ABSTRACT
The underlying principle of chemical carcinogenesis [1-3] is the concept that “chemical structure defines toxicity” Vast investigations and experimental studies are available in the literature on chemical carcinogenesis This chapter provides a brief introduction and a few illustrations of chemical carcinogenesis to appreciate the concept and practice of structural alerts and their application Well-known and representative structural alerts for genotoxicity, mutagenicity, and carcinogenicity in the literature are included
The observations in the 1700s of nasal cancer being a consequence of excessive use of tobacco snuff and skin cancer of the scrotum among chimney sweeps were the beginning of knowledge on chemical carcinogenesis A number of investigations to demonstrate the carcinogenic activity of soots and tars in experimental animals and to search for the active agents resulted in the identification of polycyclic aromatic hydrocarbons (PAHs) as the responsible carcinogens Dibenz[a,h]anthracene was the first ever carcinogen synthesized, and other PAHs, including benzo[a]pyrene, 3-methylcholanthrene, and 7,12-dimethylbenz[a]anthracene, have been extensively studied Some of the structures of chemical carcinogens discovered or identified prior to 1960 are shown in Figure 21
As the number and variety of known chemical carcinogens increased, it became evident that these chemicals lacked a common structural feature Some carcinogens, such as aromatic amine derivatives, caused tumors at distant sites (liver and urinary bladder) regardless of the route of administration, providing an early clue to the involvement of metabolism The importance of metabolic activation in chemical carcinogenesis was stated by Miller: “The great majority of chemical carcinogens were active only after metabolism to ultimate carcinogens (ie, the derivatives that actually initiate the neoplastic event) The known exceptions are the carcinogens that are alkylating or acylating agents per se Further, the data then available suggested that the ultimate forms of chemical carcinogens might all be strong electrophilic reactants The known ultimate carcinogens contain relatively electron-deficient atoms
References 90
that seek to react with nucleophilic sites, ie, atoms that have easily shared electrons These nucleophilic sites are relatively abundant in DNA’s, RNA’s, and proteins” [1] Accordingly, a common property of ultimate carcinogens is electrophilicity
Precarcinogens can be metabolized to more than one ultimate carcinogen, and there are many nucleophilic sites in target macromolecules (DNAs, RNAs, and proteins); therefore, it is possible to have multiple macromolecule-bound derivatives of each precarcinogen The possibility that the major adducts are not necessarily the most important ones in carcinogenesis, in addition to the multiplicity of adducts, gives rise to the complexity of metabolic activation in carcinogenesis Figure 22
shows examples of strong electrophilic reactants (positive ions or molecules with electron-deficient atoms) and their reactions with nucleophilic sites in DNA, RNA, and proteins (such as oxygen, nitrogen, and sulfur atoms)
A few example compounds with reported chemical carcinogenesis are shown with proposed mechanisms to illustrate the importance of metabolic bioactivation Simply alkylating agents, such as alkyl halides, were not selected, as they proceed via direct alkylation on macromolecules (DNAs, RNAs, and proteins) due to their chemical reactivity Rather, highly potent compounds inducing neoplasia (tumor) that require metabolic activation (in this sense, they are procarcinogens) are chosen Four examples shown in the literature are benzo[a]pyrene, aflatoxin B1, aristolochic acid, and aromatic amines These chemical carcinogens are, strictly speaking, precarcinogens that form the electrophilic ultimate carcinogens via metabolic activation
Benzo[a]pyrene is a ubiquitous environmental pollutant produced during combustion processes and also found in tobacco and diet [4] It is the major genotoxic component in cigarette smoke and was the cause of the high rate of cancer in Britain among chimney sweeps The majority of human exposure occurs through smoking and consumption of charbroiled foods by ingestion, inhalation, and dermal absorption Benzo[a]pyrene must be activated by enzymes such as cytochrome P450 (P450) to acquire its mutagenic and carcinogenic properties The metabolic activation pathway is shown in Figure 23 The first step is the formation of the 7,8-epoxide, followed by hydrolysis by microsomal epoxide hydrolase to the 7,8-diol metabolite The 7,8-diol metabolite is activated by P450 further and becomes the mutagenic diol epoxide The diol epoxide is extremely reactive and binds to macromolecules, including DNAs, RNAs, and proteins
Aflatoxins’ name originated from Aspergillus flavus toxins, as they are naturally occurring mycotoxins produced by a species of fungi, A. flavus Aflatoxin B1 is
acutely toxic to most animal species and is a potent mutagen and a potent hepatocarcinogen [5] A major source of human exposure to aflatoxins is dietary contamination with molded grain The formation of a DNA adduct with aflatoxin B1 through metabolic activation is shown in Figure 24 Aflatoxin B1 is first converted to the reactive 2,3-exo-epoxide via microsomal conversion The reactive 2,3-epoxide then reacts with nucleophilic sites of DNA, the N7 atom of guanine in the example
shown in Figure 24, covalently binding to DNA Aflatoxin is one of the three compounds, together with N-nitroso and azoxy compounds, named as high-potency genotoxic carcinogens in the European Medicines Agency [6] and Food and Drug Administration [7] guidelines on genotoxic impurities
Aristolochic acids are a family of structurally related nephrotoxic and carcinogenic nitrophenanthrene compounds found in Aristolochia herbal plants Some of the Aristolochia herbal plants have been used worldwide, especially in Asia, for medical purposes for over 2000 years [8] Exposure to aristolochic acids has been linked to nephropathy (kidney disease) and urothelial cancer of the upper urinary tract Aristolochic acids are enzymatically metabolized via the reduction of the nitro group to produce the reactive N-hydroxyaristolactam The N-hydroxyaristolactams undergo the postulated nitrenium intermediates, which covalently bind to the exocyclic amino groups of the purine nucleobases, as shown in Figure 25
Exposure to aromatic amines has been linked to human urinary bladder cancer Aromatic amines, and any other structures that can be metabolized into aromatic amines, are readily converted into the extremely electrophilic nitrenium ion, which in turn forms covalent adducts with biological macromolecules such as DNA [9,10] The major metabolic pathway of aniline is shown in Figure 26 The first step involves N-oxidation to N-hydroxyarylamines by P450, followed by esterification by enzymes such as O-acetyltransferase, phosphotransferase, and sulfotransferase On leaving the corresponding ester group, a reactive electrophilic intermediate, nitrenium ion, is formed, which is the ultimate carcinogen Subsequently, the nitrenium ion reacts with a macromolecular nucleophile such as DNA to form covalent adducts resulting in genetic mutation and/or cancer
These four examples clearly illustrate the significance of metabolic activation in the induction of mutation and tumors by most chemical carcinogens
Experimental determination of chemical carcinogenesis involves timeconsuming, labor-intensive, and expensive animal testing, and experimental testing data on carcinogenicity are available for only a relatively small portion of chemicals in commerce today Therefore, there is a strong demand to develop general prediction models that can be used to predict carcinogenicity for chemicals, for which no experimental data are available and possibly will not be available any time soon
The relationship of the chemical structure of a molecule and its biological activity is called the “structure-activity relationship” (SAR) In cases where the biological activity is toxic effects, the toxic properties are often related to chemical structures, specifically, to particular substructures These particular substructures are generally identified as toxicophores (or structural alerts) In 1985, Ashby [11] published a hypothetical compound structure with chemical functionalities (structural alerts) that may contribute to potential carcinogenicity, to demonstrate the relationship of chemical structures and biological activities The structure, famously known as Ashby’s polycarcinogen, is shown in Figure 27 adapted from the studies of Ashby and others [11-14] with the list of known structural alerts included in the polycarcinogen This was the first graphical display of known structural alerts to carcinogenicity
Cramer and others [15] developed a decision tree to estimate toxic hazard of chemicals based on information including SARs, metabolic mechanisms, chemical reactivity, and human exposure Chemicals are categorized into three structural
1 Aryl amines 2 Ring epoxides 3 Alkane sulfonate esters 4 Aryl nitro groups 5 Azo groups 6 Ring N-oxides 7 Dimethyl amines 8 Methylols 9 Aliphatic aldehydes 10 Ring vinyl groups 11 Aziridine groups 12 Chloramines 13 Nitrogen mustards 14 Benzyl halides 15 Alkyl urethanes 16 Alkyl nitrosamines
classes based on 33 questions The 33 questions are tabulated in Table 21, and the decision flow occurs according to “yes” or “no” answers Cramer emphasized that the decision tree and classification was intended to be used as a guide to the acquisition of toxicity/safety data, but not as a substitute for the data
Class I (low-toxicity) substances are those with a low order of oral toxicity and are a low priority for investigation Class III (serious-toxicity) substances are those that permit no strong initial presumptions of safety, or may even suggest significant toxicity, deserving high priority for investigation Class II (moderate-toxicity) substances are simply intermediates that are less clearly innocuous than those of class I substances but do not offer the basis to be in class III The structural features of class III substances indicative of a high potential for toxicity [16] are listed here Metabolic bioactivation is an inherent part of the Cramer classification as many of these class III structures undergo metabolic bioactivation to form potentially toxic chemical entities:
• Aliphatic secondary amino-, cyano-, N-nitroso-, diazo-, triazeno-, quaternary nitrogen
• Unionized substituents containing elements other than carbon, hydrogen, oxygen, nitrogen, or sulfur (divalent), for example, halogeno-compounds
• Safrole-like compounds • Fused lactone or α,β-unsaturated lactone • Three-membered heterocyclics, for example, epoxides • Unsubstituted heteroaromatic compounds • Three or more different functional groups (excluding the methoxy group
and considering acids and esters as one group) • Unsubstituted aromatic hydrocarbons • Compounds without a strongly anionic group for every 20 (or fewer) carbon
(C) atoms (for compounds not classified at earlier steps)
Another effort in the literature to predict the genotoxicity of chemicals by structure classification was demonstrated by the general toxicophore concept for mutagenicity Mutagenicity is the ability of a compound to cause mutations in DNA A compound can be mutagenic because of its direct reactivity toward DNA, reactive metabolite formation by the metabolic activation of a nonreactive compound, or intercalation Intercalation is a reversible, noncovalent process in which a compound (such as PAH) inserts itself between and parallel to base pairs of the DNA double helix, forming stable π-stacking interactions Kazius and others [17] published a list of eight general toxicophores for mutagenicity prediction by a substructure search of a data set A data set of over 4000 molecular structures with corresponding Ames test data (both mutagens and nonmutagens) was evaluated by a substructure search It showed that most mutagens were detected by applying only eight general toxicophores with a satisfactory prediction rate The names, substructures, and examples of the eight toxicophores are shown in Table 22 A more specific final set 29 toxicophores was proposed in the report that could be applied to risk assessment processes
Structural alerts were described by Benigni and others [18] as molecular substructures or reactive groups that are related to carcinogenic and mutagenic properties
TA B
LE 2
.1 Q
ue st
io ns
A sk
ed in
t he
D ec
is io
n Tr
ee A
pp ro
ac h
fo r
Es ti
m at
io n
of T
ox ic
H az
ar d
TA B
LE 2
.1 (
C on
ti nu
ed )
Q ue
st io
ns A
sk ed
in t
he D
ec is
io n
Tr ee
A pp
ro ac
h fo
r Es
ti m
at io
n of
T ox
ic H
az ar
d
TA B
LE 2
.1 (
C on
ti nu
ed )
Q ue
st io
ns A
sk ed
in t
he D
ec is
io n
Tr ee
A pp
ro ac
h fo
r Es
ti m
at io
n of
T ox
ic H
az ar
d
TA B
LE 2
.1 (
C on
ti nu
ed )
Q ue
st io
ns A
sk ed
in t
he D
ec is
io n
Tr ee
A pp
ro ac
h fo
r Es
ti m
at io
n of
T ox
ic H
az ar
d
TA B
LE 2
.1 (
C on
ti nu
ed )
Q ue
st io
ns A
sk ed
in t
he D
ec is
io n
Tr ee
A pp
ro ac
h fo
r Es
ti m
at io
n of
T ox
ic H
az ar
d
of the chemicals, and as a representation of “codification” aimed at highlighting the mechanism of action of the mutagenic and carcinogenic chemicals The identification of the structural alerts has facilitated the understanding of mechanisms and assessment of the risk posed by chemicals The carcinogens are classified into either genotoxic carcinogens or epigenetic carcinogens from the point of view of the mechanism of action Genotoxic carcinogens cause damage directly to DNA, and many known mutagens are in this category with mutation being one of the first steps in cancer development In contrast, epigenetic carcinogens do not bind covalently to
TABLE 2.2 Eight General Toxicophores for Mutagenicity Prediction
DNA, do not cause direct DNA damage, and are usually negative in standard mutagenicity assays Genotoxic carcinogens have a unifying feature of electrophilicity per se or via metabolic activation With this electrophilicity hypothesis [1,2], several chemical functional groups and substructures (structural alerts) were identified for genotoxic carcinogens based on mechanisms of action and metabolic fate However, the recognition of structural alerts for nongenotoxic carcinogens is not well established due to the lack of a unifying theory
Structural alerts for mutagenicity and carcinogenicity, known as the Benigni/ Bossa rulebase, with corresponding structures are listed in Table 23 [18-20] Each of the structural alerts is a code for a well-characterized chemical class, with its own specific mechanism of action
The chemical category principle states that similar chemicals should have similar toxicological profiles enabling placement into a chemical category Such categories can be used to predict a range of toxicological end points for chemicals for which no toxicological data exist The key step is the ability to select similar chemicals, and one of the most powerful methodologies to this end is based on a common mechanism of action The term mechanism of action refers to the chemical mechanism by which a chemical forms a covalent adduct with a biological macromolecule, such as DNA, RNA, or proteins The information on mechanisms of action needs to be defined through chemical structures This is the underlying principle of chemical carcinogenesis: “chemical structure defines toxicity”
Covalent adduct formation has been defined as the molecular initiating event, an event that is the first step in a series that can ultimately lead to a toxic response [9] Although a number of biological steps are required for a toxic response, the molecular initiating event must occur to initiate the remaining steps in the adverse outcome pathway The chemistry controlling covalent adduct formation could be utilized to construct chemical categories The knowledge of well-established mechanistic organic chemistry can be used to define the structural requirements of an exogenous chemical to form a covalent bond with a biological macromolecule In this case, the chemical mechanism refers to the type of covalent bond-forming reaction (molecular initiating event) between the exogenous chemical and a biological nucleophile In contrast, the biological mechanism refers to the entire pathway after the molecular initiating event to a toxic response This type of mechanistic chemistry analysis has been used in the development of chemical categories
It is important to understand that a chemical in a mechanistic category with a reactive structural feature, based on the ability to undergo a common molecular initiating event, does not necessarily mean that the chemical is toxic Other factors including toxicokinetic or toxicodynamic profiles could prevent the completion of the adverse outcome pathway An example is that aromatic amines with a sulfate substituent attached to the aromatic ring are not mutagenic This is related to the increased detoxification of the chemicals due to increased solubility and hence lower availability in a cell or in vivo This so-called mitigating factor [9] is an additional structural feature present within a molecule that removes the toxic effect In the example of the aromatic amine, the sulfate group is the mitigating factor as its presence within the molecule removes the mutagenic activity
TABLE 2.3 Structural Alerts for the Rodent In Vivo Micronucleus Assay
TABLE 2.3 (Continued ) Structural Alerts for the Rodent In Vivo Micronucleus Assay
TABLE 2.3 (Continued ) Structural Alerts for the Rodent In Vivo Micronucleus Assay
TABLE 2.3 (Continued ) Structural Alerts for the Rodent In Vivo Micronucleus Assay
The terms “mechanism of action” and “mode of action” were defined by Aptula and others [21] with respect to a defined domain of applicability, that is, identification of the range of compounds for which a quantitative structure-activity relationship (QSAR) can confidently be applied for the purposes of toxicity prediction In this context, the chemical mechanism of action was defined as “what the toxic chemical and the biological chemicals do together in vivo” and the biological mechanism of action comprises the events that occur at the cellular or subcellular level subsequently The mode of action was defined as “what happens to the organism” In the example of skin sensitization, the reaction of electrophiles with nucleophilic groups on proteins (eg, formation of antigenic modified proteins and Michael adducts) is the chemical mechanism of action The biological mechanism of action is the expression of antigenic modified proteins by Langerhans cells followed by T cell proliferation in lymph nodes The mode of action is the sensitization state of the individual The underlying principles are as follows: (1) the compound and the target organism together obey the laws of chemistry (including those that we may not know about), (2) the laws of chemistry determine the chemical mechanism of action, (3) the chemical mechanism of action determines the biological mechanism of action, and (4) the biological mechanism of action determines the mode of action In this example, the distinction between mode and mechanism is made: skin sensitization is a single mode of action phenomenon with several mechanisms of action according to how the electrophiles bind to protein
Seven reaction mechanistic domains were reported [9,21] that categorize the mechanistic chemistry of molecular initiating events, which occur when an exogenous chemical binds covalently to biological macromolecules These seven reaction mechanistic domains are shown in Figure 28 These classifications are based on organic reaction mechanistic principles and were developed with the purpose
TABLE 2.3 (Continued ) Structural Alerts for the Rodent In Vivo Micronucleus Assay
of applying mechanism-based QSARs to broader ranges of compounds for toxicity prediction Each of the structure alerts were assigned to one of the seven mechanistic domains The mechanistic domains and structural alerts enable in silico toxicity risk assessments of a chemical to act as a mutagen or genotoxic carcinogen (genotoxicity) based on the knowledge of the structure
An important application of the structural alerts concept was the establishment of the threshold of toxicological concern (TTC) with the aim of reducing extensive toxicological evaluation The concept that there are levels of exposure that do not cause adverse effects is inherent in setting acceptable daily intakes for chemicals with known toxicological profiles It should be recognized that it is neither possible
nor necessary to thoroughly test the toxicity of all chemicals in use or present in the environments or food supply The TTC approach establishes a generic human exposure threshold value below which there would be no appreciable risk to human health The underlying principle is that TTC can be identified for many chemicals, including those of unknown toxicity, taking into consideration their chemical structures and the known toxicity of chemicals that share similar structural characteristics [18,16]
Munro and others [3] proposed thresholds of acceptable human exposure based on the Cramer decision tree [15] of structural classes I, II, and III Kroes and others [16] reported that the TTC principle can be applied for low concentrations in food of chemicals that lack toxicity data, provided that there is a sound intake estimate They also proposed a decision tree to apply the TTC principle, shown in Table 24 Proteins, heavy metals, and polyhalogenated dibenzodioxins and related compounds were excluded from this approach due to the following reasons The most potent allergens are proteins foreign to the host that are capable of inducing immunologically mediated allergic reactions It is very difficult to predict and quantify such toxic effects Many heavy metals are known to accumulate (eg, cadmium in the kidney), and accumulation needs special consideration when there are major species differences in clearance that exceed the usual uncertainty factor for species differences Steroids and polyhalogenated dibenzo-p-dioxins and dibenzofurans are considered to be nongenotoxic
TABLE 2.4 Decision Tree for Low-Molecular-Weight Compounds with Limited Toxicity Data
carcinogens, which would show a threshold in their dose-response relationships; therefore, it is irrelevant to the linearized models used to develop the TTC
The concepts discussed in this chapter include chemical carcinogenesis, electrophilic reactivity of ultimate carcinogens, reaction mechanistic domains, metabolic activation, mechanism of action, mode of action, toxicophores, structural alerts, and the TTC The goal of this chapter is to succinctly introduce the interconnected concepts to interested readers with a few well-known examples These topics have been investigated extensively over the decades by numerous researchers, and the disciplines are still evolving with significant advancements being made constantly All of these subjects are critically important in the field of genotoxicity
TABLE 2.4 (Continued ) Decision Tree for Low-Molecular-Weight Compounds with Limited Toxicity Data
