Conceived and designed the experiments: Y-MG DPJ. Performed the experiments: Y-MG S-MK JRR MO. Analyzed the data: Y-MG DPJ. Contributed reagents/materials/analysis tools: S-MK DPJ. Wrote the paper: Y-MG DPJ.
The authors have declared that no competing interests exist.
Cell culture studies show that the antioxidant thiol protein, thioredoxin-1 (Trx1), translocates to cell nuclei during stress, facilitates DNA binding of transcription factors NF-κB and glucocorticoid receptor (GR) and potentiates signaling in immune cells. Excessive proinflammatory signaling
Transgenic mice (Tg) expressing human Trx1 (hTrx1) with added nuclear localization signal (NLS) showed broad tissue expression and nuclear localization. The role of nuclear Trx1 in inflammatory signaling was examined in Tg and wild-type (WT) mice following infection with influenza (H1N1) virus. Results showed that Tg mice had earlier and more extensive NF-κB activation, increased TNF-α and IL-6 expression, greater weight loss, slower recovery and increased mortality compared to WT. Decreased plasma glutathione (GSH) and oxidized plasma GSH/GSSG redox potential (EhGSSG) following infection in Tg mice showed that the increased nuclear thiol antioxidant caused a paradoxical downstream oxidative stress. An independent test of this nuclear reductive stress showed that glucocorticoid-induced thymocyte apoptosis was increased by NLS-Trx1.
Increased Trx1 in cell nuclei can increase severity of disease responses by potentiation of redox-sensitive transcription factor activation.
Thioredoxin-1 (Trx-1) and glutathione (GSH) are central thiol redox systems in cell nuclei and cytoplasm but are differentially regulated in these compartments. For instance, during cell stress induced by nutrient deprivation, proinflammatory signals, oxidants or reactive electrophiles, the nuclear pools are more resistant to oxidation or depletion
Compartmental regulation of NF-κB, AP-1 and Nrf-2 involves opposing redox-sensitive steps in cytoplasm and nuclei, i.e., 1) upstream cytoplasmic oxidative activation involving kinase signaling and 2) downstream Trx1-dependent reduction of the Cys of the DNA-binding domain
Modulation of NF-κB signaling by nuclear Trx1 raises the possibility that excessive nuclear Trx1 could cause hyper-responsive immune signaling. NF-κB activation is induced by viruses and viral products, including influenza and HIV, as well as other stimuli associated with oxidative stress (free radicals, UV light, gamma-irradiation)
Cytoplasmic redox regulation of GR differs from NF-κB, but increased nuclear Trx1 could similarly contribute to excessive transcriptional activity. GR is a transcription factor belonging to a family of nuclear receptors
Many studies show that oxidative inactivation can disrupt regulation of transcription and contribute to disease
All protocols involving mice were reviewed and approved by the Institutional Animal Care and Use Committee at Emory University.
All animal experiments and husbandry for the studies presented in this manuscript were conducted under the review and approval of the Emory University Institutional Animal Care and Use Committee (IACUC, approval ID: DAR-2000040-062113). Emory IACUC operates under the federal Animal Welfare Law (administered by the USDA) and regulations of the Department of Health and Human Services. Mice were maintained in the Emory University Division of Animal Resources Facility under a standard 12-h light/dark cycle and had free access to tap water and standard laboratory diet. C57BL/6 (WT) mice were purchased from Charles River Laboratories and used to maintain the transgenic mouse colony, with WT mice in experiments being littermates of transgenic mice (Tg) expressing the human Trx1 (hTrx1) gene. The hTrx1 was modified to contain c-Myc epitope tag at the N-terminus and nuclear localization signal (NLS) at the C-terminus (defined as NLS-hTrx1 Tg). Preliminary studies using transient transfection showed that NLS-Trx1 constructs with 1 or 3 copies of the NLS and a Myc epitope tag were expressed in cells in culture and showed expected activity in experiments with an NF-κB reporter assay. NLS-Trx1 constructs were used by the Emory University Transgenic Mouse Core to create a total of 6 transgenic lines. Briefly, Trx1 cDNA was inserted between the cytomegalovirus (CMV) promoter and the simian virus (SV-40) terminator of pCMV-Myc mammalian expression vector purchased from Clontech Laboratories (Mountain View, CA). The transgenic construct was microinjected into the pronucleus of fertilized eggs from C57BL/6 mice. The presence of the NLS-hTrx1 transgene was confirmed by PCR analysis using mouse genomic DNA prepared from tail biopsy as a template and the following oligonucleotide primers: forward primer,
Total tissue mRNA was isolated using RNeasy Mini kit purchased from Qiagen (Valencia, CA) following the manufacturer's protocol, and reverse transcription was performed to generate cDNA (Clontech Laboratories, Mountain View, CA). For quantitative real-time PCR, amplification was performed on an iCycler IQ Multicolor Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA) for 35 cycles as follows: 95°C for 30 seconds, 58°C for 30 seconds, and 72°C for 1 min. Quantification and melting curves were analyzed with iCycler software. Primers for cytokines were designed using a program provided by Integrated DNA Technologies (Coralville, Iowa). Details of PCR primer sequences and product sizes for mouse cytokines used in the analyses of extracted RNA for quantitative real-time PCR are as follows; IL-1β (180 bp), forward:
Subcellular fractionation of lung and kidney tissues was performed using Qproteome kit (Qiagen, Valencia, CA) following the procedures provided by the manufacturer. These organ systems were examined because of the expected sensitivity of lung to influenza virus and because of the abundance of NLS-Trx1 in the kidney. Twenty mg of kidney and lung from WT and Tg were processed to isolate cytoplasmic and nuclear fractions. Isolated fractions were then confirmed by western blotting probed with antibodies (ab) against GAPDH and lamin B for cytoplasm and nuclei, respectively. To examine abundance of other antioxidant proteins by western blotting, antibodies used were as follows: Myc (Cell Signaling Technology, Danvers, MA); hTrx1 and mTrx1 (AbFrontier, Seoul, Korea); Trx2 (antisera produced by Covance, Princeton, NJ); Prx1, Prx2 and Prx3 (Abcam, Cambridge, MA); Txnip (Invitrogen, Carlsbad, CA); actin (Sigma-Aldrich, St. Louis, MO); Alexa-Fluor-680-conjugated anti-rabbit or anti-mouse secondary antibody (Invitrogen). A band corresponding to each protein was visualized using an Odyssey scanner and Odyssey 2.1 software (Li-Cor, Lincoln, NE).
Cells isolated from WT and Tg kidneys were plated on glass cover slips. After 1 d, cells were washed, fixed, stained with anti-Myc ab followed by Cy3-conjugated goat anti-rabbit ab (Jackson Immuno Research, West Grove, PA) to visualize nuclear localization of NLS-hTrx1. Alexa Fluor 488 phalloidin was used to stain actin. Immunofluorescence was visualized using an Olympus X-70 fluorescence microscope system.
Nuclear extracts prepared from lungs of Tg and WT mice prior to and 3 d post infection were analyzed using an EMSA kit (Affymetrics, Clara, CA) by incubating a biotin-labeled or unlabeled probe containing an NF-κB DNA-binding consensus sequence (
HeLa cells purchased from ATCC were maintained with 10% FBS (37°C, 5% CO2) in DMEM. Cells at 80% confluence were co-transfected with pcDNA3.1, Trx1 WT, or Trx1 C35S
Plasma and lung tissue collected from Tg and WT mice 3 d after infection or with no infection were analyzed for GSH, glutathione disulfide (GSSG), Cys, and cystine (CySS) by high performance liquid chromatography (HPLC) with fluorescence detection
H1N1 influenza virus (A/California/04/2009) was kindly provided by Dr. Richard Webby. The virus was grown in 11-d old embryonated hen's eggs for 2 d at 37°C, harvested and stored at −80°C until use. Mice were infected with serial dilutions of A/California/04/2009 virus and the 50% lethal dose (LD50) was determined as described
Wild-type and NLS-hTrx1 transgenic mice (8 to 10 weeks) were used for infection. Isoflurane-anesthetized mice (18 WT, 17 Tg) were intranasally infected with 1× LD50 of A/California/04/2009 virus (1× the 50% lethal dose, LD50) in 50 µl of phosphate-buffered saline (PBS). Mice were observed daily to monitor changes in body weight and to record mortality (25% loss in body weight as the Institutional Animal Care and Use Committee (IACUC) endpoint for morbidity). For the analysis of lungs at early time points, additional groups of mice (6 mice for Tg, 5 mice for WT) were infected and euthanized at 2- or 3-d post infection to collect lung samples.
Thymocytes, from 4 to 6 week-old Tg and WT mice, were enriched by passage through nylon wool columns. The effects of glucocorticoid (dexamethasone, Sigma-Aldrich Corp, St. Louis, MO)-induced apoptosis were evaluated. Cell survival and death were quantified by measuring absorbance of the dye product from the nonradioactive quantitative reagent WST-1 (Roche, Basel, Switzerland) and by cell counts using a hemocytometer.
Statistical comparisons of data were carried out using the t-test of the OriginLab (Data Analysis and Graphing software, OriginLab Co.). P<0.05 was considered to be significant.
To examine tissue distribution of NLS-hTrx1 mRNA, cDNA converted from total RNA isolated from tissues of littermates with transgene-positive and -negative genotypes. Results showed greater abundance of mRNA for NLS-hTrx1 in heart, kidney, skeletal muscle, and thymus, and lower amounts in liver, lung, small intestine, and spleen (
Different organs harvested from Tg and littermate WT mice were analyzed to examine mRNA abundance of nuclear compartment-targeted hTrx1 and β-actin. Each cDNA converted from purified tissue RNA (1 µg) by reverse transcription was analyzed by real-time PCR. Quantified values (picomole) are represented as mean ± SE for 4 Tg and 4 WT mice. Symbols are as follows; H, heart; K, kidney; Li, liver; Lu, lung; SM, skeletal muscle; SI, small intestine; Sp, spleen; Th, thymus.
Characterization of the abundance and localization of the NLS-hTrx1 transgene product was facilitated by inclusion of the c-Myc epitope, which allowed discrimination from the endogenous mouse Trx1. Subcellular localization studies were performed on kidney because of the relative abundance of the transgenic product in kidney, the ability to obtain adherent cells for fluorescence microscopy and the ability to obtain quality nuclear preparations. Immunofluorescence studies of cells isolated from kidney (
Cells isolated from kidneys of Tg and WT were analyzed to examine localization by immunofluorescence (A) and western blotting (B). A, Nuclear compartmentalized expression of NLS-hTrx1 was visualized by myc antibody followed by Cy3 (orange, Tg). As a control, actin distribution in all area of cells was obtained by Alexa Fluor 488 phalloidin, shown for both Tg and WT (green). Subcellular fractions including cytoplasm and nuclei were obtained from kidney tissues of Tg and WT. Western blotting using each fraction confirmed nuclear localization of NLS-hTrx1 in Tg. Western blot analysis of lamin and GAPDH were used as markers to verify purity of nuclear and cytoplasmic fractions, respectively.
Western blot analysis of cytosolic and nuclear fractions of kidney (
We next examined whether expression of NLS-hTrx1 altered the abundance of other antioxidant proteins including mTrx1, mitochondrial Trx2, nuclear and cytoplasmic peroxiredoxins (Prx1, Prx2), mitochondrial Prx3, and thioredoxin-interacting protein (Txnip). Expression of hTrx1 had no effect on the abundance of mouse Trx1 but approximately doubled the total amount of Trx1 (
Protein abundance was examined by western blot analysis with antibodies specific to Trx1 (NLS-hTrx1 and mTrx1), Prx1, Prx2, Prx3, Txnip, and actin (A). Western blots show representative data of kidney tissues obtained from 3 Tg mice and littermate WT controls. B, Plasma base line concentrations of cysteine (Cys), cystine (CySS), glutathione (GSH), and glutathione disulfide (GSSG) in Tg and WT mice were determined by HPLC analyses. Redox potentials of Cys/CySS (EhCySS) and GSH/GSSG (EhGSSG) were calculated by using the Nernst equation
To determine whether increased nuclear Trx1 had a systemic effect on redox state, we analyzed the redox potentials (Eh) of GSH/GSSG and Cys/CySS in plasma. Results show that there were no significant differences between WT and Tg mice in concentrations of Cys, CySS, GSH, and GSSG or EhCySS and EhGSSG (
The 2009 H1N1 influenza virus (A/California/2009) administered at the LD50 was pathogenic to both the wild-type and NLS-hTrx1 transgenic mice as shown by body weight loss (
WT (n = 18) and Tg mice (n = 17) were infected with 2009 H1N1 influenza A/California/04/2009 virus (1× LD50) as described previously
To test whether increased mortality and poor recovery of NLS-hTrx1 Tg following viral infection was associated with increased amounts of inflammatory cytokines, we analyzed mRNA levels of colony stimulating factor (CSF), IL-1β, IL-6, IL-10, and TNFα in lung tissues of infected mice at 3-d post infection. Both TNFα (Tg, 143.5±51.5; WT, 16.0±7.3,
Lung tissues obtained from Tg and WT before (3 mice each for Tg and WT) and 3-d post infection (6 mice for Tg, 5 mice for WT) were examined for mRNA of cytokines (A) and NF-κB activity (B). mRNA levels of TNF-α and IL-6 were analyzed and quantified by real-time PCR. * p<0.05 for 3-d post infection in Tg compared to 3 d post infection in WT and 0 d in Tg. B, NF-κB activity was examined by EMSA using same tissues analyzed for cytokines shown in A. Bands indicated as NF-κB bound DNA probe show activity of NF-κB. Densitometry values shown as fold difference were obtained from measuring relative intensities compared to that in WT before infection (lane 2). [lanes 1–6 are as follows; 1, NF-κB probe alone; 2, WT before infection; 3, Tg before infection; 4, WT 3 days post infection; 5, Tg 3-d post infection; 6, Tg 3-d post infection (cold NF-κB probe incubation followed by labeled NF-κB probe incubation)]. C, NF-κB activity (NF-κB Luc) was examined by expression of NLS-hTrx1 WT or dominant negative mutant of Trx1, NLS-hTrx1 C35S by transient cotransfection with NF-κB luciferase and β-galactosidase. Quantified luciferase activity as a measure of NF-κB activity was normalized by β-galactosidase
IL-6 and TNFα are expressed under the control of the transcription factor NF-κB. To determine if DNA-binding activity was increased in Tg mice, we analyzed nuclear extracts obtained from WT and Tg lungs before [lanes 2 (WT) and 3 (Tg) in
Since Trx1 regulates NF-κB activation in a redox-dependent manner, we examined whether increased activation of NF-κB in NLS-hTrx1 Tg was regulated by Trx1 through its catalytic redox activity. To address this question, we used an
Influenza infection is associated with oxidative stress in the lungs. Since previous findings have shown nuclear translocation of Trx1 under oxidative stress conditions
WT mice infected with influenza H1N1 virus (A/California/04/2009) at day 3 post infection (day 3) or uninfected WT control mice (day 0) were examined for nuclear Trx1. Cytoplasm (Cyto) and nuclear (Nuc) fractions of lung tissues (A) and immune cells isolated from lungs were examined for Trx1 level by western blot analysis. The same blot was probed with a lamin A/C antibody as a nuclear protein marker. Results are representatives of 3 analyses.
To determine whether increased nuclear Trx1 protected against oxidative stress following influenza infection, we measured GSH, GSSG, Cys and CySS and calculated EhGSSG and EhCySS redox potentials in lung tissue and plasma before and 3-d post infection in WT and Tg mice. Lung EhGSSG of both WT and Tg was significantly oxidized after infection, but there was no statistical difference between WT and Tg. Lung tissue EhCySS was not measurable due to low levels of Cys and CySS in tissue. In WT, viral infection had no significant effect on plasma GSH or plasma EhGSSG. However, plasma GSH was decreased in Tg after infection [
The same lung samples described in
To determine whether increased nuclear Trx1 could similarly cause hyper-responsive immune signaling by another transcription factor with a redox-sensitive Cys in the DNA binding region, experiments were performed to examine glucocorticoid-stimulated apoptosis signaling in immature thymocytes. Dexamethasone (Dex) was used under conditions to cause limited (7%, 4 h) cell death in WT thymocytes measured by WST-1. Under identical conditions in Tg thymocytes (
Western blot shows NLS-hTrx1 expression in thymocytes of Tg mice (A) and in transfected HeLa cell lysates used as NLS-hTrx1 protein control. Isolated thymocytes from Tg (6 mice) and WT (6 mice) were incubated with dexamethasone (Dex, 0, 1, 10 µM) for 4 h. Quantification of surviving cells in each group was measured 4 h after Dex treatment by WST-1 assay and confirmed by trypan blue cell counting. (Data are shown as means ± SE, 3 independent experiments, * p<0.05).
Compartmentalized redox regulation is a key component of redox signaling and regulation of cell functions. Studies by Curran and coworkers showed that early stress response by AP-1 involved redox regulation in cell nuclei
Nuclear Trx1 is relatively reduced under non-stressed conditions
A model incorporating these concepts to account for excessive immune response due to increased nuclear Trx1 is summarized in
Virus infection and other inflammatory stimuli affect cytoplasmic redox state, e.g. oxidation of EhGSSG. Cytoplamic oxidation results in phosphorylation and degradation of I-κBα, which then translocates NF-κB to nucleus. Upon its translocation, p50 subunit (Cys62) reduced by nuclear localized Trx1 together with Ref-1 and Prx stimulates its DNA binding activity followed by subsequent gene expression, e.g. TNF-α and IL-6. Increased inflammatory cytokines further activate NF-κB as feedback stimulation.
In the absence of viral challenge, the NLS-hTrx1 mice did not show gross phenotypic differences, e.g., weight, growth, fecundity, appearance, activity, from WT, and no major effects were observed in other thiol antioxidant proteins. This suggests that the present model can be useful to study oxidative processes that occur in cell nuclei. In this regard, the model adds to the growing number of mouse models useful to translate compartment-specific mechanistic information to
The exacerbation of injury with increased nuclear Trx1 is distinct from previous Trx1- and Trx2-transgenic mouse studies that showed protection against injury
Contribution of nuclear Trx1 to hyperactivity of the immune system could provide the basis for therapeutic development to inhibit nuclear Trx1 translocation as a means to prevent excessive activation following influenza infection. With some capacity to maintain reduction of transcription factors already present in nuclei, controlling further increase might limit intensity of activation without blocking necessary activity. Therapeutics targeting Trx1 translocation may also be useful for adult respiratory distress and multi-systems organ failure where post-infection immune responses contribute to tissue injury and death
In summary, the present studies with H1N1 influenza viral infection in NLS-hTrx1 Tg mice show that increased nuclear Trx1 enhances activity of redox-sensitive transcription by NF-κB, causes exaggerated immune response and contributes to disease severity. Consistent with this response, redox-sensitive glucocorticoid-induced cell death was also exacerbated in thymocytes from Tg compared to WT littermates. Together, the results show that increased nuclear Trx1 has a critical role in stimulating intensity of immune responses. This suggests that nuclear translocation of Trx1 may be a useful therapeutic target to prevent severity of disease caused by excessive or prolonged activation of redox-sensitive transcription factors.