Metabolism and Elimination of the Endogenous DNA Adduct, 3-(2-Deoxy-β-D-erythropentofuranosyl)-pyrimido[1,2-α]purine-10(3H)-one, in the Rat*

Endogenously occurring damage to DNA is a contributing factor to the onset of several genetic diseases, including cancer. Monitoring urinary levels of DNA adducts is one approach to assess genomic exposure to endogenous damage. However, metabolism and alternative routes of elimination have not been considered as factors that may limit the detection of DNA adducts in urine. We recently demonstrated that the peroxidation-derived deoxyguanosine adduct, 3-(2-deoxy-β-d-erythropentofuranosyl)-pyrimido[1,2-α]purine-10(3H)-one (M1dG), is subject to enzymatic oxidation in vivo resulting in the formation of a major metabolite, 6-oxo-M1dG. Based on the administration of [14C]M1dG (22 μCi/kg) to Sprague-Dawley rats (n = 4), we now report that 6-oxo-M1dG is the principal metabolite of M1dG in vivo representing 45% of the total administered dose. When [14C]6-oxo-M1dG was administered to Sprague-Dawley rats, 6-oxo-M1dG was recovered unchanged (>97% stability). These studies also revealed that M1dG and 6-oxo-M1dG are subject to biliary elimination. Additionally, both M1dG and 6-oxo-M1dG exhibited a long residence time following administration (>48 h), and the major species observed in urine at late collections was 6-oxo-M1dG.

Endogenously produced DNA damage is a contributing factor to the progression of several genetic diseases, including cancer (1,2). Furthermore, there is a strong association between chronic inflammation and cancer risk (3). Assessing exposure to endogenously produced electrophiles and oxidants may have an impact in risk assessment (4). Monitoring the levels of DNA adducts in urine is a common approach used to assess exposure to genomic damaging agents. However, factors that may limit the detection of DNA adducts in urine (metabolism, routes of elimination, etc.) have not previously been investigated.
3-(2-Deoxy-␤-D-erythropentofuranosyl)-pyrimido[1,2-␣]purine-10(3H)-one (M 1 dG) 3 is an endogenous pyrimidopurinone adduct formed by reacting deoxyguanosine with malondialdehyde, a product of enzymatic and nonenzymatic lipid peroxidation reactions, or base propenal, a DNA peroxidation product (5)(6)(7)(8)(9)(10)(11). This adduct is mutagenic in bacteria and mammalian cells (12)(13)(14)(15) and is a substrate for nucleotide excision repair (13,16). It is also one of the first endogenously occurring DNA damage products to be detected in DNA of healthy humans (17). Prior attempts to quantify the levels of M 1 dG in human urine demonstrated an excretion rate of 12 fmol/kg/24 h (18). The rate of M 1 dG elimination in rats could not be determined, because the levels were below the limit of detection for the analytical method (19). The low rate of elimination in human populations and the absence of observed material in the urine of rats led us to hypothesize that factors such as metabolism (oxidation, conjugation, etc.) or alternative routes of elimination may limit the appearance of M 1 dG in urine.
We recently demonstrated that M 1 dG is subject to oxidative metabolism in the rat to a principal metabolite, 6-oxo-M 1 dG (20). However, the total recovery and extent of metabolism in these studies could not definitively be determined. To unambiguously monitor the metabolism and elimination of M 1 dG in vivo, we synthesized M 1 dG containing a carbon-14 incorporated into the purine ring for use in animal studies. This tracer allowed us to quantitatively monitor the metabolism and elimination of M 1 dG and its principal metabolite, 6-oxo-M 1 dG, in the rat. The results of this investigation are described herein.

EXPERIMENTAL PROCEDURES
All chemicals were obtained from commercial sources and used as received. Solvents were of HPLC grade purity or higher.   (21). Briefly, [8][9][10][11][12][13][14] C]2Ј-deoxyguanosine ( 14 C-dG) was obtained from Sigma (0.2 mCi, 55 mCi/mmol) as a solution in ethanol/ water (1:1). Iodoacrolein was freshly prepared from ethyl cis 3-iodoacrylate as described previously and dissolved in anhydrous N,N-dimethylformamide (22). Freshly prepared iodoacrolein was analyzed by 1 H NMR and determined to be Ͼ95% pure. 14 C-dG (0.2 mCi, 1 mg) was diluted with excess ethanol and evaporated to dryness under reduced pressure. The remaining residue was dissolved in N,N-dimethylformamide (0.4 ml) in the presence of 1.5 eq of K 2 CO 3 and heated to 65°C. To this solution, 1 eq of freshly prepared iodoacrolein was added each hour over 5 h of reaction (Fig. 1). The resulting solution was evaporated under reduced pressure and dissolved in 1.0 ml of 0.01 M potassium phosphate. The mixture was purified by HPLC using a Phenomenex Luna C18 (2) column (250 ϫ 4.6 mm, 5 m) equilibrated with 100% solvent A (0.01 M potassium phosphate, pH 7.8) at a flow rate of 1.0 ml/min. The following gradient elution was applied: 100% solvent A with 0% solvent B (methanol) for 5 min followed by a linear increase to 20% B over 30 min, holding at 20% B for 5 min and increasing to 85% B in 1 min, holding at 85% B for 5 min, decreasing to 0% B in 1 min, and re-equilibrating at initial conditions for 5 min. Fractions corresponding to M 1 dG (ϳ35 min) were collected, pooled, and concentrated to a minimal volume. Purity was assessed by HPLC-UV with ␤-RAM detection. The radiochemical purity was assessed to be Ͼ95%, with a minor impurity of residual Bile was also collected (from two animals) as outlined above, but the first interval was broken down as follows: 0 -0.5, 0.5-1, 1-2, and 2-4 h. All samples were collected into pre-tared tubes and weighed following the collection interval to determine the mass of sample collected (feces was first dried on the bench top). Samples were stored at Ϫ20°C until analysis.
Radiochemical Analysis of Biological Samples-Biological samples were processed as suggested by PerkinElmer Life Sciences. Urine (0.1-1.0 ml) and bile (0.1-0.3 ml) samples were added directly to 10 ml of Pico-Fluor TM 40 and analyzed by liquid scintillation counting (10-min counts). Fecal samples were dried on the bench top for a minimum of 24 h. The fecal material was broken into small pieces and mixed, and ϳ20 mg of feces (dry weight) was transferred to scintillation vials (n ϭ 4) and re-hydrated with 0.2 ml of water. To each sample 0.5 ml of Soluene-350 was added, mixed, and heated to 40°C for 1.5 h. A 0.5-ml aliquot of isopropyl alcohol was added and swirled, followed by dropwise addition of 0.2 ml of H 2 O 2 . The resulting mixture was heated to 40°C for 2 h and cooled, and 10 ml of Hionic-Fluor TM scintillation mixture was added. Samples were stored in the dark for a minimum of 36 h prior to counting by liquid scintillation counting (10min counts). Urine and bile samples were profiled by HPLC-UV with ␤-RAM detection, and centrifuged at 200,000 ϫ g for 1 h prior to analysis. Injections (0.025-0.3 ml) were made onto a Phenomenex Luna C18 (2) column (250 ϫ 4.6 mm, 5 m) equilibrated with 100% solvent A (0.5% formic acid in H 2 O) at a flow rate of 1.0 ml/min. The following gradient elution was applied: 100% solvent A with 0% solvent B (0.5% formic acid in methanol) for 5 min followed by a linear increase to 20% B over 30 min, holding at 20% B for 5 min, increasing to 85% B in 1 min, holding at 85% B for 5 min, decreasing to 0% B in 1 min, and re-equilibrating at initial conditions for 5 min. Column eluent first passed through the UV detector (254 nm) and onto the ␤-RAM, where it mixed with IN-FLOW TM 2:1 (IN/US Systems) at a ratio of 1:1.1 (HPLC eluent: mixture) and was analyzed in 0.5-ml flow cell with a dwell time of 0.2 s.
LC-MS/MS Analysis of Biological Samples-Tandem mass spectrometry employing selected reaction monitoring was used to verify the molecular ions attributed to the main source or radioactivity in the biological samples. Approximately 500 -1000 dpm were analyzed per injection. Urine and bile samples were diluted to an approximate concentration of 50,000 dpm/ml and injected onto a Phenomenex Luna C18 (2) col-umn (250 ϫ 2.0 mm, 5 m) equilibrated with 100% solvent A (0.5% formic acid in H 2 O) at a flow rate of 0.3 ml/min. The following gradient elution was applied: 100% solvent A with 0% solvent B (0.5% formic acid in methanol) for 5 min followed by a linear increase to 20% B over 30 min, holding at 20% B for 5 min, increasing to 85% B in 1 min, holding at 85% B for 5 min, decreasing to 0% B in 1 min, and re-equilibrating at initial conditions for 5 min. A ThermoElectron Quantum triple-quadrupole instrument with an electrospray source operated in positive ion mode was used during the analysis.
The following instrument parameters were set: spray voltage 4.3 kV; capillary temperature ϭ 250°C; sheath gas ϭ 33 p.s.i.; auxiliary gas ϭ 25; source CID off; collision pressure 1.5 mTorr. [2][3][4][5][6][7][8][9][10][11][12][13][14] C]M 1 dG was synthesized from [8-14 C]2Ј-deoxyguanosine in a reaction with iodoacrolein in the presence of a mild base (K 2 CO 3 ) (Fig. 1) (21). The material was purified by HPLC, and the radiochemical purity was determined to be Ͼ95% (supplemental Fig. S1). A small amount (Ͻ5%) of unreacted starting material ([8-14 C]dG) was observed in the dosing solution. [2-14 C]M 1 dG was diluted in sterile saline solution and administered intravenously (22 Ci/kg, 123 g/kg) to male Sprague-Dawley rats (n ϭ 4) via catheters surgically implanted in the jugular vein. Two animals contained additional catheters surgically implanted into the bile duct. The animals were housed in metabolism cages throughout the duration of the experiment to collect urine, feces, and bile (where applicable). The collected biological samples were analyzed for total radioactivity by liquid scintillation counting, and the data are summarized in Table 1. Approximately 50% of the radioactivity was recovered in the urine of both bile-catheterized and non-bile-catheterized animals. Greater than 90% of the urinary recovery was collected during the first 8 h (Fig. 2A). The remaining radioactivity was recovered in the feces of non-bile-catheterized animals and in the bile of bile-catheterized animals (Fig. 2, B and C).

RESULTS
Urine and bile samples were profiled by HPLC with radiochemical detection. Representative chromatograms from urine and bile are shown in Fig. 3. The urine profiles from both the bile-catheterized and non-bile-catheterized rats were identical. The major metabolite observed in all radiochemical profiles eluted with the authentic retention time for 6-oxo-M 1 dG and accounted for 20% of the radioactivity recovered in urine. The analysis of bile samples revealed that 70% of the radioactivity in bile was attributed to the metabolite 6-oxo-M 1 dG. Selected reaction monitoring analysis of these samples revealed the expected transitions for M 1 dG   Thus, metabolites other than 6-oxo-M 1 dG collectively accounted for Ͻ6% of the total radioactivity. 6-Oxo-M 1 dG accounted for 45% of the total recovered radioactivity. Analysis of late time point urine collections (post 24 h) from both non-bile-catheterized and bile-catheterized animals suggested 6-oxo-M 1 dG was present at the same or higher abundance than M 1 dG (Fig. 6). This was especially apparent in the  of the pyrimido ring with purified bovine xanthine oxidase (Fig. 1). The material was purified by HPLC and exhibited radiochemical purity of Ͼ99% (supplemental Fig. S2). Freshly prepared [2-14 C]6-oxo-M 1 dG was diluted in sterile saline solution and intravenously administered (15 Ci/kg, 83 g/kg) to male Sprague-Dawley rats (n ϭ 4) via the jugular catheter. Two animals also contained catheters surgically implanted into the bile duct. The animals were housed in metabolism cages throughout the duration of the experiment to collect urine, feces, and bile (where available) at intervals. All samples were collected into pre-tared tubes and weighed following collection to determine the mass of sample collected. The biological samples were processed by liquid scintillation counting (Fig. 4), and the results from the elimination data are summarized in Table 2. The nonbile-catheterized rats eliminated 30 Ϯ 5% of the dose in the urine, whereas 70 Ϯ 5% was recovered in the feces. In bilecatheterized animals, 45 Ϯ 7% was recovered in urine, 1% in feces, and 54 Ϯ 2% in bile.
In all HPLC radiochemical profiles, the principal peak observed eluted with the retention time of 6-oxo-M 1 dG (Fig. 5). No significant secondary metabolites of 6-oxo-M 1 dG were observed during any collection interval. Therefore, 6-oxo-M 1 dG was determined to be metabolically stable (Ͼ97%) and not subject to further metabolism in vivo. These data suggest that the minor metabolites observed in the [2-14 C]M 1 dG study arose directly from M 1 dG and did not represent further metabolism of 6-oxo-M 1 dG.
[2-14 C]6-Oxo-M 1 dG also exhibited significant biliary clearance (as seen in the [2-14 C]M 1 dG study), which suggested 6-oxo-M 1 dG was absorbed by the liver and subject to transport into bile. As seen in the [2-14 C]M 1 dG study, [2-14 C]6-oxo-M 1 dG also exhibited a long residence time (Ͼ48 h) following intravenous administration. The identity of 6-oxo-M 1 dG was verified by selected reaction monitoring analysis of the late time point urine collections (Fig. 6E).

DISCUSSION
The endogenous peroxidation-derived DNA adduct, M 1 dG, was recently found to undergo metabolism in vivo to a single oxidized metabolite, 6-oxo-M 1 dG (20). Our laboratory has now performed definitive in vivo analysis of the metabolism and excretion of M 1 dG and its major metabolite 6-oxo-M 1 dG by the incorporation of stable carbon-14 tracers into these molecules. Male Sprague-Dawley rats with (n ϭ 2) and without (n ϭ 2) bile catheters were intravenously dosed with [2-14 C]M 1 dG at 22 Ci/kg (123 g/kg). Approximately 50% of the administered dose was recovered in the urine, whereas the remainder was recovered in the feces and bile (Fig. 7). 6-Oxo-M 1 dG accounted for 45% of the recovered radioactivity. The significant conversion of M 1 dG to 6-oxo-M 1 dG prompted additional metabolism and elimination studies on the metabolite. Upon dosing [2-14 C]6-oxo-M 1 dG (15 Ci/ kg, 83 g/kg), 30 Ϯ 6% of the radioactivity was recovered in the urine, whereas 71 Ϯ 6% was excreted in the feces of non-bile-catheterized animals. Rats containing bile catheters deposited 45 Ϯ 7% of the radioactivity into the urine, 54 Ϯ 2% into the bile, and 1% into the feces. Profiling bile and urine samples revealed the metabolite, [2-14 C]6-oxo-M 1 dG, was cleared unchanged (Ͼ97%), with no significant metabolism to secondary metabolites.
The simplicity of M 1 dG metabolism and elimination is noteworthy. A single metabolite is formed that is neither conjugated nor further metabolized. In our study, 6-oxo-M 1 dG represents a significant contribution (45% of the total dose) in the overall mass balance of M 1 dG. It was previously hypothesized that depurination reactions may give rise to the free base M 1 G as an alternate metabolic pathway (23); however, no accumulation of this compound or its metabo-  lites was observed in our analysis. Furthermore, our investigation of 6-oxo-M 1 dG metabolism revealed that it was exceptionally stable in vivo (Ͼ97%) and not metabolized to additional products. It should be noted that 6-oxo-M 1 dG contains a possible Michael acceptor on the pyrimido ring (at C 8 ), but no evidence for nucleophilic addition or conjugation (glutathione addition) was observed in the recovered samples.
Considering the main metabolite of M 1 dG is a singly oxidized species, the extent to which M 1 dG and 6-oxo-M 1 dG were cleared through biliary excretion was unexpected. The intravenous administration of M 1 dG and 6-oxo-M 1 dG, and subsequent appearance of these products in bile and feces, strongly suggests the role of transporters in the disposition of M 1 dG and 6-oxo-M 1 dG in vivo. Nucleosides and nucleo-side derivatives are reported substrates for transport proteins, such as organic anion transporters (24) and the concentrative, Na ϩ -dependent (CNT) and equilibrative, Na ϩindependent nucleoside transporters (ENT) (25). Indeed, ENT isoforms appear to be involved in nucleoside transport into hepatocytes (26), whereas the CNT isoforms are expressed along the bile canalicular membrane and are likely involved in biliary transport (27). It would be interesting to evaluate their involvement in DNA adduct disposition in future studies.
The notion that DNA adducts were subject to biliary elimination was first suggested by the work of Wang and Hecht (28). Following an intravenous administration of N 7 ,C-8 N-nitrosopyrrolidine guanine adduct to F344 rats, 52.2% of the administered radioactivity was recovered in the urine.