|Home | About | Journals | Submit | Contact Us | Français|
Transcriptional regulation of the antioxidant response element (ARE) by Nrf2 is important for the cellular adaptive response to toxic insults. New data show that primary skin-derived fibroblasts from the long-lived Snell dwarf mutant mouse, previously shown to be resistant to many toxic stresses, have elevated levels of Nrf2 and of multiple Nrf2-sensitive ARE genes. Dwarf-derived fibroblasts exhibit many of the traits associated with enhanced activity of Nrf2/ARE, including higher levels of glutathione and resistance to plasma membrane lipid peroxidation. Treatment of control cells with arsenite, an inducer of Nrf2 activity, increases their resistance to paraquat, hydrogen peroxide, cadmium, and UV light, rendering these cells as stress resistant as untreated cells from dwarf mice. Furthermore, mRNA levels for some Nrf2-sensitive genes are elevated in at least some tissues of Snell dwarf mice, suggesting that the phenotypes observed in culture may be mirrored in vivo. Augmented activity of Nrf2 and ARE-responsive genes may coordinate many of the stress resistance traits seen in cells from these long-lived mutant mice.
The discovery of single gene mutations that extend life span, first in invertebrates (43, 48, 60) and then in mice (9, 14, 19), has provided new momentum for defining the molecular mechanisms that control the aging process. Since Harman first proposed the free radical theory of aging (23), many lines of evidence have suggested that oxidative stress plays an important role in aging. In roundworms (Caenorhabditis elegans) (44, 52) and fruit flies (Drosophila melanogaster) (7, 67, 109), mutations resulting in resistance to toxic stresses, both oxidative and otherwise, tend to result in increases in longevity. The relative importance of oxidation damage as a regulator of life span is more controversial. Longevity is often associated with resistance to oxidative injury within and among species, but most attempts to retard aging by antioxidant treatments have failed to show beneficial effects, and mutations that promote oxidative damage in mice have often had little impact on life span (25, 72, 81, 86, 96). Utilizing cells from the Snell dwarf mouse, a model of extended longevity, we are attempting to find the mechanism behind cellular stress resistance in hopes of relating this resistance to the delayed aging of the Snell dwarf animal.
Snell dwarf mice are homozygous for a mutation at the Pit-1 locus which causes improper development of the anterior pituitary, leading to low levels of growth hormone, thyroid stimulating hormone, and prolactin in young and adult mice (10, 30). These pituitary defects lead to diminished circulating levels of insulin-like growth factor 1 (IGF-1) and thyroxine, which in turn result in reduced size and hypothermia. Snell dwarf mice, like the closely related Prop1 mutant Ames dwarf (9), live approximately 40% longer than littermate controls on several different background stocks (19, 20) and show delay in many forms of aging-related pathologies (1, 3, 19).
Previous work has shown that primary dermal fibroblasts derived from the tails of young adult Snell dwarf mice are resistant to many types of cytotoxic stress (71, 93), including agents that kill cells at least in part via reactive oxygen species (ROS), such as paraquat, peroxide, cadmium, and others that cause cell death through other pathways, including UV light, the DNA alkylating agent methyl methanesulfonate (MMS), and heat. In addition, fibroblasts from dwarf mice are resistant to the nonlethal effects of culture in the presence of rotenone or at low concentrations of glucose, each of which diminishes the cells' ability to reduce an external tetrazolium dye (WST-1) (55). Relative resistance to the lethal effects of cadmium, heat, peroxide, and MMS and to the metabolic effects of rotenone and low glucose was also characteristic of fibroblasts from relatively long-lived rodent species, suggesting that this property might be involved in the evolution of longevity (24). The mechanisms that underlie the metabolic and stress resistance of cells from Snell dwarf mice and species of long-lived rodents are not yet understood.
The plasma membrane redox system (PMRS) is a group of NAD(P)H reductases which serve to pass electrons into and across the plasma membrane (35). The PMRS acts to maintain the redox status of important antioxidants such as coenzyme Q ([CoQ] ubiquinone), α-tocopherol, and ascorbate (35, 61, 74), which are thought to protect cells from exogenous oxidative stress, in particular, by protecting membranes from lipid peroxidation chain reactions (33). The PMRS may also have a function in cellular communication via reactive oxygen species (18, 37) and in maintaining cellular redox status via NAD(P)H recycling (4). The specific proteins that contribute to PMRS function are not completely characterized but include NAD(P)H:quinone oxidoreductase 1 (NQO1) and ascorbate free radical reductase (also known as cytochrome b5 reductase), as well as several activities attributed to multiple proteins (35, 63). Reduction of extracellular tetrazolium dyes, including WST-1, has been used previously as an index of integrated whole-cell PMRS activity, but the extent to which WST-1 reduction is influenced by specific plasma membrane enzyme complexes, intracellular electron donors, and extracellular electron acceptors is not fully understood (5, 28, 102).
The expression of at least one PMRS enzyme (NQO1) is under the control of a promoter known as the antioxidant response element (ARE) (54, 105). The ARE, also called the EpRE (electrophile response element), is a transcriptional promoter element thought to be important for cellular adaptation to oxidative stress (42, 76). ARE-related genes are involved in many aspects of stress resistance, including glutathione regulation, catalase expression, cellular redox control, and proteasome function (47). The activity of ARE promoters is modulated by NF-E2-related factor 2 (Nrf2), which was discovered in 1994 based on its ability to bind the NF-E2/AP1 repeat in the promoter of the beta-globin gene (68). Nrf2 was extensively characterized by Yamamoto et al., who showed that Nrf2 is a posttranslationally regulated gene whose activation of the ARE is crucial for the adaptive response to electrophiles (39-41, 50). Under normal, unstressed circumstances, Nrf2 is found mostly in the cytosol bound to kelch-like ECH-associated protein 1 (Keap1), which promotes its ubiquitination and degradation by the proteasome (111). Nrf2 can be dissociated from Keap1 indirectly through signal transduction (phosphorylation) or directly through oxidative damage to crucial cysteine residues on the Keap1 protein. After dissociation from Keap1, Nrf2 translocates to the nucleus and binds to promoters containing the ARE sequence, activating their transcription (50). There are many proteins with ARE sequences in their promoters, most of which help protect against ROS, including enzymes involved in glutathione synthesis (38) and maintenance (69), protein turnover (51), antioxidant expression (83, 84), oxidant inactivation (27), NADPH synthesis (26, 103), toxin export (26), and prevention of inflammation (85).
Despite the well-documented relationship between Nrf2 and protection from oxidative stress, relatively little has been published evaluating the role of Nrf2 in aging. Recent studies in C. elegans have suggested that the Nrf2 homologue, SKN-1, is necessary for the life span extension seen with dietary restriction (6) and insulin pathway disruption, and overexpression of SKN-1 can increase worm life span (104). Further work has suggested that a Keap1 heterozygous loss-of-function mutation in D. melanogaster can increase Nrf2 activity, increasing both oxidative stress resistance and life span in male flies only (101). Analogous studies in mice have been complicated by the lack of viability of the Keap1 knockout mouse (107) and have been mostly limited to observations of decreased Nrf2 signaling with increased age (97) and increased Nrf2 signaling with calorie restriction (80). We report here a series of studies consistent with the idea that augmented Nrf2 activity may contribute to several forms of stress resistance seen in cultured fibroblasts from long-lived Snell dwarf mice and that this phenotype may extend to the tissues of dwarf mice.
Snell dwarf (dw/dw) mice, and heterozygote (dw/+) controls were bred as the progeny of (DW/J × C3H/HeJ) dw/+ females and (DW/J × C3H/HeJ)F1 dw/dw males. Sires were treated with growth hormone and thyroxine for increased body size and fertility. Tail skin biopsies were taken from 3- to 6-month-old mice.
Tail skin biopsies (3 to 5 mm) were obtained and cultured as previously described. Briefly, skin samples were washed, diced, and digested overnight in collagenase type II (400 U/ml; 1,000 U total per tail; Gibco-Invitrogen, Carlsbad, CA) dissolved in Dulbecco's modified Eagle's medium (DMEM) supplemented with 20% heat-inactivated fetal bovine serum (Atlanta Biological, Lawrenceville, GA), antibiotics, and fungizone (complete medium) at 37° with 5% CO2 in air. After collagenase treatment, cells were dislodged, centrifuged, and resuspended in complete medium. Approximately 2.5 × 105 cells in 3 ml of medium were seeded into tissue culture flasks of 25-cm2 surface area, and this was called passage 0. Cells were fed after 3 days (2/3 volume of medium replacement) and split by trypsinization after 7 days into 75- or 175-cm2 flasks at a density of 1 × 104 cells/cm2. Each passage was split at a 7-day interval, with approximately 2/3 total volume of medium replaced at day 3. Cells used in the assays described were confluently plated at ~105 cells/cm2 in the third passage.
Cells for metabolic studies were plated confluently in 60-mm dishes. Following overnight incubation, cells were washed three times in phosphate-buffered saline (PBS) and treated in parallel with various metabolic stresses for 2 h. They were then washed in ice-cold PBS and lysed with either trichloroacetic acid (TCA) or sodium hydroxide (NaOH). Lysates were centrifuged to remove protein and neutralized to ~pH 6.5 before testing. Lysates were then separated and quantified on a high-performance liquid chromatography (HPLC) column as previously described (17, 65).
Plasma membranes were purified by the method of Navas et al. (75). Briefly, cells were trypsinized as described above and plated confluently (~6 × 106 cells) in 150-mm Falcon tissue culture dishes (Becton-Dickinson). After the cells were allowed to adhere overnight, the medium was removed, and the cells were washed with ice-cold PBS, scraped into 14-ml tubes, and spun down at 4° for 5 min at 1,000 × g. PBS was removed, followed by a 10-min incubation in swelling buffer (10 mM HEPES-KOH, 15 mM KCl, 1.5 mM Mg acetate, 1 mM dithiothreitol [DTT], 1 mM phenylmethylsulfonyl fluoride [PMSF]). Cells were then broken in a dounce homogenizer for approximately 2 min, followed by addition of concentrated swelling buffer (80 mM HEPES-KOH, 150 mM KCl, 8.5 mM Mg acetate, 1 mM DTT, 1 mM PMSF). Homogenates were spun down for 10 min at 100 × g to remove unbroken cells and nuclei, followed by a 1-h centrifugation at 40,000 × g. Pellets were resuspended in water and subjected to aqueous two-phase partition (6.6% dextran T500 [Pharmacosmos, Holbaek, Denmark], 6.6% polyethylene glycol [Integra, Renton, WA], 0.25 sucrose, 5 mM potassium phosphate, pH 7.2). Phases were mixed by inversion, followed by centrifugation at 750 × g, and the upper phase was transferred to a centrifuge tube and spun down for 1 h at 40,000 × g. Pellets were resuspended in water, flash frozen in liquid nitrogen, and stored at −80° for enzymatic assays. Plasma membrane purity was measured as previously described (16), using antibodies (Santa Cruz Biotechnology) specific for the proteins of the plasma membrane (anti-Na+,K+-ATPase α-subunit), endoplasmic reticulum (anti-ribophorin I), and mitochondria (anti-cytochrome c oxidase subunit I), using immunoblotting techniques described below.
PMRS enzymatic assays were performed as previously described on 10 μg of purified plasma membrane protein. For cytochrome c reductase (34), sample was added to a final solution containing 50 mM Tris, pH 7.6, 0.2 mM NADH, 0.1% Triton X-100, and 20 μmol/liter cytochrome c. For NADH CoQ reductase activity (34), enzyme sample was added to a solution containing 50 mM Tris, pH 7.6, 0.2 mM NADH, 0.1% Triton X-100, and 0.2 mmol/liter coenzyme Q0. NADH-ascorbate free radical (AFR) reductase (16) absorbance at 340 nm was measured kinetically after addition of 66 × 10−3 units of ascorbate oxidase in solution containing 50 mM Tris-HCl, pH 7.6, 0.2 mM NADH, 0.1% Triton X-100, 0.4 mM fresh ascorbate, and 20 μg of plasma membrane. NQO1 (16) (DT-diaphorase) activity was measured at 550 nm in solution containing 50 mM Tris-HCl, pH 7.6, 0.2 mM NADPH, 0.1% Triton X-100, 10 μM menadione, and 75 μM cytochrome c, with or without 10 μM dicoumarol. NQO1 activity was calculated as the difference between the uninhibited sample and the dicoumarol inhibited sample.
Cells were measured for lipid peroxidation as previously described (29). Briefly, cells were plated confluently (~106 cells per well) on six-well plates and allowed to adhere overnight. Cells were then washed and placed in either complete or serum-free medium containing 2% bovine serum albumin (BSA) for ~24 h. The medium was removed, the cells were washed with PBS, and the various media (in either complete medium or DMEM) were added in a solution containing 50 μM diphenyl-1-pyrenylphosphine (DPPP; Invitrogen). Media used were complete (DMEM plus 10% serum), serum-free (DMEM only), complete with stress (50 μM hydrogen peroxide or 40 μM rotenone), and glucose-free media(10% serum) and DMEM with stress (50 μM hydrogen peroxide or 10 μM cadmium chloride). Cells were incubated at 37°C in the dark for 30 min, after which they were washed twice with PBS and scraped in 125 μl of PBS, and 100 μl of cell solution was added to a 96-well plate. The plate was read for DPPP fluorescence (with excitation at 340 nm and emission at 405 nm) on a Spectramax 96-well fluorescence plate reader. The remaining 25 μl of solution was tested for protein content.
ROS production was measured by the change in fluorescence of the mitochondrial superoxide sensing dye Mitosox (Invitrogen) and the cytosolic superoxide sensing dye dihydroethidium (DHE) (Invitrogen). Briefly, cells plated confluently in six-well plates were exposed to 24-h serum deprivation, 3 h of rotenone (40 μM with serum), or no stress. Mitosox was added at 4 μM, and DHE was added at 3 μM for 30 min each before cells were washed in ice-cold PBS and scraped into tubes. Cells (100 μl) were measured for fluorescence at 485-nm excitation and 590-nm emission for Mitosox and at both 518-nm excitation with 605-nm emission and 355-nm excitation with 420-nm emission for DHE. The remaining cells were tested for protein content.
Lysates for Western blots were prepared in radioimmunoprecipitation assay (RIPA) buffer. Lysates were aliquoted and stored at −80° until the date of use. All blots were run on SDS-PAGE gels containing 10% acrylamide (Bio-Rad). All primary and secondary antibodies for Nrf2, Nrf1, and Keap1 were obtained from Santa Cruz Biotechnology. Secondary antibodies were conjugated with alkaline phosphatase, and blots were read after enhanced chemifluorescence (ECF; Amersham) addition on a Storm 840 fluorescence scanner. Quantification was performed using ImageQuant software.
Cells plated confluently in 96-well plates were washed in PBS and exposed to low doses of Nrf2 activators, arsenite, and tert-butyl hydroquinone (tBHQ). After 24-h incubations with these activators, WST-1 was added as described below (see arsenite survival); cells were incubated for 3 h and read on a 96-well plate reader.
Cells plated confluently in six-well plates were either treated with 24-h serum deprivation or 24-h low-level arsenite exposure or were left untreated (fresh medium). After 24 h, RNA was prepared using an RNeasy kit (Qiagen) following the manufacturer's instructions. RNA was frozen at −80° and stored until used for reverse transcription-PCR (RT-PCR) assays. RT-PCR was performed using a SYBR green RT-PCR kit and Quantitect primers (Qiagen) as described by the manufacturer, using a Corbett Rotor-Gene 3000 thermal cycler.
Cells were treated as in lipid peroxidation measurements, except that cells were incubated for 2 h before dye (monochlorobimane; Invitrogen) was added to the cells, and fluorescence was read at an excitation of 380 nm and an emission 465 nm.
Total glutathione (GSH) was measured using a total glutathione measurement kit (Cayman Chemicals). The principle of the kit is an enzymatic cycling reaction where GSH reacts with 5,5′-dithio-bis-2-(nitrobenzoic acid) (DTNB; Ellman's reagent), producing a yellow acid (5-thio-2-nitrobenzoic acid [TNB]). The mixed disulfide produced by the reaction is reduced by glutathione reductase to GSH and produces more TNB. The rate of color (TNB) production is proportional to the concentration of total glutathione. Briefly, cells were prepared and scraped as above, followed by manual homogenization for 30 s with a Pellet Pestle (Fisher). Homogenates were centrifuged at 10,000 × g for 15 min, followed by incubation on ice prior to testing. Fifty microliters of sample was added to 150 μl of glutathione cocktail (enzymes, cofactors, and Ellman's reagent), and the samples were read on a 96-well spectrophotometer at 405 nm every 2 min for 30 min, with constant shaking. Results were compared to control curves for GSSG (oxidized GSH) standard to obtain glutathione values.
Cells were plated confluently (3 × 106 cells per well) in 96-well plates and allowed to adhere overnight. Medium was removed, cells were washed with PBS, medium (DMEM, 20% FCS, penicillin-streptomycin, and fungizone) containing various concentrations of inhibitors was added, and cells were incubated for 1 h at 37° and 5% CO2 before addition of 5 μl of WST-1 (Roche). After 3-h incubations with WST-1, plates were read at 585 nm on a 96-well spectrophotometer to obtain inhibition results. After plates were read, inhibitory medium was removed, and cells were washed in PBS and returned to standard medium. After a 24-h recovery period cells were tested again with WST-1 to measure survival. Results were verified by measuring thymidine uptake as described previously (71).
Stress assays were performed as previously described (93), with slight modifications. Briefly, cells were plated confluently on 96-well plates, allowed to adhere overnight, and either serum starved for 24 h, glucose starved for 24 h (glucose-free testing), left in complete medium (rotenone testing), or serum starved with the addition of 5 μM arsenite. Rotenone assays were dosed with 5 μM rotenone 2 h prior to peroxide addition. Except for UV assays, cells were dosed with stressor for 6 h, after which cells were washed with PBS and returned to the previous medium without glucose, rotenone, or arsenite stress. UV plates were washed with warm PBS and placed in PBS while dosed with UV using a Stratolinker irradiation system set to the desired dosage, after which they were returned to serum-free (recovery) medium. Survival was measured the following day by WST-1 reduction.
Tissues of 4- to 6-month-old dwarf and control mice were prepared by RNeasy kit preparation, as previously described (100).
Protein content was measured by Coomassie blue (Bradford assay; Bio-Rad), as described by the manufacturer, except where detergent was used in cell lysis. Where detergent was present, a detergent-compatible (DC) protein assay (Bio-Rad) was used as described by the manufacturer.
Statistical analyses were performed using paired t tests (two-tailed) for comparison between dwarf and control cells as each dwarf and control cell line pair was tested together. Graphs show the mean of combined experiments, and error bars represent the standard error of the mean. Differences between treatments were calculated using repeated-measures analysis of variance (ANOVA). For calculation of 50% lethal doses (LD50s), mean survival was calculated at each dose, and the LD50 was calculated using probit analysis using NCSS software (NCSS, Kaysville, UT).
Unless otherwise stated, all chemicals were obtained from Sigma-Aldrich, St. Louis, MO.
Previous data have shown that primary fibroblasts from the skin of Snell dwarf mice are resistant to the nonlethal inhibition of PMRS function induced by either glucose withdrawal or rotenone exposure, as measured by reduction of the extracellular dye WST-1 (55). Resistance to PMRS inhibition is correlated in cells from normal mice with the level of resistance to the lethal effects of hydrogen peroxide and cadmium and is also seen in cells from long-lived rodent species (24, 55). Because the enzymes responsible for WST-1 reduction require NAD(P)H and because both glucose deprivation and rotenone exposure affect NAD(P)H levels, we evaluated the redox status of NAD(P)H in dwarf and control fibroblasts under control and stressed conditions (Fig. (Fig.1).1). As expected, NADH levels were reduced in low-glucose medium and increased by rotenone, but there were no significant differences between the dwarf and littermate controls in any of the conditions tested. Similarly, there were no differences between dwarf and control cells in levels of NADPH or in the ratio of NAD(P) to NAD(P)H in any condition. These data suggest that differences between dwarf and normal cells in the ability to reduce WST-1 are unlikely to be due to differences in levels of NAD(P) metabolites but do not rule out the possibility that differences in specific cellular compartments (e.g., cytosolic) in NAD(P)H could be important for exogenous dye reduction by dwarf cells under metabolic stress.
Having ruled out NAD(P)H redox status as the mechanism for resistance of dwarf cells to inhibitors of PMRS function, we next examined the enzymes that contribute to PMRS function per se. To do this, we prepared plasma membranes from dwarf and control fibroblasts and tested for in vitro activity levels of CoQ reductase, ascorbate free radical (AFR) reductase, cytochrome c reductase, and NQO1 (Fig. (Fig.2).2). We noted a significant increase in CoQ and AFR reductase activity in membranes prepared from dwarf cells. There was no difference in cytochrome c reductase activity. Consistent with prior studies (59, 91), we found very little membrane-associated NQO1 although this activity was readily detected, at equal levels, in whole-cell lysates from control and dwarf cells (not shown). The data on CoQ and AFR reductases show that some elements of the PMRS may be more active in plasma membranes from dwarf-derived cells.
The PMRS is thought to be important for protection against plasma membrane lipid peroxidation elicited by exogenous or endogenous oxidative stress. Using a dye (diphenyl-1-pyrenylphosphine, or DPPP) that intercalates into the membrane and reacts with lipid hydroperoxides, we examined the lipid peroxide formation of normal and dwarf fibroblasts at baseline and under stressed conditions. As shown in Fig. Fig.2,2, there were no differences between dwarf and normal cells in DPPP fluorescence in unstressed cells cultured with or without serum. H2O2, tert-butyl hydroperoxide, and cadmium each led to an increase in lipid peroxidation, which was significantly higher in control than in dwarf cells. Low-glucose culture led to a significant increase in lipid peroxidation only in normal cells and had no effect on dwarf cells. In separate experiments (Fig. (Fig.2),2), we evaluated intracellular production of ROS using Mitosox and DHE as indicators (89) and found no difference between normal and dwarf cells in medium with or without serum or in the presence of 40 μM rotenone. The relative resistance of stressed dwarf fibroblasts to lipid peroxidation is consistent with the data suggesting that higher levels of PMRS function in the plasma membranes of these cells.
If the PMRS helps protect cells against plasma membrane oxidation damage, inhibition of the PMRS should sensitize cells to the lethal effects of H2O2. Consistent with this idea, Fig. Fig.22 shows that exposure of control cells to rotenone or culture in low-glucose medium can each lead to a significant decrease in LD50 for H2O2. The suggestion that elevated PMRS may contribute to resistance to lethal oxidants is consistent with our previous work (55) showing (i) that individual mice whose cells are relatively resistant to the inhibition of PMRS by glucose withdrawal are also resistant to the lethal effects of H2O2 and cadmium and (ii) that resistance to the lethal effects of these two agents accompanies the resistance to PMRS inhibition noted in fibroblasts from long-lived rodent species (24).
The increase in PMRS activity could in itself account for resistance of dwarf cells to the metabolic effects of rotenone or low-glucose medium and contribute to some forms of cytotoxic stress noted in the dwarf cells. We next wished to see if broader changes in signal transduction could account for increases in PMRS activity and simultaneously for a broader spectrum of stress resistance pathways. We focused on the antioxidant response element (ARE), which has been reported to regulate expression of at least one PMRS enzyme (NQO1) (90). To test if the ARE could affect the PMRS activity involved in WST-1 reduction by skin-derived fibroblasts, we exposed cells to low levels of the ARE activators arsenite and tert-butyl hydroquinone (tBHQ). The results (Fig. (Fig.3)3) showed that ARE activation increased WST-1 reduction by 20 to 60%, depending on the cell line, consistent with a role for ARE induction in increasing PMRS-mediated WST-1 reduction.
To test if the increased PMRS activity in dwarf cells might be due to an increase in ARE activation, we measured the steady-state levels of the transcription factor Nrf2, a key mediator of ARE activation. As a positive control we used a lysate from an Nrf2-transfected 293T cell line (sc-12742; Santa Cruz Biotechnology), which showed a single band at ~66 kDa, consistent with previously published studies (13, 58, 82), as shown in Fig. Fig.4E.4E. Analysis of lysates derived from unstressed cells or after exposure to the Nrf2 activator sodium arsenite showed that dwarf cells have 25% to 35% more Nrf2 protein than cells from littermate controls (Fig. (Fig.4).4). The difference between dwarf and normal cells was significant (P < 0.01), and both cell types responded to an equal degree to arsenite stress. Many investigators have also observed a ~110-kDa band corresponding to Nrf2 in some cell lines and tissue lysates (13, 56, 57), which we also observed in our cell lysates (though not in the positive control); this 110-kDa band was elevated in dwarf cells by 23% (P = 0.07) in unstressed conditions and 18% (P = 0.05) after arsenite exposure.
Since the degradation of Nrf2 is regulated by its binding partner Keap1, we also tested to see if Keap1 was downregulated in dwarf cells, but we found no differences between genotypes or conditions in steady-state Keap1 levels or in the levels of the related transcription factor, Nrf1, with or without arsenite treatment.
To test whether the relatively small increase in Nrf2 levels in dwarf cells might affect ARE-responsive genes, we used quantitative RT-PCR to measure mRNA levels of six ARE genes, as well as of Nrf2 itself. The results (Fig. (Fig.5)5) show that Nrf2 transcription levels do not differ between dwarf and controls cells, with or without exposure to 5 μM arsenite. In contrast, there are robust and statistically significant differences between dwarf and control cells in expression of each of the ARE-related genes tested. Four of the genes show approximately 2-fold increased expression in dwarf cells prior to arsenite exposure. These include glutamate cysteine ligase modifier subunit 1 (GCLM1), an important rate-limiting enzyme of glutathione synthesis (12, 47, 77); heme-oxygenase 1 (36, 83); metallothionein 1, an important enzyme for metal-binding (88, 110); and thioredoxin reductase, an important enzyme for ascorbate recycling and protection from oxidative stress (62, 64, 73). Two others, NAD(P)H quinone oxidoreductase 1 (NQO1) and glutathione S-transferase A1 (GSTA1), showed no genotype differences in the control condition but were higher in dwarf cells after arsenite treatment. All six genes respond vigorously to arsenite treatment, as expected. The mRNA levels in the arsenite-treated cells are significantly higher in dwarf cells than in control cells for five of the six genes (the exception is GSTA1, where P < 0.1). Thus, many Nrf2-responsive genes are expressed at elevated levels in dwarf-derived cells prior to and/or after exposure to an ARE activator.
Because ARE induction is known to cause increases in reduced glutathione (49, 98), we also measured levels of total reduced thiols (as a surrogate for reduced glutathione) (Fig. (Fig.6A)6A) as well as total glutathione content (Fig. (Fig.6B)6B) in dwarf and normal cells under various conditions. Under baseline conditions, dwarf cells had significantly higher levels of both reduced and total glutathione than control cells. As expected, arsenite (5 μM for 24 h) produced a significant increase in both reduced and total glutathione; thus, both measures remained higher in dwarf than in normal cells after arsenite exposure. Because arsenite is known to damage glutathione sulfhydryl groups, we also tested whether dwarf cells were resistant to the toxicity of arsenite. The results (Fig. 6C and D) show that dwarf cells have a robust, 4-fold higher resistance to the toxic effects of arsenite, perhaps due to their increased glutathione and other ARE-related stress defenses. These increases in glutathione levels and arsenite resistance suggest that ARE upregulation could function to protect dwarf cells from some toxic stresses.
The hypothesis that elevated Nrf2 and ARE function contribute to the stress resistance of cells from dwarf mice implies that induction of ARE responses in normal cells should increase their resistance to lethal injury. We found (Fig. (Fig.7)7) that exposure of normal cells to arsenite does indeed increase their resistance to multiple cellular stresses, including oxidative stressors peroxide, paraquat, and cadmium, as well as to UV irradiation. Arsenite treatment of cells from dwarf mice also produced increases in resistance to multiple forms of lethal stress, but the effect was less dramatic than the response of controls cells shown in Fig. Fig.7;7; responses of arsenite-treated dwarf cells were thus similar to those of arsenite-treated control cells, consistent with the hypothesis that higher Nrf2 signaling may contribute to the resistance of dwarf fibroblasts to multiple forms of lethal stress.
To determine whether increased Nrf2 activity in dwarf-derived fibroblasts is a phenomenon seen only in cell culture, we evaluated levels of ARE transcription in tissues harvested from 4- to 6-month-old dwarf and control mice. The results (Fig. (Fig.8)8) show that multiple tissues of Snell dwarf mice show increased levels of ARE transcripts compared to tissues from littermate controls. Some of the transcripts (e.g., MT1) were increased in heart, liver, and brain, while others (heme oxygenase [HMOX], thioredoxin reductase [TXNRD], and GCLM) were upregulated in some tissues and not in others. These data suggest that the augmented Nrf2 activity characteristic of cultured fibroblasts of Snell dwarf mice is also present in at least some cells and tissues of intact dwarf mice although it is clear that the pattern of ARE gene expression may differ from tissue to tissue in mice.
Our previous work (55, 71, 93) has shown that cultured primary fibroblasts from long-lived rodent species and fibroblasts from long-lived mutant mice are resistant to apoptosis induced by a wide range of lethal stresses and resistant to the inhibition of the PMRS by rotenone or by low-glucose medium. Here, we advance and provide initial data to support the hypothesis that these properties of cells cultured from Snell dwarf mice are due to increased activity of the ARE-specific transcription factor Nrf2 and that this increased activity extends to dwarf tissues. We have shown that Nrf2 protein levels are higher in dwarf cells (Fig. (Fig.4);4); that plasma membranes of dwarf cells have higher activity of the two PMRS activities (Fig. (Fig.2)2) and that these membranes are resistant to stress-induced lipid peroxidation; that incubation in rotenone or in low glucose medium, each of which inhibits PMRS function, also decreases the resistance of fibroblasts to lethal effects of H2O2; that four Nrf2-sensitive genes are at higher levels in dwarf than in control cells and that five Nrf2-sensitive genes are higher in dwarf than in control cells after exposure to the Nrf2 inducer arsenite; that levels of reduced and total glutathione are higher in dwarf cells prior to and after arsenite treatment; that arsenite increases resistance of control cells to paraquat, hydrogen peroxide, cadmium, and UV light, rendering these cells as stress resistant as untreated cells from dwarf mice; and that the increase in ARE activation observed in cell culture extends to tissues harvested from whole mice. Although no one of these findings provides definitive evidence that the unusual properties of Snell dwarf fibroblasts are caused by higher Nrf2 levels, they are all fully consistent with this idea and provide a plausible link to connect alterations in resistance to lethal stresses to the resistance of dwarf cells (and, potentially, cells and tissues of long-lived species) to inhibition of PMRS activity.
Increased Nrf2 transcriptional activity has previously been suggested to contribute to the effects of caloric restriction (CR), another model for increased longevity in rodents, by showing increased ARE transcription in the livers of calorie-restricted mice (80). This report also used an Nrf2 knockout strain to suggest that Nrf2 was necessary for the anticarcinogenic effects of CR. There was no evidence that Nrf2 was required for increased life span in response to CR although interpretation was complicated by the short life span of the Nrf2 knockout mice on the control (ad libitum) diet. It is possible that the increase in life span of the Nrf2 knockout under CR conditions was due to decreased production of reactive oxygen species (66, 106), which would not be affected by Nrf2. Because no Nrf2-positive controls were included in the study, it is unknown whether the effects of CR simply restore Nrf2 knockouts to a normal life span or increase life span similarly to a normal, Nrf2-positive mouse on CR.
Information about the possible role of Nrf2 in controlling aging and life span in intact organisms is still quite limited but provocative. In C. elegans, the Nrf2 homologue SKN-1 has been found to be required for diet restriction-induced longevity (6). SKN-1 was also reported to be necessary for life extension by disruption of insulin-like signaling pathways, and overexpression of SKN-1 led to an increase in life span (104). In flies, heterozygous deletion of Keap1, which increases Nrf2 activity, increases oxidative stress resistance and life span in males (101). In mice, data suggest that Nrf2 knockouts are phenotypically normal in terms of in growth and development (11) but highly susceptible to a variety of oxidative stresses, including acetaminophen toxicity, hyperoxic lung injury, pulmonary fibrosis, glucose-induced oxidative cardiomyocyte damage, colitis-associated colorectal cancer, and many others (reviewed in reference 47). Based on the study mentioned above (80) and another report (53), Nrf2 knockout mice appear relatively short-lived due to development of immune-mediated hemolytic anemia, but large formal studies have yet to be published. It would be of interest to evaluate the life span of mice engineered to overexpress Nrf2. Knockout mice with diminished expression of Keap1 do show increased Nrf2 activity, but these mice die at an early age due to hyperkeratinosis of the esophagus and forestomach (107), and thus evaluation of life span in mice with elevated Nrf2 function may require development of systems for conditional overexpression of Nrf2 or conditional deletion of Keap1.
Our new data suggest a mechanism for the relative resistance of cells from dwarf mice to the inhibitory effects of glucose on PMRS: dwarf cells have both higher levels of PMRS function in their plasma membranes and also higher levels of electron-donating reduced thiols (mostly reduced glutathione) in both standard and low-glucose media (Fig. (Fig.6).6). Our work also shows that Nrf2 activation acts to increase WST-1 reduction, suggesting that an increase in electron donation from glutathione and parallel increase in PMRS expression may be sufficient to account for the increased resistance of PMRS to glucose shortage. We hypothesize that under low-glucose conditions, the increased level of available reduced glutathione makes up for the lack of NADH, preserving the ability of dwarf-derived fibroblasts to reduce exogenous dyes like WST-1. Our data do not offer a clear explanation for the resistance of dwarf cells to rotenone, but we hypothesize that rotenone may affect the PMRS directly to prevent WST-1 reduction and that differential expression of PMRS enzymes partly compensates for this inhibition in dwarf cells. This hypothesis is supported by previous work showing that rotenone binds to NADH oxidoreductases (22, 79, 95) other than those associated with complex I of the mitochondrial electron transport chain and by our own data (not shown) suggesting rotenone can affect PMRS activity in isolated plasma membranes.
Variations in activity of the PMRS could also influence the pace and consequences of the aging process. Previous work has shown that PMRS antioxidants and enzymatic activity decrease with age and that these effects are prevented by the effects of caloric restriction (34, 35). The decrease in membrane antioxidant levels and changes in redox defenses could leave cells from older animals more susceptible to plasma membrane damage through lipid peroxidation. Lipid peroxidation is thought to be important in several age-related diseases, such as Parkinson's disease, Alzheimer's disease, and thyroid disease (21, 70, 99). The work described here suggests that coenzyme Q reducing activity and ascorbate free-radical reducing activity, both elements of the PMRS, are each increased in dwarf membranes. In contrast, another PMRS enzyme, NQO1, was not increased in these cells, suggesting that augmented PMRS function in dwarf-derived cells likely reflects an increase in expression of cytochrome b5 reductase or another, less studied enzyme. It is not clear if cytochrome b5 reductase is activated by Nrf2 signaling although the promoter region of the murine cytochrome b5 reductase 2 gene contains at least one putative ARE site. It is also possible that increased PMRS function in dwarf membranes could be due to increases in other PMRS enzymes, including elements not yet described. The increase in PMRS function, together with the increase in glutathione and glutathione peroxidase (78), provide a plausible explanation for the protection of dwarf fibroblasts from lipid peroxidation. Inferences about the possible effects of increased PMRS function on redox homeostasis and cell growth and signaling (reviewed in reference 94) are at this point purely speculative.
Augmented Nrf2 signals, which could lead to increased levels of reduced and total glutathione, may also account for the resistance of dwarf-derived fibroblasts to lethal oxidative damage. Previous work has shown that there is a shift toward oxidation in glutathione redox state in the brain, liver, heart, kidney, eye, and testis of aged mice (87), perhaps leading to a more oxidative environment and increased susceptibility to oxidative damage in tissues from aging animals (2). Glutathione is the most prevalent of the redox molecules in the cell, and an increase of 25% as seen in dwarf cells could prevent the age-related shift toward oxidation in the cellular redox state, thus lowering cumulative levels of oxidative damage (87). Furthermore, increased glutathione synthesis extends life span in Drosophila, showing that glutathione may have a direct role in the aging process (77). Not only does increased glutathione provide the ability to increase activity of key antioxidants such as glutathione peroxidase, but also the change in cellular redox status creates a more reducing intracellular environment, helping to buffer oxidative stress.
We have previously noted in a study of genetically heterogeneous mice (55) that the animals whose cells are resistant to the effects of glucose levels of PMRS function are also relatively resistant to the lethal effects of peroxide and cadmium. Our new data suggest a mechanism for this correlation: increased expression of ARE-related genes and the PMRS. Augmented transcription of ARE-regulated genes, approximating a 2-fold change in most of the genes evaluated (Fig. (Fig.7),7), may lead to increased expression of glutathione and induction of phase II detoxification genes and thus improve resistance of dwarf cells to lethal stresses. The ~25% increase in cellular glutathione may also help to protect these cells from oxidative stress and is consistent with previous work showing that ARE-related genes glutamate cysteine ligase (GCL), GST, and their end product, GSH, were markedly increased in the livers of 3-, 12-, and 24-month-old Ames dwarf mice (8).
Our data suggest that increased Nrf2 expression in cells from dwarf mice enhances their resistance to cytotoxic and metabolic stresses. In this context, it will be valuable to learn more about how the unusual endocrine environment of the dwarf mouse alters Nrf2 expression and how these cellular properties are maintained by epigenetic mechanisms through several passages in culture. We will also need to determine how increased Nrf2 levels in dwarf cells selectively activate subsets of ARE genes. In resting fibroblasts, i.e., prior to arsenite treatment, levels of four ARE genes (HO-1, GCLM1, MT1, and TXRD) were significantly elevated, but two others (NQO1 and GSTA1) were not. These differential effects could involve phosphorylation, acetylation, or other modifications that may affect Nrf2 binding to specific subsets of ARE sites. Current work on Nrf2 has focused largely on factors that regulate its stabilization and nuclear transport, but much less is known about selective activation of the 200 or more ARE-regulated genes.
We have also shown that differences in ARE transcription extend to cells and tissues harvested directly from young adult dwarf mice. These results suggest that the well-described stress resistance phenotype observed in cells isolated from dwarf mice may also affect cells and tissues in the intact mouse, a hypothesis we plan to evaluate in further studies. There are also clear differences in tissue-specific patterns of expression of ARE-dependent genes, which may reflect combinations of cellular differentiation, hormonal influences, and prior stress exposure. It will be of interest to investigate the basis for the tissue specificity of Nrf2-dependent gene activation.
Further analysis of expression of Nrf2-sensitive genes in tissues of Snell dwarf mice and mice of other long-lived stocks may provide insights into the paths by which altered production of pituitary hormones renders these mice resistant to multiple late-life diseases and degenerative changes. Previous work on Nrf2 has suggested that phosphatidylinositol-3 kinase (PI3K), Akt, protein kinase C (PKC), and mitogen-activated protein kinase (MAPK) may be directly or indirectly important for Nrf2 activation (15, 31, 32, 45, 46, 108). PI3K and Akt are known to be directly downstream of growth factor and insulin-like signaling, and changes in the expression or activity of either kinase due to lack of signaling during development could result in wide-ranging changes including increased Nrf2 activity. Figure Figure99 summarizes this model in which altered hormonal signals, modulating Nrf2/Keap1 interaction and/or Nrf2 binding to ARE sites, might alter multiple aspects of cellular defense, including improved DNA repair (92, 93), resistance to lethal stresses, maintenance of reduced thiol levels, and augmentation of PMRS function. Documentation and clarification of these postulated steps may shed light on the cellular basis of disease resistance and longevity in pituitary dwarf mice.
We thank Maggie Lauderdale and Lisa Burmeister for technical assistance. We thank Ruth Senter and Joel Weinberg for training and technical assistance with HPLC.
This work was supported by NIH grants DK334275 T32-AG000114, AG-023122, and AG025164.
Published ahead of print on 23 November 2009.