Previous studies indicated that renal tubular epithelial cells from some long-lived avian species exhibit robust and/or unique protective mechanisms against oxidative stress relative to murine cells. Here we extend these studies to investigate the response of primary embryonic fibroblast-like cells to oxidative challenge in long- and short-lived avian species (budgerigar, Melopsittacus undulatus, longevity up to 20 years, vs Japanese quail, Coturnix coturnix japonica, longevity up to 5 years) and short- and long-lived mammalian species (house mouse, Mus musculus, longevity up to 4 years vs humans, Homo sapiens, longevity up to 122 years). Under the conditions of our assay, the oxidative-damage resistance phenotype appears to be associated with exceptional longevity in avian species, but not in mammals. Furthermore, the extreme oxidative damage resistance phenotype observed in a long-lived bird requires active gene transcription and translation, suggesting that specific gene products may have evolved in long-lived birds to facilitate resistance to oxidative stress.
Decision Editor: John Faulkner, PhD
OF the two major classes of endothermic animals, birds are substantially longer lived than mammals of similar body size (1). Given that increasing evidence implicates oxidative damage as a by-product of reactive oxygen species (ROS) produced by normal metabolism (2) in the aging process, superior avian longevity seems paradoxical, in that birds exhibit basal metabolic rates at least as high as, and often higher than, those of mammals. In the course of their longer lives, therefore, birds process much more oxygen per cell than do mammals, even long-lived mammals such as humans. For comparison, a house mouse (Mus musculus) produces about 250 kcal of metabolic energy per gram of body tissue in its life, a human approximately 800 kcal, and a bird as much as 4000 kcal (1).
Two reasonable mechanisms could explain the longevity difference between birds and mammals, assuming that oxidative damage is a major determinant of longevity. First, birds might produce fewer ROS per mole of oxygen consumed. In fact, this seems to be the case. Two laboratories independently found a substantially lower production of H2O2 in tissues of pigeons (Columbia livia) than in tissues from similar-sized Norway rats (Rattus norvegicus) (3)(4). Additionally, birds might possess superior defenses against cellular damage by ROS. The two previously mentioned pigeon–rat comparisons reached opposite conclusions regarding whether birds exhibit higher activity in a range of antioxidant enzymes. Ku and Sohal (3) found a general increase in antioxidant enzyme activity, whereas Barja and colleagues (4) as well as Herrero and Barja (5) found a general decrease in pigeons and other long-lived bird species compared with similar-sized mammals. We previously demonstrated, by directly exposing primary cultures of kidney epithelial cells to challenge by ROS, that cells from three small long-lived bird species survived better, and suffered less DNA damage, than did similar mouse cells, suggesting that in addition to producing comparatively fewer ROS, long-lived birds also possessed superior defenses of an unknown nature against oxidative damage (6). In principle, the defenses could be structural and constitutive, inducible (as is the case with stress response genes), or both. In support of the notion that structural elements of the cell are important, Pamplona and colleagues (7) found that liver mitochondrial membranes were more resistant to lipid peroxidation in pigeons than in rats, because of the lower fatty acid unsaturation in those membranes.
Not all birds are exceptionally long lived. Nonflying and weakly flying species are shorter lived than good fliers (1). For instance, the Japanese quail (Coturnix coturnix japonica), a 100-g weak flier, rarely lives beyond 4–5 years even in captivity (8), whereas the stronger-flying but similar-sized common grackle (Quiscalus quiscula) has lived more than 20 years in the wild (9). Given its short life, even in coddled captive conditions, the Japanese quail may be the shortest-lived known bird species.
In this paper, we extend our previous studies (6) to a different cell type, primary embryonic fibroblast-like cells, and we investigate whether the response to oxidative challenge differs in cells from a long-lived bird (budgerigar, Melopsittacus undulatus, longevity up to 20 years) versus a short-lived bird (Japanese quail) versus short- and long-lived mammals (mice, 3+ years; humans, longevity > 100 years). It is possible that the oxidative-damage resistance phenotype is a general property of avian cells and is not associated with exceptional longevity. In addition, using drug inhibition protocols, we asked whether active gene transcription and/or translation is required for the resistance phenotype to be expressed, in order to determine whether defenses against oxidative damage require gene action. To inhibit gene transcription, we used both actinomycin-D and α-amanitin, because they are known to work by different mechanisms and have different inhibition specificities (10). Actinomycin-D intercalates into double-stranded DNA and blocks transcript elongation by RNA polymerase I, II, and III (11), consequently having a variety of side effects. In contrast, α-amanitin selectively and specifically inhibits polymerases II and III (10) and has fewer side effects. Cycloheximide was our translation inhibitor. Cycloheximide inhibits peptidyl transferase on the 60S ribosomal subunit of eukaryotic cells (12).
Materials and Methods
Mouse embryonic fibroblasts were obtained from C57Bl/6J mice bred for the production of embryos in a specific pathogen-free colony at the University of Washington. Fertilized budgerigar eggs were obtained from a commercial breeder as well as from a captive colony maintained at the University of Idaho. Fertilized quail eggs came from M.A. Ottinger's laboratory at the University of Maryland. Human embryonic fibroblasts were isolated by using explants of skin collected from a normal embryo obtained by late first-trimester therapeutic abortions.
Cell Culture Preparation
Fresh embryo tissues from mice (gestational age 10–14 days), and from budgerigar and quail eggs (incubated for 7–14 days) were aseptically dissected by removing heads and liver tissue. Tissues were placed in the high-glucose formulation (4500 mg/l) of Dulbecco–Vogt modified Eagle's medium (DMEM; Gibco BRL, Carlsbad, CA) supplemented with 100 units/ml of penicillin, 100 μg/ml of streptomycin, and 30 mM 4-(2-hydroxyethyl)-1-piperzazineethane sulfonic acid) (HEPES) buffer (pH 7.4). Cultures were initiated from cell suspensions prepared by enzymatic dissociation of these tissue samples. Two to three grams of tissue were used (for both mice and budgerigars, this required pooling of four to six embryos; for quail, pooling of only two to three embryos was required). The tissues were minced to obtain ∼1-mm3 tissue fragments. These were washed two to three times in Ca++- and Mg++-free phosphate-buffered normal saline (PBS; pH 7.1; Gibco BRL) prior to enzymatic dissociation. Type 1 collagenase (Sigma, St. Louis, MO) was used to dissociate tissue fragments into cell suspensions. Minced tissue fragments were resuspended in 50 ml of a 1:1 mixture of Type 1 collagenase, 1 mg/ml in PBS, and DMEM with penicillin, 100 units/ml and streptomycin, 100 μg/ml (without the addition of serum).
Digestion was performed in a sterile 125-ml glass trypsinizing flask (Bellco, Vineland, NJ) with a 2-cm sterile stirring bar on a magnetic stir plate at moderate speed for 1 hour in a 37°C incubator. After the tissue fragments were allowed to settle, the supernatant was collected and transferred into 50-ml tubes containing 5 ml of heat-inactivated (56°C for 30 minutes) fetal bovine serum (FBS; Gibco BRL) to stop the collagenase action, and the tubes were placed on ice. The remaining tissue fragments were digested with another 50 ml of the collagenase DMEM mixture as above for an additional 1 hour at 37°C. At the end of this digestion, there were few if any tissue fragments remaining. The resulting cell suspension was combined with that previously collected and centrifuged at 2,600× g for 5 minutes. The resulting cell pellet was resuspended in 20 ml of DMEM supplemented with 100 units/ml of penicillin, 100 μg/ml of streptomycin, and heat-inactivated 10% FBS. This suspension was split among eight to ten 75-cm2 flasks containing 10 ml of DMEM with penicillin and streptomycin and 10% heat-inactivated FBS. When mammalian and avian species were compared, or for experiments comparing budgerigar and quail cells only, the medium used was Medium 199 (Gibco BRL) supplemented with penicillin (100 units/ml) and streptomycin (100 μg/ml), 1% chicken serum (not heat-inactivated; Sigma), 1% heat-inactivated FBS, and 2% tryptose phosphate broth (Gibco BRL), which improved growth for those cells and did not appear to affect survival studies. Flasks were incubated at 37°C in a humidified atmosphere of 5% CO2 and 95% air. Cells were fed the next day by replacing the medium with fresh medium to remove unattached debris. When the cells were confluent, usually after 2 days, they were used for experiments or they were frozen for use in future experiments.
Human embryonic fibroblasts previously prepared by using skin from late first-trimester therapeutic abortions with explant methodology (13) were grown in DMEM with penicillin and streptomycin and 10% heat-inactivated FBS.
Cultured cells were subjected to oxidative challenge, treated with inhibitors, or both. Oxidative challenge was imposed either by exposing the cells for various lengths of time to 95% oxygen or by adding various doses of hydrogen peroxide for 24 hours to the cultural medium. Hydrogen peroxide generates hydroxyl radicals in the presence of metal ions such as Fe++ or Cu++.
Flasks of cells were rinsed twice with 5 ml of 0.53 mM ethylenediamine tetra-acetic acid (EDTA) in PBS (without Ca++ and Mg++), followed by the addition of 5 ml of 0.05% trypsin in this same EDTA/PBS solution for 5 minutes at 37°C. The resulting cell suspension was mixed with 5 ml of medium containing serum, and it was counted and plated in multiple 25-cm2 flasks at 105 cells per flask. The next day (after cells had attached) they were treated with the agent of interest, or in the case of inhibitor studies, treated with the inhibitor of choice, or left as untreated controls.
The inhibitor studies were designed in two phases. In the first phase, designed to assess the toxicity of the inhibitors to otherwise untreated cells, cultures were treated with the indicated concentration of cycloheximide, actinomycin-D, or α-amanitin for 24 hours, and cell survival relative to untreated controls was quantified. In the second phase, cells were treated with the inhibitors for 24 hours, followed by challenge with oxygen or hydrogen peroxide (H2O2), combined with continuing treatment with the inhibitor. Controls in this case were also oxidatively challenged, but without the inhibitor.
For a challenge with 95% O2, the cells were placed in modular incubator chambers (Billups-Rothenberg, Del Mar, CA), gassed with a mixture of 95% O2 and 5% CO2 for 10 minutes, and sealed and incubated at 37 °C for the period indicated in the Results section; they were regassed every 24 hours for the duration of the exposure. For H2O2 challenge, the flasks were fed with freshly prepared medium containing the indicated concentration of H2O2 plus inhibitor at the indicated concentration for 24 hours. As a way to obtain cell counts, cells were again released by using trypsin as described above and counted by using a hemocytometer.
Each dose-response experiment was analyzed by using a two-factor analysis of variance. Two nominal or categorical variables (species, dose or dose, dose) and one continuous dependent variable (cell survival) were used. The null hypotheses, that the effect on cell survival of doses of oxygen or hydrogen peroxide are the same regardless of species, and that the survival effect of doses of oxygen or hydrogen peroxide are the same regardless of dose of inhibitor, were tested by using the Species × Oxidative Stressor (or inhibitor) Dose × Oxidative Stressor Dose interaction terms. In some cases, post hoc analyses of individual doses were done by using Student t tests. The analyses were performed by using Statview 4.5 software from Abacus Concepts (Berkeley, CA).
Resistance to Oxidative Stress
As suggested by our previous work and confirmed here, cell survival under oxidative stress seems to be a characteristic of long-lived birds, rather than birds generally, as shown by the fact that budgerigar cells displayed significantly better survival than quail cells when challenged with either 95% oxygen or hydrogen peroxide (Fig. 1). Budgerigar cells also survived significantly better than either long-lived (humans) or short-lived (mice) mammals at all exposures (p = .018 for O2; p < .001 for H2O2). Post hoc analyses at each given dose level also showed budgerigar cells to be significantly (p < .001) better able to survive challenge with oxygen or hydrogen peroxide than any of the other species at any dose (Fig. 1), with the exception of human cells as the lowest hydrogen peroxide dose (p = .84).
Inhibitors of gene expression frequently have toxic effects on cells. Because previous reports of greater resistance to chemical stressors have been reported for cells cultured from long-lived compared with short-lived species (14), we examined whether budgerigars and quail differed in their response to the inhibitors themselves (Fig. 2). Consistent with previous comparative results for mammals, cells from long-lived budgerigars survived treatment with cycloheximide and actinomycin-D better than cells from short-lived quail at all doses (p < .001). However, the pattern of response to α-amanitin (Fig. 2) differed markedly. Not only was there generally less toxicity to quail cells overall with α-amanitin treatment than with the other two inhibitors, but also no statistical difference between survival in quail and budgerigar cells at any dose (p = .62) was observed. This is perhaps due to α-amanitin's greater inhibitory specificity.
Inhibitor Impact on Response to Oxidative Challenge
Treatment with the translation-inhibitor cycloheximide reduced budgerigar cell survival significantly in a dose-dependent manner with either of the oxidative challenges ( p = .008 for O2; p < .001 for H2O2), but it had no statistically distinguishable effect on quail cell survival, whether challenged with 95% oxygen (p = .727) or hydrogen peroxide (p = .787; Fig. 3 and Fig. 3). To the extent that a nonsignificant trend is evident in quail cells, inhibition of protein synthesis seems to enhance cell survival, suggesting that protein synthesis may contribute to cell death.
Budgerigar cells treated with either transcription inhibitor actinomycin-D or α-amanitin became more sensitive to oxygen dose treatment (p = .0154 for actinomycin-D, p < .001 for α-amanitin; Fig. 4 and Fig. 4), whereas similarly treated quail cells were unchanged in sensitivity to oxygen for actinomycin-D treated cells (p = .90; Fig. 4). In contrast, quail cells responded to increasing doses of α-amanitin treatment with a significant decrease in sensitivity to oxygen (p = .007; Fig. 4), again suggesting that active gene expression enhances cell death rate of quail cells under oxidative stress.
We have demonstrated that primary embryonic fibroblast-like cells exhibit the same pattern of oxidative-damage resistance that adult renal epithelial cells do (6). That is, cells from a long-lived bird (budgerigar) exhibited much greater damage resistance than those from a short-lived mammal (house mouse). Furthermore, we demonstrated that such resistance is absent in a short-lived bird (Japanese quail). Admittedly, there are many ecological, behavioral, and physiological differences between budgerigars and quail besides longevity. However, given the abundant evidence that accumulating damage caused by ROS contributes to the aging rate, and that our previous work found exceptional resistance to oxidative damage in a different cell type in three distantly related but long-lived bird species, it seems a plausible hypothesis that understanding the nature of these differences in cellular resistance to oxidative challenge will reveal key features of retarded senescence. Furthermore, because exceptional resistance to oxidative damage occurs in two tissue types (embryonic fibroblasts and renal epithelium) that likely have substantially different gene expression profiles, comparing expression profiles between these tissues may aid us in identifying the key genes involved in the resistance.
Somewhat surprisingly, exceptional oxidative-damage resistance was not found in cells from a long-lived mammal, H. sapiens, under most conditions of our assay. When exposed to 95% oxygen, human cells survived no better than mouse or quail cells. The response to hydrogen peroxide was more complex. Human cells survived as well as budgerigar cells at the lowest dose, but at higher doses they survived no better than mouse cells. In contrast, a recent study (14) found that human dermal fibroblasts cells survived better than similar cells from seven shorter-lived mammal species in response to a range of exogenous oxidative stressors, including hydrogen peroxide. One possible explanation for these somewhat confusing findings may have to do with the range of oxidative stress that cells experience in intact animals. Within mammalian tissues, cells probably are exposed to 5% oxygen or less, and comparative cell growth and survival characteristics may be very different under more physiologically relevant conditions. For instance, the difference between mouse and human in number of population doublings each will undergo before replicative senescence is much reduced when cells are grown in 5% oxygen as when they are grown in 20% oxygen (air; (15)). Therefore, 95% oxygen may represent a level of oxidative challenge that overwhelms what might be comparatively effective oxidative-damage resistance at more physiological levels of exposure. Our data on hydrogen peroxide support this interpretation to a certain extent. At the lowest dose (10 μM), human cells survived substantially better than mouse or quail cells and not significantly worse than budgerigar cells. However, at higher doses this comparative relationship disappeared. In the study finding, human cells survived exposure to hydrogen peroxide better than cells from shorter-lived species; the hydrogen peroxide dose was somewhat higher than in this study, but the exposure was only for 2 hours compared with our 24-hour exposure.
Is there some reason to think that tissues in long-lived birds might experience higher tissue levels of oxidative stress than tissues in many mammals? Perhaps. Budgerigars have a resting metabolic rate ∼50% higher than that of the similar-sized house mouse (16)(17). Furthermore, their metabolic rate during cruising flight is approximately tenfold above basal, and, being nomadic, budgerigars might maintain such flight for hours at a time in nature (17). Not surprisingly then, the cardiovascular system of budgerigars also has a far greater oxygen exchange capacity than that of mice (18). Thus it is possible that budgerigar cells within metabolically active tissues are typically bathed in higher oxygen concentrations than those in most mammals. Even if this is the case, it doesn't necessarily follow that budgerigars will be exposed to higher chronic oxidative stress, because their mitochondria produce fewer mitochondrial ROS than mice per mole of oxygen consumed (19).
For whatever ultimate reason, budgerigar cells clearly possess superior defenses against oxidative damage compared with quail, mouse, or human cells over a wide range of conditions. Furthermore, this oxidative-damage resistant phenotype requires both active gene transcription and message translation. Treatment with more than 1 μg/ml of the translation inhibitor, cycloheximide, reduces budgerigar cell survival to essentially that of the other species (Fig. 1 and Fig. 3), whereas it has no similar effect on quail cells. Indeed, quail cells treated with cycloheximide may even survive somewhat better than untreated cells under 95% oxygen. Similarly, inhibition of transcription with either actinomycin-D or α-amanitin also reduces budgerigar, but not quail, cell survival. In sum, these results suggest the possibility that human cells might be modified to resist oxidative stress better than they normally do.
It seems somewhat surprising that active protein synthesis is necessary for any vestige of the exceptional oxidative-damage phenotype to be expressed, particularly when it has been shown that budgerigar cardiac (and perhaps other) cells seem structurally more resistant to lipid peroxidation as a result of less fatty acid unsaturation (20). However, our assay of oxidative-damage resistance only measures relative cell death in cells in stationary culture and may be too crude to detect more subtle differences in cellular response to oxidative challenge.
The cells of long-lived birds clearly possess exceptional resistance to oxidative damage. The identity and nature of the gene action necessary for expression of the exceptional oxidative-damage resistance in budgerigars is currently unknown. One possibility is that cells of long-lived birds avoid or repair oxidative damage to DNA by some mechanism. Another possibility is that cells of long-lived birds are more resistant to DNA damage-induced apoptosis. Some preliminary evidence supports this latter interpretation. Higher levels of the DNA damage by-product 8-oxo-2-deoxyguanosine were found in brain and heart, but not liver, tissue from young budgerigars than in that from young mice and rats (21). The exact mechanism underlying this superior survival of cells from long-lived birds under ROS bombardment is under vigorous investigation in our laboratories.
This research was supported by Grant AG01751 from the National Institute on Aging to G. Martin and an Ellison Medical Foundation grant to S. Austad.