Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2012 Feb 29.
Published in final edited form as: Biomarkers. 2009 Dec;14(8):587–595. doi: 10.3109/13547500903183954

Comparison of Three Oxidative Stress Biomarkers in a Sample of Healthy Adults

Joanne L Watters 1,2, Jessie A Satia 2,3, Kerry-Ann da Costa 2, Gunnar Boysen 4,5, Leonard B Collins 4, Jason D Morrow 6, Ginger L Milne 6, James A Swenberg 2,4,5
PMCID: PMC3290341  NIHMSID: NIHMS133086  PMID: 20001708

Abstract

Oxidative stress is a potentially important etiologic factor for many chronic diseases, including cardiovascular disease, neurodegenerative disease, and cancer, yet studies often find inconsistent results. The associations between three of the most widely-used biomarkers of oxidative stress, i.e., F2-isoprostanes for lipid peroxidation and 8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxo-dG) and the comet assay with FPG for oxidative DNA damage, were compared in a sample of 135 healthy African American and White adults. Modest associations were observed between F2-isoprostanes and the comet assay (r=0.22, p=0.01), but there were no significant correlations between 8-oxo-dG and the comet assay (r=−0.09) or F2-IsoP (r=−0.04). These results are informative for researchers seeking to compare results pertaining to oxidative stress across studies and/or assessment methods in healthy disease-free populations. The development and use of oxidative stress biomarkers is a promising field; however, additional validation studies are necessary to establish accuracy and comparability across oxidative stress biomarkers.

Keywords: Oxidation DNA damage, lipid peroxidation, 8-hydroxy-2-deoxyguanosine, comet assay, isoprostanes

Introduction

Oxidative stress is commonly described as the imbalance that occurs when reactive oxygen species (ROS) or radical-generating agent concentrations exceed the body’s defense mechanisms (Sies 1991). Oxidative stress can be caused by exogenous factors, such as smoking, as well as endogenous processes during normal cell metabolism. Humans have well-developed defense systems that generally maintain homeostasis by detoxification of these oxidative products or by DNA repair; however, defenses may be overwhelmed under conditions of elevated oxidative stress, leading to damage of lipids, proteins, and DNA. Thus, oxidative stress is postulated to be a potentially important factor in the development of many chronic diseases, including cardiovascular disease, neurodegenerative disease, and cancer (Ames 2001; Klaunig and Kamendulis 2004; Cooke, Olinski et al. 2006; Dalle-Donne, Rossi et al. 2006; Tsimikas 2006; Montuschi, Barnes et al. 2007; Musiek, McLaughlin et al. 2007).

Although oxidative stress also affects proteins, this work focuses on oxidative DNA damage and lipid peroxidation because they are the forms of oxidative stress most frequently reported in the literature. The two most widely studied measures of oxidative DNA damage are the comet assay and 8-oxo-7, 8-dihydro-2′-deoxyguanosine (8-oxo-dG) (Gallo, Khan et al. 2008). The comet assay, also called single-cell gel electrophoresis, measures DNA strand breaks and alkali-labile sites within individual cells. Breaks in DNA allow supercoiled loops of DNA to relax and if damaged, appear like a comet with a tail under the conditions of the assay (Collins, Dusinska et al. 1996; Burlinson, Tice et al. 2007). Levels of 8-oxo-dG, or 8-Hydroxy-2′-deoxyguanosine (8-OHdG), which are formed by direct attack of ROS on DNA, can be measured using several methods, including LC/MS, ELISA, and HPLC. Although 8-oxo-dG is an often used method, study results have been inconsistent due to variation in the quantification method used, sample preparation, and DNA isolation, which may increase artifactual oxidation due to traces of transition metals and oxygen (Cadet, Douki et al. 1997; Halliwell 2000). However, recent refinements in methods using UPLC-HESI-MS/MS have greatly improved sensitivity and specificity (Boysen et al, in preparation). F2-IsoPs are prostaglandin-like compounds produced by the peroxidation of arachidonic acid independently of cyclooxygenase (COX) enzymes (Milne and Morrow 2006; Montuschi, Barnes et al. 2007). Quantification of F2-IsoPs in plasma or urine has recently been described as the gold standard for measuring lipid peroxidation (Kadiiska, Gladen et al. 2005).

An increasing number of studies are examining associations of oxidative stress with chronic diseases (Klaunig and Kamendulis 2004; Tsimikas 2006; Montuschi, Barnes et al. 2007; Musiek, McLaughlin et al. 2007) and other disease risk factors, such as dietary exposures (Ames 2001; Collins 2005; Hwang and Bowen 2007; Montuschi, Barnes et al. 2007; Watters, Satia et al. 2007). However, the biomarkers of oxidative stress and the methods used to measure them often vary widely across studies, making it difficult to put results from different studies into context. Recently, the European Standards Committee on Oxidative DNA Damage (ESCODD) compared results measuring oxidative DNA damage levels in healthy males using six different laboratories (n=88) and found no overall association between samples measured by the comet assay and 8-oxo-dG by HPLC methods (ESCODD, Gedik et al. 2005). Considering the validity of many biomarkers remains to be established, the association between oxidative stress and disease in vivo in humans may be obscured due to noted inconsistencies among oxidative stress assays (Halliwell and Grootveld 1987; Dalle-Donne, Rossi et al. 2006; Montuschi, Barnes et al. 2007). Thus, further investigations into the comparability and validity of biomarkers used to assess oxidative stress are warranted.

In this report, we examine associations among three of the most widely-used biomarkers of oxidative stress, F2-isoprostanes (lipid peroxidation) and 8-oxo-dG and the comet assay with FPG (oxidative DNA damage), as well as associations with participant characteristics, in a sample of 135 healthy African American and White adults. Results from this study provide insight into the extent to which there is agreement among the three methods and inform on the appropriateness of comparing results across studies that have employed different biomarkers of oxidative stress.

Materials and Methods

Study population

Data reported here are from the DIet, Supplements, and Health (DISH) Study, a cross-sectional study of dietary antioxidants and oxidative stress in healthy African American and White adults in North Carolina. Details on the DISH study design and methods have been published elsewhere (Watters, Satia et al. 2007). Briefly, participants were recruited between March and December 2005 via flyers displayed in public venues throughout the Research Triangle area in North Carolina, including churches, gyms, campus buildings, and university -wide emails. Eligible persons were 20 to 45 years of age, generally healthy, free of diseases related to oxidative stress (i.e., cancer, diabetes, heart disease, or Alzheimer’s disease), and fluent in written and spoken English. Persons likely to have high levels of oxidative stress, such as current smokers and those with a self-reported body mass index (BMI) of 30 or greater, were ineligible. Other exclusion criteria included anorexia or bulimia nervosa, large weight change (more than 15 pounds) in the past year, inability to fast for 6 h, and pregnancy. Of the 191 respondents deemed eligible during the screening interview, 168 (88.0%) were enrolled and 164 (85.9%) completed all aspects of the study. Data for nine participants were excluded because of serum cotinine levels that were consistent with active smoking (≥15 ng/mL) and all three oxidative stress assays were not available for twenty participants, leaving a total of 135 participants for these analyses (65 African Americans, 70 whites).

Data collection

Participants completed a self-administered 12-page demographic, health, and dietary questionnaire which included 37 questions pertaining to general health, demographic, behavior, and lifestyle characteristics. During a one-time visit to the University of North Carolina (UNC) General Clinical Research Center (GCRC), participants had height, weight, and waist circumference measured, provided semi-fasting (≥6 h) blood and urine samples, participated in a dietary supplement inventory, and answered questions about the use of NSAIDS and lipid-lowering drugs, current occupation, outdoor exposure, and last menstrual cycle (women only). Blood samples that were protected from heat and light were analyzed for oxidative DNA damage, lipid peroxidation, plasma antioxidant nutrient concentrations, cholesterol, hemoglobin A1c (to confirm self-reported absence of diabetes), and serum cotinine (to validate self-reported smoking status). Plasma concentrations of retinol, tocopherols, vitamin C, and carotenoids (lutein, zeaxanthin, α-cryptoxanthin, β-cryptoxanthin, lycopene, α-carotene, β-carotene) were measured using high performance liquid chromatography (HPLC) with multi-wavelength photodiode-array absorbance detection (Craft 1996). Each participant received $100 compensation for his/her time upon completion of all study activities. This study was approved by the Institutional Review Board at UNC and written (signed) informed consent was obtained from all participants.

Measures of oxidative stress

F2-Isoprostanes

F2-IsoPs, also called 8-iso-PGF2α, are prostaglandin-like compounds produced by the peroxidation of arachidonic acid, independently of cyclooxygenase (COX) enzymes, and have been described as the ‘gold standard’ in vivo measure of lipid peroxidation (Milne and Morrow 2006; Montuschi, Barnes et al. 2007). Urinary F2-IsoPs represent a biomarker for systemic or “whole body” oxidative stress over time (Milne, Sanchez et al. 2007) and are very stable because artifactual F2-IsoPs are not formed by auto-oxidation due to the low lipid content of urine (Montuschi, Barnes et al. 2007). Urinary F2-isoprostanes (F2-IsoPs) levels were measured according to previously published methods (Milne, Sanchez et al. 2007) using spot urine samples that had been stored at −80°C, as recommended. Briefly, urinary F2-IsoPs were quantified by gas chromatography with negative ion chemical ionization mass spectrometry using [2H4]-15-F2t-IsoP as an internal standard (Milne, Sanchez et al. 2007). Compounds were analyzed as pentafluorobenzyl ester and trimethylsilyl ether derivatives by measuring the m/z 569 ion for endogenous F2-isoPs and m/z 573 ion for [2H4]-15-F2t-IsoP. All urinary F2-IsoPs levels were normalized to creatinine clearance and expressed as ng/mg creatinine.

8-oxo-7, 8-dihydro-2′-deoxyguanosine (8-oxo-dG)

The quantitative analysis of 8-oxo-dG by UPLC-HESI-MS/MS was performed as described in (Boysen et al., in preparation). In brief, DNA was isolated from tissues using the Puregene® (Qiagen, Valencia, CA) procedure with modifications. Tissues were lysed with Puregene® Cell Lysis Solution® in the presence of 20 mM 2,2,6,6-Tetramethylpiperidine 1-oxyl (TEMPO) (Acros, NJ). Proteins were precipitated by addition of Protein Precipitation Solution® containing 20mM TEMPO. RNA was hydrolyzed with RNase A (St.Louis, MO). DNA was precipitated by addition of isopropanol containing 20 mM TEMPO. Purity and amount of DNA was determined by UV photo spectrometry Abs 260/280 > 0.8. DNA solutions of 0.5 to 1 ug/ul were prepared in Tris-HCl, 20 mM MgCl2, pH 7.0 and 20mM TEMPO. Stable isotope labeled internal standard (500 fmol)[15N5]8-oxo-dG, Cambridge Isotope Laboratories (Andover, MA) was added for accurate quantitation. DNA (50 μg) was digested to single nucleotides with DNase I, Type II, Sigma-Aldrich (St.Louis, MO), Phosphodiesterase I, and Alkaline Phosphatase. Proteins were removed by Centricon YM-10 cellulose centrifuge filters (Millipore, Bedford, MA) according to manufacturer specifications. 8-oxo-dG and [15N5]8-oxo-dG were separated from normal nucleotides by reverse phase HPLC using a Beckman ODS C18 4.6 × 250 mm column, and 7% methanol - 10mM ammonium formate (pH 4.3) in HPLC grade water as the mobile phase. The fraction containing 8-oxo-dG and [15N5]8-oxo-dG was collected into a collection tube containing 300 μl 75mM TEMPO and the solvent was removed in speed vac. Samples were reconstituted in 20 μl water and measured with an UPLC (Waters, Milford, MA) coupled to a TSQ-Quantum Ultra triple-quadrupole mass analyzer (ThermoFinnigan, San Jose, CA) using a heat assisted electrospray ionization (HESI) source. A 2.1 × 100 mm HSS T3 C18, 1.8 μm column (Waters, Milford MA) was operated with a linear gradient of 1 % methanol-0.1% acetic acid in water to 5 % methanol-0.1 % acetic acid and then held for 10 min, followed by increase to 50% methanol-0.1% acetic acid in 2 min, at a flow rate of 200 μl/min. The analyte and internal standard were detected in selected reaction monitoring (SRM) mode, monitoring the transitions of the m/z 284.1 to 168.05 and m/z 289.1 to 173.05 for 8-oxo-dG and [15N5]8-oxo-dG, respectively. The electrospray conditions were as follows: spray voltage of 3000 V, vaporizer (HESI) temperature of 250 °C, sheath gas pressure 30, auxillary gas pressure 35, capillary temperature of 285°C and collision energy of 12 eV.

Comet Assay with FPG

The single cell gel electrophoresis, or comet assay, measures DNA strand breaks and alkali-labile sites at the level of a single cell in which lymphocytes are digested with lesion-specific repair endonucleases (Collins 2005). The comet assay used here was a slightly modified version in which formamidopyrimidine DNA glycosylase (FPG) (kindly provided by Dr. A.R. Collins, Oslo, Norway) was added to convert oxidized purines into strand breaks (Collins 2005; ESCODD, Gedik et al. 2005). FPG (5 μL) was diluted into 0.5 mL reaction buffer containing 40 mM HEPES, 0.1 M KCl, 0.5 mM EDTA, 0.2 mg/ml BSA, pH 8.0, and 10% glycerol and stored as 10 μL aliquots at −80°C until used. Peripheral whole blood lymphocytes were washed in PBS, counted using a hemacytometer, and cryopreserved in 1 ml RPMI-1640 containing 15% BSA and 10% DMSO. All samples were processed within 2 h of collection and stored at −80°C until assays were performed, as recommended. Lymphocytes were sandwiched between 0.5% agarose and 0.5% low-melting-point (37°C) agarose (Fisher, Fair Lawn, NJ). The resulting slides were placed into cold, freshly made lysis solution [10 mM Tris (pH 10), 2.5 M NaCl, 100 mM EDTA, 10% DMSO, and 1% Triton X-100] at 4 °C for 1 h. 50 μL FPG (10 μL aliquot diluted into 0.3 mL reaction buffer with no glycerol) was added to each slide for digestion at 37°C for 30 min. After lysis, slides were washed three times in buffer (40 mM HEPES, 0.1 M KCl, 0.5 mM EDTA, 0.2 mg/ml bovine serum albumin, pH 8.0). 50 ul FPG in this buffer was added to each slide and then treated for 20 min in electrophoresis solution [300 mM NaOH, 1 mM EDTA (pH 13)](ESCODD, Gedik et al. 2005; da Costa, Niculescu et al. 2006). After electrophoresis was performed at 25V and 300mA for 20 min, slides were incubated 3 times for 5 min in neutralization buffer [0.4 M Tris (pH 7.5)], washed with methanol, and stained with ethidium bromide. Multiple levels of quality control samples (e.g., 10% blinded duplicate samples) were included in each batch and all assays were performed by the University of North Carolina Clinical Nutrition Research Unit. Comet tail length (the distance of DNA migration from the body of the nuclear core) was visualized by using a fluorescence microscope (typically, 100 cells/sample) and SCION IMAGE software (da Costa, Niculescu et al. 2006). The comet tail moment (defined as the integrated density in the comet tail multiplied by the distance from the center of the nucleus to the center of mass of the tail) was calculated by using the NIH Image Analysis Macro software (http://dir.nhlbi.nih.gov/labs/ldn/macroanalysis.asp). The comet tail moment reported here was calculated using FPG-sensitive sites only, but was also assessed for strands breaks without the addition of FPG.

Statistical analysis

Data analyses were performed using Stata (version SE 8.2, STATACorp, College Station, TX). Descriptive statistics (means, standard deviations, medians, and percentages) were calculated for all variables. Twenty participants with one or more missing values for the oxidative stress measures were excluded. Log transformations were applied to the three oxidative stress measures: lipid peroxidation (F2-isoprostanes) and oxidative DNA damage (8-oxo-dG and the comet tail moment) to help meet normality distribution assumptions, as each was right-skewed. Furthermore, oxidative stress distributions were normalized by Z score transformation to account for differences in units among the three methods and used in all correlation analyses. Differences between participant characteristics and categories of each oxidative stress measure were evaluated using analysis of variance. The demographic and lifestyle characteristics of participants analyzed included sex, race, age, BMI, education, income, dietary supplement use, physical activity, alcohol consumption, passive smoke exposure, outdoor exposure, self-rated health status, and county of residence. To examine associations among the three oxidative stress measures, Pearson correlations, crude and partialled for relevant covariates, were computed between each pair of methods in the total sample and separately by race and gender. Age, sex, race, body mass index (calculated as weight in kilograms divided by height in meters squared), physical activity, education, days since last menses (for women only), alcohol consumption, and serum cotinine, a metabolite of tobacco exposure have been found to be associated with oxidative stress (Hercberg, Preziosi et al. 1994; Willett 1998; Galan, Viteri et al. 2005) and thus, were evaluated as potential confounders and included in all models as each affected point estimates by at least 10%. All statistical tests were two-sided and p values ≤0.05 were considered statistically significant.

Results

The mean oxidative stress values for each of the three methods, stratified by participant (demographic, lifestyle, and behavioral) characteristics, are given in Table 1. The mean age of participants was 31.6 years (8.2 SD), 50% were female, 48% were African American, and 69% had a college degree. F2-IsoP levels were statistically significantly lower in men those who exercised more than twice a week, and among those with more than 30 h of outdoor exposure per month. As with F2-IsoP, 8-oxo-dG levels were also lower among those with higher levels of physical activity (p=0.02), as well as among males, and Whites. In contrast, mean comet tail moment indicated significantly higher oxidative DNA damage levels in Whites and respondents who did not live with smoker(s), p<0.05. There were no differences among the three oxidative stress measures within categories based on age, BMI, education, income, dietary supplement use, alcohol consumption, self-rated health status, and county of residence.

Table 1.

F2-isoprostanes, 8-oxo-dG, and Comet Tail Moment by Demographic and Behavioral Characteristics (n=135)

Characteristic Distributions
n (%)		 F2-isoprostanes
(SD)		 8-oxo-dG
per 106 dG (SD)		 Mean Comet Tail
Moment (SD)
Sex
 Male 67 (50) 1.34 (0.74) 1.54 (0.85) 1.45 (0.32)
 Female 68 (50) 2.06 (1.42) 1.69 (0.94) 1.47 (0.30)
 p value1 0.0009 0.27 0.66
Race
 African American 65 (48) 1.67 (1.39) 1.74 (0.90) 1.37 (0.28)
 White 70 (52) 1.74 (0.97) 1.50 (0.88) 1.55 (0.32)
 p value 0.27 0.05 0.0003
Age (years)
 20-28 56 (41) 1.51 (0.76) 1.67 (0.93) 1.43 (0.31)
 29-37 38 (28) 1.60 (0.84) 1.45 (0.66) 1.41 (0.21)
 38-45 41 (30) 2.05 (1.75) 1.69 (1.03) 1.57 (0.37)
 p value 0.67 0.46 0.09
BMI (kg/m2)
 Normal (18.5–24.9) 80 (59) 1.78 (1.35) 1.57 (0.88) 1.48 (0.35)
 Overweight (25–29.9) 50 (37) 1.59 (0.92) 1.63 (0.72) 1.42 (0.24)
 Obese (≥30) 5 (4) 1.64 (0.49) 2.16 (2.19) 1.63 (0.22)
 p value 0.92 0.70 0.33
Education
 Some college or less 42 (31) 1.71 (1.22) 1.77 (0.98) 1.44 (0.33)
 College graduate 55 (41) 1.66 (1.03) 1.46 (0.69) 1.49 (0.29)
 Advanced Degree 38 (28) 1.77 (1.37) 1.68 (1.04) 1.45 (0.33)
 p value 0.95 0.17 0.57
Income
 Less than $20,000 25 (20) 1.58 (0.90) 1.76 (0.96) 1.43 (0.31)
 $20,000-39,000 32 (25) 1.77 (1.26) 1.59 (1.00) 1.50 (0.33)
 $40,000-59,000 40 (32) 1.73 (1.33) 1.66 (0.94) 1.47 (0.32)
 $60,000 or more 29 (23) 1.71 (1.27) 1.52 (0.78) 1.46 (0.31)
 p value 0.97 0.80 0.84
Dietary Supplement Use
 Non-Users 79 (59) 1.65 (1.07) 1.72 (0.99) 1.46 (0.31)
 Users 54 (41) 1.78 (1.35) 1.45 (0.71) 1.47 (0.31)
 p value 0.51 0.33 0.19
Physical Activity
 ≤2 times per week 43 (35) 2.18 (1.58) 1.72 (0.76) 1.49 (0.30)
 3-4 times per week 48 (39) 1.38 (0.92) 1.69 (1.02) 1.46 (0.33)
 5+ times per week 32 (26) 1.56 (0.78) 1.30 (0.71) 1.46 (0.30)
 p value 0.01 0.02 0.76
Alcohol Consumption
 Never 8 (6) 1.06 (0.33) 1.62 (0.71) 1.64 (0.37)
 Less than 1 per week 48 (36) 1.66 (0.92) 1.56 (1.00) 1.52 (0.29)
 1-6 times per week 37 (27) 1.57 (1.00) 1.57 (1.00) 1.41 (0.31)
 1 or more per day 42 (31) 1.99 (1.60) 1.71 (0.71) 1.42 (0.30)
 p value 0.22 0.33 0.08
Passive Smoke Exposure
 Lives with a smoker 8 (6) 1.81 (0.93) 1.53 (1.08) 1.23 (0.26)
 No one at home smokes 126 (94) 1.69 (1.21) 1.60 (0.87) 1.48 (0.31)
 p value 0.52 0.74 0.02
Outdoor Exposure
 < 30 h/ month 15 (11) 2.54 (2.12) 1.42 (0.50) 1.41 (0.26)
 30-59 h/ month 41 (30) 1.90 (1.24) 1.65 (0.90) 1.46 (0.29)
 60-89 h/ month 36 (27) 1.41 (0.76) 1.85 (1.20) 1.54 (0.35)
 90+ h/ month 43 (32) 1.47 (0.79) 1.45 (0.63) 1.43 (0.31)
 p value 0.05 0.54 0.46
Self-Rated Health Status
 Excellent 31 (23) 2.13 (1.59) 1.64 (0.59) 1.47 (0.34)
 Very Good 66 (49) 1.48 (0.79) 1.73 (1.11) 1.45 (0.30)
 Good, Fair, or Poor 38 (28) 1.75 (1.31) 1.39 (0.62) 1.49 (0.31)
 p value 0.09 0.20 0.76
County of Residence
 Urban (Orange, Durham, Wake) 115 (85) 1.68 (1.22) 1.61 (0.85) 1.47 (0.31)
 Rural (Franklin, Chatham, Johnston) 13 (10) 1.80 (0.92) 1.72 (1.35) 1.42 (0.29)
 None of the above 7 (5) 1.87 (1.08) 1.54 (0.78) 1.53 (0.30)
 p value 0.65 0.98 0.69
Open in a new tab

NOTE: Percentages may not add up to 100% because of rounding and missing data

1

Tests for differences between categories were calculated by ANOVA using log-transformed oxidative stress measures.

Associations among the three oxidative stress measures, based on crude and adjusted Pearson’s correlations are given in Table 2. Adjusted correlations were partialled for age, BMI, cotinine levels, alcohol intake, physical activity level, education, days since last menses (for women only) and where appropriate, sex and race. The strongest associations were observed between F2-isoprostanes and the Comet tail moment, with correlations ranging from 0.09 to 0.40 (crude) and −0.09-0.55 (adjusted). These associations were generally stronger in African Americans than in Whites: the most robust association was in African American females (r=0.55) and the weakest in the combined sample of Whites and White females (r=0.08). Correlations were also stronger for women (r=0.26, p=0.03) relative to men (r=0.20, p=0.11). There were no significant correlations between 8-oxo-dG and the comet assay or F2-IsoP; the crude correlations ranged from −0.23 to 0.12, whereas the adjusted correlations ranged from −0.39 to 0.11. Correlations between the 8-oxo-dG and the F2-IsoP measures were also relatively weak. The correlations did not differ when the comet tail moment was assessed using net FPG-sensitive sites or strand breaks plus FPG-sensitive sites. For 8-oxo-dG and the comet tail moment, the correlations for strand breaks plus FPG-sensitive sites (r=−0.01, p=0.90) were similar to those for net FPG-sensitive sites (r=−0.07, p=0.47), as well as correlations with F2-IsoPs for strand breaks plus FPG-sensitive sites (r=0.03, p=0.79) and net FPG-sensitive sites (r=0.14, p=0.16) (data for strand breaks plus FPG-sensitive sites not shown).

Table 2.

Correlations1 of oxidative stress, measured by F2-isoprostanes, 8-oxo-dG, and Comet Tail Moment, by race and sex (n=135)

8-oxo-dG and Comet Tail Moment 8-oxo-dG and F2-isoprostanes F2-isoprostanes and Comet Tail
Moment
Crude Adjusted2 Crude Adjusted Crude Adjusted
Corr. p value Corr. p value Corr. p value Corr. p value Corr. p value Corr. p value
All (n=135) −0.09 0.36 −0.07 0.47 −0.04 0.64 −0.09 0.34 0.22 0.01 0.14 0.16
Male (n=67) −0.02 0.85 −0.17 0.26 0.01 0.91 0.01 0.96 0.20 0.11 0.09 0.52
Female(n=68) −0.14 0.25 −0.04 0.79 −0.11 0.36 −0.11 0.48 0.26 0.03 0.16 0.33
Whites (n=70) −0.01 0.96 −0.07 0.63 −0.02 0.85 −0.16 0.25 0.10 0.40 0.08 0.74
Male (n=36) −0.01 0.93 −0.04 0.87 0.12 0.49 0.04 0.86 0.09 0.62 0.20 0.42
Female(n=34) 0.003 0.99 −0.18 0.45 −0.22 0.21 −0.38 0.10 0.12 0.50 0.08 0.58
African Americans (n=65) −0.08 0.51 0.01 0.95 −0.05 0.71 −0.19 0.25 0.34 0.005 0.27 0.01
Male (n=31) 0.06 0.76 0.02 0.96 −0.23 0.22 −0.39 0.13 0.31 0.09 −0.09 0.74
Female (n=34) −0.17 0.34 −0.02 0.97 −0.08 0.64 0.11 0.77 0.40 0.02 0.55 0.01
Open in a new tab
1

Crude and adjusted correlations were calculated using z-transformed variables to account for differences in units among the three methods.

2

Partial pearson’s correlations were calculated using z-transformed variables and adjusted for age, BMI, cotinine levels, alcohol intake, physical activity level, education, days since last menses (for women only) and where appropriate, sex and race.

Discussion

This study examined associations among three of the most commonly used measures of oxidative stress (i.e., urinary F2-isoprostanes, 8-oxo-dG, and the comet assay) in a sample of healthy African American and White adults ages 20 to 45 in North Carolina. We found modest associations between F2-IsoPs and the comet tail moment, but weak associations of 8-oxo-dG with both F2-IsoPs and the comet tail moment. The results of our study are informative, particularly in light of the fact that researchers often compare results pertaining to oxidative stress across studies and/or assessment methods. While the low levels of associations observed here do not necessarily minimize the internal validity of any measure within a study, they certainly raise concerns about the legitimacy of comparing different oxidative stress measures or assuming that the measures are equivalent, particularly in healthy populations.

Although these measures are widely used, few studies have directly compared oxidative stress measures using the same human samples (Gallo, Khan et al. 2008). The European Standards Committee on Oxidative DNA Damage (ESCODD) analyzed oxidative DNA damage data (8-oxo-dG and the comet with FPG) from healthy males (n=88) using six different laboratories, and similar to our results, found no overall association between methods (ESCODD, Gedik et al. 2005). One laboratory found a strong, statistically significant correlation (r=0.93, p<0.0001), however, the correlations from the remaining five labs were not significant and ranged from −0.41 to 0.59 (ESCODD, Gedik et al. 2005). To our knowledge, no study has directly compared urinary F2-IsoPs with 8-oxo-dG and the comet assay in human lymphocytes. Several studies have assessed oxidative stress using at least two of these assays in an attempt to capture a more complete estimate, however, the different methods often produced conflicting results (Sorensen, Schins et al. 2005; Kato, Ren et al. 2006; Kuhnt, Wagner et al. 2006; Rossner, Gammon et al. 2006; Chia, Hsu et al. 2008). For example, in a case-control study in Long Island, NY, there was an observed association between F2-IsoPs and breast cancer risk, but no such association was seen for 8-oxo-dG and breast cancer risk in the same study (Rossner, Gammon et al. 2006). While it is possible that lipid peroxidation, and not oxidative DNA damage, is the relevant oxidative stress pathway for breast cancer or that urinary levels did not reflect levels in the target tissue (i.e., breast), the absence of an association may also be due to measurement error. Similar inconsistent results were observed in other studies that included both 8-oxo-dG and isoprostane assays (Kato, Ren et al. 2006; Kuhnt, Wagner et al. 2006; Rossner, Gammon et al. 2006), as well as in two other studies that investigated oxidative DNA damage via both 8-oxo-dG and the comet assay (Sorensen, Schins et al. 2005; Chia, Hsu et al. 2008).

The differences seen between the lipid peroxidation and oxidative DNA damage assays are not entirely unexpected because they reflect different pathways of oxidative stress (Bartsch and Nair 2006) and were also assayed using different biological specimens (urine vs. lymphocytes). However, we are less clear about the differences between 8-oxo-dG and the mean comet tail moment. Regardless of whether the comet assay measured overall strand breaks or FPG-sensitive sites only, there was no correlation between 8-oxo-dG and the comet tail moment. The addition of FPG to the comet assay ensures the measurement of oxidative DNA damage from oxidized purines, but is not specific to 8-oxo-dG. Additional modified purines, such as 4,6-diamino-5-formamidopyrimidine (FapyAde) and 2,6-diamino-4-hydroxy-5-formamidopyrimidine (FapyGua), which are produced from a common precursor to 8-oxo-dG by one-electron reduction, are recognized by the FPG protein and may partly account for differences seen here (Dizdaroglu, Kirkali et al. 2008). Furthermore, we cannot rule out that these measures may reflect still unknown pathways or aspects of oxidative stress.

Methodological differences may also affect results. ESCODD noted that results from enzymatic methods, e.g., the comet assay with FPG, were several times lower and also more precise than those from chromatography (ESCODD, Gedik et al. 2005). Similarly, when electrospray tandem mass spectrometry and enzyme-linked immunosorbent assay (ELISA) were used to quantify 8-oxo-dG in 140 workers exposed to polyaromatic hydrocarbons, there was only a significant difference between cases and controls using the LC/MS/MS data, even though the association between methods was significant (Hu, Wu et al. 2004). A multi-laboratory validation study of oxidative stress biomarkers using an acute CCl4 rat model, sponsored by the National Institute of Environmental Health Sciences (NIEHS), concluded that the best candidate biomarkers were plasma malondialdehyde (MDA), urinary 8-oxo-dG, and plasma or urinary isoprostanes (Kadiiska, Gladen et al. 2005).

Our results suggest that there were significant differences by sex, race, physical activity, and passive cigarette smoke exposure according to the oxidative stress assessment method used. For example, African Americans had lower oxidative stress levels when measured by the comet assay (p=0.0003) and F2-IsoPs (p=0.27), but higher 8-oxo-dG (p=0.05) than Whites. Studies examining oxidative stress with intermediate markers, such as dietary antioxidants, or disease endpoints, have often found conflicting results. These differences may in part, explain conflicting results seen across oxidative stress measures.

Considering that there are few intermediate measures of chronic disease risk, the identification and validation of such markers is crucial. The criteria necessary for optimal biomarkers of oxidative stress has been extensively discussed (Mayne 2003; Blumberg 2004; Dalle-Donne, Rossi et al. 2006; Swenberg, Fryar-Tita et al. 2008) and include specificity, accuracy, precision, and reliability. Artifactual oxidation during collection and storage remains a constant challenge (Mayne 2003; Blumberg 2004; ESCODD, Gedik et al. 2005; Dalle-Donne, Rossi et al. 2006). Adding antioxidant agents, such as TEMPO, ensuring proper handling techniques, and storage conditions should be considered to minimize potential artifact. ESCODD identified a reasonable range of 0.3 - 4.2 8-oxo-dG per 106 guanines (ESCODD, Gedik et al. 2005; Milne, Sanchez et al. 2007) and 1.6 ± 0.6 ng per mg creatinine has been defined as the normal range of F2-IsoPs in healthy adults (ESCODD, Gedik et al. 2005; Milne, Sanchez et al. 2007). Almost of all values measured here were within the documented ranges of acceptable values for healthy people, yet there was still only modest agreement between methods used.

We acknowledge some limitations in this analysis. Due to the cross-sectional nature of the study, there were no repeated measures and thus, seasonal and diurnal variability could not be assessed. In addition, there was little variation in oxidative stress levels because the original study examined healthy, non-obese, nonsmokers 20 to 45 years of age. The values presented here for all three biomarkers represent a relatively narrow range of natural variability in normal values. Thus, potential correlations may be hard to detect without greater sample sizes or without widening the distribution of values by inclusion of subjects with known oxidative stress (e.g. smokers or diabetics). Furthermore, we cannot exclude the possibility that correlations were not seen because these methods were not sensitive enough to detect differences within healthy ranges of oxidative stress. These analyses were limited to only three measures of oxidative stress; however, there are other viable methods that are worth examining. We expressed the results of the comet assay using tail moment; however, some studies that suggest % tail DNA better captures damage, especially at low levels (Collins, Oscoz et al. 2008), yet others support tail moment as the most stable estimate of DNA damage (Lee, Oh et al. 2004). Also, since these estimates reflect DNA damage in lymphocytes and lipid peroxidation in urine, they may not represent the status of oxidative stress in other tissues and cells within the body. However, both biospecimens represent “whole body” exposure as lymphocytes circulate throughout the body and urine undergoes metabolism in the excretion process. Furthermore, because urine has a low lipid content, F2-IsoPs are very stable and not subject to auto-oxidation formation (Morrow, Hill et al. 1990; Dalle-Donne, Rossi et al. 2006). There are also notable strengths to this study. All assays were performed by experienced laboratories with a demonstrated record of successful use of these techniques. Samples were processed within 2 h of collection and proper storage techniques were employed. In addition, the study population included equal numbers of African Americans and Whites and had sufficient power to examine associations by race and sex.

Although the development and use of oxidative stress biomarkers is an exciting and promising field, additional validation studies are necessary to assess their accuracy and the information they contribute to various disease-related endpoints. We compared three commonly used measures of oxidative stress, F2-isoprostanes, 8-oxo-dG, and the comet assay, in healthy adults and found modest correlations between F2-IsoPs and the comet assay, but little association between 8-oxo-dG and either F2-IsoPs or the comet assay. This work highlights the limitations of comparing results of oxidative stress across studies or measurement tools in healthy disease-free populations.

Acknowledgements

This study was supported in part by the following grants: RO3 CA108276, K22 CA96556, T32 CA72319, P30ES010126, DK56350, and RR00046. The authors would also like to gratefully acknowledge Corneliu N. Craciunescu’s technical assistance regarding the comet assay.

We are saddened to report that Dr. Morrow and Dr. Satia have passed away.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

References

  1. Ames BN. DNA damage from micronutrient deficiencies is likely to be a major cause of cancer. Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis. 2001;475(1-2):7–20. doi: 10.1016/s0027-5107(01)00070-7. [DOI] [PubMed] [Google Scholar]
  2. Bartsch H, Nair J. Chronic inflammation and oxidative stress in the genesis and perpetuation of cancer: role of lipid peroxidation, DNA damage, and repair. Langenbeck’s Archives of Surgery. 2006;391(5):499–510. doi: 10.1007/s00423-006-0073-1. [DOI] [PubMed] [Google Scholar]
  3. Blumberg J. Use of Biomarkers of Oxidative Stress in Research Studies. 2004;134(11):3188S–3189. doi: 10.1093/jn/134.11.3188S. [DOI] [PubMed] [Google Scholar]
  4. Burlinson B, Tice RR, et al. Fourth International Workgroup on Genotoxicity testing: Results of the in vivo Comet assay workgroup. Mutation Research/Genetic Toxicology and Environmental Mutagenesis. 2007;627(1):31–35. doi: 10.1016/j.mrgentox.2006.08.011. [DOI] [PubMed] [Google Scholar]
  5. Cadet J, Douki T, et al. Artifacts associated with the measurement of oxidized DNA bases. Environ Health Perspect. 1997;105(10):1034–1039. doi: 10.1289/ehp.105-1470384. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Chia T, Hsu CY, et al. Oxidative Damage of Workers in Secondary Metal Recovery Plants Affected by Smoking Status and Joining the Smelting Work. Industrial Health. 2008;46(2):174–182. doi: 10.2486/indhealth.46.174. [DOI] [PubMed] [Google Scholar]
  7. Collins AR. Assays for oxidative stress and antioxidant status: applications to research into the biological effectiveness of polyphenols. 2005;81(1):261S–267. doi: 10.1093/ajcn/81.1.261S. [DOI] [PubMed] [Google Scholar]
  8. Collins AR, Dusinska M, et al. Oxidative damage to DNA: do we have a reliable biomarker? Environ Health Perspect. 1996;104(Suppl 3):465–9. doi: 10.1289/ehp.96104s3465. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Collins AR, Oscoz AA, et al. The comet assay: topical issues. Mutagenesis. 2008 doi: 10.1093/mutage/gem051. gem051. [DOI] [PubMed] [Google Scholar]
  10. Cooke MS, Olinski R, et al. Does measurement of oxidative damage to DNA have clinical significance? Clinica Chimica Acta. 2006;365(1-2):30–49. doi: 10.1016/j.cca.2005.09.009. [DOI] [PubMed] [Google Scholar]
  11. Craft N. High resolution HPLC method for the simultaneous analysis of carotenoids, retinoids, and tocopherols. FASEB J A527 (abs.) 1996 [Google Scholar]
  12. da Costa K-A, Niculescu MD, et al. Choline deficiency increases lymphocyte apoptosis and DNA damage in humans. 2006;84(1):88–94. doi: 10.1093/ajcn/84.1.88. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Dalle-Donne I, Rossi R, et al. Biomarkers of Oxidative Damage in Human Disease. 2006;52(4):601–623. doi: 10.1373/clinchem.2005.061408. [DOI] [PubMed] [Google Scholar]
  14. Dizdaroglu M, Kirkali G, et al. Formamidopyrimidines in DNA: Mechanisms of formation, repair, and biological effects. Free Radical Biology and Medicine. 2008 doi: 10.1016/j.freeradbiomed.2008.07.004. In press. [DOI] [PubMed] [Google Scholar]
  15. ESCODD. Gedik CM, et al. Establishing the background level of base oxidation in human lymphocyte DNA: results of an interlaboratory validation study. FASEB J. 2005;19:82–84. doi: 10.1096/fj.04-1767fje. [DOI] [PubMed] [Google Scholar]
  16. Galan P, Viteri FE, et al. Serum concentrations of [beta]-carotene, vitamins C and E, zinc and selenium are influenced by sex, age, diet, smoking status, alcohol consumption and corpulence in a general French adult population. Eur J Clin Nutr. 2005;59(10):1181–1190. doi: 10.1038/sj.ejcn.1602230. [DOI] [PubMed] [Google Scholar]
  17. Gallo V, Khan A, et al. Validation of biomarkers for the study of environmental carcinogens: a review. Biomarkers. 2008;13(5):505–534. doi: 10.1080/13547500802054611. [DOI] [PubMed] [Google Scholar]
  18. Halliwell B. Why and how should we measure oxidative DNA damage in nutritional studies? How far have we come? Am J Clin Nutr. 2000;72(5):1082–1087. doi: 10.1093/ajcn/72.5.1082. [DOI] [PubMed] [Google Scholar]
  19. Halliwell B, Grootveld M. The measurement of free radical reactions in humans : Some thoughts for future experimentation. FEBS Letters. 1987;213(1):9–14. doi: 10.1016/0014-5793(87)81455-2. [DOI] [PubMed] [Google Scholar]
  20. Hercberg S, Preziosi P, et al. Vitamin status of a healthy French population: dietary intakes and biochemical markers. Int J Vitam Nutr Res. 1994;64:220–32. [PubMed] [Google Scholar]
  21. Hu C-W, Wu M-T, et al. Comparison of analyses of urinary 8-hydroxy-2-deoxyguanosine by isotope-dilution liquid chromatography with electrospray tandem mass spectrometry and by enzyme-linked immunosorbent assay. 2004;18(4):505–510. doi: 10.1002/rcm.1367. [DOI] [PubMed] [Google Scholar]
  22. Hwang E-S, Bowen PE. DNA Damage, a Biomarker of Carcinogenesis: Its Measurement and Modulation by Diet and Environment. 2007;47(1):27–50. doi: 10.1080/10408390600550299. [DOI] [PubMed] [Google Scholar]
  23. Kadiiska MB, Gladen BC, et al. Biomarkers of Oxidative Stress Study II: Are oxidation products of lipids, proteins, and DNA markers of CCl4 poisoning? Free Radical Biology and Medicine. 2005;38(6):698–710. doi: 10.1016/j.freeradbiomed.2004.09.017. [DOI] [PubMed] [Google Scholar]
  24. Kato I, Ren J, et al. Intra- and inter-individual variability in measurements of biomarkers for oxidative damage in vivo: Nutrition and Breast Health Study. Biomarkers. 2006;11(2):143–152. doi: 10.1080/13547500600565693. [DOI] [PubMed] [Google Scholar]
  25. Klaunig JE, Kamendulis LM. The Role of Oxidative Stress in Carcinogenesis. 2004. pp. 239–267. [DOI] [PubMed]
  26. Kuhnt K, Wagner A, et al. Dietary supplementation with 11trans- and 12trans-18:1 and oxidative stress in humans. Am J Clin Nutr. 2006;84(5):981–988. doi: 10.1093/ajcn/84.5.981. [DOI] [PubMed] [Google Scholar]
  27. Lee E, Oh E, et al. Use of the Tail Moment of the Lymphocytes to Evaluate DNA Damage in Human Biomonitoring Studies. Toxicol. Sci. 2004;81(1):121–132. doi: 10.1093/toxsci/kfh184. [DOI] [PubMed] [Google Scholar]
  28. Mayne ST. Antioxidant Nutrients and Chronic Disease: Use of Biomarkers of Exposure and Oxidative Stress Status in Epidemiologic Research. 2003. pp. 933S–940. [DOI] [PubMed]
  29. Milne GL, Morrow JD. Isoprostanes and Related Compounds: Update 2006. Antioxidants & Redox Signaling. 2006;8(7-8):1379–1384. doi: 10.1089/ars.2006.8.1379. [DOI] [PubMed] [Google Scholar]
  30. Milne GL, Sanchez SC, et al. Quantification of F2-isoprostanes as a biomarker of oxidative stress. Nat. Protocols. 2007;2(1):221–226. doi: 10.1038/nprot.2006.375. [DOI] [PubMed] [Google Scholar]
  31. Montuschi P, Barnes P, et al. Insights into oxidative stress: the isoprostanes. Current Medicinal Chemistry. 2007;14(6):703–717. doi: 10.2174/092986707780059607. [DOI] [PubMed] [Google Scholar]
  32. Morrow J, Hill K, et al. A series of prostaglandin F2-like compounds are produced in vivo in humans by a noncyclooxygenase, free radical-catalyzed mechanism. Proc Natl Acad Sci. 1990;87:9383–9387. doi: 10.1073/pnas.87.23.9383. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Musiek ES, McLaughlin B, et al. Electrophilic Cyclopentenone Isoprostanes in Neurodegeneration. Journal of Molecular Neuroscience. 2007;33(1):80–86. doi: 10.1007/s12031-007-0042-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Rossner P, Jr., Gammon MD, et al. Relationship between Urinary 15-F2t-Isoprostane and 8-Oxodeoxyguanosine Levels and Breast Cancer Risk. Cancer Epidemiology Biomarkers & Prevention. 2006;15(4):639–644. doi: 10.1158/1055-9965.EPI-05-0554. [DOI] [PubMed] [Google Scholar]
  35. Sies H. Oxidative stress: Oxidants and Antioxidants. Academic Press; San Diego, CA: 1991. Oxidative Stress: introduction; pp. 15–22. [Google Scholar]
  36. Sorensen M, Schins RPF, et al. Transition Metals in Personal Samples of PM2.5 and Oxidative Stress in Human Volunteers. Cancer Epidemiology Biomarkers & Prevention. 2005;14(5):1340–1343. doi: 10.1158/1055-9965.EPI-04-0899. [DOI] [PubMed] [Google Scholar]
  37. Swenberg JA, Fryar-Tita E, et al. Biomarkers in Toxicology and Risk Assessment: Informing Critical Dose-Response Relationships. Chem. Res. Toxicol. 2008;21(1):253–265. doi: 10.1021/tx700408t. [DOI] [PubMed] [Google Scholar]
  38. Tsimikas S. Oxidative Biomarkers in the Diagnosis and Prognosis of Cardiovascular Disease. The American Journal of Cardiology. 2006;98(11, Supplement 1):S9–S17. doi: 10.1016/j.amjcard.2006.09.015. [DOI] [PubMed] [Google Scholar]
  39. Watters JL, Satia JA, et al. Associations of antioxidant nutrients and oxidative DNA damage in healthy African American and White Adults. Cancer Epidemiology Biomarkers & Prevention. 2007 doi: 10.1158/1055-9965.EPI-06-1030. [DOI] [PubMed] [Google Scholar]
  40. Willett WC. Nutritional epidemiology. Oxford University Press; New York, NY: 1998. [Google Scholar]

RESOURCES