High Excretion of Etheno Adducts in Liver Fluke–Infected Patients: Protection by Praziquantel against DNA Damage

Chronic infection by the liver fluke (Opisthorchis viverrini; OV) is a strong risk factor for developing cholangiocarcinoma (CCA). This tumor, which arises from infected bile duct epithelial cells, is a major public health problem in northeast Thailand (1). Persistent cellular oxidative/nitrative stress and enhanced lipid peroxidation (LPO), leading to macromolecular damage and disruption of signaling pathways, are implicated in the development of human malignancies associated with chronic inflammatory processes (2-9). The oxidation of lipids by reactive oxygen and nitrogen species results in byproducts such as trans-4-hydroxy-2-nonenal (HNE), malondialdehyde (MDA), and crotonaldehyde. These intermediates can react with DNA bases to form exocyclic DNA adducts (7, 10-12). Reaction of DNA bases with the major LPO end-product HNE yields inter alia the etheno-DNA adducts 1,N⁶-etheno-2′-deoxyadenosine (εdA) and 3,N⁴-etheno-2′-deoxycytidine (εdC). Because LPO products, especially HNE, are increasingly implicated in carcinogenesis (3, 13-16), εdA and εdC appear to be promising tools for quantifying promutagenic DNA damage in early, premalignant stages of the carcinogenesis process. We have explored previously the usefulness of etheno-DNA adducts measured either in tissues or excreted in urine and in cancer etiology and clinical studies (reviewed in ref. 17). We could show that εdA and εdC may serve as potential risk markers for cancer-prone diseases caused by metal storage, alcohol abuse, viral infections, chronic inflammatory processes such as chronic pancreatitis, and inflammatory bowel diseases (18, 19).

The aim of the present study was to clarify whether chronic OV infection can induce LPO-derived DNA damage due to oxidative/nitrative stress in vivo in infested Thai patients. As noninvasive indicators, we measured the urinary excretion of two LPO-derived etheno deoxyribonucleosides. εdA and εdC levels were quantified in collected urine samples by immunoprecipitation/ high-performance liquid chromatography (HPLC)/fluorescence detection and ³²P-postlabeling TLC, respectively. εdA and εdC in urine are thought to derive mainly from adducted DNA via repair processes in affected tissues and possibly also from adduct formation in the deoxyribonucleoside pool or DNA fragmentation (Fig. 1). Excreted etheno adducts were related to oxidative/nitrative stress-induced MDA and nitrate/nitrite levels in urine. Plasma alkaline phosphatase (ALP) activity was determined, and the relationship between etheno adducts and degree of liver damage was investigated. Another aim was to verify whether a single dose of Praziquantel, known very effective antiparasitic drug when administered to OV-infected subjects, could lower urinary εdA and εdC levels, concomitantly with those of MDA, nitrate/nitrite, and ALP activity.

The study was done in healthy volunteers (n = 20; 9 males and 11 females) and OV-infected subjects (n = 50; 26 males and 24 females). Healthy subjects (age range, 35-59 years; mean ± SE, 47.55 ± 1.74 years) were defined as persons who live in a nonendemic area of OV infection and showed negative results (both at present and in the past) for OV egg counts in feces. OV-infected subjects (age range, 28-78 years; mean ± SE, 53.03 ± 2.04 years) were persons who lived in an endemic area (Phuvieng District near Ubonrat Dam, Khon Kaen Province, northeast Thailand) and had a positive OV egg count in feces. Persons with acute infections showing nitrate/nitrite and leukocyte positivity in urine were excluded; patients with chronic inflammatory conditions caused by hepatitis B virus and TB infections and those with diabetic mellitus were also omitted. An informed consent form was signed by each study subject. The study protocol was approved by the Ethics Group of the Human Research Committee (HE 480316), Khon Kaen University, Thailand.

Morning urine samples (30 mL) were collected and immediately centrifuged at 3,000 rpm at 4°C for 10 min and then transferred into stock tubes. Samples were stored at −80°C for up to 6 months before analysis.

Peripheral 10 mL blood samples were obtained by sterilized venipuncture, collected in tubes containing acid citrate/dextrose, and centrifuged at 2,000 rpm at 4°C for 15 min; plasma was then isolated and was stored at −80°C until use.

Briefly, urine samples were passed through a 0.22 μm filter, and a 2 mL sample was spiked with [2,8-³H]1,N⁶-ethenoadenosine used as internal standard. The urinary protein was precipitated by adding cold absolute ethanol and centrifuged at 4,000 rpm for 10 min after standing at −20°C for 30 min; the supernatant was concentrated by vacuum centrifugation. The dried sample was redissolved in 1 mL water and loaded onto a LC-18 solid-phase silica column (500 mg, 3 mL; particle size, 45 μm; pore size, 0.8 cm³/g pore volume; Supelco). The columns were washed with 10 mL of 20 mmol/L NaH₂PO₄ (pH 6.8; prepared by mixing of 20 mmol/L Na₂HPO₄·2H₂O and 20 mmol/L NaH₂PO₄·H₂O; pH was adjusted by either solution) followed by 10 mL water and 10 mL of 10% methanol (v/v) at 4°C to remove the bulk of normal nucleosides. The columns were brought to room temperature (20°C) and εdA and εdC were eluted twice with 2.5 mL methanol/water (1:1, v/v) and concentrated by vacuum centrifugation. Before sample loading, the columns were prewashed with 10 mL absolute methanol followed by 10 mL water. The dried samples were resolved by HPLC on a reverse-phase Hypersil ODS column (250 × 8.0 mm, 5 μm) connected to an automated fraction collector (Agilent 1100 series) using a linear water (A)/acetronitrile (B) gradient as follows: 0-8 min, 100% A; 8-20 min, 0-20% B in A; 20-30 min, 20% B isocratically; 30-40 min, 20-50% B in A; 40-45 min, 50% B isocratically; and 45-50 min, 100% B. The flow rate was 2 mL/min. Fractions containing the internal standard (19.6-20.6 min retention time) and εdA and εdC (20.7-21.9 min retention time) were collected separately. The radioactivity of the earlier eluting internal standard fraction was counted for the correction of recovery. The fractions containing εdA and εdC were concentrated in vacuo. The dried samples were halved for εdA and εdC analysis. The HPLC system was cleaned after each sample analysis, and blank samples were randomly run to ensure that no residue contamination was present.

The εdA levels were measured in urine samples by a modified immunoprecipitation/HPLC/fluorescence detection method (20). Briefly, the dried samples from the preparation step were used for immunoprecipitation, which was done in Tris-HCl buffer [10 mmol/L Tris (pH 7.5), 140 mmol/L NaCl, 3 mmol/L NaN₃] containing 1% bovine serum albumin and 0.1% rabbit IgG (Sigma-Aldrich) and monoclonal antibody EM A-1 (provided by Dr. P. Lorenz, Institute for Cell Biology, University of Essen). The antigen-antibody complex was precipitated with saturated ammonium sulfate. The precipitate was washed with 10% saturated ammonium sulfate/water and εdA eluted using 50% methanol/water and later concentrated in vacuo. The residue was analyzed by a HPLC system consisting of a HP1100 pump, HP1046A fluorescence detector (Hewlett-Packard) and 250 × 4.6 mm Lichrospher 100 RP 18E 5 μm (Bischoff), using a linear gradient orthophosphate [17 mmol/L (NH₄)₃PO₄ (pH 5), buffer/methanol 9:1 to 8:2 in 30 min at a flow rate of 1 mL/min]. The εdA (18 min retention time) peak was detected at excitation λ = 230 nm and emission λ = 410 nm and quantified by a standard peak using εdA standard for one concentration of 25 fmol/injection (Sigma-Aldrich). The recovery of this step was also determined by adding the internal standard before starting the precipitation of the sample. The on-column detection limit of εdA in this HPLC system was ∼5 fmol εdA/injection. Urinary creatinine concentration for normalizing the εdA and εdC data was measured by a picric acid–based method using a kit provided by Sigma (Sigma-Aldrich) according to the supplier's protocol.

The dried samples were subjected to ³²P-postlabeling and TLC analysis of εdC according to a previously described method that was slightly modified (21). In brief, after addition of 5 μL kinase buffer [125 mmol/L Tris-HCl, 25 mmol/L MgCl₂, and 25 mmol/L DTT (pH 6.8)], 2 μL [γ-³²P]ATP (10 μCi), and 0.5 μg Dm-deoxynucleoside kinase (kindly provided by Anna Karlsson, Karolinska Institute), the mixture was incubated at 37°C for 2 h, centrifuged, and spotted onto the left side of prewashed polyethyleneimine-cellulose plates at 1.5 cm from the margins. The polyethyleneimine-cellulose plates were developed by two-dimensional TLC [D1: 1 mol/L acetic acid (pH 3.5); D2: saturated ammonium sulfate (pH 3.5)]. The plates were exposed to X-ray films (Fuji medical X-ray film, 100 NIF, 18 × 20 cm) in cassettes supported with intensifying screens at −80°C for 3 h. Sections of the TLC plates corresponding to the spots of the [³²P]εdC-5′-monophosphate and [³²P]BrdU-5′-monophosohate (used as internal standard) were cut out and radioactivity was determined by liquid scintillation counting. Background counts were determined from appropriate blank areas on TLC plates and subtracted. A total of 1 fmol εdC standard was also labeled in parallel and separated in a similar way. The detection limit for εdC was 0.1 fmol detectable in 500 μL human urine.

Urinary MDA was measured by the thiobarbituric acid–based method using a kit provided by Cayman Chemical according to the supplier's protocol. Urinary nitrate/nitrite was measured by a simple Griess-based method using a kit provided by Cayman Chemical according to the supplier's protocol.

Plasma ALP activity, a marker of hepatobiliary tract damage, was analyzed by a standard automated spectrophotometer (Automate RA100) using a commercial kit (Thermo Trace) according to the supplier's protocol.

Statistical analysis of the data was done with the StatView 5.0 program (SAS Institute). The comparison between OV-infected and healthy control group was conducted using the Mann Whitney U test between OV-infected before and after treatment with Praziquantel was conducted using the Wilcoxon's rank test, and the correlations between variables were determined using the Spearman's rank test.

To clarify the role of oxidative stress and LPO-derived DNA damage in the pathogenesis of opisthorchiasis-associated CCA, urinary excretion of two LPO-derived etheno DNA adducts was measured in 20 healthy volunteers and 50 OV-infected Thai subjects. εdA and εdC are thought to be excreted following DNA repair and other processes probably occurring in the infected hepatobiliary tract (Fig. 1). εdA and εdC were quantified in collected urine samples by immunoprecipitation/HPLC/fluorescence detection and ³²P-postlabeling-TLC, respectively. These ultrasensitive and specific detection methods were developed in our laboratory (20, 21) to permit application in human biomonitoring (7, 22). Both methods allow the detection of background adduct levels in urine of asymptomatic subjects; thus, any disease-related increase can be reliably determined.

Figure 2A to E show box plots and mean levels of urinary εdA, εdC, MDA, and nitrate/nitrite and plasma ALP activity in OV-infected patients compared with healthy subjects. Mean εdA and εdC levels in urine (fmol/μmol creatinine) were significantly 3 to 4 times higher in urine of OV-infected patients than in healthy controls (εdA, 22.70 ± 3.68 versus 5.16 ± 1.33, P < 0.001; εdC, 6.53 ± 2.30 versus 1.95 ± 0.44, P < 0.01). Huge 200- to 300-fold interindividual variations for εdA and εdC excretion were found in OV-infected subjects. As expected, the inflammatory indicators, MDA, nitrate/nitrite, and ALP, were concomitantly increased by OV infection: higher levels of MDA in urine (μmol/g creatinine) were found (2.07 ± 0.25 versus 1.45 ± 0.10, P < 0.05) and urinary nitrate/nitrite (μmol/g creatinine) was elevated in OV-infected subjects (294.06 ± 23.80 versus 166.52 ± 18.25, P < 0.01). The plasma ALP activity in OV-infected subjects was significantly higher than in healthy controls (44.91 ± 1.47 versus 27.85 ± 2.75 units/L, P < 0.0001).

A comparison of urinary εdA, εdC, MDA, and nitrate/nitrite levels and plasma ALP activity in OV-infected subjects before and 2 months after treatment by a single dose (40 mg/kg body weight) of the antiparasitic drug Praziquantel is shown in Table 1. εdA levels were significantly lowered by OV elimination (22.70 ± 3.68 versus 17.42 ± 2.98, P < 0.05) but remained still significantly higher than in healthy controls (5.16 ± 1.33 fmol/μmol creatinine, P < 0.001). There was a nonsignificant decrease of urinary εdC levels after Praziquantel treatment (6.53 ± 2.30 versus 3.52 ± 0.70, P = 0.08), which was also still significantly higher than in healthy controls (1.95 ± 0.44 fmol/μmol creatinine, P < 0.01). Praziquantel treatment of OV-infected patients significantly decreased urinary levels of MDA (2.07 ± 0.25 versus 1.59 ± 0.25, P < 0.001), urinary nitrate/nitrite (294.06 ± 23.80 versus 271.57 ± 25.92, P < 0.05), and plasma ALP activity (44.91 ± 1.47 versus 41.98 ± 1.60, P < 0.05). However, after treatment, urinary nitrate/nitrite (P < 0.01) and plasma ALP (P < 0.0001) still remained higher than in controls; for the latter, values were within the normal clinical range.

The correlations between variables measured in urine of OV-infected subjects (before and after treatment with Praziquantel) and untreated healthy controls all combined are shown in Table 2. There was a significant positive correlation between urinary εdA and εdC levels (r = 0.31, P < 0.005). Urinary εdA levels were significantly correlated with urinary MDA and with plasma ALP activity; no such correlations were found for urinary εdC. For all investigated variables, neither gender-related nor age-related differences were apparent; also, egg counts in feces were not correlated (data not shown).

Chronic infection by the liver fluke OV is a strong risk factor for developing CCA and thus a major health problem in northeast Thailand (risk ratio = 5; ref. 23). To clarify the role of oxidative/nitrative stress and LPO-derived DNA damage in opisthorchiasis-associated CCA, excretion of two LPO-derived etheno DNA adducts was measured in urine samples collected from healthy volunteers and OV-infested Thai subjects before and after treatment by the antiparasitic drug Praziquantel. εdA and εdC levels were quantified by ultrasensitive and specific methods developed in our laboratory (immunoprecipitation/HPLC/fluorescence detection and by ³²P-postlabeling-TLC, respectively). Excreted etheno adducts were correlated with oxidative/nitrative stress-induced MDA and nitrate/nitrite levels in urine and with plasma ALP activity. Mean εdA and εdC levels were measured in the same urine sample and found to be 3 to 4 times higher in OV-infected patients than in healthy controls. Three indicators for inflammatory conditions, MDA, nitrate/nitrite, and ALP, were also increased up to 2-fold by OV infection and were (except nitrate/nitrite) positively related to εdA excretion.

Our results showing a significant increase of urinary etheno adduct levels in OV-infected subjects and a significant reduction after treatment by the antiparasitic Praziquantel provide strong support that OV-related infection and inflammation are key events for causing LPO-derived DNA damage in vivo. This assumption is further supported by our previous results showing that OV infection in hamsters also induced other types of DNA damage (8-nitroguanine and 8-oxo-deoxyguanosine in bile duct epithelial cells; ref. 24). Such DNA damage in the hepatobiliary tract is likely to contribute to the development of CCA in OV-infected patients (25). Praziquantel as shown previously by us prevents OV-induced CCA not only by elimination of parasites but also by its anti-inflammatory effect by inhibiting inducible nitric oxide synthase–dependent DNA damage (26). We conclude that the reduction by Praziquantel of urinary LPO-derived etheno-DNA adducts and the inflammatory indicators (MDA, nitrate/nitrite, and ALP) is caused by its suppression of oxidative/nitrative stress, acting together with its known antiparasitic effectiveness (Fig. 1). Three decades ago, we were the first to show that Praziquantel is devoid of genotoxic activity, which encouraged the pharmaceutical companies (Merck, Bayer) to obtain approval for its use as an antihelminthic drug in humans (27).

Mechanical injury from migrating flukes contributes to biliary damage in the human host (28). Also, the host immune response and immunopathologic processes are thought to mediate hepatobiliary damage in opisthorchiasis (29). In agreement with these mechanisms, in our study, urinary εdA levels were positively associated with plasma ALP activity, which was significantly higher in OV-infected patients than in controls. We conclude that inflammation-induced LPO together with mechanical injury causes damage in target hepatobiliary cells acting as a driving force to malignancy.

The lack of a positive correlation between urinary nitrate/nitrite and etheno adducts may be due to uncontrolled confounders (e.g. dietary intake, stress condition, etc.), which are known to affect urinary nitrate/nitrite levels. Nevertheless, nitric oxide via peroxynitrite has been shown to induce LPO and produced LPO-derived etheno-DNA adducts as shown in a mouse model for nitric oxide overproduction and inflammation-related DNA damage (30, 31). Moreover, we have shown previously that, in OV-infected hamsters, nitric oxide–mediated DNA damage is implicated in cholangiocarcinogenesis (32).

Although direct proof does not exist, εdA and εdC in urine are thought to derive mainly from adducts formed in DNA in the affected hepatobiliary tract via repair processes. Other sources (e.g., from adduct formation in the deoxyribonucleoside pool and/or by DNA fragmentation after apoptosis; Fig. 1) cannot be ruled out. Alkyl-N-purine-DNA glycosylase and AlkB protein remove εdA, the mismatch-specific thymidine-DNA glycosylase, and AlkB protein removes εdC. Previous studies showed that εdA adducts in cellular DNA in vitro are repaired more quickly by base excision repair than are εdC adducts (33-35). Hence, all three (thus far) known repair pathways by glycosylases and AlkB do not yield etheno-deoxyribonucleosides as excised products. Recent work indicates that another exocyclic HNE-derived DNA adduct (HNE-dG) is formed in human and animal tissues in addition to the simple etheno adducts, εdA and εdC (36). Nucleotide excision repair was found to be a major pathway for eliminating the HNE-dG adduct in both human and Escherichia coli cells (37, 38). We therefore hypothesize that during (long-patch) nucleotide excision repair, εdA and εdC may be coeliminated and end up as etheno-deoxyribonucleoside excision products excreted in urine. Experimental proof for this assumption is now required.

There is growing support that etheno-DNA adducts both in target tissue and excreted in urine appear to be useful markers for assessing oxidative/nitrative stress-derived DNA damage in early, premalignant stages of the carcinogenesis process. Thus, they could serve as potential biomarkers for the progression of cancer-prone diseases, especially those that have an inflammatory component in their etiopathogenesis (7, 18). εdA excreted in urine was massively increased in patients who had developed chronic hepatitis, cirrhosis, and hepatocellular carcinoma due to hepatitis B virus infection (17). In Japanese women, urinary εdA levels were positively associated with NaCl excretion, suggesting salt-induced inflammation and LPO to occur in the stomach in conjunction with Helicobacter pylori infection (20). Recently, we have shown an increase 9- to 13-fold excretion of urinary etheno adducts in β-thalassemia patients who are due to an iron overload at increased risk for developing hepatocellular carcinoma (39).

We conclude that chronic OV infection through oxidative/nitrative stress leads to massive urinary excretion of the etheno-bridged deoxyribonucleosides, reflecting a high LPO-derived DNA damage in vivo. These promutagenic etheno-DNA adducts, together with other lesions in bile duct epithelial cells, may increase the risk of OV-infected patients to develop CCA at a later stage. A relationship between marker and disease causation is further supported by the protective effect of the antiparasitic drug Praziquantel against LPO-derived DNA damage and CCA. Urinary εdA and εdC levels should be explored (a) as noninvasive risk markers for developing opisthorchiasis-related CCA and (b) as promising biomarkers to assess the efficacy of preventive and therapeutic interventions.

No potential conflicts of interest were disclosed.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank all volunteers who participated in this study, Susanna Fuladdjusch for secretarial help, and Robert W. Owen for editorial assistance.

This article is dedicated to Jagadeesan Nair who died prematurely in 2007.