Mutagenicity testing of 9-N-substituted adenines and their N-oxidation products.

Adenine together with certain 9-N-substituted derivatives such as 9-methyl, 9-benzyl, 9-benzhydryl, and 9-trityl were tested against Salmonella typhimurium strains TA97, TA98, and TA100 in the absence and presence of rat hepatic S9 prepared from Aroclor 1254 pretreated rats. All compounds were positive toward TA98 in the presence of the metabolic activating system, whereas they all lacked mutagenic activity in the absence of S9, and toward TA97 and TA100 with or without S9 when tested at 100 ng/plate. A similar pattern was observed for the corresponding 1-N-oxides. 6-Hydroxylaminopurine was not mutagenic toward TA100 at 100 ng/plate, whereas it was toxic toward TA97 and TA98 at this level. When tested at 1 ng/plate, hydroxylaminopurine was still toxic to TA98 but produced twice the spontaneous reversion rate to TA97 without metabolic activation. Surprisingly, 9-methyl-6-hydroxylaminopurine was only active toward TA98 in the presence of S9, whereas 9-benzyl-6-hydroxylaminopurine was highly active toward TA97 and TA100 in the absence of S9 and even more active in the presence of S9. This compound was inactive toward TA98 in the absence of S9. The results generally support the concept that nuclear N-oxidation of aminoazaheterocycles is a detoxication process, whereas N-hydroxylation of the exo amino group is a toxication reaction.


Introduction
Aminoazaheterocylcles are a class of compounds characterized by an aromatic nuclear nitrogen atom(s) together with an ortho exo-amino group. These compounds have diverse pharmacological properties, are widely used in clinical medicine, and occur as natural products [(1-3) Fig. 1] or are formed as metabolites of drugs ( Fig. 2). They may also be produced during the cooking of foods or pyrolysis of proteins and amino acids [(4,5) Fig. 3].
The metabolism of many of these compounds has been studied (1)(2)(3), and generally it has been observed that compounds used in clinical medicine are not genotoxic and are metabolized by oxidation at one or more nuclear nitrogens (6)(7)(8)(9)(10)(11)(12), whereas those produced by pyrolysis are potent mutagens (13) and are metabolized by N-hydroxylation of the exo amino group (14)(15)(16). This has led to the concept (1-3) that nuclear N-oxidation of aminoazaheterocycles is a detoxication process, whereas N-hydroxylation of the exo amino group is a toxication reaction (Fig. 4).
Ortho aminoazaheterocycles are usually oxidized at either one or other ofthe constituent nitrogens. Only in the case of the nonmutagenic 2-aminopyridine has both pathways of N-oxidation been observed (17,18); di-Noxygenation has never been found with these compounds. Direct proof of the toxication/detoxication concept of artho aminoazaheterocycle metabolism and toxicity has therefore not been possible. Adenine (Fig. 5) is a naturally occurring artho aminoazaheterocycle that is capable of being chemically converted to both the 1-N-oxide (19) and 6-hydroxylaminopurine (20). Similar pairs of N-oxidation products may be prepared from 9-substituted adenines which themselves differ in their physicochemical properties (12). Certain 1-N-oxides and 6-hydroxylamine derivatives of 9-substituted purine have been compared with their parent compounds and used to probe the toxication/detoxication concept using various Salmonella typhimurium strains with and without a hepatic activation system.
6-Hydroxylamino-9-benzylpurine (HBP) was synthesized by the reaction of 9-benzyl-6-chloropurine with an ethanolic solution of hydroxylamine. Hydroxylamine hydrochloride (7.8 g) dissolved in boiling, absolute ethanol (130 mL) was mixed with a solution ofpotassium hydroxide (6.8 g) in hot ethanol (25 mL). The potassium chloride that formed was filtered off and washed three times with hot ethanol (15 mL). The ethanolic hydroxylamine solution obtained was added to a boiling solution of 9-benzyl-6chloropurine (1.87 g), prepared according to the method of Montgomery and Temple (21), in absolute ethanol (78 mL). The mixture was refluxed for 6 hr, after which the solution was left to cool overnight. HBP precipitated; was collected, and recrystallized from absolute ethanol (yield 78%, melting point 226-228°C with decomposition).

Determination of Mutagenic Potential
The compounds were tested against Salmonella typhimurium strains TA97, TA98, and TA100 exactly as described by Maron and Ames (22) at a concentration of 100 ng/plate or less dissolved in dimethylsulfoxide (1 ,ug/ mL). The viability ofthe systems were checked by incorporating either 9-aminoacridine (20 jig/plate; TA97), 2-nitrofluorene (2 ,ug/plate; TA98), or methyl-N-nitro-Nnitrosoguanidine (2 ,ug/plate; TA100). 2-Aminoanthracene (5 ,ug/plate) was used with all strains in the presence of an S9 metabolic activating system. The S9 metabolic activating system was prepared frormL rats that had been pretreated with Aroclor 1254 (a single IP dose of 500 mg/kg, the animals were killed on day 5 after administration) and contained S9 at a 10% v/v suspension. GuaRsulian Nucleic acid base
pounds (23). The results of the mutagenicity study carried Cycloguanil out in the absence or presence of a metabolic activating Antimalarial system are presented in Table 1.
The results in the absence of a metabolic activating system may be summarized as follows: a) the parent purines did not demonstrate enhanced mutagenicity when jX NH. where available, were generally more toxic and more mutagenic than the base or the N-oxide toward TA97 and TA100 but not toward TA98. This suggests that they may be NHJ,<J1°NH. acting by either an intercalating or base-pair mechanism. o)0 N However, this was not true for the 9 In the presence of a metabolic activating system, the results observed were a) the parent compounds were ural aminoazahetero-inactive toward TA97 and TA100, whereas they were active toward TA98; b) the 1-N-oxides were also inactive towards TA97 and TA100 but again were active towards TA98; c) the hydroxylamines were active toward both TA97 and TA100 as well as TA98. Additionally, in all cases where mutagenic activity was demonstrated in the absence of a metabolic system, an enhanced mutagenic effect was observed after incorporation of the metabolic activating system. In the case of compounds that failed to demonstrate mutagenic activity (e.g., bases and 1-N-oxides in the absence of an activating system) addition of the system allowed mutagenic activity to be expressed.  Previous studies on the metabolism -of 9-substitutedadenines showed that they were only coverted to 1-Noxides by a phenobarbital-inducible cytochrome P450 species, presumably CYP2B. This enzyme failed to oxidize adenine to adenine-1-N-oxide (11,12,24). Using rats pretreated with either 3-methylcholanthrene or isosafrole (inducers of CYP1), Clement and Kunze (25) demonstrated that adenine, at high substrate concentrations, was con-verted to 6-hydroxylaminopurine. It is likely that the activating system prepared from rats pretreated with Aroclor 1254 (an inducer of both CYP1 and CYP2B), used in the present study is also capable of metabolizing adenines to the corresponding 6-hydroxylaminopurines. However, as the 6-hydroxylaminopurines tested generally showed an enhanced activity in the presence ofthe activating system, it may be that further oxidation of the 6-hydroxylamino function is occurring, producing a nitroso or nitro compound or an intermediate nitrenium ion or nitroxide radical (26,27). Such intermediates are known to react with cellular macromolecules and may initiate their mutagenic effect via this mechanism (28,29).
The results obtained in the present work are generally consistent with previous observations on the genotoxicity of N-oxidized derivatives of aminopurines. Thus, 6-hydroxylaminopurine induced mitotic inhibition and nuclear degeneration in mouse sarcoma 180 cells in vitro and prolonged the survival time of mice bearing sarcoma 180 ascites cells (30,31). This substance is genotoxic to both eukaryotes and prokaryotes (32)(33)(34)(35)(36).
Our results agree with McCartney et al. (35), who also found that adenine was inactive toward TA100, whereas 6-hydroxylaminopurine was mutagenic. This latter compound was found by these authors to be inactive toward TA98, whereas in our experiments we found that HP was extremely toxic toward this strain.
It has been proposed that HP acts by forming a desoxynucleoside triphosphate, which is incorporated into DNA (36), although other evidence (35) indicates that the mutagenic activity of 6-hydroxylaminopurine is caused by direct adduct formation rather than by base incorporation. This is supported by results obtained using a series of E. lable 1. Mutagenicity of adenine derivatives toward S. typhimurium strains with ( + ) and without ( -) a metabolic activating system.a TA97 TA98 TA100 Compound Spontaneous reversion rate I a. Adenine b. Adenine-1-N-oxide c. 6-Hydroxylaminopurine (1 ng/plate) II a. 9-Methyladenine b. 9-Methyladenine-1-N-oxide c. 9-Methyl-6-hydroxylaminopurine III a. 9-Benzyladenine b. 9-Benzyladenine-1-N-oxide c. 9-Benzyl-6-hydroxylaminopurine (50 ng/plate) IV a. 9-Benzhydryladenine b. 9-Benzhydryladenine-1-N-oxide c. 9-Benzhydryl-6-hydroxylaminopurine aResults are presented as histidine revertants/plate (mean ± SD oftriplicates). Each study was repeated at least once. Unless otherwise stated, each compound was tested at a concentration of 100 ng/plate. Compounds were dissolved in dimethylsulfoxide (1 ,ug/mL). The activation system contained hepatic S9 fraction (10% v/v) from Aroclor 1254 pretreated rats. coli mutants, which led Murray (37) to propose that "the major mechanism for hydroxylaminopurine is due to the reaction of these compounds with the 06 position of guanine and the 04 position of thymine." Administration of 6-hydroxylaminopurine to 10 female Wistar rats produced only one tumor at 15 months and was completely inactive in Sprague-Dawley rats; the authors (38) concluded that "its oncogenicity is marginal at best." Barrett (33), citing the same publication (38), interprets the data as meaning that HP is weakly carcinogenic; this is supported by the present studies where HP only displayed a weak mutagenic response. 6-Hydroxylaminopurine is also teratogenic (30).
Studies on the toxicity of adenine-1-N-oxide are more controversial. In an early publication (39), it was stated that "Administration of adenine-1-N-oxide for 6 months failed to produce tumors in any tissues of 13 rats during the experimental period of 15 months." In a later paper (38), it was reported that "Subsequent assays have led to tumours in both Sprague-Dawley and Wistar rats; at a dose level of 10 mg/week for 26 weeks, tumours have been induced at the site of injection of 33 of 41 rats." Adenine-1-N-oxide was inactive as a sulfate acceptor (40), a process that has been proposed as the second activating step for hydroxylamines and hydroxamic acids (26,27), although in vivo acetylation to 1-acetoxyadenine produced a reactive species toward Bacillus subtilis-transforming DNA and can be considered as a weak mutagen (41). The role of purine N-oxides in cancer has been reviewed (43).
The limited results we have obtained in the Ames mutagenicity test are not inconsistent with the metabolic-Noxygenation hypothesis of toxication/detoxication of aminoazaheterocycles, which is a good guide to genotoxicity. Clearly, more bases and their corresponding N-oxides and hydroxylamines need to be prepared and tested at various concentrations using additional sophisticated bacterial test strains lacking specific xenobiotic-metabolizing enzyme systems before the concept can be fully substantiated.