Intra-person Variation of Urinary Biomarkers of Oxidative Stress and Inflammation

Abstract

Background

Oxidative stress and inflammation have been linked to many chronic diseases including cancer and cardiovascular diseases. Urinary levels of F₂-isoprostanes (F₂-IsoPs), 2,3-dinor-5,6-dihydro-15-F_2t-IsoP(15-F_2t-IsoP-M), a major metabolite of F₂-IsoPs, prostaglandin E₂ metabolite (PGE-M), and leukotriene E₄ (LTE₄), have been proposed as biomarkers for oxidative stress and inflammation. However, little information is available regarding the intra-person variation of these biomarkers, hindering their application in epidemiological studies.

Methods

We evaluated the intra-person variation of these four urinary biomarkers among 48 randomly chosen participants of a validation study of a population-based cohort, the Shanghai Men's Health Study (SMHS). Four spot urine samples, collected during each season over a one-year period, were measured for these biomarkers.

Results

The intraclass correlation coefficients (ICCs) for F₂-IsoPs, 15-F_2t-IsoP-M, PGE-M, and LTE₄ were 0.69, 0.76, 0.67, and 0.64, respectively. The Spearman correlation coefficients, derived by using bootstrap analysis of single spot measurements and the average of the other three seasonal measurements, were 0.47, 0.60, 0.61, and 0.57 for F₂-IsoPs, 15-F_2t-IsoP-M, PGE-M, and LTE₄. Except for high correlations between F₂-IsoPs and 15-F_2t-IsoP-M (r =0.65), the other biomarkers were moderately correlated (r =0.21-0.44).

Conclusions

Our study results suggest that these four urinary biomarkers have relatively low intra-person variation over a one-year period.

Impact

Spot measurement of F₂-IsoPs, 15-F_2t-IsoP-M, PGE-M, and LTE₄ could be useful as biomarkers of oxidative stress and inflammation status for epidemiological studies.

Introduction

Oxidative stress, the adverse effect of oxidants on physiological function, has been implicated in the pathogenesis of a variety of human conditions and diseases, such as cancers, cardiovascular disease, neurodegenerative disease, and aging (1, 2). F₂-isoprostanes (F₂-IsoPs) are a unique series of prostaglandin-like compounds formed in vivo via a non-enzymatic mechanism involving the free radical-initiated peroxidation of arachidonic acid (3). F₂-IsoPs are further metabolized to form 2,3-dinor-5,6-dihydro-15-F_2t-IsoP (15-F_2t-IsoP-M), a major end product of F₂-IsoPs excreted in urine (4). Urine is considered to be an ideal biological material for the measurement of F₂-IsoPs because it, unlike plasma, does not contain high lipid content, which minimizes concern about the artifactual generation of isoprostanes by lipid autoxidation during sampling (5). Urinary F₂-IsoPs measured by mass spectrometric methods is considered as a reliable and accurate biomarker of oxidative stress in vivo (6, 7). Measurement of its end metabolite 15-F_2t-IsoP-M in urine may offer an additional advantage over its parent compounds, potentially providing a better integrated index of oxidative stress status in vivo (5). We have recently reported in a nested case-control study that elevated levels of urinary 15-F_2t-IsoP-M are associated with increased risk of breast cancer among obese women (8).

Cumulative evidence from both in vitro and animal studies suggests that cyclooxygenase-2 (COX-2) may be involved in the development and progression of cancer (9) and other diseases (10, 11). It is believed that the pro-inflammatory effects of the COX-2 pathway are largely mediated through PGE₂. PGE₂ is quickly converted to 11α-hydroxy-9,15-dioxo-2,3,4,5-tetranor-prostane-1,20-dioic acid (PGE-M), a major PGE₂ metabolite, and excreted in urine (12). It is generally accepted that the most accurate approach for the assessment of the endogenous production of prostaglandins in humans is to quantify excreted prostaglandin metabolites in urine (12). Leukotrienes (LTs) also play a major role in the inflammatory process (13). LTs are synthesized via 5-lipoxygenase (5-LOX), which catalyzes a two-step conversion of arachidonic acid to LTA₄ (14). The final and biologically active metabolites of the 5-LOX cascade are LTB₄ and cysteinyl LTs (LTC₄, LTD₄, and LTE₄), which are derived from the unstable intermediate LTA₄. LTs are potent mediators of inflammation (15). LTE₄ is more stable than the other LTs and is excreted in urine, where it can be readily measured. It has been suggested that urinary LTE₄ is a reliable marker of endogenous cysteinyl LT formation (16).

Although measurements by recently developed mass-spectrometry-based methods have shown high accuracy and sensitivity, to the best of our knowledge, no study has evaluated the specific intrapersonal variations of urinary levels of F2-IsoPs, 15-F2t-IsoP-M, PGE-M, and LTE4. Understanding long-term intrapersonal variations is essential to the implementation and interpretation of epidemiological research on the associations between these biomarkers and health outcomes as most epidemiological studies have only collected one biospecimen. Using the resources collected by a dietary validation study conducted within the Shanghai Men's Health Study (SMHS), we evaluated intrapersonal variation of these four biomarkers and their correlations with selected oxidative stress and inflammation-related conditions.

Materials and Methods

Subjects and Design

The parent study, the SMHS, an ongoing, population-based prospective cohort study of 61,500 men aged 40 to 74 years, was designed to investigate the associations of lifestyle factors with risk of cancers and other major chronic diseases. Recruitment for the SMHS was initiated in April 2002 and was completed in June 2006 with a response rate of 74.0%. A total of 196 subjects were randomly selected from the SMHS and completed a validation study between April 2003 and May 2004 to evaluate the performance of the SMHS food frequency and physical activity questionnaires. Participants completed 12 monthly 24-hour dietary and physical activity recalls. Spot urine and blood samples were collected each quarter during the one-year study period. Urine samples were collected using a sterilized cup containing 125 mg of ascorbic acid, transported in a portable, insulated bag with ice packs (at approximately 0 to 4°C), processed within 6 hours of collection, and stored at -80 °C. The overall participation rate for the validation study was 69.3%. Taking into account both statistical power considerations and budget constraints, we randomly selected 48 men from among validation study participants who had provided 4 spot urine and blood samples throughout the year for the current study.

Laboratory Methods

The levels of four urinary biomarkers were measured by accurate and precise laboratory methods. F₂-IsoPs and 15-F_2t-IsoP-M were measured using gas chromatography/negative ion chemical ionization mass spectrometry (GC/NICI-MS) as previously described (7, 17). Briefly, urine was extracted using C18 and silica solid phase extraction cartridges after addition of an isotopically labeled internal standard. Endogenous F₂-IsoPs or 15-F_2t-IsoP-M were then converted to the pentafluorobenzyl ester, trimethylsilyl ether derivatives for GC/NICI-MS analysis. This analysis was performed by using an Agilent 5973 GC/MS instrument with the column temperature programmed from 190°C to 300°C at 20°C per minute. The lower limit of detection for these assays is in the range of 5 pg.

PGE-M was measured using a liquid chromatography/tandem mass spectrometric (LC/MS) method described previously (18). Concisely, 1mL urine was acidified to pH 3 with HCl and endogenous PGE-M was then converted to the O-methyloxime derivative by treatment of methyloxime HCl. The sample was then extracted with a C18 solid phase extraction column. The eluate was dried, reconstituted in mobile phase, and filtered for analysis using LC/MS. The lower limit of detection of PGE-M is 40 pg.

LTE₄ was measured by ultra performance liquid chromatography (UPLC)/MS; the specific procedure was described by Duffield-Lillico (19).

Statistical Analysis

We compared the basic demographic characteristics and selected lifestyle factors of participants in the current study with participants of the parent cohort and 196 participants of the validation study using the t test for continuous variables and the chi-square test for categorical variables. The measurements of each urinary biomarker were first ranked; then repeated measures ANOVA was performed to evaluate variation among the four seasons, because these biomarkers were non-normally distributed. Intraclass correlation coefficients (ICCs) were calculated to evaluate intrapersonal variation of the biomarkers by using the following one-way random effect model: $\mathit{\text{ICC}}={\sigma }_B^2/({\sigma }_B^2+{\sigma }_W^2)=\mathit{[}\mathit{\text{between group variance}}\mathit{/}\mathit{(}\mathit{\text{between}} \mathit{\text{group}} \mathit{\text{variance}} + \mathit{\text{within}} \mathit{\text{group}} \mathit{\text{variance}}\mathit{)}\mathit{]}$ (20). To evaluate whether a single spot urine sample can reflect long-term levels of the biomarkers, Spearman correlation coefficients between each seasonal measurement and the average of other three seasonal measurements were estimated using the bootstrap method with 2,000 repeats for each biomarker.

The relationship between the average of the four seasonal urine measurements of these biomarkers and selected lifestyle factors was estimated by Spearman correlation coefficients. Smoking and alcohol consumption were analyzed based on number of cigarettes/day and ounces/day, respectively. Physical activity was measured as energy expenditure, which was calculated by multiplying the time (in hours) spent on each activity by the corresponding MET (metabolic equivalent task) value obtained from the Compendium of Physical Activity (MET-h/day/year) (21). Information on diabetes (yes or no) and hypertension (yes or no) was gathered at the baseline interview. A co-morbidity score was generated from baseline information according to the Charlson index (22, 23). Observations 5 standard deviations away from the mean were considered to be outliers (24); three values were excluded for this reason, one for 15-F_2t-IsoP-M and two for PGE-M. Statistical analyses were performed using SAS version 9.2 for Windows software (SAS Institute, Inc., Cary, NC). Statistical significance was considered to be two-sided P < 0.05.

Results

Participants of the current study were similar to the 196 validation study participants (data not shown) and to the parent SMHS cohort regarding age, body mass index, waist-to-hip ratio, systolic blood pressure, diastolic blood pressure, current smoking, current alcohol consumption, and total energy intake at baseline interview (Table 1), suggesting that the biomarker study participants are representative of the whole cohort.

Table 1. Comparison of participants in the current study with all cohort members, Shanghai Men's Health Study.

Characteristic	Current study subjects (n = 48)	Cohort members (n = 61,500)	P value†
Age (years) at recruitment (mean (SD*))	54.81 (9.19)	54.88(9.74)	0.9638
Body mass index (mean (SD))	24.00(2.89)	23.72(3.08)	0.5318
Waist-to-hip ratio (mean (SD))	0. 89(0.06)	0.90(0.06)	0.4441
Systolic blood pressure (mmHg) (mean(SD))	129.10 (17.85)	127.40(17.93)	0.5057
Diastolic blood pressure (mmHg) (mean(SD))	84.00 (11.44)	82.34(10.41)	0.2684
Current smoking (%)	56.25	58.63	0.7378
Amount of smoking for current smokers (cigarettes/day) (Median(IQR‡))	12 (10)	20 (10)	0.1335
Current alcohol consumption (%)	29.17	29.28	0.9862
Amount of alcohol consumption for current drinkers (ounces/day)(Median(IQR))	0.92(0.37)	0.92(1.03)	0.6820
Total energy intake (kcal/day) (mean(SD))	1981.90(451.80)	1908.80(485.00)	0.2965

^*SD: standard deviation.

^†One-sample t test for continuous variables except for the Wilcoxon Rank-Sum Test for current smoking and current alcohol consumption, chi-square test for categorical variables.

^‡Interquartile range=Q3-Q1.

The medians of urinary levels of F₂-IsoPs, 15-F_2t-IsoP-M, PGE-M, and LTE₄ were 1.90, 0.53, 13.04, and 0.09 (ng/mg creatinine), respectively. These values were consistent with previously reported values of these urinary compounds in healthy Caucasian-American adults (7, 25). No significant seasonal variations were observed, although the median level of F₂-IsoPs appeared to be slightly higher in winter than in the other three seasons (Table 2). However, the interquatile range for this marker did not vary by season. The correlations between a single measurement and the average of the three remaining measurements for the four urinary biomarkers were relatively high (0.47 ≤ r ≤ 0.61). The ICCs were 0.69, 0.76, 0.67, and 0.64 for F₂-IsoPs, 15-F_2t-IsoP-M, PGE-M, and LTE₄.

Table 2. Median levels (interquartile range) of urinary biomarkers by season, Spearman correlation coefficients (95% confidence intervals) and ICC (95% confidence intervals).

Biomarkers	Median (IQR‖) (ng/mg creatinine)						Correlation (95%CI) ‡				r § (95%CI)	ICC (95%CI)
	Winter	Spring	Summer	Fall	Average*	P value †	Winter	Spring	Summer	Fall
F2-IsoPs	2.21 (1.27)	1.88 (1.46)	1.89 (1.17)	1.80 (1.12)	1.90 (1.21)	0.1782	0.64 (0.43,0.78)	0.74 (0.58,0.85)	0.66 (0.45,0.79)	0.67 (0.47,0.80)	0.47 (0.45,0.50)	0.69 (0.59,0.77)
15-F2t-IsoP-M	0.55 (0.32)	0.53 (0.34)	0.51 (0.34)	0.52 (0.37)	0.53 (0.32)	0.8766	0.63 (0.42,0.78)	0.69 (0.49,0.82)	0.57 (0.32,0.74)	0.61 (0.38,0.77)	0.60 (0.56,0.64)	0.76 (0.68,0.82)
PGE-M	12.07 (10.16)	12.36 (9.72)	13.85 (12.01)	13.17 (14.93)	13.04 (11.96)	0.5307	0.66 (0.46,0.80)	0.51 (0.25,0.70)	0.60 (0.38,0.76)	0.70 (0.51,0.82)	0.61 (0.59,0.64)	0.67 (0.57,0.75)
LTE4	0.09 (0.11)	0.08 (0.08)	0.09 (0.11)	0.10 (0.08)	0.09 (0.08)	0.9047	0.56 (0.31,0.74)	0.59 (0.35,0.76)	0.56 (0.31,0.74)	0.61 (0.38,0.77)	0.57 (0.54,0.60)	0.64 (0.53,0.73)

^*Average (median) of four seasonal samples.

^†Repeated measures ANOVA after ranked measurements of each urinary biomarkers.

^‡Spearman correlation coefficient between each individual measurement and the average of the other three seasonal measurements.

^§Spearman correlations between a randomly chosen individual measurement and the average of the other three seasonal measurements, estimated using the bootstrap method.

^‖Interquartile range=Q3-Q1.

The four biomarkers were positively correlated with each other (0.21 ≤ r ≤0.65) (Table 3). F₂-IsoPs and LTE₄ were positively associated with amount of smoking (r = 0.39 and 0.40, respectively); 15-F_2t-IsoP-M was positively correlated with amount of alcohol consumption (r = 0.32) (Table 4).

Table 3. Correlations^* (95% confidence intervals) between urinary biomarkers.

Biomarkers	F2-IsoPs	15-F2t-IsoP-M	PGE_M	LTE4
F2-IsoPs	1.00
15-F2t-IsoP-M	0.65(0.56,0.73) †	1.00
PGE_M	0.44(0.31,0.55) †	0.44(0.31,0.56) †	1.00
LTE4	0.37(0.23,0.49) †	0.21(0.06,0.36) †	0.21(0.06,0.35) †	1.00

^*Spearman correlation coefficient.

^†P <0.05.

Table 4. Correlations^* (95% confidence intervals) between urinary biomarkers and selected lifestyle factors.

Characteristics	F2-IsoPs	15-F2t-IsoP-M	PGE-M	LTE4
Age †	0.06(-0.24,0.34)	0.15(-0.15,0.43)	0.11(-0.19,0.39)	-0.22(-0.49,0.09)
Amount of smoking ‡	0.39(0.12,0.61) §	0.25(-0.05,0.51)	0.10(-0.20,0.38)	0.40(0.11,0.63) §
Amount of alcohol consumption	0.13(-0.16,0.41)	0.32(0.03,0.56) §	0.25(-0.05,0.51)	-0.05(-0.35,0.25)
Body mass index	-0.27(-0.52,0.03)	-0.08(-0.37,0.22)	-0.17(-0.44,0.13)	-0.04(-0.34,0.27)
Physical activity	0.11(-0.18,0.39)	-0.07(-0.36,0.23)	0.00(-0.29,0.29)	0.18(-0.13,0.46)
Diabetes	0.06(-0.24,0.34)	-0.08(-0.37,0.22)	0.03(-0.27,0.32)	0.00(-0.30,0.31)
Hypertension	0.21(-0.08,0.47)	-0.10(-0.38,0.21)	0.03(-0.27,0.32)	0.14(-0.17,0.42)
Co-morbidity	0.14(-0.16,0.41)	0.18(-0.12,0.45)	0.01(-0.29,0.30)	0.11(-0.20,0.40)

^*Spearman correlation coefficient between the average of the four seasonal measurements of urinary biomarkers and selected lifestyle factors after adjustment for age and amount of cigarette smoking (cigarettes/day).

^†Adjusted for amount of cigarette smoking (cigarettes/day).

^‡Adjusted for age and amount of alcohol consumption (ounce/day).

^§P <0.05.

Given the observed ICC, we also estimated attenuation factors if single measurements of the biomarkers were used in an epidemiological study. Assuming a true relative risk of 2.0, by multiplying the natural logarithm of the specified true relative risks with the ICC and exponentiating the result (26), we estimated the observed relative risk would be 80.0% of the true value for F₂-IsoPs, PGE_M, and LTE₄, and 85.0% for 15-F_2t-IsoP-M (data not shown in tables).

Discussion

F₂-IsoPs, 15-F_2t-IsoP-M, PGE-M, and LTE₄ are oxidation products of arachidonic acid and markers of oxidative stress and/or inflammation. In this study, we found little seasonal or intrapersonal variation in these four biomarkers. The ICCs for the 4 seasonal measurements were reasonable high, and a single measurement correlated very well with the average of the three other measurements. Our results suggest that a single measurement of a spot urine sample of these biomarkers can reflect levels of these biomarkers over a 1-year period of time.

Previous studies have suggested that PGE-M is a potential biomarker for the detection of colorectal cancer risk, non-small cell lung cancer, and gastric cancer risk (18, 27-29). The level of urinary LTE₄ is elevated in persons with bronchial asthma, Crohn's disease, ulcerative colitis, and other diseases (19, 30). Smoking is also a known risk factor for some of these diseases (31). In the current study, we found that F₂-IsoPs and LTE₄ were positively correlated with amount of smoking, which is consistent with other studies (19, 32) and suggests that the effect of smoking on the above-mentioned diseases is mediated, at least in part, through oxidative and inflammation pathways.

Meagher et al. found a dose-dependent increase in urinary F₂-IsoPs excretion with respect to alcohol consumption in ten healthy volunteers (33). Although there was no significant relationship between F₂-IsoPs and alcohol consumption in the current study, we found that 15-F_2t-IsoP-M, the main urinary metabolite of F₂-IsoPs, was positively associated with alcohol consumption.

Exposure misclassification is one of the major concerns for epidemiological studies and generally attenuates the risk estimate when the misclassification is non-differential. In our study, we found that using a single urinary measurement of F₂-IsoPs, 15-F_2t-IsoP-M, PGE-M, and LTE₄ would result in moderately underestimated risk estimates (e.g., 15%-20%, if the true relative risk were 2). These results suggest that these biomarkers can be very useful in epidemiological studies.

Our study has several strengths, such as its population-based cohort study design and high response rates. The participants of the current study were randomly selected from the parent cohort and are representative of the parent study. The 4 urine samples collected over a one-year period provided a unique opportunity to evaluate intrapersonal and seasonal variation for four newly emerged biomarkers of oxidative stress and inflammation. The main limitation of the study is the relatively small number of samples involved, which prevented an evaluation of associations of these biomarkers with additional lifestyle factors and conditions that are known to be associated with oxidative stress and inflammation. Because variation of biomarkers may be influenced by exposure level, the findings of our study may not be directly generalizable to other populations.

In summary, we found that urinary levels of F₂-IsoPs, 15-F_2t-IsoP-M, PGE-M, and LTE₄ are stable and that measurements based on a single spot urine sample reflect well the level of these biomarkers for over one year among middle-aged and elderly Chinese men. Our study suggests that urinary levels of F₂-IsoPs, 15-F_2t-IsoP-M, PGE-M, and LTE₄ can be used as biomarkers in large epidemiological studies.