Formamidopyrimidine adducts are detected using the comet assay in human cells treated with reactive metabolites of 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK)

Abstract

4-(Methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) is the most lung-specific of the carcinogens present in tobacco smoke. Its bioactivation in cells leads to a small amount of methylation or pyridyloxobutylation DNA damage. Considering its great sensitivity, the comet assay seems a technique of choice to investigate NNK-related damage. Several strategies were used to impart some specificity to the assay: (1) using analogs that produce a limited variety of DNA lesions, as they mimic either the methylation or the pyridyloxobutylation pathway; (2) using cells with different bioactivation abilities; (3) using alkali conversion and/or enzymes specific for cleaving particular classes of damage; (4) using different lysis conditions to convert a specific class of DNA lesions into enzyme-sensitive lesions. We determined that several NNK-associated lesions can be detected with some specificity with the comet assay. For the methylation pathway, they are AP sites and the more frequent formamidopyrimidine (fapy) adducts. These fapy adducts correspond to N7-methylguanines generated in the cells that were ring-opened during the assay by the lysis solution at pH 10. For the pyridyloxobutylation pathway, alkylphosphotriesters and a roughly equal frequency of fapy sites were detected. By analogy to the methylation damage, these fapy adducts are thought to be the ring-opened form of N7-pyridyloxobutylguanines (N7-pobG). N7-pobG are unstable and this constitutes the first indirect demonstration of their formation in cells. But contrary to N7-m-fapy, the lysis time or pH did not influence the frequency of N7-pob-fapy adducts detected, suggesting that they already exist in the cells and are not related to the experimental conditions. These N7-pob-fapy have a strong mutagenic potential and we think that the comet assay, in spite of its limitations, is a good way to study them considering their low frequency and the inherent instability of the adduct from which they originate.

Introduction

N-Nitrosamines are thought to play a major role in carcinogenesis related to tobacco consumption. Among the tobacco specific nitrosamines, 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) is the most abundant and the strongest carcinogen [1], [2]. It exhibits lung-selective toxicity and induces primarily adenocarcinoma of the lung on laboratory animals [3]. This led to the assumption that NNK accounts for recent increases in the relative incidence of human lung adenocarcinoma in smokers [4], [5].

Bioactivation to reactive intermediates is necessary for NNK to be genotoxic and its activation has been extensively studied [6] in freshly isolated human lung cells [7] and in peripheral human lung microsomes [8]. The two α carbons of the nitrosamino group of NNK can be hydroxylated, yielding unstable intermediates that can lead to DNA alkylation (Fig. 1). The α-methylene hydroxylation of NNK leads to the formation of methyldiazohydroxides, which can methylate DNA and the α-methyl hydroxylation of NNK results in the formation of oxobutyldiazohydroxides, which can alkylate DNA through a reaction of pyridyloxobutylation. Both types of α-carbon hydroxylation clearly exist in human lung cells [3], [7] and both pathways seem to participate in NNK-related lung carcinogenesis in humans [3], [9]. But the extent to which they do, remains to be determined. These pathways can be investigated separately through the use of analogs that generate either one type of reactive species or the other. For instance, N-nitroso(acetoxymethyl)methylamine (NDMAOAc) and 4-(acetoxymethylnitrosamino)-1-(3-pyridyl)-1-butanone (NNKOAc) mimic respectively the methylation or the pyridyloxobutylation pathway (Fig. 1). NNK bioactivation, that is the biotransformation in alkylating reactive species, is not as frequent as other possible biotransformations corresponding for example to detoxification pathways, but it is thought to be responsible for the carcinogenetic effect of NNK. In fact, the relative occurrence of bioactivation in reactive species seems significantly higher in alveolar type II cells and macrophages than in other pulmonary cell types, making them potential target for NNK toxicity [7].

DNA adducts generated by NNK through the methylation or the pyridyloxobutylation pathway (Table 1) have been extensively studied to determinate which of them could be relevant to carcinogenesis [3]. Laboratory animals treated with NNK show the following adducts in their lung: N7-methylguanine (N7-mG), O⁶-methylguanine (O⁶-mG) and to a lesser extend of O⁴-methylthymine (O⁴-mT), all these are due to the methylation pathway. Some pyridyloxobutyl adducts have also been detected such as O⁶-pyridyloxobutylguanine (O⁶-pobG) or phosphate adducts (pyridyloxobutylphosphotriesters) [6], [10], [11]. NNK generates single-strand breaks [6] but exactly how they are produced remains undetermined. It has been suggested that they derived from the methylation pathway by spontaneous or enzymatic depurination of N7-mG or N3-mA [3] but they more likely result from the pyridyloxobutylation pathway by the generation of pyridyloxobutylphosphotriesters [10], [11]. The O⁶-alkyl adducts (O⁶-mG and O⁶-pobG) are considered as major players in NNK-related carcinogenicity because of their mutagenetic potential and the fact that they are repaired both by the same O⁶-methylguanine methyltransferase which is easily overwhelmed by low adduct levels [3]. Still, carcinogenicity of the pyridyloxobutylation pathway has been also associated with the presence of another type of damage. This adduct type, that accounts for at least 50% of the DNA binding obtained through this pathway, is the most prevalent adduct from the pyridyloxobutylation pathway and remained uncharacterized until recently. It was therefore only qualified according to its ability to release 4-hydroxy-1-(3-pyridyloxobutyl)-1-butanone (HPB) (Fig. 1) upon acid or neutral thermal hydrolysis. Recently, several HPB-releasing adducts have been obtained and identified in DNA and/or deoxyguanosine treated with NNKOAc in vitro: N7-pyridyloxobutylguanine (N7-pobG) [12] and O²-pyridyloxobutylcytosine (O²-pobC) [13], the former being by far the most abundant [14]. Other adducts have also been characterized that do not release HPB. They derive from the pyridyloxobutylation of the N² position of guanine [12] and more frequently of the O² position of thymine [13]. None of these newly characterized adduct (HPB-releasing or not) have yet been demonstrated to occur in animals or cells, probably because of their low frequency of occurrence and/or their instability, which makes them difficult to observe [12], [14].

The comet assay is a commonly used, sensitive method to investigate DNA damage in individual cells [15]. It has already been used to detect NNK-associated damage in the metabolically competent lymphoblastoid cell line MCL-5 [16] and in white blood cells where NNK seems to act synergistically with UVA to generate DNA damage [17]. In both studies, the type of damage detected could not be identified. In fact, this technique completely lacks specificity in that it can determine only one parameter: the average break frequency. On the contrary, a technique like ligation-mediated PCR (LMPCR), which is commonly used in DNA damage studies, gives several information concerning the breaks: the exact genomic base position along with some chemical information such as the existence of a 5′ phosphate termination. Nevertheless, the comet assay is extremely sensitive, detecting breaks as infrequent as 1 break/10⁸ bases [18] versus 1 break/5 × 10⁴ bases for LMPCR [19] and can therefore be useful to detect DNA damage that is infrequent. In fact, specificity must be sacrificed for sensitivity in cases like NNK-associated damage where damage frequency can be limited by factors such as the ability of the cell to biotransform the exogenous pre-mutagen into intracellular DNA-damaging species. Moreover, some specificity can be imparted to the assay by the following treatments: (1) using analogs that produce a limited variety of DNA lesions such as NDMAOAc or NNKOAc; (2) using a variety of cells with different activation abilities: U937, a histiocytic lymphoma cell line that bioactivates NNK [20], NCI-H23, a lung adenocarcinoma cell line originally obtained from a smoker and normal lymphocytes; (3) generating the single-strand breaks from DNA using chemical treatment (comet assay at different pH) and/or enzymes specific for cleaving particular classes of DNA lesions; (4) using chemical treatments (different lysis conditions) that convert a specific class of DNA lesions into enzyme-sensitive lesions. We used all four specificity enhancement methods to distinguish various types of NNK-associated damage that can be specifically studied using the comet assay.