Oxidative Stress–Induced Single-Strand Breaks in Chromosomal Telomeres of Human Retinal Pigment Epithelial Cells In Vitro

Abstract

purpose. To demonstrate that chronic hyperoxia induces single-stranded breaks in chromosomal telomeres as a measure of oxidative DNA damage in cultured RPE cells.

methods. RPE340 cells were cultured in 40% and 20% (control) O₂. DNA damage was assessed by mean terminal restriction fragment (TRF) length, and the S1 nuclease assay was used to determine the frequency of single-strand breaks in telomeric DNA. The degree of oxidative stress in cells was estimated by flow cytometric analysis of reactive oxygen intermediate (ROI)-induced 2′,7′-dichlorodihydrofluorescein diacetate fluorescence and Northern blot analysis of heme oxygenase-1 (HO-1) mRNA induction.

results. The mean TRF length of cells grown in 40% O₂ shortened at a faster rate than those grown in 20% O₂. The S1 nuclease assay showed that the accelerated mean TRF length shortening was due to an increased accumulation of single-stranded breaks in telomeric DNA. The degree of ROI production and HO-1 mRNA induction was greater in cells treated with 40% than 20% O₂, an effect that was also larger in old than young passaged cells.

conclusions. RPE340 cells in vitro grown in chronic hyperoxia exhibited evidence of DNA damage with accelerated telomeric shortening via an increased accumulation of single-strand breaks in telomeric DNA. These changes could provide insight into aging of RPE cells by oxidative DNA damage.

Oxidative stress is an important potential mechanism to explain aging of the retinal pigment epithelium (RPE) because the RPE is exposed to high ambient Po ₂s (70–90 mm Hg), ^{1 2 3} and its high level of oxygen metabolism generates large quantities of reactive oxygen intermediates (ROIs). ⁴ Oxidative stress induces a variety of structural and functional changes to lipids, proteins, and DNA. ^{5 6} It is unclear, however, how these oxidatively damaged target macromolecules contribute to the cell’s behavior during age-related disease. Although oxidative damage to lipids and proteins has been extensively studied in RPE cells, little attention has been paid to the oxidative damage of DNA. Telomeres are one potential region of DNA that warrants attention. Telomeres are specialized structures at the end of chromosomes that contain characteristic repetitive G-rich DNA sequences (TTAGGG) that when damaged or shortened can induce an altered cellular phenotype that could promote aging. ⁷  

Recently, we developed a model of chronic oxidative stress in RPE cells in vitro through exposure to hyperoxia. These conditions led to phenotypic changes in RPE cells, including the altered expression of oxidative stress response genes. ⁸ The purpose of this study was to test the hypothesis that chronic hyperoxia in vitro leads directly to DNA damage in RPE cells. We chose to examine oxidative stress–related, single-stranded breaks occurring in chromosomal telomeres as a sensitive method to detect DNA damage in general. Von Zglinicki et al., ⁹ in a study of oxidative damage to WI38 human lung fibroblasts, demonstrated that single-stranded breaks are selectively retained in telomeres as opposed to microsatellite sequences throughout the genome when exposed to hyperoxia. This retention of single-stranded breaks was convincingly demonstrated to be the result of slow DNA repair in telomeres compared with control DNA sequences. ^{9 10 11 12} In this respect then, telomeres appear to be especially sensitive to oxidative damage in the same fashion as mitochondrial DNA, which also suffers from a lack of adequate DNA repair. ⁴  

Materials and Methods

Cell Culture

Cultures were maintained in a three-gas incubator (NuAire, Plymouth, MN), which maintains the oxygen concentration to within an accuracy of ±0.2%. The gas tension was measured weekly with a Fyrite kit (Bacharach, Pittsburgh, PA) and confirmed to be 20% or 40% oxygen. The Po ₂ in the culture medium was measured with a YSI 5331 oxygen probe (YSI Incorporated, Yellow Springs, OH) and found to be 160 ± 4 mm Hg (n = 3) when in 20% oxygen (control) and 295 ± 3 mm Hg (n = 3) in 40% oxygen. The RPE340 cell line from one globe of a 1-year-old trauma victim was propagated as previously described and maintained in Dulbecco’s modified Eagle’s medium/Nutrient mixture F12 with 15 mM Hepes buffer (DMEM/F12; BioWhittaker, Walkersville, MD) + 10% fetal bovine serum (FBS; UBI Upstate, Lake Placid, NY), 0.348% additional sodium bicarbonate, 2 mM l-glutamine solution (GIBCO, Grand Island, NY) at 20% and 40% oxygen conditioned in 10% CO₂ at 37°C. ¹³ WI38 cells (ATCC, Manassas, VA) were cultured in Eagle’s minimum essential medium with Earle’s balanced salt solution, 2 mM l-glutamine (EMEM), 1 mM sodium pyruvate (GIBCO), and 10% FBS at 20% and 40% oxygen conditioned in 5% CO₂ at 37°C. This culture medium was selected because it is considered the optimum growth medium for WI38 cells (personal communication, ATCC), and it was used for WI38 cells in other oxidative stress–related work. ¹¹ For experiments, cells were grown in 75-cm² flasks at an initial seeding density of 10,000/cm² unless stated. Cells were passaged before reaching confluence to avoid contact inhibition. At each passage, cell number was counted using a Coulter counter Z1 (Coulter, Miami, FL), and population doubling level (PDL) was determined as current PDL = last PDL + log₂(collected cell number/seeded cell number). 

Southern Blot Analysis of Mean Terminal Restriction Fragment Length

Genomic DNA was extracted using the Qiagen blood and cell culture DNA midi kit (Qiagen, Santa Clarita, CA). DNA samples at known PDL were limit-digested using HinfI and RsaI (Boehringer Mannheim, Indianapolis, IN) and electrophoresed on a 1% agarose gel for 500 to 600 V-h. Southern blotting was performed using an alkaline phosphatase–labeled telomeric specific probe (TTAGGG)₄ according to manufacturer’s instructions (GIBCO) and exposed to autoradiographic film as previously described. ⁷ Autoradiograms were scanned by densitometry, and the mean terminal restriction fragment (TRF) length was calculated according to L = ∑ OD_i/∑ (OD_i/L _i ), where OD_i is the densitometer output (arbitrary units), and L _i is the length of the DNA at position i.  

S1 Nuclease Assay

The accumulation of single-strand breaks in DNA results in a progressive shift of the hybridization signal to a shorter molecule size with increasing S1 nuclease concentration. After limited digestion with HinfI and RsaI, DNA was incubated with S1 nuclease (0.01 to 2 U/μg DNA; Boehringer Mannheim) for 30 minutes at 37°C. The reaction was terminated with 25 mM EDTA. Samples were electrophoresed and Southern blotted as described above. Resultant autoradiograms were scanned to quantify the profiles of each disperse band, and S1 break content was determined as previously described. ¹² Briefly, the frequency of S1 nuclease sensitive site (n _x ) per Mbp DNA detected by using an S1 nuclease concentration of x U/μg DNA was calculated by n _x = (L ₀ − L _x )/L ₀ L _x , where L ₀ is the mean fragment length of DNA undigested with S1 nuclease, and L _x is the mean fragment length at a nuclease concentration of x U of S1 nuclease. The degree of mean TRF length shortening depends on the number of single-stranded breaks, which have been correlated to the sites that are digested by S1 nuclease. In preliminary experiments, a dose-dependent decrease in mean TRF length due to single-stranded breaks was seen until 0.1 to 0.2 U/μg DNA S1 nuclease, and concentrations >0.2 U/μg DNA induced a potent decrease in TRF length due to nonspecific double-stranded DNA breaks. Thus, 0.1 U/μg DNA S1 nuclease was used to assess the number of S1 sites. 

RNA Extraction and Northern Blot Analysis

Total RNA was extracted using TRIZOL reagent (GIBCO) according to the manufacture’s recommendations. Fifteen-microgram aliquots of each sample were electrophoresed in 1% agarose gels, transferred to nylon membranes, and prehybridized for 3 hours at 42°C in 50% formamide, 5× Denhardt’s solution, 5× SSC, 100 μg/ml salmon sperm DNA, and 0.1% SDS. The membranes were hybridized for 18 hours at 42°C with 25 ng of the ³²P-labeled cDNA probe of heme oxygenase-1 (HO-1), a generous gift from Augustine Choi (Yale University, New Haven, CT). ¹⁴ Blots were washed in 0.1% SDS/0.1× SSC three times at room temperature and then once at 50°C before being subjected to phosphorimage analysis (Molecular Dynamics, Sunnyvale, CA). The blots were stripped and hybridized with a 28S rRNA cDNA probe. ¹⁵ Hybridization signals were normalized against 28S rRNA. 

Flow Cytometry

Cells were grown in DMEM/F12 + 10% FBS (RPE340 cells) and EMEM+ 10% FBS (WI38 cells) for 3 days in control conditions (20% oxygen). At 80% to 90% confluence, half of the flasks were transferred to hyperoxia (40% oxygen) for 24 hours. Cells were rinsed twice with Hanks’ balanced salt solution (HBSS) and treated with 10 μM of 2′,7′-dichlorodihydrofluorescein diacetate (H2DCFDA; Molecular Probes, Eugene, OR) for 30 minutes. Cells were then trypsinized and resuspended in 1% paraformaldehyde/HBSS, and fluorescence was measured with a FACScan (Becton Dickinson, San Jose, CA). ¹⁶  

Statistical Analysis

Statistical significance of data was determined using the two-tailed Student’s t-test. P < 0.05 was considered significant. 

Results

Mean TRF Length of RPE340 Cells and WI38 Fibroblasts Grown in 20% and 40% O₂

Figure 1A is a representative Southern blot used to calculate the mean TRF length for RPE340 and WI38 cells, whereas Figure 1B graphically depicts the decrease in mean TRF with increasing PDL obtained from the Southern blot analysis of RPE340 cells in Figure 1A . Figure 2 summarizes three independent experiments showing the decrease in mean TRF length after hyperoxia treatment. The mean TRF length of control treated cells (20% O₂) decreased gradually to a minimum length of 4.9 kbp in RPE340 cells, whereas WI38 cells showed shortening of the mean TRF length to 5.2 kbp. The mean TRF length shortening by RPE340 cells (41 bp per PDL) was accelerated by hyperoxia, an effect that was greater in cells with an old (128 bp/PDL) than a young starting PDL (54 bp/PDL). The mean TRF shortening in WI38 cells grown in control (20% O₂) was 112 bp per PDL, which was accelerated by hyperoxia treatment. This acceleration was also greater in old (255 bp/PDL) than young (139 bp/PDL) WI38 cells. 

S1 Nuclease–Sensitive Site (S1S) in the Mean TRF of Cells Grown in 20% and 40% O₂

S1 nuclease digestion was performed to assess the amount of single-stranded breaks in telomeric DNA induced by oxidative stress. In control 20% O₂, the number of S1 nuclease sensitive sites (S1S) in the mean TRF length increased gradually with PDL up to 2.6-fold of the starting PDL in RPE340 (Figs. 3A and 4A) and 9.9-fold in WI38 cells (Figs. 3B and 4B) . In 40% O₂, S1S increased rapidly with PDL up to 6.8-fold in RPE340 cells and up to 27.0-fold in WI38 cells compared with the initial S1S frequency of cells grown in control 20% O₂. The S1S accumulation in RPE340 cells after hyperoxia was dependent on the starting PDL of the cell, with an average increase in S1S of 2.7, 8.5, and 77.9/Mbp DNA/PDL for starting PDL 36, 43, and 49, respectively. 

ROI Production and HO-1 mRNA Expression by Hyperoxia in Young and Old PDL Cells

The accelerated reduction in mean TRF shortening with hyperoxia by cells with older starting PDL prompted us to measure the degree of oxidative stress on cells with several different starting PDLs grown in hyperoxia. The development of ROI production in RPE340 cells grown in 40% O₂ was assessed by the development of fluorescence from the reaction of 2′,7′-dichlorodihydrofluorescein diacetate with H₂O₂. After 24 hours of incubation, hyperoxia increased the fluorescence 50% in young PDL RPE340 cells compared with young cells grown in the 20% O₂ control (P < 0.0005). The fluorescence in old PDL RPE340 cells treated with hyperoxia increased 40% over old PDL cells in the 20% O₂ control (P < 0.001). Hyperoxia also increased the fluorescence of young WI38 cells 80% over young cells grown in 20% O₂ (P < 0.005) and increased 30% in old WI38 cells compared with old PDL cells grown in 20% O₂ (P < 0.01). The replicative age of the cell was a factor in determining the degree of fluorescence. In 20% O₂, the fluorescence in old PDL RPE340 cells was 230% higher than young cells (P < 0.0001; Fig. 5A ). Likewise, old PDL WI38 cells showed a 50% increase in fluorescence compared with young cells (P < 0.0005) grown in the control 20% O₂ (Fig. 5B) . 

The induction of HO-1 mRNA was used as a second marker of oxidative stress induced by hyperoxia treatment. ^{17 18} Young RPE340 cells grown in 40% O₂ had a 40% increased steady state HO-1 mRNA expression over young RPE340 cells grown in 20% O₂ (P < 0.05). Similarly, HO-1 mRNA in old PDL RPE 340 cells was upregulated 30% in 40% O₂ compared with old PDL RPE340 cells grown in the control 20% O₂ (P < 0.05). The expression of HO-1 mRNA was approximately 200% higher in old than young PDL RPE340 cells grown in 20% O₂ (P < 0.01, Figs. 6A 6C ). In WI38 cells, HO-1 mRNA increased 70% in young cells with hyperoxia treatment (P < 0.05), but did not change in old cells (P = 0.49). In 20% O₂, the expression of HO-1 mRNA by old WI38 cells was increased 210% over young cells (P < 0.05, Figs. 6B 6D ). 

Discussion

These data support our hypothesis that chronic hyperoxia induces oxidative stress–related, single-stranded breaks in telomeric DNA in RPE cells in vitro. We demonstrate also that cells with longer replicative life spans, which were more susceptible to oxidative stress, showed increased levels of single-stranded telomeric breaks. This observation suggests, but does not prove, that the accumulation of DNA damage in older cells was a result of oxidative stress. 

Von Zglinicki et al. ^{9 10 11} in a series of publications showed that hyperoxia dramatically shortened the mean TRF length in human WI38 fibroblasts by inducing single-stranded breaks. This effect depended on culture age, duration of hyperoxia, and slow repair of single-stranded breaks in telomeric DNA compared with other regions of DNA. Chen et al. ¹⁹ showed that oxygen generated 8-oxoguanine in human fibroblasts in vitro. The ability of hyperoxia to damage DNA may be related to site-specific Fenton reactions or the production of OH radicals from H₂O₂ catalyzed by DNA-bound Fe²⁺. Henle et al. ²⁰ recently identified the G-rich DNA sequences of telomeric repeats as especially sensitive sites for Fe²⁺/H₂O₂-mediated DNA oxidation. Although this study demonstrated the vulnerability in a human telomere insert, these results should be generalizable to any G-rich DNA fragment. 

Compared with RPE340 cells, the increase in single-stranded breaks in WI38 cells was dramatically accelerated after hyperoxia treatment. We chose 40% oxygen incubation because in our previous work, this concentration provided the highest oxygen exposure without inducing cellular toxicity to RPE340 or WI38 cells. ⁸ Presumably, the difference in single-stranded telomeric breaks between the two cell types is due to different antioxidant capabilities, which predispose WI38 cells to oxidative damage. ^{21 22} The basal ROI level was higher in WI38 than RPE340 cells (data not shown), which suggests a relatively decreased antioxidant capacity in WI38 cells. It is possible that small increases in ROI production from hyperoxic treatment surpassed a critical level of ROI that promoted rapid single-stranded break formation in telomeric DNA of WI38 cells. Alternatively, the differences in medium composition for the two cell lines could account for the different rates of telomeric DNA damage from hyperoxia. In preliminary experiments, however, we determined that RPE340 cells had no significant change in ROI production when grown in DMEM/F12- or EMEM-based formulations. 

A contributing factor to oxidative aging concerns the cell’s antioxidant defense mechanisms. The RPE is well documented to contain a significant antioxidant system. With aging, the RPE’s defense system appears to be reduced. For example, Liles et al. ²¹ found an age-related decrease in RPE catalase activity, whereas Frank et al. ²³ observed an age-related decline in HO-1 immunoreactivity in the RPE with aging and in age-related macular degeneration specimens. Furthermore, Samiec et al. ²⁴ noted that reduced glutathione content with age plays an essential role in susceptibility to oxidative damage to the RPE. Thus, it is possible that aged RPE cells in vivo have a decreased antioxidant defense that would promote accumulation of oxidative damage to DNA. 

Our data highlight how oxidative stress can induce significant DNA damage. Although telomeres appear to be particularly sensitive to oxidative damage and served as a useful marker for our work, the important targets for oxidative damage are not yet known. For example, Barreau et al. ²⁵ showed that mitochondrial DNA 4977 deletion increased significantly with aging in the RPE in vivo and speculated that oxidative stress was a likely etiology. Therefore, a systematic examination of DNA damage and their repair mechanisms from aging eyes for a variety of DNA damage like single-stranded breaks or mutational deletions would likely provide insights into oxidative aging of the RPE. 

 Supported by National Institutes of Health Grants EY00344 (JTH) and EY06473 (LMH), UCD Health System Awards (JTH), Manpower Award (JTH), an unrestricted RPB grant from Research to Prevent Blindness to the Department of Ophthalmology and Nippon Eye Bank Association (SH).

 Submitted for publication February 9, 2001; accepted April 13, 2001.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

The decrease in mean TRF length by hyperoxic treatment. (A) A representative Southern blot using an alkaline phosphatase-labeled (TTAGGG)₄ oligonucleotide probe. Molecular weight markers are indicated on the left. (B) Graph of mean TRF length versus PDL obtained from the Southern blot analysis of RPE340 cells depicted in (A): (□), control (20% O₂); (•), 40% O₂.

Figure 2.

 

Graph summarizing the accelerated mean TRF length decrease by hyperoxia. Alteration of mean TRF length of RPE340 cells (A) and WI38 cells (B) in control (20% O₂; □) or 40% O₂ by PDL. Values are shown as the mean of three independent blots. For RPE340 cells, data are shown for four different starting PDLs in 40% O₂: (•), PDL 36; (▪), PDL 41; (▴), PDL 43; (♦), PDL 45. For WI38 cells, data are shown for two different starting PDLs in 40% O₂: (•), PDL 34; (▴), PDL 38.

Figure 3.

 

Southern blot of TRF after digestion with S1 nuclease. HinfI- and RsaI-digested DNA was subjected to incubation with or without S1 nuclease (0.1 U/μg DNA). RPE340 cells (A) and WI38 cells (B) after treatment with control (20% O₂) and 40% O₂. Molecular weight markers are indicated on the left.

Figure 4.

 

Analysis of S1 nuclease assay of RPE340 cells (A) and WI38 cells (B). The mean TRF length of cells treated or untreated with 0.1 U/μg DNA of S1 nuclease was determined and the frequency of single-stranded DNA breaks was calculated. For RPE340 cells, data are shown for control (20% O₂; □) and three different starting PDLs in 40% O₂: (•), PDL 36; (▪), PDL 43; (▴), PDL 49. For WI38 cells, data are shown for control (20% O₂; □) and 40% O₂ with starting PDL 34 (•). Values in (A) are shown as mean of three independent blots and values in (B) are from a single data set.

Figure 5.

 

Flow cytometry of young and old PDL RPE340 cells (A) and WI38 cells (B). Cells were preincubated for 24 hours in control (20% O₂). After a medium change, cells were treated with 20% or 40% O₂ for 24 hours before adding H2DCFDA dye. (□), young PDL cells; (▪), old PDL cells. Values are the mean ± SEM of three independent experiments.

Figure 6.

 

HO-1 mRNA expression in young and old PDL RPE340 cells (A) and WI38 cells (B). Young (lanes 1 and 2) and old PDL cells (lanes 3 and 4) were preincubated for 24 hours in 20% O₂ after changing the medium, followed by treatment with control 20% O₂ (lanes 1 and 3) and 40% O₂ (lanes 2 and 4) for 24 hours before extracting RNA. Signal intensity was measured by phosphorimager analysis and shown as the mean ± SEM of three independent experiments for RPE340 cells (C) and WI38 cells (D). (□), young PDL cells; (▪), old PDL cells.