream regulators of STAT3 phos- reliant on the MEF cell background, we also exposed HeLa cells to co- phorylation could be responsible for the altered STAT3 phosphorylation treatment with oncostatin M + H2O2 and demonstrated that the pres- and transcriptional activation following exposure to LIF + H2O2.There ence of H2O2 could sustain the STAT3 phosphorylation in response to were no changes in the levels of JAK1 or 2, or in the kinetics of the cytokine oncostatin M (Fig. 1C, and quantitative data in Supp. phosphorylation of JAK1 (pY1022) or JAK2 (pY1007) in the combined

Fig. 1A). Similarly, to ensure that the responses were not speciﬁc for presence of LIF + H2O2 detected by immunoblotting (Supp. Fig. 3). H2O2, we also exposed MEFs to the co-treatment of LIF + menadione, The pan JAK inhibitor I prevented pY705 phosphorylation in response a superoxide generator, and again demonstrated the sustained STAT3 to LIF or LIF + H2O2 treatment (Fig. 3A, and quantitative data in Supp. phosphorylation (Fig. 1D, and quantitative data in Supp. Fig. 1B), Fig. 4A), consistent with the central role for the JAKs as the key media- emphasizing that different initiators of oxidative stress can promote tors of STAT3 Y705 phosphorylation status. Since oxidative stress can prolonged STAT3 activation in different cell types in the presence of also upregulate Src family kinase activity [34], we tested whether the

IL6 family cytokines. Furthermore, replacing LIF + H2O2-containing Src family kinase inhibitor SU6656 could alter STAT3 phosphorylation. medium with control serum-free medium could reverse the changes, Markedly reduced levels of STAT3 Y705 phosphorylation upon implying that the observed increases in STAT3 phosphorylation are LIF + H2O2 treatment were observed in the presence of SU6656

A LIF H2O2 LIF + H2O2 B 0 15 60 120 15 60 120 15 60 120 (min) 4 15 min 60 min 120 min 4 15 min 60 min 120 min

STAT3 ∗∗ (pY705) 3 ∗∗ 3 ∗∗ ∗∗∗ STAT3 ∗∗ ∗∗∗ (pS727) 2 2 ∗ ∗∗ STAT3 ∗∗∗ ∗∗ 1 1 ∗ Relative Density (A.U) ∗ Relative Density (A.U) (pY705 STAT3/STAT3) ∗∗

(pS727 STAT3/STAT3) ∗

α-tubulin 0 0 LIF H2O2 LIF + LIF H2O2 LIF + H2O2 H2O2 Control Control

HeLa LIF + CDE LIF + H2O2 OSM H O OSM + H O LIF MEN MEN 2 2 2 2 0 15 60 120 15 60 (min) 0 15 60 120 15 60 120 15 60 120 (min) 0 15 60 120 15 60 120 15 60 120 (min) SF Media: - 0 30 60 0 30 60 STAT3 STAT3 STAT3 (pY705) (pY705) (pY705) STAT3 STAT3 STAT3 (pS727) (pS727) (pS727)

STAT3 STAT3 STAT3

α-tubulin α-tubulin α-tubulin

Fig. 1. Oxidative stress drives reversible persistent cytokine-stimulated STAT3 phosphorylation. (A & B) WT MEFs were treated with 10 ng/ml LIF, 1 mM H2O2 or LIF + H2O2 for the times indicated. (A) Lysates were subjected to immunoblot analysis for activated STAT3 (pY705 or pS727), total STAT3, or α-tubulin as a loading control. (B) Quantitative analysis of the immunoblots for STAT3, STAT3 pY705 and STAT3 pS727 (n = 3) was performed using ImageJ software. Results for STAT3 pY705 and STAT3 pS727 are expressed relative to those for total STAT3. Data represent the mean ± SEM. Asterisks denote statistically signiﬁcant differences between cells treated compared to untreated cells (*, p ≤ 0.05; **, p ≤ 0.01; ***, p ≤ 0.001). (C) HeLa cells were treated with 10 ng/ml oncostatin M (OSM), 1 mM H2O2 or OSM + H2O2 for the times indicated, or (D) MEFs were treated with 10 ng/ml LIF, 50 μM menadione (MEN) or

LIF + MEN. Lysates were subjected to immunoblot analysis as per (B). (E) MEFs were treated with LIF + H2O2 (0–120 min as indicated) prior to exposure to serum-free (SF) media (0–60 min as indicated). Lysates were subjected to immunoblot analysis as per (B). 486 I.H.W. Ng et al. / Biochimica et Biophysica Acta 1843 (2014) 483–494

20 0 min 15 min 60 min 120 min

Relative Quantity 5 ∗∗ SOCS3/GAPDH mRNA (A.U) ∗∗ ∗∗ ∗∗ 0 LIF H2O2 LIF + LIF H2O2 LIF + LIF H2O2 LIF + H2O2 H2O2 H2O2

Control 20 μM STATTIC STAT3-/- MEF

Fig. 2. Oxidative stress slows cytokine-stimulated STAT3-mediated changes in SOCS3 expression. WT MEFs, WT MEFs pre-treated (30 min) with 20 μMSTATTIC,orSTAT3−/− MEFs were treated with agents as per the legend to Fig. 1. SOCS3 and GAPDH mRNA levels were measured by quantitative real-time PCR and values normalised to the levels detected in the cells under control conditions (n = 3). Data represent the mean ± SEM. Asterisks denote statistically signiﬁcant differences as per the legend to Fig. 1B.

(Fig. 3B, and quantitative data in Supp. Fig. 4B), implicating the involve- clearly being active in inhibiting phosphorylation of their respective ment of Src family kinases in upstream regulatory events leading to speciﬁc pathway components (Fig. 4A, B and C, and quantitative data STAT3 Y705 phosphorylation induced by LIF in the absence or presence in Supp. Fig. 5A, B and C, respectively). Thus, proline-directed kinases of H2O2. other than these MAPKs are likely responsible for mediating the in- As LIF + H2O2 treatment also increased pS727 STAT3 levels creases in pS727 STAT3 observed in response to LIF, H2O2 or LIF + H2O2. (Fig. 1A), we evaluated the mitogen-activated protein kinases (MAPKs) that have been implicated as direct regulators of STAT3 pS727 [33]. We examined the effects of inhibitors of the different 3.3. H2O2 induces delayed LIF-dependent STAT3 nuclear translocation and MAPK pathways (JNK Inhibitor VIII directed at the JNKs, SB203580 di- mislocalisation of STAT3 phosphatase TC-PTP (TC45) to the cytoplasm rected at p38 MAPK, and UO126 directed at MEK1/2 within the

ERK pathway). None of the inhibitors altered the LIF-, H2O2-or As STAT3 phosphorylation is required for cytokine-stimulated STAT3 LIF + H2O2-stimulated increases in pY705 or pS727 STAT3, although nuclear localisation [3], we evaluated STAT3 localisation, under conditions of combined cytokine stimulation and oxidative stress, using Control 0.5 µM JAK Inhibitor I A both total STAT3 and STAT3 pY705 antibodies. In contrast to H2O2 LIF + LIF + alone that had no effect, LIF induced a rapid increase in nuclear and LIF H2O2 LIF H2O2 H2O2 H2O2 phosphorylated STAT3 in the ﬁrst 15 min of exposure followed by a

0 15 60 15 60 15 60 0 15 60 15 60 15 60 (min) decrease in nuclear levels (Fig. 5). Combined LIF + H2O2 treatment STAT3 appeared to delay nuclear STAT3 localisation but resulted in prolonged (pY705) nuclear STAT3 levels over the 2 h treatment that correlated with the in- creasing nuclear pY705 staining over the 2 h treatment (Fig. 5A and C). This was conﬁrmed by quantitative analysis (Fig. 5B and D). Clearly, LIF- STAT3 induced nuclear translocation kinetics are altered in the presence of

H2O2. In addition to upregulated tyrosine kinase activity, sustained STAT3 α -tubulin phosphorylation can also result from reduced STAT3 phosphatase activity [35,36], with oxidative stress known to impact on tyrosine phosphatases in particular [37]. Since the nuclear phosphatase T-cell protein B Control 10 µM SU6656 tyrosine phosphatase (TC-PTP) can dephosphorylate nuclear pY705- LIF + LIF + STAT3 [9,35,38,39], we examined TC-PTP under conditions of LIF, H2O2 LIF H2O2 LIF H2O2 H2O2 H2O2 or LIF + H2O2 exposure. Immunoblot analysis revealed a significant 0156015601560 0156015601560 (min) (up to 60%) decrease in the total levels of detectable endogenous TC- PTP following exposure of MEFs to H O or LIF + H O (Fig. 6A, and STAT3 2 2 2 2 (pY705) quantitative data in Fig. 6B); loss could be reduced by the inclusion of the proteasomal inhibitor MG132 (Fig. 6C, and quantitative data in Fig. 6D), thus implying proteasomal degradation of TC-PTP under the STAT3 conditions of oxidative stress. The nuclear/cytoplasmic distribution of TC-PTP under conditions of α-tubulin oxidative stress, a factor of key importance in providing access to its nuclear or cytoplasmic substrates [25], was also examined by transiently transfecting MEFs to express myc epitope tagged-TC-PTP and Fig. 3. JAK and Src family tyrosine kinases contribute to persistent STAT3 Y705 phosphor- performing immunostaining in the absence or presence of cytokine ylation driven by cytokine and oxidative stress stimulation. MEFs were pre-treated (1 h) and H O . Our analysis confirmed the exclusive nuclear localisation of with DMSO as a control, (A) 0.5 μM JAK Inhibitor I, or (B) 10 μMSU6656beforeexposure 2 2 myc-TC-PTP in either the absence or presence of LIF alone, but revealed to 10 ng/ml LIF, 1 mM H2O2 or LIF + H2O2 for the times indicated. Lysates were subjected to immunoblot analysis as per the legend to Fig. 1. significant (p ≤ 0.001) relocalisation of myc-TC-PTP from the nucleus I.H.W. Ng et al. / Biochimica et Biophysica Acta 1843 (2014) 483–494 487

ABCControl 20 µM JNK Inhibitor VIII Control 20 µM SB 203580 Control 20 µM UO126 LIF + LIF + LIF + LIF + LIF + LIF + LIF H2O2 LIF H2O2 LIF H2O2 LIF H2O2 LIF H O LIF H O H2O2 H2O2 H2O2 H2O2 2 2 2 2 H2O2 H2O2 0156015601560 0156015601560(min) 0156015601560 0156015601560(min) 0156015601560 0156015601560 (min) STAT3 STAT3 (pY705) (pY705) STAT3 (pY705) STAT3 STAT3 STAT3 (pS727) (pS727) (pS727)

STAT3 STAT3 STAT3 MAPKAP2 JNK1/2 (pT222) ERK1/2 (pT183/pY185) (pT202/pY204) MAPKAP2 (pT334) JNK1/2 ERK1/2 MAPKAP2

p38 α c-Jun -tubulin (pT180/pY182)

α-tubulin p38

α-tubulin

Fig. 4. Inhibition of MAPK pathways does not abrogate persistent STAT3 phosphorylation in the presence of cytokine and oxidative stress. MEFs were pre-treated with DMSO, (A) 20 μM

JNK Inhibitor VIII, (B) 20 μM SB203580 (p38 inhibitor) or (C) 20 μM UO126 (MEK1/2 inhibitor) for 1 h before treatment with 10 ng/ml LIF, 1 mM H2O2 or LIF + H2O2 (15 or 60 min, as indicated). Cell lysates were prepared and subjected to immunoblot analysis for activated STAT3 (pY705 and pS727), total STAT3, or α-tubulin as a loading control. (A) Immunoblot analysis for phosphorylated JNK1/2 (pT183/pY185), total JNK1/2, and total c-Jun, confirmed JNK Inhibitor VIII actions. (B) Immunoblot analysis for phosphorylated MAPKAP2 (pT222 and pT334), total MAPKAP2, phosphorylated p38 (pT180/pY182), and total p38, confirmed SB203580 actions. (C) Immunoblot analysis for phosphorylated ERK1/2 (pT202/pY204) and total ERK1/2, confirmed UO126 actions.

to the cytosol at 60 min after treatment with H2O2 in either the absence explore more broadly the possibility that oxidative stress disrupts or presence of LIF (Fig. 6E, and quantitative data in Fig. 6F). importin-dependent nuclear transport. Importins-β1and-α3and Ran have previously been implicated in STAT3 nuclear import during 3.4. Oxidative stress disrupts nuclear transport protein distributions and cytokine-stimulation [40–43],andweconﬁrmed the reliance of functions STAT3 nuclear entry on importin-β1 and importin-α3byusingthe Chariot peptide system to deliver importin-β1 and importin-α3an- With our results highlighting altered cytoplasmic/nuclear distri- tibodies to cells. This showed the reduced nuclear accumulation bution of both STAT3 and TC-PTP, we extended our analyses to of STAT3 in response to LIF stimulation (Supp. Fig. 6). We next

A 0 min15 min 60 min 120 min C 0 min15 min 60 min 120 min

STAT3 pSTAT3

LIF LIF

H2O2 H2O2

LIF + LIF +

H2O2 H2O2

B 15 min 60 min 120 min D 15 min 60 min 120 min

6 200 ∗∗∗ n = n = 65 ∗∗∗ ∗∗∗ 72 n =

n = ) n = 74 2 150 68 65 4 ∗∗∗ n = STAT3 77 100 ∗∗∗ Fn/c n = 75 2

∗∗∗ pSTAT3 Y705 n = n = ∗∗∗ ∗∗∗ ∗∗∗ (mean nuclear

fluorescence/µm ∗∗∗ 63 68 n = n = 50 n = n = n = n = n = n = n = n = 72 73 69 61 65 n = 70 63 66 57 69 58 0 0 LIF H O LIF + 2 2 LIF H2O2 LIF + H O 2 2 H2O2 Control Control

Fig. 5. Oxidative stress delays cytokine-stimulated STAT3 nuclear accumulation. MEFs were treated with agents as per the legend to Fig. 1. Confocal laser scanning microscopy analysis using cells immunostained with either (A) anti-STAT3-C20 or (C) anti-pSTAT3 Y705 primary and Cy2-coupled secondary antibodies. Bars represent 10 μm. (B) Quantitative analysis of

STAT3 localisation using the ImageJ software to determine the nuclear to cytoplasmic ratio (Fn/c), representing measurements on at least 8 cells from 10 different microscopic fields for each time point for the nuclear (Fn) and cytoplasmic (Fc) fluorescence above background. (D) Quantitative analysis of pSTAT3 Y705 nuclear staining using the ImageJ software to determine the mean nuclear fluorescence intensity per μm2. Data represent the mean ± SEM. Asterisks denote statistically significant differences as per the legend to Fig. 1B. 488 I.H.W. Ng et al. / Biochimica et Biophysica Acta 1843 (2014) 483–494

LIF H2O2 LIF + H2O2 A B 15 min 60 min 120 min 0 15 60 120 15 60 120 15 60 120 (min) 1.5

TC-PTP

1.0 STAT3 ∗ (pY705) ∗ ∗∗ ∗∗ 0.5 ∗∗ ∗∗ STAT3 (TC-PTP/tubulin) Relative Density (A.U)

α-tubulin 0 LIF H2O2 LIF + H2O2 Control

Control 20 µM MG132 CD1.5 15 min 60 min LIF + LIF + LIF H2O2 LIF H2O2 H2O2 H2O2 01560156015600156015601560(min) 1.0 TC-PTP ∗∗ ∗∗ ∗∗ ∗∗ STAT3 0.5 (TC-PTP/tubulin) Relative Density (A.U) α-tubulin 0 LIF H2O2 LIF + LIF H2O2 LIF + H2O2 H2O2 Control Control

Control 20 μM MG132 E 0 min 15 min 60 min F MYC-TC-PTP

15 min 60 min LIF 10 n= n= 51 n= 43 n= 46 n= 41 8 63 n/c F 6

H2O2 4

Myc-TC-PTP 2 ∗∗∗ ∗∗∗ n= n= 41 56 0

LIF + LIF H2O2 LIF +

H2O2 H2O2 Control

Fig. 6. Oxidative stress induces TC-PTP degradation and mislocalisation to the cytoplasm. MEFs were treated with agents as per the legend to Fig. 1. (A & B) Cell lysates were prepared and subjected to immunoblot analysis for TC-PTP, activated STAT3 (pY705), total STAT3, and α-tubulin as a loading control. (C & D) MEFs were pre-treated with either DMSO or 20 μM MG132

(proteasomal inhibitor) for 6 h before treatment with 10 ng/ml LIF, 1 mM H2O2 or LIF + H2O2 (as indicated). Cell lysates were prepared and subjected to immunoblot analysis for TC-PTP, STAT3, or α-tubulin as a loading control. (B & D) Quantitative analysis of the immunoblots for the levels of TC-PTP (n = 3) was performed using ImageJ software as per the legend to Fig. 1B, where data represent the mean ± SEM. Asterisks denote statistically signiﬁcant differences as per the legend to Fig. 1B. (E & F) MEFs were transiently transfected to express

Myc-tagged TC-PTP prior to treatment with 10 ng/ml LIF, 1 mM H2O2 or LIF + H2O2 for 15 or 60 min as indicated. (E) Confocal laser scanning microscopy analysis on cells immunostained using anti-myc primary and Cy2-coupled secondary antibodies. Bar represents 10 μm. (F) Image analysis of Myc-TC-PTP localisation using ImageJ software as described in the legend to Fig. 5B; data represent the mean ± SEM. Asterisks denote statistically signiﬁcant differences between treated and control cells as per the legend to Fig. 1B.

examined the subcellular localisations of importins-β1and-α3and Ran proteins could be observed within 5 min of H2O2 treatment RaninresponsetotreatmentwithLIF,H2O2 or LIF + H2O2. Whilst (Ivan Ng, unpublished data). Quantitative image analysis to deter- importin-β1 localisation remained predominantly cytoplasmic/ mine Fn/c values conﬁrmed these results, with importin-α3and nuclear envelope-localised throughout 60 min of each treatment Ran showing signiﬁcantly (p b 0.001) altered Fn/c values in the (Fig. 7A, and quantitative data in Fig. 7B), both importin-α3and presence of H2O2 regardless of LIF stimulation (Fig. 7D and F). Com- Ran showed altered localisation even after 15 min of H2O2 treat- parable nuclear and cytoplasmic mislocalisation of importin-α3and ment, with importin-α3 accumulating in the nucleus (Fig. 7C, and Ran respectively was observed following exposure to 50 μMmena- quantitative data in Fig. 7D) and Ran mislocalising to the cytoplasm dione to induce oxidative stress (Supp. Fig. 7). Thus, oxidative stress in the absence or presence of LIF (Fig. 7E, and quantitative data in has clear effects on localisation of key components of the cellular Fig. 7F). Indeed, the aberrant localisation of both importin-α3and nuclear transport machinery. I.H.W. Ng et al. / Biochimica et Biophysica Acta 1843 (2014) 483–494 489 A 0 min 15 min 60 min IMP-β 1

LIF 15 min B n = 0.3 60 n = n = 58 n = 60 min n = 53 n = n = 55 66 55 45 n/c

F 0.2 α

H O 2 2 0.1 Importin- 1

LIF H2O2 LIF + H O LIF + 2 2 Control H2O2

C 015min min60 min IMP-α 3

LIF ∗∗∗ D 15 min ∗∗∗ 60 min 3 n = 67 n = n = 45 n = 45 n/c 66 F

3 2

H2O2 α

n = n = n = 1 53 58 63 Importin-

0 LIF + LIF H2O2 LIF + H2O2 H2O2 Control

E 015min min60 min Ran

LIF ∗∗∗ 15 min F ∗∗∗ 60 min 3 n = 51 n = n = 53 56

n/c 2 F H2O2 Ran 1 n = n = n = n = 56 43 48 51

0 LIF + LIF H2O2 LIF + H2O2 H2O2 Control

Fig. 7. Oxidative stress disrupts importin-α3 and Ran but not importin-β1 subcellular localisation. MEFs were treated with agents as per the legend to Fig. 1. Confocal laser scanning microscopy analysis was performed on immunostained cells using specific primary antibodies for (A & B) importin-β1(IMP-β1), (C & D) importin-α3(IMP-α3) or (E & F) Ran, followed by the appropriate fluorescently labelled secondary antibodies. Bars represent 10 μm. (B, D & F) Quantitative results for importin-β1, importin-α3 and Ran localisation using ImageJ software to determine the Fn/c as per the legend to Fig. 5B; data represent the mean ± SEM. Asterisks denote statistically significant differences between treated and control cells as per the legend to Fig. 1B.

To investigate further the extent to which the effect of H2O2 on LIF- Strikingly, even though H2O2 pre-treatment for 15 or 60 min caused re- dependent STAT3 nuclear translocation might relate to the mis- distribution of Ran to the cytoplasm and importin-α3 to the nucleus to localisation of key components of STAT3's nuclear transport pathway, similar extents, LIF-induced translocation of STAT3 from the cytoplasm we tested the effect of pre-treatment with H2O2 on STAT3 trafﬁcking. to the nucleus was not abrogated (Fig. 8A, and quantitative data in 490 I.H.W. Ng et al. / Biochimica et Biophysica Acta 1843 (2014) 483–494

A H2O2: No pre-treatment 15 min pre-treatment 60 min pre-treatment + LIF: 0 min 15 min 60 min 0 min 15 min 60 min 0 min 15 min 60 min STAT3

Ran

IMP-α 3

+ LIF 0 min + LIF 15 min + LIF 60 min B ∗∗∗ ∗∗

6 n = 54 ∗∗∗ 5 ∗∗∗ ∗∗∗ ∗∗∗ 4 n = n = 46 ∗∗∗ 48 n = n = n = 54 n = Fn/c 46 48 3 n = n = 57 n = 42 n = 47 n = n = 56 45 67 66 2 n = n = 50 45 n = n = n = n = n = 42 n = n = n = n = 1 51 48 51 53 56 43 50 n = 56 n = n = 48 53 47 0

H2O2 Pre-treatment: 0 min 15 min 60 min 0 min 15 min 60 min 0 min 15 min 60 min STAT3 Ran Importin-α 3

C LIF: 0 min 15 min 60 min 15 min pre-treatment

+ H2O2: 0min 0min 0min 15 min 30 min 45 min STAT3

Ran

IMP-α3

∗∗∗ D 6 n = 53 ∗∗∗ n = n = 5 48 51 n = 42 ∗∗∗ n = 4 48 n = 50 n = 51 Fn/c 3 n = n = 42 n = 56 53 2 n = 43 n = n = n = n = 42 41 1 n = n = n = 48 42 44 47 53 0 LIF pre-treatment: 015 60 15 0603015 0604515 (min) + H2O2: 00 0 15 30 45 001530450 001530450 STAT3 Ran Importin-α 3 I.H.W. Ng et al. / Biochimica et Biophysica Acta 1843 (2014) 483–494 491

H2O2 A Pre-treatment : 01560 B (min) + LIF 0 min + LIF 15 min + LIF 60 min + LIF: 0 15 60 0 15 60 0 15 60 4 STAT3 ∗∗∗ (pY705) ∗∗∗ 3 ∗ STAT3

ERK1/2 2 (pT202/pY204) Relative Density (A.U) ERK1/2 (pY705 STAT3/STAT3) 1

α-tubulin 0 H2O2 Pre-treatment: 0min 15min 60min

LIF : 01560 15 C (min) D + H2O2 : ---15 30 45 4 ∗∗∗ STAT3 (pY705) 3 STAT3

ERK1/2 2 (pT202/pY204)

ERK1/2 Relative Density (A.U) 1 (pY705 STAT3/STAT3)

α-tubulin 0 LIF pre-treatment: 015 60 15

+ H2O2: 00 0 15 30 45

Fig. 9. Phosphorylated pY705 STAT3 levels are sustained in different cytokine and oxidative stress pre-treatment protocols. (A) MEFs were pre-treated with 1 mM H2O2 (0–60 min as indicated) prior to exposure to 10 ng/ml LIF for 15 or 60 min in the continued presence of 1 mM H2O2. (C) MEFs were either treated with 10 ng/ml LIF (0–60 min as indicated) or pre-treated with 10 ng/ml LIF (15 min) prior to exposure to 1 mM H2O2 for 15 or 60 min in the continued presence of 10 ng/ml LIF. (A & C) Cell lysates were prepared and subjected to immunoblot analysis for activated STAT3 (pY705), total STAT3, activated ERK1/2 (pT202/pY204), total ERK1/2 or α-tubulin as a loading control. (B & D) Quantitative analysis of the immunoblots for STAT3 and STAT3 pY705 (n = 3) was performed using ImageJ software, where data represent the mean ± SEM. Asterisks denote statistically signiﬁcant differences as per the legend to Fig. 1B.

Fig. 8B). Further, H2O2 pre-treatment for 15 min resulted in higher HIV-1 Rev protein [26], were also analysed. All of the nuclear import STAT3 nuclear accumulation 60 min after LIF addition compared to cargoes showed signiﬁcantly reduced (p ≤ 0.001) nuclear accumula- cells not pre-treated with H2O2. Experiments were also performed tion, in contrast to GFP alone which showed no effect of H2O2 treatment, where H2O2 was added 15 min post-LIF treatment, conditions which implying no impact of oxidative stress on passive diffusion through the again mislocalised Ran and importin-α3 to the cytoplasm and nucleus nuclear pore (Fig. 10A, and quantitative data in Fig. 10B). We conﬁrmed respectively (Fig. 8C, and quantitative data in Fig. 8D); in this case, that CRM1-mediated nuclear export of GFP-Rev-NES was sensitive to

H2O2 treatment post-LIF addition prolonged the high levels of STAT3 the CRM1 inhibitor leptomycin B (Supp. Fig. 8). GFP-Rev-NES was unaf- in the nucleus (Fig. 8C and D; Fn/c ~4–5) up to 60 min. Evaluation of fected in the presence of H2O2 (Fig. 10), implying that nuclear export pY705 STAT3 levels under these different conditions confirmed was still functional and this was further confirmed through the evalua- sustained phosphorylation of STAT3 (Fig. 9) as the likely basis of the tion of the export of endogenous RanBP1 that continued in the presence prolonged nuclear retention of STAT3. of H2O2 (Supp. Fig. 9). Thus, the main impact on nucleocytoplasmic To confirm that oxidative stress-induced alterations in the nuclear transport of Ran mislocalisation due to H2O2 treatment appears to be transport machinery impacted on cellular nuclear transport generally, on importin-dependent nuclear import rather than nuclear export. we examined the effect of H2O2 on the subcellular localisation of the NLS-containing GFP-tagged nuclear import cargoes GFP-conjugated 4. Discussion SV40 large T-antigen (GFP-T-ag) and human cytomegalovirus polymerase UL54 (GFP-pUL54), representative of importin-α/β1-dependent Cytokine-stimulated STAT3 phosphorylation has been typically transport [27,28], and the importin-β1-specific nuclear import cargo shown to be both rapid and transient under normal, non-stressed con- GFP-VP3 (chicken anaemia virus viral protein 3) [29,30]; since VP3 ditions. Detailed studies both in vitro and in vivo have revealed a coor- also contains a CRM1 (exportin 1)-recognised nuclear export signal dinated negative regulatory network contributing to the kinetics of this (NES), we used a NES-inactivated mutant derivative, whilst a VP3 con- activation, highlighting the importance of proteins such as TC-PTP [9] struct containing an intact NES, in addition to the NLS/NES-containing and SOCS3 [11]. A persistent activation of JAK/STAT3 signalling has

Fig. 8. Oxidative stress impacts on LIF-dependent STAT3 nuclear accumulation either pre- or post-cytokine addition. (A & B) MEFs were pre-treated with 1 mM H2O2 for the indicated times prior to exposure to 10 ng/ml LIF for 15 or 60 min in the continued presence of H2O2. (A) Confocal laser scanning microscopy analysis was performed on cells immunostained using speciﬁc antibodies for STAT3, Ran and importin-α3(IMP-α3) as per the legends to Figs. 5Aand7. (B) Quantitative results for STAT3, Ran and importin-α3 localisation using ImageJ software to determine the Fn/c as per the legend to Fig. 5B; data represent the mean ± SEM. Asterisks denote statistically signiﬁcant differences for cells analysed under control conditions

(**, p ≤ 0.01; ***, p ≤ 0.001). (C & D) MEFs were pre-treated with 10 ng/ml LIF for 15 min prior to exposure to 1 mM H2O2 (15–45 min as indicated) in the continued presence of LIF. (C) Confocal laser scanning microscopic analysis was performed on immunostained cells as per (A & B) above. (D) Quantitative results for STAT3, Ran and importin-α3 localisation using ImageJ software to determine the Fn/c as per the legend to Fig. 5B; data represent the mean ± SEM. (A & C) Bars represent 10 μm. 492 I.H.W. Ng et al. / Biochimica et Biophysica Acta 1843 (2014) 483–494

GFP-T-ag GFP-UL54 GFP-VP3 GFP-VP3 A GFP GFP-Rev-NES (114-135) (1145-1161) (74-121) NES mut

Control

1 mM H2O2 (60 min)

B ∗∗∗ 15 ∗∗∗ -H 2O2

n = + H2O2 51 n = 49 ∗∗∗ 10 ∗∗∗ GFP n = n = (Fn/c) 50 60 5 n = n = 55 46 n = n = n = n = 53 47 n = n = 48 52 43 49 0 GFPGFP GFP GFP GFP GFP Rev-NES T-ag UL54 VP3 VP3 NES mut

Fig. 10. Oxidative stress inhibits importin-α/β-dependent nuclear protein import. MEFs were transiently transfected to express GFP-tagged cargo proteins as indicated prior to treatment with 1 mM H2O2 for 60 min. (A) Live cell confocal laser scanning microscopic imaging. Bar represents 10 μm. (B) Quantitative results for the various GFP-tagged proteins using ImageJ software to determine the Fn/c as per the legend to Fig. 5B; data represent the mean ± SEM. Asterisks denote statistically significant differences for treated compared to untreated cells. been implicated in the disease progression, and so a deeper understand- As summarized in Fig. 11A, the key cytokine-stimulated events regulat- ing of the molecular drivers of this aberrant regulation is urgently ing STAT3 have been well described. However, as summarized in required. Oxidative stress induced by H2O2, superoxide or ultraviolet Fig. 11B, we have shown the simultaneous exposure to H2O2 and the radiation can activate STAT3 in the absence of cytokine stimulation cytokine LIF drives a striking and persistent phosphorylation of STAT3, [15,16,44,45]. Our studies here have addressed for the first time the but that the activation of SOCS3 expression as a direct target gene of impact of oxidative stress on cytokine-driven STAT3 signalling events. STAT3 is slowed and the nuclear import machinery exemplified by

Fig. 11. Schematic diagram of the impact of oxidative stress on cytokine-driven STAT3 activation. (A) Under normal (non-stressed) conditions, the cytokine LIF activates the JAK-STAT3 signal transduction pathway in which negative regulators such as TC-PTP and SOCS3 limit the duration of signalling events. (B) During oxidative stress, multiple regulatory events of the STAT3 pathway are altered: TC-PTP is mislocalised and its levels are downregulated, Ran and importin-α3 distributions are altered. The result of these changes is seen in the sustained phosphorylation and nuclear localisation of STAT3. I.H.W. Ng et al. / Biochimica et Biophysica Acta 1843 (2014) 483–494 493 cytosolic/nuclear distributions of importin-α3 and Ran is disrupted. In The significance of our findings is reinforced by the intense interest combination, these changes slow the nuclear accumulation of STAT3 directed towards STAT3 as part of altered signalling events in a range but also allow for its persistent nuclear retention. of pathologies, and a concerted effort to define signalling nodes Studies on the impact of stress on nuclear transport events have for therapeutic targeting [58,59]. Some studies have suggested that previously defined changes in importin localisation and the Ran oxidative stress per se can induce STAT3 Y705 phosphorylation as a gradient [46–49]. Importantly, these disruptions noted under the downstream consequence of growth factor activation [34,60].Our conditions of ultraviolet radiation, oxidative stress, and heat shock results reveal a requirement for JAKs and Src family kinases in the acti- are predicted to have a profound influence on the nuclear transport vation of STAT3 driven by oxidative stress and cytokine stimulation and of the wide range of nuclear-destined cargoes dependent on suggest that targeting these kinases will contribute to alleviating detri- these importin proteins. Previous studies have shown that STAT3 mental consequences of STAT3 activation noted in disease. Taken nuclear import relies on importin-α3 that interacts with STAT3's together, the results presented in our study reveal that the disruption NLS (150RKRQDLEQKMK162 within the coiled-coil domain [42]), of several regulatory mechanisms leads to a delayed nuclear accumula- importin-β1, and a Ran gradient [41]. Here, we demonstrate the tion and transcriptional response by STAT3. striking redistribution of importin-α3 and Ran, as well as a relocalisation of the phosphatase TC-PTP and a delay in the nuclear Acknowledgements entry of STAT3 during oxidative stress. Paradoxically, a cytokine- driven nuclear entry of STAT3 remains possible despite the We thank V. Poli (University of Turin, Italy) for the wild-type and importin-α3 and Ran disruptions and can be observed even when STAT3−/− murine embryonic fibroblasts. This work was supported by an the oxidative stress insult precedes cytokine stimulation. Australian Research Council (ARC) Discovery Grant (DP130100804). With such disruptions to the classical nuclear import pathway, the IHWN is a recipient of a Monash University: Monash Graduate Scholar- possibility for a non-Ran dependent nuclear transport mechanism for ship and Faculty of Medicine International Postgraduate Research critical proteins such as the STAT3 transcription factor requires careful Scholarship and DAJ is an NHMRC Senior Principal Research Fellow consideration. Activated STAT3 can accumulate in the nucleus through (#APP1002486). non-canonical nuclear import pathways [50–53]. Indeed, an import pathway requiring receptor-mediated endocytosis trafficking has been Appendix A. Supplementary data proposed as critical for signalling by cytokines that are weak STAT3 activators. Thus, the formation of signalling endosomes protects and Supplementary data to this article can be found online at http://dx. prolongs the activation of signalling proteins, but this is not required doi.org/10.1016/j.bbamcr.2013.11.015. for the potent activation of STAT3 by interleukin-6 family cytokines [51–53]. Alternative stress-evoked mechanisms have been proposed References as critical for the nuclear transport of Hsp70 during heat stress [54,55]; whether these may apply to other types of stress, including [1] J.E. Darnell Jr., I.M. Kerr, G.R. Stark, Jak-STAT pathways and transcriptional activation in response to IFNs and other extracellular signaling proteins, Science 264 (1994) oxidative stress, remains to be determined. This will be particularly im- 1415–1421. portant in understanding how critical mediators of stress responses [2] C. Mertens, J.E. Darnell Jr., SnapShot: JAK-STAT signaling, Cell 131 (2007) 612. such as signalling proteins and transcription factors can continue to [3] D.E. Levy, J.E. Darnell Jr., Stats: transcriptional control and biological impact, Nat. Rev. Mol. Cell Biol. 3 (2002) 651–662. enter the nucleus of the stressed cell. Indeed, a number of protein [4] P.C. Heinrich, I. Behrmann, S. Haan, H.M. Hermanns, G. Muller-Newen, F. Schaper, kinases are reported to show enhanced nuclear retention in response Principles of interleukin (IL)-6-type cytokine signalling and its regulation, Biochem. to stress [56], suggesting their nuclear entry will also require mecha- J. 374 (2003) 1–20. [5] V. Poli, STAT3 function in vivo, in: P.B. Sehgal, D.E. Levy, T. Hirano (Eds.), Signal nisms that are not reliant on the classical Ran-dependent importin Transducers and Activators of Transcription (STATs), Kluwer Academic Publishers, pathways. The Netherlands, 2003, pp. 492–512. The sustained nuclear retention of cytokine-activated STAT3 we ob- [6] H. Yu, R. Jove, The STATs of cancer-new molecular targets come of age, Nat. Rev. – served during oxidative stress also challenges our understanding of the Cancer 4 (2004) 97 105. [7] R. Garcia, T.L. Bowman, G. Niu, H. Yu, S. Minton, C.A. Muro-Cacho, C.E. Cox, R. STAT3 export pathway. Importantly, the dephosphorylation of STAT3 Falcone, R. Fairclough, S. Parsons, A. Laudano, A. Gazit, A. Levitzki, A. Kraker, R. Y705 is required for its nuclear export by CRM1 [57]. The expression Jove, Constitutive activation of Stat3 by the Src and JAK tyrosine kinases participates and localisation of TC-PTP are critical regulators of nuclear STAT3 in growth regulation of human breast carcinoma cells, Oncogene 20 (2001) 2499–2513. Y705 phosphorylation [9,35,38,39]. The mislocalisation of TC-PTP has [8] T. Mustelin, T. Vang, N. Bottini, Protein tyrosine phosphatases and the immune been reported previously for cells subjected to hyperosmotic stress response, Nat. Rev. Immunol. 5 (2005) 43–57. [25]. In that study, as the hyperosmotic stress-induced nuclear exit of [9] D. Xu, C.K. Qu, Protein tyrosine phosphatases in the JAK/STAT pathway, Front. Biosci. 13 (2008) 4925–4932. TC-PTP was not inhibited by leptomycin B, the nuclear export of TC- [10] U. Vinkemeier, Getting the message across, STAT! Design principles of a molecular PTP appeared to be independent of the exportin CRM1 [25]. Similarly, signaling circuit, J. Cell Biol. 167 (2004) 197–201. we observed TC-PTP mislocalisation during oxidative stress by tracking [11] J.J. Babon, N.A. Nicola, The biology and mechanism of action of suppressor of cytokine signaling 3, Growth Factors 30 (2012) 207–219. the localisation of myc-tagged TC-PTP. Although further studies will [12] N. Khaper, S. Bryan, S. Dhingra, R. Singal, A. Bajaj, C.M. Pathak, P.K. Singal, Targeting require more extensive evaluation of factors impacting on the the vicious inflammation-oxidative stress cycle for the management of heart failure, localisation of the endogenous TC-PTP under conditions of oxidative Antioxid. Redox Signal. 13 (2010) 1033–1049. [13] L.A. Sena, N.S. Chandel, Physiological roles of mitochondrial reactive oxygen species, stress (i.e., rather than following exogenously expressed myc-tagged Mol. Cell 48 (2012) 158–167. TC-PTP protein, anti-TC-PTP antibodies suitable for use in immunostain- [14] C. Zgheib, M. Kurdi, F.A. Zouein, B.W. Gunter, B.A. Stanley, J. Zgheib, D.G. Romero, S.B. ing protocols will be needed), our observed loss of the endogenous TC- King, N. Paolocci, G.W. Booz, Acyloxy nitroso compounds inhibit LIF signaling in PTP protein during oxidative stress, adds to the key stress-stimulated endothelial cells and cardiac myocytes: evidence that STAT3 signaling is redox-sensitive, PLoS One 7 (2012) e43313. events impeding STAT3 dephosphorylation in the nucleus (see [15] S.K. Park, M.K. Dahmer, M.W. Quasney, MAPK and JAK-STAT signaling pathways are Fig. 11B). Our analysis of the nuclear export of the CRM1-dependent involved in the oxidative stress-induced decrease in expression of surfactant protein – cargoes GFP-Rev-NES or RanBP1 clearly demonstrated that impairment genes, Cell. Physiol. Biochem. 30 (2012) 334 346. fi [16] G. Millonig, I. Ganzleben, T. Peccerella, G. Casanovas, L. Brodziak-Jarosz, K. of the Ran gradient was not suf cient to affect their predominantly Breitkopf-Heinlein, T.P. Dick, H.K. Seitz, M.U. Muckenthaler, S. Mueller, Sustained cytoplasmic localisations. Thus, the prolonged STAT3 phosphorylation submicromolar H2O2 levels induce hepcidin via signal transducer and activator of would appear to be the key to the enhanced nuclear localisation of transcription 3 (STAT3), J. Biol. Chem. 287 (2012) 37472–37482. [17] B.G. Szczepankiewicz, C. Kosogof, L.T. Nelson, G. Liu, B. Liu, H. Zhao, M.D. Serby, Z. STAT3 under the conditions of combined cytokine and oxidative stress Xin, M. Liu, R.J. Gum, D.L. Haasch, S. Wang, J.E. Clampit, E.F. Johnson, T.H. Lubben, stimulation. M.A. Stashko, E.T. Olejniczak, C. Sun, S.A. Dorwin, K. Haskins, C. Abad-Zapatero, 494 I.H.W. Ng et al. / Biochimica et Biophysica Acta 1843 (2014) 483–494

E.H. Fry, C.W. Hutchins, H.L. Sham, C.M. Rondinone, J.M. Trevillyan, [39] T. Yamamoto, Y. Sekine, K. Kashima, A. Kubota, N. Sato, N. Aoki, T. Matsuda, The Aminopyridine-based c-Jun N-terminal kinase inhibitors with cellular activity and nuclear isoform of protein-tyrosine phosphatase TC-PTP regulates interleukin- minimal cross-kinase activity, J. Med. Chem. 49 (2006) 3563–3580. 6-mediated signaling pathway through STAT3 dephosphorylation, Biochem. [18] M.F. Favata, K.Y. Horiuchi, E.J. Manos, A.J. Daulerio, D.A. Stradley, W.S. Feeser, D.E. Biophys. Res. Commun. 297 (2002) 811–817. Van Dyk, W.J. Pitts, R.A. Earl, F. Hobbs, R.A. Copeland, R.L. Magolda, P.A. Scherle, [40] S. Huang, I.S. Chang, W. Lin, W. Ye, R.Z. Luo, Z. Lu, Y. Lu, K. Zhang, W.S. Liao, T. Tao, J.M. Trzaskos, Identification of a novel inhibitor of mitogen-activated protein kinase R.C. Bast Jr., X. Chen, Y. Yu, ARHI (DIRAS3), an imprinted tumour suppressor gene, kinase, J. Biol. Chem. 273 (1998) 18623–18632. binds to importins and blocks nuclear import of cargo proteins, Biosci. Rep. 30 [19] A. Cuenda, J. Rouse, Y.N. Doza, R. Meier, P. Cohen, T.F. Gallagher, P.R. Young, J.C. Lee, (2010) 159–168. SB 203580 is a specific inhibitor of a MAP kinase homologue which is stimulated by [41] V. Cimica, H.C. Chen, J.K. Iyer, N.C. Reich, Dynamics of the STAT3 transcription factor: cellular stresses and interleukin-1, FEBS Lett. 364 (1995) 229–233. nuclear import dependent on Ran and importin-beta1, PLoS One 6 (2011) e20188. [20] J.E. Thompson, R.M. Cubbon, R.T. Cummings, L.S. Wicker, R. Frankshun, B.R. [42] L. Liu, K.M. McBride, N.C. Reich, STAT3 nuclear import is independent of tyrosine Cunningham, P.M. Cameron, P.T. Meinke, N. Liverton, Y. Weng, J.A. DeMartino, phosphorylation and mediated by importin-alpha3, Proc. Natl. Acad. Sci. U. S. A. Photochemical preparation of a pyridone containing tetracycle: a Jak protein kinase 102 (2005) 8150–8155. inhibitor, Bioorg. Med. Chem. Lett. 12 (2002) 1219–1223. [43] R. Ushijima, N. Sakaguchi, A. Kano, A. Maruyama, Y. Miyamoto, T. Sekimoto, Y. [21] R.A. Blake, M.A. Broome, X. Liu, J. Wu, M. Gishizky, L. Sun, S.A. Courtneidge, SU6656, Yoneda, K. Ogino, T. Tachibana, Extracellular signal-dependent nuclear import of a selective src family kinase inhibitor, used to probe growth factor signaling, Mol. STAT3 is mediated by various importin alphas, Biochem. Biophys. Res. Commun. Cell. Biol. 20 (2000) 9018–9027. 330 (2005) 880–886. [22] J. Schust, B. Sperl, A. Hollis, T.U. Mayer, T. Berg, Stattic: a small-molecule inhibitor of [44] M. Carballo, M. Conde, R. El Bekay, J. Martin-Nieto, M.J. Camacho, J. Monteseirin, J. STAT3 activation and dimerization, Chem. Biol. 13 (2006) 1235–1242. Conde, F.J. Bedoya, F. Sobrino, Oxidative stress triggers STAT3 tyrosine phosphoryla- [23] M.C. Morris, S. Deshayes, F. Heitz, G. Divita, Cell-penetrating peptides: from molec- tion and nuclear translocation in human lymphocytes, J. Biol. Chem. 274 (1999) ular mechanisms to therapeutics, Biol. Cell 100 (2008) 201–217. 17580–17586. [24] M.C. Morris, J. Depollier, J. Mery, F. Heitz, G. Divita, A peptide carrier for the delivery [45] T. Bito, N. Sumita, T. Masaki, T. Shirakawa, M. Ueda, R. Yoshiki, Y. Tokura, C. of biologically active proteins into mammalian cells, Nat. Biotechnol. 19 (2001) Nishigori, Ultraviolet light induces Stat3 activation in human keratinocytes and 1173–1176. fibroblasts through reactive oxygen species and DNA damage, Exp. Dermatol. 19 [25] M.H. Lam, B.J. Michell, M.T. Fodero-Tavoletti, B.E. Kemp, N.K. Tonks, T. Tiganis, Cellu- (2010) 654–660. lar stress regulates the nucleocytoplasmic distribution of the protein-tyrosine phos- [46] M. Kodiha, A. Chu, N. Matusiewicz, U. Stochaj, Multiple mechanisms promote the in- phatase TCPTP, J. Biol. Chem. 276 (2001) 37700–37707. hibition of classical nuclear import upon exposure to severe oxidative stress, Cell [26] M.H. Lam, B. Henderson, M.T. Gillespie, D.A. Jans, Dynamics of leptomycin Death Differ. 11 (2004) 862–874. B-sensitive nucleocytoplasmic flux of parathyroid hormone-related protein, Traffic [47] Y. Miyamoto, T. Saiwaki, J. Yamashita, Y. Yasuda, I. Kotera, S. Shibata, M. Shigeta, Y. 2(2001)812–819. Hiraoka, T. Haraguchi, Y. Yoneda, Cellular stresses induce the nuclear accumulation [27] S. Hubner, C.Y. Xiao, D.A. Jans, The protein kinase CK2 site (Ser111/112) enhances of importin alpha and cause a conventional nuclear import block, J. Cell Biol. 165 recognition of the simian virus 40 large T-antigen nuclear localization sequence (2004) 617–623. by importin, J. Biol. Chem. 272 (1997) 17191–17195. [48] N. Crampton, M. Kodiha, S. Shrivastava, R. Umar, U. Stochaj, Oxidative stress inhibits [28] G. Alvisi, A. Ripalti, A. Ngankeu, M. Giannandrea, S.G. Caraffi, M.M. Dias, D.A. Jans, nuclear protein export by multiple mechanisms that target FG nucleoporins and Human cytomegalovirus DNA polymerase catalytic subunit pUL54 possesses inde- Crm1, Mol. Biol. Cell 20 (2009) 5106–5116. pendently acting nuclear localization and pp UL44 binding motifs, Traffic 7 (2006) [49] Y. Yasuda, Y. Miyamoto, T. Saiwaki, Y. Yoneda, Mechanism of the stress-induced 1322–1332. collapse of the Ran distribution, Exp. Cell Res. 312 (2006) 512–520. [29] H.V. Kuusisto, K.M. Wagstaff, G. Alvisi, D.A. Jans, The C-terminus of apoptin repre- [50] J.L. Lee, M.J. Wang, J.Y. Chen, Acetylation and activation of STAT3 mediated by sents a unique tumor cell-enhanced nuclear targeting module, Int. J. Cancer 123 nuclear translocation of CD44, J. Cell Biol. 185 (2009) 949–957. (2008) 2965–2969. [51] S. Kermorgant, P.J. Parker, Receptor trafficking controls weak signal delivery: a strat- [30] I.K. Poon, C. Oro, M.M. Dias, J. Zhang, D.A. Jans, Apoptin nuclear accumulation is egy used by c-Met for STAT3 nuclear accumulation, J. Cell Biol. 182 (2008) 855–863. modulated by a CRM1-recognized nuclear export signal that is active in normal [52] P.B. Sehgal, Paradigm shifts in the cell biology of STAT signaling, Semin. Cell Dev. but not in tumor cells, Cancer Res. 65 (2005) 7059–7064. Biol.19(2008)329–340. [31] Y. Yashiroda, M. Yoshida, Nucleo-cytoplasmic transport of proteins as a target for [53] A.H. Bild, J. Turkson, R. Jove, Cytoplasmic transport of Stat3 by receptor-mediated therapeutic drugs, Curr. Med. Chem. 10 (2003) 741–748. endocytosis, EMBO J. 21 (2002) 3255–3263. [32] I.H. Ng, D.C. Ng, D.A. Jans, M.A. Bogoyevitch, Selective STAT3-alpha or -beta [54] N. Imamoto, S. Kose, Heat-shock stress activates a novel nuclear import pathway expression reveals spliceform-specific phosphorylation kinetics, nuclear retention mediated by Hikeshi, Nucleus 3 (2012) 422–428. and distinct gene expression outcomes, Biochem. J. 447 (2012) 125–136. [55] S. Kose, M. Furuta, N. Imamoto, Hikeshi, a nuclear import carrier for Hsp70s, protects [33] T. Decker, P. Kovarik, Serine phosphorylation of STATs, Oncogene 19 (2000) cells from heat shock-induced nuclear damage, Cell 149 (2012) 578–589. 2628–2637. [56] M. Kodiha, U. Stochaj, Nuclear transport: a switch for the oxidative stress-signaling [34] G. Sun, D.J. Kemble, To C or not to C: direct and indirect redox regulation of Src circuit? J. Signal Transduct. 2012 (2012) 208650. protein tyrosine kinase, Cell Cycle 8 (2009) 2353–2355. [57] Y. Huang, J. Qiu, S. Dong, M.S. Redell, V. Poli, M.A. Mancini, D.J. Tweardy, Stat3 iso- [35] B.J. Shields, F. Wiede, E.N. Gurzov, K. Wee, C. Hauser, H.J. Zhu, T.J. Molloy, S.A. forms, alpha and beta, demonstrate distinct intracellular dynamics with prolonged O'Toole, R.J. Daly, R.L. Sutherland, C.A. Mitchell, C.A. McLean, T. Tiganis, TCPTP regu- nuclear retention of Stat3beta mapping to its unique C-terminal end, J. Biol. Chem. lates SFK and STAT3 signaling and is lost in triple-negative breast cancers, Mol. Cell. 282 (2007) 34958–34967. Biol. 33 (2013) 557–570. [58] E. Bournazou, J. Bromberg, Targeting the tumor microenvironment: JAK-STAT3 [36] T. Yi, D. Lindner, The role and target potential of protein tyrosine phosphatases in signaling, JAK-STAT 2 (2013) e23828. cancer, Curr. Oncol. Rep. 10 (2008) 114–121. [59] P. Sansone, J. Bromberg, Targeting the interleukin-6/Jak/stat pathway in human [37] C. Blanchetot, L.G. Tertoolen, J. den Hertog, Regulation of receptor protein-tyrosine malignancies, J. Clin. Oncol. 30 (2012) 1005–1014. phosphatase alpha by oxidative stress, EMBO J. 21 (2002) 493–503. [60] S. Filosto, E.M. Khan, E. Tognon, C. Becker, M. Ashfaq, T. Ravid, T. Goldkorn, EGF re- [38] R. Wakahara, H. Kunimoto, K. Tanino, H. Kojima, A. Inoue, H. Shintaku, K. Nakajima, ceptor exposed to oxidative stress acquires abnormal phosphorylation and aberrant Phospho-Ser727 of STAT3 regulates STAT3 activity by enhancing dephosphorylation activated conformation that impairs canonical dimerization, PLoS One 6 (2011) of phospho-Tyr705 largely through TC45, Genes Cells 17 (2012) 132–145. e23240.