DNA damage levels in prostate cancer cases and controls

Abstract

This study used the alkaline Comet assay to evaluate whether basal or H ₂ O ₂ -induced DNA damage is associated with prostate cancer (CaP) risk. Using lymphocyte samples from 158 CaP cases and 128 controls, collected in an ongoing case–control study, our results showed that basal DNA damage did not differ between cases and controls. However, the H ₂ O ₂ -induced DNA damage level was significantly higher in incident cases (mean ± SD; 6.61 ± 4.43, n = 102) than controls (5.30 ± 3.60, n = 128) or prevalent cases (4.47 ± 3.19; n = 56). Incident cases with a positive smoking history had significantly higher H ₂ O ₂ -induced DNA damage than never-smokers (7.57 ± 4.82 versus 4.52 ± 2.40; P < 0.001). Above-median H ₂ O ₂ -induced DNA damage was associated with a 1.61-fold increase in CaP risk [95% confidence interval (CI) = 0.92–2.81], after adjustment for age, race, benign prostatic hyperplasia (BPH), smoking history and family history (FH). Using the lowest quartile of H ₂ O ₂ -induced DNA damage as the referent group, the adjusted ORs for the 25th, 50th and 75th quartiles were 0.90 (95% CI = 0.39–2.05), 1.06 (95% CI = 0.48–2.35) and 2.05 (95% CI = 0.96–4.37), respectively ( P = 0.046, test for linear trend). The association between CaP and DNA damage was modified by age, smoking history, family history and body mass index. Our results suggest that DNA damage may be associated with CaP risk. However, larger case–control and follow-up studies are warranted to further evaluate the potential application of the alkaline Comet assay in CaP risk assessment and prevention.

Introduction

Prostate cancer (CaP) is the most common cancer in American men ( 1 ). In 2005, ∼232 090 USA men will be diagnosed with CaP, and 30 350 will die from it ( 1 ). The well-established risk factors are ethnicity/race and family history (FH). CaP incidence increases with age ( 1 ). An accumulation of DNA damage and a decline in DNA repair during aging may lead to CaP development ( 2 ). The mutational spectra of the androgen receptor gene and p53 gene in tumor tissue suggest that both endogenous and exogenous carcinogens play critical roles in human prostate carcinogenesis ( 3 – 5 ). Persistent DNA damage and mutations may result from increased exposure to carcinogens and suboptimal DNA repair ( 6 ).

Two previous studies suggest that age-related structural changes in the DNA from CaP tissue are likely a result of oxidative damage induced by hydroxyl radicals ( 2 , 7 ). More than 70% of CaP cases are diagnosed in men over age 65 ( 1 ). A previous study reported that the ratio of mutagenic 8-hydroxy to non-mutagenic base lesions in prostate DNA increased ∼3-fold in men between the ages of 16 and 83 ( 2 ). Another study suggests that substantial age-related changes in DNA structure play a role in the development of CaP in older men ( 7 ). Although their results must be validated, they speculate that prostatic epithelial cells may be subjected to chronic oxidative stresses, perhaps as a consequence of the high-level expression of oxidizing enzymes and/or exposure to androgen and/or other steroid hormone metabolites ( 8 , 9 ).

Cytogenetic assays, such as chromosomal aberrations, micronuclei and sister chromatid exchanges, have been used to measure cellular DNA damage ( 10 ). This study tested single-cell gel electrophoresis, or the Comet assay, which estimates DNA strand breakage and alkali-labile sites by measuring DNA migration from the nucleus in a single cell. Intact DNA does not migrate far from the nucleus, but relaxed, broken DNA does, resulting in images that look like comets with tails that consist of DNA fragments. The assay has been applied to several population-based studies ( 11 – 16 ). One study used the Comet assay to demonstrate that smokers with elevated DNA damage in their non-neoplastic urothelial cells may be more susceptible to urinary bladder cancer ( 17 ). Patients with cervical dysplasia exhibited higher DNA damage in both cervical epithelial cells and peripheral blood lymphocytes ( 18 ). Breast cancer cases and women with FH of breast cancer have elevated levels of basal DNA damage, after treatment with either N -methyl- N -nitrosoguanidine or ionizing radiation, and following repair ( 19 , 20 ). To the best of our knowledge, this study is the first using the Comet assay to evaluate whether DNA damage is associated with CaP risk.

Materials and methods

Study population

CaP cases and controls were recruited from the Departments of Urology and Internal Medicine of the Wake Forest University School of Medicine using sequential patient populations as described previously ( 21 ). All study subjects received a detailed description of the study protocol and signed their informed consent, as approved by the medical center's Institutional Review Board. The general eligibility criteria were (i) able to comprehend informed consent and (ii) without previously diagnosed cancer. The exclusion criteria were (i) clinical diagnosis of autoimmune diseases; (ii) chronic inflammatory conditions; and (iii) infections within the past 6 weeks. Blood samples were collected from all subjects.

Two groups of cases were recruited from the Urology Clinic: (i) incident ( n = 102)—newly diagnosed, untreated cases; and (ii) prevalent ( n = 56)—cases diagnosed with CaP within 5 years and free of cancer or treatments for at least 6 months prior to study entry. Case status was confirmed by medical records and pathology reports. Controls were frequency-matched to cases by age (±5 years). Two groups of controls were recruited from the Urology and Internal Medicine clinics: men with (i) normal prostate-specific antigen (PSA) levels (<4.0 ng/ml) and normal digital rectal exam (DRE); and (ii) abnormal PSA or DRE but negative biopsy results for CaP. Controls recruited from the Urology clinic were mainly follow-up patients with urological problems, such as benign prostatic hyperplasia (BPH), while those from the Internal Medicine clinic were mainly patients coming in for an annual PSA screening without active medical conditions. A self-administered questionnaire collected information on: (i) demographic factors, such as age, race, weight and height; (ii) medical history and medication use; (iii) smoking history; and (iv) FH of cancer. Men who had at least one first-degree relative with CaP were considered to have a positive FH. The Block food frequency questionnaire (1998 version) was used to estimate nutrient intakes. The overall response rates for cases and controls were 89 and 95%, respectively.

Alkaline Comet assay

We modified the Comet assay protocol as described previously ( 20 , 22 ). As part of the quality control effort, aliquots of cryopreserved lymphocytes from a healthy individual were used as an internal control for each batch of assay samples. To adjust for anticipated batch-to-batch assay variations, each batch of coded sample consisted of 50% from CaP cases and 50% from controls. These samples were thawed, suspended slowly in thawing medium (50% fetal bovine serum, 40% RPMI 1640 and 10% dextrose), and centrifuged at 300 g at 4°C for 10 min. Using trypan blue exclusion, we selected samples with viability >90% for the Comet assay. Cell pellets were resuspended and incubated for 10 min in ice-cold phosphate-buffered saline (PBS) without calcium and magnesium (basal) or containing H ₂ O ₂ (final concentration of 500 μM). Lymphocytes were resuspended at a final concentration of 3 × 10 ⁵ cells/ml in RPMI 1640 without phenol red; 50 μl of cell suspension (with or without H ₂ O ₂ treatment) was mixed with 500 μl of low-melting-point agarose (0.5% low-melting-point agarose in PBS), and 50 μl were spread onto two Comet slide wells (Trevigen, Gaithersburg, MD).

Slides were allowed to solidify at 4°C for 30 min and then placed in a prechilled lysis solution (Trevigen) for 1 h. After lysis, slides were treated with an alkali solution (300 mM NaOH and 1 mM EDTA) at room temperature for 1 h and electrophoresed in TBE at ∼0.8 V/cm for 15 min. Slides were fixed in 100% ethanol for 5 min and stored at 4°C until image analysis. Cellular DNA was stained with 1:10 000 SYBR Green in TE buffer, followed by the addition of antifade solution (0.37 mM p -phenylenediamine dihydrochloride/90% glycerol in PBS, Ca ²⁺ , Mg ²⁺ free) and a coverslip. The slides were analyzed using the LAI Comet Assay Analysis System (Loats Associates, Westminster, MD), and the results presented as the mean comet-tail moment of 50 cells. The comet-tail moment is defined as the product of the fraction of cellular DNA in the comet tail and the tail length. A higher comet-tail moment value represents a greater number of cellular DNA strand breaks.

Statistical analysis

The characteristics of CaP cases and controls were compared using Student's t -test, χ ² -test or Fisher's exact-test. Analysis of covariance was used to compare differences in mean comet-tail moments between cases and controls, after adjusting for characteristics or two-way interactions between case–control status and characteristics. Logistic regression was used to calculate crude and adjusted odds ratios (OR) and 95% confidence intervals (CI). Likelihood ratio test (LRT) was used to test for interactions. All statistical analyses were carried out using the Statistical Analysis System (SAS Institute, Cary, NC) for personal computers and the S-Plus Statistical Package (Insightful Corporation, Seattle, WA).

Results

Table I summarizes study subjects' demographic characteristics. Age, race, smoking history, FH and body mass index (BMI) did not differ significantly between cases and controls. However, smoking status differed significantly between cases and controls, with a higher percentage of cases being current smokers than controls (18% versus 7%). The proportion of subjects with a history of BPH also differed significantly, with a higher percentage of controls having a history of BPH than did cases (56% versus 42%; P = 0.05).

As part of the laboratory assay quality-control effort, we included a cryopreserved aliquot of lymphocytes from a healthy subject with each batch of assay samples to monitor batch-to-batch assay variability. In 24 batches, the mean ± SD of basal and H ₂ O ₂ -induced comet-tail moments were 0.44 ± 0.15 and 2.44 ± 0.74, respectively, and the coefficient of variation was 35% for basal and 30% for H ₂ O ₂ -induced DNA damage. We also plotted the mean basal and H ₂ O ₂ -induced comet-tail moments of cases and controls separately by assay batch and did not observe any unusual time-dependent drift (data not shown). Overall, our technical variability supports the use of the Comet assay in molecular epidemiology studies.

We first evaluated whether DNA damage differed between the two groups of cases ( Table II ). For 102 newly diagnosed cases before any treatment, basal and H ₂ O ₂ -induced damage levels were 1.31 ± 1.22 and 6.61 ± 4.43, respectively. For 56 cases diagnosed within 5 years and free of cancer or treatments for at least 6 months before study entry, basal and H ₂ O ₂ -induced damage levels were 1.03 ± 0.97 and 4.47 ± 3.19, respectively. No significant difference was evident for basal damage levels between these two groups ( P = 0.14). However, mean H ₂ O ₂ -induced damage levels differed significantly between them ( P < 0.01). Therefore, data from incident cases were used for subsequent case–control comparisons.

In comparing incident cases and controls, there was no overall difference in basal DNA damage between the two groups or after stratifying by age, race, smoking history, smoking status, FH, BPH history and BMI. The incident cases had higher mean H ₂ O ₂ -induced DNA damage compared with controls (6.61 ±4.43 versus 5.30 ± 3.60; P = 0.01) ( Table II ). Within the incident case group, subjects with a positive smoking history had significantly higher H ₂ O ₂ -induced comet-tail moments compared with never-smokers (7.57 ± 4.82 versus 4.52 ± 2.40; P < 0.01). Both former and current smokers had significantly higher H ₂ O ₂ -induced damage than never-smokers ( P < 0.01). To explore whether dietary behavior change following cancer diagnosis affected the observed lower DNA damage levels in prevalent cases than in incident cases, we pilot-tested the dietary intake and supplementation data using the Block food frequency questionnaire. Prevalent cases ( n = 40) have significantly higher dietary intakes of antioxidants than incident cases ( n = 66), including vitamins B1, B6 and C, riboflavin, folate, alpha- and beta-carotene and fruit servings. Furthermore, a higher percentage of prevalent cases than incident cases used supplements, including vitamin A ( P = 0.07), vitamin D ( P = 0.06), vitamin E ( P = 0.09), beta-carotene ( P = 0.02) and selenium ( P = 0.01).

In case-only analysis, we also evaluated whether DNA damage was related to tumor aggressiveness, as determined by Gleason score, tumor stage and pretreatment PSA level. Our results showed that in incident cases, neither basal nor H ₂ O ₂ -induced DNA damage levels differed by biopsy Gleason score (≤7 versus >7; P = 0.26 and 0.38). The mean ± SD basal damage value was 1.31 ± 1.26 ( n = 81) for cases with Gleason score ≤7; and 0.90 ± 0.93 ( n = 13) for cases with Gleason score >7. The mean ± SD H ₂ O ₂ -induced damage value was 6.57 ± 4.68 ( n = 81) and 5.82 ± 2.40 ( n = 13) for cases with Gleason score ≤7 and >7, respectively. Similarly, neither basal nor H ₂ O ₂ -induced DNA damage levels differed by tumor stage ( P = 0.72 and 0.67). Neither basal nor H ₂ O ₂ -induced DNA damage levels in controls and incident cases differed by PSA level. In controls, the average basal comet-tail moments for subjects with PSA ≤2 and >2–4 were 1.13 ± 1.02 and 1.07 ± 0.94, respectively ( P = 0.81). The average H ₂ O ₂ -induced comet-tail moments were 5.10 ± 3.42 and 5.99 ± 4.31, respectively ( P = 0.26). In incident cases, the average basal comet-tail moment levels for subjects with PSA <10 and ≥10 were 1.20 ± 1.28 and 1.39 ± 1.02, respectively ( P = 0.54). The mean H ₂ O ₂ -induced comet-tail moment levels were 5.99 ± 3.94 and 8.11 ± 5.77, respectively ( P = 0.13).

When DNA damage level was dichotomized by the median comet-tail moment of controls, there was a slightly higher but non-significant association (OR = 1.20, 95% CI = 0.70–2.06) between basal DNA damage and CaP risk ( Table III ). When we stratified basal damage values by quartiles of controls, there was no dose-dependent association between DNA damage and CaP risk ( P = 0.19, test for linear trend). However, above-median H ₂ O ₂ -induced DNA damage level was associated with a higher but non-significant 1.61-fold CaP risk (95% CI = 0.92–2.81), after adjusting for age, race, smoking history, FH and BPH. When we stratified by quartiles, there was a significant dose-dependent association between increasing H ₂ O ₂ -induced damage and elevated CaP risk ( P = 0.046, test for linear trend). Using the lowest quartile of DNA damage as the referent group, the adjusted ORs for the 25th, 50th and 75th quartiles were 0.90 (95% CI = 0.39–2.05), 1.06 (95% CI = 0.48–2.35) and 2.05 (95% CI = 0.96–4.37), respectively.

We also evaluated whether the association between DNA damage and CaP risk was modified by age, smoking history, FH and BMI. As shown in Table IV , significant interaction was observed between age and level of basal DNA damage ( P = 0.047). The association was stronger in the younger age group (OR = 1.82, 95% CI = 0.84–3.95) than older (OR = 0.59, 95% CI = 0.25–1.38) age group. The interaction between basal DNA damage and BMI was marginally significant ( P = 0.061). A stronger association between basal damage and CaP risk was observed in subjects with a higher BMI (OR = 2.07; 95% CI = 0.89–4.78) than those with a lower BMI (OR = 0.72; 95% CI = 0.31–1.66). Marginally significant interaction was also observed between H ₂ O ₂ -induced DNA damage and smoking history ( P = 0.098). The association between H ₂ O ₂ -induced damage and CaP was observed only in those with a positive smoking history (OR = 2.34; 95% CI = 1.15–4.77), not in non-smokers (OR = 0.71; 95% CI = 0.27–1.92).

Discussion

Increasing evidence suggests the roles of DNA damage/repair in human CaP risk ( 2 , 7 , 8 , 20 – 25 ). Oxidative stress and accumulated genomic damage may contribute to prostate carcinogenesis ( 9 , 26 ). The results from this study showed that the H ₂ O ₂ -induced DNA damage level was significantly higher in incident cases than controls and prevalent cases. However, the mean basal DNA damage level did not differ between cases and controls. There was a dose-dependent association between increasing H ₂ O ₂ -induced DNA damage and higher CaP risk. Our data also demonstrated that the CaP/damage association may be modified by age, smoking history and BMI. In summary, our current results suggest that elevated DNA damage may be associated with human CaP risk.

Although the role of smoking in CaP is unclear, several studies have shown a positive association between smoking and fatal CaP ( 27 – 30 ). Cigarette smoke contains such carcinogens as polycyclic aromatic hydrocarbons and heterocyclic aromatic amines ( 31 ), which can both be activated by prostate cells ( 32 ). As shown in Table II , a positive smoking history has a significant effect on DNA damage in incident cases but not in controls. Data from previous studies suggest that current smokers tend to have higher DNA repair capacity compared with former or non-smokers ( 33 ), and smokers had higher expression levels of several DNA repair genes than non-smokers ( 34 ). Wei et al. ( 35 ) showed that heavy smokers have more proficient DNA repair than light smokers. They suggest that current smokers may have an adaptive response to tobacco carcinogens, upregulating their DNA repair in response to chronic tobacco-related insults ( 36 , 37 ). In theory, differential ‘adaptive responses’ or ‘damage-induced responses’ to chronic exposure may result from genetic variations in drug metabolism and/or DNA repair. The evolving hypothesis is that subjects with lower cancer risk may have up-regulated detoxification and DNA repair enzymes in response to chronic exposures, and subjects with higher cancer risk lack this defense mechanism. Alternatively, extremely high exposures may overwhelm or inhibit the system. To support this hypothesis, our results in Table IV suggest that the association between H ₂ O ₂ -induced damage and CaP is stronger in subjects with a positive smoking history than never-smokers, which suggests that H ₂ O ₂ -induced damage may serve as a susceptibility risk marker, particularly in smokers.

The role of BMI in CaP risk is under intensive investigation. However, the results from a number of previous studies were inconsistent. Some showed no association between CaP and BMI ( 38 – 40 ); others found a slightly increased risk with higher BMI ( 41 – 44 ). One interesting study even showed that men with higher BMI (≥30 kg/m ² ) had lower CaP risk than men with lower BMI (23–24.9 kg/m ² ) ( 45 ). In addition, one study found that obese men (BMI ≥ 30) had a significantly higher CaP mortality rate than non-obese men (BMI < 25) ( 46 ). Our current results showed a marginally significant interaction between BMI and basal damage in CaP risk ( Table IV ), suggesting that basal DNA damage may serve a CaP risk marker in subjects with high BMI. Intriguingly, a similar interaction was observed in a previous study of breast cancer ( 20 ). Future larger studies are warranted to test whether combining BMI and DNA damage data will refine the risk model for CaP development and/or progression.

Although incident cases are usually preferred in case–control studies, DNA damage measurements in lymphocytes may be modulated by tumor-associated factors, such as tumor-associated antigens and cytokines. Therefore, we also evaluated DNA damage levels in samples collected from cancer-free subjects who were diagnosed previously with CaP. To avoid potential survival bias and treatment effects, we limited our recruitment to cases diagnosed with CaP within 5 years and free of treatments and disease for at least 6 months before study entry. Our results showed that basal DNA damage levels were similar in prevalent and incident cases. However, the H ₂ O ₂ -induced DNA damage level was significantly higher in incident cases. We considered two possible explanations: (i) prevalent cases may have changed to healthier dietary habits and/or lifestyles; and/or (ii) tumor-associated factors may be absent in prevalent cases. Regarding the first hypothesis, our preliminary data provided some evidence that prevalent cases had significantly higher intakes of dietary antioxidants and supplements than incident cases. Although these data only indirectly support our first hypothesis, they have potential implications in CaP prevention, based on previous results showing that dietary micronutrients may modulate DNA damage and repair ( 47 – 54 ). Previous studies using the Comet DNA damage assay as the biomarker have shown that increased lycopene and vitamin C intake from tomato products, grape juice, black tea polyphenols and soy milk protect against DNA damage in lymphocytes ( 47 – 50 ). Increased vitamin E intake has been shown to decrease DNA damage induced by polyunsaturated fatty acids ( 51 ), and vitamin C, vitamin E and β-carotene supplementation decreased endogenous and H ₂ O ₂ -induced DNA damage in human lymphocytes ( 52 ). Selenium has also been shown to protect cells against DNA damage and to induce DNA repair in response to damage ( 53 , 54 ).

We also considered the second hypothesis that tumor-associated factors may modulate H ₂ O ₂ -induced DNA damage. For example, two related tumor-associated factors, transforming growth factor beta and interleukin-6, are targeted because both play roles in DNA damage/repair in addition to their other functions ( 55 , 56 ), and their plasma levels decrease 6–8 weeks after prostatectomy ( 57 ). Removing these tumor-associated factors may lower DNA damage levels in prevalent cases in addition to dietary/lifestyle change. Therefore, whether higher DNA damage in incident cases reflects CaP susceptibility and/or tumor status is unclear. In summary, higher dietary antioxidant intakes and/or removal of tumor-associated factors may contribute to lower DNA damage in prevalent cases. Future follow-up studies are warranted to further examine these two possibilities and the potential application of the Comet assay in CaP susceptibility and prevention.

The strengths of this study include a well-controlled sample processing procedure (blood was processed within 2 h after phlebotomy) and storage (cryopreserved lymphocytes had >90–95% viability), extensive laboratory quality-control programs, and objective comet image analysis. Considering its limitation, peripheral lymphocytes are most often used, because they are easily obtained from blood, but whether they reflect DNA damage/repair in the target tissue must be evaluated. One study reported that comet-tail lengths in cervical epithelial cells and peripheral blood leukocytes significantly differed between cancer patients and controls ( 18 ). In addition, normal adjacent breast tissue contained higher levels of DNA adducts, and altered DNA repair was observed in cultured skin fibroblasts of breast cancer patients ( 58 , 59 ). Results from these studies suggest that genetic defects in DNA repair may contribute to higher levels of DNA damage in lymphocytes and target tissue in cancer patients. Since obtaining normal primary prostate epithelial cells from controls is generally not feasible, we are currently evaluating DNA damage levels by Comet assay in lymphocytes with corresponding histologically normal primary prostate cells from CaP cases.

With a limited sample size, our current data suggest that elevated DNA damage (alkaline Comet assay) may be associated with CaP. Larger case–control and follow-up studies are warranted to further test the potential application of the alkaline Comet assay in CaP risk assessment and prevention.

The authors are grateful to study participants. We also want to acknowledge the contributions of Frank M.Torti, MD; Robert Lee, MD; Charles J.Rosser, MD; Dean G.Assimos, MD; Elizabeth Albertson, MD; Dominck J.Carbone, MD; William Rice, MD; Francis O'Brien, MD; Ray Morrow, MD; Franklyn Millman, MD; Nadine Shelton, Joel Anderson, Shirley Cothren, Eunkyung Chang, the General Clinical Research Center, the Urology Clinic and the Internal Medicine Clinic. This study was supported by a grant from the American Cancer Society (No. CNE-101119 to J.J.H.), a pilot grant from the Comprehensive Cancer Center of Wake Forest University (CA12197 to J.J.H.) and a grant from the National Research Foundation to the Wake Forest University's General Clinical Research Center (M01-RR07122).

Conflict of Interest Statement : None declared.

References

Author notes

1Department of Cancer Biology, 2Department of Urology and 3Comprehensive Cancer Center, Wake Forest University School of Medicine, Winston-Salem, NC 27157, USA, 4School of Public Health and 5Stanley S. Scott Cancer Center, Louisiana State University Health Sciences Center, New Orleans, LA 70112, USA