Urinary excretion of acrylamide and metabolites in Fischer 344 rats and B6C3F1 mice administered a single dose of acrylamide

Toxicology Letters 169 (2007) 34–42

Urinary excretion of acrylamide and metabolites in Fischer 344 rats and B6C3F1 mice administered a single dose of acrylamide Daniel R. Doerge a,∗ , Nathan C. Twaddle a , Melanie I. Boettcher b , L. Patrice McDaniel a , J¨urgen Angerer b a

b

National Center for Toxicological Research, 3900 NCTR Road, Jefferson, AR 72079, United States Institute and Outpatient Clinic of Occupational, Social, and Environmental Medicine, University of Erlangen-Nuremberg, Schillerstrasse 25/29, D-91054 Erlangen, Germany Received 5 October 2006; received in revised form 6 December 2006; accepted 6 December 2006 Available online 16 December 2006

Abstract Acrylamide (AA) is a widely studied industrial chemical that is neurotoxic, mutagenic to somatic and germ cells, and carcinogenic in mice and rats. AA is also formed during cooking in many commonly consumed starchy foods. Our previous toxicokinetic investigations of AA and its genotoxic metabolite, glycidamide (GA), in rodents showed that AA is highly bioavailable from oral routes of administration, is widely distributed to tissues, and that the dietary route, in particular, favors metabolism to GA. Formation and accumulation of mutagenic GA–DNA adducts in many tissues support the hypothesis that AA is carcinogenic in rodent bioassays through metabolism to GA. The current investigation describes the quantification of 24 h urinary metabolites, including free AA and GA and their mercapturic acid conjugates (AAMA and GAMA, respectively), using LC/MS/MS in F344 rats and B6C3F1 mice following a dose of 0.1 mg/kg bw given by intravenous, gavage, and dietary routes of administration. Similar groups of rodents were used previously for serum/tissue toxicokinetic and adduct determinations (DNA and hemoglobin). The goal was to investigate relationships between urinary and circulating biomarkers of exposure, toxicokinetic parameters for AA and GA, and tissue GA–DNA adducts in rodents from single doses of AA. Significant linear correlations were observed between urinary levels of AA with AAMA and GA with GAMA in the current data sets for rats and mice. Concentrations of AA and AAMA correlated significantly with average AUC values determined previously for AA in groups of rats and mice similarly dosed with AA. Urinary GA and GAMA concentrations showed significant correlations with average AUC values for GA and liver GA–DNA adducts determined previously in rats and mice similarly dosed with AA. Despite statistical significance, considerable inter-animal variability was observed in all urinary measurements, which limited the degree of correlation with either average toxicokinetic or biomarker data collected from different groups of animals. These results suggest that urinary measurements of AA and its metabolites may be useful for prediction of internal exposures to AA and GA. © 2007 Elsevier Ireland Ltd. All rights reserved. Keywords: Acrylamide; Glycidamide; Urine; Mass spectrometry; Biomarkers

1. Introduction

∗ Corresponding author. Tel.: +1 870 543 7943; fax: +1 870 543 7720. E-mail address: [email protected] (D.R. Doerge).

Acrylamide (AA, see structure in Fig. 1) is an important industrial chemical with annual worldwide production estimated at >200 million kg. The recent discovery that AA is formed during cooking (Tareke et al.,

0378-4274/$ – see front matter © 2007 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.toxlet.2006.12.002

D.R. Doerge et al. / Toxicology Letters 169 (2007) 34–42

35

Fig. 1. Structures of acrylamide, its metabolites, and selected circulating and urinary biomarkers.

2002) of many common foods (Roach et al., 2003) has renewed interest worldwide in the potential for human toxicity. The basis for concern comes from observations that: AA is neurotoxic in experimental animals (reviewed in LoPachin et al., 2003) and humans (Calleman et al., 1994); its primary oxidative metabolite, glycidamide (GA, see structure in Fig. 1), forms several DNA adducts in vivo (Segerb¨ack et al., 1995; Gamboa da Costa et al., 2003) that accumulate upon repeated exposure (Doerge et al., 2005a); is clastogenic in vivo (Paulsson et al., 2003a; Ghanayem et al., 2005a); is mutagenic in somatic cells in vitro (Besaratinia and Pfeifer, 2004) and in vivo (Manjanatha et al., 2006); is mutagenic in germ cells in vivo (Ghanayem et al., 2005b); and chronic exposure to AA produces tumors in several organs of rats (Johnson et al., 1986; Friedman et al., 1995). Previous investigations of the toxicokinetics of AA and GA in rats and mice focused on use of a single low dose (0.1 mg/kg bw) via intravenous, gavage, and dietary administration (Doerge et al., 2005b,c). These studies demonstrated that AA is highly bioavailable in rats and mice from oral routes of administration and that oral routes favored metabolism to GA, the puta-

tive mutagen and carcinogen. Another previous study showed that levels of GA–DNA adducts in liver and GA–hemoglobin (Hb) adducts were significantly correlated with the AUCs (areas under the time concentration curves) in serum for GA (Tareke et al., 2006). These significant correlations from rodent studies were also used to estimate steady state DNA adduct levels in humans using the limited available steady state GA–Hb adduct data from continual background exposure to AA through the diet (Tareke et al., 2006). Urinary excretion of AA and metabolites also provides another useful and more readily accessible biomarker for evaluating the detoxification of AA and its conversion to GA (Fig. 1). Similar to steady state adduct biomarkers, continual exposure to AA could lead to steady state concentrations of urinary metabolites in urine, albeit dependent on different kinetic parameters. For example, the short half-life reported by Fuhr et al. (2006) of 2.4 ± 0.4 h for AA suggests that this analyte would be less likely to achieve steady state compared to those with longer excretion half-lifes like AAMA (17.4 ± 3.9 h) and GAMA (25.1 ± 6.4 h) relative to a meal-associated dosing rate (3 meals/day). Previous

36

D.R. Doerge et al. / Toxicology Letters 169 (2007) 34–42

studies have evaluated the urinary metabolites of AA in rodents following single gavage doses (50 mg/kg bw; Sumner et al., 1992) and in humans following either chronic consumption through the normal diet and smoking (Boettcher et al., 2005; Bjellaas et al., 2005; Kellert et al., 2006) or acute administration of small doses (1 mg bw, Boettcher et al., 2006; 0.94 mg bw, Fuhr et al., 2006; 0.5–3 mg/kg bw, Fennell et al., 2006). NMR studies have identified and quantified the major metabolites of AA in rodent urine as the mercapturic acids of AA (AAMA) and GA (GAMA) and GA (Sumner et al., 1992, 2003). Glyceramide, the hydrolysis product from GA, has also been identified and quantified using NMR in some, but not all, rodent urine studies (Sumner et al., 1992, 2003; Fennell et al., 2005). By contrast, in human urine, LC/MS/MS methodology has been used to identify and quantify AA, AAMA and GAMA (Fuhr et al., 2006; Fennell et al., 2006); GA, AAMA–sulfoxide and small amounts of AA–cysteine have also been quantified in human urine using LC/MS/MS (Fennell et al., 2006). Comprehensive measurements of biomarkers and toxicokinetics in rodent models should be most useful for extrapolation of AA toxicities observed in laboratory animals to possible effects in humans, whose dietary consumption has been estimated to be less than 1 ␮g/kg bw/day (Boon et al., 2005; JECFA, in press). This paper describes the relationships between urinary biomarkers of AA exposure and previously determined serum toxicokinetic parameters for AA and GA, tissue GA–DNA adducts (Doerge et al., 2005b,c) and Hb adducts of AA and GA (Tareke et al., 2006) in rodents after single dose administrations of AA at 0.1 mg/kg bw. This dose was selected as the lowest possible that would produce blood levels amenable to analysis at peak and 90% elimination times based on analytical sensitivity. The experimental design of the study was to use different routes of exposure to produce a range of internal exposures to AA and GA in male and female rats and mice in conjunction with quantification of the major urinary metabolites (AA, GA, AAMA and GAMA). The goal of this investigation was to determine the suitability of several widely measured human urinary biomarkers for the prediction of internal exposures to AA and GA following exposure to AA through the diet.

Toronto Research Chemicals (North York, Ontario); labeled C3 -AA (99 atom%) was obtained from Cambridge Isotope Laboratories (Andover, MA); 13 C3 -GA (99 atom%, no AA content) was obtained from Toronto Research Chemicals (North York, Ontario). All other solvents were of analytical grade and Milli-Q water was used throughout. AAMA, GAMA and their deuterated analogs were prepared and characterized as previously described (Boettcher and Angerer, 2005). 13

2.2. Animal handling procedures Procedures involving care and handling of rats were reviewed and approved by the National Center for Toxicological Research (NCTR) Laboratory Animal Care and Use Committee using the same experimental design previously reported (Doerge et al., 2005b,c). Briefly, in those studies serum was collected from groups of mice (3 per sex) at different time points or serially from individual rats (5–7 per sex) except for the intravenous exposure, which were collected as described for the mice. In all cases, erythrocytes and liver were collected at the terminal time points. All rodents were obtained as weanlings from the NCTR colony (PND 21) and placed on an irradiated basal diet (5LG6 meal containing approximately 10 ppb AA, Purina Mills Co., St. Louis, MO; Twaddle et al., 2004a) until PND 50 when dosing started. Briefly, groups of male and female B6C3F1 mice (n = 4 per sex) and F344 rats (n = 2 per sex) were exposed to a single dose of AA (0.1 mg/kg bw) using intravenous, gavage, and dietary routes of administration and then individual rats and pairs of male or female mice were placed in plastic metabolic cages for collection of 24 h urine samples on ice. Mice were paired because individuals produced insufficient urine volume to be compatible with the collection apparatus. Total urine volume was measured for each cage in order to calculate total amounts of excreted. Urine was collected similarly from unexposed groups of rodents. All urine samples were kept frozen at −20 ◦ C until analyzed. The AA-fortified diets used for the dietary administration of AA to mice and rats (1.5 and 2.5 ppm, respectively) were prepared and analyzed as previously described (Doerge et al., 2005b,c) and analyzed using isotope dilution LC/MS/MS as previously described (Twaddle et al., 2004a). Actual doses delivered were determined from individual measurements of feed consumption and body weights and ranged between 0.06 and 0.16 mg/kg bw for rats and 0.15–0.65 mg/kg bw for mice. Concentrations of urinary metabolites following dietary administration of AA were normalized to 0.1 mg/kg bw for individual rodents in order to compare with gavage and intravenous groups (i.e., if the measured dose administered in feed to an individual rodent was 0.15 mg/kg, each urinary metabolite concentration was divided by 1.5).

2. Materials and methods 2.3. LC-ES/MS/MS analysis of AA and GA in urine 2.1. Reagents AA (>99.9%) was obtained from Sigma (St. Louis, MO); GA (98.5%, with AA present at approximately 1%) from

Analysis of AA and GA in urine was performed by using a high throughput LC-ES/MS/MS method previously validated for use with rodent serum and tissues (Twaddle et al., 2004b).

D.R. Doerge et al. / Toxicology Letters 169 (2007) 34–42

37

Table 1 Molar percentages of the total dose administered represented by acrylamide and metabolites in 24 h mouse urine Sex/route

AA

GA

AAMA

GAMA

GA/AAa

Totalb

M-IV F-IV M-gavage F-gavage M-diet F-diet

3–5 1–2 0.6–0.7 0.1–1 0 0

14–21 20–31 16–18 12–28 19–49 13–21

28–51 13–16 5–9 5–7 20–31 6–10

10–13 14–25 9–22 6–12 20–21 3–8

0.5–1 2.5–3.2 4–4.8 3.8–5.1 2–2.3 2.9

62–82 60–63 33–48 23–48 60–101 21–39

a b

This ratio represents GA + GAMA/AA + AAMA. This parameter is the total molar percentage of the dose administered represented by the sum of AA + AAMA + GA + GAMA excreted in urine.

Briefly, to each thawed urine sample (10 ␮l) was added labeled internal standards and the samples were diluted to a total volume of 200 ␮l with water. Then the samples were purified using solid phase extraction in 96-well plates and analyzed using LCES/MS/MS in the multiple reaction monitoring mode by using specific transitions for labeled and unlabeled AA and GA. The method was validated using blank rat and mouse urine spiked with AA and GA at 1 ␮M. Method intraday and interday precision and accuracy were in the range of 2–14% and 110–112%, respectively. The limit of detection was 0.01 ␮M for AA and 0.1 ␮M for GA. 2.4. LC-ES/MS/MS analysis of mercapturic acid conjugates in urine Analysis of AAMA and GAMA in urine was performed by using an LC-ES/MS/MS method previously validated for use with human urine (Boettcher and Angerer, 2005). Briefly, to each thawed urine sample (400–800 ␮l) was added labeled internal standards and diluted to a total volume of 4 ml with water. Method intraday and interday precision were in the range of 2–6%. The limit of detection was approximately 0.01 ␮M for AAMA and GAMA. 2.5. Statistical analysis Statistical significance was assessed using linear regression analysis with AUCs as the independent variable and

significant associations were confirmed by performing Spearman rank correlations (SigmaStat). For measurements where assignment of independent and dependent variables was not possible, Spearman rank correlations were used because this technique makes no assumptions about normality of the input data. Group comparisons were made using the two-tailed t-test. In all cases, significance was assumed for p < 0.05.

3. Results 3.1. Urinary levels of AA and metabolites Urinary concentrations of AA and its metabolites were measured in rats and mice dosed identically to those previously used to measure serum toxicokinetic parameters (Doerge et al., 2005a,b) and hemoglobin and DNA adduct biomarkers (Tareke et al., 2006). Measurable concentrations of AAMA and GAMA were observed in all rodents, including those consuming an irradiated diet that contained approximately 10 ppb AA (Twaddle et al., 2004a); by contrast, no detectable levels of AA or GA were observed in control urine. Dosing rodents with AA by intravenous, gavage, or diet at a dose of 0.1 mg/kg bw increased the concentrations in urine of all analytes tested. Total urine volumes were used to calculate the total molar percentages of individual analytes relative to the total dose of AA administered

Table 2 Molar percentages of the total dose administered represented by acrylamide and metabolites in 24 h rat urine Sex/route

AA

GA

AAMA

GAMA

GA/AAa

Totalb

M-IV F-IV M-gavage F-gavage M-diet F-dietc

2–4% 2–3 2 1–2 0.4–1.3 0.4

3–4% 5 6 4–7 0.8–4.3 5.7

42–49% 32–35 31 28–30 4–9 4

10–17% 11–18 27–29 20–22 4–14 8

0.3–0.4 0.4–0.7 1–1.1 0.8–1 1.1–1.9 3.3

57–74% 54–57 64–66 53–57 10–28 18

a

This ratio represents GA + GAMA/AA + AAMA. This parameter is the total molar percentage of the dose administered represented by the sum of AA + AAMA + GA + GAMA excreted in urine. c The urine volume collected from one rat was anomalously low so the total metabolites excreted were 1% of the dose administered and those values were therefore excluded from consideration. b

38

D.R. Doerge et al. / Toxicology Letters 169 (2007) 34–42

Fig. 2. Correlations between urinary concentrations of acrylamide, glycidamide and their mercapturic acids in rats and mice. Open symbols = glycidamide; filled symbols = acrylamide; diamonds = gavage; squares = IV; triangles = diet; x = controls. Rat plot: AA vs. AAMA, r2 = 0.78, p < 0.001; GA vs. GAMA, r2 = 0.81, p < 0.001. Mouse plot: AA vs. AAMA, r2 = 0.86, p < 0.001; GA vs. GAMA, r2 = 0.57, p < 0.001.

(Tables 1 and 2). As shown in Fig. 2, the concentration of either AA or GA in rat and mouse urine correlated significantly with the respective mercapturic acid concentration. Similar correlations were observed when total excreted metabolite data were substituted (data not shown). Mice typically excreted a greater fraction of the total dose as GA-derived species than did rats (Tables 1 and 2). The fraction of total dose excreted as GA-derived species was generally lower for the IV route relative to the oral routes for rats and mice. For individual rats and mice, and for all routes of administration, the fractions excreted as GA in urine were consistently higher than corresponding AA fraction. The fraction excreted as GA was generally greater in mice than rats. In rats, the fraction excreted as AAMA was consistently greater than that for GAMA; however, in mice the fraction excreted as GAMA was often greater than that for AAMA. For rats and mice combined, the percentage of total dose excreted as the sum of AA and its metabo-

Fig. 3. Correlations between urinary concentrations of acrylamide, glycidamide, and their mercapturic acids with the respective average serum AUC in rats. Open symbols = glycidamide; filled symbols = acrylamide; diamonds = gavage; squares = IV; triangles = diet; x = controls. Top plot: AUC–AA vs. AAMA, r2 = 0.83, p < 0.001; AUC-GA vs. GAMA, r2 = 0.32, p < 0.02. Bottom plot: AUC-AA vs. AA, r2 = 0.74, p < 0.001; AUC-GA vs. GA, r2 = 0.53, p < 0.001.

lites analyzed here (i.e., AA + GA + AAMA + GAMA) over 24 h for intravenous, gavage and dietary administration averaged 64 ± 9, 49 ± 14, 40 ± 29%, respectively. 3.2. Correlations between urinary metabolites and AUCs The average serum AUCs for AA and GA resulting from intravenous, gavage, and dietary administration of 0.1 mg/kg bw were previously determined for groups of rats and mice (Doerge et al., 2005b,c). Urinary concentrations of either AAMA or AA from the rodents in the current study were significantly correlated with the average AUC for AA and concentrations of either GAMA or GA were significantly correlated with the average AUC for GA (Figs. 3 and 4). Similar correlations were observed when total excreted metabolite data were substituted (data not shown).

D.R. Doerge et al. / Toxicology Letters 169 (2007) 34–42

39

observed when total excreted metabolite data were substituted (data not shown). Because significant correlations were also observed previously between average GA–Hb adducts, but not AA-Hb adducts, and average GA–DNA adducts (liver N7-GA-Gua; Tareke et al., 2006), it was of interest if levels of urinary GA-derived metabolites could also be useful for prediction of DNA adducts. Significant linear correlations were observed between urinary concentrations of either GA or GAMA with average DNA adducts in rats (p = 0.001 and 0.02, respectively; data not shown); in mice, a significant linear correlation was observed between GAMA, but not GA and average GA–DNA adducts (p = 0.03 and 0.2, respectively; data not shown). In both rats and mice, significant linear correlations were observed between either AA or AAMA with average GA–DNA adduct levels (p = 0.005 and 0.004, respectively; data not shown). Similar correlations were observed when total excreted metabolite data were substituted (data not shown). 4. Discussion

Fig. 4. Correlations between urinary concentrations of acrylamide, glycidamide, and their mercapturic acids with the respective average serum AUC in mice. Open symbols = glycidamide; filled symbols = acrylamide; diamonds = gavage; squares = IV; triangles = diet; x = controls. Top plot: AUC-AA vs. AAMA, r2 = 0.41, p < 0.01; AUCGA vs. GAMA, r2 = 0.56, p < 0.0001. Bottom plot: AUC-AA vs. AA, r2 = 0.41, p < 0.011; AUC-GA vs. GA, r2 = 0.34, p < 0.022.

3.3. Correlations between urinary metabolites and biomarkers Because significant correlations were also observed previously between average AUCs for AA or GA and average Hb adduct levels (AA–Val and GA–Val, respectively; Tareke et al., 2006), it was possible that urinary measurements would also correlate with Hb adducts. Although significant rank correlations were observed in rats for AA-Val with either urinary AA or AAMA (p = 0.001 for both, r2 = 0.73 and 0.68, respectively; data not shown), a weaker correlation was observed between GA–Val and urinary GA (p = 0.04, r2 = 0.27; data not shown) and a non-significant correlation was observed between GA–Val and urinary GAMA (data not shown). The only significant rank correlation observed in mice was between urinary AA and AA–Val (p = 0.006, r2 = 0.27; data not shown). Similar correlations were

The results from this study recapitulate some of the main findings of our previous serum toxicokinetic and biomarker evaluations in that mice generally excreted in urine greater amounts of GA-derived metabolites than rats, relative to AA-derived, and oral routes of administration resulted in greater formation of GA-derived species than intravenous (Tables 1 and 2). Unchanged AA was also excreted in urine at levels at or below 5% of the total dose and generally a higher percentage was excreted as AA following intravenous administration relative to the oral routes. The average molar percentages of total administered AA dose that were excreted in urine over 24 h as all four analytes measured in this study ranged from 40% to 64% for rats and mice. These values are consistent with previous 24 h urine measurements: 62% of total AA-derived radioactivity in rats (10 mg/kg bw intravenous, Miller et al., 1982); 50–51% of total measured AA-derived species in rats and mice (50 mg/kg bw gavage; Sumner et al., 1992); 31–62% of total measured AA-derived species in rats (50, 50, 150 mg/kg and 3 ppm in air for intraperitoneal injection, gavage, dermal, and inhalation exposure, respectively (Sumner et al., 2003) and 50% in rats (3 mg/kg bw gavage; Fennell et al., 2005). In these latter three papers, Fennell and co-workers used 13 C NMR to identify and quantify 13 Clabeled metabolites, but not 13 C-AA itself, in urine from male rats and mice. Table 3, which compares these previous results with the average male rodent gavage data from the current study, shows a consistent inverse

40

D.R. Doerge et al. / Toxicology Letters 169 (2007) 34–42

Table 3 Dose-dependency of urinary metabolite excretion over 24 h as percentages of orally administered acrylamide doses to male rodents and people Species dose

AA

AAMA

GA

GAMA

Glyceramide

a GA

a GA/b AA

Mouse 50 mg/kgc Mouse 0.1 mg/kgd Rat 50 mg/kgc Rat 3 mg/kge Rat 50 mg/kgf Rat 0.1 mg/kgg Human 3 mg/kge Human 12 ␮g/kgh Human 0.5–3 mg/kgi

NQ 0.65 NQ NQ NQ 2 NQ 4.4 ± 1.5 3–5

21 ± 1.1 7 34 ± 1.8 29 ± 4.5 71 31 26 ± 6.4 50 ± 9.4 35–43i

8.6 ± 1.1 16 2.8 ± 0.50 ND 7.3 6 0.79 ± 0.24 ND 0.4–0.7

17 ± 0.60 16 12 ± 0.60 21 ±2.1 19.8 28 ND 5.9 ± 1.2 0.7–0.8

2.7 ± 0.60 NM 1.2 ± 0.40 ND NM NM 3.3 ± 1.1 NM NM

28 32 16 21 28 34 4.1 5.9 1.3–1.5

1.3 4.2 0.47 0.72 0.39 1.0 0.16 0.11 0.03

NQ: observed but not quantified; ND: not detected; NM: not measured. a This sum represents GA + GAMA + glyceramide. b This sum represents AAMA + AAMA-sulfoxide (AA-cysteine is very small). c Sumner et al., 1992; Tables 3 (gavage, male rats) and 4 (gavage, male mice). d Table 1 (gavage, male mice this paper). e Fennell et al., 2005; Tables 1 (gavage, male rats) and 5 (gavage, men). f Sumner et al., 2003; Table 3 (gavage male rats). g Table 2 (gavage, male rats this paper). h Fuhr et al., 2006; Table 1 (men and women consuming potato chips). i Fennell et al., 2006; Table 1 (gavage, men).

dose-relationship for either the sum of GA-derived metabolites (GA + GAMA + glyceramide; GA) or the ratio of the sum of GA-derived metabolites to the sum of AA-derived species (GA + GAMA + glyceramide/ AA + AAMA; GA/AA). That is, as the AA dose is decreased, the amount of AA metabolized to GA increases in rats and mice. These comparisons show that urinary measurements faithfully reflect the effect of AA dose previously observed on rodent serum AUCs (Twaddle et al., 2004b; Doerge et al., 2005b,c), tissue DNA adduct levels (Doerge et al., 2005a), and Hb adducts (Bergmark et al., 1991). The data in Table 3 also suggest that humans exhibit lower metabolism to GA at low oral doses (≤0.1 mg/kg bw AA) than rodents in that humans excrete more AA-derived species and rodents excrete more GA-derived species, as previously discussed for a dose of 3 mg/kg bw (Fennell et al., 2005). The current investigation of urinary concentrations and total excretion amounts of AA and its metabolites was undertaken in rodent models for which detailed analyses of serum and tissue toxicokinetics, Hb adducts and DNA adducts have already been reported (Doerge et al., 2005a,b,c; Tareke et al., 2006). While measurement of Hb adducts is a well-established procedure for the estimation of internal exposures to AA and GA in humans (Paulsson et al., 2003b; Schettgen et al., 2004; Fennell et al., 2005; Vesper et al., 2005, 2006), several studies of AA and its metabolites in human urine have been published recently (Boettcher et al., 2005, 2006; Bjellaas et al., 2005; Fennell et al., 2005, 2006; Fuhr

et al., 2006). The high degree of variability in ratios of AA- and GA-derived biomarkers from non-smoking human Hb (Schettgen et al., 2004; Vesper et al., 2005, 2006) and urine (Bjellaas et al., 2005; Boettcher et al., 2005) has made it difficult to assess underlying differences in exposure or metabolic phenotype that could lead to differences in GA formation. This question is critical because GA has been extensively linked with the genotoxic properties of AA, which include DNA adduct formation in all rodent tissues examined (Segerb¨ack et al., 1995; Gamboa da Costa et al., 2003), micronucleus formation in rodents (Paulsson et al., 2003a; Ghanayem et al., 2005a), mutagenicity at hprt and cII loci in Big Blue mice that appears to be causally related with formation of N7-GA-Gua, the major GA–DNA adduct (Manjanatha et al., 2006) and presumably carcinogenicity in multiple tissues from male and female F344 rats following lifetime exposure (Johnson et al., 1986; Friedman et al., 1995). A possible strategy for assessing cancer risks from dietary AA is to predict target tissue GA–DNA adduct levels in humans and to compare them with those from rodent carcinogenicity bioassays in order to reduce the uncertainty inherent in inter-species extrapolation. Significant linear correlations were observed between urinary measurements reported here and the corresponding average AUC values, which were determined previously in other groups of rodents (Doerge et al., 2005b,c). It should be noted that the strongest correlations reported here were those between either AA and AAMA or GA and GAMA concentrations, which were

D.R. Doerge et al. / Toxicology Letters 169 (2007) 34–42

determined in common urine samples (Fig. 2). The lower degree of correlation observed between AA- and GAderived urinary metabolites and group averages of the respective AUCs, Hb adducts and N7-GA-Gua levels from previous measurements probably reflects the limitations that are inherent in comparisons that include a high degree of inter-animal variability and small group sizes. This conclusion is also consistent with the uniformly strong correlations between GA AUCs and either GA–Hb or GA–DNA adducts previously determined in the same groups of rats and mice following single and multiple dosing with AA (Tareke et al., 2006). The direct relationship between GA serum concentration and either DNA or Hb adduct formation described above (as opposed to the successive steps of conjugation with glutathione, metabolism to mercapturic acids and excretion in urine) would also predict better correlations between the two adduct measurements (Fig. 1). Further complications to using only urinary data sets to predict human internal exposures come from the observations of dose and species differences in urinary profiles of AA-derived compounds. At a common dose of 3 mg/kg bw, human urine contains a significant proportion of total GA-derived metabolites as glyceramide (approximately 80%) whereas rat urine did not (Fennell et al., 2005). This observation does not permit an evaluation of whether hydrolysis of GA to glyceramide occurred during sample collection or was enzymatic, involving epoxide hydrolase; however, there is some evidence that in whole human blood, soluble epoxide hydrolase is not involved (Paulsson et al., 2005). Similarly, only Fennell et al. (2006) has identified and quantified AAMA-sulfoxide as a prominent human, but not rodent, urinary metabolite derived from glutathione conjugation of AA (approximately 25% of AAMA values). To date, no analyses of AA and metabolites in human urine from solely dietary exposures has included glyceramide, which could be an important indicator of GA formation that should be included in future evaluations of human internal exposures. The deficiencies inherent in using any one type of biomarker data suggest that the best approach to estimating human internal exposure and any pharmacodynamic consequences may come from the integration of all types of data in a physiologically based pharmacokinetic (PBPK) model based on measurements of toxicokinetics in experimental animals, urinary metabolites in animals and humans and adduct biomarkers in animals and humans. Using such a comprehensive approach should be the best way to evaluate the underlying diversity in the human population, including age, gender, exposure and metabolic factors, into a meaningful risk assessment.

41

5. Conflict of interest statement None. Acknowledgements The high-level technical expertise of Ms. Michelle Vanlandingham, Ms. Carol Cain and Mr. Andy Matson, Bionetics Corporation, Jefferson, AR, is gratefully acknowledged. This research was supported in part by Interagency Agreement #224-93-0001 between the National Center for Toxicological Research/U.S. Food and Drug Administration and the National Institute for Environmental Health Sciences/National Toxicology Program. We thank the DFG (German Research Foundation) for their financial support of the project (AN 107/17-117-2). Helpful discussions with Drs. John F. Young and Ralph L. Kodell, both from NCTR, are greatly appreciated. The views presented in this article do not necessarily reflect those of the U.S. Food and Drug Administration. References Bergmark, E., Calleman, C.J., Costa, L.G., 1991. Formation of hemoglobin adducts of acrylamide and its epoxide metabolite glycidamide in the rat. Toxicol. Appl. Pharmacol. 111, 352–363. Besaratinia, A., Pfeifer, G.P., 2004. Genotoxicity of acrylamide and glycidamide. J. Nat. Cancer Inst. 96, 1023–1029. Bjellaas, T., Janak, K., Lundanes, E., Kronberg, L., Becher, G., 2005. Determination and quantification of urinary metabolites after dietary exposure to acrylamide. Xenobiotica 35, 1003–1018. Boettcher, M.I., Angerer, J., 2005. Determination of the major mercapturic acids of acrylamide and glycidamide in human urine by LC-ESI–MS/MS. J. Chromatogr. B 824, 283–294. Boettcher, M.I., Schettgen, T., K¨utting, B., Pischetsreider, M., Angerer, J., 2005. Mercapturic acids of acrylamide and glycidamide as biomarkers of the internal exposure to acrylamide in the general population. Mutat. Res. 580, 167–176. Boettcher, M.I., Bolt, H.M., Drexler, H., Angerer, J., 2006. Excretion of mercapturic acids of acrylamide and glycidamide in human urine after single oral administration of deuterium-labelled acrylamide. Arch. Toxicol. 80, 55–61. Boon, P.E., de Mul, A., van der Voet, H., van Donkersgoed, G., Brette, M., Van Klavern, J., 2005. Calculations of dietary exposure to acrylamide. Mutat. Res. 580, 143–155. Calleman, C.J., Wu, Y., He, F., Tian, G., Bergmark, E., Zhang, S., Deng, H., Wang, Y., Crofton, K.M., Fennell, T.R., Costa, L.G., 1994. Relationships between biomarkers of exposure and neurological effects in a group of workers exposed to acrylamide. Toxicol. Appl. Pharmacol. 126, 361–371. Doerge, D.R., Gamboa da Costa, G., McDaniel, L.P., Churchwell, M.I., Twaddle, N.C., Beland, F.A., 2005a. DNA adducts derived from administration of acrylamide and glycidamide to mice and rats. Mutat. Res. 580, 131–142. Doerge, D.R., Young, J.F., McDaniel, L.P., Twaddle, N.C., Churchwell, M.I., 2005b. Toxicokinetics of acrylamide and gly-

42

D.R. Doerge et al. / Toxicology Letters 169 (2007) 34–42

cidamide in B6C3F1 mice. Toxicol. Appl. Pharmacol. 202, 258–267. Doerge, D.R., Young, J.F., McDaniel, P., Twaddle, N.C., Churchwell, M.I., 2005c. Toxicokinetics of acrylamide and glycidamide in Fischer 344 rats. Toxicol. Appl. Pharmacol. 208, 199–209. Fennell, T.R., Sumner, S.C.J., Snyder, R.W., Burgess, J., Spicer, R., Bridson, W.E., Friedman, M.A., 2005. Metabolism and hemoglobin adduct formation of acrylamide in humans. Toxicol. Sci. 85, 447–459. Fennell, T.R., Sumner, S.C.J., Snyder, R.W., Burgess, J., Friedman, M.A., 2006. Kinetics of elimination of urinary metabolites of acrylamide in humans. Toxicol. Sci. 93, 256–267. Friedman, M.A., Dulak, L.H., Stedman, M.A., 1995. A lifetime oncogenicity study in rats with acrylamide. Fundam. Appl. Toxicol. 27, 95–105. Fuhr, U., Boettcher, M.I., Kinzig-Schippers, M., Weyer, A., Jetter, A., Lazar, A., Taubert, D., Tomalik-Scharte, D., Pournara, P., Jakob, V., Harlfinger, S., Klassen, T., Berkessel, A., Angerer, J., S¨orgel, F., Sch¨omig, E., 2006. Toxicokinetics of acrylamide in humans after ingestion of a defined dose in a test meal to improve risk assessment for acrylamide carcinogenicity. Cancer Epidemiol. Biomarkers Prev. 15, 266–271. Gamboa da Costa, G., Churchwell, M.I., Hamilton, L.P., Beland, F.A., Marques, M.M., Doerge, D.R., 2003. DNA adduct formation from acrylamide via conversion to glycidamide in adult and neonatal mice. Chem. Res. Toxicol. 16, 1328–1337. Ghanayem, B.I., Witt, K.L., Kissling, G.E., Tice, R.R., Recio, L., 2005a. Absence of acrylamide-induced genotoxicity in CYP2E1null mice: evidence consistent with a glycidamide-mediated effect. Mutat. Res. 578, 284–297. Ghanayem, B.I., Witt, K.L., El-Hadri, L., Hoffler, U., Kissling, G.E., Shelby, M.D., Bishop, J.B., 2005b. Comparison of germ cell mutagenicity in male CYP2E1-null and wild-type mice treated with acrylamide: evidence supporting a glycidamide-mediated effect. Biol. Reprod. 72, 157–163. Johnson, K.A., Gorzinski, S.J., Bodner, K.M., Campbell, R.A., Wolf, C.H., Friedman, M.A., Mast, R.W., 1986. Chronic toxicity and oncogenicity study on acrylamide incorporated in the drinking water of Fischer 344 rats. Toxicol. Appl. Pharmacol. 85, 154–168. 64th Joint WHO/FAO Expert Committee on Food Additives (JECFA), in press. Evaluation of acrylamide in food. World Health Organ. Tech. Rep. Ser. Kellert, M., Scholz, K., Wagner, S., Dekant, W., V¨olkel, W., 2006. Quantitation of mercapturic acids from acrylamide and glycidamide in human urine using a column switching tool with two trap columns and electrospray tandem mass spectrometry. J. Chromatogr. A 1131, 58–66. LoPachin, R.M., Balaban, C.D., Ross, J.F., 2003. Acrylamide axonopathy revisited. Toxicol. Appl. Pharmacol. 188, 135–153. Manjanatha, M.G., Aidoo, A., Shelton, S.D., Bishop, M.E., McDaniel, L.P., Lyn-Cook, L.E., Doerge, D.R., 2006. Genotoxicity of acrylamide and its metabolite glycidamide administered in drinking water to male and female Big Blue mice. Environ. Mol. Mutagen. 47, 6–17. Miller, M.J., Carter, D.E., Sipes, I.G., 1982. Pharmacokinetics of acrylamide in Fisher-344 rats. Toxicol. Appl. Pharmacol. 63, 36–44. Paulsson, B., Kotova, N., Graw´e, J., Henderson, A., Granath, F., Golding, B., T¨ornqvist, M., 2003a. Induction of micronuclei in mouse

and rat by glycidamide, genotoxic metabolite of acrylamide. Mutat. Res. 535, 15–24. Paulsson, B., Athanassiadis, I., Rydgerg, P., T¨ornqvist, M., 2003b. Hemoglobin adducts from glycidamide: acetonization of hydrophilic groups for reproducible gas chromatography/tandem mass spectrometric analysis. Rapid Commun. Mass Spectrom. 17, 1859–1865. Paulsson, B., Rannug, A., Henderson, A.P., Golding, B.T., T¨ornqvist, M., Warholm, M., 2005. In vitro studies of the influence of glutathione transferases and epoxide hydrolase on the detoxification of acrylamide and glycidamide in blood. Mutat. Res. 580, 53–59. Roach, J.A.G., Andrezejewski, D., Gay, M.L., Northrup, D., Musser, S.M., 2003. Rugged LC/MS/MS survey analysis for acrylamide in foods. J. Agric. Food Chem. 51, 7547–7554. Schettgen, T., Rossbach, B., K¨utting, B., Letzel, S., Drexler, H., Angerer, J., 2004. Determination of haemoglobin adducts of acrylamide and glycidamide in smoking and non-smoking persons of the general population. Int. J. Hyg. Environ. Health 207, 1–9. Segerb¨ack, D., Calleman, C.J., Schroeder, J.L., Costa, L.G., Faustman, E.M., 1995. Formation of N-7-(2-carbamoyl-2hydroxyethyl)guanine in DNA of the mouse and the rat following intraperitoneal administration of [14 C]acrylamide. Carcinogenesis 16, 1161–1165. Sumner, S.C.J., MacNeela, J.P., Fennell, T.R., 1992. Characterization and quantification of urinary metabolites of [1,2,3-13 C]acrylamide in rats and mice using 13 C nuclear magnetic resonance spectroscopy. Chem. Res. Toxicol. 5, 81–89. Sumner, S.C.J., Williams, C.C., Snyder, R.W., Krol, W.L., Asgharian, B., Fennell, T.R., 2003. Acrylamide: a comparison of metabolism and hemoglobin aducts in rodents following dermal, intraperitoneal, oral, or inhalation exposure. Toxicol. Sci. 75, 260–270. Tareke, E., Rydberg, P., Karlsson, P., Eriksson, S., T¨ornqvist, M., 2002. Analysis of acrylamide, a carcinogen formed in heated foodstuffs. J. Agric. Food Chem. 50, 4998–5006. Tareke, E., Twaddle, N.C., McDaniel, P., Churchwell, M.I., Young, J.F., Doerge, D.R., 2006. Relationships between biomarkers of exposure and toxicokinetics in Fischer 344 rats and B6C3F1 mice administered single doses of acrylamide and glycidamide and multiple doses of acrylamide. Toxicol. Appl. Pharmacol. 217, 63–75. Twaddle, N.C., Churchwell, M.C., McDaniel, L.P., Doerge, D.R., 2004a. Autoclave sterilization produces acrylamide in rodent diets: implications for toxicity testing. J. Agric. Food Chem. 52, 4344–4349. Twaddle, N.C., Hamilton, L.P., Gamboa da Costa, G., Churchwell, M.I., Beland, F.A., Doerge, D.R., 2004b. Determination of acrylamide and glycidamide serum toxicokinetics in B6C3F1 mice using LC-ES/MS/MS. Cancer Lett. 207, 9–17. Vesper, H.W., Licea-Perez, H., Meyers, T., Ospina, M., Meyers, G.L., 2005. Pilot study on the impact of potato chips consumption on biomarkers of acrylamide exposure. Adv. Exp. Med. Biol. 561, 89–96. Vesper, H.W., Ospina, M., Meyers, T., Ingrahm, L., Smith, A., Gray, J.G., Meyers, G.L., 2006. Automated method for measuring globin adducts of acrylamide and glycidamide at optimized Edman reaction conditions. Rapid Commun. Mass Spectrom. 20, 959–964.