Mutation Research 580 (2005) 91–101
Reaction-kinetic parameters of glycidamide as determinants of mutagenic potency V. Silvaria , J. Haglunda , D. Jenssenb , B.T. Goldingc , L. Ehrenberga , M. T¨ornqvista,∗ a
Department of Environmental Chemistry, Stockholm University, S-10691 Stockholm, Sweden b Department of Genetics, Microbiology and Toxicology, Stockholm University, S-10691 Stockholm, Sweden c Department of Chemistry, Bedson Building, University of Newcastle upon Tyne, NEI 7RU, UK Received 9 October 2004; received in revised form 9 November 2004; accepted 13 November 2004 Available online 6 January 2005
Abstract Values for reaction-kinetic parameters of electrophiles can be used to predict mutagenic potency. One approach employs the Swain–Scott relationship for comparative kinetic studies of electrophilic agents reacting with nucleophiles. In this way glycidamide (GA), the putatively mutagenic/carcinogenic metabolite of acrylamide, was assessed by determining the rates of reaction with different nucleophiles. The rate constants (kNu ) were determined using the “supernucleophile” cob(I)alamin [Cbl(I)] as an analytical tool. The Swain–Scott parameters for GA were compared with those of ethylene oxide (EO). The substrate constants, s values, for GA and for EO were found to be 1.0 and 0.93, respectively. The reaction rates at low values of nucleophilic strength (n = 1–3), corresponding to oxygens in DNA, were determined to be 2–3.5 times higher for GA compared to EO. GA was also more reactive than EO towards other nucleophiles (n = 0-6.4). The mutagenic potency of GA was estimated in Chinese hamster ovary cells (hprt mutations in CHO-AA8 cells per dose unit with gamma-radiation as reference standard). The potency of GA was estimated to be about three mutations per 105 cells and mMh corresponding to about 400 rad-equ./mMh. A preliminary comparison of the mutagenic potency
Abbreviations: Cbl, cobalamin; Cbl(I), cob(I)alamin; CHO cells, Chinese hamster ovary cells; EO, ethylene oxide; EO–Cbl, ethylene oxide–cobalamin; GA, glycidamide; GA–Cbl, glycidamide–cobalamin; HBSS++ , Hank’s balanced salt solution with HEPES; HEPES, N-(2-hydroxyethyl)piperazine-N -2-ethanesulfonic acid; IS, internal standard; IS-Cbl, internal standard-cobalamin; LC–MS/MS, liquid chromatography–tandem mass spectrometry; mMh, millimolar-hour; n, nucleophilic strength; OH-Cbl, hydroxocobalamin; PO, propylene oxide; PO–Cbl, propylene oxide–cobalamin; Rad-equ, rad-equivalence; s, substrate constant; SN 2, bimolecular nucleophilic substitution; TFA, trifluoroacetic acid; VMA, valine methylamide ∗
Corresponding author. Tel.: +46 8 163769; fax: +46 8 163979. E-mail address:
[email protected] (M. T¨ornqvist).
1383-5718/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.mrgentox.2004.11.004
92
V. Silvari et al. / Mutation Research 580 (2005) 91–101
(per mMh and as rad-equivalents) of GA and EO shows an approximately seven times higher potency for GA. A higher mutagenic potency of GA compared to EO is compatible with expectation from reaction-kinetic data of the two compounds. The data confirmed that GA is not a strong mutagen, which is in line with what is expected for simple oxiranes. The present study shows the value of cob(I)alamin for the determination of reaction-kinetic parameters and their use for prediction of mutagenic potency. © 2004 Elsevier B.V. All rights reserved. Keywords: Glycidamide; Ethylene oxide; Reaction-kinetics; Mutagenic potency; Cobalamin
1. Introduction Qualitative and quantitative aspects of “genotoxic effects” are areas of interest in environmental chemistry aiming at estimating risks of the harmful effects of chemicals. Primarily this concerns mutation, considering that mutation in germ-line cells may cause heritable diseases that are manifested in generations ahead, and that mutations in somatic cells are key events in chemical induction of cancer. The main pathway leading to mutation comprises changes in DNA through reaction of a compound or its reactive metabolite(s) with the nucleophilic atoms in DNA. Important features of these reactions of electrophiles concern the reactivity pattern, as well as the absolute reactivity towards specific nucleophilic sites in DNA molecules. Both these parameters are basic to relationships between genotoxic response and dose of the electrophile. Characterisation of putatively mutagenic compounds with respect to reactivity in nucleophilic substitution reactions is expected to generate background data for prediction of the mutagenic potency, at least approximately, at low and intermediate doses. In 1953, Swain and Scott proposed an empirical linear freeenergy relationship kNu log = s × n (1) k H2 O which is useful for comparisons of electrophilic compounds [1]. The dependence of a reaction rate (kNu ) on the nucleophilic strength (n) is presented in terms of a substrate (or selectivity) constant s, which describes the reaction rate relative to that obtained with a standard nucleophile, such as water, in the Swain–Scott system (Eq. (1)). This approach, as well as a few alternative ways of looking at structure–activity relationships were described by Ross [2] (cf. [3]).
Interest in studies of the parameters of Eq. (1) and similar relationships were stimulated by the work of Lawley and Brookes (1963) [4] and by Loveless’s suggestion in 1969 [5] that mutagenicity was associated with the ability to alkylate guanine-O6 atoms (low n value) in DNA. Several alkylating agents, e.g. simple epoxides like ethylene oxide (EO) and methylating and ethylating agents, have been characterised with regard to their reactivity towards nucleophiles. When the relationship (1) was applied to mono-functional alkylating agents, it was found that the mutagenic potency per dose unit was approximately proportional to the reactivity at low n values (n = 1–3), assumed to correspond to oxygen atoms in DNA [6,7]. In this context the dose is expressed as concentration over time, e.g. in millimolar-hours (mMh) [8]. Furthermore, this kind of structure–activity relationship has been shown to be approximately valid for tumorigenic potency for a series of monofunctional alkylating agents, where a correlation to s values has been observed [9,10]. The present study aims at an evaluation of glycidamide (GA) as the putatively mutagenic and, therefore, cancer risk-increasing metabolite operating in exposure to acrylamide. The finding that acrylamide is formed at relatively high levels in cooking [11,12] has led to concern about the exposure and associated health risks to the general population. GA has been shown, also in humans, to be metabolically formed by epoxidation of the double bond of acrylamide [13,14]. GA is electrophilically reactive and gives rise to 2-carbamoyl2-hydroxyethyl adducts to nucleophilic atoms (Fig. 1). A further characterisation of the reactivity of GA is called for in studies of genotoxic mechanisms and in evaluation of the ability of acrylamide exposure to increase cancer risk. The present investigation involved reaction of GA with model nucleophiles with various nucleophilic
V. Silvari et al. / Mutation Research 580 (2005) 91–101
93
Fig. 1. Structures of glycidamide, ethylene oxide and propylene oxide and their adducts in reaction with the studied nucleophiles or cob(I)alamin following attack on the least hindered carbon in the 1,2-epoxide function.
strengths. The rate constants for reactions with the nucleophiles were determined by measurement of the rate of decrease of the concentration of GA. A novel method, using the “supernucleophile” cob(I)alamin [Cbl(I)] for instant trapping of electrophiles, has been applied for accurate measurement of concentrations of the electrophiles by conversion to the corresponding alkyl–cobalamins [15]. For comparison of the analytical method ethylene oxide (EO), giving 2hydroxyethylation in nucleophilic substitution reactions, was studied in parallel (Fig. 1). EO is an unsubstituted oxirane that has earlier been carefully studied with regard to both reaction-kinetic parameters and genotoxicity [3,10,16,17]. In the present study, the Swain–Scott parameters (kH2 O , s and kn≈2 ) for GA were determined and compared with those of EO. The relative genotoxic potency predicted from the reactionkinetic studies was compared with that obtained in in
vitro mutation tests in Chinese hamster ovary cells (hprt mutations in CHO-AA8 cells).
2. Materials and methods 2.1. Materials Aniline, ethylene oxide (EO; CAS-Nr 75-21-68), Hank’s balanced salt solution (HBSS), hydroxocobalamin (OH-Cbl), N-(2-hydroxyethyl)piperazine-N -2ethanesulfonic acid (HEPES), sodium borohydride (NaBH4 ), sodium acetate, sodium hydrogen carbonate, sodium thiosulfate and trifluoroacetic acid (TFA) were obtained from Sigma (St. Louis, MO). Propylene oxide (PO; CAS-Nr 75-56-9) was obtained from Fluka (Buchs, Switzerland). Glycidamide (GA; CASNr 5694-00-8) [18] and valine methylamide (VMA)
94
V. Silvari et al. / Mutation Research 580 (2005) 91–101
[19] were synthesized in our laboratory. Solvents used were of analytical grade. Caution: GA, EO and PO are hazardous and should be handled carefully. 2.2. Preparation of cob(I)alamin Cob(I)alamin [Cbl(I)] was prepared according to Haglund et al. [15] for the determination of rate constants of oxiranes, with slight modifications. Briefly, OH-Cbl dissolved in water (5 mM, 200 L) was added, with a reduction catalyst Co(NO3 )2 (25 mM, 10 L), to a septum-sealed amber vial. The solution was degassed with argon for a few minutes and NaBH4 (150 mM, 25 L) was added to the solution, prior to the addition of oxirane (see below), to produce the reduced form of cobalamin, Cbl(I). 2.3. Reactions All reactions were performed in septum-sealed vials (2 or 4 mL). Stock solutions of the oxiranes (GA and EO) were prepared in water. The hydrolysis of GA was studied by adding the oxirane (1–5 mM final concentration) by syringe to water (pH 6). The mixture was incubated at 37 ◦ C. Aliquots (200 L) were sampled at different time points using a syringe and mixed with an equal volume of internal standard (IS) solution. As internal standard PO (1–2 mM in MeOH) was used (evaluated by Haglund et al., to be published). Finally, 100–300 L of the mixed solution were added to the Cbl(I) solution to form the alkyl–cobalamins of GA, EO and of PO (Fig. 1). After addition of reaction solution and IS to the cob(I)alamin vial, the sample was kept under argon for 2 min. After another 15 min the cap was opened to reoxidise the unmodified cobalamin to OH-Cbl. The vials were stored in the freezer (−20 ◦ C) until analysed by HPLC–UV with regard to the alkyl–cobalamins (GA–Cbl, EO–Cbl, IS-Cbl; see Fig. 1). GA and EO were incubated with model nucleophiles, with nucleophilic strengths, n, in the Swain–Scott system according the following: water, 0 (by definition, [1]), carbonate, 2.62 [20], acetate, 2.72 [2], valine methylamide, 4.2 [21], aniline, 4.35 [20] and thiosulfate, 6.36 [2]. Different concentrations of the nucleophiles were selected to obtain reasonable time periods of the studies. The concentration of the nucle-
ophile was at least 10 times in excess compared to the electrophile in all reactions. The incubations were done at 37 ◦ C and the following reaction conditions: Acetate (500 mM in water, at pH 6.5), VMA (20 mM in water, at pH 8), aniline (45 mM in water, at pH 6.8) and thiosulfate (10 mM in water, at pH 6.5). In addition, GA was studied with HCO3 − (0.5 M in water, at pH 8). The pH was measured using a calibrated pH electrode. In addition, the rate of disappearance of GA and EO in HBSS++ medium (pH 7.4) was determined (HBSS with addition of 10 mM HEPES). The reactions were followed as described above. 2.4. Chromatography conditions All the reactions were followed by HPLC analysis carried out on a Perkin-Elmer series 200 liquid chromatography pump (Norwalk, USA), connected to an injector equipped with a 10 L loop and to an Applied Biosystem’s 759 A UV detector (California, USA). GA–Cbl, EO–Cbl and IS-Cbl, were analysed using a Kromasil-RP-C18 column (150 mm × 4.6 mm i.d.; particle size 5 m; Hichrom Limited, Reading, UK). The samples were eluted in gradient conditions with a mobile phase constituted by solution A (8% acetonitrile) and solution B (80% acetonitrile), each with 0.1% TFA, at a flow rate of 1 mL/min. The pumps were kept at 15% B for 5 min followed by a linear increase to 60% B within the next 20 min. Finally 100% B was used to rinse the column followed by regeneration at 15% B. All samples were analysed using this system (or slight modification) except for the reactions of EO and GA with thiosulfate where 50 L of the samples were analysed. This required coupling of the column to a pre-column and switching valve. A Kromasil-RP-C18 pre-column (50 mm × 4.6 mm i.d.; particle size 5 m; Hichrom Limited, Reading, UK) was connected to the analytical column by a switching valve (Waters-Millipore, MA, USA). The same gradient was used but during the first 5 min the switching valve was directed to waste to get rid of the excess of unmodified OH-Cbl. At time 5.00 min the valve position was switched and the solvent was passed to the analytical column. The mobile phases were degassed prior to use. The detector wavelength was set at 267 nm. For the data acquisition and integration, the programme ELDS 1.1 for Windows was used.
V. Silvari et al. / Mutation Research 580 (2005) 91–101
2.5. Determination of the (pseudo) first-order rate constant kH for GA 2O
Swain–Scott relationship (Eq. (1)) and the secondorder rate constants for the reactions with the respective nucleophiles.
The natural logarithms of the ratios of the alkyl–cobalamins formed (GA–Cbl and IS-Cbl or EO–Cbl and IS-Cbl) were plotted versus time. The pseudo first-order rate constants were obtained from the slopes of the linear regression line according to (Eqs. (2) and (3)): ln C = −k t + ln C0 or
GA–Cbl ln IS-Cbl
(2)
GA–Cbl = k t + ln IS-Cbl
(3) 0
The half-life of GA in water at 37 ◦ C was calculated from (Eq. (4)): t1/2 =
ln 2 kH 2O
(4)
In this work we used the rate constant for the hydrolysis of EO, previously determined using the cobalamin method [15], which was in agreement with values determined by other methods [16,22]. 2.6. Calculation of second-order rate constants and s values of GA and EO The first-order rate constants for GA and EO, kacetate , kaniline , kVMA , kthiosulfate , kHBSS+HEPES and kcarbonate (only for GA) were determined experimentally as described above and corrected by subtraction of the hydrolysis rate in water (kH ). The second-order 2O rate constants were calculated as follows: kNu =
kNu [Nu]
(5)
where the concentration of the nucleophile [Nu] corresponds to the concentration of free base according to the Henderson–Hasselbalch equation: pH = pKa − log
HA A−
95
(6)
The second-order rate constants of GA obtained were compared with those of EO. The values of the substrate constants, s, of GA (sGA ) and of EO (sEO ) were determined using the
2.7. Quantification of mutagenic potency Mutation experiments were performed on CHOAA8 cells, and the gene encoding hypoxanthine– guanine–phosphoribosyl transferase (HPRT) was used for detection of induced mutations according to a quantitative assay described elsewhere [23]. The mutant frequency induced by GA per dose unit was determined and compared with the genotoxic potency of gamma-radiation, determined in a parallel experiment. The treatment with GA and the cultivation of cells were done as described by Johansson et al. [24]. Treatment with GA in the mutagenicity assay was conducted for 1 h in Hank’s balanced salt solution with 10 mM HEPES (HBSS++ ). In the mutation experiment with ionizing radiation, irradiation was performed using a 137 Cs source at a dose rate of 56 rad/min (0.56 Gy/min). Mutant frequencies were adjusted with the corresponding cloning efficiency for each treatment group [23]. Standard errors of mean (S.E.) for induced mutation were estimated from the linear regression analysis of each treatment. In separate experiments the stability of GA (and EO for comparison) in Hank’s balanced salt solution at treatment conditions was studied, which generated data allowing calculation of the doses (concentration × time) of GA during treatment. The data on mutagenicity of GA was inferred from Johansson et al. [24] using the treatment doses for GA calculated in this work. Furthermore, the results on mutation frequency from gamma-radiation allowed comparison of mutagenic potency in radiation-dose equivalents [25,26].
3. Results 3.1. The cob(I)alamin method The solution used for trapping of EO and GA by Cbl(I), with PO as internal standard (IS), was analysed by HPLC–UV and showed formation of GA–Cbl and EO–Cbl, respectively, together with ISCbl (Figs. 2 and 3). The alkyl–cobalamins in the HPLC analyses were quantified as the areas of the peaks.
96
V. Silvari et al. / Mutation Research 580 (2005) 91–101
Fig. 2. Chromatogram showing the alkyl–cobalamins formed in reaction between cob(I)alamin and glycidamide and propylene oxide (IS), respectively.
for GA was r2 > 0.95, for example see Fig. 4). The kH 2O −1 ◦ determined to be 0.016 h at 37 C and pH 6, corresponding to a half-life in water of 43 h. This is compatible with half-lives of similar oxiranes, e.g. the half-life of EO is 65–70 h [15,16,22]. Previously determined value on kH for EO [15] was used for calculations in 2O the present study. The second-order rate constants for GA and EO, obtained according to Eq. (5) after subtraction of the hydrolysis rate, are summarised in Table 1. GA was shown to be more reactive than EO towards all the nucleophiles (n values = 0-6.4). The reactivity of GA towards thiosulfate was too high to permit accurate measurements during the experimental conditions. From the reaction between GA and the other nucleophiles, GA is expected to react with thiosulfate 10 times faster than EO, resulting in a half-life of 3 min at the conditions used in our study.
3.3. Determination of substrate constants according to the Swain–Scott relationship Fig. 3. Chromatogram showing the alkyl–cobalamins formed after reaction between cob(I)alamin and ethylene oxide and propylene oxide (IS), respectively.
3.2. Rate constants The plots of ln(GA–Cbl/IS-Cbl) versus time comprised 5–6 time points taken within one half-life (with
Relative reactivities in nucleophilic substitution reactions are described quantitatively by substrate constants, s, according to Swain and Scott’s Eq. (1). The s values for EO and GA were determined from regression of log kNu versus n of the studied nucleophiles (Fig. 5, cf. Eq. (1)). The n values (Table 1) used are taken from the literature (see 2.3). The s value of EO
Fig. 4. Data from the measurement of hydrolysis of GA using cob(I)alamin. The ratios of the areas of the formed GA–Cbl and IS-Cbl are plotted versus time.
V. Silvari et al. / Mutation Research 580 (2005) 91–101
97
Table 1 Summary of the rate constants of glycidamide and ethylene oxide obtained in the reaction-kinetic studies Nucleophile (in water)
n value
[Nu]
Glycidamide (h−1 ) ktot
H2 O HCO3 − Acetate VMA
0 (by definition) 2.6 2.7 4.2
Aniline Thiosulfate HBSS++
4.4 6.3 Not known
∗ ∗∗
55.5 M 0.5 M 0.5 M 25 mM for GA* 50 mM for EO* 45 mM 10 mM Not known
Ethylene oxide kNu (M−1 h−1 )
0.016 0.104 0.16 0.18
3 × 10−4
0.31 – 0.1**
6.5 – Not calculated
0.18 0.29 8.2
(h−1 ) ktot
kNu (M−1 h−1 )
0.01 – 0.05 0.08
2 × 10−4 – 0.08 1.6
0.17 1.3 0.02**
3.6 130 Not calculated
The concentrations of VMA were corrected to 20 mM for GA and 40 mM for EO according to pH and pKa . corresponds to the disappearance of the oxiranes due to reactions with all components in HBSS++ . The rate constant ktot
was redetermined in this work to validate the approach using the cobalamin method for determination of rate constants. The s value obtained for EO, 0.93, is in agreement with values of 0.9-0.96 earlier determined ([3]; cf. ref [10]) (Fig. 5). The s value obtained from the plot (Fig. 5) for GA was 1.0. The deviations from the regression line in Fig. 5 are assumed to partly depend on uncertainties in the used n values obtained from the literature, therefore log kEO was plotted versus log kGA , only using kNu of reactions with nucleophiles performed with both GA and EO (Fig. 6). This gives a slope corresponding to sEO /sGA of 0.93. Using the sEO obtained in the present study (0.93), sGA will then be 1.0.
Fig. 5. The logarithm of the kGA and kEO obtained are plotted against the n values of the studied nucleophiles (cf. Table 1). The thiosulfate value is not included in the calculation of the regression line of GA (see Section 3.2).
3.4. Rate of disappearance of GA and EO in HBSS++ cell medium The rate of disappearance of GA and EO in Hank’s balanced salt solution with addition of HEPES (HBSS++ ) was determined since this medium was used in the mutation tests. HBSS++ is a complex mixture of different components with different nucleophilic strengths. The half-lives of GA and EO in HBSS++ were 7 and 35 h, respectively. 3.5. The mutagenicity of GA and EO The mutagenicity of GA was determined in the HPRT locus in Chinese hamster cells. The mutagenic potency of GA was determined per dose unit and as radiation-dose equivalents (inferred from mutation experiments with gamma-radiation). To obtain quantitative comparability it is important to refer the relative genotoxic potency to treatment dose (in mMh). From the determination of the half-lives of GA and EO in the media for treatment of cells in the mutation experiment the treatment dose could be calculated. EO could be considered approximately stable during the 1 h treatment time in the mutation tests, and the dose could therefore be calculated as initial concentration × treatment time. However, due to non-negligible elimination of GA during 1 h treatment the dose (D) was calculated according to Eq. (7) where k is the first order rate constant for disappearance of GA and C0 is the initial concentration of GA. D=
C0 (1 − e−kt ) k
(7)
98
V. Silvari et al. / Mutation Research 580 (2005) 91–101
Fig. 6. The logarithm of the kGA obtained is plotted against the logarithm of kEO obtained (cf. Table 1). The rate constants kH2 O and kVMA for EO are adapted from a previous study [15].
On this basis the mutagenic potency of GA was calculated to be 3.45 ± 2.1 mutations per 105 cells and mMh, using mutation data from [24]. With gammaradiation, giving linear dose–response, a mutation frequency of 0.0083 ± 0.0034 mutants per 105 cells and rad was obtained. For GA, with gamma-radiation as reference standard, a relative genotoxic potency of about 400 ± 300 rad-equ/mMh was inferred.
4. Discussion
ophile Cbl(I) (with n ≈ 10) reacts 104 to 106 times faster than relatively strong standard nucleophiles and is, as earlier demonstrated, therefore advantageous for trapping of reactive compounds [15]. Measuring the remaining concentration of electrophile implies that all reactions causing elimination of the electrophile is included, hereby considered in the calculations of rate ) by subtraction of the hydrolysis rate. constants (kNu In the present work, Cbl(I) was also applied for measurement of concentration of electrophiles in the media used for mutation tests, for the determination of treatment doses.
4.1. The cob(I)alamin method 4.2. Nucleophilic substitution reactions Cob(I)alamin has been applied to analysis of GA and EO in reaction-kinetic studies. The reactions were followed by the measurement of the decreasing concentrations of the oxiranes. The oxiranes were trapped and analysed as alkyl–Cbl by HPLC–UV. Using this approach the analyte is the same irrespective of nucleophile studied. If higher analytical sensitivity is needed, the analysis of the alkyl–cobalamins formed may be performed using liquid-chromatography–tandem mass spectrometry (LC–MS/MS) [27,28]. The supernucle-
The reactions with Cbl(I) can be formulated as nucleophilic displacement processes of the SN 2 type [29]. As the Cbl moiety is sterically very demanding it is likely that GA is subject to attack at the least hindered carbon, which would lead to formation of a pair of diastereoisomers from a racemic 1,2-epoxide (Fig. 2). In earlier studies, reaction products with the more substituted carbon (C-2) in GA have not been demonstrated and are therefore assumed to be minor
V. Silvari et al. / Mutation Research 580 (2005) 91–101
99
products [14,30,31]. An exception may be reaction at oxirane C-2 in GA with glutathione, which was reported as one product in urine of rats [32]. In the case of the achiral EO the two carbons have equal reactivity and Cbl(I) necessarily gives a single product, 2-hydroxyethyl–cobalamin (Fig. 3). The nucleophiles used for characterisation of reaction-kinetic parameters were chosen on the basis of their suitability and nucleophilic strength, with the aim of creating a test battery useful for reaction-kinetic studies of oxiranes or other electrophiles (see Table 1). The n values have been determined experimentally by different methods at different laboratories and the values are encumbered with some uncertainties. To avoid influence of these uncertainties in the determination of the substrate constant s for GA, the data on rate constants were not only plotted as log k versus n (Fig. 5), but also as log kEO versus log kGA (see Eq. (8); Fig. 6): sEO log kEO = × log kGA (8) sGA
Through measurement of the disappearance of GA (and EO) in the medium used in mutation tests the mutagenic frequency per dose unit was obtained in this study. Furthermore, gamma-radiation was used as reference standard [25] in the mutation tests, which allows expression of mutagenic potency of GA as radequivalence (rad-equ/mMh). A mutation study of EO in the same system gave a value of about 0.5 mutants per 105 cells and mMh and a radiation-dose equivalence of 60 rad-equ/mMh (Jenssen et al., in preparation). These data permit a preliminary comparison of the mutagenic potency of GA and EO, which arrives at a figure of about 7 (the statistical uncertainty may amount to a factor of about 2) times higher mutagenic potency of GA per dose unit, both expressed as mMh and as radequivalents. Determination of mutagenic potency per dose unit (mMh) and radiation-dose equivalents facilitate comparison between chemicals and test systems [10]. The obtained relative mutagenic potency is compatible with relative reactivity at low n values.
The s value of 1.0, obtained for GA is similar to the s value 0.93 for EO (this work), and the value 0.96 previously determined for EO and PO [10]. More effective mutagens like alkylnitrosoureas, have lower s values (0.26 for N-ethyl-N-nitrosourea and 0.42 for Nmethyl-N-nitrosourea) [10]. GA, compared to EO, showed higher reactivity toward all nucleophiles studied (n = 0-6.4). The higher reactivity of GA may be due to the electron-withdrawing action of the carbamoyl substituent.
5. Conclusions
4.3. Reaction-kinetic parameters and relative mutagenic potency At low n values (n ≈ 1–3) GA is about 2–3.5 times more reactive than EO. Reactivity at low n values corresponding to oxygen atoms in DNA, e.g. O6 -guanine, is considered to be critical for mutation [33]. Reaction of oxiranes with phosphate oxygens in DNA (n ≈ 1–1.5) gives 2-hydroxyalkyl adducts, which favours the formation of a dioxafosfolan ring and ensuing strand breaks [34]. This would mean that GA, due to its higher reactivity, is expected to give a relatively higher frequency of strand breaks per dose unit compared to EO. In in vitro studies GA has been shown to give rise to strand breaks [24].
A battery of nucleophiles useful in reaction-kinetic studies of alkylating agents, has been applied for the determination of reaction-kinetic parameters of the simple oxiranes GA and EO. The cobalamin method has been successfully applied for this purpose. The usefulness of reaction-kinetic parameters as determinants of mutagenic potency was further demonstrated. Substrate (selectivity) constants of GA and EO were determined as well as their absolute reactivity toward nucleophiles with different nucleophilic strengths (06.4). It was shown that GA has an s value of about 1.0, similar to other simple oxiranes. GA, however, showed a higher reactivity than EO, with 2–3.5 times higher reactivities at low n values. The data thus predict that GA is expected to have slightly higher mutagenic potency compared to EO. According to mutation tests the mutagenic potency of GA is higher (about 7 times) than that of EO. The ratio is the same irrespective of whether potency is measured per dose unit in terms of mMh or in terms of radiation-dose equivalents. This estimate is comparable with expectations from the reaction-kinetic data for these compounds, considering that this is a preliminary determination of relative mutagenic potency.
100
V. Silvari et al. / Mutation Research 580 (2005) 91–101
These data show that GA is a rather weak mutagen, in line with what is expected for simple oxiranes. This kind of data are useful to evaluate GA as a cancer risk-increasing agent in exposure to acrylamide at low doses, which in combination with in vivo doses of GA could be used for cancer risk estimation (cf. [25,26]). Acknowledgements The authors are grateful to Dr. Siv Osterman-Golkar for valuable viewpoints, Ronald Davies (B.Sc.) for contribution with synthesis and Ulla Hedebrant (M.Sc.) for technical assistance. The work was financially supported by the European Commission (Contract no. HPRN-CT-2002-00195 ‘Cobalamins and Mimics’), The Swedish Research Council Formas, The National Board for Laboratory Animals, The Swedish Cancer Society, and The Swedish Cancer and Allergy Fund. References
[10]
[11]
[12]
[13]
[14]
[15]
[16] [17] [18]
[1] C.G. Swain, C.B. Scott, Quantitative correlation of relative rates. Comparison of hydroxide ion with other nucleophilic reagents towards alkyl halides, esters, epoxides and acyl halides, J. Am. Chem. Soc. 75 (1953) 141–147. [2] W.C.J. Ross, Biological Alkylating Agents, Butterworths, London, 1962. [3] S. Osterman-Golkar, Studies on the reaction kinetics of biologically active electrophilic reagents as a basis for risk estimates, Stockholm University, 1975. [4] P.D. Lawley, P. Brookes, Further studies on the alkylation of nucleic acids and their constituent nucleotides, Biochem. J. 89 (1963) 127–138. [5] A. Loveless, Possible evidence of O6 -alkylation of deoxyguanosine to the mutagenicity and carcinogenicity of nitrosamines and nitrosamides, Nature 223 (1969). [6] I. Turtoczky, L. Ehrenberg, Reaction rates and biological action of alkylating agents. Preliminary report on bactericidal and mutagenic action in E. coli, Mutat. Res. 8 (1969) 229–238. [7] S. Osterman-Golkar, L. Ehrenberg, C.-A. Wachtmeister, Reaction kinetics and biological action in barley of mono-functional methanesulfonic esters, Radiat. Bot. 10 (1970) 303–327. [8] L. Ehrenberg, E. Moustacchi, S. Osterman-Golkar, International Commission for Protection Against Environmental Mutagens and Carcinogens. Dosimetry of genotoxic agents and dose–response relationships of their effects, Mutat. Res. 123 (1983) 121–182. [9] A. Barbin, H. Bartsch, Nucleophilic selectivity as a determinant of carcinogenic potency (TD50) in rodents: a comparison of
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
mono- and bi-functional alkylating agents and vinyl chloride metabolites, Mutat. Res. 215 (1989) 95–106. E.W. Vogel, M.J.M. Nivard, L.A.B. Ballering, H. Bartsch, A. Barbin, J. Nair, et al., DNA damage and repair in mutagenesis and carcinogenesis: implications of structure–activity relationships for cross-species extrapolation, Mutat. Res. 113 (1996) 177–218. E. Tareke, P. Rydberg, P. Karlsson, S. Eriksson, M. T¨ornqvist, Acrylamide: a cooking carcinogen? Chem. Res. Toxicol. 13 (2000) 517–522. E. Tareke, P. Rydberg, P. Karlsson, S. Eriksson, M. T¨ornqvist, Analysis of acrylamide, a carcinogen formed in heated foodstuffs, J. Agric. Food Chem. 50 (2002) 4998–5006. E. Bergmark, C.J. Calleman, F. He, L.G. Costa, Determination of hemoglobin adducts in humans occupationally exposed to acrylamide, Toxicol. Appl. Pharmacol. 120 (1993) 45– 54. C.J. Calleman, E. Bergmark, L.G. Costa, Acrylamide is metabolized to glycidamide in the rat: evidence from hemoglobin adduct formation, Chem. Res. Toxicol. 3 (1990) 406–412. J. Haglund, A.L. Magnusson, L. Ehrenberg, M. T¨ornqvist, Introduction of cob(I)alamin as an analytical tool: application to reaction-kinetic studies of oxiranes, Toxicol. Environ. Chem. 85 (2003) 81–94. J. Bronsted, M. Kilpatrick, M. Kilpatrick, Kinetic studies on ethylene oxide, J. Am. Chem. Soc. 51 (1929) 428–461. L. Ehrenberg, S. Hussain, Genetic toxicity of some important epoxides, Mutat. Res. 86 (1981) 1–113. B. Paulsson, F. Granath, J. Grawe, L. Ehrenberg, M. T¨ornqvist, The multiplicative model for cancer risk assessment: applicability to acrylamide, Carcinogenesis 22 (2001) 817–819. H. Helleberg, R. Westerholm, L. Ehrenberg, M. T¨ornqvist, Synthesis and characterisation of N-alkylated l-valine methylamide products from diol epoxid metabolites of fluoranthene and benzo[a]pyrene, Acta Chem. Scand. 52 (1998) 797–805. J. Koskikallio, Nucleophilic reactivity. Part III. Kinetics of reactions of methyl perchlorate with amines in water and ethanol, Acta Chem. Scand. 23 (1969) 1490–1494. V. Poirier, C.J. Calleman, The nucleophilic reactivity of valine methylamide. A model of the N-terminal valine residues of hemoglobin, Acta Chem. Scand., B 37 (1983) 817–822. P. Virtanen, Kinetics of the reactions of ethylene oxide with nucleophiles, Ann. Acad. Sci. Fennicae Ser. A II 124 (1963) 1–89. D. Jenssen, A quantitative test for mutagenicity in V79 Chinese hamster cells, in: W.N.C.R.B.J. Kilbey (Ed.), Handbook of Mutagenicity Test Procedures, Elsevier, Amsterdam, 1984, pp. 269–290. F. Johansson, T. Lundell, P. Rydberg, K. Erixon, D. Jenssen, Mutagenicity and DNA repair of glycidamide-induced adducts in mammalian cells, Mutat. Res. (2004) (this volume). L. Ehrenberg, Purposes and methods of comparing effects of radiation and chemicals, in: Radiological Equivalents of Chemical Pollutants, IAEA, Vienna, 1980, pp. 23–36. F.N. Granath, C.E. Vaca, L.G. Ehrenberg, M.A. T¨ornqvist, Cancer risk estimation of genotoxic chemicals based on target dose and a multiplicative model, Risk Anal. 19 (1999) 309–320.
V. Silvari et al. / Mutation Research 580 (2005) 91–101 [27] T. Alsberg, J. Minten, J. Haglund, M. T¨ornqvist, Determination of hydroxyalkyl derivatives of cobalamin (Vitamin B12) using reversed phase high performance liquid chromatography with electrospray tandem mass spectrometry and ultraviolet diode array detection, Rapid Commun. Mass Spectrom. 15 (2001) 2438–2445. [28] J. Haglund, W. Van Dongen, F. Lemiere, E.L. Esmans, Analysis of DNA-phosphate adducts in vitro using miniaturized LC–ESI–MS/MS and column switching: phosphotriesters and alkyl cobalamins, J. Am. Soc. Mass Spectrom. 15 (2004) 593–606. [29] G.N. Schrauzer, E. Deutsch, Reactions of cobalt(I) supernucleophiles. The alkylation of Vitamin B12s , cobaloximes(I), and related compounds, J. Am. Chem. Soc. 91 (1969) 3341–3350. [30] B. Paulsson, I. Athanassiadis, P. Rydberg, M. T¨ornqvist, Hemoglobin adducts from glycidamide: acetonization of hydrophilic groups for reproducible gas chromatography/tandem mass spectrometric analysis, Rapid Commun. Mass Spectrom. 17 (2003) 1859–1865.
101
[31] D. Segerb¨ack, C.J. Calleman, J.L. Schroeder, L.G. Costa, E.M. Faustman, Formation of N-7-(2-carbamoyl-2hydroxyethyl)guanine in DNA of the mouse and the rat following intraperitoneal administration of [14C]acrylamide, Carcinogenesis 16 (1995) 1161–1165. [32] S.C. Sumner, J.P. MacNeela, T.R. Fennell, Characterization and quantitation of urinary metabolites of [1,2,313C]acrylamide in rats and mice using 13C nuclear magnetic resonance spectroscopy, Chem. Res. Toxicol. 5 (1992) 81– 89. [33] D. Jenssen, J. Johansson, T. Helleday, C. Lundin, K. Erixon, Pathways involved in mammalian mutagenesis, in: M. RuizRubio, E. Alejandre-Dur´an, T. Rold´an-Arjona (Eds.), DNA Repair and Mutagenesis, Research Signpost, India, 2002, pp. 97–110. [34] S. Walles, L. Ehrenberg, Effects of beta-hydroxyethylation and beta-methoxyethylation on DNA in vitro, Acta Chem. Scand. 22 (1968) 2727–2729.
