Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2012 May;19(5):847-58.
doi: 10.1038/cdd.2011.165. Epub 2011 Nov 18.

Mutation of ATF4 Mediates Resistance of Neuronal Cell Lines Against Oxidative Stress by Inducing xCT Expression

Affiliations
Free PMC article

Mutation of ATF4 Mediates Resistance of Neuronal Cell Lines Against Oxidative Stress by Inducing xCT Expression

J Lewerenz et al. Cell Death Differ. .
Free PMC article

Abstract

Selecting neuronal cell lines for resistance against oxidative stress might recapitulate some adaptive processes in neurodegenerative diseases where oxidative stress is involved like Parkinson's disease. We recently reported that in hippocampal HT22 cells selected for resistance against oxidative glutamate toxicity, the cystine/glutamate antiporter system x(c)(-), which imports cystine for synthesis of the antioxidant glutathione, and its specific subunit, xCT, are upregulated. (Lewerenz et al., J Neurochem 98(3):916-25). Here, we show that in these glutamate-resistant HT22 cells upregulation of xCT mediates glutamate resistance, and xCT expression is induced by upregulation of the transcription factor ATF4. The mechanism of ATF4 upregulation consists of a 13 bp deletion in the upstream open reading frame (uORF2) overlapping the ATF4 open reading frame. The resulting uORF2-ATF4 fusion protein is efficiently translated even at a low phosphorylation levels of the translation initiation factor eIF2α, a condition under which ATF4 translation is normally suppressed. A similar ATF4 mutation associated with prominent upregulation of xCT expression was identified in PC12 cells selected for resistance against amyloid β-peptide. Our data indicate that ATF4 has a central role in regulating xCT expression and resistance against oxidative stress. ATF4 mutations might have broader significance as upregulation of xCT is found in tumor cells and associated with anticancer drug resistance.

Figures

Figure 1
Figure 1
Upregulation of xCT mediates glutamate resistance in HT22-R cells. (a) HT22-wt and –R cells were exposed to glutamate at the indicated concentrations, and survival was measured by the MTT assay after 24 h. The graph represents the mean of four experiments with the viability at 0 mM glutamate normalized to 100% as mean±S.E.M. (b) xCT mRNA in HT22-wt and –R cells was quantified by qPCR and normalized to the expression of the housekeeping genes HRPT and β-actin. The results of two independent experiments, each performed in triplicate, are shown as mean±S.E.M. with the expression in HT22-wt cells normalized to 1. (c) Total GSH in HT22-wt and –R cells was quantified enzymatically and normalized to cellular protein. The graph shows the mean±S.E.M. of three independent experiments each performed in duplicate with the GSH in HT22-wt cells normalized to 100%. (d) HT22-R cells were exposed to glutamate at the indicated concentrations in the presence of 50 μM sulfasalazine (SAS), 25 μM (S)-4-carboxyphenylglycine (CPG) or vehicle (Ctrl). After 24 h viability was quantified by the MTT assay and the viability of cells treated with the system xc inhibitor or vehicle alone was normalized to 100%. The graph shows the mean of three independent experiments as mean±S.E.M. (e and f) HT22-R cells were transiently transfected with the control vector (Ctrl) or vectors expressing two different xCT-specific miRNAs (#1, #2). xCT mRNA was quantified by qPCR (e) and transfected cells were exposed to the indicated concentrations of glutamate (f). Viability was assessed by the MTT assay and viability of cells exposed to 0 mM glutamate normalized to 100%. The graph shows the results of four experiments each performed in triplicate (e) and the mean of six experiments (f) as mean±S.E.M. Statistical analysis was performed using two-way ANOVA (a, d, f), one-way ANOVA (e) with Bonferroni's multiple comparisons post test or two-sided unpaired t-test (b and c). *P<0.05, **P<0.01, ***P<0.001
Figure 2
Figure 2
An ATF4 isoform with an apparent molecular weight ∼5 kD higher than regular ATF4 mediates the upregulation of xCT and glutamate resistance in HT22-R cells. (a) System xc activity is upregulated in glutamate-resistant HT22 cells and HT22 cells treated with tBHQ. Glutamate-sensitive sodium-independent 35S-cystine uptake was measured in wild-type HT22 cells (HT22-wt), wild-type HT22 cells treated for 24 h with 25 μM tBHQ (wt + tBHQ) and glutamate-resistant HT22 cells (HT22-R). Individual experiments were performed in triplicate. The graph represents the mean of four (tBHQ) and five (HT22-R) independent experiments with the mean uptake in wild-type HT22 cells for each experiment normalized to 100% as mean±S.E.M. (b) Left panel: xCT protein is upregulated by tBHQ treatment in HT22-wt cells and in HT22-R cells. Equal amounts of purified membrane protein from HT22-wt cells, tBHQ-treated HT22-wt cells and HT22-R cells were analyzed by western blotting using an anti-xCT antibody. Actin served as loading control. The graph represents densitometric data of four independent protein preparations with xCT expression after tBHQ treatment normalized to 1 as mean±S.E.M. xCT was undetectable in HT22-wt without treatment. A representative blot is shown below. Right panel: 4F2hc is upregulated in HT22-R cells but not by tBHQ in HT22-wt cells. Western blots were hybridized with an anti-4F2hc antibody. The graph shows quantitative data of two independent protein preparations as mean±S.E.M. (c) xCT is transcriptionally upregulated. HT22-wt cells and HT22-R cells were co-transfected with a 4700 bp-xCT promoter luciferase construct or pGl3 and pSV-β-Gal vector. tBHQ (25 μM) was added to HT22-wt cells after transfection. Luciferase and β-galactosidase activity was measured after 24 h and normalized to pGL3 transfection. The luciferase induction in HT22-wt cells was normalized to 100%. The graph represents data from four independent experiments as mean±S.E.M. (d) Increased activity of the xCT promoter in HT22-R cells originates from the proximal AARE. HT22-wt and HT22-R cells were co-transfected with xCT promoter–luciferase constructs of the indicated lengths and the pSV-β-Gal vector. Luciferase and β-galactosidase activity were measured after 24 h. Luciferase/β-galactosidase ratio of HT22-wt cells co-transfected with pGL3 was normalized to 1. Deletion of the most proximal base pair (−94) of the proximal AARE abolishes the difference between HT22-wt and HT22-R cells. The graph represents data from four independent experiments as mean±S.E.M. xCT promoter activity in HT22-R cells was significantly higher (P<0.01) in all constructs >93 bp. (e) ATF4, but not Nrf2, is upregulated in HT22-R cells. Western blots of nuclear extracts of HT22-wt cells or HT22-R cells were hybridized with antibodies against ATF4 and Nrf2. Blots were probed with an anti-actin antibody as a loading control. Graphs represent quantitative data from five independent protein preparations and western blots for HT22-wt and -R as mean±S.E.M. To ensure specificity of the antisera, cells treated with 25 μM tBHQ (wt+tBHQ) or arginine-free medium (wt-Arg) for 24 h were included in some experiments (wt+ tBHQ: anti-ATF4 in two, anti-Nrf2 in four, wt-arg in two). Representative western blots are shown below the graphs. Note the higher molecular weight of ATF4 in HT22-R cells. (f) EMSA shows increased binding of nuclear proteins from HT22-R cells to the AARE present in the xCT promoter. Nuclear extracts from HT22-wt and -R cells were allowed to bind to the radiolabeled AARE and separated by PAGE. The arrow indicates ATF4-specific binding, the asterisk non-specific binding and the arrowhead free unbound probe. (gi) HT22-R cells were co-transfected with siRNA against ATF4 or control siRNA together with either pcDNA3 (Ctrl) or pcDNA3-xCT (xCT) vector. After 24 h, cells were replated for protein analysis, system xc activity measurement and oxidative glutamate toxicity. (g) Western blot analysis of transfected HT22-R cells. Nuclear extracts were evaluated by western blot for expression of ATF4 (ATF4) and membrane fractions were evaluated for xCT expression (xCT). Actin served as a loading control. (h) System xc activity in co-transfected HT22-R cells was measured as glutamate-sensitive 35S-cystine uptake. (i) Survival in response to oxidative glutamate toxicity in co-transfected HT22-R cells. Cells in 96-well plates were treated with the indicated concentrations of glutamate for 24 h and survival measured by the MTT assay. (g–i) All graphs represent data of the same four consecutive transfections as mean±S.E.M. Representative western blots are shown in (g). Results obtained by transfection with both control siRNA and control vector were normalized to 1 in (g) or 100% in (h). Viability of cells not treated with glutamate was normalized to 100% in (i). Statistical analysis was performed by one-way ANOVA (ac, e, g, h) or two-way ANOVA (i) with Bonferroni's multiple comparisons post test, *P<0.05, **P<0.01, ***P<0.001 compared with HT22-wt (a–e). (g–i) *P<0.05, **P<0.01, ***P<0.001 compared with HT22-R cells transfected with control siRNA and control vector, #P<0.05, ##P<0.01, ###P<0.001 compared with cells transfected with ATF4 siRNA and control vector
Figure 3
Figure 3
ATF4 in HT22-R cells has the same half-life and a higher synthesis rate as ATF4 in HT22-wt cells, and the higher molecular weight is not due to post-translational modification. (a) HT22-wt and HT22-R cells were treated with 5 μg/ml cycloheximide (CHX) for the indicated times. Whole-cell extracts were analyzed by western blotting with an antibody against ATF4. Actin served as a loading control. Samples of HT22-R cells were diluted 1 : 3 to adjust for higher ATF4 expression. A representative western blot is shown. The graph represents pooled data from three independent protein preparations with the ATF4/actin ratio without protein synthesis inhibition normalized to 100% as mean±S.E.M. Non-linear regression was performed using one phase exponential decay. (b) HT22-wt and HT22-R cells were treated with 10 μM MG132 for the indicated times and whole-cell extracts were analyzed for ATF4 and Nrf2 expression. For ATF4, HT22-R samples were diluted 1 : 3 to adjust for higher basal ATF4 levels. An unspecific band obtained with the Nrf2 antibody is indicated (Unsp.). The graphs represent three and seven independent experiments for ATF4 and Nrf2, respectively. ATF4 expression was normalized to actin and the relative increase in ATF4 protein was calculated by subtraction of the baseline ATF4/actin ratio at 0 min. ATF4 or Nrf2 to actin ratios obtained in HT22wt cells at 40 min were normalized to 100% for every single experiment and the graphs show the mean±S.E.M. of pooled and normalized data. Statistical analysis was performed by linear regression. (c) qPCR on HT22-wt and HT22-R cDNA shows equal expression of ATF4 mRNA in both the cell lines. The graph represents the mean±S.E.M. of three independent experiments. (d) Recombinant ATF4 does not have a higher molecular weight in HT22-R cells. HT22-wt and HT22-R cells were transfected with wild-type-ATF4 in the pRK7 vector (ATF4-wt) or ATF4-HA. Whole-cell extracts were analyzed by western blotting using antibodies against ATF4 (α-ATF4) or HA (α-HA). For detection of overexpressed ATF4-wt only, samples of HT22-wt and HT22-R cells were diluted 1 : 80 and 1 : 20, respectively, until endogenous ATF4 was undetectable in mock-transfected control cells (not shown). The experiment was done three times with identical results
Figure 4
Figure 4
The high molecular weight and ‘classical' ATF4 isoforms are inversely regulated by eIF2α phosphorylation. (a) Whole-cells extracts of HT22-wt and HT22-R cells treated with either vehicle (Ctrl) or salubrinal (Salu) at the indicated concentrations for 24 h or 5 μg/ml tunicamycin (Tuni) for 2 h were analyzed for eIF2α phosphorylation (peIF2α) relative to eIF2α expression and expression of the ‘classical' and high molecular weight ATF4 isoforms (ATF4∼60 kD and ATF4∼65 kD, respectively) relative to actin. Relative eIF2α phosphorylation and ATF4 expression were normalized to those in HT22-wt cells treated with vehicle. (b) HT22-wt and HT22-R cells were co-transfected with a luciferase construct fused to the ATF4 5′UTR and the pSV-β-GAL vector. Cells were treated either with vehicle (Ctrl) or 60 μM salubrinal (Salu). The luciferase/β-galactosidase ratio of untreated HT22-wt cells was normalized to 100%. (c and d) HT22-wt or HT22-R cells were treated with 30 μM Salubrinal (Salu) or vehicle (Ctrl) for 24 h. (c) System xc activity was measured as HCA-sensitive sodium-independent 3H-glutamate uptake. Uptake in vehicle-treated cells was normalized to 100%. (d) Cells in 96-well plates were treated with glutamate at the indicated concentrations for 24 h in quadruplicate. Viability was measured by the MTT assay and relative survival calculated by normalizing to the MTT value of cells not treated with glutamate. The graphs show data from three (a and d) or four (b and c) independent experiments as mean±S.E.M. Statistical analysis was performed by two-way ANOVA and Bonferroni post test, *P<0.05. **P<0.01,***P<0.001 compared with control cells, #P<0.05, HT22-R cells compared with HT22-wt without salubrinal. Representative blots are shown below the graphs in (a)
Figure 5
Figure 5
High molecular weight ATF4 is the result of a 13-bp deletion within uORF2 in one allele of the ATF4 gene. (a) Scheme of ATF4 mRNA with uORF1, uORF2 and the ATF4 ORF. When eIF2α phosphorylation is low, uORF2 is preferentially translated (upper), whereas the ATF4 ORF is preferentially translated in the presence of high eIF2α phosphorylation. (b) A 13 bp deletion (red) in uORF2 leads to fusion of uORF2 and the ATF4 ORF. The high molecular weight ATF4 (ATF4h) ORF encodes a 377 AA protein. The eight new AAs that are encoded by uORF2 due to the frame shift are shown in green. The color reproduction of this figure is available at the Cell Death and Differentiation journal online
Figure 6
Figure 6
Wild-type and high molecular weight ATF4 equally induce system xc activity, elevate cellular GSH levels and protect HT22 cells against oxidative glutamate toxicity. HT22-wt cells were transiently transfected with a vector encoding ATF4-wt (ATF4), high molecular weight ATF4 (ATF4h) or empty pRK7 vector (Ctrl). (a) Western blotting shows similar expression of ATF4 and ATF4h upon transfection. ATF4 and ATF4h increase (b) xCT mRNA expression when quantified by qPCR and (c) system xc activity measured as HCA-sensitive uptake of radiolabeled glutamate. Overexpression of ATF4 and ATF4h increase (d) total GSH and (e) resistance against oxidative glutamate toxicity. (e) At 24 h after plating, HT22-wt cells transfected with empty vector (Ctrl), ATF4-wt (ATF4) or high molecular weight ATF4 (ATF4h) were exposed to the indicated concentrations of glutamate and viability was quantified by the MTT assay 24 h later. Graphs show the results of (b) two experiments each performed in triplicate or three (a, c, e) or five (d) independent experiments. (a) ATF4h expression was normalized to ATF4 expression after correction for actin expression as a loading control. (b and c) xCT mRNA was corrected for HRPT and actin β expression, system xc activity and GSH were corrected for total protein and results were normalized to HT22-wt cells transfected with empty pRK7. (e) Viability at 0 mM glutamate was normalized to 100%. Statistical analysis was performed by (b and c) one-way ANOVA or (e) two-way ANOVA with Bonferroni's multiple comparisons post test, *P<0.05, **P<0.01, ***P<0.001 compared with HT22-wt transfected with empty vector
Figure 7
Figure 7
System xc activity, xCT protein and a high molecular weight ATF4 isoform with similar properties as in HT22-R are upregulated in clonal PC12 cells selected for resistance against amyloid-β peptide. (a) Wild-type PC12 (PC12-wt) and clones 1, 7 and 8 of Aβ-resistant PC12 cells (PC12r1, PC12r7, PC12r8) were plated in 24-well dishes and system xc activity was measured as glutamate-sensitive 35S-cystine uptake. Uptake was normalized to PC12-wt cells. The graph represents five independent experiments, two of which included PC12-r8, as mean±S.E.M. (b and c) PC12-wt and PC12-r1 cells were fractionated and (b) membrane protein analyzed for xCT expression and (c) nuclear extracts analyzed for ATF4 and Nrf2 expression by western blotting. Western blotting against actin (b) or the histone band visualized by Poinceau S-staining (c) served as loading controls. Nrf2 runs at ∼110 kD. An unspecific band is indicated (unsp.). The graphs represent the results of three (b) and four (c) independent protein preparations with the relative expression in PC12-wt cells normalized to 1 as mean±S.E.M. Statistical analysis was performed by (a) one-way ANOVA with Bonferroni's multiple comparisons post tests (PC12-r1 compared with -wt, -r7 and -r8 ) or (b and c) one sample sample t test compared with 1, *P<0.05, **P<0.01,***P<0.001, n.s.=not significant. (d) PC12-wt and PC12-r1 cells were seeded in 60-mm dishes and treated as for HT22 cells in Figure 4. Nuclei were analyzed by western blotting for ATF4 expression. Actin served as a loading control. Expression levels of low (clear boxes, ATF4l) and high (black boxes, ATF4h) molecular weight ATF4 isoforms were analyzed by densitometry. ATF4l expression in control cultures was normalized to 1 for both the cell lines, and ATF4h levels in PC12r1 are given relative to ATF4l expression in controls. The graph represents five (PC12-wt) and seven (PC12-r1) independent experiments. Representative western blots are shown below
Figure 8
Figure 8
A 1-bp deletion in the uORF2 leads to the high molecular weight ATF4 in PC12r1 cells. A 1-bp deletion (red) in the region where uORF2 and the ATF4 ORF overlap leads to fusion of uORF2 and the ATF4 ORF. The high molecular weight ATF4 (ATF4h) ORF encodes a 379 AA protein consisting of the 58 N-terminal AAs of uORF2 and the ATF4 ORF lacking the first 26 AAs. The color reproduction of this figure is available at the Cell Death and Differentiation journal online

Similar articles

See all similar articles

Cited by 23 articles

See all "Cited by" articles

Publication types

Feedback