Chromosomal instability in an oxygen-tolerant variant of Chinese hamster ovary cells

The RNA surveillance protein SMG1 activates p53 in response to DNA double-strand breaks but not exogenously oxidized mRNA

DNA damage, stalled replication forks, errors in mRNA splicing, and availability of nutrients activate specific phosphatidylinositiol-3 kinase-like kinases (PIKKs) that in turn phosphorylate downstream targets such as p53 on serine 15. While the PIKK proteins ATM and ATR respond to specific DNA lesions, SMG1 responds to errors in mRNA splicing and when cells are exposed to genotoxic stress. Yet, whether genotoxic stress activates SMG1 through specific types of DNA lesions or RNA damage remains poorly understood. Here, we demonstrate that siRNA oligonucleotides targeting the mRNA surveillance proteins SMG1, Upf1, Upf2, or the PIKK protein ATM attenuated p53 (ser15) phosphorylation in cells damaged by high oxygen (hyperoxia), a model of persistent oxidative stress that damages nucleotides. In contrast, loss of SMG1 or ATM, but not Upf1 or Upf2 reduced p53 (ser15) phosphorylation in response to DNA double strand breaks produced by expression of the endonuclease I-PpoI. To determine whether SMG1-dependent activation of p53 was in response to oxidative mRNA damage, mRNA encoding green fluorescence protein (GFP) transcribed in vitro was oxidized by Fenton chemistry and transfected into cells. Although oxidation of GFP mRNA resulted in dose-dependent fragmentation of the mRNA and reduced expression of GFP, it did not stimulate p53 or the p53-target gene p21. These findings establish SMG1 activates p53 in response to DNA double-strand breaks independent of the RNA surveillance proteins Upf1 or Upf2; however, these proteins can stimulate p53 in response to oxidative stress but not necessarily oxidized RNA.

Effects of ambient oxygen concentration on the proliferation and viability of cultured human corneal epithelial cells

Ambient oxygen (O(2)) affects the metabolism and other functions of corneal epithelial cells. The effects of O(2) concentration on the proliferation and viability of corneal epithelial cells in culture were investigated. Simian virus 40-transformed human corneal epithelial (HCE) cells were maintained at 37 degrees C in a humidified incubator containing 5% CO(2) and 95% air. The cells were subsequently transferred to a multigas incubator and exposed to 5% CO(2) and either 1, 21, or 60% O(2) plus 94, 74, or 35% N(2), respectively. Cell proliferation was evaluated by determination of cell number and measurement of the incorporation of bromodeoxyuridine. Cell lysis was quantified by measurement of the release of lactate dehydrogenase. Apoptosis was evaluated by flow cytometric analysis of cells stained with annexin V and propidium iodide as well as by immunoblot analysis of cleavage of caspase-7. The phosphorylation (activation) of Akt was also detected by immunoblot analysis. Hyperoxia (60% O(2)) inhibited the increase in cell number and the incorporation of bromodeoxyuridine apparent in HCE cells exposed to normoxia (21% O(2)). It also induced the release of lactate dehydrogenase, an increase in the proportion of apoptotic (annexin V(+), propidium iodide(-)) cells, the cleavage of caspase-7, and the phosphorylation of Akt. None of these effects was observed in cells exposed to hypoxia (1% O(2)). The amounts of the cleaved forms of caspase-3, 6, and 9 did not differ among HCE cells cultured under 1, 21, or 60% O(2). These results indicate that hyperoxia inhibited the proliferation of, and induced death by apoptosis in, cultured human corneal epithelial cells. The antiapoptotic protein Akt was also activated in cells exposed to hyperoxia, possibly reflecting a protective response to oxygen toxicity.

Loss of Gadd45a does not modify the pulmonary response to oxidative stress

It is well established that exposure to high levels of oxygen (hyperoxia) injures and kills microvascular endothelial and alveolar type I epithelial cells. In contrast, significant death of airway and type II epithelial cells is not observed at mortality, suggesting that these cell types may express genes that protect against oxidative stress and damage. During a search for genes induced by hyperoxia, we previously reported that airway and alveolar type II epithelial cells uniquely express the growth arrest and DNA damage ( Gadd) 45a gene. Because Gadd45a has been implicated in protection against genotoxic stress, adult Gadd45a (+/+) and Gadd45a (−/−) mice were exposed to hyperoxia to investigate whether it protected epithelial cells against oxidative stress. During hyperoxia, Gadd45a deficiency did not affect loss of airway epithelial expression of Clara cell secretory protein or type II epithelial cell expression of pro-surfactant protein C. Likewise, Gadd45a deficiency did not alter recruitment of inflammatory cells, edema, or overall mortality. Consistent with Gadd45a not affecting the oxidative stress response, p21Cip1/WAF1and heme oxygenase-1 were comparably induced in Gadd45a (+/+) and Gadd45a (−/−) mice. Additionally, Gadd45a deficiency did not affect oxidative DNA damage or apoptosis as assessed by oxidized guanine and terminal deoxyneucleotidyl transferase-mediated dUTP nick-end labeling staining. Overexpression of Gadd45a in human lung adenocarcinoma cells did not affect viability or survival during exposure, whereas it was protective against UV-radiation. We conclude that increased tolerance of airway and type II epithelial cells to hyperoxia is not attributed solely to expression of Gadd45a.

Hyperoxia-induced DNA damage causes decreased DNA methylation in human lung epithelial-like A549 cells

The effect of hyperoxia on levels of DNA damage and global DNA methylation was examined in lung epithelial-like A549 cells. DNA damage was assessed by the single-cell gel electrophoresis (comet assay) and DNA methylation status by the cytosine extension assays. Cells exposed to ionizing radiation (0, 1, 2, 4, or 8 Gy) showed increasing rates of percentage of DNA in the tail and tail length with increasing radiation dose. When cells were exposed to room air (normoxia) for 1 day and 95% O2 (hyperoxia) for 1, 2, 3, 4, and 5 days, data indicated that hyperoxia caused time-dependent increases in levels of (a) single strand breaks, (b) double strand breaks, and (c) 8-oxoguanine. Decreased DNA methylation also was observed at day 5 of hyperoxic exposure, suggesting that hyperoxia-induced DNA damage can influence patterns of DNA methylation in a lung-derived cell line.

In vivo exposure to hyperoxia induces DNA damage in a population of alveolar type II epithelial cells

It is well established that hyperoxia injures and kills alveolar endothelial and type I epithelial cells of the lung. Although type II epithelial cells remain morphologically intact, it remains unclear whether they are also damaged. DNA integrity was investigated in adult mice whose type II cells were identified by their endogenous expression of pro-surfactant protein C or transgenic expression of enhanced green fluorescent protein. In mice exposed to room air, punctate perinuclear 8-oxoguanine staining was detected in approximately 4% of all alveolar cells and in 30% of type II cells. After 48 or 72 h of hyperoxia, 8-oxoguanine was detected in 11% of all alveolar cells and in >60% of type II cells. 8-Oxoguanine colocalized by confocal microscopy with the mitochondrial transmembrane protein cytochrome oxidase subunit 1. Type II cells isolated from hyperoxic lungs exhibited nuclear DNA strand breaks by comet assay even though they were viable and morphologically indistinguishable from cells isolated from lungs exposed to room air. These data reveal that type II cells exposed to in vivo hyperoxia have oxidized and fragmented DNA. Because type II cells are essential for lung remodeling, our findings raise the possibility that they are proficient in DNA repair.

Hyperoxia activates the ATR-Chk1 pathway and phosphorylates p53 at multiple sites

Hyperoxia has been shown to cause DNA damage resulting in growth arrest of cells in p53-dependent, as well as p53-independent, pathways. Although H2O2 and other peroxides have been shown to induce ataxia telangiectasia-mutated (ATM)-dependent p53 phosphorylation in response to DNA damage, the signal transduction mechanisms in response to hyperoxia are currently unknown. Here we demonstrate that hyperoxia phosphorylates the Ser15 residue of p53 independently of ATM. Hyperoxia phosphorylated p53 (Ser15) in DNA-dependent protein kinase null (DNA-PK-/-) cells, indicating that it may not depend on DNA-PK for phosphorylation of p53 (Ser15). We show that Ser37 and Ser392 residues of p53 are also phosphorylated in an ATM-independent manner in hyperoxia. In contrast, H2O2 did not phosphorylate Ser37 in either ATM+/+ or ATM-/- cells. Furthermore, H2O2 failed to phosphorylate Ser15 in ATM-/- cells. Additionally, overexpression of kinase-inactive ATM-and-Rad3-related (ATR) in HEK293T cells diminished Ser15, Ser37, and Ser392 phosphorylation compared with vector-only transfected cells. In contrast, wild-type ATR overexpression did not diminish Ser15, Ser37, or Ser392 phosphorylation. We also show that checkpoint kinase 1 (Chk1) is phosphorylated on Ser345 in response to hyperoxia, which could be inhibited by caffeine or wortmannin, potent inhibitors of phosphoinositide 3-kinase-related kinases. Hyperoxia also phosphorylated Chk1 in ATM+/+ as well as in ATM-/- cells, demonstrating an ATM-independent mechanism in Chk1 phosphorylation. Together, our data suggest that hyperoxia activates the ATR-Chk1 pathway and phosphorylates p53 at multiple sites in an ATM-independent manner, which is different from other forms of oxidative stress such as H2O2 or UV light.

Activation of the G2 cell cycle checkpoint enhances survival of epithelial cells exposed to hyperoxia

Reactive oxygen species produced during hyperoxia damage DNA, inhibit proliferation in G1- through p53-dependent activation of p21(Cip1/WAF1/Sdi1), and kill cells. Because checkpoint activation protects cells from genotoxic stress, we investigated cell proliferation and survival of the murine type II epithelial cell line MLE15 during hyperoxia. These cells were chosen for study because they express Simian large and small-T antigens, which transform cells in part by disrupting the p53-dependent G1 checkpoint. Cell counts, 5-bromo-2'-deoxyuridine labeling, and flow cytometry revealed that hyperoxia slowed cell cycle progression after one replication, resulting in a pronounced G2 arrest by 72 h. Addition of caffeine, which inactivates the G2 checkpoint, diminished the percentage of hyperoxic cells in G2 and increased the percentage in sub-G1 and G1. Abrogation of the G2 checkpoint was associated with enhanced oxygen-induced DNA strand breaks and cell death. Caffeine did not affect DNA integrity or viability of cells exposed to room air. Similarly, caffeine abrogated the G2 checkpoint in hyperoxic A549 epithelial cells and enhanced oxygen-induced toxicity. These data indicate that hyperoxia rapidly inhibits proliferation after one cell cycle and that the G2 checkpoint is critical for limiting DNA damage and cell death.

Protection of human lung cells against hyperoxia using the DNA base excision repair genes hOgg1 and Fpg

Hyperoxia causes pulmonary toxicity in part by injuring alveolar epithelial cells. Previous studies have shown that toxic oxygen-derived species damage DNA and this damage is recognized and repaired by either human enzyme 8-oxoguanine DNA glycosylase (hOgg1) or Escherichia coli enzyme formamidopyrimidine DNA glycosylase (Fpg). To determine whether these DNA repair proteins can reduce O(2)-mediated DNA damage in lung cells, A549 lung epithelial cells were transduced with either hOgg1 or Fpg using a retroviral vector containing enhanced green fluorescent protein. Expression of each gene in the transduced cells was confirmed by fluorescent microscopy, Northern blotting, Western blotting, and an enzymatic oligonucleotide cleavage assay. A549 cells expressing either hOgg1 or Fpg were protected from hyperoxia as evidenced by a decrease in DNA damage and a corresponding increase in cell survival. Further, we determined that overexpression of hOgg1 or Fpg partially mitigated the toxic effects of hydrogen peroxide in lung cells. Our data suggest that increased expression of DNA base excision repair genes might represent a new approach for protecting critical lung cells from the toxic effects of hyperoxia.

DNA damage and cell cycle checkpoints in hyperoxic lung injury: braking to facilitate repair

The beneficial use of supplemental oxygen therapies to increase arterial blood oxygen levels and reduce tissue hypoxia is offset by the knowledge that it injures and kills cells, resulting in increased morbidity and mortality. Although many studies have focused on understanding how hyperoxia kills cells, recent findings reveal that it also inhibits proliferation through activation of cell cycle checkpoints rather than through overt cytotoxicity. Cell cycle checkpoints are thought to be protective because they allow additional time for injured cells to repair damaged DNA and other essential molecules. During recovery in room air, the lung undergoes a burst of proliferation to replace injured and dead cells. Failure to terminate this proliferation has been associated with fibrosis. These observations suggest that growth-suppressive signals, which inhibit proliferation of injured cells and terminate proliferation when tissue repair has been completed, may play an important role in the pulmonary response to hyperoxia. Because DNA replication is coupled with DNA repair, activation of cell cycle checkpoints during hyperoxia may be a mechanism by which cells protect themselves from oxidant genotoxic stress. This review examines the effect of hyperoxia on DNA integrity, pulmonary cell proliferation, and cell cycle checkpoints activated by DNA damage.

Cellular oxygen toxicity. Oxidant injury without apoptosis

All forms of aerobic life are faced with the threat of oxidation from molecular oxygen (O2) and have evolved antioxidant defenses to cope with this potential problem. However, cellular antioxidants can become overwhelmed by oxidative insults, including supraphysiologic concentrations of O2 (hyperoxia). Oxidative cell injury involves the modification of cellular macromolecules by reactive oxygen intermediates (ROI), often leading to cell death. O2 therapy, which is a widely used component of life-saving intensive care, can cause lung injury. It is generally thought that hyperoxia injures cells by virtue of the accumulation of toxic levels of ROI, including H2O2 and the superoxide anion (O2-), which are not adequately scavenged by endogenous antioxidant defenses. These oxidants are cytotoxic and have been shown to kill cells via apoptosis, or programmed cell death. If hyperoxia-induced cell death is a result of increased ROI, then O2 toxicity should kill cells via apoptosis. We studied cultured epithelial cells in 95% O2 and assayed apoptosis using a DNA-binding fluorescent dye, in situ end-labeling of DNA, and electron microscopy. Using all approaches we found that hyperoxia kills cells via necrosis, not apoptosis. In contrast, lethal concentrations of either H2O2 or O2- cause apoptosis. Paradoxically, apoptosis is a prominent event in the lungs of animals injured by breathing 100% O2. These data indicate that O2 toxicity is somewhat distinct from other forms of oxidative injury and suggest that apoptosis in vivo is not a direct effect of O2.

Effects of hyperoxia and caffeine on the expression of fragile site at Xq27.3

To enhance the cytogenetic expression of the fragile X chromosome, we studied the effects of hyperoxia and caffeine on the induction of fragile Xq27.3. A lymphoblastoid cell line (GM 06912) derived from a fragile X male proband was cultured in RPMI 1640 containing 16% dialyzed fetal calf serum. The cells were synchronously subjected to one of 3 different atmospheric oxygen tensions (21%, 21.3 kPa, normoxic; 40%, 40.5 kPa, hyperoxic; or 60%, 60.8 kPa, hyperoxic) during the last 24 hours of the 72 hour culture, immediately after the addition of 2'-deoxy-5-fluorouridine (FUdR) at 25 ng/ml. To study the enhancing effect of caffeine, with or without hyperoxia, a second set of cultures was additionally subjected to caffeine (2.5 mM) during the last 6 hours of the culture. When the fragility of hyperoxic cells (38.1 kPa dissolved oxygen) was compared to that of normoxic control cells (13.3 kPa dissolved oxygen), the difference was significant (P < 0.05). These data suggest that there is a mean increase in the fragile Xq27.3 expressivity as the dissolved oxygen tension increases. Additionally, we observed that caffeine, with or without hyperoxia, significantly (P < 0.05) suppressed the expression of the fragile X site in this lymphoblastoid cell line.

Mechanism of hyperoxia-induced chromosomal breakage in Chinese hamster cells

Exposure of cell cultures to hyperoxia, i.e., an atmosphere containing more than 20% O2, results in various genotoxic effects. The most prominent effect of hyperoxia is its clastogenicity. In this paper, earlier published data, obtained from research devoted to the mechanism of hyperoxia-induced clastogenesis, are reviewed. In addition, new data are presented concerning the hyperoxia-sensitivity of the DNA-repair deficient Chinese hamster cell lines xrs1, irs1, and EM9. None of these ionizing radiation-sensitive mutants showed hypersensitivity to hyperoxia, as measured by chromosomal aberration induction and loss of clonogenic cell survival. From the normal hyperoxia-sensitivity of xrs1, it may be concluded that DNA double strand breaks, of the type that are induced by ionizing radiation, do not play a role in chromosomal aberration formation by hyperoxia. In addition, since xrs1 is hypersensitive to drugs that inhibit topoisomerase II, it seems rather unlikely that exposure to hyperoxia affects topoisomerase II activity. Based on circumstantial evidence we hypothesize that perturbation of poly(ADP-ribose) metabolism may play a critical role in the mechanism of hyperoxia-induced clastogenesis.

Cell culture models for oxidative stress: superoxide and hydrogen peroxide versus normobaric hyperoxia

According to the free radical theory of aging, loss of cellular function during aging is a consequence of accumulating subcellular damage inflicted by activated oxygen species. In cells, the deleterious effects of activated oxygen species may become manifest when the balance between radical formation and destruction (removal) is disturbed creating a situation denoted as 'oxidative stress'. Cell culture systems are especially useful to study the effects of oxidative stress, in terms of both toxicity and cellular adaptive responses. A better understanding of such processes may be pertinent to fully comprehend the cellular aging process. This article reviews three model systems for oxidative stress: extracellular sources of O2-. and H2O2, and normobaric hyperoxia (elevated ambient oxygen). Methodological and practical aspects of these exposure models are discussed, as well as their prominent effects as observed in cultures of Chinese hamster cell lines. Since chronic exposure models are to be preferred, it is argued that normobaric hyperoxia is a particularly relevant oxidative stress model for in vitro cellular aging studies.

Effect of iron chelators on the cytotoxic and genotoxic action of hyperoxia in Chinese hamster ovary cells

The iron chelators o-phenanthroline and desferrioxamine were tested for their ability to protect Chinese hamster ovary cells against the cytotoxic and genotoxic effects of normobaric hyperoxia. Desferrioxamine added at sub-toxic concentrations (up to 2.5 microM) over a period of several days had no protective effect on hyperoxia-induced clonogenic cell killing and growth inhibition. The clastogenic effect of hyperoxia was strongly potentiated by desferrioxamine, while the induction of sister-chromatid exchanges (SCEs) by hyperoxia was unaffected. Similarly, o-phenanthroline (up to 0.25 microM) had no protective effect on hyperoxia-induced cell killing, growth inhibition, and SCE induction, while also this compound potentiated the clastogenic effect of hyperoxia. These results do not support a critical role for cellular iron in the mechanism of toxicity by normobaric hyperoxia in CHO cells. However, the results may still be consistent with a critical involvement of particular iron fraction(s) not susceptible to the chelators used. Furthermore, our results show that concentrations of iron chelators known to protect against short-term (up to 1 h) toxic exposure to oxidative stress become toxic themselves when applied chronically, i.e., in the order of days.

DNA damage by oxygen-derived species. Its mechanism and measurement in mammalian systems

When cells are exposed to oxidative stress, DNA damage frequently occurs. The molecular mechanisms causing this damage may include activation of nucleases and direct reaction of hydroxyl radicals with the DNA. Several oxygen-derived species can attack DNA, producing distinctive patterns of chemical modification. Observation of these patterns and measurement of some of the products formed has been used to determine the role of different oxygen-derived species in DNA cleavage reactions, to assess the extent of oxidative damage to DNA in vivo and to investigate the mechanism of DNA damage by ionizing radiation and chemical carcinogens.

Homologous recombination is elevated in some Werner-like syndromes but not during normal in vitro or in vivo senescence of mammalian cells

Werner syndrome (WS) is a recessive genetic condition associated with markedly reduced replicative lifespans of cells in culture, high chromosomal instability in vivo and in vitro, and premature appearance of many characteristics of normal aging, including an increased incidence of cancer. We have monitored plasmid homologous recombination frequencies in diploid fibroblasts from 6 Werner or Werner-like syndrome patients, following transfection with a plasmid substrate containing 2 overlapping fragments of the TN5 Neor gene. Plasmid DNA recovered from these cells was then assayed for homologous recombination by (a) transformation of recA- bacteria to Ampr (indicating total viable plasmid) or Neor (indicating viable recombinant plasmid), and (b) by limited-cycle polymerase chain reaction (PCR) to co-amplify a recombinant fragment containing the overlap region, and a control region of the same plasmid, without bacterial transformation. Bacterial assay data indicated that recombination rates in 3 of the 6 WS strains were significantly elevated above normal controls; 4 of 6 appeared elevated by PCR assay. The highest-recombination WS strain showed evidence of reduced degradation of transfected plasmid DNA. For this small sample of WS strains, clinical severity of WS was not well correlated with recombination rate as determined by either assay (Pearson r = 0.78, not significant, for PCR assay); elevated recombination may, however, define a subset of WS at greatest risk for cancer and/or atherosclerosis. PCR assay of a hyperoxia-resistant HeLa cell line, displaying substantially increased chromosome breakage, indicated increased recombination between direct-repeat fragments. Nevertheless, elevated recombination in WS strains is unlikely to be secondary to impaired replicative capacity characteristic of WS cells, or to defective repair of chromosome damage which is increased in WS, since recombination in non-WS strains was unaffected by passage level or repeated UV irradiation.

Oxygen toxicity in control and H2O2-resistant Chinese hamster fibroblast cell lines

Following exposure to 95% oxygen, clonogenic cell survival was assayed and qualitative morphologic changes were observed in a Chinese hamster fibroblast cell line (HA-1). The time in 95% O2 necessary to clonogenically inactivate 90% of the cells was inversely related to the cell density of the cultures at the beginning of hyperoxic exposure (from 1 to 6 X 10(4) cells/cm2). The O2-induced loss in clonogenicity and evidence of morphologic injury were shown to be significantly delayed (17-22 h) in an H2O2-resistant variant of the parental HA-1 cell line. After the delay in onset of clonogenic cell killing or morphologic injury, the process of injury proceeded in a similar fashion in both cell lines. The H2O2-resistant cell line demonstrated significantly greater catalase activity (20-fold), CuZn superoxide dismutase activity (2-fold), and Se-dependent glutathione peroxidase activity (1.5-fold). The greater activities of CuZn superoxide dismutase and catalase were accompanied by similarly greater quantities of immunoreactive protein as determined by immunoblotting. These data demonstrate that the cells adapted and/or selected for growth in a highly peroxidative environment also became refractory to O2-induced toxicity, which may be related to increased expression of antioxidant enzymes. However, the magnitude of this cross-resistance to O2 toxicity was less than the magnitude of the cellular resistance to the toxicity of exogenous H2O2, suggesting that in this system the toxicity of 95% oxygen is not identical to H2O2-mediated cytotoxicity.

Effects of lethal exposure to hyperoxia and to hydrogen peroxide on NAD(H) and ATP pools in Chinese hamster ovary cells

Cell death by oxidative stress has been proposed to be based on suicidal NAD depletion, typically followed by ATP depletion, caused by the NAD-consuming enzyme poly(ADP)ribose polymerase, which becomes activated by the presence of excessive DNA-strand breaks. In this study NAD+, NADH and ATP levels as well as DNA-strand breaks (assayed by alkaline elution) were determined in Chinese hamster ovary (CHO) cells treated with either H2O2 or hyperoxia to a level of more than 80% clonogenic cell killing. With H2O2 extensive DNA damage and NAD depletion were observed, while at a higher H2O2 dosage ATP also became depleted. In agreement with results of others, the poly(ADP)ribose polymerase inhibitor 3-aminobenzamide completely prevented NAD depletion. However, both H2O2-induced ATP depletion and cell killing were unaffected by the inhibitor, suggesting that ATP depletion may be a more critical factor than NAD depletion in H2O2-induced killing of CHO cells. With hyperoxia, only moderate DNA damage (2 X background) and no NAD depletion were observed, whereas ATP became largely (70%) depleted. We conclude that (1) there is no direct relation between ATP and NAD depletion in CHO cells subjected to toxic doses of H2O2 or hyperoxia; (2) H2O2-induced NAD depletion is not by itself sufficient to kill CHO cells; (3) killing of CHO cells by hyperoxia is not due to NAD depletion, but may be due to depletion of ATP.

Chromosomal instability and progressive loss of chromosomes in HeLa cells during adaptation to hyperoxic growth conditions

By sequential selection for resistance to stepwise increased levels of atmospheric O2, a genetic variant of HeLa cells was obtained capable of stable proliferation under an atmosphere containing 80% O2 (HeLa-80). This cell strain has previously been characterized in terms of growth characteristics, morphology and antioxidant status (Joenje et al., 1985). In an attempt to find cytogenetic clues possibly related to the O2-tolerant character, metaphases of HeLa-80 cells were analyzed and compared to the parental (HeLa-20) strain. Numerical analysis revealed a progressive decrease in the number of chromosomes per cell during selection for O2 resistance, from a modal number of 112 in HeLa-20 cells to 84 in HeLa-80 cells. Cytogenetic endpoints for genetic damage revealed increased frequencies in HeLa-80 cells of both chromosomal aberrations (29.7 versus 6.9% aberrant cells) and sister-chromatid exchanges (0.46 +/- 0.13 versus 0.31 +/- 0.10 SCE/chromosome). G-banded metaphases failed to reveal cytogenetic evidence of gene amplification (homogeneously staining regions, double minutes) in the karyotype of HeLa-80 cells.