Abstract
Human health risk assessment consists of bringing to bear a large body of in vitro, animal, and epidemiologic studies on the question of whether environmental exposures to a substance are a potential risk to humans. The body of scientific information is typically less than definitive and often contains apparent contradictions. Often various possible conclusions about potential human risks may be drawn from the data and these may vary from very strong to tenuous. The task, therefore, is to communicate the uncertainties in the inferences from the data effectively, giving proper consideration to contrary data and alternative scientifically plausible interpretations. We propose an approach, Hypothesis-Based Weight of Evidence (HBWoE), to organize, evaluate, and communicate the large body of available relevant data on a given chemical, using naphthalene as an example. The goal for our use of the term “weight of evidence” (WoE) is broad in that we express the relative degrees of credence that should be placed in alternative possible interpretations of the naphthalene data and hypothesized carcinogenic modes of action (MoAs), expressed in a way that shows how such credence is tied to specific scientific interpretations, considering consistencies, inconsistencies, and contradictions within the data set.
| Abbreviations: | ||
| ATSDR, | = | Agency for Toxic Substances and Disease Registry; |
| CYP, | = | cytochrome P450; |
| DD, | = | dihydrodiol dehydrogenase; |
| DNA, | = | deoxyribonucleic acid; |
| EH, | = | epoxide hydrolase; |
| HBWoE, | = | Hypothesis-Based Weight of Evidence; |
| IARC, | = | International Agency for Research on Cancer; |
| IP, | = | intraperitoneal; |
| MNL, | = | mononuclear leucocyte; |
| MoA, | = | mode of action; GSH, glutathione; |
| MTD, | = | maximum tolerable dose; |
| NTP, | = | National Toxicology Program; |
| PAHs, | = | polycyclic aromatic hydrocarbons; |
| pHLMs, | = | pooled human liver microsomes; |
| PIP, | = | phenyl-1-pentyne; |
| ROS, | = | reactive oxygen species; |
| SCE, | = | sister-chromatid exchange; |
| US | = | EPA, United States Environmental Protection Agency; |
| WoE, | = | weight of evidence. |
Contents
Abstract 671
Abbreviations 671
Introduction and background 672
HBWoE approach 675
Proposed MoAs for naphthalene 677
(1) Postulated cytotoxic MoA for naphthalene 678
(2) US EPA’s postulated genotoxic MoA for naphthalene 678
(3) Postulated dual MoA for naphthalene 678
HBWoE evaluation for naphthalene MoA in animals 678
Overview of questions 1 and 2 678
Overview of question 3 679
Key event 1—Primary metabolism by CYP2F 680
Current data and hypothesized MoA 681
HBWoE evaluation of the current data 681
Key event 2—Cytotoxicity resulting from CYP2F metabolism of naphthalene 683
Current data and hypothesized MoA 683
HBWoE evaluation of the current data 683
Key event 3—Chronic inflammation/regenerative hyperplasia resulting from CYP2F metabolism of naphthalene 684
Current data and hypothesized MoA 684
HBWoE evaluation of the current data 685
Key event 4—Genotoxicity 685
Current data and hypothesized MoA 685
HBWoE evaluation of the current data 686
Key event 5—Tumor formation 689
Current data and hypothesized MoA 689
HBWoE evaluation of the current data 689
HBWoE conclusion regarding MoA in animals and possible explanation of observed outcomes 690
HBWoE—Plausibility of MoA extrapolation to humans 690
Epidemiology data are lacking 690
Current in vitro data for naphthalene activity in humans 690
HBWoE evaluation of the current data 691
Several questions that should be addressed before any MoA for naphthalene can be supported 692
Discussion and conclusions 692
Declaration of interest 695
References 695
Introduction and background
One of the challenges in human health risk assessment is that the body of scientific information available to support it is rarely definitive—information is typically indirect (animal studies are applied to humans, high-dose studies are extrapolated to low doses), it is often incomplete (few strains/species are tested, epidemiologic studies may lack good exposure information), and it frequently contains apparent contradictions (endpoints may be discordant among studies, genotoxicity tests may include positive and negative outcomes).
This challenge is exemplified through application of the current naphthalene carcinogenesis data set to hypothesize several carcinogenic modes of action (MoAs) for naphthalene. Inhalation of naphthalene causes olfactory epithelial nasal tumors in rats (but not in mice) and benign lung adenomas in mice (but not in rats) (CitationNTP, 1992, Citation2000), as summarized recently by CitationNorth et al. (2008). and summarize the bioassay data. There are no other animal carcinogenesis studies for naphthalene. There are no systematic epidemiologic studies of naphthalene exposure and cancer. There is an evident lack of a tumor effect in humans strong enough to have called attention to itself, despite the lack of systematic investigation, for occupationally exposed people and for people in the general population. Nasal tumors are rare in humans, although lung tumors are not. Studies of respiratory tract cancers in humans and their potential causative agents have not identified naphthalene exposure as associated with tumor risk, although this has not been explicitly examined. In short, although there is no positive human evidence for naphthalene’s carcinogenicity, the lack of evidence is not a compelling refutation.
Table 1. Incidence of neoplastic and non-neoplastic lesions in lung and nasal tissue of mice exposed to naphthalene in 2-year inhalation study (CitationNTP, 1992).
Table 2. Incidence of neoplastic and non-neoplastic lesions in nasal tissue of rats exposed to naphthalene in 2-year inhalation study (CitationNTP, 2000).
The question of naphthalene’s carcinogenicity in humans accordingly depends entirely on the experimental evidence from rats and mice, along with some MoA information that may bear on whether the rat nasal tumors or the mouse lung tumors are likely to be indicative of a potential for human risk. Several things are striking about the tumors that naphthalene inhalation induced in rats and mice in the existing bioassays. First, the tumors are of a particular histological type, and they are confined to very particular epithelial tissues of the respiratory tract that are directly exposed to naphthalene vapors: in mice, the adenomas are localized to bronchioles, and in rats there are distinct but similar tumors in olfactory and respiratory epithelial areas of the nasal cavity. This suggests a very specific and local MoA.
Second, in both rats and mice, the tissues in which tumors occur are subject (at tumorigenic exposure levels) to widespread cytotoxicity and inflammation. In the case of many other chemicals, target tissue toxicity with cell killing and regenerative hyperplasia is thought to be the immediate and primary carcinogenic process, with the forced proliferation of cells under extreme stress being the means by which tumor-inducing somatic mutations are generated and clonal expansion of mutated cells promoted; chloroform is an example of this type of chemical (CitationUS EPA, 2001). Moreover, tissues other than nasal and lung epithelia do not show such cytotoxicity and hyperplasia following naphthalene exposure, and they do not have tumors (see below regarding species specificity). That is, the tumors occur where and only where there is marked tissue toxicity, strongly suggesting a causal role.
Third, the particular tissues subject to such toxicity are also sites of concentrated and localized metabolic activity toward naphthalene. (See for the proposed pathways of metabolism of naphthalene.) In nasal and lung tissues there is cytochrome P450 (CYP)-mediated metabolic activation of naphthalene to its epoxide, a reactive compound. The epoxide can be conjugated with glutathione (GSH) (a step that appears to detoxify naphthalene and promote its excretion), but at high levels of exposure, GSH can become locally depleted in the respiratory epithelial tissues, resulting in respiratory toxicity (CitationPhimister et al., 2004). Moreover, the epoxide can be further metabolized to other potentially reactive compounds, notably 1,2-naphthoquinone, which is known to react with critical macromolecules or possibly to undergo redox cycling (CitationBuckpitt et al., 2002), potentially leading to oxidative stress (CitationBagchi et al., 1998a, Citation1998b, Citation2000, Citation2002). It appears likely that such further metabolism of the epoxide is accelerated under conditions of GSH depletion, when the capacity to conjugate and remove the epoxide is reduced.
Figure 1. Proposed scheme for naphthalene metabolism and reactive metabolites (ATSDR, 2005; CitationBogen et al., 2008).
Furthermore, inhibiting metabolism of naphthalene extinguishes the cytotoxicity. Naphthalene is metabolized in other tissues as well, notably the liver, but in these tissues there is no marked GSH depletion or evidence of tissue injury (CitationPhimister et al., 2004). Exposure to naphthalene by intraperitoneal (IP) injection results in the same pattern of metabolic activation and cytotoxicity in the same respiratory tract epithelia, showing that the localization of effects in rodents is attributable to localized high metabolic activity rather than to the direct inhalation exposure of the tissues. It is also notable that a particular CYP isozyme, CYP2F, is largely localized in just the tissues in question and has considerable metabolic activity toward naphthalene, raising the possibility that the localization of tumor response may be dependent on this particular isozyme.
The high degree of localization of naphthalene metabolic enzymes, the production of metabolites capable of causing toxicity, GSH depletion in tissues where cytotoxicity occurs, the cytotoxic reactions of the tissues themselves, and the local production of tumors—all along with the lack of tumors in tissues not having these phenomena—suggest that these processes are involved in the mode of naphthalene’s carcinogenic action in rodents, and that the balance of the activities of the enzymes responsible for these cellular processes is ultimately what determines the potential for naphthalene to cause tissue injury, and this balance likely varies across tissues and species.
There are, at least apparently, incongruous observations that clearly reflect differences across species and tissues: naphthalene metabolism is very active in mouse nasal tissue (including CYP2F activity), and indeed there is nasal tissue toxicity in mouse nose that, unlike the toxicity in rat nose, is not associated with the appearance of tumors. If tissue toxicity from local naphthalene metabolism is indeed critical and sufficient to the generation of rat nasal tumors, as has been put forward, then one must account for why the tissue toxicity in mouse nose does not have a similar carcinogenic role. Similarly, there is some metabolism and mild tissue toxicity (in the form of inflammation rather than frank cytotoxicity) in rat lung, at least in the National Toxicology Program (NTP) inhalation bioassay (CitationNTP, 2000; CitationAbdo et al., 2001), and to the degree that such processes in mice are deemed part of the mouse lung tumor response, it is necessary to understand what is different about the rat lung that makes somewhat similar processes, although much more mild in severity, not carcinogenic.
It is noteworthy that the two animal tumor responses in question—the mouse lung tumors and the rat nasal tumors—are associated with similar suites of phenomena, including localized and pronounced metabolic activation, GSH depletion (mouse lung and nasal, and GSH conjugation in rat nasal), CYP2F involvement, and manifestations of cytotoxicity. The apparently parallel dependence of mouse lung tumors and rat nasal tumors on this constellation of factors suggests a very plausible common MoA. It is nonetheless important to evaluate the evidence regarding whether the responses in rats and mice flow from essentially the same set of causative factors. To the extent that the tumor responses represent manifestation of a common underlying MoA, they support one another in providing evidence for the MoA, and for the necessity of its key events. It is important, however, not to look just at the target tissues and the positive associations of factors and tumorigenicity within those tissues, but also to examine what happens in the homologous tissues of nontarget species and in nontarget tissues where some of the same phenomena also occur. The similarities and differences in the responses, and the potential causative biological processes, provide a means to define more precisely the key steps, and to sort out what is necessary from what is merely correlated with the causative factors.
In the present paper, we have two aims: first, to introduce the Hypothesis-Based Weight of Evidence (HBWoE) approach as a general, useful tool for hazard identification and for characterizing the uncertainties in inference of human risk potential; and second, to apply the HBWoE approach to look systematically and in a comparative way at the various key elements in the hypothesized MoA for carcinogenesis by inhaled naphthalene, trying thereby to identify what are indeed the responsible and necessary elements, highlighting the uncertainties and apparent contradictions, and suggesting possible explanations for the array of observed outcomes. Our analysis is not intended to present all of the questions that should be considered in the naphthalene case, nor is it intended to prove or disprove any one hypothesis, but is an example of an evaluation and logical presentation of the naphthalene carcinogenesis data and some key questions that arise from that evaluation.
HBWoE approach
We need to grapple with less-than-definitive information on what adverse effects are possible with sufficient exposure to an agent. As noted above, relevant available information is typically less than completely dispositive: it is often incomplete and frequently contains internal inconsistencies. Extrapolations, justified by hypothesized generalizations, are often necessary in applying animal results to humans and in inferring possible risks at low environmental exposures from those observed at the high exposure levels present in the studies where the effects have been observed.
Chemical risk assessment’s reliance on animal studies (toxicologic testing)—especially studies in mammals—is based on the high degree of homology of anatomy, physiology, and biochemistry across mammalian species. This raises the prima facie case that a response to a chemical seen in animals might predict a similar response in other mammals, including humans.
But, especially with the advances in toxicological science and the ability to examine toxicological processes at the biochemical level, there are many exceptions to the expectation that different mammalian species should respond alike. There are important genetically based differences in biochemical pathways, quantitative differences in activity of key metabolic steps, differences in affinities of receptors, and other factors that can act to make it so that the qualitative toxicologic responses seen in one species may not appear in another, or they may appear only at markedly higher dose levels. There are several clear examples of species- and strain-specific responses not expected to appear in other animals or in humans. For example, some laboratory strains are prone to certain diseases—with the underlying basis for this sensitivity identified in some cases but not in others—and this leads to high background rates for those responses (compared to other strains and species) even in the absence of dosing.
On the other hand, there are cases where (at least apparently) a similar MoA occurs in different tissues in different species. So, if properly interpreted, an effect on one tissue might be evidence that a different tissue is at risk in humans. (This is most evident for genotoxic agents, where the mode is fairly generalizable across tissues, depending chiefly on the ability to be locally activated.) The possibility of this is often used to say that “site concordance is not required”—but it is really only a demonstration that different-site responses, that nonetheless represent a common kind of response, are sometimes possible. It does not constitute evidence that all responses of whatever kind in different species are all linked to a common causal factor. At some level there are indeed elements of commonality among tissues of the basic processes of control of cell division and differentiation—the somatically heritable disruption of which are the hallmarks of malignantly transformed cells—but if one too readily invokes this commonality to conclude that carcinogenic effects in one tissue in animals indicate carcinogenic risks to other tissues in humans, one is then left with accounting for the observed lack of elevated cancer responses in other tissues in the animal studies. That is, if carcinogenic influence of an agent is so general as to apply across tissues, why do we typically observe organ and histological specificity of carcinogenic effects in animal studies? If such specificity is attributed to pharmacokinetics, to localized metabolic activity, to location of molecular targets, or to tissue-specificity of biochemical processes affected by the agent, then why should such causes of site-specificity not apply in humans?
The use of animal tests that show toxicity from a chemical as a basis for identifying a hazard in humans is a matter of inference, not of direct observation. Presuming that humans will be at risk of the same (or related) outcomes amounts to the following generalizations:
By adducing an animal response as evidence of the potential for human toxicity, one is asserting there is some causal factor or process in common between the responding test species and humans that (1) accounts for the positive response in animals; and (2) constitutes a potential cause of the corresponding effect in humans. It is this (hypothesized) commonality that makes the animals a legitimate model for the possible but unobserved effect in humans. In essence, it is the assertion that there is some generally operable process that the observed rodents and humans share (presumably, by virtue of their common ancestry).
If this is so, then other species related to rodents and humans should also share the feature, and presumably its effect, unless they have deviated away from the common pattern. Therefore, looking for the feature or factor in other species constitutes a kind of test of the hypothesized generalization. If one indeed finds it in most other (mammalian) species that have been adequately examined, then the inference that humans also share this nearly universal feature (and hence also share the risk) is strengthened.
If, in contrast, the feature is not generally seen in other species, this demonstrates that its presumed presence in humans should also be questioned. In this case, the inference of potential human risk comes down to evaluating the proposition that the observed animal effect nonetheless applies to humans, owing to features that the responding animals and humans share, but that are not shared by other (nonresponding) animals. This evaluation entails trying to specify what is necessary and what is sufficient for the agent’s carcinogenic effect as observed in the animals and then determining whether humans have the requisite effects (whereas nonresponding animals or nonresponding tissues do not).
It is important to articulate the reasoning by which one proposes that a carcinogenic effect in an animal bioassay result serves as evidence that humans might be at carcinogenic risk from (sufficient) exposure to the same agent. It is only by putting the hypothesized basis for inferring potential human risk explicitly forward that one can then evaluate evidence for and against the proposition. Evidence has “weight” only insofar as it addresses the credence that should be placed in a stated hypothesis. Otherwise it is only data; what makes “data” into “evidence” is its being brought to bear on evaluation of an articulated hypothesis.
The practical task in evaluating WoE for the generalizability of a carcinogenic response involves noting the pattern of known responders and nonresponders and then weighing alternative explanations of the pattern. When faced with the question of the applicability of a particular animal response to humans, evaluating the WoE for human carcinogenicity entails (1) characterizing how firmly such an MoA can be established as the likely cause of the animal tumor responses; (2) evaluating how firmly particular biological phenomena (say, GSH depletion, inflammation, or reactive oxygen species [ROS] generation) can be concluded to be necessary parts of the carcinogenic response (i.e., without which there would be no such response); (3) evaluating how well it is established that humans do or do not have the capacity to show those same necessary phenomena after sufficient exposure; and (4) if humans are qualitatively capable of producing the necessary phenomena, evaluating how well it is established that foreseeable human exposures might be quantitatively sufficient to cause those phenomena, and hence presumably to cause human cancer risk.
Considering the similarities and differences in the responses, and in the potential causative biological processes—as they are observed in different species and different tissues within those species—provides a means to better define the key steps and to sort out what is necessary from what is merely correlated with the causative factors. These comparisons can increase confidence that the set of factors producing carcinogenesis has been identified, since they produce the effect in similar settings when the key events occur. Parallel manifestations (e.g., the mouse lung and rat nasal tumors for naphthalene, and their apparent association with local metabolism and cytotoxicity) also provide opportunities to test whether hypothesized key elements are indeed essential, since differences (e.g., between noses and lungs in rats and mice) can point out elements that may be ancillary, rather than central, to the tumorigenic process. Of course, it is also possible that the various tumor responses represent distinct, or at least partially distinct, tumorigenic processes, and this possibility needs to be evaluated. If this is so, one then needs to evaluate the evidence that each of the processes and sets of key events hypothesized to account for one of the animal tumor responses would also be expected to occur in humans.
As applied to naphthalene, the cross comparison of species is particularly important because some phenomena—at least at first examination—appear to be incongruous with the MoA outlined above. It is notable that mice do not get nasal tumors and rats do not get lung tumors as a result of naphthalene inhalation. At the very least, this casts doubt on the degree of generalizability of the animal tumor responses. Since rats do not get lung tumors, why do we suppose that humans are not like them but are rather like the mice (who do get lung tumors)? Since mice do not get nasal tumors, why do we suppose that humans are not like them but rather like the rats (who do get nasal tumors)? The fact that the responsiveness of specific tissues does not generalize between rats and mice calls into question the basis for presuming that the responses do generalize to humans. Of course it may be so that humans do have a parallel causative process either to the rat noses or to the mouse lungs, but the strength of that inference is weakened by the observation that the phenomena we are proposing to generalize from rodents to humans do not succeed in generalizing from one rodent species to another.
Evidently, something is different between rats and mice that dictates their different carcinogenic responsiveness to naphthalene inhalation in nasal and lung tissues. We are forced to refine the initial basic hypothesis that, by virtue of common mammalian anatomy, physiology, and biochemistry, whatever happens in one mammal will happen in another (including humans). The WoE evaluation then has to focus on identifying what the differences between rats and mice are that dictate their different outcomes, and concomitantly, it has to hypothesize what is deemed to be causative for the responses seen, what among those causes is missing for the species in which the response is not seen, and (if the animal endpoint is proposed as evidence of potential human risk) what the state of affairs in humans may be that would make the corresponding tissue—or some other tissue or anatomical site—susceptible to a similar process as the one proposed as causing the tumors in the responding animals.
We emphasize that the point is not definitively to prove or disprove such refined hypotheses, for such will rarely be possible. Instead, the aim is to evaluate the credibility of the hypothesized basis for an inference of potential human risk, in view of (1) the plausibility of the hypothesis (in view of what we know about biology and the invoked causal processes and their variation among species); (2) the degree to which we need to fill data gaps or complete inferences with assumptions that have not been empirically verified; (3) the plausibility of those specific assumptions and inferences, that is, an assessment of the a priori likelihood that they would prove true if tested; (4) the degree to which specific hypothesized causative phenomena have indeed been observed rather than merely assumed or proposed; (5) the degree to which the causative elements have been tested and affirmed in other species or in other tissues; (6) the degree to which the observed lack of the same causative elements in other settings is associated with a lack of observed effect (i.e., the complementary affirmation of its role); and (7) the degree to which the hypothesized elements have not merely been crafted to accommodate or explain apparent incongruities in the data that would seem to be inconsistent with a more straightforward version of the hypothesis. By making explicit the logic and content of the hypothesized basis for using an animal response as an indicator of potential human hazard, and by evaluating that logic using the numbered points above against all the available data, one can come to a transparently explained evaluation of how credible the proposed existence of the human hazard potential ought to be judged.
The last point about accommodating incongruities bears some comment. It is clearly important that the hypothesized generalization of the toxicity phenomena be in accord with the data and observations we have, but it is important to beware of what epistemologists call accommodation: adding fillips or provisos or other small contingencies onto the basic hypothesis to bring it into line with what would otherwise be contradictory findings. It is often possible to “save” a hypothesis from refutation by small modifications, but one needs to be aware that in doing so, the fact that the modified hypothesis fits all the data is not really much of an affirmation, at least until the newly added provisos are used as predictions about further outcomes that can then be tested. In our discussion below we have referred to such modification or extension of what is hypothesized as the making of “ad hoc assumptions.” What makes an assumption ad hoc (as opposed to just an ordinary assumption, a reasonable assertion about something that has not been directly observed) is that ad hoc assumptions are made with their consequences in mind and chosen so that those consequences match what has already been observed. There is nothing inherently wrong with ad hoc assumptions—indeed they are necessary to updating our understanding as we bring to bear complex results on initially simple hypotheses—but in doing so we need to recognize that they are part of hypothesis formulation and not of hypothesis testing.
One must also be careful not to ascribe a causative role to some element simply because it is highly correlated with the carcinogenic response. If one examined enough features, it is not surprising that some would be shared by different species or tissues, but much less so by other tissues. Ascribing the tumors in target tissues to such a feature merely because of the co-occurrence would be a logical error; indeed, it would be an ad hoc assumption justified by the fact that, if it were indeed causal, then it would explain the pattern of tumor occurrence and nonoccurrence. As an example, one might have used the observation of tumors in tissues with direct air contact in the animal inhalation bioassays of naphthalene to hypothesize that such direct contact is necessary. If this were so, it would help explain tissue specificity of the tumorigenic effects that were seen, and it is certainly biologically plausible and analogous to what has been seen in certain other chemicals. But the observation of the same cytotoxicity localized in the same respiratory tract epithelium after IP exposure to naphthalene, along with the extinguishing of cytotoxicity by inhibition of naphthalene metabolism, strongly suggests that local metabolic activation, rather than direct air contact, is responsible for the site specificity. That is, the proposed air-contact requirement can be seen to have been somewhat ad hoc (although very reasonable and plausible), and it is only by recognizing it as a hypothesis—and by testing that hypothesis to see how it stands up to other conditions—that one comes to fuller understanding. That understanding is important because naphthalene-exposed humans presumably share the air-contact aspect with the rats and mice, but do not share the patterns of metabolic activity. Co-occurrence patterns combined with some biological plausibility is a good way to create a hypothesis about necessary elements in an MoA, but in the end, its credibility depends on the degree to which the essentiality of the phenomena for carcinogenic response can be tested and affirmed.
Proposed MoAs for naphthalene
For our HBWoE analysis we have considered three potential MoAs for naphthalene—the United States Environmental Protection Agency’s (US EPA’s) proposed genotoxic MoA (CitationUS EPA, 2004), and two other potential MoAs that we believe may be more plausible in animals and humans, based on the available scientific data. The three MoAs are as follows.
(1) Postulated cytotoxic MoA for naphthalene
One MoA put forth involves CYP2F-mediated metabolism of naphthalene to cytotoxic, but nongenotoxic, metabolites in the target tissues for carcinogenesis. These naphthalene metabolites lead to cytotoxicity in mouse lung and rat nasal tissue, characterized by chronic inflammation and regenerative hyperplasia, which in turn leads to tumor formation in those tissues. This MoA is supported by observations in mouse lung and rat nasal tissue whereby CYP2F levels are elevated and naphthalene is shown to be metabolized to its epoxide. For this proposed MoA, any genotoxic effects observed upon naphthalene treatment are considered secondary (typically cytogenetic) and are the result of naphthalene-induced cytotoxicity.
(2) US EPA’s postulated genotoxic MoA for naphthalene
Based on the same set of data used to derive the cytotoxic MoA, US EPA proposed a genotoxic MoA for naphthalene. US EPA noted that a number of naphthalene metabolites (naphthalene-1,2-oxide, 1,2-naphthoquinone, and 1,4-naphthoquinone) are potentially involved in the MoA by which naphthalene affects mouse lung epithelial tissue and mouse and rat nasal tissue, and that these metabolites may damage tissue macromolecules either by directly reacting with deoxyribonucleic acid (DNA) or proteins or by the generation of ROS. It also acknowledged that the current data potentially support a nongenotoxic MoA for naphthalene, but that an understanding of this MoA is inadequate for determining why rats, but not mice, develop nasal tumors, although naphthalene causes cytotoxicity in nasal tissue of both species. Based on this uncertainty, and the idea that other events preceding tumor formation potentially have not yet been identified, it concluded that for naphthalene, “a possible genotoxic mode of carcinogenic action cannot be discounted as there are some data indicating genotoxicity of 1,2-naphthoquinone.”
Under this view, a genotoxic MoA could be adopted for animals and potentially extrapolated as such to humans.
(3) Postulated dual MoA for naphthalene
Another MoA put forth was discussed by CitationBogen (2008). Bogen proposed a “dual” MoA including both cytotoxicity and genotoxicity that happen concurrently, suggesting that a relationship exists whereby naphthalene-induced increased cytotoxicity and DNA damage can occur as independent events, but that the DNA damage occurs at the same time as the cytotoxicity and is not an initiating event. The proposed MoA is based on data from a study by CitationWilson et al. (1996) in which cytotoxicity and genotoxicity (sister-chromatid exchange [SCE]) were assayed in human mononuclear leukocytes (MNLs) exposed to naphthalene and various naphthalene metabolites in vitro. It is also based on the known potential for genotoxicity of the downstream naphthalene metabolites (i.e., 1,2-naphthoquinone and 1,4-naphthoquinone) that has been shown in vitro, and recently in vivo (in mouse skin) (CitationSaeed et al., 2009), in combination with histopathology findings in the rat and mouse bioassays that clearly link region-specific cytotoxicity with observed sites of tumorigenesis. The proposal was put forth, in part, to illustrate that a purely genotoxic MoA will overestimate risk when there is a clear cytotoxic component to the MoA.
HBWoE evaluation for naphthalene MoA in animals
In the HBWoE analysis, our aim is to look systematically and comparatively at the various key elements in the metabolism and the hypothesized MoAs for inhaled naphthalene, to try to identify the responsible and necessary elements for carcinogenesis. For each key event, and for each species and tissue, we apply three key lines of questioning:
What is necessary in the proposed naphthalene MoA? What is sufficient and are other elements also necessary?
For those events or processes proposed as critical to the observed carcinogenic effects of naphthalene, what other observable manifestations should they have (in other tissues or species)? Are these other manifestations indeed found?
If either the operation or the necessity of these proposed critical events were disproven, how else would one account for the array of outcomes?
summarizes an overview of the tracing of the logic that we applied in evaluating the WoE for the hypothesized naphthalene MoA in animals. and list each key event put forth as part of the various proposed MoAs for naphthalene in animals, and summarize the outcome of our HBWoE evaluation of that data.
Table 3. Weight of evidence for key events in hypothesized naphthalene carcinogenic MoA in animals.
Table 4. Weight of evidence for genotoxic event in MoA for naphthalene (1,2-naphthoquinone) carcinogenesis.
Figure 2. Overall evaluation of the weight of evidence for the hypothesized naphthalene modes of action in animals.
Overview of questions 1 and 2
The following proposed key events—metabolism of naphthalene to the epoxide and other metabolites by CYP2F (and/or other CYPs), cytotoxicity (including GSH depletion in cells where cytotoxicity occurs), chronic inflammation, and regenerative hyperplasia—are plausibly necessary but not evidently sufficient for the animal MoA. If these are the only critical events in the MoA, we would expect to observe nasal tumors in mice, but this was not observed in the mouse bioassay. Further explanations or assumptions are therefore necessary to help account for the lack of tumors in the mouse nose. This requires that we look harder at the current data, perhaps to modify the proposed MoAs to reflect all available data.
With regard to genotoxicity as a potential key event, no evidence suggests that genotoxicity is necessary (i.e., critical early event) for naphthalene-induced tissue injury and carcinogenesis. If genotoxicity were a critical early event (i.e., before cytotoxicity), one would expect tumors to be observed more systemically, i.e., in other tissues where naphthalene is present and metabolized by CYP2F and where detoxification mechanisms are potentially depleted (i.e., the mouse nose), but this was not observed in the animal bioassays. In vitro and in vivo evidence does suggest that 1,2-naphthoquinone (1,2-naphthoquinone) is genotoxic (ATSDR, 2005; CitationSaeed et al., 2009). Is there a role for genotoxicity of 1,2-naphthoquinone in the animal MoA? If so, does it occur subsequent to or concurrent with cytotoxicity?
Overview of question 3
1,2-naphthoquinone is a downstream metabolite, possibly resulting from GSH depletion and epoxide hydrolase (EH) and dihydrodiol dehydrogenase (DD) activities (see ). Generation of substantial amounts of 1,2-naphthoquinone potentially requires exposure to large amounts of naphthalene—sufficient to saturate higher-capacity competing detoxification pathways and cause GSH depletion, changes that would also be cytotoxic. Therefore, the potential for 1,2-naphthoquinone to damage DNA likely occurs concurrent with or subsequent to the cytotoxic effects. This is possibly why naphthalene was not mutagenic in the Ames assays (i.e., the level of cytotoxicity was sufficiently high by the time the 1,2-naphthoquinone was formed, such that cell death occurred before the tester strains were able to fix mutations at an observable rate), but direct exposure to the 1,2-naphthoquinone was mutagenic in Salmonella (CitationFlowers-Geary et al., 1996). Understanding the balance of activities of the enzymes—namely CYP2F (or other CYPs), GSH transferase, EH, DD, and DNA repair enzymes—involved in generation of the 1,2-naphthoquinone (or other naphthalene metabolites) and repair of any DNA damage generated by these metabolites, is critical in evaluating an MoA for naphthalene, particularly because this balance likely varies across species and tissues. We propose that this balance is disrupted in the mouse lung and nose, and rat nose, but due to differences in the relative amounts and efficiencies of the enzymes involved in each tissue, the ultimate disruption is different in each tissue, such that it leads to different effects.
Below we discuss each key event by first summarizing the current data for that key event, and then how the data bear on the WoE for that key event (also summarized in and ). For each key event, we present some examples of the questions that become apparent once the hypothetical elements of the MoA are evaluated against the available data. We further assess how the ability to address those questions, the assumptions that still need to be made, the apparent inconsistencies, etc., affect the overall WoE for the animal MoA. It is important to point out that, in the context of the HBWoE, we track the presence or absence of key events in species and tissues where tumors did not occur. This is key in the HBWoE approach in that all of the data, positive and negative, need to be considered in order to propose a plausible MoA that is consistent with all available data. For each species/tissue combination, for each proposed key event (shown in and ), we have selected a category that represents the WoE for the involvement of that key event in the naphthalene carcinogenesis MoA in animals. The WoE is qualitatively categorized, as defined in and , as “strongly supporting,” “strongly refuting,” “some evidence,” “no evidence,” “ad hoc assumptions required,” or “data gap.”
Key event 1—Primary metabolism by CYP2F
Although the animal data suggest CYP2F is involved in the epoxidation of naphthalene, the available data do not rule out the potential involvement of other CYPs in the metabolism of naphthalene. Collocation of CYP2F activity and the toxic effects observed suggest a causal role, but one must consider the possibility that the collocation is mere happenstance, with after-the-fact significance attributed to it mostly because it would constitute a tidy explanation of site specificity.
Current data and hypothesized MoA
As summarized in recent reviews on the metabolism of naphthalene (CitationBuckpitt et al., 2002; CitationBogen et al., 2008), available data indicate the toxicity of naphthalene is attributable to its metabolic activation by one or more forms of CYP. Metabolism of naphthalene to the naphthalene 1,2-epoxide is believed to be the principal first step in the cytotoxicity of naphthalene and subsequent cellular proliferation and carcinogenicity.
CYP2F2 enzyme isolated from mouse liver has been shown to catalyze epoxidation of naphthalene (CitationNagata et al., 1990). Immunolocalization studies and naphthalene metabolism assays in microdissected airways in mice indicate that CYP2F2 in distal airways (the most susceptible site) correlates with the highest rate of naphthalene metabolism (measured by presence of the dihydrodiol and GSH conjugates) (CitationBuckpitt et al., 1995). This study also showed that very low CYP2F levels in Clara cells of rat and hamster were correlated with low rates of naphthalene metabolism. Immunoblot analysis also showed that mice have 30- to 40-fold higher levels of CYP2F than rats in terminal bronchioles and trachea (CitationBaldwin et al., 2004). A straight comparison of the rat and mouse lung data suggests potential involvement of CYP2F in the primary metabolism of naphthalene because the low levels of CYP2F in rat lung are consistent with the lack of lung tumors in rats. Other data have been put forth as supportive of the involvement of CYP2F in naphthalene carcinogenicity. For example, wild-type mice treated (IP injection) with naphthalene and a CYP2F inhibitor (5-phenyl-1-pentyne [5-PIP]) showed no signs of lung or olfactory toxicity, as compared to the mice treated with naphthalene and no inhibitor where toxicity was observed (CitationGenter et al., 2006). Furthermore, immunolocalization studies indicated that the rat nasal mucosa (primarily the olfactory epithelium) contains high levels of CYP2F4 (CitationBaldwin et al., 2004). Naphthalene has been shown to be metabolized to its epoxide in rat nasal mucosa preparations (CitationBuckpitt et al., 1992). In addition, CitationLee et al. (2005) showed that upon naphthalene inhalation in rats, the location of severe nasal injury (olfactory region) correlated with the highest rate of naphthalene metabolism to the epoxide. A recent study by CitationMorris and Buckpitt (2009), which examined in vitro metabolism and uptake of naphthalene in the upper respiratory tract of rats, found that naphthalene metabolism in the rat nose was inhibited after pretreatment with the CYP inhibitor 5-PIP, suggesting that CYP2F is likely responsible, at least in part, for naphthalene metabolism in the rat nose. The study also found that naphthalene uptake was significantly reduced by 5-PIP treatment, suggesting that nasal metabolism of naphthalene is directly related to naphthalene uptake.
These data have been used to hypothesize that CYP2F metabolism of naphthalene to its epoxide is the first, and obligatory, key event in the animal MoA for naphthalene carcinogenesis.
HBWoE evaluation of the current data
The data discussed above suggest that CYP2F metabolism of naphthalene to its epoxide is the first key event in the animal MoA for naphthalene carcinogenesis (). However, the available data do not rule out the potential involvement of other CYPs in the metabolism of naphthalene. In addition, it is not clear why mouse nasal tissue express CYP2F at levels comparable to, or slightly higher than, rat nasal tissue, yet mouse nasal tumors were not observed in the mouse bioassay. When one looks carefully at the naphthalene metabolism studies carried out in mice and rats, and at the CYP2F data for other tissues and species, three questions become evident:
If CYP2F is the primary metabolizing enzyme for naphthalene that is ultimately responsible for tumor formation, and CYP2F is expressed in mouse nasal tissue at levels higher than in rat nasal tissue, why does naphthalene induce olfactory neuroblastomas in rats but not in mice?
CYP2F is expressed in mouse and rat liver and is active toward naphthalene; so why does naphthalene not induce liver tumors in mice and rats?
Since the CYP2F inhibitor (5-PIP) also inhibits CYP2E1, could CYP2E1 be involved in the metabolism of naphthalene to its toxic metabolite(s)?
1. If CYP2F is the primary metabolizing enzyme for naphthalene that is ultimately responsible for tumor formation, and CYP2F is expressed in mouse nasal tissue at levels higher than in rat nasal tissue, why does naphthalene induce olfactory neuroblastomas in rats but not in mice?
CitationBaldwin et al. (2004) showed that mouse nasal tissue expresses CYP2F2 at levels lower than does mouse lung, but at slightly higher levels than in rat nasal tissue. Based on this observation, and the proposed CYP2F metabolism of naphthalene to its epoxide as a key event, one would expect to find nasal tumors in mice as a response to naphthalene exposure, especially in light of the observation of subsequent key events in mouse nasal tissue (cytotoxicity, chronic inflammation, and regenerative hyperplasia), discussed later in this paper. However, naphthalene did not induce nasal tumors in mice. Ad hoc assumptions one might make with regard to this question is that more naphthalene might reach the nasal epithelium in the rat than in mice, and/or that other CYPs or other metabolic enzymes (that are not as active in mouse nasal tissue) are important in the activation of naphthalene to toxic metabolites in rat nasal tissue. Although these assumptions are not unreasonable, they are put forth to explain why the observed data do not fit with the current hypothesized MoA, and require further investigation with the idea that once we understand the basis for why tumors are not observed in the mouse nose that this may result in a modification of the proposed MoA. Therefore, this is an important question to consider in interpretation of the metabolic data for naphthalene in proposing an animal MoA and extrapolation to humans.
As shown in , to suggest the mouse nasal data fit with the proposed MoA requires that ad hoc assumptions be made; thus, the categorization “ad hoc assumptions required.”
2. CYP2F is expressed in mouse and rat liver and is active toward naphthalene; so why does naphthalene not induce liver tumors in mice and rats?
CYP2F is active in mouse and rat liver, and has been shown to metabolize naphthalene to its epoxide in mouse liver; however, tumors are not observed in mouse liver, nor are any of the other hypothesized key events in the proposed naphthalene MoA in mouse lung (chronic inflammation and regenerative hyperplasia). As discussed, the relative amounts and efficiencies of the enzymes involved in naphthalene metabolism, detoxification, and possibly DNA repair in each tissue ultimately determines the extent to which naphthalene can induce injury in each tissues. CitationPhimister et al. (2004) showed that GSH depletion precedes and is critical to the observed naphthalene-induced cytotoxicity in mouse lung, and that GSH is not depleted in the liver at naphthalene inhalation exposure concentrations that deplete GSH in the lung. Therefore, it is likely that at exposure concentrations used in the mouse bioassay, naphthalene is sufficiently detoxified in the liver so that GSH is not depleted and tissue injury and subsequent tumor formation do not occur. However, we still need to make the assumption that naphthalene is being sufficiently detoxified in the liver of other species in order for the observations in the liver to fit the proposed MoA. Again, this is not an unreasonable assumption, but one needs to realize that it is being made and that further testing may be necessary. Therefore, as shown in , for the liver, the mouse data for this key event are categorized as “some evidence” and for the rat as “ad hoc assumptions required.”
3. Since the CYP2F inhibitor (5-PIP) also inhibits CYP2E1, could CYP2E1 be involved in the metabolism of naphthalene to its toxic metabolite(s)?
Wild-type mice treated (IP injection) with naphthalene and a CYP2F inhibitor (5-PIP) showed no signs of lung or olfactory toxicity, as compared to the mice treated with naphthalene and no inhibitor where toxicity was observed in lung (hemorrhagic lungs) and olfactory tissue (100% sloughing of olfactory mucosa) (CitationGenter et al., 2006). This study has been frequently cited in the literature as support that CYP2F is the CYP involved in metabolic activation of naphthalene to its toxic metabolites in mouse lung. Although these studies are supportive, especially in combination with the immunolocalization studies (CitationBuckpitt et al., 1995) correlating CYP2F with the naphthalene epoxide at the site of injury in mouse lung, it is important to consider that the CYP2F inhibitor (5-PIP) also inhibits CYP2E1, which is also present in mouse lung and rat nose. Therefore, one needs to consider the possibility that CYP2E1 could be involved in naphthalene metabolism. However, the relative activities of these enzymes toward naphthalene needs consideration. Recombinant CYP2F from mice and rats efficiently generates the naphthalene oxide: CYP2F4 from rats (KM = 3 µM, Vmax = 107 min−1) (CitationBaldwin et al., 2005); and CYP2F2 from mice (KM = 3 µM, Vmax = 104 min−1) (CitationShultz et al., 1999). We did not find any studies that evaluated the activity of rodent CYP2E1 toward naphthalene. However, recombinant CYP2E1 from humans metabolized naphthalene to its 1-naphthol metabolite at a much lower efficiency (KM = 10.1 µM, Vmax = 8.4 min−1) (CitationCho et al., 2006). These data suggest that CYP2F metabolizes naphthalene more efficiently than CYP2E1, but studies comparing both CYPs in humans and rodents would provide stronger evidence as to the relevance of these two CYPs in various species. Studies in CYP2E1 and CYP2F knockout mice, or with inhibitors specific to CYP2E1, would provide useful information in this regard. Accordingly, as shown in , the mouse lung data provide “some evidence” that CYP2F is the primary CYP involved in metabolism of naphthalene to the epoxide, but more studies are necessary before we can rule out the involvement of CYP2E1 in metabolic activation of naphthalene to its toxic metabolites in mouse lung. CYP2E1 is present in mouse nasal and lung tissues (CitationSimmonds et al., 2004; CitationGreen et al., 2001). One possibility might be that CYP2E1 and CY2F2 are both involved in metabolic activation of naphthalene and cytotoxicity. There is some precedent for considering this possibility; a recent study of 1,1-DCE bioactivation and cytotoxicity in murine Clara cells (CitationSimmonds et al., 2004) suggested that both CYP2F2 and CYP2E1 are involved in 1,1-dichloroethylene (1,1-DCE) bioactivation and lung toxicity.
Immunolocalization studies indicated that the rat nasal mucosa (primarily the olfactory epithelium) contains high levels of CYP2F4 (CitationBaldwin et al., 2004). In addition, a recent study by CitationMorris and Buckpitt (2009) found that naphthalene metabolism in the rat nose was inhibited after pretreatment with 5-PIP, suggesting that CYP2F is likely responsible, at least in part, for naphthalene metabolism in the rat nose. Therefore, as shown in , the rat nasal data provide “some evidence” that CYP2F4 is responsible for naphthalene metabolism to its epoxide, but there are no definitive data that CYP2F4 is responsible for naphthalene metabolism in the rat nose that ultimately leads to tumor formation. One might ask: Is CYP2E1 present in rat nasal tissue, and has it been shown to activate naphthalene to the epoxide in any rat tissue? CYP2E1 is present in rat nasal tissue (CitationThornton-Manning and Dahl, 1997), but as discussed above, to our knowledge its ability to activate naphthalene to the epoxide in any rat tissue is unknown. Therefore, if CYP2E1 knockout and/or inhibitor studies suggest it may be involved in naphthalene metabolism, the potential involvement of CYP2E1 in metabolism of naphthalene in the rat nose would require further investigation.
As discussed, CYP2F expression is very low in rat Clara cells compared to mice, and is correlated with low rates of naphthalene metabolism. These studies support the hypothesis that CYP2F is the primary metabolizing enzyme for naphthalene because the low level of this CYP in rat lung is consistent with the lack of naphthalene-induced tumor formation in rat lung. However, the fact that the involvement of CYP2F and possibly CYP2E1 in naphthalene-induced mouse lung tumor and rat nasal tumor formation needs further investigation, requires that the rat lung data be interpreted as “some evidence” of a CYP2F4 involvement. As discussed below, the data are not entirely clear as to whether naphthalene causes mild cytotoxicity in rat lung, at least according to the NTP bioassay (CitationNTP, 2000; CitationAbdo et al., 2001). If future studies suggest that inhalation of naphthalene does induce rat lung cytotoxicity, then the potential involvement of CYP2E1 might be investigated.
To clarify, and in accordance with our methodology, we are not implying that CYP2E1 must be involved in naphthalene metabolism, but only that its role has not been definitively ruled out. And in this case, particularly given the inconsistent observations across species and tissues, settling on CYP2F and ruling out potential roles for other CYPs (not just CYP2E1) is not sufficiently supported with the current available data.
Key event 2—Cytotoxicity resulting from CYP2F metabolism of naphthalene
It is well known that naphthalene cytotoxicity requires metabolic activation and that unmetabolized naphthalene does not cause cytotoxicity (CitationBuckpitt et al., 2002; CitationATSDR, 2005; CitationBogen et al., 2008). However, as introduced in the previous section, it is not clear whether the cytotoxicity is due to naphthalene metabolites generated solely by CYP2F activity. In addition, although the animal data suggest cytotoxicity is a key event in naphthalene carcinogenesis, tumors are not observed in all tissues in which naphthalene-induced cytotoxicity is observed.
Current data and hypothesized MoA
Rats and mice have similar levels of CYP2F expression in nasal tissue (CitationBaldwin et al., 2004), and naphthalene induces a cytotoxic response in the mouse (CitationNTP, 1992; CitationPlopper, 1992a,b) and rat (CitationNTP, 2000) olfactory epithelium at similar naphthalene exposure levels. However, olfactory neuroblastomas were observed in rats but not in mice.
Sites of naphthalene-induced Clara cell cytotoxicity in mouse lung are similar to sites of lung tumor formation (CitationPlopper et al., 1992a, Citation1992b) observed in the naphthalene inhalation mouse bioassay (CitationNTP, 1992), in a dose-dependent manner. In addition, immunolocalization studies in mice indicated that CYP2F2 levels in Clara cells are elevated at sites of naphthalene-induced injury in mouse lung (CitationBuckpitt et al., 1995). These data suggest that CYP2F2 activation of naphthalene may be responsible, at least in part, for the cytotoxic effects of naphthalene in mouse lung. CitationPhimister et al. (2004) showed that naphthalene inhalation causes GSH depletion in the mouse respiratory tract correlated with an increase in naphthalene protein adduct formation in mouse airway, and with an increase in cellular injury and toxicity at the same site, with GSH depletion preceding injury.
Naphthalene inhalation caused increased alveolar epithelial hyperplasia and mild chronic inflammation (also seen in the chamber control) in the rat lung (CitationNTP, 2000; CitationAbdo et al., 2001). The authors noted that since mild chronic inflammation is often seen in chamber control rats, and that there was no dose-response observed, it is not clear whether the changes were exposure related. However, they did not completely rule out the possibility of an exposure-related association. Earlier studies by CitationPlopper et al. (1992a, Citation1992b) showed, via intraperitoneal injection, that naphthalene does not cause rat Clara Cell toxicity. We only found one other study that examined lung cytotoxicity in rats via the inhalation route (CitationWest et al., 2001); this study did not observe lung toxicity at concentrations as high as 100 ppm, but the authors only looked at toxicity from a 4-hour exposure.
HBWoE evaluation of the current data
A closer look at the studies described above raises the following questions:
Could naphthalene-induced GSH depletion and cytotoxicity in mouse lung be due to CYP2E1 metabolism of naphthalene (or another CYP), or a combination of CYP2F2 and another CYP? Similarly, is naphthalene-induced cytotoxicity in the rat and mouse nose the result of CYP2F2 metabolism or metabolism by another CYP?
If there is mild naphthalene-induced injury in rat alveolar epithelial cells (and so far the data do not strongly suggest this), is it due to CYP2F or might it be due to another CYP?
If CYP2F2 activity ultimately is responsible for the cytotoxic response in the mouse nose, why does this response (in addition to subsequent key events discussed later) not lead to tumor formation in the mouse nose? Are other events occurring in the rat nose that are not occurring in the mouse nose, which could explain why naphthalene causes tumors in rat but not mouse nasal tissue?
1. Could naphthalene-induced GSH depletion and cytotoxicity in mouse lung be due to CYP2E1 metabolism of naphthalene (or another CYP), or a combination of CYP2F2 and another CYP? Similarly, is naphthalene-induced cytotoxicity in the rat and mouse nose the result of CYP2F2 metabolism or metabolism by another CYP?
Although CitationPhimister et al. (2004) showed that naphthalene inhalation causes GSH depletion in the mouse respiratory tract, which is correlated with an increase in naphthalene protein adduct formation in mouse airway, and an increase in cellular injury and toxicity at the same site, this study did not specifically examine whether CYP2F naphthalene bioactivation was correlated with GSH depletion. This study, in combination with the naphthalene metabolism studies discussed earlier, is fairly supportive that CYP2F activity is likely involved, at least in part, in the formation of naphthalene metabolites responsible for naphthalene-induced GSH depletion and mouse lung cytotoxicity. However, as discussed above, the potential for involvement of CYP2E1 cannot be ruled out, and a recent study (CitationSimmonds et al., 2004) showed that both CYP2F2 and CYP2E1 are involved in 1,1-DCE bioactivation and lung toxicity in mice. In addition, naphthalene (but not 1-naphthol) cytotoxicity in human MNLs was increased in the presence of CYP2E1-induced rat liver microsomes (CitationWilson et al., 1996), suggesting that CYP2E1 may be involved in formation of the epoxide but that another CYP may be involved in further metabolism of the 1-naphthol to the quinones. Therefore, as shown in , the mouse lung data provide “some evidence” that CYP2F is involved in cytotoxicity in the mouse lung, since the data currently cannot rule out the involvement of other CYPs.
In addition, the rat nose data provide “some evidence” that CYP2F-induced cytotoxicity is a key event because there is CYP2F4 expression in the rat nose, and increased CYP levels are correlated with formation of the naphthalene epoxide and cytotoxicity in the rat nose upon naphthalene exposure; however, as with the mouse lung, there is a question as to the potential involvement of other CYPs (such as CYP2E1) in the primary metabolism of naphthalene that leads to cytotoxicity. Similarly, there is a question as to the potential involvement of CYP2E1 in naphthalene-induced cytotoxicity in the mouse nose. However, interpretation of the mouse nasal tissue data, which have shown that these tissues have CYP2F levels similar to rats (actually slightly higher in mice) (CitationBaldwin et al., 2004), is more complex (as discussed below), given that tumors have not been observed in the mouse nose.
2. If there is mild naphthalene-induced injury in rat alveolar epithelial cells (and so far the data do not strongly suggest this), is it due to CYP2F or might it be due to another CYP?
Although it is very likely that the NTP rat lung bioassay results (CitationNTP, 2000; CitationAbdo et al., 2001) were not exposure related, based on the specific observations and conclusions in the study (discussed above), and considering the results by CitationPlopper et al. (1992a, Citation1992b) and CitationWest et al. (2001) that found no toxicity in rat lung, it would be useful to further examine rat lung cytotoxicity from long-term exposure to naphthalene, via the inhalation route, to more definitively rule out any potential relationship to naphthalene exposure. If mild rat lung toxicity is found to be associated with naphthalene exposure, there is the question as to whether CYP2F or other CYPs could potentially be involved in low-level primary metabolism of naphthalene in rat lung. As shown in , the current rat lung data provide “some evidence” of CYP2F being involved in naphthalene metabolism that ultimately leads to cytotoxicity (i.e., low levels of CYP2F in rat lung and possibly low levels of cytotoxicity).
3. If CYP2F2 activity ultimately is responsible for the cytotoxic response in the mouse nose, why does this response (in addition to subsequent key events in the mouse nose, discussed later) not lead to tumor formation in the mouse nose? Are other events occurring in the rat nose that are not occurring in the mouse nose, which could explain why naphthalene causes tumors in rat but not mouse nasal tissue?
Based on observations in mouse nasal tissue, as shown in , the mouse nasal data are categorized as “ad hoc assumptions required,” i.e., although the expression of CYP2F is correlated with naphthalene-induced cytotoxicity in the mouse nose, the lack of tumor formation in mouse versus rat nose suggests there is some naphthalene-induced event in the rat nose that is not occurring in the mouse nose, and we need to make ad hoc assumptions about what these events might be in order to fit the available data. Since the levels of cytotoxicity and CYP2F are similar in both rat and mouse nose, this suggests that CYP2F expression may be tightly correlated with cytotoxicity, but that further metabolism of the epoxide may be necessary for tumor formation in the rat nose. This is further discussed below.
Key event 3—Chronic inflammation/regenerative hyperplasia resulting from CYP2F metabolism of naphthalene
Current data and hypothesized MoA
As summarized in the mouse NTP bioassays (CitationNTP, 1992; CitationAbdo et al., 1992) and in the review by CitationNorth et al. (2008), naphthalene-induced lung adenomas/carcinomas in mice were accompanied by pulmonary chronic inflammation in a dose-dependent manner. In addition, studies of naphthalene-induced injury in mouse lung via IP injection have reported regenerative repair (hyperplasia) (CitationVan Winkle et al., 1995, Citation1997, Citation2001, Citation2004). Furthermore, as summarized in the mouse (CitationNTP, 1992; CitationAbdo et al., 1992) and rat (CitationNTP, 2000; CitationAbdo et al., 2001) naphthalene bioassays, naphthalene induces similar non-neoplastic lesions in rat and mouse nasal tissue. In both species, chronic inflammation and respiratory epithelial hyperplasia were observed in the nasal tissue. Chronic inflammation and hyperplasia were nearly 100% at all doses in rats and mice. Atypical hyperplasia was observed in the rat olfactory epithelial tissue where neuroblastomas were also observed.
Based on these data, chronic inflammation and regenerative hyperplasia have been hypothesized to be likely key events in the carcinogenic MoA for naphthalene in animals.
HBWoE evaluation of the current data
A closer look at the data discussed above raises the following questions:
Does chronic inflammation and regenerative hyperplasia ultimately result from CYP2F metabolism of naphthalene?
Is there further metabolism of the naphthalene epoxide in the rat olfactory epithelium that does not happen in the mouse nose, which allows for further clonal expansion of precancerous hyperplastic cell populations, resulting in atypical hyperplasia and tumor formation?
1. Does chronic inflammation and regenerative hyperplasia ultimately result from CYP2F metabolism of naphthalene?
As shown in , the mouse lung and rat nasal data provide “some evidence” that chronic inflammation and regenerative hyperplasia are key events in the naphthalene MoA in animals, but for the same reasons discussed above for the other hypothesized key events, it is not clear if this is due only to CYP2F metabolism of naphthalene, or whether CYP2E1 (or another CYP) could be involved. As discussed in the previous sections, rat lung has low levels of CYP2F4, and naphthalene inhalation caused a mild increase in alveolar epithelial hyperplasia and mild chronic inflammation (also seen in the chamber control) in the rat lung (CitationNTP, 2000; CitationAbdo et al., 2001). However, it is not clear whether this observation was exposure related, and other studies have shown rat lung to be refractory to naphthalene-induced cytotoxicity (CitationPlopper et al., 1992a, Citation1992b; CitationWest et al., 2001). Therefore, the rat lung data provide “some evidence” () of CYP2F-induced chronic inflammation and regenerative hyperplasia as key events in the naphthalene MoA, because low CYP2F4 levels in rat lung lead to very low (if any) levels of these effects.
2. Is there further metabolism of the naphthalene epoxide in the rat nose that does not happen in the mouse nose, which allows for further clonal expansion of precancerous hyperplastic cell populations, resulting in atypical hyperplasia?
“Atypical” hyperplasia and olfactory neuroblastomas were observed in rat but not mouse olfactory epithelium (nearly 100% at all doses in rat). As shown in , the mouse nasal data are categorized as “ad hoc assumptions required”; that is, if chronic inflammation and regenerative hyperplasia are key events in the naphthalene MoA, and these key events occur in mouse nasal tissue but do not lead to tumor formation, ad hoc assumptions are necessary to explain this observation. Hyperplasia is defined as an increase in the number of normal cells in a normal arrangement within a tissue; it exhibits a normal growth pattern and maturation sequence, and is not generally considered “preneoplastic” (CitationEustis, 1989). Atypical hyperplasia (dysplasia), on the other hand, is defined as a proliferation of cells characterized by cellular atypia, alteration in the maturation sequence, or abnormal differentiation of cells within a tissue, and frequently indicates the emergence of a population of cells that may become cancerous (CitationEustis, 1989). As discussed in CitationBogen et al. (2008) and CitationLong et al. (2003), atypical hyperplasia as observed in the rat tumor assay was considered “an unusual proliferative lesion… [and] may represent a precursor for nasal olfactory carcinogenesis.” As shown in , the rat nasal data provide “some evidence” that CYP2F-induced chronic inflammation and regenerative hyperplasia are key events in the naphthalene MoA, again with the question as to the potential involvement of other CYPs or other metabolic enzymes. However, the olfactory epithelial data provide strong evidence that, regardless of the CYP involved in the primary metabolism of naphthalene, “atypical” hyperplasia may be the key event necessary for carcinoma formation.
Now the question is: What causes atypical hyperplasia in the rat nose, and why is this not happening in the mouse nose? Is it possible that atypical hyperplasia in the rat nose is due to secondary genotoxic action of downstream naphthalene metabolites (e.g., 1,2-naphthoquinone) on susceptible cells via direct reaction with DNA, or possibly through redox cycling to induce oxidative stress and subsequent DNA damage? If so, then what enzymes are involved in this metabolism in the rat nose, and how do activities of these enzymes compare to the mouse nose and other potential target tissues and species?
We propose that the balance of these activities differs between the mouse and rat nose, ultimately resulting in the observed difference in tumor formation. These questions are further addressed in the next section on genotoxicity.
Key event 4—Genotoxicity
Current data and hypothesized MoA
An understanding of the potential genotoxicity of naphthalene and/or its metabolites is critical in developing an accurate MoA for naphthalene carcinogenesis. The Agency for Toxic Substances and Disease Registry (CitationATSDR, 2005) presented a review of over 40 naphthalene genotoxicity studies, 80% of which reported no evidence for genotoxicity for naphthalene. And the majority of positive studies are consistent with secondary effects produced by cytotoxicity. Several recent reviews (CitationSchreiner, 2003; CitationBrusick, 2008; CitationBrusick et al., 2008) have discussed the naphthalene genotoxicity data and that the evidence does not suggest that naphthalene or any of its metabolites drive naphthalene tumorigenicity. However, as outlined in proposed MoA (2) above, US EPA has proposed a genotoxic MoA for naphthalene (CitationUS EPA, 2004), based on data indicating genotoxicity of the naphthalene metabolite 1,2-naphthoquinone. The basis of US EPA’s proposed MoA is derived, in part, from a 1,2-naphthoquinone mutagenesis study by CitationFlowers-Geary et al. (1996), in which 1,2-naphthoquinone was found to cause a small increase in reversion mutations in Salmonella. In addition, 1,2-naphthoquinone and 1,4-naphthoquinone induced SCE in human MNLs (CitationWilson et al., 1996).
Recent studies have shown that 1,2-naphthoquinone reacts with DNA in vitro (CitationSaeed et al., 2007) and in vivo in mouse skin (CitationSaeed et al., 2009) to form depurinating N3-adenine and N7-guanine adducts. CitationSaeed et al. (2007) also observed stable DNA adducts in mouse skin that appeared to derive from the 1,2-naphthoquinone. Although depurinating adducts are known to cause mutations in DNA, at least in vitro, a recent review of the literature evaluating the biological significance for the involvement of N7-guanine adducts in mutagenesis found that there is insufficient evidence that N7-guanine adducts, or their apurinic sites, are the cause of mutations in cells and tissues, and that the stable DNA adducts may be more relevant (CitationBoysen et al., 2009). Therefore, the relevance of naphthalene-induced stable DNA adducts needs further investigation. Studies by CitationKim et al. (1995) and CitationKim et al. (2000) suggest that naphthalene can induce stable N6 adducts of deoxyadenosine from the tetrahydrodiol epoxide (“diol epoxide”) of naphthalene. As shown in , the 1,2-naphthoquinone and diol epoxide adducts are well downstream of the initial naphthalene epoxide and would not likely form until GSH has been depleted and cytotoxicity has already occurred.
No in vivo studies have examined possible DNA adducts in naphthalene-exposed rat or mouse lung or nasal epithelial tissue. As we have mentioned, the balance of naphthalene-metabolizing, -detoxifying, and repair enzymes likely varies between species and tissues, and therefore the potential for formation of DNA adducts likely varies as well and should be examined in each relevant tissue (or cell type). Therefore, as shown in , each species and tissue combination is categorized as “data gap.” In addition, studies have shown that naphthalene metabolites (naphthalene epoxide, 1,2-naphthoquinone, and 1,4-naphthoquinone) react readily with proteins in rats and mice, and in a dose-dependent fashion (CitationZheng et al., 1997; CitationWaidyanatha et al., 2002; CitationWaidyanatha and Rappaport, 2008; CitationBuckpitt et al., 2002), suggesting that covalent binding of naphthalene metabolites to cellular proteins occurs and mediates naphthalene cytotoxicity. Furthermore, these studies, in combination with the genotoxicity studies, suggest that it is very likely that cytotoxicity occurs in cells well before any 1,2-naphthoquinone is available to react directly with DNA or possibly to undergo redox cycling to generate ROS that could react with DNA.
Therefore, as shown in (and described more below), the in vitro data are “strongly refuting,” and the animal data provide “no evidence” that naphthalene-induced genotoxicity occurs prior to cytotoxicity as a cancer initiating event. In addition, since the CitationSaeed et al. (2009) data only suggest that 1,2-naphthoquinone forms DNA adducts in mouse skin, there is “no evidence” that formation of these adducts are key events in mouse lung or rat nasal tumor formation, even as genotoxic events that may occur subsequent to or concurrent with cytotoxicity.
Below we evaluate the elements of the proposed genotoxic MoA against the available data.
HBWoE evaluation of the current data
Although the current genotoxicity data support the potential for genotoxic effects of 1,2-naphthoquinone, no data suggest the presence of these adducts in naphthalene-exposed rat nasal or mouse lung tissue, nor are there data to support that if these adducts were found in those tissues they invoked a tumor-initiating event. The key question we need to try to answer, or at least try to evaluate as plausible, is: Does the 1,2-naphthoquinone react with DNA and cause mutations as an early event in the induction of tumors? Although we realize that other downstream naphthalene metabolites (such as the diol epoxide) may be potentially genotoxic, to simplify our discussion, we focus on the 1,2-naphthoquinone in this section. To try to answer this question, we start by asking the following more pointed questions of the available naphthalene data:
Are there factors that could have influenced the in vitro genotoxicity testing results that we should consider more carefully?
What enzymes are involved in metabolizing the naphthalene epoxide to the 1,2-naphthoquinone? And how might the activities of these enzymes vary among species and tissues that might help explain differences in responses to naphthalene?
If 1,2-naphthoquinone is genotoxic, what is the temporal relationship between its cytotoxic and genotoxic events in vivo?
Do downstream naphthalene metabolites lead to stable DNA adducts in relevant species and tissues? And if so, are these adducts mutagenic?
1. Are there factors that could have influenced the genotoxicity testing results that we should consider more carefully?
Although 80% of the naphthalene genotoxicity studies were negative, some factors should be considered in the context of weighing these data against the proposed MoA hypothesis. For example, many of the negative studies were in bacterial assays (e.g., Ames tests). Is it possible that these studies are negative because CYP2F or CYP2E1 (or other CYP) activities are not present within induced rat liver S9 to generate the primary naphthalene epoxide? Furthermore, what about activities of other enzymes in induced rat liver S9, such as GSH S-transferase that eliminates naphthalene, or EH and DD that generate the 1,2-naphthoquinone? Although no studies confirm the activities of these enzymes toward naphthalene in induced rat liver S9, it is very likely that these activities are present because these enzymes are known to be present in rat liver. In addition, is it possible genotoxicity occurs in the bacterial assays concurrently with or subsequent to cytotoxicity so that the bacterial survival rates are too low to observe a significant mutation frequency? The answer to this question is yes, this is possible, and would therefore point to an MoA that is cytotoxic or dual cytotoxic/genotoxic (this is discussed more below). It is very possible that naphthalene does not cause mutations in the Ames assays because it would require a very large quantity of naphthalene (an amount that would likely be cytotoxic to bacteria) to generate enough 1,2-naphthoquinone to be able to react with DNA, or possibly for the 1,2-naphthoquinone to redox cycle to generate DNA damage via oxygen radicals, as an early event in naphthalene toxicity. Therefore, as shown in , the in vitro data are “strongly refuting” that naphthalene-induced genotoxicity occurs prior to cytotoxicity as a cancer initiating event. In fact, the in vitro data provide “some evidence” that if there is naphthalene-induced genotoxicity, perhaps via the 1,2-naphthoquinone metabolite, it occurs concurrent with or subsequent to cytotoxicity, and is therefore not an initiating event.
2. What enzymes are involved in metabolizing the naphthalene epoxide to the 1,2-naphthoquinone? And how might the activities of these enzymes vary among species and tissues that might help explain differences in responses to naphthalene?
As shown in , the naphthalene epoxide is metabolized by EH to the 1,2-dihydroxy-1,2-dihydronaphthalene (“dihydrodiol”), and then further by DD to the 1,2-naphthoquinone. Much of the focus on the MoA for naphthalene carcinogenesis has been on formation of the primary naphthalene metabolite, the naphthalene epoxide, by CYP2F. However, since there are studies indicating the involvement of the 1,2-napthoquinone in naphthalene cytotoxicity, and the potential for its genotoxic effects, understanding the tissue and species variability of the enzymes involved in formation of the 1,2-naphthoquinone, or other downstream naphthalene metabolites (such as the diol epoxide), is critical toward evaluating an MoA for naphthalene carcinogenicity. CitationSaeed et al. (2009) suggests that metabolism of 1-naphthol to the 1,2-naphthoquinone may be metabolically similar to metabolism of estradiol to catechol estrogen quinones, which have been shown to react with DNA to form N7-guanine and N3-adenine adducts. As discussed by CitationBrusick et al. (2008), CYP1B1 has been shown to metabolize estradiol to catechol estrogen quinones, resulting in N7-guanine and N3-adenine adduct formation in vitro (CitationBelous et al., 2007). Therefore, examining the involvement of CYP1B1 in formation of the 1,2-naphthoquinone seems warranted. Again, it is very likely that species and tissue variability in the balance between the levels and activities of CYP2F (and/or CYP2E1, CYP1B1, or other CYPs) and other enzymes (GSH S-transferase, EH, DD) involved in formation of the 1,2-naphthoquinone or other downstream metabolites, or enzymes involved in repair of DNA damage induced by these metabolites, will help explain differences in species and tissue responses to naphthalene.
For example, as shown in a study by CitationGreen et al. (2001), the activity of EH toward metabolism of the styrene metabolite, styrene epoxide, is about 10-fold higher in the rat than in the mouse nose. The authors suggest that this could be why the rat nose is less susceptible to styrene toxicity than the mouse nose, because the rat is able to more rapidly detoxify the epoxide. A higher activity of EH toward naphthalene would likely have a different outcome. If the activity of EH toward naphthalene is also higher in the rat nose, this might suggest formation of the dihydrodiol and 1,2-naphthoquinone at a higher rate in the rat versus the mouse nose. This might explain the difference in response to naphthalene injury between the mouse and rat nose. There are similar levels of CYP2F in rat and mouse nasal tissue, leading to similar levels of cytotoxicity and hyperplasia. However, upon high levels of naphthalene exposure, GSH depletion, and cytotoxicity, higher levels of EH activity in the rat nose could lead to higher levels of 1,2-naphthoquinone that could lead to genotoxic effects (either from direct reaction with DNA or redox cycling to generate DNA damage via ROS) on already hyperplastic tissue, potentially leading to atypical hyperplasia and tumors. So, unlike styrene for which EH is detoxifying, for naphthalene, high EH in rat nasal tissue may lead to increased levels of the toxic 1,2-naphthoquinone. illustrates potential combinations (yet to be tested) of CYP2F levels, GSH depletion, and EH and DD activities that could lead to the observed naphthalene-induced effects in rodents. If EH is more active toward naphthalene in the rat than the mouse nose, this could explain why there are no tumors in the mouse nose: As shown in row 3 of , GSH depletion may lead to formation of the epoxide in the mouse nose, but EH may not as actively metabolize the epoxide to the 1,2-naphthoquinone as in the rat nose, which could result in more epoxide spontaneously being converted to the 1-naphthol and further to the 1,4-naphthoquinone (), which does not appear to be as genotoxic or cytotoxic, therefore resulting in a lesser response in mouse nose, but not because of CYP2F differences (as discussed, the levels of CYP2F are not very different between rat and mouse nose), but possibly because of lower EH (and possibly DD) in the mouse nose.
Figure 3. Potential combinations of activities that might lead to observed effects.
Also shown in (row 2), it is plausible that EH and DD are lower in the mouse lung than in the rat nose, which could result in lower levels of the 1,2-naphthoquinone formed in the mouse lung, and could be why more adenomas are observed in mouse lung than carcinomas. That is, naphthalene exposure in the mouse lung may result in formation of the epoxide by CYP2F2 metabolism, and then once GSH is depleted (or at the same time), EH metabolizes some of the epoxide to the 1,2-naphthoquinone, but not to the extent that it is formed in the rat nose, so that atypical hyperplasia does not occur. The lack of tumors in the rat lung may potentially be explained by low CYP2F4 so that there is never enough epoxide to deplete GSH and cause sufficient cytotoxicity, or for EH to metabolize to the 1,2-naphthoquinone to cause cytotoxicity and/or genotoxicity (row 4 of ).
Although these are only possible explanations for the observed differences across species, they illustrate the idea that it is very likely that the simple presence or absence of CYP2F activity is not the only critical element of metabolism in the naphthalene MoA. CYP2F is obviously a very important element because formation of the epoxide is necessary before metabolism to 1,2-naphthoquinone is possible, but the extent to which the epoxide is metabolized to the 1,2-naphthoquinone in various species and tissues is also critical. Therefore, the formation of the 1,2-naphthoquinone likely depends on the balance of activities of CYP2F (and/or CYP2E1, CYP1B1, or other CYPs), GSH S-transferase, EH, and DD, and possibly other metabolic enzymes. Since GSH depletion has been shown to precede naphthalene-induced cytotoxicity (CitationPhimister et al., 2004), it is plausible that CYP2F formation of the epoxide followed by GSH depletion are the first (and obligatory) steps in naphthalene-induced injury, but that the potential outcome of the injury will vary depending on the extent of activities of downstream enzymes. An understanding of the balance of activities of these enzymes in different target tissues (e.g., mouse lung and rat nose), and in individual cells, could potentially explain the differences in species and tissue responses to naphthalene injury. And further, understanding of that balance in humans would allow for a more scientifically defensible extrapolation of the MoA from animals to humans (discussed more below).
3. What is the temporal relationship between the cytotoxic and genotoxic events of 1,2-naphthoquinone in vivo?
As discussed, the balance of activities of the various enzymes potentially involved in the formation of the 1,2-naphthoquinone will affect the extent to which the quinone is formed, which is critical to the cytotoxicity and potential genotoxicity of naphthalene. If the balance is such that the 1,2-naphthoquinone is formed in sufficient amounts that it is available to react with cellular proteins to induce cytotoxicity, and/or react directly with DNA or redox cycle to generate ROS that might induce genotoxicity, what is the relationship between these cytotoxic and genotoxic events? In other words, at what point in the naphthalene toxicity pathway could the 1,2-naphthoquinone cause DNA damage?
Prior to cytotoxicity (as an early event)?
Concurrent with cytotoxicity?
Following cytotoxicity (as a secondary event)?
For the 1,2-naphthoquinone to damage DNA as an early event, it would need to be generated and react with DNA early in the metabolic pathway of naphthalene. As discussed, there are sufficient data to suggest that GSH depletion precedes naphthalene-induced cytotoxicity in mouse lung and nasal tissues and increases protein adduct formation (CitationPhimister et al., 2004), suggesting that once GSH is depleted, the levels of the epoxide and potential downstream metabolites (such as the 1,2-naphthoquinone) begin to form and react with cellular proteins to induce cytotoxic effects. Therefore, once GSH is depleted, the questions become: what is the balance between (1) epoxide reaction with proteins?; (2) epoxide metabolism to the dihydrodiol and the 1,2-naphthoquinone by EH and DD?; (3) once the 1,2-naphthoquinone is formed, what is the likelihood that it will react with DNA versus proteins?; and (4) how readily is any DNA damage repaired?
The answers to these questions are likely species and tissue specific, particularly (2), which will depend on the levels and activities of EH and DD that likely have tissue and species variability. The reaction of naphthalene epoxide and 1,2-naphthoquinone with cellular proteins may be relatively more similar across species and tissues because the reaction is not dependent on further metabolism that could vary among species. Naphthalene epoxide and 1,2-naphthoquinone protein adducts have been studied in mouse lung and nasal tissues (CitationZheng et al., 1997; CitationPhimister et al., 2004), rat and monkey nasal tissue (CitationDeStefano-Shields et al., 2010), and in blood of rats and mice after IP dosing of animals with naphthalene (CitationWaidyanatha et al., 2002; CitationWaidyanatha and Rappaport, 2008). These studies suggest that both the epoxide and the 1,2-naphthoquinone react readily with proteins. CitationPhimister et al. (2004) also showed that naphthalene-induced GSH depletion and cytotoxicity was accompanied by increased naphthalene protein adduct formation. Although studies have shown that 1,2-naphthoquinone reacts with DNA in vitro and in vivo in mouse skin (CitationSaeed et al., 2007, Citation2009), no studies have evaluated formation of DNA adducts in naphthalene-exposed mouse lung or rat nasal tissue. Furthermore, genotoxicity studies of naphthalene are predominantly negative, and the positive results suggest secondary DNA damage (e.g., SCE, chromosomal aberrations, DNA fragmentation), suggesting that naphthalene and its metabolites do not induce viable mutations that lead to tumor-initiating events, and that the potential for 1,2-naphthoquinone to cause DNA damage is likely secondary to cytotoxicity. The balance between naphthalene epoxide and 1,2-naphthoquinone formation likely varies between tissue and species and is likely related to the metabolic enzymes involved in generation of the 1,2-naphthoquinone from the epoxide (e.g., EH and DD), but once formed, direct reaction of the 1,2-naphthoquinone with proteins is likely favored over reaction with DNA, likely leading to cytotoxicity before genotoxic events occur.
CitationBagchi et al. (1998a, Citation1998b, Citation2000, Citation2002) have shown that naphthalene toxicity in rodent liver and brain may be, at least in part, potentially due to redox cycling of the 1,2-naphthoquinone or 1,4-naphthoquinone to generate ROS, leading to lipid peroxidation, oxidative stress, and oxidative DNA damage. It is unclear whether the apparent DNA damage in these assays was due to direct effects of naphthalene metabolites or a secondary effect from reactive oxygen species generated from GSH depletion. However, these results indicate the possibility that the quinones could induce oxidative DNA damage concurrently with the induction of quinone protein or DNA adducts; if this does occur, it could suggest a dual cytotoxic/genotoxic MoA for naphthalene.
An understanding of the relative potentials, in various tissues and species, for: detoxification of potentially genotoxic naphthalene metabolites; 1,2-naphthoquinone to form from the epoxide; redox cycling of the quinones to generate ROS; reactivity of these metabolites with DNA versus proteins; and repair of these DNA adducts is critical in evaluating a potential MoA for naphthalene. The large body of negative genotoxicity tests for naphthalene, in combination with studies suggesting that the quinones react readily with proteins, suggests that direct reaction of the 1,2-naphthoquinone with DNA, or redox cycling to generate ROS that can react with DNA, is not likely to occur prior to cytotoxicity. Therefore, in animals, the current data support a MoA for naphthalene tumorigenesis based on cytotoxicity that occurs prior to or concurrent with genotoxicity. The WoE does not support a purely genotoxic MoA in animals.
4. Do downstream naphthalene metabolites lead to stable DNA adducts in relevant species and tissues? And if so, are these adducts mutagenic?
The potential for formation of stable DNA adducts from the naphthalene diol epoxide, or from the 1,2-naphthoquinone, deserves further investigation in mouse and rat nasal and lung tissues. If these adducts are present in these tissues, the potential for these adducts to result in mutations would be important to understand, asking similar questions to those described above. However, since the diol epoxide and the 1,2-naphthoquinone are downstream metabolites of naphthalene metabolism, they are likely to form only after exposure to large doses of naphthalene that result in GSH depletion and cytotoxicity.
Key event 5—Tumor formation
Current data and hypothesized MoA
As summarized in and , and in the review by CitationNorth et al. (2008), a statistically significant increase in alveolar/ bronchiolar adenomas was observed in female mice at the highest exposure concentration (30 ppm), and although an increase was observed in males, it did not reach statistical significance. In addition, alveolar/bronchiolar carcinomas in mice did not reach statistical significance for either sex, and mouse nasal tumors were not observed for either sex. In rats, a statistically significant increase in nasal respiratory epithelial adenomas was observed in males but not females, and a statistically significant increase in olfactory epithelial neuroblastomas was observed in females, at the highest concentration (60 ppm), but not males.
HBWoE evaluation of the current data
The naphthalene tumor data raise the following questions:
What accounts for sex differences in tumor incidence in mice and rats?
What accounts for predominance of adenomas in mouse lung versus neuroblastomas in rat olfactory?
1. What accounts for sex differences in tumor incidence in mice and rats?
It is not clear what accounts specifically for the sex differences in naphthalene-induced tumor incidence in mice and rats. A recent study (CitationOliver et al., 2009) found that female mice were more susceptible to Clara cell injury than male mice, and that at low (50 mg/kg body weight) and medium (100 mg/kg body weight) naphthalene doses (IP injection), cell proliferation and mitosis were more abundant in female than male bronchiolar epithelium and peribronchiolar interstitium, but normal lung tissue regeneration was detected in both sexes. At the highest naphthalene dose (200 mg/kg body weight), lung regeneration was delayed in female mice, while male mice showed a faster, more timely, regenerative response. The authors suggested that this delayed repair response in female mice may be attributed to the more severe damage in female mouse lung tissue, and may be why female mice are more susceptible to naphthalene-induced lung injury in the NTP bioassay (CitationNTP, 1992). Although this study provides histopathologic evidence supporting an explanation for the observed differences, further studies are necessary to understand the specific mechanistic differences that may be responsible for the tumor responses in male and female mice.
Therefore, as shown in , these data account for “some evidence” supporting the observed tumor incidence in rat nose and mouse lung. The rationale for tissues where no tumors have been observed has already been evaluated as part of the previous key events, and is therefore the same as for the previous key events.
2. What accounts for predominance of adenomas in mouse lung versus neuroblastomas in rat olfactory?
As described above, the difference between the predominance of adenomas in mouse lung and neuroblastomas in rat olfactory is possibly related to species and tissue differences in the balance between detoxification of the naphthalene epoxide (GSH S-transferase) and the enzymes involved in the formation of the potentially toxic downstream metabolites of naphthalene (such as the EH and DD activities to generate the dihydrodiol or the 1,2-naphthoquinone); different metabolites may be present in the rat nose that are not present in the mouse lung (or at different levels), potentially leading to the observed differences in adenomas in mouse lung versus carcinomas in rat olfactory.
HBWoE conclusion regarding MoA in animals and possible explanation of observed outcomes
As shown in , the overall WoE is categorized as “ad hoc assumptions required” for all key events in the proposed MoA. Although the rat nasal and mouse lung metabolism data provide “some evidence” of the involvement of CYP2F, the data do not definitively support the sole involvement of CYP2F, nor do they rule out the involvement of other CYPs, such as CYP2E1. Overall categorization as “ad hoc assumptions required” is predominantly due to CYP2F expression, cytotoxicity, chronic inflammation, and regenerative hyperplasia in mouse nasal tissue that does not lead to tumor formation in the mouse nose. CYP2F expression in the rat liver does not lead to the subsequent key events upon naphthalene exposure, and GSH depletion may be the reason for this, but it has not been specifically tested. Therefore, the observations in the mouse nose and rat liver require that we make ad hoc assumptions about the presence or absence of other events in these tissues that are necessary for tumor formation, to explain the observed effects. If the overall WoE for a particular MoA requires ad hoc assumptions for its support, the MoA should be reevaluated based on the available data, perhaps modified to reflect all of the available data, and additional studies carried out to test the hypothesis of the modified MoA. CYP2F knockout and CYP2E1 knockout mice and/or rats, or CYP inhibition studies, using inhibitors more specific toward either CYP, would provide useful information with regard to the involvement of these CYPs in naphthalene metabolic activation and ultimately in naphthalene-induced cytotoxicity and tumor formation. In addition, studies that evaluate the tissue- and species-specific balance of enzymes and activities (CYP2F/CYP2E1 or other CYPs, GSH S-transferase, EH, DD) involved in the formation of the 1,2-naphthoquinone or other downstream metabolites, such as the diol epoxide, could provide an understanding of species and tissue differences in response to naphthalene injury. Furthermore, the naphthalene genotoxicity data suggest that although the 1,2-naphthoquinone potentially induces genotoxicity, genotoxic events are not likely to occur prior to cytotoxic events, and are more likely to occur concurrent with or subsequent to cytotoxicity. Although further investigation, in target cells (mouse and rat lung and nasal), would help to better understand the temporal relationship between the induction of 1,2-naphthoquinone (or other naphthalene metabolites) cytotoxic and genotoxic events, and to more definitively propose a MoA, the current weight of evidence does not support a solely genotoxic MoA for naphthalene carcinogenesis.
HBWoE—Plausibility of MoA extrapolation to humans
Epidemiology data are lacking
To our knowledge, there are no cohort or case-control studies of naphthalene and cancer risk. There are only a few reports of individual cancer cases with prior naphthalene exposure (CitationGriego et al., 2008). These include four smokers with laryngeal cancer among naphthalene purification workers in East Germany and 23 colon cancer cases, half of whom reported taking a treatment containing naphthalene (CitationGriego et al., 2008). In no case can the cancer be definitively attributed to naphthalene. Also, none of these cancers occurred in the lung or nose.
Exposure to polycyclic aromatic hydrocarbons (PAHs) in general has been associated with a modest increase in lung cancer risk (CitationBosetti et al., 2007), but it is unclear if any of this risk is attributable to naphthalene. In contrast, to our knowledge, no studies have examined risks of nasal cancer from PAHs. It is conceivable that this is because nasal cancer is extremely rare, which makes it difficult to study epidemiologically, but nasal cancer has been studied and found to be associated with other environmental and occupational exposures in epidemiological studies (e.g., insoluble nickel, see CitationGoodman et al., 2009). Were PAHs in general, and naphthalene in particular, associated with nasal cancer risk, it certainly could have been studied. Even though the chances of a statistically significant association is low (owing to nasal cancer’s low background rate), any occurrence of this rare cancer would have been notable. Yet, it has never been noted. Although this lack of evidence does not definitively show that naphthalene is not associated with nasal cancers in humans, it is certainly suggestive.
Current in vitro data for naphthalene activity in humans
CitationBuckpitt and Bahnson (1986) showed that naphthalene is metabolized to its epoxide (as measured by presence of the dihydrodiol) in microsomal fractions of human lung tissue at about 3% the rate observed in rodents. Recombinant human lung CYP2F1 has been shown to metabolize naphthalene to its epoxide metabolite in human lymphoblastoid cells at very low rates (CitationBogen et al., 2008; CitationLanza et al., 1999), and CYP2F1 mRNA has been identified in human respiratory tissue, but results in much lower expression than CYP2F4 in rats (CitationBogen et al., 2008; CitationRaunio et al., 1999; CitationDing and Kaminsky, 2003). Currently, no studies have examined the involvement of CYP2F1 in the activation of naphthalene to its epoxide in actual human lung tissue, i.e., in the presence of other CYPs and metabolic and detoxifying enzymes (GSH S-transferase, EH, and DD) in the human lung.
It is important to point out that the naphthalene concentrations used in the bioassays exceeded the maximum tolerable dose (MTD) (CitationNorth et al., 2008), where nearly 100% chronic inflammation occurred in both rats and mice and in both sexes at all doses, indicating that cytotoxicity occurred at all doses. Although this does not in itself constitute proof, especially given the high exposure concentrations, these data suggest that cytotoxicity played a significant role in the observed naphthalene-induced tumor formation in both species. Therefore, since the concentrations used in these bioassays were about 3000-fold greater than naphthalene levels measured in ambient air (CitationGriego et al., 2008), low-dose extrapolation of these data to concentrations humans might be exposed to is problematic.
HBWoE evaluation of the current data
A closer look at the data discussed above raises the following questions:
Are CYP2F, CYP2E1, or other CYPs present, and have they been shown to activate naphthalene to its epoxide or other naphthalene metabolites, in human lung or nasal tissue? And if so, what are the relative amounts of protein and efficiencies of these activities toward naphthalene in human lung or nasal tissue?
What do we know about activities of EH and DD in human lung or nasal tissue? And what might these activities tell us (particularly in relation to levels of CYP2F or other CYPs) about the potential for carcinogenesis in these tissues?
1. Are CYP2F, CYP2E1, or other CYPs present, and have they been shown to activate naphthalene to its epoxide or other naphthalene metabolites, in human lung or nasal tissue? And if so, what are the relative amounts of protein and efficiencies of these activities toward naphthalene in human lung or nasal tissue?
Although to our knowledge CYP2E1 activity toward naphthalene has not been tested directly in human lung tissue, CYP2E1 is present in human lung tissue (CitationDing and Kaminsky, 2003) and was shown to activate naphthalene to its epoxide in pooled human liver microsomes (pHLMs), since this activity was decreased in the presence of the potent CYP2E inhibitor 4-methylpyrazole (CitationChang et al., 2006). CitationCho et al. (2006) showed that CYP2E1 activated naphthalene to 1-naphthol and 2-naphthol. CitationCho et al. (2006) also showed that CYP1A2 was the most active isoform for producing the dihydrodiol and 1-naphthol metabolites, CYP1A2 and 2D6*1 were the most active isoform for producing 1,4-naphthoquinone, CYP3A4 was most effective for 2-naphthol production, and CYP2A6 and CYP3A4 were most active in metabolizing the dihydrodiol. Expression of these CYP isoforms has been observed in human lung (CitationDing and Kaminsky, 2003).
Therefore, the current available data indicate that CYP2E1, CYP1A2, CYP3A4, and CYP2A6 in human lung tissue could be active toward naphthalene. The amount and catalytic activity of these CYPs in human lung, relative to CYP2F1 (or other CYPs) and other metabolic and detoxifying enzymes that are active toward naphthalene metabolites in human lung, ultimately determines the extent of involvement of these individual CYPs toward naphthalene-induced tissue injury. To our knowledge, experiments examining these activities in human lung tissue, or in individual human lung cells, have not been done.
A recent study by CitationFukami et al. (2008) evaluated the activity of human CYP2A13 toward naphthalene. CYP2A13 is expressed in the human respiratory tract, with highest levels in the nasal mucosa, followed by the lung and trachea (CitationSu et al., 2000). CitationFukami et al. (2008) found that human CYP2A13, assayed in vitro from an Esherichia coli membrane preparation, catalyzed formation of 1-naphthol and 2-naphthol from naphthalene. These data suggest that in human nasal tissue, if CYP2A13 is found to be an important CYP isozyme for naphthalene metabolism, perhaps metabolism to the naphthols results in little, if any, metabolism to the toxic 1,2-naphthoquinone (see ), which might explain the apparent lack of naphthalene-induced injury in human nasal mucosa.
Immunolocalization studies did not detect CYP2F in rhesus macaques pulmonary tissue, and detected only low levels (20- to 30-fold less than in rodents) in nasal ethmoturbinates (CitationBaldwin et al., 2004). Studies of naphthalene metabolism from microsomal preparations made from mouse, hamster, rat, and monkey lung tissue showed that naphthalene was metabolized to the dihydrodiol at a rate approximately 1% that of the mouse (CitationBuckpitt et al., 1992). CitationBoland et al. (2004) also showed that metabolism of naphthalene to the dihydrodiol in dissected airways of rhesus macaques was about 100-fold less than observed in rodents. These data, in combination with the human lung microsome data from CitationBuckpitt and Bahnson (1986), suggest that regardless of the CYP isoform involved in naphthalene metabolism in primates (including humans), the rate of metabolism is potentially much lower in primates, suggesting that primate airways may not be as susceptible as rodent airways to naphthalene injury. Experiments in individual lung cells would help support this interpretation.
2. What do we know about activities of EH and DD in human lung or nasal tissue? And what might these activities tell us (particularly in relation to levels of CYP2F or other CYPs) about the potential for carcinogenesis in these tissues?
Based on the available data for humans and nonhuman primates, and the proposed MoA for animals (cytotoxic or dual cytotoxic/genotoxic), we can consider the plausibility of the MoA in humans. The data suggest that whether naphthalene is metabolized by CYP2F, CYP2E1, or other CYPs in human lung, the rate of metabolism to the epoxide is very low, likely 1% that observed in the mouse lung (at least as shown in whole human lung microsomes and regionally in monkey dissected airways). Nevertheless, there is concern that the higher levels of EH and DD in human lung compared to rodents could result in a higher susceptibility of human lung to naphthalene-induced injury (CitationBuckpitt et al., 2002). However, since the rate of the first, and obligatory, step in the metabolism of naphthalene is very low, it is possible that GSH is not depleted, so that the epoxide is never available for metabolism to the dihydrodiol and 1,2-napthoquinone by EH and DD, regardless of the high levels of these enzyme activities in human lung. This is illustrated in (row 5) along with the proposed enzyme combinations in animals and primates. Therefore, differences between humans and animals in the balance of the enzymes and rates of activities involved in formation of the epoxide and 1,2-naphthoquinone, particularly the very low rate of metabolism to the epoxide in humans compared to rodents, suggest that the animal MoA for naphthalene is not likely in human lung tissue, and furthermore that naphthalene is not likely to induce lung tumors in humans. The pathways shown in are suggestions based on the available data. Further investigation (at the cellular level) would lead to a better understanding of these pathways across species and tissues, leading to a better understanding of the potential for any naphthalene-induced lung tissue damage in humans, and why an association between naphthalene exposure and nasal cancer has never been noted in humans.
Several questions that should be addressed before any MoA for naphthalene can be supported
Is CYP2F the only CYP involved in metabolism of naphthalene to its epoxide in target tissues? Or are other CYPs possibly involved?
Once GSH is depleted in target tissues, what enzymes are involved in metabolism of the epoxide to the 1,2-naphthoquinone (or other potentially toxic metabolites)?
How might the balance of enzymes involved in epoxide formation (CYP2F/CYP2E1 or other CYP activity), GSH depletion, 1,2-naphthoquinone formation (EH and DD), or diol epoxide formation, vary among target tissues? And might these distinctions explain the observed variations in responses to naphthalene in target tissues and species?
Is 1,2-naphthoquinone (or the diol epoxide) genotoxic in target tissues? And if so, when are these events likely to occur? Evidence so far does not suggest genotoxicity as an early event. Is genotoxicity occurring concurrent with or subsequent to cytotoxicity?
Once we have a better understanding of what might be happening in animals, similar testing in potential human target tissues or cells (lung and nasal) would help inform any extrapolation to a potential human MoA.
Discussion and conclusions
The purpose of WoE is not (and should not be) simply to arrive at decisions in the face of incomplete data, but rather to describe where on the spectrum from established human risk to disproven human risk the available data put our understanding of potential hazard. Existing schemes (International Agency for Research on Cancer [IARC], US EPA) try to use WoE categories to express where along this spectrum the evidence lies. These schemes are not very specific about how data are to be used, however, and often the question “How much evidence is necessary?” arises. Although it is often not explicit, the basic rationale for these schemes is based on the “universality” idea discussed at the outset of this paper—that observations in one setting suggest that similar results would be obtained in other settings, including the setting of the human population being protected. MoA does come in, but usually as a secondary factor and not in a particularly formal way. That is, existence of multiple positive studies is taken as evidence that effect is not specific to one species/system/study and may therefore be general. The justifying arguments for this assertion usually flow from policy, precedent, analogy with other cases, etc., rather than from case-specific inferences.
In WoE, judgment is necessary, but what the rationale and reasoning for conclusions may be is rarely explicit. This leads to disputation of the judgments based largely on ad hominem considerations—who is judging (and whether they are “unbiased”) rather than on the soundness of their judgments per se. Our approach is aimed at making the connection between judgments and case-specific evidence more explicit. In this way, the discussion can focus on the scientific interpretation of specific observations and the degree to which that interpretation is supported by those observations, rather than on who is making the interpretation. It should foster a more scientific and objective evaluation of WoE.
To the degree that the issues come down to evaluation of MoA, the soundness of the MoA conclusions for the animal studies and the question of whether humans have the same MoA elements, the HBWoE method is complementary to the Human Relevance/MoA framework (CitationBoobis et al., 2006; CitationCohen et al., 2003; CitationMeek et al., 2003; CitationSonich-Mullin et al., 2001). Our approach makes more explicit how one should evaluate the MoA information, and it shows the value of looking beyond just the single-animal model in which the response is seen to consider what happens (and what does not happen) in other nontarget species and tissues. It calls attention to the role of inconsistent information, not just to the plausibility of the proposed MoA elements in setting where they produced the endpoint of interest. It emphasizes the role of wider scientific understanding in judging what reasonable inferences are, and it points out the pitfalls of post hoc reasoning about the potential role of MoA elements (remembering the example of direct air contact explaining the location of tumors).
HBWoE comes down to evaluation of alternative “accounts.” An account (which we put forth in this context as a technical term) is a proposed set of explanations for the set of observed phenomena across the body of relevant observations. The explanations need not be proven—what is important is that one set out what is being proposed as causal and generalizable phenomena, what is the proposed basis for deviations that lead to observations that do not fit (i.e., that would otherwise be counterexamples or refutations), what assumptions are made that are ad hoc (to explain particulars, but for which the evidence consists of their plausibility and the observations they are adduced to explain), what further assumptions have to be made (and how reasonable they are), and what is relegated to error, happenstance, or other causes not relevant to the question at hand. There are competing accounts, and one should evaluate the main ones as to how the evidence supports them; what is necessary to assume; and overall, how the WoE for each suggests how compelling the account is.
The importance of analogy, or considering alternate “accounts,” is made explicit by Austin Bradford Hill in his seminal paper on distinguishing causality from association:
None of my nine viewpoints can bring indisputable evidence for or against the cause-and-effect hypothesis and none can be required as a sine qua non. What they can do, with greater or less strength, is to help us to make up our minds on the fundamental question—is there any other way of explaining the set of facts before us, is there any other answer equally, or more, likely than cause and effect? (Bradford Hill, 1965) [emphasis added]
The essence of the accounts is that they constitute being explicit about Bradford Hill’s “ways of explaining the set of facts before us.” They are not conclusions or findings, but rather provisional proposals for the reasons behind the set of observations we have at hand, set out in a way that makes clear where assumptions, interpretations, and tentative inferences have been drawn. It is by comparing alternative accounts—alternative hypotheses about what causal effects actually exist—and assessing: their comparative success at explaining phenomena; their comparative need for assumptions to fill in gaps (and the comparative reasonableness of those assumptions); and their comparative invocation of ad hoc suppositions that are necessary to accommodate what might otherwise be inexplicable results, that we can judge how compelling each alternative should be deemed, and hence with what degree of confidence we can judge the hypothesized causal processes (and their consequences for human risk estimation) to be supported by the factual record.
An account is most compelling when it is not only in good accord with the data, but helps explain the data in a parsimonious way (not elaborating a separate reason for every observation but instead finding common reasons for sets of observations), and, moreover, achieves this explanatory ability much more readily than any competing account. The alternative accounts reflect the set of hypotheses and what reasons we are in effect accepting as true for each set of available associated observations if we subscribe to the account.
For naphthalene, one account has the animal tumors not indicative of a potential human risk, by virtue of an MoA that would not operate in humans, especially at anticipated exposure levels. This account is that the mouse lung tumors and the rat nasal tumors are produced by a common MoA that includes marked cytotoxicity in those tissues, and tumor formation is secondary to such toxicity, not occurring in its absence. The tissue cytotoxicity is produced by marked metabolic activation of naphthalene to reactive metabolites, probably involving CYP2F, locally in the target tissues. It is only because of GSH depletion that sufficient levels of the reactive metabolites are generated, so this is a high-dose phenomenon even in the rodents. It is not clear whether there is any contribution of genotoxicity, but this would appear to be an issue only at high doses, at which GSH depletion leads to metabolites not produced in abundance at lower doses (where epoxide is readily conjugated and removed), resulting in genotoxicity that occurs concurrently with cytotoxicity.
Under this account, humans are understood as not susceptible to this MoA because they have insufficient metabolic activation to the naphthalene epoxide to deplete GSH or to create sufficient levels of reactive metabolites so as to produce cytotoxicity, genotoxicity, or both, and without such toxicity no carcinogenic risk is induced.
This account—if true—explains why the animal tumors occur in the particular tissues they do (local metabolic activity sufficient to deplete GSH, causing metabolic activation of the epoxide, possibly by EH and DD, to the toxic 1,2-naphthoquinone or other metabolites); it accounts for the collocation of CYP2F, cytotoxicity, and tumors. It accounts for why there are similar syndromes for the tumors in mice and in rats, and it specifically accounts for why tumor responses do not occur in other tissues. It accounts for why inhibition of CYP metabolism extinguishes the cytotoxicity and why IP exposure leads to cytotoxicity in the same tissues as does inhalation. It explains why tumors are only seen at the high doses. It explains why humans do not have a clear relationship between naphthalene exposure and cytotoxicity in any tissues (they have insufficient metabolic activation). It explains the lack of rodent liver tumors, despite naphthalene metabolism by liver, by invoking the very high GSH levels in liver, in accord with data showing lack of GSH depletion, and it figures that this means that the cytotoxic metabolites are not produced significantly when naphthalene epoxide is readily cleared by GSH conjugation. It explains the weak in vivo genotoxicity of naphthalene despite the fact that naphthalene metabolites are genotoxic if given directly in vitro and in bacterial assays.
Incongruous elements in this account are that the mouse nose has the metabolic activation and cytotoxicity as well, but no tumors. Evidently, there is something about mouse nasal tissue that makes the cytotoxicity less able to cause transformation of cells. The reason is not clear, and one must invoke a reason to fully believe in the other aspects of this account. It is noteworthy that a number of other structurally related chemicals also cause mouse lung tumors and mouse nasal toxicity but no nasal tumors (CitationCruzan et al., 2009), so the pattern is repeatable and is not likely an error of the particular naphthalene experiments. There is also some evident systematic effect in operation, the nature of which is not at present clear. It may be that key metabolites necessary for tumorigenesis, but not for cytotoxicity, are not as readily formed in mouse nose (e.g., possibly the 1,2-naphthoquinone). Or it may be that, for some reason, mouse nasal epithelium is especially resistant to the effects of cytotoxicity on tumorigenesis. These are largely speculations invoked to explain the observed phenomena (which is not to say that they are wrong, only that they are ad hoc and unproven). It is not clear how any such mouse nose-specific effect would bear on the possibility of human risk from naphthalene. Here we speculate that it could be the initial rate of epoxide formation, followed by GSH depletion (if the rate of epoxide formation is high enough to deplete GSH), and then by activities of EH and DD on the epoxide to form the 1,2-naphthoquinone that is important for tumorigenesis, and that the 1,2-naphthoquinone may not form as readily in the mouse versus rat nasal mucosa. There is evidence to suggest this may be true (i.e., higher EH activity toward styrene epoxide in the rat versus mouse nose has been observed). We also suggest the possibility that with relatively low levels of 1,2-naphthoquinone (as may be the case in the mouse versus rat nose) cytotoxicity may occur without carcinogenicity. Since evidence suggests that the epoxide formation is much lower in human lung compared with mouse lung or mouse and rat nasal tissue, much lower levels of the 1,2-naphthoquinone (or possibly none) are likely in human lung, suggesting minimal naphthalene-induced tissue injury in humans.
A related incongruous element is the lack of tumors in rat lung. Here, the degree of naphthalene metabolism is not as marked as in mice, and in fact is very mild, and the cytotoxicity is only inflammation rather than cell killing. And it may be that the observed effects are unrelated to the exposure. Or, it may be that the effect is related to exposure but simply not strong enough. The latter is consistent with the idea of it being a high-dose-only phenomenon, but the insufficiency in rat lung (although the lower metabolism and toxicity response can be observed) is somewhat ad hoc, in that our proposal that the observed levels of these processes are insufficient to cause tumors is based in part on the observed lack of tumors—the phenomenon we are seeking to explain. If rat lung tumors had been seen in the bioassay, even with the lower metabolism and toxicity we could have explained this as a consequence of the toxicity, which, although mild, was evidently sufficient, and this would have been seen as consistent with the mouse lung tumors. That is, the degree of effect proposed as necessary is an adjustment to the hypothesis invoked to explain why it is sufficient in mice but (evidently) not in rat lung.
To some extent, the focus on cytotoxicity and CYP2F as key events could be considered ad hoc in that they are suggested after the fact by the collocation of these with the tumors in the animal inhalation bioassays. That is, they were not observed first and used as predictions of where tumors would occur. It is conceivable, though unlikely, that they are only correlated with the site of tumors by happenstance but are not involved in the MoA. The fact that there are similar syndromes in both the rat nasal and mouse lung tissues reduces the reasonableness of the happenstance explanation. Cytotoxicity is well recognized as a key factor in tumorigenesis generally, independently from its pattern in the naphthalene bioassays, and if it were not a key event, one is left with no explanation for why naphthalene caused no tumors in other tissues where it is not cytotoxic but to which it is transported systemically and in which it may be metabolized. The focus on CYP metabolism is supported by experimental evidence, but CYPs other than 2F have activity toward naphthalene, and indeed may generate at least some of the same metabolites. It may be that the fact that 2F has prominently high activity in both target tissues but not elsewhere is coincidence, though this seems unlikely. At present, though, it is not clear what is unique about 2F-mediated metabolism that makes it especially cytotoxic and effective in tumorigenesis. It may be unique metabolites or action of 2F in further downstream metabolism of naphthalene epoxide or its metabolites. But it may also be that CYP-mediated metabolism of naphthalene by any CYP, as long as it is of sufficient magnitude, and as long as GSH depletion allows the formation of sufficient levels of reactive metabolites, would suffice to cause cytotoxicity. The low levels of human 2F activity may be directly important, but even if it is not specifically 2F, human metabolism of naphthalene to its epoxide by other CYPs is not marked enough to lead to GSH depletion and substantial reactive metabolite formation, even in those tissues with the greatest CYP activity and greatest activity toward naphthalene.
There is a second, contrasting account in which the observed animal tumor responses do indicate a cancer-induction risk to humans from low-level naphthalene exposures. There are actually two related versions of this account—a general one simply invokes the commonality among tissues and species in the cellular processes that control differentiation and proliferation (with the logic being that the rat nasal and mouse lung tumors show that such processes can be altered by naphthalene to lead to the production of malignant cells, and so some version of this alteration might occur in some tissues in humans as well). A more specific one asserts that genotoxicity of naphthalene via DNA-reactive metabolites generated by CYP-mediated metabolism are early initiating events sufficient to cause the tumorigenic effect seen in animals, and that they would act in a similar way in human tissues in which they were produced. The general version of this account provides no reason for why the specific tissues having tumor response in the animals are the ones affected, while other tissues (that share the presumed-to-be-general cellular control processes) are not affected. One must presume some yet unknown factor causes the rat and mouse tissue-site specificity (presumably a different factor in each species, since the tissue specificity differs) and one must also presume that some human tissues have the functional equivalent of at least one of these factors, making some unspecified human tissue to have a risk-generating process that mimics whatever it is that produces naphthalene-induced tumors in the animal bioassays. The more specific genotoxicity hypothesis has no ready explanation for why certain tissues in animals that have substantial naphthalene metabolism (for example, liver) do not have tumors. It must be presumed that the array of metabolites is wrong in the liver, or that some other unspecified factor is missing or different, but that this factor is in its nonprotective state in some human tissues with metabolic activity toward naphthalene. At the same time, the alternative account ascribes the apparent tying of mouse lung tumors and rat nasal tumors to the particular CYP2F-mediated metabolism of naphthalene, the collocation of cytotoxicity and tumorigenesis, and the similarities in this association of factors in the mouse and rat responding tissues and the lack of these factors in nonresponding tissues as coincidences that are not causally important, or at the very least not necessary parts of the tumorigenic response (because if they were necessary, their lack in humans would preclude the applicability of the animal bioassay results to projecting human risk).
There is a very large degree of ad hoc argument in this second account. It has to invoke a number of unnamed and unobserved factors, the specific nature of which is not suggested and which have no positive evidence for their existence, and have as essentially their only argument that, if there indeed were such factors having the properties imagined, then the observed outcomes in rats and mice would occur and there would be some relevance of these to a corresponding process in humans. That is, the properties of the causal model are chosen so as to produce the particular outcomes already observed. This does not disprove their possibility, but when it comes to assessing the WoE, it is clear that the inferences about human risk are not coming from the data themselves, but from the assumptions invoked after the fact, without evidence that would make the animal tumors have something in common with a human risk-generating process. Therefore, the WoE for this account (i.e., an initiating genotoxic MoA) is weak compared to the more substantial WoE supporting either a cytotoxic or dual cytotoxic/genotoxic MoA for naphthalene carcinogenesis.
Declaration of interest
The authors’ affiliation is as shown on the first page. This paper was prepared with financial support to Gradient, a private environmental consulting firm, from ExxonMobil Biomedical Sciences and the American Petroleum Institute (API), a trade association representing producers, refiners, and distributors of petroleum products. The work reported in the paper was conducted during the normal course of employment by Gradient. The authors have the sole responsibility for the writing and contents of this paper. The views and opinions expressed are not necessarily those of API or ExxonMobil.
Related Research Data
References
- Abdo KM, Eustis SL, McDonald M, Jokinen MP, Adkins B, Haseman JK. (1992). Naphthalene: A respiratory tract toxicant and carcinogen for mice. Inhal Toxicol 4:393–409.
- Abdo KM, Grumbein S, Chou BJ, Herbert R. (2001). Toxicity and carcinogenicity study in F344 rats following 2 years of whole-body exposure to naphthalene vapors. Inhal Toxicol 13:931–950.
- Agency for Toxic Substances and Disease Registry (ATSDR). (2005). Toxicological profile for naphthalene, 1-methylnaphthalene, and 2-methylnaphthalene. Atlanta: ATSDR, 347 pp, August.
- Bagchi D, Bagchi M, Balmoori J, Vuchetich PJ, Stohs SJ. (1998a). Induction of oxidative stress and DNA damage by chronic administration of naphthalene to rats. Res Commun Mol Pathol Pharmacol 101:249–257.
- Bagchi M, Bagchi D, Balmoori J, Ye X, Stohs SJ. (1998b). Naphthalene-induced oxidative stress and DNA damage in cultured macrophage J774A.1 cells. Free Radic Biol Med 25:137–143.
- Bagchi D, Balmoori J, Bagchi M, Ye X, Williams CB, Stohs SJ. (2000). Role of p53 tumor suppressor gene in the toxicity of TCDD, endrin, naphthalene, and chromium (VI) in liver and brain tissues of mice. Free Radic Biol Med 28:895–903.
- Bagchi D, Balmoori J, Bagchi M, Ye X, Williams CB, Stohs SJ. (2002). Comparative effects of TCDD, endrin, naphthalene and chromium (VI) on oxidative stress and tissue damage in the liver and brain tissues of mice. Toxicology 175:73–82.
- Baldwin RM, Jewell WT, Fanucci MV, Plopper C, Buckpitt A. (2004). Comparison of pulmonary/nasal CYP2F expression levels in rodents and rhesus macaque. J Pharmacol Exp Ther 309:127–136.
- Baldwin, RM, Shultz MA, Buckpitt AR. (2005). Bioactivation of the pulmonary toxicants naphthalene and 1-nitronaphthalene by rat CYP2F4. J Pharmacol Exp Ther 312:857–865.
- Belous AR, Hachey DL, Dawling S, Roodi N, Parl FF. (2007). Cytochrome P450 1B1-mediated estrogen metabolism results in estrogen-deoxyribonucleoside adduct formation. Cancer Res 67:812–817.
- Bogen KT. (2008). An adjustment factor for mode-of-action uncertainty with dual-mode carcinogens: The case of naphthalene-induced nasal tumors in rats. Risk Anal 28:1033–1051.
- Bogen KT, Benson JM, Yost GS, Morris JB, Dahl AR, Clewell HJ, Krishnan K, Omiecinski CJ. (2008). Naphthalene metabolism in relation to target tissue anatomy, physiology, cytotoxicity and tumorigenic mechanism of action. Regul. Toxicol. Pharmacol. 51(2 Suppl 1):27–36.
- Boland B, Lin CY, Morin D, Miller L, Plopper C, Buckpitt A. (2004). Site-specific metabolism of naphthalene and 1-nitronaphthalene in dissected airways of rhesus macaques. J Pharmacol Exp Ther 310:546–554.
- Boobis AR, Cohen SM, Dellarco V, McGregor D, Meek ME, Vickers C, Willcocks D, Farland W. (2006). IPCS framework for analyzing the relevance of a cancer mode of action for humans. Crit Rev Toxicol 36:781–792.
- Bosetti C, Boffetta P, La Vecchia C. (2007). Occupational exposures to polycyclic aromatic hydrocarbons, and respiratory and urinary tract cancers: A quantitative review to 2005. Ann Oncol 18:431–446.
- Boysen G, Pachkowski BF, Nakamura J, Swenberg JA. (2009). The formation and biological significance of N7-guanine adducts. Mutat Res 678:76–94.
- Brusick D. (2008). Critical assessment of the genetic toxicity of naphthalene. Regul Toxicol Pharmacol 51(2 Suppl 1):37–42.
- Brusick D, Small MS, Cavalieri EL, Chakravarti D, Ding X, Longfellow DG, Nakamura J, Rogan EC, Swenberg JA. (2008). Possible genotoxic modes of action for naphthalene. Regul Toxicol Pharmacol 51(2 Suppl 1):43–50.
- Buckpitt AR, Bahnson LS. (1986). Naphthalene metabolism by human lung microsomal enzymes. Toxicology 41:331–341.
- Buckpitt A, Buonarati M, Avey LB, Chang AM, Morin D, Plopper CG. (1992). Relationship of cytochrome P450 activity to Clara cell cytotoxicity. II. Comparison of stereoselectivity of naphthalene epoxidation in lung and nasal mucosa of mouse, hamster, rat and rhesus monkey. J Pharmacol Exp Ther 261:364–372.
- Buckpitt A, Chang AM, Weir A, Van Winkle L, Duan X, Philpot R, Plopper C. (1995). Relationship of cytochrome P450 activity to Clara cell cytotoxicity. IV. Metabolism of naphthalene and naphthalene oxide in microdissected airways from mice, rats, and hamsters. Mol Pharmacol 47:74–81.
- Buckpitt A, Boland B, Isbell M, Morin D, Shultz M, Baldwin R, Chan K, Karlsson A, Lin C, Taff A, West J, Fanucci M, Van Winkle L, Plopper C. (2002). Naphthalene-induced respiratory tract toxicity: Metabolic mechanisms of toxicity. Drug Metab Rev 34:791–820.
- Chang JH, Kochansky CJ, Shou M. (2006). The role of P-glycoprotein in the bioactivation of raloxifene. Drug Metab Dispos 34:2073–2078.
- Cho TM, Rose RL, Hodgson E. (2006). In vitro metabolism of naphthalene by human liver microsomal cytochrome P450 enzymes. Drug Metab Dispos 34:176–183.
- Cohen SM, Meek ME, Klaunig JE, Patton DE, Fenner-Crisp PA. (2003). The human relevance of information on carcinogenic modes of action: Overview. Crit Rev Toxicol 33:581–589.
- Cruzan G, Bus J, Gingell R, Carlson G. (2009). Mouse specific lung tumors from CYP2F2-mediated cytotoxic metabolism: An endpoint/toxic response where data from multiple chemicals converge to support a mode of action. Regul Toxicol Pharmacol 55:205–218.
- DeStefano-Shields C, Morin D, Buckpitt A. (2010). Formation of covalently bound protein adducts from the cytotoxicant naphthalene in nasal epithelium: Species comparisons. Environ Health Perspect 118:647–652.
- Ding X, Kaminsky LS. (2003). Human extrahepatic cytochromes P450: Function in xenobiotic metabolism and tissue-selective chemical toxicity in the respiratory and gastrointestinal tracts. Annu Rev Pharmacol Toxicol 43:149–173.
- Eustis SL. (1989). The sequential development of cancer: A morphological perspective. Toxicol Lett 49:267–281.
- Flowers-Geary L, Bleczinki W, Harvey RG, Penning TM. (1996). Cytotoxicity and mutagenicity of polycyclic aromatic hydrocarbon ortho-quinones produced by dihydrodiol dehydrogenase. Chem Biol Interact 99:55–72.
- Fukami T, Katoh M, Yamazaki H, Yokoi T, Nakajima M. (2008). Human cytochrome P450 2A13 efficiently metabolizes chemicals in air pollutants: Naphthalene, styrene, and toluene. Chem Res Toxicol 21:720–725.
- Genter MB, Marlowe J, Kerzee JK, Dragin N, Puga A, Dalton TP, Nebert DW. (2006). Naphthalene toxicity in mice and aryl hydrocarbon receptor-mediated CYPs. Biochem Biophys Res Commun 348:120–123.
- Goodman JE, Prueitt RL, Dodge DG, Thakali S. (2009). Carcinogenicity assessment of water-soluble nickel compounds. Crit Rev Toxicol 39:365–417.
- Green T, Lee R, Toghill A, Meadowcraft S, Lund V, Foster J. (2001). The toxicity of styrene to the nasal epithelium of mice and rats: Studies on the mode of action and relevance to humans. Chem Biol Interact 137:185–202.
- Griego FY, Bogen KT, Price PS, Weed DL. (2008). Exposure, epidemiology and human cancer incidence of naphthalene. Regul Toxicol Pharmacol 51(2 Suppl 1):22–26.
- Hill AB. (1965). The environment and disease: Association or causation? Proc R Soc Med 58:295–300.
- Kim SJ, Jajoo HK, Kim HY, Zhou L, Horton P, Harris CM, Harris TM. (1995). An efficient route to N6 deoxyadenosine adducts of diol epoxides of carcinogenic polycyclic aromatic hydrocarbons. Bioorg Med Chem 3:811–822.
- Kim HY, Finneman JI, Harris CM, Harris TM. (2000). Studies of the mechanisms of adduction of 2'-deoxyadenosine with styrene oxide and polycyclic aromatic hydrocarbon dihydrodiol epoxides. Chem Res Toxicol 13:625–637.
- Lanza DL, Code E, Crespi CL, Gonzalez FJ, Yost GS. (1999). Specific dehydrogenation of 3-methylindole and epoxidation of naphthalene by recombinant human CYP2F1 expressed in lymphoblastoid cells. Drug Metab Dispos 27:798–803.
- Lee MG, Phimister A, Morin D, Buckpitt A, Plopper C. (2005). In situ naphthalene bioactivation and nasal airflow cause region-specific injury patterns in the nasal mucosa of rats exposed to naphthalene by inhalation. J Pharmacol Exp Ther 314:103–110.
- Long PH, Herbert RA, Peckham JC, Grumbein SL, Shackelford CC, Abdo K. (2003). Morphology of nasal lesions in F344/N rats following chronic inhalation exposure to naphthalene vapors. Toxicol Pathol 31:655–664.
- Meek ME, Bucher JR, Cohen SM, Dellarco V, Hill RN, Lehman-McKeeman LD, Longfellow DG, Pastoor T, Seed J, Patton DE. (2003). A framework for human relevance analysis of information on carcinogenic modes of action. Crit Rev Toxicol 33:591–653.
- Morris JB, Buckpitt AR. (2009). Upper respiratory tract uptake of naphthalene. Toxicol Sci 111:383–391.
- Nagata K, Martin BM, Gillette JR, Sasame HA. (1990). Isozymes of cytochrome P-450 that metabolize naphthalene in liver and lung of untreated mice. Drug Metab Dispos 18:557–564.
- National Toxicology Program (NTP). (1992). Toxicology and Carcinogenesis Studies of Naphthalene (CAS No. 91-20-3) in B6C3F1 Mice (Inhalation Studies). Bethesda, MD: US Public Health Service, National Institutes of Health. NTP-TR-410; NIH publication no. 92-3141.
- NTP. (2000). Toxicology and Carcinogenesis Studies of Naphthalene (CASD No. 91-20-3) in F344/N Rats (Inhalation Studies). National Toxicology Program. Rockville, MD: US Department of Health and Human Services, National Institutes of Health. Technical report series no. 500.
- North DW, Abdo KM, Benson JM, Dahl AR, Morris JB, Renne R, Witschi H. (2008). A review of whole animal bioassays of the carcinogenic potential of naphthalene. Regul Toxicol Pharmacol 51(2 Suppl 1):6–14.
- Oliver JR, Kushwah R, Wu J, Cutz E, Yeger H, Waddell TK, Hu J. (2009). Gender differences in pulmonary regenerative response to naphthalene-induced bronchiolar epithelial cell injury. Cell Prolif 2009 Oct;42(5):672-87.
- Phimister A, Lee M, Morin D, Buckpitt AR, Plopper CG. (2004). Glutathione depletion is a major determinant of inhaled naphthalene respiratory toxicity and naphthalene metabolism in mice. Toxicol Sci 82:268–278.
- Plopper CG, Suverkropp C, Morin D, Nishio S, Buckpitt A. (1992a). Relationship of cytochrome P-450 activity to Clara cell cytotoxicity. I. Histopathologic comparison of the respiratory tract of mice, rats and hamsters after parenteral administration of naphthalene. J Pharmacol Exp Ther 261:353–363.
- Plopper CG, Macklin J, Nishio SJ, Hyde DM, Buckpitt AR. (1992b). Relationship of cytochrome P-450 activity to Clara cell cytotoxicity. III. Morphometric comparison of changes in the epithelial populations of terminal bronchioles and lobar bronchi in mice, hamsters, and rats after parenteral administration of naphthalene. Lab Invest 67:553–565.
- Raunio H, Hakkola J, Hukkanen J, Lassila A, Paivarinta L, Pelkonen O, Anttila S, Piipari R, Boobis A, Edwards RJ. (1999). Expression of xenobiotic-metabolizing CYPs in human pulmonary tissue. Exp Toxicol Pathol 51:412–417.
- Saeed M, Higginbotham S, Rogan E, Cavalieri E. (2007). Formation of depurinating N3adenine and N7guanine adducts after reaction of 1,2-naphthoquinone or enzyme-activated 1,2-dihydroxynaphthalene with DNA. Implications for the mechanism of tumor initiation by naphthalene. Chem Biol Interact 165:175–188.
- Saeed M, Higginbotham S, Gaikwad N, Chakravarti D, Rogan E, Cavalieri E. (2009). Depurinating naphthalene-DNA adducts in mouse skin related to cancer initiation. Free Radic Biol Med 47:1075–1081.
- Schreiner CA. (2003). Genetic toxicity of naphthalene: A review. J Toxicol Environ Health B Crit Rev 6:161–183.
- Shultz MA, Choudary PV, Buckpitt AR. (1999). Role of murine cytochrome P-450 2F2 in metabolic activation of naphthalene and metabolism of other xenobiotics. J Pharmacol Exp Ther 290:281–288.
- Simmonds AC, Ghanayem BI, Sharma A, Reilly CA, Millen B, Yost GS, Forkert, PG. (2004). Bioactivation of 1,1-dichloroethylene by CYP2E1 and CYP2F2 in murine lung. J Pharmacol Exp Ther. 310:855–864.
- Sonich-Mullin C, Fielder R, Wiltse J, Baetcke K, Dempsey J, Fenner-Crisp P, Grant D, Hartley M, Knaap A, Kroese D, Mangelsdorf I, Meek E, Rice JM, Younes M. (2001). IPCS conceptual framework for evaluating a mode of action for chemical carcinogenesis. Regul Toxicol Pharmacol 34:146–152.
- Su T, Bao Z, Zhang QY, Smith TJ, Hong JY, Ding X. (2000). Human cytochrome P450 CYP2A13: Predominant expression in the respiratory tract and its high efficiency metabolic activation of a tobacco-specific carcinogen, 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone. Cancer Res 60:5074–5079.
- Thornton-Manning JR, Dahl AR. (1997). Metabolic capacity of nasal tissue interspecies comparisons of xenobiotic-metabolizing enzymes. Mutat Res 380:43–59.
- United States Environmental Protection Agency (US EPA). (2001). Toxicological Review of Chloroform (CAS No. 67-66-3), In Support of Summary Information on the Integrated Risk Information System (IRIS). October. Available at: http://www.epa.gov/ncea/iris/toxreviews/0025tr.pdf.
- US EPA. (2004). Toxicological review of naphthalene (CAS No. 91-20-3) In Support of Summary Information on the Integrated Risk Information System (IRIS). NCEA-S-1707, June 2004, External review draft. Washington, DC: Office of Research and Development.
- Van Winkle L, Isaac J, Plopper C. (1997). Distribution of epidermal growth factor receptor and ligands during bronchiolar epithelial repair from naphthalene-induced Clara cell injury in the mouse. Am J Pathol 151:443–459.
- Van Winkle LS, Buckpitt AR, Nishio SJ, Isaac JM, Plopper CG. (1995). Cellular responses in naphthalene-induced Clara cell injury and bronchiolar epithelial repair in mice. Am J Physiol 269(6 Pt 1):L800–L818.
- Van Winkle LS, Evans MJ, Brown CD, Willits NH, Pinkerton KE, Plopper CG. (2001). Prior exposure to aged and diluted sidestream cigarette smoke impairs bronchiolar injury and repair. Toxicol Sci 60:152–164.
- Van Winkle LS, Brown CD, Shimizu JA, Gunderson AD, Evans MJ, Plopper CG. (2004). Impaired recovery from naphthalene-induced bronchiolar epithelial injury in mice exposed to aged and diluted sidestream cigarette smoke. Toxicol Lett 154:1–9.
- Waidyanatha S, Rappaport SM. (2008). Hemoglobin and albumin adducts of naphthalene-1,2-oxide, 1,2-naphthoquinone and 1,4-naphthoquinone in Swiss Webster mice. Chem Biol Interact 172:105–114.
- Waidyanatha S, Troester MA, Lindstrom AB, Rappaport SM. (2002). Measurement of hemoglobin and albumin adducts of naphthalene-1,2-oxide, 1,2-naphthoquinone and 1,4-naphthoquinone after administration of naphthalene to F344 rats. Chem Biol Interact 141:189–210.
- West JA, Pakehham G, Morin D, Fleschneer CA, Buckpitt AR, Plopper CG. (2001). Inhaled naphthalene causes dose dependent Clara cell cytotoxicity in mice but not in rats. Toxicol Appl Pharmacol 173:114–119.
- Wilson AS, Davis CD, Williams DP, Buckpitt AR, Pirmohamed M, Park BK. (1996). Characterization of the toxic metabolite(s) of naphthalene. Toxicology 114:233–242.
- Zheng J, Cho M, Jones AD, Hammock BD. (1997). Evidence of quinone metabolites of naphthalene covalently bound to sulfur nucleophiles of proteins of murine Clara cells after exposure to naphthalene. Chem Res Toxicol 10:1008–1014.