Abstract
Oncogenic mutations within RAS genes and inactivation of p53 are the most common events in cancer. Earlier, we reported that activated Ras contributes to chromosome instability, especially in p53-deficient cells. Here we show that an increase in intracellular reactive oxygen species (ROS) and oxidative DNA damage represents a major mechanism of Ras-induced mutagenesis. Introduction of oncogenic H- or N-Ras caused elevated intracellular ROS, accumulation of 8-oxo-2′-deoxyguanosine, and increased number of chromosome breaks in mitotic cells, which were prevented by antioxidant N-acetyl-l-cysteine. By using Ras mutants that selectively activate either of the three major targets of Ras (Raf, RalGDS, and phosphatidylinositol-3-kinase) as well as dominant-negative Rac1 and RalA mutants and inhibitors of mitogen-activated protein kinase (MAPK)/extracellular signal–regulated kinases kinase-1 and p38 MAPKs, we have shown that several Ras effectors independently mediate ROS up-regulation. Introduction of oncogenic RAS resulted in repression of transcription from sestrin family genes SESN1 and SESN3, which encode antioxidant modulators of peroxiredoxins. Inhibition of mRNAs from these genes in control cells by RNA interference substantially increased ROS levels and mutagenesis. Ectopic expression of SESN1 and SESN3 from lentiviral constructs interfered with Ras-induced ROS increase, suggesting their important contribution to the effect. The stability of Ras-induced increase in ROS was dependent on a p53 function: in the p53-positive cells displaying activation of p53 in response to Ras, only transient (4–7 days) elevation of ROS was observed, whereas in the p53-deficient cells the up-regulation was permanent. The reversion to normal ROS levels in the Ras-expressing p53-positive cells correlated with up-regulation of p53-responsive genes, including reactivation of SESN1 gene. Thus, changes in expression of sestrins can represent an important determinant of genetic instability in neoplastic cells showing simultaneous dysfunctions of Ras and p53. [Cancer Res 2007;67(10):4671–8]
Introduction
The proteins of the Ras family (H-, K-, and N-Ras) function as key regulators of signal transduction pathways that control cell proliferation, survival, migration, and differentiation (1–5). Various extracellular signals reaching cell-surface receptors stimulate the conversion of Ras proteins from the inactive GDP-bound to the active GTP-bound form. In the GTP-bound form, Ras stimulates downstream effectors, which, in turn, affect activities of numerous proteins, including large group of transcription factors. The biological effects of activated Ras proteins are mediated through several effectors that include Raf serine/threonine kinases, phosphatidylinositol-3-kinases (PI3K), and RalGDS, guanine nucleotide exchange factor (GEF) for small GTPases RalA and RalB (2–6). Mutations at residues 12, 13, or 61 that constitutively activate Ras proteins are found in 95% to 98% of pancreatic cancers and 25% to 40% of many other tumor types (1, 7, 8). Substantial experimental data indicate that aberrant Ras expression plays a critical role in oncogenesis causing stimulation of cell proliferation, angiogenesis, inhibition of apoptosis, and increased cell motility (i.e., features that are responsible for tumor growth, invasion, and metastasis; refs. 2–5).
In addition to these effects, the expression of Ras oncoproteins causes genetic instability (9–11), a feature that is considered as an engine for steadfast tumor progression (12–14). We have previously found that the mutagenic effect of Ras is especially prominent in p53-deficient cells (11) and can be connected with both attenuation of DNA damage cell cycle checkpoints and up-regulation of reactive oxygen species (ROS; refs. 15, 16). In fact, expression of Ras oncoproteins increases ROS content (15, 17–20) and interferes with DNA damage–induced arrest in G1 and G2 (11, 16), although the mechanisms of these effects are poorly understood.
In this article, we present the data showing a critical role of ROS up-regulation in Ras-induced mutagenesis. We found that functional p53 is capable of counteracting the Ras-induced increase in ROS levels. We also propose a novel mechanism by which oncogenic Ras up-regulates ROS and compromises genetic stability, which is connected to transcriptional repression of sestrin family genes. The products of these genes participate in antioxidant defense by modulating regeneration of peroxiredoxins (21). The presented results highlight the importance of antioxidant mechanisms in controlling genetic stability.
Materials and Methods
DNA constructs. The following previously developed and described constructs were used: pBabe-puro/Ras retroviral vector expressing activated human N-Ras D13 mutant (11, 22); the retroviral vector pLXSN-neo containing human activated H-Ras V12 and its restricted effector specificity mutants V12S35, V12G37, and V12C40 (ref. 23; provided by J. Downward, Cancer Research UK, London, United Kingdom); pLXSN-neo/RalA-V23 with constitutively active V23 RalA mutant and lentiviral pLV-CMV/RalA-N28 expressing dominant-negative N28 RalA mutant (16); and lentiviral pLSLP vectors bearing short hairpin RNAs (shRNA) against p53 or human SESN1 and SESN2 genes (21, 24). Complementary 60-bp hairpin oligonucleotides containing 19-nucleotide regions corresponding to the human SESN3 gene (5′-GACGAGGAGAAGAGCATTT-3′ and 5′-CCAGAGAGAGATCCAGAAA-3′) were designed according to siRNA-scale algorithm5
and cloned into pLSLP vector for shRNA, as described earlier (24). The pMSCV-puro retroviral vectors expressing activated V23 RalB and dominant-negative N17 Rac1 mutants were obtained from Dr. I. Zborovskaya (Russian Cancer Research Center, Moscow, Russia). For expression of human SESN1 (T2 and T3 isoforms), SESN2, and SESN3 genes, the appropriate cDNA was introduced into cells using lentiviral expression vector pLV-CMV, as described earlier (24).Cell cultures. Human p53-deficient MDAH041 immortalized fibroblasts derived from a Li-Fraumeni syndrome patient (25) and rat immortalized embryonic fibroblasts REF52 and Rat1 were used. Their derivatives with tetracycline-repressible expression of activated N-RasD13 mutant (MDAH041/tet-Ras, REF52/tet-Ras, and Rat1/tet-Ras cell lines, respectively) were described by us earlier (11, 22). The cell sublines with constitutive expression of various H-Ras, N-Ras, RalA, and Rac1 mutants were previously described (11, 16). Briefly, the retroviral DNA constructs were transfected into Phoenix-ampho packaging cells using LipofectAMINE PLUS reagent (Invitrogen). Virus-containing supernatants collected 24 to 8 h after transfection were used to infect recipient cells in the presence of polybrene (8 μg/mL). Infected cell cultures were selected in medium containing genticin (Invitrogen; 0.5 mg/mL for 14–20 days) or puromycin (Sigma; 1 μg/mL for 5–6 days).
High-titer stocks of recombinant lentiviruses pseudotyped with VSV-G protein were produced by cotransfection into 293T cells of the pLV or pLSLP constructs together with the pCMVδ8.2R and pVSV-G packaging plasmids. Virus-containing supernatants were collected 24 to 48 h after transfection and used for infection of recipient cells.
All the cells were cultured in DMEM supplemented with 10% fetal bovine serum and penicillin/streptomycin. The MDAH041/tet-Ras, REF52/tet-Ras, and Rat1/tet-Ras cells were permanently maintained in medium supplemented with 1 μg/mL of tetracycline (Sigma), and the expression of RAS oncogene was induced by removal of tetracycline from culture medium 2 to 15 days before the experiments. All experiments with the MDAH041 cell sublines sustaining constitutive expression of various Ras, Ral, or Rac1 mutants were done 7 to 20 days after retrovirus vector–mediated gene transfer.
Measurement of intracellular ROS. Intracellular ROS was measured as described earlier (26). Briefly, the cells were incubated with 2′-7′ dichlorodihydrofluorescein diacetate (DCF-DA; Molecular Probes) at 5 μmol/L for 30 min, washed with PBS, trypsinized, collected in 1 mL of PBS, transferred to polystyrene tubes with cell-strainer caps (Falcon), and subjected to fluorescence-activated cell sorting (FACS; Becton Dickinson FACScan) using CellQuest 3.2 (Becton Dickinson) software for acquisition and analysis.
DNA oxidation. DNA oxidation was assessed by measuring 8-oxo-2′-deoxyguanosine content with Biotrin OxyDNA Kit (Biotrin). For the assay, 106 cells were processed for FACS analysis according to protocol provided by the manufacturer.
Chromosome analysis. Chromosomal slides were prepared by routine methods. To determine the frequency of chromosomal breaks, 50 to 100 metaphases were karyotyped in each case. The unrepaired chromatid and chromosome breaks and gaps were scored.
Detection of mRNA for sestrin genes by reverse transcription-PCR. Total mRNA was isolated with SV Total RNA Isolation System (Promega) according to the manufacturer's protocols. Reverse transcription-PCR (RT-PCR) analysis was done as described (24). To monitor expression of corresponding genes by RT-PCR e, we used the following primers: for hSESN1 (all isoforms), forward primer 5′-CTTCTGGAGGCAGTTCAAGC-3′ and reverse primer 5′-TGAATGGCAGCCTGTCTTCAC-3′; for SESN1 T1 isoform, forward 5′-CGGATGGGTTGAATAAGCTACT-3′ and common (T1, T2, and T3 isoforms) reverse 5′-CCAATGTAGTGACGATAATGTAGG-3′ primer; for SESN1 T2 isoform, forward 5′-CGACCAGGACGAGGAACTT-3′; for SESN1 T3 isoform, forward 5′-CCTGCAAGCTCTATGACGTTATG-3′; for SESN2, forward 5′-CAAGCTCGGAATTAATGTGCC-3′ and reverse 5′-CTCACACCATTAAGCATGGAG-3′; and for SESN3, forward 5′-GTTCACTGTATGTTTGGAATCAGG-3′ and reverse 5′-GGGTGATACTTCAGGTCAAATG-3′. For rat genes, we used the following primers: for Sesn1, forward 5′-GGACGAGGAACTTGGAATCA-3′ and reverse 5′-CCCTCGATGTGTTCTTTGGT-3′; for Sesn2, forward 5′-ACAGGTTTCCTCCGACACAC-3′ and reverse 5′-CTCTGGAGTCTGCCTGTTCC-3′; and for Sesn3, forward 5′-CCTGGACAACATCACACAGG-3′ and reverse 5′-ACTAGCTCAGGCAAGGACCA-3′. To detect in human and rat cells α-tubulin mRNA, we used the same forward (5′-GTTGGTCTGGAATTCTGTCAG-3′) and reverse (5′-AAGAAGTCCAAGCTGGAGTTC-3′) primers.
Western blot analysis. Whole-cell extracts were lysed in ice-cold radioimmunoprecipitation assay buffer [50 mmol/L Tris-HCl (pH 7.4), 150 mmol/L NaCl, 1% deoxycholate-Na, 1% NP40, 0.1% SDS, 100 mmol/L phenylmethylsulfonyl fluoride, 1 mmol/L pepstatin A, and 1 mmol/L E64]. Protein concentration in the extracts was determined with a protein assay system (Bio-Rad). Twenty micrograms of protein were separated on a 10% SDS polyacrylamide gel and transferred to polyvinylidene difluoride membrane (Amersham GE Healthcare). The membranes were probed with antibodies specific to p21Ras (OP21, Calbiochem), p53 (PAb421, Calbiochem), p21Cip1/Waf1 (F-5, Santa Cruz Biotechnology), and p44/p42 extracellular signal–regulated kinases (ERK) phosphorylated at Thr202/Tyr204 (E10, Cell Signaling Technology). Membranes were treated with secondary antimouse-HRP or antirabbit-HRP antibodies (Amersham GE Healthcare). Filters were developed with enhanced chemiluminescence reagents (Amersham GE Healthcare) according to the manufacturer's protocol.
Ral activity assay. Cells were lysed in RAB buffer (Upstate Biotechnology) and the amount of GTP-bound Ral was determined using Ral Activation Assay Kit (Upstate Biotechnology) according to the manufacturer's protocols.
Results
Ras-induced chromosome breaks are associated with elevated intracellular ROS and enhanced DNA oxidation. To estimate contribution of increased intracellular ROS in Ras-induced genomic instability, we used human MDAH041 and rat REF52 and Rat1 cell lines. Constitutive or conditional tetracycline-regulated expression of activated Ras (H-RasV12 and N-RasD13) coincided with increased intracellular ROS levels (DCF fluorescence), accumulation of the DNA oxidation product 8-oxo-2′-deoxyguanosine, and elevated frequency of chromosome breaks in mitoses (Fig. 1; Table 1). Treatment with 2 to 10 mmol/L of antioxidant N-acetyl-l-cysteine (NAC) for 12 to 24 h did not influence the level of Ras protein (Supplementary Fig. S1) but abolished the increase in intracellular ROS and led to almost complete attenuation of Ras-induced mutagenesis (Table 1), suggesting that the increase in intracellular ROS can boost chromosome instability in cells expressing oncogenic Ras.
Effect of expression of activated N-RasD13 on ROS levels and oxidized DNA content in human and rat cells. A, influence of constitutive expression of pBabe-puro/Ras construct (MDAH041/Ras cells) or incubation of MDAH041/tet-Ras, REF52/tet-Ras, and Rat1/tet-Ras cells in tetracycline-free medium for 5 d on expression of N-Ras protein (Western blot analysis was done as described in Materials and Methods). B, distribution of 104 cells according to the intensity of DCF-DA fluorescence in control and Ras-expressing cultures. Result of one of typical experiments. C, indices of total DCF-DA fluorescence of 104 cells. Summarized data of two to three independent experiments. D, indices of total fluorescence of 104 cells stained with FITC-conjugated antibodies to 8-oxo-2′-deoxyguanosine (8-oxo-DG). Summarized result of two experiments.
Effect of expression of activated N-RasD13 on ROS levels and oxidized DNA content in human and rat cells. A, influence of constitutive expression of pBabe-puro/Ras construct (MDAH041/Ras cells) or incubation of MDAH041/tet-Ras, REF52/tet-Ras, and Rat1/tet-Ras cells in tetracycline-free medium for 5 d on expression of N-Ras protein (Western blot analysis was done as described in Materials and Methods). B, distribution of 104 cells according to the intensity of DCF-DA fluorescence in control and Ras-expressing cultures. Result of one of typical experiments. C, indices of total DCF-DA fluorescence of 104 cells. Summarized data of two to three independent experiments. D, indices of total fluorescence of 104 cells stained with FITC-conjugated antibodies to 8-oxo-2′-deoxyguanosine (8-oxo-DG). Summarized result of two experiments.
Effects of Ras expression and NAC treatment on ROS levels and frequency of chromosome breaks in dividing cells
Cell line . | Exogenous Ras expression . | NAC . | DCF fluorescence . | Chromosome breaks . | . | |
---|---|---|---|---|---|---|
. | . | . | . | % Metaphases . | Breaks per metaphase . | |
MDAH041 | - | - | 57 ± 9 | 33 ± 1.2 | 0.67 ± 0.05 | |
+ | 52 ± 8 | 17 ± 1.0 | 0.36 ± 0.04 | |||
MDAH041/Ras* | + | - | 105 ± 12 | 58 ± 1.8 | 1.44 ± 0.10 | |
+ | 65 ± 8 | 29 ± 3.0 | 0.72 ± 0.05 | |||
MDAH041/tet-Ras† | - | - | 65 ± 12 | 39 ± 3.2 | 1.20 ± 0.34 | |
+ | - | 125 ± 18 | 63 ± 4.0 | 2.30 ± 0.52 | ||
REF52/tet-Ras† | - | - | 24 ± 6 | 13 ± 1.4 | 0.16 ± 0.02 | |
+ | 21 ± 5 | 10 ± 1.0 | 0.14 ± 0.02 | |||
+ | - | 49 ± 8 | 22 ± 2.0 | 0.58 ± 0.02 | ||
+ | 28 ± 7 | 8 ± 0.1 | 0.08 ± 0.01 | |||
Rat1/tet-Ras† | - | - | 9 ± 4 | 13 ± 1.0 | 0.16 ± 0.08 | |
+ | 7 ± 4 | 11 ± 1.0 | 0.13 ± 0.02 | |||
+ | - | 66 ± 16 | 28 ± 2.0 | 0.56 ± 0.10 | ||
+ | 25 ± 8 | 10 ± 1.0 | 0.16 ± 0.04 |
Cell line . | Exogenous Ras expression . | NAC . | DCF fluorescence . | Chromosome breaks . | . | |
---|---|---|---|---|---|---|
. | . | . | . | % Metaphases . | Breaks per metaphase . | |
MDAH041 | - | - | 57 ± 9 | 33 ± 1.2 | 0.67 ± 0.05 | |
+ | 52 ± 8 | 17 ± 1.0 | 0.36 ± 0.04 | |||
MDAH041/Ras* | + | - | 105 ± 12 | 58 ± 1.8 | 1.44 ± 0.10 | |
+ | 65 ± 8 | 29 ± 3.0 | 0.72 ± 0.05 | |||
MDAH041/tet-Ras† | - | - | 65 ± 12 | 39 ± 3.2 | 1.20 ± 0.34 | |
+ | - | 125 ± 18 | 63 ± 4.0 | 2.30 ± 0.52 | ||
REF52/tet-Ras† | - | - | 24 ± 6 | 13 ± 1.4 | 0.16 ± 0.02 | |
+ | 21 ± 5 | 10 ± 1.0 | 0.14 ± 0.02 | |||
+ | - | 49 ± 8 | 22 ± 2.0 | 0.58 ± 0.02 | ||
+ | 28 ± 7 | 8 ± 0.1 | 0.08 ± 0.01 | |||
Rat1/tet-Ras† | - | - | 9 ± 4 | 13 ± 1.0 | 0.16 ± 0.08 | |
+ | 7 ± 4 | 11 ± 1.0 | 0.13 ± 0.02 | |||
+ | - | 66 ± 16 | 28 ± 2.0 | 0.56 ± 0.10 | ||
+ | 25 ± 8 | 10 ± 1.0 | 0.16 ± 0.04 |
The MDAH041/Ras cells were analyzed 9 to 12 d after retroviral transfer of H-RasV12 (4–7 d after selection in the medium with puromycin). The data of two experiments are presented; in each experiment, 30 to 50 control and Ras-expressing metaphases were scored.
Cell cultures were analyzed 5 to 7 d after removal of tetracycline from culture medium (REF52/tet-Ras cells were studied on the 5th day, i.e. before the development of Ras-induced cell cycle arrest; ref. 22). For each cell line, the data of two experiments are presented; in each experiment, 25 to 50 metaphases were scored.
Several Ras effectors are responsible for ROS up-regulation and genetic instability. To assess the role of different Ras effectors in ROS regulation, we introduced into MDAH041, REF52, and Rat1 cells the restricted effector specificity H-Ras mutants V12S35, V12G37, and V12C40, which selectively activate Raf, RalGDS, or PI3K, respectively (23). In all tested cell lines, expression of each of these mutants caused notable increase in ROS levels, although its magnitude was lower than that observed after introduction of full-value activated Ras V12 (Fig. 2A,, B, and E). Of note, in the cells expressing the H-Ras mutants, the frequency of chromosome breaks directly correlated with the ROS levels (Fig. 2A). This observation highlights a major role of elevated ROS in Ras-induced genetic instability.
Effect of changes in activity of various Ras effectors on expression of sestrin family genes, ROS levels, and mutagenesis. A, effect of constitutive expression of various H-Ras mutants on activity of ERK1/2 kinases; levels of SESN1, SESN2, SESN3 mRNAs; DCF-DA fluorescence; and chromosome breaks in human MDAH041 cells. The cell cultures were analyzed, as described in Materials and Methods, 7 d after retroviral transfer of corresponding pLXSN-neo constructs and selection in the medium with genticin. B, effect of various H-Ras mutants on expression of SESN1, SESN2, and SESN3 mRNAs and DCF-DA fluorescence in rat REF52 cells. The cell cultures were analyzed 7 d after introduction of corresponding retroviral pLXSN-neo vectors. C, effect of constitutive expression of dominant-negative Rac1N17 mutant on ROS levels in Ras-expressing MDAH041/tet-Ras, REF52/tet-Ras, and Rat1/tet-Ras cells. The cells were infected with pMSCV-puro/Rac1N17 vector and selected with puromycin in tetracycline-containing medium for 3 d. The DCF-DA fluorescence of Ras-expressing cells was analyzed 5 d after tetracycline withdrawal. D, effect of constitutively active V23 and dominant-negative N28 RalA mutants on amounts of functionally active GTP-bound forms of RalA and levels of ROS in human MDAH041 cells. The cell cultures expressing the RalA mutants were obtained and analyzed as described in Materials and Methods. E, effect of various H-Ras and RalA mutants on ROS levels in Rat1 cells. The cell cultures were obtained and analyzed as described in Materials and Methods. Typical results of one of two to three independent experiments (A–E).
Effect of changes in activity of various Ras effectors on expression of sestrin family genes, ROS levels, and mutagenesis. A, effect of constitutive expression of various H-Ras mutants on activity of ERK1/2 kinases; levels of SESN1, SESN2, SESN3 mRNAs; DCF-DA fluorescence; and chromosome breaks in human MDAH041 cells. The cell cultures were analyzed, as described in Materials and Methods, 7 d after retroviral transfer of corresponding pLXSN-neo constructs and selection in the medium with genticin. B, effect of various H-Ras mutants on expression of SESN1, SESN2, and SESN3 mRNAs and DCF-DA fluorescence in rat REF52 cells. The cell cultures were analyzed 7 d after introduction of corresponding retroviral pLXSN-neo vectors. C, effect of constitutive expression of dominant-negative Rac1N17 mutant on ROS levels in Ras-expressing MDAH041/tet-Ras, REF52/tet-Ras, and Rat1/tet-Ras cells. The cells were infected with pMSCV-puro/Rac1N17 vector and selected with puromycin in tetracycline-containing medium for 3 d. The DCF-DA fluorescence of Ras-expressing cells was analyzed 5 d after tetracycline withdrawal. D, effect of constitutively active V23 and dominant-negative N28 RalA mutants on amounts of functionally active GTP-bound forms of RalA and levels of ROS in human MDAH041 cells. The cell cultures expressing the RalA mutants were obtained and analyzed as described in Materials and Methods. E, effect of various H-Ras and RalA mutants on ROS levels in Rat1 cells. The cell cultures were obtained and analyzed as described in Materials and Methods. Typical results of one of two to three independent experiments (A–E).
The effect of V12C40 mutant can be connected with activation of the PI3K-Rac1-NADPH-oxidase pathway, which is known to be responsible for ROS generation in cells with activated Ras (17, 18, 20). Accordingly, we have found that inactivation of Rac1 function with dominant-negative Rac1 mutant caused a decrease of ROS levels in the cell lines expressing activated RasV12 (Fig. 2C). Tentatively, up-regulation of ROS in response to the mutant Ras V12G37 can be mediated through activation of the Ras-RalGDS-RalA pathway. We found that in Rat1 cells, the constitutively active RalA mutant V23 (but not the RalB mutant V23) caused an increase in ROS level similar to that induced by the mutant H-Ras V12G37, whereas dominant-negative RalA mutant N28 led to decreased ROS in the cells expressing activated Ras (Fig. 2E). However, contribution of this pathway was less prominent in human MDAH041 cells showing a modest effect of mutant H-Ras V12G37 compared with that in the Rat1 cells. In the MDAH041 cells, the constitutively active or dominant-negative RalA mutants caused no notable changes in ROS levels, although both mutants induced corresponding alterations in activity of RalA, as determined by measurement of its GTP-bound forms (Fig. 2D).
Role of sestrins in Ras-induced ROS up-regulation and mutagenesis. Recently, we identified a pathway that controls activity of peroxiredoxins, the enzymes responsible for decomposition of hydrogen peroxide produced in response to receptor signaling (27–29). Sestrins form a small family of redox enzymes, which, together with sulfiredoxin, participate in regeneration of overoxidized peroxiredoxins, thus controlling their antioxidant function (21). To test for possible involvement of this pathway in Ras-mediated up-regulation of ROS levels, we studied the effects of various Ras mutants on the expression of sestrin family genes. We found that in all tested cell lines, constitutive expression of V12 H-Ras as well as conditional expression of D13 N-Ras causes a decrease in levels of mRNAs derived from the sestrin family genes (Figs. 2A–B and 3). In all cases, the inhibition was most prominent for SESN3 gene; SESN2, on the contrary, showed minor repression in human MDAH041 cells and no changes in rat REF52 and Rat1 cells (Figs. 2A–B and 3). In the human MDAH041 cells, among three alternative transcripts encoded by SESN1 gene (30), the T2 and T3 SESN1 mRNAs showed clear down-regulation in Ras-expressing cells whereas the expression T1 was not affected (Figs. 2A and 3). In Ras-expressing rat REF52 and Rat1 cells, the only known SESN1 transcript was found to be strongly inhibited (Figs. 2B and 3).
Effect of conditional expression of N-RasD13 and inhibitors of MEK1 (PD98059) and p38 (SB203580) on amounts of SESN1, SESN2, and SESN3 mRNAs and ROS levels. Ras-expressing REF52/tet-Ras and Rat1/tet-Ras cells were analyzed on the 5th day after tetracycline withdrawal; MDAH041/tet-Ras cells were incubated in tetracycline-free medium for 7 d; PD98059 (10 μmol/L; Promega) and SB203580 (0.5 μmol/L; Promega) were added for the last 48 h; and DCF-DA fluorescence and mRNAs were detected as described in Materials and Methods.
Effect of conditional expression of N-RasD13 and inhibitors of MEK1 (PD98059) and p38 (SB203580) on amounts of SESN1, SESN2, and SESN3 mRNAs and ROS levels. Ras-expressing REF52/tet-Ras and Rat1/tet-Ras cells were analyzed on the 5th day after tetracycline withdrawal; MDAH041/tet-Ras cells were incubated in tetracycline-free medium for 7 d; PD98059 (10 μmol/L; Promega) and SB203580 (0.5 μmol/L; Promega) were added for the last 48 h; and DCF-DA fluorescence and mRNAs were detected as described in Materials and Methods.
The results obtained with restricted effector specificity H-Ras mutants and inhibitors of MAPK/ERK kinase (MEK)-1 (PD98059) and p38 (SB203580) suggest that the repression of SESN3 gene is mainly connected with activation of Ras-mitogen-activated protein kinase (MAPK) pathways. In fact, the V12S35 Ras mutant retaining the ability to interact with Raf (23) and activate ERK1/2 kinases (Fig. 2A) was capable of repressing SESN3 gene transcription whereas the V12G37 and V12C40 H-Ras mutants, interacting with RalGDS and PI3K, respectively (23), caused substantially weaker ERK1/2 phosphorylation (Fig. 2A) and induced only minor repression of SESN3 gene transcription, both in human and in rat cells (Fig. 2A and B). Furthermore, inhibition of MEK1 by PD98059 or p38α and p38β by SB203580 attenuated substantially Ras-induced SESN3 gene repression (Fig. 3). Probably both Ras-Raf-MEK-ERK and Ras-p38 pathways contribute to SESN3 repression. On the other hand, inhibition of MEK1 and p38 differentially affected Ras-induced down-regulation of distinct SESN1 transcripts: PD98059 attenuated repression of T2 but did not influence T3 expression; the SB203580 counteracted the effect of Ras on expression of T3 rather than of T2 (Fig. 3). MAPK-independent pathways seem to be also involved in regulation of SESN1 gene: the RalGDS-interacting V12G37 Ras mutant repressed SESN1 gene in rat (Fig. 2B) and human cells, where the strongest down-regulation was observed for the T3 transcript (Fig. 2A).
To estimate possible contribution of repression of sestrins to Ras-induced ROS up-regulation, we independently expressed various isoforms of sestrins from introduced lentiviral constructs. The MDAH041/tet-Ras cells bearing SESN1, SESN2, or SESN3 constructs were more resistant to ROS increase upon switched-on expression of Ras (Fig. 4A). In Ras-expressing cells bearing exogenous SESN1 T2 or T3 cDNAs, the ROS level was increased 1.3- to 1.4-fold, compared with a 3-fold increase in the Ras-expressing control. We also studied the effects of inhibition of SESN1, SESN2, and SESN3 genes by lentiviral transduction of corresponding shRNA-expressing constructs (Supplementary Fig. S2). An inhibition of each of the sestrin genes in human MDAH041 cells caused ROS up-regulation and stimulation of mutagenesis similar to those observed in their Ras-expressing derivatives (Tables 1 and 2). Similar results were obtained with rat REF52 and Rat1 cells expressing an shRNA (21) to a conservative region of SESN1 and SESN3 mRNA, which simultaneously targets both transcripts (Supplementary Fig. S2; Table 2).
A, effects of sestrins on Ras-induced ROS up-regulation. MDAH041/tet-Ras cells infected with lentiviral vectors expressing SESN1, SESN2, or SESN3 cDNAs were introduced into MDAH041/tet-Ras cells using lentiviral vectors. On day 2 after infection, tetracycline was removed from culture medium and DCF fluorescence was assayed on day 7. Left, distribution of 104 cells according to the intensity of DCF-DA fluorescence in tested cell cultures; right, average indices of total DCF-DA fluorescence of 104 cells calculated from two independent experiments. B, kinetics of changes in ROS levels in Ras-expressing p53-positive REF52/tet-Ras cells, p53-deficient REF52/tet-Ras/si-p53 cells, and p53-negative MDAH041/tet-Ras cells. The cell cultures were analyzed 2 to 15 d after tetracycline withdrawal. Average result of two independent experiments. C, changes in expression of Ras, p53, and p21Cip1/Waf1 proteins and SESN1, SESN2, and SESN3 mRNAs in REF52/tet-Ras and REF52/tet-Ras/si-p53 cell cultures incubated without tetracycline for 2 to 15 d. Western blot analysis of Ras, p53, and p21Cip1/Waf1 proteins and RT-PCR detection of mRNAs.
A, effects of sestrins on Ras-induced ROS up-regulation. MDAH041/tet-Ras cells infected with lentiviral vectors expressing SESN1, SESN2, or SESN3 cDNAs were introduced into MDAH041/tet-Ras cells using lentiviral vectors. On day 2 after infection, tetracycline was removed from culture medium and DCF fluorescence was assayed on day 7. Left, distribution of 104 cells according to the intensity of DCF-DA fluorescence in tested cell cultures; right, average indices of total DCF-DA fluorescence of 104 cells calculated from two independent experiments. B, kinetics of changes in ROS levels in Ras-expressing p53-positive REF52/tet-Ras cells, p53-deficient REF52/tet-Ras/si-p53 cells, and p53-negative MDAH041/tet-Ras cells. The cell cultures were analyzed 2 to 15 d after tetracycline withdrawal. Average result of two independent experiments. C, changes in expression of Ras, p53, and p21Cip1/Waf1 proteins and SESN1, SESN2, and SESN3 mRNAs in REF52/tet-Ras and REF52/tet-Ras/si-p53 cell cultures incubated without tetracycline for 2 to 15 d. Western blot analysis of Ras, p53, and p21Cip1/Waf1 proteins and RT-PCR detection of mRNAs.
Effect of inhibition of sestrin by shRNAs on ROS levels and frequency of chromosome breaks
Cell line . | DCF fluorescence . | Chromosome breaks . | . | |
---|---|---|---|---|
. | . | % Metaphases . | Breaks per metaphase . | |
MDAH041 | 57 ± 9 | 33 ± 1.2 | 0.67 ± 0.05 | |
MDAH041/shRNA SESN1 | 92 ± 9 | 54 | 1.7 | |
MDAH041/shRNA SESN2 | 101 ± 11 | 58 | 1.8 | |
MDAH041/shRNA SESN3 | 105 ± 10 | 51 | 1.6 | |
REF52 | 23 | 12 | 0.22 | |
REF52/shRNA SESN1/3 | 35 | 26 | 0.40 | |
REF52/shRNA SESN2 | 39 | 30 | 0.34 | |
Rat1 | 10 | 12 | 0.12 | |
Rat1/shRNA SESN1/3 | 39 | 24 | 0.36 | |
Rat1/shRNA SESN2 | 45 | 28 | 0.52 |
Cell line . | DCF fluorescence . | Chromosome breaks . | . | |
---|---|---|---|---|
. | . | % Metaphases . | Breaks per metaphase . | |
MDAH041 | 57 ± 9 | 33 ± 1.2 | 0.67 ± 0.05 | |
MDAH041/shRNA SESN1 | 92 ± 9 | 54 | 1.7 | |
MDAH041/shRNA SESN2 | 101 ± 11 | 58 | 1.8 | |
MDAH041/shRNA SESN3 | 105 ± 10 | 51 | 1.6 | |
REF52 | 23 | 12 | 0.22 | |
REF52/shRNA SESN1/3 | 35 | 26 | 0.40 | |
REF52/shRNA SESN2 | 39 | 30 | 0.34 | |
Rat1 | 10 | 12 | 0.12 | |
Rat1/shRNA SESN1/3 | 39 | 24 | 0.36 | |
Rat1/shRNA SESN2 | 45 | 28 | 0.52 |
NOTE: The cells were analyzed 7 to 10 d after lentiviral transfer of corresponding shRNAs. In each cell culture, 30 to 50 metaphases were analyzed.
p53-mediated reversion of the high ROS levels induced by oncogenic Ras. We monitored changes in intracellular ROS levels at different time intervals following onset of oncogenic Ras expression. We observed in the p53-deficient MDAH041/tet-Ras cells a stable increase in ROS. In the p53-positive REF52/tet-Ras cells, the increase was transient, and by day 15 the intracellular ROS content retreated to the initial level (Fig. 4B). The observed reversion in ROS level correlated with gradual accumulation of p53 and its target gene product p21Cip1/Waf1, as well as an increased expression of SESN1, a p53-responsive antioxidant gene (Fig. 4C). To test whether the reversion was dependent on a function of p53, we inhibited p53 in REF52/tet-Ras cells by expression of shRNA. There was no induction of p53 and p21Cip1/Waf1 and SESN1 in response to oncogenic H-Ras expression. Concomitantly, there was no significant reversion of elevated ROS as compared with that observed in the p53-positive REF52 cells (Fig. 4B).
Discussion
Previous studies indicate that expression of activated Ras oncogene can cause an increase in intracellular ROS levels (15) and genetic instability (11), although underlying mechanisms as well as possible connection between these effects were not established. Here we present evidence for a substantial role of ROS up-regulation in Ras-induced mutagenesis. Indeed, in all tested cell lines with constitutive and conditional expression of oncogenic Ras, we observed a direct correlation between elevated levels of intracellular ROS, accumulation of 8-oxo-2′-deoxyguanosine (a mutagenic product of DNA oxidation; reviewed in ref. 31), and frequent chromosome breaks in mitotic cells. Moreover, treatment with antioxidant NAC was capable of restoring normal levels of intracellular ROS, which also abrogated the Ras-induced mutagenesis. Thus, the increased ROS level can be a major cause of genetic instability in cells expressing oncogenic Ras.
Oncogenic Ras can induce an increase in intracellular ROS levels by stimulation of membrane-bound NAD(P)H-dependent oxidases that generate superoxide anion from molecular oxygen (32, 33). NAD(P)H-dependent oxidases belong to a family of multi-subunit enzymes composed of a membrane-bound heme-containing cytochrome b558 complex (a catalytic gp91 tightly associated with cytochrome b light chain p22), cytosolic p47 or Noxo1 (Nox-organizer), and p67 or Noxa1 (Nox-activator) proteins and a small GTPase of the Rac family (34, 35). Several lines of evidence suggested that expression of activated Ras can cause an increase in ROS levels through activation of the PI3K-Rac1/NAD(P)H oxidase pathway (17, 19, 20). Here, we found that in all tested cell lines, the transduction of dominant-negative Rac1N17, which inhibits the function of NAD(P)H oxidase (36), caused only partial (30–40%) reduction of Ras-induced increase in intracellular ROS levels. In addition, transduction of mutant RasV12C40 that selectively activates PI3K resulted in only moderate increase in intracellular ROS. Therefore, we assumed that there might be additional mechanisms for Ras-induced ROS-up-regulation that are not related to activation of NAD(P)H oxidase through PI3K and Rac.
We decided to test whether Ras can act downstream of NAD(P)H oxidase, possibly by interfering with the decomposition of produced hydrogen peroxide. Cytosolic peroxiredoxins play a major role in decomposition of endogenously produced H2O2 (37, 38) whereas their activity is regulated by combined action of sulfiredoxin (21, 39, 40) and proteins of the sestrin family (21, 24). We found that expression of H-RAS and N-RAS oncogenes in human and rat fibroblast cells is accompanied by a decrease in expression of mRNAs derived from the sestrin family genes. SESN3 gene showed the strongest repression in all tested cell lines. A decrease in expression of SESN1 was also prominent whereas SESN2 gene showed only minor repression in human cells and no visible changes in rat cells. The data obtained with restricted effector specificity Ras mutants that selectively activate the Raf-ERK, PI3K, and RalGDS pathways, as well as with selective inhibitors of MEK1 and p38 MAPKs, suggest that activation of the Raf-MEK-ERK and p38 pathways may play a major role in repression of the SESN3 gene. Repression of the SESN1 gene might be subject to a more complex regulation. Both the RasV12S35 mutant that selectively activates the Raf-MEK-ERK pathway and the RasV12G37 mutant that stimulates RalGDS were similarly active in the down-regulation of rat SESN1 mRNA and human SESN1 mRNA isoform T3. The data obtained with SB203580 inhibitor suggest that the Ras-p38 pathway might also participate in the down-regulation of SESN1 mRNA isoform T3. At the same time, the mutant RasV12S35 was more effective than RasV12G37 in repressing human SESN1 mRNA isoform T2. The result of experiment with MEK1 inhibitor PD98059 confirmed a major role of the Raf-MEK-ERK pathway in the down-regulation of human SESN1 mRNA isoform T2. The detailed mechanisms by which Ras-induced activation of the MAPK and RalGDS pathways mediate repression of SESN3 and SESN1 genes remain to be established.
To assess possible biological significance of the Ras-induced decrease in expression of SESN3 and SESN1 mRNAs, we used two approaches. First, we studied how elimination of this effect can influence ROS up-regulation. For this purpose, we analyzed ROS levels in Ras-expressing cells, which additionally express various exogenous sestrin cDNAs introduced by lentiviral transfer. This experiment showed that restoration of expression of each of the sestrin genes interfered with the ROS up-regulation, suggesting a substantial role of sestrin gene repression in Ras-induced ROS increase. Then, using RNA interference, we studied the effects of repression of sestrin genes. In human MDAH041 cells, a shRNA-induced decrease in expression of each of the SESN1, SESN2, and SESN3 genes leads to a nearly 2-fold increase in ROS levels and a 2- to 3-fold increase in frequency of chromosome breaks in dividing cells, which is comparable to the effects produced by activated RAS. Similar effects were observed in rat cells with simultaneously inhibited expression of SESN1 and SESN3 mRNAs. These data suggest that down-regulation of sestrins could contribute to the increased ROS level and accelerated mutagenesis in cells expressing oncogenic RAS.
The p53 function is known to be up-regulated in some types of Ras-expressing cells (41) including the REF52 cell line (11, 22). The effect of p53 on ROS level seems to depend on the degree of stress and the degree of p53 activation, being pro-oxidant at high stresses and antioxidant at nearly physiologic conditions (24). Earlier, we reported that the rate of Ras-induced mutagenesis is higher in p53-deficient cells as compared with isogenic cells expressing functional p53 (11). In part, the synergy in the induction of chromosome instability by oncogenic RAS in the p53-null background is explained by the ability of Ras to attenuate the p53-independent components of the DNA damage–induced G1 and G2 phase checkpoints (11, 16). In the present study, we found a connection between up-regulated ROS and inhibited expression of sestrin family genes, which might represent an additional factor enhancing Ras-induced mutagenesis in p53-deficient cells. In the cell lines with unimpaired p53 activation (particularly in the REF52 cell line), the increase in ROS is temporal, returning to normal level after 10 to 15 days of oncogenic Ras expression. By contrast, in the REF52 cells with inhibited p53, as well as in the human p53-deficient MDAH041 cells, the Ras-induced increase in ROS level is permanent. Possibly, the observed reversion of elevated ROS to normal levels in Ras-expressing REF52/tet-Ras cells is connected with the up-regulation of a set of the transcriptional target genes of p53, which include those with antioxidant function, such as GPX1 (42, 43), SOD2 (42), ALDH4A1 (44), SESN1, and SESN2 (21, 24). In fact, in the Ras-expressing p53-positive REF52 cells, we observed a gradual restoration of SESN1 gene expression, which was not seen in the corresponding p53-deficient cultures. The p53-mediated feedback mechanisms counteracting Ras-induced ROS up-regulation can prevent the previously observed accelerated mutagenesis in p53-deficient cells expressing oncogenic Ras (11). Collectively, our results highlight the important role of Ras-induced inhibition of antioxidant defense in acceleration of genomic instability and cancer progression.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
B.P. Kopnin and P.M. Chumakov contributed equally to this work.
Acknowledgments
Grant support: Russian Foundation for Basic Research grants 05-04-48969 (L.S. Agapova) and 05-04-48269 (B.P. Kopnin), International Research Scholars Program of the Howard Hughes Medical Institute grants 55000321 (B.P. Kopnin) and 55005603 (P.M. Chumakov), and NIH R01 grants AG25278 and R01 CA10490 (P.M. Chumakov).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.