Monitoring in vivo metabolism and elimination of the endogenous DNA adduct, M₁dG (3-(2-deoxy-β-D-erythro-pentofuranosyl)pyrimido[1,2-α]purin-10(3H)-one), by accelerator mass spectrometry

Abstract

Our laboratory is investigating the in vitro and in vivo metabolic processing of endogenously formed DNA adducts as a means of evaluating candidate urinary biomarkers. In particular, we have focused our studies on the metabolism and disposition of the peroxidation-derived pyrimidopurinone deoxyguanosine (dG) adduct, 3-(2-deoxy-β-D-erythro-pentofuranosyl)pyrimido[1,2-α]purin-10(3H)-one (M₁dG) and its principal metabolite, 6-oxo-M₁dG. We now report the metabolic processing of M₁dG at concentrations 4 to 8 orders of magnitude lower in concentration than previously analyzed, by the use of accelerator mass spectrometry analysis. Administration of 2.0 nCi/kg [¹⁴C]M₁dG resulted in 49% of the ¹⁴C recovered in urine whereas 51% was recovered in feces. In urine samples, approximately 40% of the ¹⁴C corresponded to the metabolite, 6-oxo-M₁dG. Following iv administration of 0.5 and 54 pCi/kg [¹⁴C]M₁dG, approximately 25% of the urinary recovery corresponded to the metabolite, 6-oxo-M₁dG. Thus, upon administration of trace amounts of M₁dG, a significant percentage of 6-oxo-M₁dG was produced suggesting that 6-oxo-M₁dG maybe a useful urinary marker of exposure to endogenous oxidative damage.

Introduction

Oxidative damage to DNA is considered a contributing factor to the onset of several genetic diseases including cancer (1, 2). Monitoring genomic exposure to oxidative damage may be useful in risk assessment studies (3, 4). Endogenously formed electrophiles arise from enzymatic and non-enzymatic sources and are involved in the etiology of disease. The principal mediators of endogenous damage are radicals. Radicals can directly damage DNA and other cellular macromolecules to disrupt native structure and function. Radical reactions (enzymatic and non-enzymatic) with polyunsaturated lipids are a precursor to the production of several highly reactive lipid hydroperoxides, 9- and 13-hydroperoxyoctadecadienoic acid, and aldehydic electrophiles, such as malondialdehyde, 4-hydroxy-nonenal, 4-oxo-nonenal, and others (5). When radicals or aldehydes react with DNA, modifications result that can lead to mutation. DNA adducts formed from endogenous sources are generally considered endpoints of reaction; thus, monitoring their formation may provide a useful index of genomic exposure to oxidative damage.

M₁dG is a major endogenous, peroxidation-derived adduct of dG formed in reactions with either malondialdehyde (a product of lipid peroxidation, DNA peroxidation, and prostaglandin synthesis) or base propenal (a product of DNA peroxidation) (6–9). M₁dG is a substrate for the nucleotide excision repair pathway, which accounts for its presence in human urine as the deoxynucleoside (10, 11). Urinary elimination of DNA adducts is a common approach used to assess exposure to genetic insult from endogenous sources. Several laboratories have implemented strategies for the non-invasive quantification of endogenous adducts (12–16). An increasing body of evidence suggests that factors affecting the appearance of DNA adducts in urine (i.e., metabolism, biliary elimination, etc.) may limit their direct detection. The metabolic stability of M₁dG has been the subject of several investigations (17–19).

Results from in vitro and in vivo metabolism experiments have demonstrated that M₁dG is subject to enzymatic oxidation to a single metabolite, 6-oxo-M₁dG (17,19). In vivo metabolism and elimination studies were performed with µCi quantities of carbon-14 labeled M₁dG ([¹⁴C]M₁dG). Following iv administration of [¹⁴C]M₁dG, 45% of the isotope was recovered as the metabolite, 6-oxo-M₁dG, with the balance of radioactivity attributed to [¹⁴C]M₁dG (19). No other significant metabolites were observed in bile or urine samples. 6-Oxo-M₁dG accounted for 20% of the total radioactivity in the early time-point urine samples; however, at late time-points, there was an approximate equal ratio of M₁dG to 6-oxo-M₁dG. Based solely on the in vitro determined K_m for M₁dG metabolism in rat liver cytosol (K_m = 370 µM) this observed increase in metabolite production may not be expected (17). Additionally, since endogenously formed adducts are produced in very low amounts, we were curious to know if M₁dG would be metabolized when administered at very low quantities that might approximate physiological levels. To address this question, we employed nCi and pCi dosing procedures with [¹⁴C]M₁dG and monitored metabolism and elimination. Analysis of biological samples from these studies was possible by the use of accelerator mass spectrometry (AMS).

Experimental Procedures

Materials

All chemicals were obtained from commercial sources and used as received. Solvents were of HPLC grade purity or higher. [8-¹⁴C]2´-deoxyguanosine was purchased from Sigma (St. Louis, MO). [¹⁴C]M₁dG was synthesized as previously described (20). HPLC-UV separations were performed with either a Waters 2695 autosampler and binary pump or a Waters 1585 binary pump and 717plus autosampler, and a Waters 2487 dual wavelength UV detector (Milford, MA).

Administration of [¹⁴C]M₁dG to Rats

Animal protocols were performed under approval of Vanderbilt University and in accordance with the Institutional Animal Care and Use Committee policies. Male Sprague Dawley rats (225–250 g), with vascular catheters surgically implanted in the jugular vein were obtained from Charles River Laboratories (Wilmington, MA) and housed in shoebox cages. Animals were transferred to metabolism cages prior to dosing and allowed to feed ad libitum throughout the experiment. The [¹⁴C]M₁dG dosing solution (>99% pure) was prepared in sterile saline to the desired specific activity (3.0 dpm/µL, 0.1 dpm/µL and 0.001 dpm/µL) and administered via the jugular catheter in approximately 0.3 mL of total volume over 45 s. All doses were followed by 0.4 mL of sterile heparinized saline to flush the catheters.

One cohort of 4 animals was used for the 2.0 nCi/kg (11 ng/kg) dosing. Three animals received the [¹⁴C]M₁dG dosing solution (3.0 dpm/µL), and 1 animal received a vehicle control (sterile saline). Urine was collected over intervals (pre-dose, 0–4, 4–8, 8–12, 12–16, 16–20 and 20–24 h). All samples were stored at −20°C following collection.

A second cohort of 4 animals was used for the 0.5 pCi/kg (3.0 pg/kg) and 54 pCi/kg (300 pg/kg) studies. Three animals received the 0.5 pCi/kg [¹⁴C]M₁dG dosing solution (0.001 dpm/µL), and 1 animal received the vehicle control (saline). After 12 h, the animals were removed from the metabolism cages and returned to shoebox cages overnight. Metabolism cages were thoroughly washed following the 0.5 pCi/kg dose. Approximately 24 h after receiving the 0.5 pCi/kg dose, the same 3 experimental animals received 54 pCi/kg [¹⁴C]M₁dG dosing solution (0.1 dpm/µL) and the same control animal received a vehicle administration (saline). Urine was collected over intervals (pre-dose, 0–2 and 2–4, and 4–8). All samples were stored at −20°C. Repeat dosing was implemented in this experiment to minimize the pain and distress of laboratory animals, and to comply with the institutional guidelines for the use and care of animals in research

Fecal samples were dried on the bench top overnight. All pellets from a given time-point were pooled, placed in a sealable plastic bag, crushed with a rubber mallet, and stored at −20°C. For AMS analysis, fecal samples were weighed, diluted in 1 N NaOH, and stored at −20°C.

HPLC Fraction Collection

Urine samples were centrifuged for 10 minutes at 14,000 g. Supernatants were removed and directly injected onto a Phenomenex Luna C18(2) column (4.6 × 250 nm, 5 µm) equilibrated with 90% solvent A (10 mM potassium phosphate in H₂O, pH 7.8) and 10% solvent B (Methanol) at a flow rate of 1.0 mL/min. The solvent was programmed as follows: a linear gradient from the equilibration conditions to 20% B in 10 min, holding at 20% B for 10 min, increasing to 90% B in 0.1 min, holding for 5 min, decreasing to 10% B in 0.1 min and re-equilibrating to initial conditions. Standards (spiked into solvent A) applied to this gradient eluted at 12.7 min (6-oxo-M₁dG) and 15.8 min (M₁dG). Standards spiked into blank urine samples demonstrated slight shifts in retention to 11.5 min (6-oxo-M₁dG) and 15.7 min (M₁dG). The shift in 6-oxo-M₁dG retention time is likely due to ionization effects (pK_a = 7.0, data not reported). Injections (300 µL) were made for the 2.0 nCi/kg study and 200 µL injections were made for the 54 pCi/kg and 0.5 pCi/kg studies. Fractions were collected from 10.5 min to 12.5 min for 6-oxo-M₁dG and from 15 to 17 min for M₁dG. Samples were evaporated to dryness under a stream of nitrogen in a heated water bath and stored at −20 °C.

AMS Analysis

AMS analyses were conducted at the MIT BEAMS Lab using the instrument and procedures described elsewhere (21–24). Urine samples and fecal extracts were diluted as required to bring the total carbon and ¹⁴C concentrations within the dynamic range of the AMS system and analyzed without further processing. HPLC fractions were reconstituted with appropriate measured volumes of H₂O. Aliquots of urine, fecal, and HPLC samples (1.50 µL) were absorbed into pellets of packed CuO powder that had previously been exposed for 30 min to an atmosphere of O₂ at 720 °C. After drying briefly in a vacuum oven, the sample-loaded CuO pellets in their holders were transferred to a laser-induced combustion interface (4) for subsequent AMS analysis. Each run of samples was organized to include quantitation standards (0.0030 dpm/µL) and blanks (1 µg/µL albumin). Quantitation was performed by integrating peaks produced in the continuous trace of ¹⁴C detector count rate vs. time generated during operation of the combustion interface, which produces and delivers the CO₂ of combustion to the AMS ion source, and taking the product of the sample:standard peak area ratio multiplied by the standard concentration as the concentration of the sample. All samples were analyzed in duplicate.

Results

nCi Dosing

[¹⁴C]M₁dG was synthesized as previously described (19, 20), and administered iv to male Sprague Dawley rats (2.0 nCi/kg). Animals were housed in metabolism cages to collect urine and feces at intervals over the course of the experiment. Urine and fecal samples were subject to AMS analysis for total ¹⁴C-recovery (Table 1). Forty-nine percent of the recovered ¹⁴C was found in urine while 51% was found in feces. Approximately 84% of the urinary radioactivity was collected in the first 4 hours and isotope was observed out to 24 h of collection (Figure 1). This elimination profile was very similar to the previously observed time-course from the µCi dosing study (19). No ¹⁴C was detected in any biological samples from the vehicle control animal at amounts measurably above the background levels of ¹⁴C (standard isotopic abundance) present in biological samples

Table 1.

Total ¹⁴C recovery in urine and feces following administration of 2.0 nCi/kg [¹⁴C]M₁dG to male Sprague Dawley rats (n = 3). Expressed as means ± SD.

	AMS Analysis
Dose (dpm)	1027.00 ± 25.51
Urine	584.11 ± 91.10
Feces	603.97 ± 29.01
Recovery (%)	115.70 ± 3.67

Figure 1.

¹⁴C Recovery in urine following administration of 2.0 nCi/kg [¹⁴C]M₁dG to male Sprague Dawley rats (n = 3).

To determine the source of ¹⁴C in the urine, samples were separated by HPLC and fractions corresponding to M₁dG to 6-oxo-M₁dG were collected (Figure 2A) and analyzed by AMS (Figure 2B). The average observed ratio of M₁dG to 6-oxo-M₁dG over the 24-hour collection was 1.4:1, which demonstrated 40% of the urinary radioactivity corresponded to 6-oxo-M₁dG. The mean ¹⁴C-recovery from fraction analysis was 82% with a standard deviation of 12%.

Figure 2.

(A) Reverse phase HPLC separation of M₁dG and 6-oxo-M₁dG standards on a C18 column with a potassium phosphate, pH 7.8, and methanol gradient. (B) Ratio of M₁dG to 6-oxo-M₁dG in rat urine over time following administration of 2.0 nCi/kg [¹⁴C]M₁dG to male Sprague Dawley rats (n = 3).

pCi Dosing

Based on the high ¹⁴C signal-to-noise observed in each urine fraction following administration of 2.0 nCi/kg [¹⁴C]M₁dG, we reasoned the administration of [¹⁴C]M₁dG could be reduced and still be reliably detected. Therefore, we lowered the iv dosing of Sprague Dawley rats to 0.5 and 54 pCi/kg [¹⁴C]M₁dG, which approached our AMS limit of detection for ¹⁴C in urine. Urine was collected at intervals during the experiment to assess ¹⁴C recovery. Total ¹⁴C content in urine samples was assessed directly by AMS analysis. The data from the 54 pCi/kg dosing is summarized in Figure 3: approximately 42% of the administered ¹⁴C was recovered in urine. Direct AMS analysis of urine samples from the 0.5 pCi/kg dose did not offer significant signal above background, and no total urinary ¹⁴C recovery could be determined. Significant ¹⁴C enrichment was observed in all fractions following HPLC purification. Analyte was detected in some urine samples following AMS analysis of the 0.5 pCi/kg samples, albeit at a low signal-to-noise. Only M₁dG and 6-oxo-M₁dG fractions that displayed a signal-to-noise ≥ 3 were reported. The average observed ratio of M₁dG and 6-oxo-M₁dG was approximately 3:1 following administration at either 54 pCi/kg or 0.5 pCi/kg [2-¹⁴C]M₁dG (Table 2). Thus, 25% of the urinary ¹⁴C was recovered as 6-oxo-M₁dG following administration of M₁dG at the AMS limit of detection. The mean ¹⁴C recovery from fraction analysis performed on the 54 pCi/kg urine samples was 55% with a standard deviation of 20%. A similar decrease in recovery was observed with late time-point urine collections in the nCi dosing study. Recoveries of nearly 86–99% were obtained during the first 8 h of collection and recoveries of 56–78% were obtained during the final 8 h of collection following administration of 2.0 nCi/kg M₁dG. Thus, due to the very low analyte levels, the lower recoveries in the pCi dosing studies were attributed to non-specific loss.

Figure 3.

¹⁴C Recovery in urine following administration of 54 pCi/kg [¹⁴C]M₁dG to male Sprague Dawley rats (n = 3).

Table 2.

Percent of [¹⁴C]M₁dG converted to [¹⁴C]6-oxo-M₁dG in urine following administration of 0.5 pCi/kg or 54 pCi/kg [¹⁴C]M₁dG to male Sprague Dawley rats (n = 3). Expressed as means ± SD.

	Collection Interval (h)	Percent 6-oxo-M1dG
54 pCi/kg	0–2	24.5 ± 3.0
	2–4	26.1 ± 2.4
0.5 pCi/kg	0–2	25.4 ± 10.5
Mean & Standard Deviation		25.4 ± 5.9

Discussion

Non-invasive analysis of endogenously occurring DNA damage is of interest for use in clinical and animal studies. Accordingly, several laboratories have developed analytical methodologies to assess adduct levels in biological matrixes (12–16). The appearance of adducts in urine is typically attributed to repair-dependent removal from DNA (12, 16, 25). However, it should be noted that urinary analysis is an indirect method of assessment and the true origins of the adducts are unknown. One cannot rule out the possibility that adducts arising from either the nucleotide pool or the biological processing of nucleic acids during apoptotic or necrotic cell death mechanisms likely contribute to the steady-state elimination rates of DNA adducts in urine. Another important and largely unstudied factor is the bioavailability of DNA adducts present in the diet. Clearly, a comprehensive assessment of the origins of endogenous DNA adducts appearing in urine is a prerequisite to their clinical use as biomarkers of endogenously occurring DNA damage.

Independent of the route of formation, adducts are potentially subject to metabolic processing within the tissue of origin. Alternatively, adducts may be released from tissues and traverse through the systemic circulation prior to elimination. It stands to reason that DNA adducts in the circulation would be subject to similar metabolism and elimination pathways as other purine derivatives. Purine-based therapeutics such as acyclovir, penciclovir, zaleplon, and others readily undergo oxidative metabolism and active transport (26–34). Our results on the oxidative metabolism of M₁dG and M₁G suggest metabolism and alternate routes of elimination may also contribute to the complexity of DNA adduct analysis in biological matrixes (18, 19).

The use of ¹⁴C in DNA adduct analysis has been limited, principally due to the poor yields in chemical synthesis and low rates of adduct formation in DNA, which is coupled to the need of sufficient amounts of material for traditional liquid scintillation counting. The advent of AMS analysis has greatly reduced the required isotope levels for experimental analysis, by altering the mode of detection from decay counting to mass analysis (24). Our study suggests that less than 1000 dpm per animal are required for sound in vivo metabolism and elimination studies involved in evaluating candidate urinary biomarkers. Results from this type of analysis provide, at a minimum, recovery and elimination data, and when coupled to fraction collection analysis offers an estimation of percent metabolism of the administered material. Thus, AMS analysis may be applicable to many toxicology problems and significantly impact the development and evaluation of biomarkers.

Incorporation of a stable ¹⁴C tracer into M₁dG for use in vivo in rats has allowed us to investigate the metabolic processing of M₁dG at very low concentrations. Our findings revealed that when [¹⁴C]M₁dG was administered at 2.0 nCi/kg (approximately 2.5 ng per animal) approximately 40% of the urinary ¹⁴C was recovered as 6-oxo-M₁dG. When the administration level of [¹⁴C]M₁dG was further lowered to 0.5 and 54 pCi/kg (approximately 750 fg and 75 pg per animal, respectively) approximately 25% of the urinary ¹⁴C was recovered as 6-oxo-M₁dG. Based only on the calculated K_m for M₁dG metabolism in rat liver cytosol (370 µM) (17), the in vivo oxidation at very low concentrations may not be expected. However, in vivo the high K_m is likely mitigated by the efficiency of enzyme turnover (v_max/K_m = 0.001 min⁻¹ mg rat liver cytosol⁻¹) (17). The reasons for the observed differences in percent conversion at the different dosing levels are unknown. One possibility for the decreased conversion of M₁dG to 6-oxo-M₁dG maybe related to the enzyme efficiency. It is possible that at the lower level of administration (0.5 pCi/kg and 54 pCi/kg) the effective concentration at the site of metabolism was not sufficient to promote turnover at saturating conditions. Thus, a decrease in the percent conversion may be observed at lower levels of abundance.

The observed contribution of 6-oxo-M₁dG to the total M₁dG levels in rat urine approximated our previous estimations based on ¹⁴C decay counting (19). The fact that 25–40% of the administered [¹⁴C]M₁dG was recovered from urine as 6-oxo-M₁dG in the present studies strongly suggests that M₁dG metabolism is likely to occur under physiological conditions of M₁dG production. A prior estimate of the M₁dG elimination rate in human urine demonstrated the adduct was eliminated at 12 fmol/kg/24 h (10). This elimination rate corresponds to approximately 3.6 pg/kg/24 h. Our administration of [¹⁴C]M₁dG at 0.5 and 54 pCi/kg (750 fg and 75 pg, respectively) brackets the anticipated daily physiological clearance of M₁dG. The rate of elimination for M₁dG in the rat has not been determined. However, if we assume the average elimination rate to be similar in rats as observed in humans, then based on the elimination of 30 mL of urine per day, a total clearance of 100 fg may be anticipated in the rat.

It should be noted that our discussion of the percent conversion to 6-oxo-M₁dG following nCi and pCi dosing is limited to urinary analysis. Prior studies have shown that biliary elimination is a key route of excretion for 6-oxo-M₁dG (19). Following iv administration of [¹⁴C]6-oxo-M₁dG to Sprague Dawley rats, [¹⁴C]6-oxo-M₁dG was cleared unchanged in urine and bile; 55–70% of the total dose was eliminated in either bile or feces (19). When [¹⁴C]M₁dG was administered iv to Sprague Dawley rats, approximately 50% of the dose was cleared in bile or feces; the majority of radioactivity (70%) cleared in bile corresponded to [¹⁴C]6-oxo-M₁dG (19). If we apply the same biliary conversion to the nCi and pCi dosing studies then greater than 50% of the administered material in the current studies would have been converted to 6-oxo-M₁dG. This represents a significant portion of the total M₁dG.

Based on these and prior experiments it is clear that iv administered M₁dG is taken-up and likely accumulates in the liver. The physiological and biochemical explanation for this is unknown. However, based on the significant biliary elimination of iv administered M₁dG, it is reasonable to assume that transport proteins are involved in the disposition of M₁dG and 6-oxo-M₁dG into and out of the liver. Ultimately, this process leads to oxidative metabolism of exogenously administered M₁dG, even at concentrations that approach the analytical limit of sensitivity by AMS. Thus, there exists a distinct possibility that oxidation occurs at the projected rates of M₁dG formation in animals and humans. Our laboratory is currently developing analytical methods to probe for the appearance of M₁dG and 6-oxo-M₁dG in human elimination products.

Scheme 1.

Proposed route of metabolic processing for M₁dG. dG, 2´-deoxyguanosine; NER, nucleotide excision repair.

Abbreviations

M₁dG

3-(2-Deoxy-β-D-erythro-pentofuranosyl)pyrimido[1,2-α]purin-10(3H)-one

¹⁴C

Carbon-14

AMS

accelerator mass spectrometry

dG

2´-deoxyguanosine

[¹⁴C]M₁dG

3-(2-Deoxy-β-D-erythro-pentofuranosyl)pyrimido[1,2-α]purin-10(3H)-one-2-¹⁴C