Nasal Cytotoxic and Carcinogenic Activities of Systemically Distributed Organic Chemicals

Abstract

Toxicity and carcinogenicity in the mucosa of the nasal passages in rodents has been produced by a variety of organic chemicals which are systemically distributed. In this review, 14 such chemicals or classes were identified that produced rodent nasal cytotoxicity, but not carcinogenicity, and 11 were identified that produced nasal carcinogenicity. Most chemicals that affect the nasal mucosa were either concentrated in that tissue or readily activated there, or both. All chemicals with effects in the nasal mucosa that were DNA-reactive, were also carcinogenic, if adequately tested. None of the rodent nasal cytotoxins has been identified as a human systemic nasal toxin. This may reflect the lesser biotransformation activity of human nasal mucosa compared to rodent and the much lower levels of human exposures. None of the rodent carcinogens lacking DNA reactivity has been identified as a nasal carcinogen or other cancer hazard to humans. Some DNA-reactive rodent carcinogens that affect the nasal mucosa, as well as other tissues, have been associated with cancer at various sites in humans, but not the nasal cavity. Thus, findings in only the rodent nasal mucosa do not necessarily predict either a toxic or carcinogenic hazard to that tissue in humans.

Introduction

The nasal passages of rodents are increasingly recognized as an important target site of action for certain toxins and carcinogens. Up through 1997, excellent reviews of xenobiotic effects in rodent nasal mucosa (NM) were provided by several investigators (Dahl and Hadley, 1991; Brittebo, 1997; Monticello and Morgan, 1997; Schuller, 1997). The need for an understanding of carcinogenic susceptibility factors in rodent and human nasal tissues in order to develop plausible modes of action has been highlighted (Bogdanffy et al., 1997). The purpose of this review is to expand on these earlier reports with new information on this topic directed to evaluation of human risk from systemic exposure to rodent nasal cytotoxins and carcinogens. After acceptance of this manuscript, two other relevant papers were published (Harkema et al., 2006; Genter, 2006). For each chemical identified as producing either nasal cytotoxicity or carcinogenicity, in this review, authoritative sources were accessed and literature searches were conducted to identify human effects.

Nasal Mucosa Anatomy and Physiology

In animals and humans, the elaborate nasal passages are covered by the NM, which is composed of several types of epithelium. The most anterior portion, the nasal vestibule, is covered by stratified squamous epithelium. Posterior to the nasal vestibule, the epithelium is composed of nonciliated cuboidal transitional epithelium. Immediately posterior to the transitional epithelium, the epithelium is ciliated and assumes the structure of pseudostratified respiratory epithelium (RE). Farthest posterior, in the craniodorso-posterior aspect of the nasal cavity, the region designated as the olfactory mucosa (OM), is lined with specialized olfactory epithelium (OE). The complex OE consists of 3 layers of cells: (i) a lower basal cell compartment; (ii) a receptor or sensory cell compartment containing the olfactory (sensory) neurons, their precursor basal cells, together with the cells of Bowman’s glands (BGs) and ducts; and (iii) a supporting (or sustentacular) cell compartment (Harkema, 1990, 1991; Legrier et al., 2001). The subepithelial BGs within the lamina propria of the OE constitute exocrine glands whose excretions are passed through excretory ducts to the surface of the OE to moisten the mucosal surface (Uraih and Maronpot, 1990; Harkema, 1991). The main anatomical difference in the NM across animal species and humans is in the percentage of OE lining the NM surface; the OE occupies 50% of the surface in rodents, but only 10% in humans (Harkema, 1990), reflecting the high reliance of rodents on the sense of olfaction. Conversely, the transitional epithelium occupies only 10% and the RE 35% of the NM surface in rats and 25% and 60% respectively, in humans (Reznik, 1983; Uraih and Maronpot, 1990; Harkema, 1990, 1991; Morgan et al., 1991; Mery et al., 1994; Menco and Morrison, 2003). As described here, the OE is a site of toxicity of many xenobiotics.

The receptor neurons of the OE are bipolar cells whose dendrites extend to the epithelial surface and whose axons connect to the olfactory bulb. The cilia of the olfactory neurons form dendritic knobs that contain odorant receptors (Zhang and Firestein, 2002). The olfactory neurons have the unusual ability for neurons to periodically replace themselves (Menco, 1983; Harkema, 1990; Reznik, 1990; Uraih and Maronpot, 1990; Menco and Morrison, 2003). At birth the basal cell progenitors of the olfactory neurons in mice exhibit high mitotic activity that diminishes significantly by 3 months after birth and remains at a lower level throughout life (Legrier et al., 2001). In the rat, the life span of olfactory neurons is approximately 20–28 days (Uraih and Maronpot, 1990; Legrier et al., 2001), while in the mouse, the life span of some neurons is 1 year (Hinds et al., 1984).

The entire OE has a high level of cell turnover (Moulton, 1974) with OE cells being renewed by their progenitors, the globose basal cells (Moulton, 1974; Schwartz-Levey et al., 1991; Caggiano et al., 1994; Huard and Schwob, 1995). The rate of replacement of the OE has been suggested to reflect injury by airborne toxins (Farbman, 1990; Calderon-Garciduenas et al., 1998).

The rodent NM possesses substantial chemical biotrans-formation capability (Hadley and Dahl, 1982; Dahl, 1985; Longo et al., 1988; Dahl and Hadley, 1991; Bereziat et al., 1995; Genter et al., 1995b; Thornton-Manning and Dahl, 1997; Genter, 2004), which may play a role in olfaction (Anholt, 1989; Nef et al., 1989; Carr et al., 1990). Species differ in their specific levels of biotransformation capability (Hadley and Dahl, 1983; Thornton-Manning and Dahl, 1997). Relatively little is known about NM biotransformation activity in humans, but biotransformation enzymes are definitely present (Gervasi et al., 1991; Lewis et al., 1994; Yokose et al., 1999; Zhang et al., 2005). Comparisons of rodent and human nasal tissues generally reveal the former to have higher biotransformation activities (Feng et al., 1990; Bogdanffy et al., 1998; Heydens et al., 1999). In the rat NM, the ethmoid turbinates, which are a frequent site for tumor formation, have very high chemical biotransformation activity compared to other species (Hadley and Dahl, 1982; Hadley and Dahl, 1983; Dahl, 1985; Longo et al., 1988; Sabourin et al., 1988).

In particular, in rodents, the phase I oxidation enzymes, the cytochromes P450 (CYPs) are present at high levels in the NM, being mainly expressed in the RE and OE (Brittebo, 1997; Thornton-Manning and Dahl, 1997; Ling et al., 2004). Indeed, the most highly expressed genes in the NM are those of Phase I and II biotransformation enzymes (Hester et al., 2002). The NM CYPs include 1A1, 2A, 2C, 2E1, 2G1, 3A, and 4B (Brittebo, 1997; Thornton-Manning and Dahl, 1997; Gu et al., 1999; Wang et al., 2002). The CYP2A subfamily enzymes are involved in the biotransformation of many xenobiotics. In the rat, CYP2A3 is expressed in the NM (Robottom-Ferreira et al., 2003), whereas in mouse NM CYP2A5 is present (Piras et al., 2003) and CYP2A6 in humans (Liu et al., 1996). Among the CYPs, CYP2G1 activity (Nef et al., 1990; Hua et al., 1997) or transcripts (Yu et al., 2005) are expressed in the OE of mice, rats, and rabbits (Ling et al., 2004), but in humans the gene has multiple mutations and there is no functional enzyme activity (Sheng et al., 2000). In one study, human nasal tissue had only 0.15% of the CYP (7-ethoxycoumarin-O-deethylase) activity of that of the rat (Feng et al., 1990). Another family of oxidation enzymes, flavin-containing monooxygenases, is present at higher levels in the rat OM than in the liver (Genter and Ali, 1998).

Phase II conjugating enzymes such as glutathione-S-transferase and glucuronyl transferase are also present in the NM (Thornton-Manning and Dahl, 1997). N-acetyltransferase transcripts, specifically the N-acetyltransferase 1 enzyme, which is widely distributed in the body, is present in the RE and OE, as well as the BGs (Debiec-Rychter et al., 1996). In fact, activity in the OM of Long–Evans rats is markedly higher than in the liver (Genter, 2004). Other enzymes, such as epoxide hydrolase, are higher in rats when compared to other species such as mice (Green et al., 2001). The activities of some of the epoxide hydrolases are similar to those of liver with respect to polycyclic aromatic hydrocarbon biotransformation (Bond, 1983). These hydrolases often act to detoxify reactive arene oxides, but in other instances they can be involved in multistep activation of some aromatic hydrocarbons (Levin et al., 1980). In contrast to the situation with epoxide hydrolase, carboxylesterase activity is reported to be slightly higher in mice than in rats (Stott and McKenna, 1985). In both species, activity was greater in the OE than in the RE (Bogdanffy et al., 1987; Frederick et al., 1994), being present in the OE in sustentacular cells and BGs (Robinson et al., 2002). Considering the proficiency of the rodent NM in chemical biotransformation, this tissue might reasonably be regarded as a veritable second liver. Moreover, like the liver, it is positioned at a portal of entry for xenobiotics. Several of the enzyme systems present in the NM are known to be inducible (Bond, 1983; Gillner et al., 1987; Longo and Ingelman-Sundberg, 1993; Bereziat et al., 1995; Nikula et al., 1995; Thornton-Manning and Dahl, 1997). However, the hepatic CYP2A5 inducers pyrazole and phenobarbital did not affect message or protein expression in mouse OM (Piras et al., 2003) and 3-methylcholanthrene and pyrazole did not induce CYP2A3 in rat NM (Robottom-Ferreira et al., 2003).

Antioxidant enzymes, including superoxide dismutase, catalase, glutathione peroxidase and DT-diaphorase were reported to be at higher levels in rat NM than in the lung of rats (Reed et al., 2003). Enzymes were higher in the RE than OE.

The NM also possesses cellular protective systems, including the antioxidants ascorbate and α-tocopherol (Reed et al., 2003) and a thiol-specific protein (Peshenko et al., 1998). Heat shock protein 70 (Genter and Ali, 1998; Simpson et al., 2005) is also expressed. These protective systems would seem to reflect the situation of the NM as a tissue with exposure to airborne environmental chemicals.

As a consequence of species differences in enzymatic activities, the effect of a specific chemical in the NM may differ between species depending upon the specific pathway involved in the biotransformation of the compound. Where such information was found for a specific chemical, it was included in this review. However, broad generalizations about susceptibility of NM to chemical toxicity are not possible.

Uptake of chemicals into systemic circulation through the NM is well recognized, and nasal administration is used as a delivery route for several drugs (Hussain, 1998). Additionally, transport of chemicals from the OE to the olfactory bulb takes place along neuronal axons (Shipley, 1985; Hastings and Evans, 1991; Tjälve et al., 1996).

Nasal Mucosa Pathology

In spite of the dynamic condition of the NM, spontaneous nasal tumors are rare in rodents. In the Fischer 344 rat, which is used in National Toxicology Program (NTP) bioassays, the spontaneous nasal tumor incidence is usually <0.5% (Rao et al., 1990; Haseman and Clark, 1990; Haseman and Hailey, 1997) and similar to that seen in the Wistar rat (Feron et al., 1990). B6C3F1 mice, as reported in the NTP database, have an even lower incidence (Haseman and Elwell, 1996). The majority of spontaneous nasal tumors are squamous cell carcinomas (Feron et al., 1990; Haseman and Clark, 1990; Rao et al., 1990; Haseman and Elwell, 1996; Haseman and Hailey, 1997), which arise in the nares. The higher background of NM tumors in the rat implicates a genetic susceptibility, which may result from a variety of causes including differences in biotransformation activities or cell proliferation, and probably contributes to the greater inducibility of such tumors in the rat.

A variety of chemicals, administered orally or perenterally, has been found to produce nasal cytotoxicity (Table 1) and some to induce tumors of the rodent nasal cavity (Table 2). Chemicals that have carcinogenic activity to the anterior squamous epithelium and are active upon direct contact with the NM when administered by inhalation (Feron et al., 2001), such as formaldehyde, are not detailed in this review, although inhalation is an important route of human exposure and many volatile carcinogens act directly on the NM. Also, some chemicals that are cytotoxic to the NM by the inhalation route produce similar effects when delivered systemically (Keller et al., 1997; Lee et al., 2005). Nevertheless, here we focus on organic chemicals that have cytotoxic and carcinogenic activity in the NM upon systemic distribution following oral or perenteral administration.

Many, but not all, such nasal cytotoxins have carcinogenic activity. Among chemicals with carcinogenic activity for the RE and OE, some are clearly DNA reactive, whereas others have not been shown to bind to DNA, but were reported to bind to protein and produce cytotoxicity. Cytotoxicity and reparative compensatory cell proliferation may be important contributors to carcinogenic activity in the NM, even with DNA-reactive agents, as shown for a variety of tissues (Butterworth et al., 1992; Cunningham, 2000; Williams et al., 2000).

Many of the organic chemicals that are cytotoxic or have carcinogenic activity in the NM with systemic distribution are single ring compounds or liberate single ring components following biotransformation. Most such chemicals produce greater effects in the NM of the rat than in the mouse and males are often more susceptible than females. The organ specificity of these carcinogens appears in part to result from high and sometimes specific biotransformation within the RE and OE (Hadley and Dahl, 1983; Dahl and Hadley, 1991). The characteristics of these agents are reviewed herein and the relevance of their effects in experimental animals to human cancer risk assessment is discussed.

Rodent Nasal Cytotoxins

At least 14 chemicals or classes of compounds reviewed in this section have produced nasal cytotoxicity, but either have not produced nasal tumors or have not been tested for systemic carcinogenicity but are deemed unlikely to be nasal carcinogens by that route.

Acetaminophen (APAP) or paracetamol (Figure 1) is a widely used analgesic and is a biotransformant of phenacetin (Figure 30), which is carcinogenic to the NM (see below). With oral doses of 600 mg/kg bw ip, APAP was reported to be toxic to the NM of mice (Hart et al., 1995; Genter et al., 1998). The toxicity in the mouse NM is not dependent upon hepatic biotransformation (Gu et al., 2004), but rather involves local bioactivation to a N-acetyl-p-benzoquinoneimine (Miner and Kissinger, 1979), the chemically reactive species which binds to protein (Muldrew et al., 2002) (Figure 1), mediated by CYPs 2E1, 2A5 and 2G1 (Hinson, 1983; Gu et al.,1998; Chen et al., 1998). APAP was not genotoxic in most systems (International Agency for Research on Cancer, 1999c). Whereas APAP was reported to bind to mouse liver and kidney DNA (Rogers et al., 1997), no binding was reported in rat liver (Hasegawa et al., 1988) or rat liver, colon or bladder (Williams et al., 2002). In oral carcinogenicity studies of APAP in mice and rats at doses up to 6000 ppm in the diet, equivocal effects in female rats were reported for some tissues as assessed by the International Agency for Research on Cancer (International Agency for Research on Cancer,1999c), but no nasal tumors were observed (Hiraga and Fujii, 1985; National Toxicology Program, 1993a). In a review of about a dozen epidemiological studies, APAP use was not noted to be associated with any nasal toxicity and was not considered to be carcinogenic in humans (International Agency for Research on Cancer, 1999c). Recently, however, heavy use has been reported to be associated with a slight increase in renal cancer (Kaye et al., 2001).

Antimicrotubule Agents

Vincristine sulfate, vinblastine sulfate, vindesine sulfate (Figure 2), and paclitaxel (formally called taxol) (Figure 3) at doses of ~2–12 mg/kg bw administered intravenously produced apoptosis in the OE of BALB/c mice by 24 hours (Kai et al.,2002, 2004, 2005). This progressed to atrophy by 15 days. Whole-body radio-luminography revealed concentrations of vincristine sulfate in nasal tissues about 2-fold higher than that in blood (Kai et al., 2005).

In limited studies, evidence was not found for carcinogenicity of vinblastine sulfate (International Agency for Research on Cancer, 1987h) or vincristine sulfate (International Agency for Research on Cancer, 1987i). In spite of use as anticancer medicines, no information was found on human nasal cytotoxicity or carcinogenicity.

Aryl methyl sulfones (Figure 4) are biotransformants formed from chloro aryl hydrocarbons derived from compounds such as 1,1,1-trichloro-2,2-bis(p-chlorophenyl)ethane (DDT) and polychlorinated biphenyls (PCBs) following conjugation with glutathione (GSH) and degradation via mercapturic acid derivatives. 2,6-dichlorophenyl methylsulfone, as a single ip dose, produced necrosis preferentially in the BG and neuroepithelium in the dorsomedial olfactory region (Franzén et al., 2003), probably mediated by CYP2A5 activation, one of the prominent CYPs in the mouse (Zhuo et al., 2004; Franzén et al., 2006). By contrast, only minor damage occurred at this site in rats dosed with the 2,5-chlorinated isomer.

Assessments of nasal cytotoxicity or carcinogenicity to humans have not been reported.

Benzophenone (Figure 5) occurs naturally and is used the manufacture of a wide variety of industrial chemicals. In a 14-week study, mice receiving up to 20,000 ppm benzophenone in the diet exhibited no changes in the NM. With administration for 2 years of doses of 1200 ppm in the diet, male and female mice developed metaplasia of the OE, but rats were not affected (National Toxicology Program, 2006).

In these carcinogenicity studies, benzophenone produced increases in male mouse liver neoplasms and male rat kidney neoplasms, likely through epigenetic mechanisms, but tumors were not produced in the NM. Assessments of nasal cytotoxicity or carcinogenicity to humans were not found (Hazardous Substances Data Bank, 2006).

Bromobenzene (Figure 6) is used in organic synthesis, especially to make phenyl magnesium bromide; as an additive to motor oils; as a solvent, especially for crystallizations on a large scale and where a heavy liquid is desirable. It elicited nasal toxicity when administered ip at > 4.8 mmol/kg body weight and showed high levels of tissue binding particularly to the BGs (Brittebo et al., 1990). Human exposures have occurred in occupational situations and at low levels (normally <1 ppb) in drinking water (Environmental Working Group and EWG Action Fund, 2006). The U.S. Environmental Protection Agency estimates that exposures of children to 4000 ppb for 10 days is not expected to cause any adverse, noncarcinogenic effects. However, in February, 2005, the U.S. Environmental Protection Agency included bromobenzene in the Drinking Water Contaminant Candidate List 2 (US EPA, 2005). While bromobenzene is highly toxic, there is no evidence for carcinogenicity to the NM or to any other site (National Research Council, 2006).

Chloroform (CHCl₃) (Figure 7) has been used as a solvent and is generated as a trace contaminant in the chlorination of drinking water with a maximum allowable concentration as total trihalomethanes of 80 ppb. Oral administration to female F344 rats of 34 mg/kg/day or greater produced degeneration of the OE and superficial BGs (Larson et al., 1995), which was not associated with detectable olfaction deficit (Dorman et al., 1997). The OE, RE, and BGs have substantial chloroform biotransformation capabilities (Lofberg and Tjalve, 1986; Constan et al., 1999), leading to generation of phosgene, which in turn can react with proteins (Fabrizi et al., 2003) (Figure 7). Chloroform administration produced liver and kidney tumors in mice and kidney tumors in rats, but no nasal tumors (International Agency for Research on Cancer, 1999a). Also, chloroform did not produce nasal tumors in sensitive TP53^+/− mice (Storer et al., 2001). IARC (International Agency for Research on Cancer, 1999a) concluded that in humans the evidence for carcinogenicity of chloroform was inadequate. No indication of any effect on the nasal passages was found (Hazardous Substances Data Bank, 2006).

Coumarin (Figure 8) is a naturally occurring fragrant compound that has been used in consumer products and as a food additive. A single ip injection of 50 mg/kg bw to Wistar rats and C57BL/6 mice produced reductions at 48 hours in CYP2A and CYP2G, the major CYPs in mice (Zhuo et al., 2004), in the OE, but not in the liver (Gu et al., 1997). By 12 hours, necrosis of the OE and underlying BG was evident and this progressed up to 48 hours. The principal biotransformants resulted from 3- or 7-hydroxylation or formation of a transient 3,4-epoxide which, rather than rearranging to form 3-hydroxycoumarin, yields o-hydroxyphenylacetaldehyde, which can undergo further oxidation to the corresponding acetic acid derivative (Born et al., 2002) (Figure 8). In the mouse, OM-specific CYP2G1 is the major enzyme responsible for biotransformation, producing 7-hydroxycoumarin and o-hydroxyphenylacetaldehyde as the predominant products (Zhuo et al., 1999). Other CYPs were also active.

Although there is some evidence for induction of various neoplasms in rodents by coumarin, no tumoriginicity has been reported in the nasal cavity (National Toxicology Program, 1993b). Coumarin has not been reported to be associated with nasal toxicity or carcinogenicity to humans.

Dichlorobenzenes (substituted). 2,6-Dichlorobenzonitrile (DCBN) (dichlorobenil) and 2,6-dichlorothiobenzamide (chlorthiamid) are dichlorobenzene (DCB) derivatives (Figure 9), which have been used as herbicides.

In mice, single ip injections of DCBN (12, 25, 50 mg/kg) produced necrosis in the dorsomedial OE (Brandt et al., 1990). DCBN (25 mg/kg) and chlorthiamid (12 mg/kg) were toxic to the OE of mice with oral administration (Eriksson and Brittebo, 1995b; Mancuso et al., 1997). The injury produced by acute doses of DCBN persisted for up to 6 months. DCBN bound specifically to the epithelium of mouse BGs (Eriksson and Brittebo, 1995a). Subsequently, methylsulfonyl-2,5-DCB and methylsulfonyl-2,6-DCB were reported to be localized in the OM of female NMRI mice, but only the latter was toxic to the OE (Bahrami et al., 1999). The parent compounds, 1,3- and 1,4-DCBs, and 1,2,3-trichlorobenzene, were nontoxic to mouse OM (Bahrami et al., 1999). However, 1,4-DCB produced epithelial necrosis of the nasal turbinates in male and female rats when given by gavage for 13 weeks at 1,200 and 1,500 mg/kg (National Toxicology Program, 1987). In a study of a wide range of 2,6-DCBs, 2,6-dichloronitrobenzene and dichlorobenzaldehyde were identified as having OM toxicity, but a simple explanation of nasal toxicity based on structure activity relationships could not be identified (Carlsson et al., 2004).

CYP activity of the NM underlies the toxicities of these compounds, mediating selective covalent binding of the toxicants to the NM, especially the dorsal medial meatus (Brittebo, 1997). CYP 2A5 and CYP 2G1 activate DCBN to a reactive intermediate (Brandt et al., 1990), which may be a 2,3-arene oxide (Genter et al., 1995b; Ding et al., 1996). The region of injury lacks microsomal epoxide hydroxylase (Genter et al., 1995b), which suggests an inability to detoxify an epoxide.

DCBN induced liver tumors in rats and several types of tumors in mice, including liver tumors, but there was no indication of nasal tumors (Cox, 1997). 1,4-DCB produced kidney tumors in rats, but no nasal tumors (National Toxicology Program, 1987).

None of these DCBs has been reported to be associated with nasal toxicity or carcinogenicity in humans.

Aliphatic Nitriles

A variety of synthetic organonitriles are in industrial use, including DCBN discussed previously. Aliphatic nitriles occur in plants, for example, β-aminopropionitrile, which is the toxic compound in the sweet pea (Lathyrus odoratus). Several nitriles are neurotoxic, including allylnitrile and cis-crotononitrile and one, acrylonitrile, produced brain tumors in rats β,β′-iminodipropionitrile (IDPN) (Figure 10) is a synthetic saturated alkyl nitrile. It is neurotoxic in humans and animals and produced nasal toxicity in Long–Evans rats after either a single ip dose of 200 mg/kg or 3 consecutive daily doses and sacrifices at up to 56 days (Genter et al., 1992). IDPN-induced OE degeneration occurred in regions expressing CYP2E1, indicating that toxicity resulted from bioactivation (Genter et al., 1994). IDPN is reported to liberate cyanide (Froines et al., 1985; Dahl and Waruszewski, 1989), which could be a toxicant. The N-hydroxylated form of IDPN (Figure 10) also produced olfactory mucosal degeneration in SD rats at >100 mg/kg bw, suggesting that it may be an in vivo biotransformant (Crofton et al., 1996).

The OM injury produced by IDPN results in degeneration of nerve terminals in the glomeruli of the olfactory bulb (Boadas-Vaello et al., 2005). Two other nitriles, allylnitrile and cis-crotononitrile, when given orally to rats on 3 consecutive days produced degradation in the olfactory bulbs indicative of injury to the OM (Boadas-Vaello et al., 2005). Reports of experimental carcinogenicity of IDPN or information on human nasal cytotoxicity or carcinogenicity were not found.

Methylacrylonitrile (Figure 10) is an unsaturated alkyl nitrile that is widely used in the preparation of homopolymers and copolymers. When administered by gavage to F344 rats, the dose of 30 mg/kg bw given for 2 years, produced atrophy and metaplasia of the OE in bothmales and females, but no neoplasms (Nyska and Ghanayem, 2003). No information was found on human nasal cytotoxicity or carcinogenicity (Hazardous Substances Data Bank, 2006).

Methimazole (1-methylimidazole-2-thiol) (Figure 11) is a nitrogen heterocyclic medicine that has been used for the treatment of hyperthyroidism. In Long–Evans rats, administration of 25 mg/kg by ip injection or 50 mg/kg by intragastric instillation as single doses with animals sacrificed 32 hours later, caused almost complete destruction of the OE (Genter et al., 1995a). Methimazole was also toxic to the OE in mice (Bergman and Brittebo, 1999); in NMRI mice, 2 doses of 50 mg given by ip injection 3 days apart produced damage to the OE and BGs (Bergman et al., 2002). The damage was rapidly repaired with only minor changes 3 months later (Bergman et al., 2002). Methimazole showed selective covalent binding in the BGs as well as the bronchial epithelium in the lungs and centrilobular regions of the liver following an intravenous (iv) injection in mice (Bergman and Brittebo, 1999). The related chemicals, 2-methylimidazole and 4-methylimidazole, did not produce nasal toxicity but induced thyroid lesions (Chan et al., 2006), as does methimazole. Methimazole also produced a transient loss of the sense of smell in rats (Genter et al.,1996; Xu and Slotnick, 1999), as does carbimazole (Genter, 1998), a carbethoxy derivative of methimazole, which is converted in the body to methimiazole (Wishart et al., 2006).

Pretreatment of mice with the CYP inhibitor metyrapone completely abolished covalent binding of methimazole to the OE and bulb (Bergman and Brittebo, 1999), indicating a requirement for biotransformation. Methimazole is probably bioactivated to an S-oxide (Genter et al., 1995a) by a flavin-containing monooxygenase, an enzyme abundant in the NM (Genter and Ali, 1998). Methimazole produced rapid depletion of nonprotein sulfhydral groups, mainly GSH, in the OE of mice (Bergstrom et al., 2003).

Methimazole is a nitrogen heterocyclic, but it seems from analog studies that oxidation of the exocyclic thiol is critical for nasal toxicity (Genter et al., 1995a), although this may not provide a complete explanation and ring epoxidation and subsequent cleavage may also play a role (Mizutani et al., 2000).

In a carcinogenicity study in rats, neoplasms were increased only in the thyroid gland (Lilly, 1996), while a study in mice yielded no increase in neoplasia (Jemec, 1970). The development of the thyroid tumors in rats is likely as a result of the goitrogenic effects of the drug (Capen, 1994). No cytotoxicity to human NM has been noted (Bartalena et al., 1996), although loss of the sense of smell has been reported (Schiffman and Gatlin, 1993), as also with exposure to carbimazole (Erikssen et al., 1975; Neundorfer, 1987), similarly to rats (Genter et al., 1996; Genter, 1998). Detailed studies of possible OM degeneration in humans have not been undertaken. No information was found on human carcinogenicity.

3-Methylindole (3-MI) (Figure 12) is produced from tryptophan by fermentation in the intestinal tract (Wiethoff et al., 2001) and is present in cigarette smoke (Hoffmann and Rathkamp, 1970). Intraperitoneal injection (ip) of 400 mg/kg to C57BL mice produced OE necrosis (Turk et al., 1986, 1987), which resulted in olfaction deficits (Peele et al., 1991; Miller and O’Bryan, 2003). The BGs and sustentacular cells were predominantly affected (Miller and O’Bryan, 2003). 3-MI is biotransformed by lung CYPs to form free radicals (Bray and Kubow, 1985).

Bioactivated 3-MI formed adducts with isolated or cellular DNA (Regal et al., 2001). Its reactive intermediate, 3-methyleneindolenine (Figure 12) was generated by CYP2F (Wang et al., 1998) and bound to DNA, whereas an epoxide may be involved in reaction with softer nucleophiles such as glutathione and proteins (Skordos et al., 1998; Lanza et al., 1999). Administration of 3-MI to rats and rabbits produced protein thioether adducts in lung, kidney, and liver, as demonstrated by an enzyme-linked immunosorbent assay (Kaster and Yost, 1997), but NM was not studied. The specific contribution of either or both of the above reactive products to nasal toxicity in mice is not resolved. No report of 3-MI carcinogenicity testing was found and nasal cytotoxicity or carcinogenicity related to 3-MI exposure in humans has not been reported.

Naphthalene (Figure 13), the next higher homolog of benzene (Figure 19), is a major constituent of coal tar and creosote, and is formed in most incomplete combustion processes. Its main use is in the production of phthalic anhydride, while its use as a moth repellant is declining. Naphthalene is quite volatile and most human exposures probably occur via inhalation. When administered ip at 200 mg/kg bw to SD rats, it produced severe injury throughout the olfactory region (Lee et al., 2005). Naphthalene was tested systemically for carcinogenicity by oral administration in one study in rats, by intraperitoneal administration in newborn mice and in rats and by other routes. All these studies were considered too limited for an evaluation of the experimental carcinogenicity of naphthalene by IARC (International Agency for Research on Cancer, 2002), which considered it a possible human carcinogen. Subsequently, induction of olfactory neuroblastomas and respiratory epithelial adenomas by inhalation exposure was reported (Long et al., 2003). The lack of clear carcinogenicity to the NM by oral or perenteral routes may reflect systemic detoxification. No report of nasal cytotoxicity or carcinogenicity in humans was found (Hazardous Substances Data Bank, 2006).

Polychlorinated biphenyls (PCBs) (Figure 14) Among dioxin-like compounds tested for carcinogenicity, the PCB binary mixtures of either 3,3′,4,4′,5-pentachlorobiphenyl (PCB 126) and 2,2′,4,4′,5,5′-hexachlorobiphenyl (PCB 153) or PCB 126 and 2,2′,4,4′,5,5′-hexachlorobiphenyl (PCB 118) produced nasal lesions in female SD rats (Nyska et al., 2005). Inflamation occurred with PCB 126 and PCB 153 at doses of 300 μg/kg bw, while RE hyperplasia occurred only at the highest dose of 1000 μg/kg bw. Although PCBs induce liver tumors in rodents, induction of nasal neoplasms has not been reported (International Agency for Research on Cancer, 1987f).

Up to 1987, no report of nasal cytotoxicity or carcinogenicity related to PCB exposure in humans was available (International Agency for Research on Cancer, 1987f) and no subsequent reports were found.

3-Trifluoromethylpyridine (Figure 15) is a starting material for the synthesis of herbicides. With inhalation, it produced nasal toxicity in rats (Gaskell et al., 1990), and, when given orally, accumulated in the ethmoid turbinates and dorsal meatus of the nasal passages and bound to proteins (Hext and Lock, 1992), suggesting that it would be cytotoxic with systemic distribution. The structurally related aminodichloropyridine is cytotoxic and carcinogenic to the NM with systemic distribution (see below), but no report of 3-trifluoromethylpyridine carcinogenicity testing or human toxicity was found.

Overview of Cytotoxicity

Of the 14 compounds or classes of compounds discussed here and summarized in Table 1, which produced nasal cytotoxicity with oral or parenteral administration, none produced nasal neoplasms, although several increased tumor incidence at other sites, namely benzophenone, chloroform, coumarin, DCBN, methylindole and PCB. However, the conditions for induction of nasal tumors may not have been met, and a detailed search for nasal tumors may not always have been performed, especially in earlier studies. In fact, it is possible that with appropriate systemic testing, naphthalene might be a nasal carcinogen. In humans, no nasal cytotoxicity or carcinogenicity was found.

Mechanisms of Rodent Nasal Toxicity

The principal compounds discussed here produce nasal cytotoxicity in rodents, which is attributable to biotransformation in the NM. Most have been tested for carcinogenicity, but none has been reported to produce nasal tumors, although, as noted above, adequate testing was not always undertaken or available. Nevertheless, the absence of nasal tumors associated with the cytotoxicity induced by these chemicals is similar to the findings in inhalation studies conducted by the NTP; in their database, noncarcinogens produced inflammatory and proliferative lesions similar to those elicited by nasal carcinogens (Ward et al., 1993). This indicates that noncarcinogenic nasal toxicants lack some property of compounds that have carcinogenic activity in the NM (see below). Notably, none of these cytotoxins has been shown to bind to DNA of the NM, although, again, specific studies were not always undertaken or available.

Where appropriate investigations were conducted, many of the chemicals bound selectively to the NM, which, no doubt, underlies their nasal cytotoxicity. Perhaps, gene expression studies (Hester et al., 2002, 2005) would shed some light on tissue responses that may differ between agents lacking nasal carcinogenicity and those that are carcinogenic to the NM (discussed below).

Human Effects of Rodent Nasal Cytotoxins

None of the chemicals reviewed above which elicit rodent nasal cytotoxicity following oral or perenteral administration has been associated with a similar toxicity or carcinogenicity in humans, although human exposures have occurred, albeit at much lower levels either orally or by inhalation. The smaller proportion of OM in humans and lesser biotransformation capacity undoubtedly also contribute to human insensitivity.

Rodent Nasal Carcinogens

The principal chemicals reviewed here all have exerted some degree of carcinogenic activity, as well as cytotoxicity, in the rodent NM under certain conditions of oral or perenteral administration.

Aniline Derivatives

Among a variety carcinogenic single ring aromatic amines a few have been identified as nasal carcinogens, including phenacetin which is discussed separately next. p-Cresidine (2-methoxy-5-methylaniline, Figure 16) is a substituted aniline used mainly in the dye industry. It has been found as a trace contaminant in commercial batches of FD&C Red. No. 40. When administered in the diet at 0.5 and 1%, p-cresidine induced olfactory neuroblastomas in both genders of F344 rats, as well as squamous cell and transitional cell carcinomas of the urinary bladder (National Toxicology Program, 1979b; Resnik et al., 1981;Mortensen et al., 2002).

p-Cresidine was also carcinogenic in B6C3F1 mice, causing carcinomas of the urinary bladder in both genders and hepatocellular carcinomas in females, but not nasal tumors (National Toxicology Program, 1979b). p-Cresidine is used as the positive control in the TP53^+/− transgenic mouse bioassay in which it produces bladder neoplasms in both genders and hepatocellular carcinomas in females, but no nasal tumors (Storer et al., 2001), although NM atrophy, necrosis, and degeneration have been reported in DNA repair-deficient C57BL/6XPA^−/− mice (Mortensen et al., 2002). These findings in the mouse appear to indicate that mouse NM does not bioactivate p-cresidine. This could be examined by assessment of DNA adduct formation. p-Cresidine is only weakly active or negative in most genotoxicity assays (Ashby et al., 1991), but did produce DNA damage in mouse bladder mucosa measured by the single cell gel electrophoresis (Comet) assay (Sasaki et al., 1998).

No epidemiologic report was available from the IARC review of p-cresidine (International Agency for Research on Cancer, 1982b) and none was found in a literature search. Nevertheless, p-cresidine was listed as reasonably anticipated to be a human carcinogen by NTP (National Toxicology Program, 2005).

2,6-Dimethylaniline (2,6-DMA, Figure 17), or 2,6-xylidine, is used mainly in dye manufacture. It is also present in tobacco smoke (Bryant et al., 1988). It is a biotransformant of lidocaine (Figure 18; (Keenaghan and Boyes, 1972). 2,6-DMA induced nasal tumors in rats (National Toxicology Program, 1990b) in a complex study in which 5-week-old CD rats were given 2,6-DMA in the diet at 0, 300, 1000, or 3000 ppm and at 16 weeks they were mated and the females continued on the diets during pregnancy and lactation. The offspring, after weaning at 3 weeks, were then continued on the same diets as their parents for 104 weeks. Nasal adenomas and carcinomas were seen in both male (1000 and 3000 ppm) and female (3000 ppm only) progeny, along with some other tumors. No findings were reported on the parental generation. No carcinogenicity study in mice has been reported. Since the compound is volatile and loss from the feed occurred, there was discussion in the report (National Toxicology Program, 1990b) that the neoplasms may have been a result of inhalation exposure. This issue has not been resolved, but subsequent findings support the likelihood of the nasal effects being due to systemic distribution. For example, 2,6-DMA was reported to have ‘promoting’ activity in the OM when given in the diet at 3000 ppm for 52 weeks after N-bis(2-hydroxypropyl)nitrosamine (Koujitani et al., 1999). Rather than representing promoting activity, however, this effect may be due to enhancement of carcinogenicity by toxicity or a syncarcinogenic effect (Williams and Iatropoulos, 2001) resulting from summation of the genotoxicity of the 2,6-DMA together with that of the nitrosamine, as reported for other combinations of DNA-reactive carcinogens (Williams and Furuya, 1984).

The nasal carcinogenic activity of 2,6-DMA is highly structure-dependent, in as much as the 2,4- and 2,5-DMAs, although carcinogenic in rats and mice, were not reported to produce nasal neoplasms (Weisburger et al., 1978). Studies on the genotoxicity of 2,6-DMA, reviewed by IARC (International Agency for Research on Cancer, 1993a), showed it to be weakly mutagenic in vitro but not in vivo. Nevertheless, covalent binding to DNA of 2,6-DMA in rat NM has been demonstrated using radiolabeled compound (Short et al., 1989) and by nucleotide postlabeling (Gonçalves et al., 2001; Jeffrey et al., 2002).

2,6-DMA is likely bioactivated through N-hydroxylation (Figure 17), as with other aromatic amines (Miller, 1998). Synthetically prepared N-hydroxy-2,6-DMA was highly mutagenic to S. typhimurium and bound to DNA in vitro without bioactivation (Gonçalves et al., 2001; Jeffrey et al., 2002).

2,6-DMA is a biotransformant of lidocaine (Figure 18), a widely used local anesthetic, which is also used to treat arrhythmias, and is used as a veterinary drug. Following the administration of lidocaine, 2,6-DMA has been identified in human tissues (Keenaghan and Boyes, 1972; Parker et al., 1996), and urine (Nelson et al., 1977). Also, hemoglobin adducts have been identified by GC/MS in humans receiving lidocaine (Bryant et al., 1994). 2,6-DMA has also been found in milk from cows and humans treated with lidocaine (Puente and Josephy, 2001). Following iv administration of lidocaine to patients, methemoglobinemia has been observed (Weiss et al., 1987), which is consistent with the formation of an N-hydroxy biotransformant (Kiese, 1966).

No study on the carcinogenicity of lidocaine is described in the summary of preclinical studies in the PDR (2006), although biotransformants, presumably 2,6-DMA, are described as carcinogenic in laboratory animals (PDR, 2006).

No report of an association of nasal tumors in humans has been made for exposure to either 2,6-DMA or lidocaine. This is in spite of theoretical concerns raised by the fact that lidocaine is used as a nasal spray for the treatment of migraine headaches, as well as a local anesthetic during nasal surgery (Genter, 2004). Smokers exhibit NM alterations and have increased risk of sinonasal squamous cell carcinoma (Feron et al., 2001). In cigarette smokers, in addition to the well established increases in 4-aminobiphenyl-hemoglobin adducts, 2,6-DMA-hemoglobin adducts have been reported to be increased (Bryant et al., 1988), although, curiously, adducts were 3 times higher in nonsmokers.

2,6-Diethylaniline and 2,4,6-trimethylaniline (mesidine), are structurally similar to 2,6-DMA There are few studies on 2,6-diethylaniline despite the fact that it is a key biotransformant of the chloroacetanilide herbicides (Feng et al., 1990), discussed next, which are carcinogenic to the nasal cavity. 2,4,6-trimethylaniline produced mouse liver DNA strand breaks in the single cell gel electrophoresis assay (Przybojewska, 1999), but was not mutagenic and evidence for its carcinogenic activity was considered not evaluable by IARC (International Agency for Research on Cancer, 1982a). A structural analog, 2,4,5-trimethylaniline, has limited evidence for carcinogenicity in rodents, and was weakly mutagenic to Salmonella with bioactivation (Kugler-Steigmeier et al., 1989). In contrast to the weak mutagenicity in bacteria, 2,4,5-trimethylaniline was quite mutagenic in the Drosophila wing spot assay and to cultured fibroblasts (Kugler-Steigmeier et al., 1989). o-Anisidine (2-methoxyaniline), which lacks the 5-methyl group of p-cresidine, was carcinogenic in rodents and positive for bacterial mutagenicity (International Agency for Research on Cancer, 1982c), although negative for DNA adduct formation (Ashby et al., 1994). p-Anisidine (4-methoxy aniline) was not adequately tested in rats and was negative for bacterial mutagenicity (International Agency for Research on Cancer, 1982c).

Aniline, the simplest aromatic amine, and a variety of its derivatives are in use in industry. It had carcinogenic activity in rats, producing mainly splenic sarcomas, whereas it was not carcinogenic in mice (International Agency for Research on Cancer, 1987a). This probably reflects the greater susceptibility of rats to aromatic amine-induced methemoglobinemia (Kiese, 1966), which leads to splenic congestion and necrosis due to the culling of abnormal erythrocytes. In in vitro genotoxicity assays, aniline was negative (Williams et al., 1989; Przybojewska, 1999). Several related monocyclic aromatic amines have also been tested. p-Chloroaniline has shown genotoxic activity (Williams et al., 1989) and was carcinogenic in both rats and mice (International Agency for Research on Cancer, 1993b), although not in the nasal cavity (Chhabra et al., 1991).

Thirty-seven aniline derivatives have been tested in the hepatocyte DNA repair assay (Yoshimi et al., 1988), which responds specifically to DNA-reactive chemicals (Williams et al., 1989). Of these, 6 were positive, as follows: 2,4-DMA, 2,4,6-trimethylaniline (mesidine), 3,5-diaminobenzoic acid, 3,4-diaminochlorobenzene, 2-chloro-4-methylaniline and 4-chloro-N-methylaniline. Of these 6, some produced liver or bladder tumors, but not nasal tumors. Most notably, aniline and o-chloro-, o-methoxy-, o-ethyl-derivatives were all negative, as was phenetidine and 2,4,6-trichloroaniline. Given the high degree of correlation between positive results in hepatocyte DNA repair and carcinogenicity (Williams et al., 1989), it is expected that the positive chemicals would have carcinogenic activity at some site. Study of NM DNA adducts could provide insight as to whether this tissue would be a target for any of these chemicals.

o-Toluidine (2-methylaniline), as its hydrochloride, induced tumors at several sites, but not in the nasal cavity, in both male and female rats and mice starting at doses of up to 6,000 ppm (Weisburger et al., 1978; National Toxicology Program, 1979a; Hecht et al., 1982). IARC (International Agency for Research on Cancer, 1987d) concluded that there was sufficient evidence for carcinogenicity in animals. More recent additional studies in mice, rats and dogs confirmed the carcinogenicity of o-toluidine, but did not report on the occurrence of nasal neoplasms (Pliss, 2004). In genotoxicity assays, it produced a variety of positive effects, but results were often conflicting (International Agency for Research on Cancer, 1987d). o-Toluidine is a biotransformant of the local anesthetic prilocaine (PDR, 2006). In humans, hemoglobin adducts of o-toluidine have been reported in smokers and at lower levels in non-smokers (Bryant et al., 1988), and in workers in a chemical manufacturing facility (Ward et al., 1996).

It is possible that the substitutions in the anilines that lacked activity in the NM do not permit facile enzymic N-hydroxylation or sulfation to take place, as occurs with 2,6-DMA. In addition, if such biotransformants are formed, it may be that they require sufficient intrinsic reactivity to bind to DNA in order to be carcinogenic. As with 2,6-DMA, no report of any nasal effect in humans was found for these other anilines.

The structurally more complex polycyclic aromatic amines, for example the carcinogens β-naphthylamine, 4-aminobiphenyl, or 4,4′-methylene-bis(2-chloroaniline) (MOCA), are well-documented genotoxins (McQueen and Williams, 1990; Kadlubar and Badawi, 1995), as are the related complex heterocyclic derivatives formed as food pyrolysis products (Sugimura et al., 2000), as a consequence of N-hydroxylation and bioactivation (Miller, 1998). The simpler monocyclic derivatives discussed here generally show less genotoxicity. Based on a literature search, no polycyclic aromatic amine has been reported to produce nasal tumors in rodents, although in chronic toxicity studies a deliberate search for neoplasia may not always have been made. However, polycyclic aromatic amines can be bioactivated in the NM, since 4,4′-methylene-bis(2-chloroaniline) was found to bind to rat NM DNA (Jeffrey et al., 2002).

Benzene is the simplest aromatic hydrocarbon (Figure 19). It is widely used as an industrial solvent and is present in gasoline (~1%), automobile emissions, cigarette smoke, drinking water and a variety of foods. For most people, the level of exposure to benzene is likely to be higher from inhalation than ingestion. Benzene has been tested extensively for carcinogenicity and while positive, only 1 study reported nasal tumors resulting from oral administration (Maltoni et al., 1989). Benzene at a dose of 500 mg/kg bw for 78 weeks to both SD and Wistar rats was reported to induce nasal tumors, as well as Zymbal gland carcinomas, carcinomas of the oral cavity, skin, forestomach, and mammary glands, angiosarcomas of the liver, hemolymphoreticular neoplasias, tumors of the lung, and possibly hepatomas (Maltoni et al., 1989). Another study, however, using doses up to 200 mg/kg bw for 2 years, did not report nasal tumors (Huff et al., 1989). Homogenates from the NM had greater biotransformation capacity towards benzene than those from the liver (Low et al., 2003). Moreover, high concentrations of benzene were found to produce an inflamatory response and DNA fragmentation in cultured human nasal respiratory mucosa (Gosepath et al., 2003). However, unlike the related compounds naphthalene and bromobenzene discussed here, no report of rodent nasal cytotoxicity by benzene was found. Benzene yielded mixed results in genotoxicity test systems (International Agency for Research on Cancer, 2000b), with the exception of bone marrow. DNA binding was either absent or extremely low in several tissues in rats (Reddy et al., 1989), although binding was reported under extreme conditions (twice-daily treatment for 1 to 7 days with 440 mg/kg benzene) (Bodell et al., 1996). A scheme for the biotransformation of benzene based upon (Snyder and Hedli, 1996) is shown in Figure 19. The benzene biotransformant hydroquinone also has been tested for carcinogenicity in mice and rats with some evidence for increased kidney neoplasms in rats and liver neoplasms in mice at high doses, but without effect in the NM (National Toxicology Program, 1989). Epigenetic modes of action for the hydroquinone tumor increases have been proposed (Whysner et al., 1995).

Most human studies have not reported nasal effects of benzene (International Agency for Research on Cancer, 1987b). A cohort study of Chinese workers in occupations where benzene exposure is possible reported a suggestive increase in nasopharyngeal cancer mortality in males, but not females (Yin et al., 1996), and without a positive trend for cumulative exposure (Hayes et al., 1996). Moreover, nasopharyngeal cancer is endemic in China (see later) and confounding factors may be involved. The authors noted that tobacco use is frequent among Chinese men, but not women. Tobacco smoke contains benzene, 2,6-DMA, and carcinogenic nitrosamines (see later).

Chloroacetanilides

Alachlor, acetochlor, and butachlor (Figure 20) are herbicides that are 2,6-dialkyl aniline derivatives in which the nitrogen is disubstituted. At chronic dosages of >15 mg/kg bw/day, they have induced nasal and other tumors in rats, but not in mice (US EPA, 1997; Heydens et al., 1999; Genter et al., 2000; Green et al., 2000; Genter et al., 2004). The tumors appear to arise from OE cells (Genter et al., 2000). With exposure to 126 mg/kg bw/day alachlor in the diet, histologic changes in the NM were present by 3 months and the earliest NM tumors by 5 months (Genter et al., 2002a).

Chloroacetanilides were negative in most genotoxicity assays, including the comet assay for DNA damage in NM (Heydens et al., 1999). A study of binding of radiolabeled alachlor to rat nasal tissues showed minimal binding to DNA, although binding to proteins was substantial and twice that found in the liver (Heydens et al., 1999). At carcinogenic dosages, alachlor induced a substantial increase in cell proliferation in rat OE, but not RE at 60 days of dosing (Heydens et al., 1999).

The oncogenicity of chloroacetanilides has been linked to the formation of 2,6-dialkylbenzoquinoneimines (Feng et al., 1990; Li et al., 1992) (Figure 20), which can be formed in rats by several chemicals of this group (Jefferies et al., 1998). Whole-body autoradiograms of the distribution of radiolabeled alachlor, following oral administration to rats, showed that it was specifically associated with the NM (Feng et al., 1990), which was not the case with mice, which are insensitive to the induction of nasal tumors by these compounds. This species difference appears to result from differences in enterohepatic circulation and tissue specific biotransformation (Hadley and Dahl, 1983; Li et al., 1992), a consequence of which was the greater formation of the reactive diethyliminoquinone biotransformant (Figure 21) in the rat NM. This product binds extensively to glutathione (Jefferies et al., 1998) and thiols in proteins (Lambert et al., 1999), resulting in cytotoxicity and sustained proliferation of the NM (Heydens et al., 1999).

In studies of species differences, the formation of the quinone in liver and NM was 30 times more efficient in the rat than monkey tissues (Li et al., 1992) and up to 3,000 times more efficient than human tissues (Green et al., 2000). Given that mouse nasal tissue is less proficient than rat in forming critical metabolites and human nasal tissue is even less proficient than mouse (Heydens et al., 1999), the fact that the mouse is insensitive to chloracetanilide NM carcinogenicity suggests that humans would be unlikely to be responsive. In humans, no association with nasal toxicity or nasal or other tumors (Heydens et al., 1999) has been reported for chloracetanilide compounds.

Chloropyridines

3,5-Aminodichloropyridine (ADCP, Figure 22) is a component of several candidate respiratory disease drugs, whose mode of action involves inhibition of phosphodiesterase IV, which mediates hydrolysis of cAMP (Huang et al., 2001). One of these, N-(3,5-dichloropyrid-4-yl)-3-cyclopentyloxy-4-methoxybenzamide (piclamilast, RP73401, Figure 22), when given by single dose oral administration at > 50 mg/kg bw, produced toxicity to sustentacular cells and BGs of the OE of SD rats, but not mice (Pino et al., 1999). With administration by inhalation, it induced neuroblastomas of the OE of SD rats (Pino et al., 1999). A related compound, roflumilast (Figure 22), also produced nasal toxicity in rats (Jeffrey et al., 2002). The ADCP moiety, indicated in the structures, is suspected to be the toxic biotransformant of these molecules based on structural similarities to the monocyclic aromatic amines (Figure 22). Neither roflumilast nor ADCP formed DNA adducts in the rat NM (Jeffrey et al., 2002), however, indicating that the nasal toxicity and carcinogenicity does not involve DNA reactivity. Likewise, the carcinogenic activity of piclamilast probably does not involve DNA binding. A potential pathway of bioactivation of ADCP is formation of an epoxide (Figure 23), which could result in protein binding leading to cytotoxicity, increased cell proliferation and neoplasia, as has been demonstrated for the chloracetanilides discussed previously. Based on the available data, it seems likely that with sufficient exposure from systemic distribution, ADCP would elicit nasal tumors in rats. Patients in clinical trials of roflumilast (Timmer et al., 2002) have not been reported to experience any nasal effects. There are no available data for piclamilast.

Among pyridine analogs, 3-trifluoromethylpyridine (Figure 15) was toxic to the NM (see above). A large number of pyridine monomethyl- and dimethyl-substituted derivatives were not mutagenic in bacteria (Ho et al., 1981). 3-(Chloromethyl)pyridine was mutagenic in the presence of a bioactivation system, although its 2-analog was not. In the L5178Y mouse lymphoma mutagenicity assay (Dearfield et al., 1993), 2- and 3-chloropyridine produced small increases in mutants. The mutagenicity of 2-chloropyridine was greater with bioactivation and it was also mutagenic in Drosophila (Batiste-Alentorn et al., 1995). Pyridine itself produced liver tumors in male and female B6C3F1 mice, while findings were equivocal in rats (National Toxicology Program, 1996). Although 3-(chloromethyl)pyridine had carcinogenic activity in both rats and mice (National Toxicology Program, 1978b), its 2-chloromethyl analog did not (National Toxicology Program, 1978a). 2-Chloro-5-trifluoromethylpyridine, which is related to 3-trifluoromethylpyridine, which elicited nasal toxocity (see above), was weakly mutagenic in L5178Y cells (Dearfield et al., 1993). None of these pyridines, however, showed evidence of nasal toxicity or carcinogenicity, although pyridine administered by inhalation increased immunoreactivity of carboxyesterase in OM of F344 rats (Nikula et al., 1995).

Dimethylvinyl chloride (1-chloro-2-methylpropene) (Figure 24), a synthetic intermediate, is a structural analog of the human carcinogen vinyl chloride. When administered to groups of male and female F344/N rats and B6C3F1 mice at dosages of up to 200 mg/kg body weight, 5 days per week by gavage in corn oil for about 100 weeks it produced high incidences of nasal neoplasms in F344 rats, but not in B6C3F₁ mice, although neoplasms at other sites were increased (National Toxicology Program, 1979d). This may be explained by a greater amount of unmetabolized drug eliminated by exhalation in rats (30%) than mice (5%) (Ghanayem and Burka, 1987), thereby potentially exposing the nasal tissues.

Hexamethylphosphoramide (HMPA) (Figure 25), has been used as a chemosterilant for a number of insect pests and as a solvent for organic synthesis. It is highly toxic and carcinogenic by the inhalation route, producing mainly squamous cell carcinomas of the anterior portions of the nasal cavity of Sprague–Dawley (SD) rats (Lee and Trochimowicz, 1982). When given orally to SD rats at >100 ppm in drinking water or >15 mg/kg bw/day by intragastric instillation for ~90 days, nasal epithelial denudation and inflammation were observed in the maxilloturbinates and nasoturbinates (Keller et al., 1997). HMPA is activated by CYP2A13, which is expressed at the highest levels in human NM (Su et al., 2000).

HMPA has produced some positive genotoxicity findings (Ashby et al., 1985; International Agency for Research on Cancer, 1999b). Although oral carcinogenicity has not been studied, given the extent of data, it seems likely that an adequate study would reveal nasal carcinogenicity, since clearly HMPA is systemically distributed to the NM where it is toxic. No data were found on human toxicity, genetic effects, or carcinogenicity. Based on experimental findings, HMPA was classified as possibly carcinogenic to humans (International Agency for Research on Cancer, 1999b).

Iodinated glycerol (Organidin), which has been used as an expectorant, has as its major component 3-iodo-1,2-propanediol (Figure 26), along with an isomeric mixture of 2-(1 or 2-iodoethyl)-1,3-dioxolane-4-methanol. It was positive in some genotoxicity assays and, when administered at 250 mg/kg in male rats, it produced nasal cavity tumors in 2 rats (National Toxicology Program, 1990a).

Nitrosamines (e.g., Figure 27) are synthetic and naturally occurring compounds. They are a thoroughly studied type of DNA-reactive carcinogen, which show marked organ specificity reflecting local biotransformation (Preussmann and Steward, 1984; Magee, 1996; Lewis et al., 1997). Many of this chemical class, examples of which are given in Figures 27 and 28, targeted the NM (Preussmann and Steward, 1984; Schuller, 1997), including, for example, dimethylnitrosamine in Chinese hamsters and rats (International Agency for Research on Cancer, 1987c), diethylnitrosamine in hamsters (Herrold, 1964), and rabbits (Reznik and Padberg, 1991), di-n-propylnitrosamine in hamsters (Pour et al., 1974) and the tobacco-specific nitroso compound 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) (Figure 28) in rats (Prokopczyk et al., 1991) and hamsters (Hoffmann et al., 1981), 3-(methylnitroso)propionitrile in rats (Wenke et al., 1984) and the cyclic nitrosamines, N-nitrosopiperidine (NPIP, Figure 29, left) (Lijinsky and Taylor, 1975) and N-nitrosopyrrolidine (Figure 29, right) (Gray et al., 1991) in rats. The olfactory tumors induced were primarily carcinomas.

Nitrosamines are well-established alkylating agents after bioactivation, which involves hydroxylation on the carbon α to the nitroso group (Magee et al., 1975). Accordingly, they are DNA-reactive (Williams and Laspia, 1979; Muller and Rajewsky, 1980; Boucheron et al., 1987; Magee, 1996; Hecht, 1999).

NNK is bioactivated in mice by CYP2A5, which is a significantly better catalyst of NNK α-hydroxylation than is the closely related human enzyme, CYP2A6 (Felicia et al., 2000). A generalized scheme for the biotransformation of NNK, based on Hecht (Hecht, 1994) and Sturla et al (Sturla et al., 2005), is shown in Figure 28. Evidence has been provided that DNA pyridyloxobutylation is important in the rat NM carcinogenicity of NNK and N-nitrosonornicotine (Trushin et al., 1994). By chemical reaction several DNA adducts are formed, the major product being 7-(pyridyloxobutyl-1-yl)guanine (Sturla et al., 2005) (Figure 28). The rodent NM is rich in the enzymes supporting the biotransformation of nitrosamines (Brittebo et al., 1981), including NNK (Hong et al., 1992; Hecht, 1994; Su et al., 2000). This involves CYP2A-mediated oxidation and p-de-ethylation activity (CYP2A6 > 2E1) (Bereziat et al., 1995), which in rats is higher in the NM than in the liver or lung (Hadley and Dahl, 1982). Human CYP2A13 is the most efficient catalyst of α-hydroxylation of NNK (Wong et al., 2005).

NPIP (Figure 29, left) is a potent rat nasal carcinogen (Lijinsky and Reuber, 1981) whereas N-nitrosopyrrolidine (Figure 29, right) is highly carcinogenic in the liver but only weakly so in the nasal cavity (Lijinsky and Reuber, 1981). This difference has been explained by more efficient α-hydroxylation of NPIP due to 20–40 fold greater catalytic efficiency (V_max/K_m) of rat nasal microsomes for NPIP as a substrate, probably owing to the presence of CYP2A3, although some role has been ascribed to CYP2G1 (Wong et al., 2003).

Bioactivation in the NM clearly underlies the carcinogenicity of nitrosamines in that tissue (International Agency for Research on Cancer, 1985; Belinsky et al., 1990; Hong et al., 1992; Bereziat et al., 1995). Ethanol administration to rats has been reported to induce the α-carbon-hydroxylation activity in the NM (Castonguay et al., 1984).

No nitrosamine has been directly implicated as a nasal carcinogen in humans. Smokers exhibit NM alterations and have increased risks of sinonasal squamous cell carcinoma (Feron et al., 2001; International Agency for Research on Cancer, 2004) and nasopharyngeal carcinoma (International Agency for Research on Cancer, 2004). A role for NNK is possible (International Agency for Research on Cancer, 2004), although other rodent carcinogens, including the nasal carcinogen 2,6-dimethylaniline (see above), are found in cigarette smoke.

Phenacetin (PA, Figure 30), is an N-acetylated, single ring, aromatic amine, which was formerly widely used as an analgesic, but now is replaced by the related APAP, discussed above. Daily intragastric instillation administration of PA to SD rats at doses of up to 1.25 g/kg/day for up to 2 weeks resulted in degenerative changes in the OE and necrosis of BGs associated with increases in cell proliferation only in the OE (Bogdanffy et al., 1989). The dose-response relationship for cell proliferation, which reached an increase of over 700% at the high dose, was similar to that of nasal tumor formation, suggesting that the primary site of PA toxicity within the NM is the OE, with restorative cell proliferation being confined to the epithelial cell layer. The data indicated that early cell proliferative responses may be important in the genesis of nasal tumors.

PA was carcinogenic in mice and rats (International Agency for Research on Cancer, 1987e). In SD rats, nasal cavity adenocarcinomas, transitional cell carcinomas and squamous cell carcinomas, as well as urinary bladder tumors were produced by administration of 2.5 and 1.25 % in the diet for 18 months (Isaka et al., 1979). In B6C3F₁ mice, renal tumors were produced in males and urinary bladder tumors in both genders, but no nasal tumors (Nakanishi et al., 1982). Interestingly, PA was not carcinogenic in the sensitive TP53^+/− transgenic mouse bioassay (Storer et al., 2001).

The biotransformation of PA is principally by O-de-ethylation mediated by CYP1A2 (Belle et al., 2000) in the centrilobular regions of the liver (Pang et al., 1988) to yield APAP (Figure 1 in Hinson, 1983). In addition, N-deacetylation and N-hydroxylation reactions occur. Biotransformation in the rat NM has been reported (Brittebo, 1987). Although APAP was also toxic to the NM, as discussed here, there is no evidence that the toxicity of PA is due to formation of APAP.

Binding of ¹⁴C- and ³H-PA to DNA in vivo has been measured, but the results are complicated by its biotransformation and subsequent incorporation of the label into nucleic acid (Nery, 1971). PA bound selectively to the nasal turbinates (Brittebo and Ahlman, 1984), suggests local bioactivation, but the identity of bound product(s) has not been established. No data are available on the formation of DNA adducts, although it can be inferred from the levels of in vitro modification of DNA achievable with biotransformants of PA, that adducts could be readily detected given the sensitivity of the ³²P-postlabeling technique to detect 1 adduct in 10¹⁰ bases (Phillips et al., 2000). PA could bind to DNA through either N-hydroxyPA or N-hydroxyphenetidine. N-Hydroxyphenetidine bound to DNA directly at both pH 7.0 and 5.0, with binding about 4 times better at the lower pH (Mulder et al., 1984). CYP1A2 appears to be the only high affinity human liver PA O-deethylase (Venkatakrishnan et al., 1998). Additional studies of the biotransformation and binding of [ring-³H]PA in the NM were studied in vitro and in vivo in male SD rats (Peng et al., 1993). Whole-body autoradiography showed irreversible binding to the BGs in the OM after high, but not low, doses of [³H]PA. In the other tissues, the distribution of radioactivity was not changed when the dose was increased. GSH depletion by pretreatment with phorone resulted in binding to the BGs even after a trace dose of [³H]PA. The data also suggest that in situ bioactivation and binding of PA in the rat NM at high doses may play a role in the pathogenesis of the nasal tumors.

As reviewed by IARC (International Agency for Research on Cancer, 1987e), in genotoxicity assays, PA induced revertants in Salmonella typhimurium, only with strain TA100 and hamster microsomes. PA did not cause mutations to ouabain resistance or transform C3H/10T1/2 cells (Patierno et al., 1989). Nevertheless, it produced micronuclei in peripheral erythrocytes of mice (Higashikuni et al., 1992). Of the biotransformants of PA, N-hydroxyPA and N-acetoxyPA were both mutagenic in the presence of hamster S9 and N-hydroxyphenetidine and N-acetoxyPA were active per se in TA 100 strains. Thus, PA appears to be genotoxic, possibly via bioactivation to a quinoneimine (Hinson, 1983) or to N-hydroxyPA (Mulder et al., 1984).

PA-containing analgesics have been concluded by IARC to cause cancer of the renal pelvis and bladder in humans (International Agency for Research on Cancer, 1987e), but no association with nasal tumors has been noted. No mention of nasal toxicity in humans was found in the literature.

Procarbazine (Matulane, Figure 31) is a methylhydrazine derivative and is indicated for use in combination with other anticancer drugs for the treatment of Hodgkin’s lymphoma. Given as its hydrochloride, it produced olfactory neuroblastomas in both rats and mice, as well as other nasal neoplasms (National Toxicology Program, 1979c). While O⁶-methylguanine levels arising from procarbazine have been measured in blood leukocytes of rats (Valavanis et al., 1994), no comparable measurements in the NM seem to have been made. Such methylation requires metabolic or spontaneous oxidation to azoprocarbazine and possible diazomethane intermediates that can methylate DNA (Figure 31) (Lee and Dixon, 1978; Ogawa et al., 2003). Procarbazine is mutagenic in all test systems in vitro and in vivo (International Agency for Research on Cancer, 1981).

According to the last IARC review (International Agency for Research on Cancer, 1987g), no data in humans were found for procarbazine as a single agent. In various combinations, it was associated with appearance of acute non-lymphocytic leukemia. In nonhuman primates, procarbazine also produced acute nonlymphocytic leukemia (Thorgeirsson et al., 1994). No indication of nasal cytotoxicity or carcinogenicity was found (PDR, 2006).

Salted fish is a regional food item. It is produced and consumed primarily in Southeast Asia and northern Europe. Chinese-style salted fish is prepared by treating fish with dry salt or an aqueous salt solution and often subsequently drying in the sun. It is usually softened by partial decomposition before or during salting.

Beginning with a small-scale study (Huang et al., 1978), several experiments have demonstrated that feeding of high concentrations (i.e., >5%) of Chinese salted fish in the diet induced nasal cavity tumors in rats (Yu et al., 1989; Zheng et al., 1994). Low levels of several volatile nitrosamines, many of which are nasal carcinogens (see above), have been detected in Chinese salted fish (Huang et al., 1981; Tannenbaum et al., 1985) and high levels of N-nitrosodimethylamine have been reported in some samples (International Agency for Research on Cancer, 1993c).

The IARC (International Agency for Research on Cancer, 1993c) concluded that there was sufficient evidence in humans for causation of nasopharyngeal carcinomas by Chinese-style salted fish. This was subsequently further supported by a population-based case-control study (Yuan et al., 2000).

Overview of Carcinogenicity

At least 11 compounds or classes of compounds have been documented to be nasal carcinogens in rodents. Of these, only salted fish was associated with carcinogenicity in humans.

Mechanisms of Rodent Nasal Carcinogenicity

Several potential mechanisms exist to explain the selective production of nasal tumors by agents distributed systemically to the NM. In the case of nitrosamines, the effect on the NM is related to the bioactivation to DNA-reactive chemical species in this tissue, as is likely for PA and 2,6-DMA, and possibly benzene and procarbazine. In contrast to the established DNA-reactivity of some nasal carcinogens, the chloroacetanilides and aminodichloropyridines are also biotransformed to chemical-reactive species that bind to some cellular macromolecules, particularly proteins, but not to DNA. The superfamilies of both the Phase I and Phase II enzymes, which mediate these processes are becoming better understood. In vitro assays for genotoxicity have shown that when bioactivation is required, enzyme preparations from different species or organs, or after enzyme induction, can show marked differences in activity. For example, highly expressed forms of CYP enzymes, such as CYP2G, occur in rabbit nasal microsomes which bioactivate nasal carcinogens such as N-nitrosodiethylamine and PA (Ding and Coon, 1988). The rat NM contains relatively high levels of CYP enzymes, including CYP2E1 (Brittebo, 1997; Thornton-Manning and Dahl, 1997) and CYP2G (Gu et al., 1997). In contrast, the latter is functionally rare in humans (Sheng et al., 2000). These mixed function oxidases are inducible (Turk et al., 1987). Thus, greater bioactivation to DNA-reactive or DNA-damaging products appears to account for the susceptibility of the NM to some agents such as nitrosamines and 2,6-DMA.

The cellular events following DNA binding in the NM, however, have not been studied in detail, in contrast to carcinogenesis in many other tissues. Gene alterations in rodent nasal squamous cell carcinomas (Recio, 1997), and in NM tumors (Genter et al., 2000) have been studied. Also genomic analysis has been conducted on alachlor-induced oncogenesis in rat OM (Genter et al., 2002b). With dosing for up to 1 month, 148 genes and expressed sequence tags (ESTs) were up-regulated. A major subgroup of these was genes related to control of extracellular matrix including metalloproteinase genes, whose products appear to be important in nasal tumor progression (Genter et al., 2005). Other up-regulated genes included several related to cell cycle and proliferation. In tumors, cell proliferation genes were also up-regulated, as were genes encoding for proteins associated with oxidative stress responses. More advanced malignant tumors displayed gene changes indicative of activation of the wnt pathway (Genter et al., 2002b). With regard to tumor suppressor genes, the fact that TP53^+/− mice do not exhibit increased susceptibility to induction of nasal tumors by p-cresidine may suggest that TP53 is not a critical gene in the pathogenesis of this tumor, although lack of bioactivation in NM could also be an explanation.

In some instances DNA may not be the critical target. For example, the chloroacetanilides and ADCP do not bind to DNA. If sufficient toxicity occurs by any mechanism, such as protein binding (Heydens et al., 1999), increased and sustained cell proliferation may result. While much evidence suggests that such proliferation is important in the development of tumors (Butterworth et al., 1992), it alone is not sufficient in some circumstances (Umemura et al., 1992; Ward et al., 1993), as evidenced by the nasal toxins that are not carcinogenic to the NM. Thus, further mechanistic research is clearly needed.

Human Effects of Rodent Nasal Carcinogens

In most human populations, neoplasms of the nasal cavity and paranasal sinuses are uncommon, representing less than 0.5 % of invasive cancers (Bhattacharyya, 2002). The most frequent types are papillomas, inverted papillomas, and squamous cell carcinomas (Bhattacharyya, 2002). In southern China and Southeast Asia, nasopharyngeal cancer is endemic (Chan et al., 2004).

The etiology of any of these types of nasal neoplasia is poorly understood, except in a few situations in which airborne exposure is associated with increased risks. Inhalation of wood dust (International Agency for Research on Cancer, 1995), and formaldehyde (International Agency for Research on Cancer, 2005), other occupational exposures, such as occur in the nickel refining (International Agency for Research on Cancer, 1990b) and chromate industries (International Agency for Research on Cancer, 1990a), and cigarette smoking (International Agency for Research on Cancer, 2004), have been concluded by IARC working groups to be causative of sinonasal or nasopharyngeal cancer. The report of a causal association of nasopharyngeal cancer with formaldehyde exposure (Hauptmann et al., 2004), however, has been challenged (Marsh and Youk, 2005). Moreover, the occupational history of 70 patients with sinonasal inverted papilloma, however, did not reveal association of the above agents with 95% of cases (Barbieri et al., 2005).

Dietary factors may also contribute to the development of nasopharyngeal neoplasms (Zheng et al., 1992), including consumption of Chinese salted fish (International Agency for Research on Cancer, 1993c), which also causes nasal tumors in rats (see above). Epstein–Barr virus (International Agency for Research on Cancer, 1997) and human papilloma virus infections (Katori et al., 2006) have also been implicated.

The mortality from nasopharyangeal cancer is reported to be increased in Chinese men in occupations where benzene exposure is possible (Yin et al., 1996). However, as noted, nasopharyngeal cancer is endemic in China, and the incidence may have been confounded by the presence of other risk factors including genetic susceptibility, carcinogens in salted fish, cigarette smoking, and Epstein–Barr viral infection (Chan et al., 2004).

The analgesic PA, which is no longer on the market, when abused caused kidney disease and renal pelvis and bladder cancer (International Agency for Research on Cancer, 1987e), but not nasal tumors, in contrast to its effects in rats, indicating possible differences in bioactivation in the NM. In PA-consuming individuals, there is an increased relative risk of all cancers of 1.9 (confidence interval: 1.1–3.3), as well as increased death and morbidity from other conditions (Dubach et al., 1991). IARC classifies PA as having limited evidence of carcinogenicity in humans, whereas there is sufficient evidence for analgesic mixtures containing PA (International Agency for Research on Cancer, 1987e; Mery et al., 1994; Menco and Morrison, 2003). In cases of chronic abuse of PA, in 16 urothelial carcinomas, the TP53 tumor suppressor gene showed variability in the mutation pattern (Petersen et al., 1993), which the authors concluded to suggest that the tumors arose through chronic tissue damage rather than from promutagenic DNA lesions.

As reviewed by IARC (International Agency for Research on Cancer, 1999c), an international cohort of workers exposed to phenoxy herbicides and chlorophenols exhibited and elevated risk for cancers of the nose and nasal cavities (standardized mortality ratio: 2.9; 95% CI 0.6–8.2; Saracci et al., 1991). The authors noted, however, that the calculated risk was based on small numbers of deaths (3 exposed workers versus 0 in nonexposed workers). Moreover, it has been suggested that the conclusion of increased risk was not appropriate because the excess was not established by prior hypothesis and was not statistically significant (Peto, 1991).

2,6-DMA is considered by IARC to be possibly carcinogenic to humans, although no human data were available (International Agency for Research on Cancer, 1993a). Also, a warning on the carcinogenicity of 2,6-DMA, as a minor biotransformant of lidocaine, is included in the labeling of lidocaine (PDR, 2006). The identification of 4-hydroxy-2,6-DMA glucuronide as a urinary biotransformant in humans (Tam et al., 1990) allows for the possibility of genotoxicity through formation of a quinoneimine intermediate. 2,6-DMA-hemoglobin adducts have been measured in smokers and non-smokers (Bryant et al., 1988). Curiously, levels were about 3 times higher in nonsmokers, although this was not true for all aromatic amines. Human liver slices biotransform lidocaine to 2,6-DMA (Parker et al., 1996) and levels of 2,6-DMA-hemoglobin adducts were increased in patients taking lidocaine (Bryant et al., 1994). Analysis of hemoglobin adducts, therefore, may represent a sensitive way to detect exposure.

Thus far, PA is the only monocyclic aromatic amine to be conclusively associated with cancer in humans. The IARC review (International Agency for Research on Cancer, 2000a) of o-toluidine and its p-chloro-analog concluded they were probable (2A) human carcinogens, although there was inadequate evidence for carcinogenicity in humans for o-toluidine (International Agency for Research on Cancer, 1987d). The human data have been the subject of vigorous debate (Freudenthal and Anderson, 1995), centering on the problem that workers are exposed to other chemicals at the same time as o-toluidine; accordingly, it is difficult to associate tumors with any particular chemical.

Biomonitoring of lidocaine, as mentioned above, has been achieved by measuring increases in hemoglobin adducts. Measurements of o-toluidine protein adducts in rats (DeBord et al., 1992) revealed that hemoglobin adducts were more stable than those of albumin, having half-lives of about 12.3 and 2.6 days respectively, which interestingly is not as long as the life span of an erythrocyte of about 120 days, as might have been expected. DNA adduct formation by o-toluidine has been investigated in vitro (Marques et al., 1996), but not successfully in vivo owing to the chromatographic similarity of the adducts to normal nucleotides (F. Beland, personal communication). In all these experiments care has to be taken to use non-smokers and the analyses must be conducted in clean ambient environments to avoid contamination (Luceri et al., 1993).

The NM of humans, like that of rodents, has xenobiotic biotransformation capability (Gervasi et al., 1991; Wong et al., 2005). Importantly, where comparisons have been made (Feng et al., 1990; Li et al., 1992), rodent NM has higher biotransformation activity than does human. Moreover, the great majority of humans have loss-of-function mutations in CYP2G genes (Sheng et al., 2000), which encode abundant OE-specific enzymes in animals, particularly rats. Thus, humans may be protected by their lesser bioactivation.

Molecular changes in human nasal tumors have been the subject of a few reports. In a series of Japanese patients, over expression of p53 protein was found in about 60% of nasal squamous cell carcinomas and adenomas, but not in normal mucosa, benign or premalignant lesions (Fang et al., 1998). The overexpression correlated well with heavy smoking. Similarly, increased p53 immunoreactivity was found in nasal inverted papillomas (Schwerer et al., 2001). A study of nasal biopsies of children in Mexico City with nasal pathology revealed strong transmural p53 staining in about 25% of the samples (Calderon-Garciduenas et al., 2001). This was suggested to reflect a response to air pollution toxicity. TP53 mutations in PA-associated urothelial carcinomas (Petersen et al., 1993) have been reported, but show variability. Microsatellite instability was found in about 40% of RE sinonasal carcinomas (Uryu et al., 2005). Precancerous lesions of sinonasal inverted papillomas displayed increased expression of matrix metalloproteinases 2 and 9 (Katori et al., 2006).

In summary, of the at least fourteen compounds or classes of organic agents that induce nasal cytotoxicity or neoplasia in rodents with systemic distribution, only Chinese-style salted fish was associated with a comparable effect in humans. This may reflect lesser biotransformation activity in human NM and lower exposure of humans. Little is known about the molecular pathogenesis of nasal tumors in humans or rodents.

Overall Conclusions

Experimental evidence documents that some systemically distributed organic compounds can, organotropically, produce cytotoxicity and exert carcinogenic activity in the NM of rodents, particularly in rats. Many of these agents specifically accumulate in the nasal regions of rodents, tissues well endowed with biotransformation capability, and thereby lead to generation of chemically reactive molecular entities. Similar processes are less likely to occur in humans. Thus, the present analysis suggests that compounds lacking DNA reactivity that induce only rodent nasal tumors with systemic administration may represent no cancer hazard, either nasal or otherwise, to humans under usual exposure conditions. For other compounds, especially those that are DNA reactive and also produce tumors at other sites, there may be risks to humans of cancer induction at some site, again, depending upon exposure.