• E-mail
• Print
• Comment
• Font Size
• Digg
• del.icio.us
• Discuss article

Methylated Arsenicals: The Implications of Metabolism and Carcinogenicity Studies in Rodents to Human Risk Assessment

Posted on: Friday, 21 April 2006, 06:00 CDT

By Cohen, Samuel M; Arnold, Lora L; Eldan, Michal; Lewis, Ari S; Beck, Barbara D

Monomethylarsonic acid (MMA^sup V^) and dimethylarsinic acid (DMA^sup V^) are active ingredients in pesticidal products used mainly for weed control. MMA^sup V^ and DMA^sup V^ are also metabolites of inorganic arsenic, formed intracellularly, primarily in liver cells in a metabolic process of repeated reductions and oxidative methylations. Inorganic arsenic is a known human carcinogen, inducing tumors of the skin, urinary bladder, and lung. However, a good animal model has not yet been found. Although the metabolic process of inorganic arsenic appears to enhance the excretion of arsenic from the body, it also involves formation of methylated compounds of trivalent arsenic as intermediates. Trivalent arsenicals (whether inorganic or organic) are highly reactive compounds that can cause cytotoxicity and indirect genotoxicity in vitro. DMA^sup V^ was found to be a bladder carcinogen only in rats and only when administered in the diet or drinking water at high doses. It was negative in a two-year bioassay in mice. MMA^sup V^ was negative in 2-year bioassays in rats and mice. The mode of action for DMA^sup V^-induced bladder cancer in rats appears to not involve DNA reactivity, but rather involves cytotoxicity with consequent regenerative proliferation, ultimately leading to the formation of carcinoma. This critical review responds to the question of whether DMA^sup V^-induced bladder cancer in rats can be extrapolated to humans, based on detailed comparisons between inorganic and organic arsenicals, including their metabolism and disposition in various animal species. The further metabolism and disposition of MMA^sup V^ and DMA^sup V^ formed endogenously during the metabolism of inorganic arsenic is different from the metabolism and disposition of MMA^sup V^ and DMA^sup V^ from exogenous exposure. The trivalent arsenicals that are cytotoxic and indirectly genotoxic in vitro are hardly formed in an organism exposed to MMA^sup V^ or DMA^sup V^ because of poor cellular uptake and limited metabolism of the ingested compounds. Furthermore, the evidence strongly supports a nonlinear dose-response relationship for the biologic processes involved in the carcinogenicity of arsenicals. Based on an overall review of the evidence, using a margin-of- exposure approach for MMA^sup V^ and DMA^sup V^ risk assessment is appropriate. At anticipated environmental exposures to MMA^sup V^ and DMA^sup V^, there is not likely to be a carcinogenic risk to humans.

Keywords Arsenic Metabolism, Bladder Carcinogenesis, Cell Proliferation, Cytotoxicity, Methylated Arsenicals, Risk Assessment

I. INTRODUCTION

Arsenic is a naturally occurring element that exists in the environment in a number of different forms, each with its own unique physical, chemical, and toxicological characteristics. Inorganic arsenic (As^sub i^), most often in tri valent form (arsenite, As^sup III^^sub i^) or pentavalent form (arsenate, As^sub i^^sup V^), is the most abundant form of arsenic in nature, and is commonly present in soil, water, and food (NRC, 1999; Mok and Wai, 1994; YanChu, 1994). The natural content of arsenic in soils, globally, ranges from 0.01 to over 600 ppm, with an average of about 2 to 20 ppm depending on the country and source of information (Yan-Chu, 1994). Groundwater in several parts of the world contains substantial amounts of arsenic, primarily due to release of naturally occurring arsenic from subsurface rock formations (Nordstrom, 2002). Human exposure to naturally occurring arsenic in some world regions is significant because of the use of arsenic-contaminated groundwater as a primary source of drinking water (Chiou et al., 2001; Guo et al., 2004; NRC, 1999; Wu et al., 1989). In contrast, methylated arsenic compounds, including monomethylarsonic acid (MMA^sup V^)* and dimethylarsinic acid (DMAV) and their salts, are rarely detected in groundwater and, as a result, human exposure to these compounds from environmental sources is expected to be minimal (NRC, 1999).

Once ingested, inorganic arsenic is easily absorbed into the blood and taken up by cells in tissues, primarily the liver, where it undergoes a series of reductions and oxidative methylations (Figure 1a) to form the pentavalent organic metabolites MMA^sup V^ and DMA^sup V^, which are more easily excreted in the urine than inorganic arsenic itself. The methylation pathway was long considered to be a detoxification process (Gebel, 2002; Vahter, 1983). However, studies in the last few years have demonstrated the presence of unstable trivalent intermediates in the urine of humans exposed to drinking water containing high levels of inorganic arsenic (Aposhian et al., 2000a, 2000b; Le et al., 2000a). These trivalent intermediates, monomethylarsonous acid (MMA^sup III^) and dimethylarsinous acid (DMA^sup III^), formed during the metabolism of inorganic arsenic, are distinct compounds having structures that are spatially completely different from that of the pentavalent compounds; are highly reactive, and may play a role in the carcinogenicity of inorganic arsenic in humans. In contrast to inorganic arsenic, exogenous MMA^sup V^ and DMA^sup V^ show limited absorption and metabolism in humans and in most animals (except the rat), and are excreted mostly as parent compound (Buchet et al., 1981; Hughes and Kenyon, 1998; Marafante et al., 1987; Yamauchi et al., 1988; Yoshida et al., 1998). There is a special interest in MMA^sup V^ and DMA^sup V^ because the same compounds are manufactured for use mainly as herbicides for weed control on cotton, nonbearing orchards, turf, and in noncrop areas.

Based on epidemiological data, inorganic arsenic is a known carcinogen, causing tumors in skin, lung, bladder, and possibly other tissues in humans exposed to high levels (Chen et al., 1988, 1992, 2003b; NRC, 1999; Tseng et al., 1968). The mechanism or mode of action by which arsenic causes cancer in humans has not been clarified (Abernathy et al., 1996; Beck et al., 2001; Brown and Ross, 2002; Kitchin, 2001). The study of the toxic effects of inorganic arsenic has been complicated by the lack of a reliable animal model (Byron et al., 1967; Chan and Huff, 1997; Huff et al., 2000; IARC, 1980; NRC, 1999). It is plausible that the mechanism(s) vary according to tumor site and that multiple, interacting mechanisms exist, including interactions with proteins by reaction with sulfhydryl groups, oxidative damage secondary to the generation of free radicals, depletion of glutathione, inhibition of DNA repair mechanisms, and/or cytotoxicity and regeneration. Direct genotoxicity (DNA reactivity) does not appear to be involved. Inorganic arsenic is not a carcinogen in any known animal model, except for the recently demonstrated transplacental effect in mice administered extremely high doses of As^sup III^^sub i^. This effect is possibly secondary to alterations in estrogen regulation (Waalkes et al., 2004a).

FIG. 1. (a) Classical metabolism of inorganic arsenic (from Le et al., 2000). In most animal species, including humans, the intracellular metabolism of inorganic arsenic involves extensive metabolism to DMA^sup V^. Humans also excrete relatively large amounts of MMA^sup V^. The rat is the only species that excretes significant amounts of TMAO. (b) Newly proposed metabolism of inorganic arsenic (from Hayakawa et al., 2005). (ATG, arsenic triglutathione; MADG, monomethylarsenic digluatathione; DMAG, dimethylarsinic glutathione.

DMA^sup V^ was found to be carcinogenic to rat bladder, but not to any organ in mice. The mode of action for the DMA^sup V^-induced bladder tumors was identified as cytotoxicity with necrosis of the bladder urothelium followed by a regenerative process and leading to sustained increased cell proliferation (i.e., hyperplasia). The relevance (if any) of this finding to the carcinogenicity of inorganic arsenic in humans is questionable. It is also questionable whether the cytotoxicity and carcinogenicity of DMA^sup V^ in rats is relevant to the risk of exposure to exogenous DMA^sup V^ and MMA^sup V^ in humans. To assess whether the cytotoxicity and carcinogenicity of DMA^sup V^ in rats is relevant to humans, an understanding of the mode of action, the metabolism, and the toxicokinetics of the different arsenic compounds in different animal species is necessary.

Arsenic compounds vary in their metabolism and disposition, depending on the specific compound, the exposure route, and the animal species (Vahter, 1983). These factors are significant in determining the differences in toxicity between the various arsenic compounds and between the toxic effects produced in the various animal species. The metabolism and disposition of ingested methylated compounds differ significantly from those of ingested inorganic arsenic, and the metabolism and disposition of methylated compounds vary depending on whether the exposure to the compounds is exogenous (e.g., via ingestion) or endogenous (i.e., via metabolism of inorganic arsenic).

This critical review analyzes the differences between the metabolism and toxicity of methylated compounds of arsenic that are exogenous, and the same compounds that are generated intracellularly during the metabolism of ingested inorganic arsenic. These metabolic and toxicologie differences are examined in the context \of the mode of action of DMA^sup V^-induced rat bladder tumors, and the DMA^sup V^ dose-response, considering species sensitivities, with implications for human risk assessment.

II. METABOLISM AND DISPOSITION OF INORGANIC ARSENIC, MMA^sup V^, AND DMA^sup V^

A. Metabolism and Disposition of Inorganic Arsenic In Vivo

In humans and most experimental animals, inorganic arsenic (As^sup III^^sub i^ and As^sub i^^sup V^) is rapidly absorbed into the blood after ingestion. The majority of the inorganic arsenic (>90%) is cleared rapidly from the blood, with a half-life of 1-2 hours. Arsenic leaves the body primarily through release into the urine in two distinct phases estimated at 20 and 300 hours (Vahter, 1983). In general, once absorbed, As^sub i^^sup V^ is reduced to As^sup III^^sub i^ mainly in the blood and liver, with some reduction probably occurring in the gastrointestinal tract and the stomach (Gregus and Nerneti, 2005; Herbel et al., 2002, NRC, 1999; Nmeti and Gregus, 2005). As^sup III^^sub i^ is then taken into cells and methylated intracellularly to form MMA^sup V^ and DMA^sup V^. DMA^sup V^ is the primary metabolite excreted into the urine in most mammals. Only humans excrete significant amounts of MMA^sup V^ (Vahter, 1994). The rat behaves significantly differently from other mammals regarding arsenic metabolism and disposition, primarily because of retention of arsenic by red blood cells and extensive conversion to the trimethylated form, trimethylarsine oxide (TMAO). The primary site of arsenic metabolism in mammals is the liver, although there is also high methylating activity in testes, kidney, and lung tissues (Aposhian, 1997; Georis et al., 1990; Healy et al., 1998; Lerman and Clarkson, 1983; Lerman et al., 1983; Styblo et al., 2002; Vahter, 1999).

The metabolism of inorganic arsenic takes place via the sequence of reactions described in Figure Ia, involving a series of sequential reduction and oxidative methylation reactions. Reduction of pentavalent arsenic species to the trivalent form involves thiol reducing agents such as glutathione (GSH) orthioredoxin. S- Adenosylmethionine (SAM) supplies the methyl group for methylation of arsenic compounds by specific methyltransferases (MT) (Buchet and Lauwerys, 1988; Cullen et al., 1984; Hirata et al., 1990; Marafante et al., 1985; Thompson, 1993; Vahter and Envall, 1983; Vahter and Marafante, 1988). Recently, a rat liver arsenic methyltransfera.se, Cytl9, was identified that can catalyze the conversion of As^sup III^^sub i^ to mono- and dimethylated arsenic (Lin et al., 2002, Thomas et al., 2004; Waters et al., 2004a).

The rat liver protein Cytl9, which appears to catalyze the generation of methylated metabolites formed from arsenite (Lin et al., 2002; Thomas et al., 2004; Waters et al., 2004a), using thioredoxin, thioredoxin reductase, and NADPH as reducing agents, has a Km value for conversion of MMA^sup V^ to DMA^sup V^ that is orders of magnitude lower than that estimated for the arsenic methyltransferase from rabbit liver. Cytl9 appears to combine the functions of methyltransferase and reductase in a single protein and involves an arsenic-GSH complex as substrate.

More recently, a new metabolic pathway for methylation of inorganic arsenic has been suggested (Hayakawa et al., 2005) based on the nonenzymatic formation of glutathione complexes with inorganic arsenic, leading to the formation of triglutathione arsenic, which is then methylated by Cytl9 (Figure Ib). The methylated glutathione arsenicals were hydrolyzed when the glutathione concentration was less than 1 mM to form the trivalent methylated arsenicals (MMA^sup III^ and DMA^sup III^), which were further oxidized to the pentavalent forms (MMA^sup V^ and DMA^sup V^). The formation of triglutathione arsenic occurred at glutathione concentrations of 2 mM or higher. MMA^sup V^ and DMA^sup V^ are not substrates according to this metabolic scheme. Support for this alternative pathway comes from the observation of large amounts of arsenic triglutathione and monomethylarsenic diglutathione as major metabolites in bile and small amounts in urine of rats exposed to As^sup III^^sub i^ orally, subcutaneously, or intravenously (Csanaky and Gregus, 2005; Cui et al., 2004; Hayakawa et al., 2005; KaIa et al., 2000, 2004; Kobayashi et al., 2005; Thomas et al., 2004).

Methylation reactions appear to facilitate excretion, thereby decreasing the toxic effects of inorganic arsenic (Gebel, 2002; Vahter, 1994,1999; Vahter and Concha, 2001). However, as discussed earlier, recent data indicate that the trivalent methylated arsenicals formed during the intracellular metabolism of inorganic arsenic (Figure 1a) are highly reactive and may play a role in the toxicity of inorganic arsenic.

While the general metabolic pathway of inorganic arsenic is common to most organisms, some interspecies differences exist with respect to the efficiency with which inorganic arsenic is metabolized and excreted in urine. In humans, following exposure to inorganic arsenic, 40% to 70% of the dose is absorbed, processed, and excreted within 48 hours (Vahter, 1983). In mice and rabbits, a greater fraction (75-95%) of the ingested dose is excreted within 48 hours than in humans, whereas in rats, much less (5-20%) of the dose is excreted during the same time period (Vahter, 1983).

The extent to which inorganic arsenic is metabolized is determined through examination of the relative amounts of arsenic metabolites in urine. In most animal species, including humans, the intracellular metabolism of inorganic arsenic involves extensive metabolism to DMA^sup V^(Vahter, 1994). Humans also excrete relatively large amounts of MMA^sup V^. Urine in humans typically contains 10-20% inorganic arsenic, 10-20% MMA^sup V^, and 60-80% DMA^sup V^ (Vahter, 1994), indicating relatively efficient methylation. The rat is the only species that excretes significant amounts of TMAO (Cohen et al., 2002b; Yoshida et al., 1998).

In rats, unlike in other species, clearance of inorganic arsenic is particularly slow. The slow clearance is due to accumulation of DMA in the red blood cells after biotransformation of inorganic arsenic in the liver (Shiobara et al., 2001; Vahter, 1983, 1999; Winski and Carter, 1995). Retention in rat red blood cells appears to be mediated primarily by binding of trivalent arsenicals, especially DMA^sup III^, to hemoglobin (Lu et al., 2004; Winski and Carter, 1995). In addition, the rat retains significantly more MMA^sup III^ in its biliary excretions than other species, allowing for enhanced retention and extensive metabolism by the liver (Csanaky and Gregus, 2002).

Despite the fact that methylation is generally considered to enhance excretion of arsenic compounds, species that do not methylate inorganic arsenic (i.e., guinea pigs [Csanaky and Gregus, 2002; Healy et al., 1997], and marmoset monkeys and chimpanzees [Vahter, 1999; Vahter et al., 1982]) do not retain more arsenic in the body than species that do methylate inorganic arsenic. Thus, factors other than methylation may influence excretion of arsenic compounds, at least in these species (Vahter, 1999; Vahteretal., 1982).

Advances in analytical methodology have allowed for the detection of MMA^sup III^ and DMA^sup III^ in urine (Le et al., 2000a). Investigators initially identified trivalent arsenic compounds in the urine of populations chronically exposed to drinking water containing high concentrations of inorganic arsenic by treating subjects with 2,3-dimercaptopropane-l-sulfonic acid (DMPS), a chelating agent that enhances the release of the trivalent species from the body by efficiently sequestering metals in the trivalent state (Aposhian, 1998). For example, in a study of individuals chronically exposed to inorganic arsenic in drinking water, treatment with DMPS increased urinary arsenic excretion five-fold (Aposhian et al., 200Ob). DMPS treatment also altered the proportion of arsenic metabolites in urine; the percentage of MMA^sup V^ in urine increased from 15% to 45%, and DMA^sup V^ levels sharply decreased. MMA^sup III^ was detected in the urine of 18 of 20 arsenicexposed subjects following treatment with DMPS (Aposhian et al., 200Ob). Using a different analytical methodology, Le et al. (200Ob) analyzed urines from a population in Inner Mongolia and detected MMA^sup III^ in 51 of 164 urine samples and DMA^sup III^ in 2 of the 164 samples. TMAO was not detected (Le et al., 200Ob).

Methylated compounds of trivalent arsenic have been detected in the urine of populations chronically exposed to inorganic arsenic, even without DMPS pretreatment. Aposhian et al. (200Oa) found that urine from an arsenic-exposed population in Romania contained 7-11% MMA^sup III^ and 13-14% MMA^sup V^. The percentage of subjects with detectable urine levels of MMA^sup III^ increased with arsenic exposure, but even in the high dose group (161 g/L arsenic in drinking water), only 30% of the subjects had detectable levels of MMA^sup III^ (Aposhian et al., 2000a). Similar levels of MMA^sup III^ were found in urine collected from arsenic-exposed populations in Mexico and West Bengal, India, but the incidence of individuals with detectable levels varied (Del Razo et al., 2001; Mandai et al., 2001, 2004). For example, Mandai et al. (2001) detected MMA^sup III^ in 48%, and DMA^sup III^ in 72% of exposed individuals. However, the validity of the analytical method used in some of these studies has recently been questioned (Hansen et al., 2004).

TMAO has generally not been detected in human urine (Le et al., 200Ob). It was reported in one study in which one individual (42 year old male, 80 kg) was exposed to an extremely high oral dose of DMA^sup V^, containing 8 mg radiolabeled arsenic (74As)/kg (Marafante et al, 1987). In contrast, TMAO is a major metabolite in rat urine, especially after DMA^sup V^ administration. TMAO was found also in the urine of mice and hamsters f\ollowing administration of DMA^sup V^, but at considerably lower concentrations than in rats (Marafante et al., 1987; Vahter et al., 1984; Yamauchi and Yamamura, 1984).

FIG. 2. Intracellular metabolism of inorganic arsenic.

B. Disposition and lntracellular Metabolism of Trivalent Versus Pentavalent Inorganic Arsenic

Cellular uptake of trivalent inorganic arsenic (arsenite) is more extensive than that of pentavalent inorganic arsenic (arsenate) (Bertolero et al., 1987; Delnomdedieu et al., 1995; Fischer et al., 1985; Georis et al., 1990; Lerman et al., 1983; Mles et al., 1998; Styblo et al., 1999b) (Figure 2). As^sub i^^sup V^ is hardly taken up into cells. When As^sub i^^sup V^s enters the bloodstream it is quickly reduced to As^sup III^^sub i^ (Gregus and Nerneti, 2005; Nerneti and Gregus, 2005), which is taken up extensively into cells. In a human oral epidermal carcinoma cell line, cellular uptake of As^sup III^^sub i^ appeared to involve simple diffusion, while As^sub i^^sup V^ uptake involved an energydependent transport system similar to that of phosphate uptake (Huang and Lee, 1996). The lower cellular uptake of pentavalent arsenic as compared to trivalent arsenic compounds is an important factor in the lower cytotoxicity of pentavalent arsenic compounds (Styblo et al., 2000; Vega et al., 2001 ). Furthermore, trivalent inorganic arsenic is a preferred substrate for arsenite methyltransferase over pentavalent inorganic arsenic. In an in vitro assay system involving rat liver cytosol, 92% of As^sup III^^sub i^, but only 33% of As^sub i^^sup V^, was converted to DMA^sup V^ after a 90-minute incubation (Styblo et al., 1995). Thus, trivalent inorganic arsenic is metabolized more extensively than pentavalent inorganic arsenic, due to both extensive cellular uptake and higher affinity for arsenite methyltransferase (Aposhian, 1997, 1998; Styblo et al., 2000).

C. Metabolism and Disposition of MMA^sup V^ In Vivo

The results of in vivo metabolism studies with MMAV in various animal species are presented in Table 1. In striking contrast to the significant metabolism of exogenous inorganic arsenic in humans, methylated arsenicals undergo limited further metabolism (i.e., reduction and oxidative methylation) in most species (except rat) and are excreted in the urine largely as the parent compound. MMA^sup V^ that is directly administered (orally or intravenously) to humans, mice, rabbits, sheep, goats, or hamsters, is readily absorbed into the body, and rapidly excreted, with limited further metabolism (Buchet et al., 1981; Hughes and Kenyon, 1998; Jaghabiret al., 1991 ; Shariatpanahi andAnderson, 1984; Yamauchietal., 1988). Consequently, only small amounts of DMA^sup V^ are found in the urine after MMA^sup V^ exposure. For example, within 4 days of exposure, human volunteers excreted in their urine 78% of a single oral dose of MMA^sup V^, of which 87% was unchanged MMA^sup V^ and only 13% was converted to DMA^sup V^ (Buchet et al., 1981). In mice, 73-78% of a single intravenous dose of MMA^sup V^ was excreted in the urine within 24 hours, and only 2-8% of the dose was further metabolized to DMA^sup V^. The remaining 92-98% of the dose was excreted as unchanged MMA^sup V^ (Hughes and Kenyon, 1998). In hamsters, 4% of an MMA^sup V^ dose was methylated and excreted as DMA^sup V^, and only 0.3% was further methylated and excreted as TMAO (Yamauchi et al., 1988). Rats are unique in that they methylate MMA^sup V^ more extensively than other species. When rats were exposed to MMA^sup V^ in drinking water for 7 months, their urine contained 65% MMA^sup V^, 27% DMA^sup V^, 4.1% TMAO, and 0.1% tetramethylarsonium (Yoshida et al., 1998). In a recent publication, Hughes et al. (2004) reported that mice administered 0.4 mg MMA^sup V^/kg excreted mostly unchanged MMA^sup V^ and hardly any DMA^sup V^ or DMA^sup III^, whereas administration of 0.4 mg/kg of trivalent MMA^sup III^ resulted in extensive generation and excretion, in urine, of DMA^sup V^ and some DMA^sup III^ (Hughes et al., 2004).

TABLE 1

Results of in vivo studies of the metabolism of exogenous MMA^sup V^

D. In Vitro Studies for lntracellular Metabolism of MMA

The in vivo studies just described show that orally administered (or "exogenous") MMA^sup V^ is excreted in urine largely unchanged, and that only a small fraction of MMA^sup V^ administered orally undergoes further methylation to DMA^sup V^ and TMAO. In vitro studies provide a mechanistic basis for these findings and show that the limited methylation is due to a combination of poor cellular uptake and limited intracellular metabolism.

Cellular uptake of MMA^sup III^ was severalfold greater than that of MMA^sup V^ in the in vitro studies using rat and human liver cells (Styblo et al., 1999a). In rabbit erythrocytes, cellular uptake of MMA^sup V^ was less than that of inorganic arsenic compounds, but more than that of DMA^sup V^ (Delnomdedieu et al., 1995). In peritoneal macrophages, isolated from CDFl mice, the cellular uptake of MMA^sup V^ was reported to be about five times less than uptake of DMA^sup V^ or TMAO (Sakurai et al., 1998). In general, the uptake of the methylated compounds of pentavalent arsenic is less efficient than that of inorganic arsenic compounds and of that of methylated compounds of trivalent arsenic. The uptake of the specific compound depends on the specific cell type.

Two major enzymes are involved in the metabolism of MMA: MMA- reductase and MMA-methyltransferases. The reduction of MMA^sup V^ to MMA^sup III^ by MMA-reductase was found to be the rate-limiting step in the methylation and metabolism of inorganic arsenic. The affinity of MMA-reductase for MMA^sup V^ (K^sub m^ =2.16 10^sup -3^) in homogenized rabbit liver is lower than the affinity of methyltransferases to either arsenite (K^sub m^ = 5.5 10^sup -6^) or MMA^sup III^ (9.2 10^sup -6^) by over two orders of magnitude (Sampayo-Reyes et al., 2000; Zakharyan et al., 1999b). Based on the K^sub m^ values a concentration of MMA^sub V^ in the millimolar range needs to be present before a significant amount of MMA^sup V^ can be reduced to MMA^sup III^ (Zakharyan et al., 1999b).

The affinity of MMA-methyltransferase to trivalent compounds of arsenic is also higher than to the pentavalent compounds. MMA- methyltransferases from rabbit liver or Chang human hepatocytes were shown by Zakharyan et al. (1999a) to have a higher affinity for MMA^sup III^ than for MMA^sup V^. In an in vitro assay system of rat liver cytosol, 99.6% of MMA^sup III^, but only 3% of MMA^sup V^, was converted to DMA^sup V^ (Styblo et al., 1995). Arsenite is also a preferred substrate for methyltransferase. In hamster, the K^sub m^ of MMA-methyltransferase has been demonstrated to be 450-fold higher than the K^sub m^ of arsenite-methyltransferase (Wildfang et al., 1998).

E. The Metabolism and Disposition of DMA^sup V^

1. Animals (Except Rats)

Similarly to MMA^sup V^, in most experimental animals, DMA^sup V^ that is directly administered (via oral or intravenous route) is excreted mainly unchanged, with very little or no metabolism to TMAO. The first reports on the metabolism of exogenous DMA^sup V^ were published in the early 1980s (Vahter, 1983; Vahter et al., 1984; Yamauchi and Yamamura, 1984). Hamsters given an oral dose of 50 mg DMA^sup V^/kg body weight (i.e., 1440 g As per animal) excreted 80% of the dose (45% in the urine and 34.7% in the feces) within 24 hours. Approximately 65% of the total dose was excreted as unchanged DMA^sup V^ and approximately 15% was excreted as a trimethylarsenic compound (Yamauchi and Yamamura, 1984).

Mice and hamsters administered radiolabeled DMA^sup V^ (40 mg As/ kg body weight) retained less than 1 % of the dose two days after administration. Of the 99% of the dose excreted, mice excreted 69% in the urine and 29% in the feces, whereas hamsters excreted 57% in the urine and 42% in the feces. In both species, about 80-85% of the dose was eliminated in the form of unmetabolized DMA^sup V^. In mice and hamsters, only 3.5% and 6.4% of the dose was excreted in the form of TMAO, respectively (Marafante et al., 1987). Similarly, over 80% of the total DMA^sup V^ administered to mice in a single intravenous dose was excreted within 24 hours, primarily as unchanged DMA^sup V^. A small amount of the dose was detected in urine as an "unstable DMA* complex" (Hughes and Kenyon, 1998). In a more recent study, Hughes et al. (2000) did not detect any methylated or demethylated DMA^sup V^ metabolites in the blood or in any organ of mice treated intravenously with DMA^sup V^ up to 8 hours after the administration of radiolabeled DMA^sup V^. The results of in vivo metabolism studies with DMA^sup V^ in various animal species, as evaluated by the relative amounts of arsenic compounds in urine, are summarized in Table 2.

2. Rats

Rat metabolism of DMA^sup V^, as well as rat metabolism of other arsenic compounds, differs from that of other species in a number of respects. Following exposure to DMA^sup V^, rats retain DMA more than any other species, although the amount of the retained DMA is lower than the amount retained after exposure of rats to inorganic arsenic because of the faster excretion of DMA^sup V^ (Lu et al., 2004; Winski and Carter, 1995). Additionally, the rat metabolizes DMA^sup V^ to TMAO more extensively than other species (Figure 3). Yoshida et al. (1997) demonstrated that following administration of a single dose of radiolabeled DMA^sup V^ to rats, most of the dose was excreted within the first 10 hours, with 80% of the dose excreted within the first 4 hours after administration. The major form of arsenic excreted during the first 4 hours was unmetabolized DMA^sup V^, but the portion of DMA^sup V^ decreased with time, and TMAO accounted for over 50% of all arsenic excreted between 6 and 24 hours after administration. Following intraperitoneal administration, the proportions of DMA^sup V^and TMAO found in the urine were similar, but excretion was more rapid (Yoshida et al., 1997). Yoshida et al. ( 1998) demonstrated that one week after exposure to DMA^sup V^ (100 mg As/L) in drinking water, 44.9% of all the arsenic excreted was eliminated as unchanged DMA^sup V^, with 40% excreted in the form of further methylated metabolite TMAO and 0.4% as TMA. After 7 months of exposure, the proportion of DMA^sup V^ excreted in the urine increased to 60.7%, and TMAO decreased to 23.9% (Yoshida et al., 1998). When NCI-Black-Reiter (NBR) rats were exposed for 4 weeks to drinking water containing 0.05% N-butyl-N-(4- hydroxybutyl)nitrosamine (BBN), followed by exposure for 32 weeks to drinking water containing 100 ppm DMA^sup V^, the concentrations of arsenic metabolites measured in urine were: 31.1 mg/L DMA^sup V^, 17 mg/L TMAO, 7.4 mg/L of two unidentified arsenic compounds, 0.15 mg/ L arsenobetaine (AsBe), 0.09 mg/L As^sub i^^sup V^, 0.05 mg/L MMA^sup V^, and <0.01 mg/L As^sup III^^sub i^, (Li et al., 1998). In a similar experiment that exposed rats to 0.05% BBN in the drinking water for 4 weeks, followed by exposure to only 50 ppm DMA^sup V^ in the drinking water for 32 weeks, the concentrations of urinary metabolites were 10.9 mg/L DMA^sup V^, 5.7 mg/L TMAO, 0.3 mg/L AsBe, 1.3 mg/L As^sup III^^sub i^, and 1.2 mg/L unidentified metabolites (Wanibuchi et al., 1996). The relatively high level of As^sup III^^sub i^ reported in this study probably was in error, based on later analytical improvements developed in this laboratory that were reported by Yoshida et al. (1998). Analysis of 24-hour urine samples collected from rats fed 100 ppm DMA^sup V^ in the diet showed that the major metabolites were DMA^sup V^ (66.4 2.7 M) and TMAO (73.2 9.5 M) (Cohen et al., 2002a), and analysis of freshly voided urine collected from female rats treated with 2,10,40, or 100 ppm DMA^sup V^ showed that urinary levels of DMA^sup V^ and TMAO increased with the dose of administered DMA^sup V^ (Arnold et al., 2003a). When special precautions were taken to preserve the unstable methylated trivalent arsenicals, significant amounts of DMA^sup III^ were also detected in rat urine after feeding the rats with diet containing 100 ppm of DMA^sup V^ (Cohen et al., 2002b).

TABLE 2

Results of in vivo studies of the metabolism of exogenous DMA(v)

TABLE 2

Results of in vivo studies of the metabolism of exogenous DMA(v)

FIG. 3. Relative percent of metabolites found in the urine of various species 48 hours after an exogenously administered dose of DMA^sup V^. a, Dose 50 mg/kg body weight (Yoshidaet al., 1997); 4.0% of metabolities excreted was reported to be AS^sub i^. This is probably in error, based on later analytical improvements developed in this laboratory and reported by Yoshida et al. (1998). Percent of total dose excreted in the urine not calculated, b, Dose 40 mg As/ kg body weight administered as DMA^sup V^ (Marafante et al., 1987); 57% of total dose excreted in the urine. c, Dose 40 mg As/kg body weight administered as DMA^sup V^ (Marafante et al., 1987); 69% of total dose excreted in the urine. d, Dose 0.1 mg As/kg body weight administered as DMA^sup V^ (Marafante et al., 1987); 84% of total dose excreted in the urine.

The unique rat metabolism of DMA^sup V^, which results in the generation of high levels of TMAO (Vahter, 1999), is consistent with the rat being particularly susceptible to DMA^sup V^-induced cytotoxicity and tumorigenicity. The rat is the only known species in which DMA^sup V^ administration results in high urinary concentrations of TMAO. During the formation of TMAO the highly reactive DMA^sup III^ is formed as a metabolic intermediate. The rat is also the only known species in which DMA^sup V^ administration results in urinary concentrations of DMA^sup III^ that are equivalent to DMA^sup III^ concentrations that are cytotoxic to urothelial cells in vitro.

Demethylation of DMA^sup V^ to arsenite has not been observed in vivo in rat studies (Stevens et al., 1977; Vahter et al., 1984; Yamauchi and Yamamura, 1984) except for Yoshida et al. (1997), who reported observations of low levels (<1% of the administered dose) of inorganic arsenic in the urine of rats administered a single dose of DMA^sup V^. The authors of this single report noted that demethylation was not immediate, but occurred over time. They attributed the demethylation phenomenon to the presence of intestinal microorganisms and to an artifact of the urine collection method (Yoshida et al., 1997). However, based on method refinements in a subsequent report (Yoshida et al., 1998), they later noted that the presumed arsenite from their previous study was not arsenite but was tetramethylarsonium, highlighting that the difficulties associated with certain analytical methodologies can lead to erroneous conclusions. Any reports of demethylation in mammalian organisms require careful and thorough evaluation of the analytical methodology. It is generally accepted that the arsenic-carbon bond is quite strong and is not broken during mammalian metabolism (Cullen and Reimer, 1989).

3. Humans

In human volunteers, 75% of a single oral dose of DMA^sup V^ was excreted in urine as unchanged DMA^sup V^ within 4 days. There was no evidence for further methylation or demethylation (Buchet et al., 1981). There is only a single report in the literature of TMAO detection in human urine. Following exposure of one individual (42- year-old male, 80 kg) to an extremely high oral dose of DMA^sup V^ containing 8 mg radiolabeled arsenic (74As) (Marafante et al., 1987), 80% of the dose was excreted in the urine as DMA^sup V^ within 3 days of administration, and 4% was reported to be TMAO. More recently, using a very sensitive analytical method with a detection limit of 2 gfL, Le et al. (2000b) analyzed 164 urine samples collected from 41 people in Inner Mongolia, China, exposed to inorganic arsenic in the drinking water, at levels of 510-660 g/ L. No TMAO was detected in any of these samples, and DMA^sup III^ was reported in only two samples (Le et al., 2000b). Using the same analytical method, Lu et al. (2003) detected TMAO in urine from rats fed DMA^sup V^. In a recent analysis of leukemic patients administered pharmacologie doses of arsenic trioxide (0.15 mg/kg), less than 0.1 % of the dose was excreted as TMAO (Wang et al., 2004; personal communication, X. Chris Le). Another case of TMAO detection in human urine was reported by Francesconi et al. (2002), who exposed a single volunteer to 1220 g arsenic in the form of dimethylated arsenosugar. However, the relevance of this finding with respect to ingestion of DMA^sup V^ by humans is questionable.

F. In Vitro Studies for lntracellular Metabolism of DMA^sup V^

Differences between animal species in uptake of the pentavalent (DMA^sup V^) and the trivalent (DMA^sup III^) dimethylated compounds of arsenic into red blood cells have been investigated in vitro as a possible explanation for species differences in retention and metabolism of DMA^sup V^. In most animal species, DMA^sup V^ does not enter the cells and is excreted unchanged (Hughes and Kenyon, 1998; Marafante et al., 1987; Vahter, 1983; Vahteretal., 1984; Yamauchi and Yamamura, 1984). Delnomdedieu et al. (1995) found that cellular uptake of DMA^sup V^ in rabbit erythrocytes was 44%, compared to 81-88% uptake of inorganic arsenic compounds and 65% uptake of MMA. These authors also determined that there was no further methylation of DMA^sup V^ in rabbit erythrocytes. Shiobara et al. (2001) examined the uptake of DMA^sup V^ and DMA^sup III^ into the red blood cells of the rat, hamster, mouse, and human. They found that DMA^sup III^ was taken up more efficiently than DMA^sup V^ in all tested species, and that DMA^sup III^ was taken up most efficiently in the rat cells and least efficiently in the human cells. Intracellular retention of DMA^sup III^ was also shown to be the highest in the rat red blood cells compared to the human or the hamster, primarily because of the much greater binding of DMA^sup III^ to sulfhydryl groups of rat hemoglobin compared to other species (Lu et al., 2004).

FIG. 4. Relative percentages of different arsenic compounds found in the urine of humans four days after exogenous ingestion of 500 g arsenic as either sodium arsenite (As^sub i^), MMA^sup V^, or DMA^sup V^ (Buchet et al., 1981): a, 45.1% of total dose excreted in urine: b, 78.3% of total dose excreted in urine; c, 75.1% of total dose excreted in urine.

In addition, the efficiency of the oxidation of DMA^sup III^ to DMA^sup V^ is in the order hamster > human > rat. Thus, the concentrations of DMA^sup III^ in the rat red blood cells are highest, due to higher uptake, higher retention, and lower oxidation. The authors concluded that differences between animal species in uptake of DMA^sup V^ by red blood cells contribute to differences in their reduction and methylation capacity.

In summary, inorganic arsenic, particularly the trivalent form, is efficiently transported into cells where it is reduced and methylated sequentially to MMA^sup V^ and DMA^sup V^. Methylation of inorganic compounds has been shown to facilitate excretion of arsenic compounds in vivo (Gebel, 2002; Vanter, 1994, 1999: Vahter and Concha, 2001 ), although the importance of methylation for elimination of inorganic arsenic remains unclear, given its ready excretion in some species that do not methylate arsenic (Csansky and Gregus, 2002; Healy et al., 1997). The relatively large amounts of DMA^sup V^ measured in human urine after ingestion of inorganic arsenic are indicative of the extensive metabolism of inorganic arsenic.

Directly ingested (exogenous) MMA^sup V^ and DMA^sup V^ have a different disposition than MMA^sup V^ and DMA^sup V^ that are produced in the cells during the metabolism of inorganic arsenic in all tested species, including humans (Figure 4). In contrast to methylated comp\ounds generated endogenously, exogenous MMA^sup V^ and DMA^sup V^ do not readily enter the cell and are excreted mostly unchanged, with only a small fraction undergoing further methylation, the amount of which depends on the animal species. In view of these metabolic differences between exogenous and endogenous MMA^sup V^ and DMA^sup V^, information on the toxicity of MMA^sup V^ and DMA^sup V^ produced from metabolism of inorganic arsenic should not be used to assess the toxicity of exogenous MMA^sup V^ and DMA^sup V^.

The differences in metabolism and disposition of inorganic and organic arsenic among different animal species should be considered when interpreting toxicity and carcinogenicity studies of arsenic compounds ( Aposhian, 1997; Thomas et al., 2001). In particular, the metabolism of arsenic compounds in the rat is unique due to accumulation of DMA in rat erythrocytes (leading to much slower arsenic excretion than in humans and other species), and the greater ability of the rat to metabolize DMA^sup V^ to TMAO. Therefore, studies in rats have limited relevance to other animal species and specifically to humans when evaluating risks associated with exposure to DMA^sup V^. Given the striking differences in metabolism and disposition of arsenicals in general, Aposhian (1997) has questioned the relevance of arsenic studies in rats to humans.

III. CARCINOGENICITY OF ARSENIC

A. Carcinogenicity of Inorganic Arsenic

Adverse health effects from ingestion of high levels of inorganic arsenic in drinking water have been observed in populations in several countries, including Taiwan (Guo, 2004; Wu et al., 1989), Chile (Ferreccio et al., 2000), Bangladesh (Rahman et al., 2001 ), and Inner Mongolia (Tucker et al., 2001). Specifically, increased incidences of cancer of the skin, bladder, and lung resulted from exposure to concentrations of several hundred micrograms inorganic arsenic per liter drinking water (for reviews see Brown and Ross, 2002; Schoen et al., 2004). For example, Wu et al. ( 1989) observed increased risks for skin, bladder, and lung cancer in males from a population in Taiwan drinking 600 g/L of arsenic in water (reference group <300 g/L). Chiou et al. (2001) found elevated risks for skin and bladder cancer in Taiwanese males exposed to drinking water containing arsenic concentrations higher than 100 g/L (the highest group tested, which included concentrations up to 3590 g/L). Recently, Guo (2004) reported elevated lung cancer mortality in a Taiwanese population of males and females exposed to drinking water containing concentrations of arsenic higher than 640 g/L. In the United States, however, studies have not found an association between cancer and ingestion of drinking water containing inorganic arsenic up to average levels of 190 g/L (the highest concentration tested) (Lewis et al., 1999; NRC, 1999; Schoen et al., 2004).

Most epidemiologic studies use measurements of total inorganic arsenic, combining arsenite and arsenate. This is reasonable since most ingested arsenate is rapidly reduced to arsenite (Aposhian, 1997; Cullen and Reimer, 1989). For essentially the same reason, most in vivo and in vitro studies evaluating the toxicity or carcinogenicity of inorganic arsenic use arsenite rather than arsenate.

In contrast to the definitive human data, results from animal studies are more ambiguous (Huff et al., 2000; Wang et al., 2002; Wanibuchi et al., 2004). Animal studies done several years ago examining long-term exposure to high doses of inorganic arsenic, as sodium arsenite (up to 250 mg/L) or sodium arsenate (up to 400 mg/ L), were negative in rats, mice, beagles, and cynomologus monkeys, indicating inorganic arsenic by itself is not a carcinogen in animals (Kitchin, 2001; NRC, 1999). Two-year bioassays using today's standards have not been performed. It is noteworthy that Waalkes et al. (2003) showed that exposure to sodium arsenite in pregnant mice during gestation days 8-18 at doses of 42.5 and 85 mg/L in drinking water resulted in a dose-dependent increase in tumors of the liver and adrenal gland in male offspring and of the ovary in female offspring. When the mice were treated with dermal tetradecanoylphorbol acetate (TPA) following the transplacental arsenic exposure, lung lesions were slightly increased in males and females (Waalkes et al., 2004b). Like most of the studies with arsenic compounds, the doses in these studies were very high compared to human exposures. In addition, strain and species specificity of these transplacental effects has yet to be determined. A later study by the same researchers suggests that this transplacental effect may be secondary to alterations in estrogen regulation rather than a direct effect of the arsenic (Waalkes et al., 2004a). If the tumors induced transplacentally were directly due to the estrogen rather than the arsenic, the application of this model to arsenic-induced skin, lung, or bladder cancer in humans may be limited. Given that so far this is the only animal model of inorganic arsenic carcinogenesis, it is essential that additional research be performed to determine possible mechanisms.

Inorganic arsenic (arsenite) administered in the drinking water (0.01%) to C57BL DBA2 mice for 4 weeks produced evidence of urinary bladder urothelial simple hyperplasia (Luster and Simeonova, 2004; Simeonova et al., 2000). Longer term studies have not been reported. Two-year bioassays in mice are required (Huff et al., 2000) to delineate the implication of this important observation for bladder carcinogenesis.

Other studies examining the carcinogenic potential of inorganic arsenic have demonstrated that As^sub i^ must be administered in combination with other carcinogenic factors to elicit a tumorigenic response, suggesting that As^sub i^ is acting as a cocarcinogen. Specifically, Germolec et al. (1998) and Moser et al. (2002) demonstrated that sodium arsenite could act as a copromoter or cocarcinogen in the development of skin tumors in Tg.AC mice, which carry the H-ras mutation with a zetaglobin promoter, when they were treated with low doses of dermal TPA and sodium arsenite in water. However, the validity of the Tg. AC model to evaluate carcinogenic activity has been questioned by the International Agency for Research on Cancer (IARC, Working Group, 1999), the National Toxicology Program (NTP, 2002), and others (Goodman, 2001).

Rossman et al. (2001) found that treatment with sodium arsenite in water increased the incidence of skin tumors induced by ultraviolet radiation in hairless mice. Treatment with arsenite alone did not cause skin tumors in hairless mice, indicating that the inorganic arsenic alone is not carcinogenic and may require other factors to cause cancer. Subsequently, the researchers also demonstrated that the increase in skin tumors was dosedependent (Burns et al., 2004).

Oral administration of arsenite to p53(+/-) knockout mice did not show an increased incidence of tumors at any site. No increase was seen even in the lymphomas and sarcomas that are known to commonly occur spontaneously in these mice (Popovicova et al., 2000). In addition, sodium arsenite had no effect on skin tumors in this model when co-administered with 4-vinyl-one-cyclohexene diepoxide or on bladder tumors when co-administered with p-cresidine (Popovicova et al., 2000).

B. MMA^sup V^ Carcinogenicity Studies

Long-term animal bioassays of MMA^sup V^ provide clear evidence that MMA^sup V^ is not carcinogenic. In 2-year cancer bioassays in rats and mice, dietary MMA^sup V^ did not induce any tumors at doses up to and above the maximum tolerated dose of 400 ppm in feed (Arnold et al., 2003b), which is comparable to an intake of arsenic of 83 mg/kg/day in male mice, 104 mg/kg/day in female mice, 93 mg/ kg/day in male rats, and 101 mg/kg/day in female rats. In another 2- year bioassay, male rats were treated with 50 and 200 ppm MMA^sup V^ in the drinking water (Shen et al., 2003). There was a statistically significant increase in hyperplasia in the bladder epithelium at 50 and 200 ppm and a statistically significant increase in the numbers and areas of glutathione S-transferase placental form positive foci in the liver at 200 ppm. No MMA^sup V^ treatment-related tumors were found.

The lack of Carcinogenicity of MMA^sup V^ in the rat is consistent with the limited amount of DMA^sup V^ formed from exogenously administered MMA^sup V^. Susceptibility of other species would be expected to be even less, given their limited conversion of exogenously ingested MMA^sup V^ to DMA^sup V^.

MMA^sup V^, DMA^sup V^, or TMAO administered at 100 ppm in the drinking water increased rat liver foci number and area in the Ito medium term assay, which involves a complicated protocol lasting 8 weeks, including pretreatment with N,N-diethylnitrosamine (DEN) and performance of a partial hepatectomy during the time the test agent is administered (Nishikawa et al., 2002). The complexities inherent in the assay preclude obtaining meaningful mechanistic information, and the relevance of this assay to carcinogenesis in humans is unknown.

C. DMA^sup V^ Carcinogenicity Studies

DMA^sup V^ induced bladder tumors in rats when administered at high doses (100 ppm) in the diet in a two-year bioassay (van Gemert and Eldan, 1999; Arnold et al., submitted) (Table 3). The increased incidence of bladder tumors was statistically significant in female rats. The small increase in tumors in male rats was not statistically significant, but was likely treatment related, since spontaneous bladder tumors in this strain of rats are rare and there was a significant number of male rats with urothelial hyperplasia. Treatment-related tumors were not found in the lower dose groups (2, 10, and 40 ppm), but there was a statistically significant increase in hyperplasia of the bladder urothelium at 40 ppm in female and male rats. There was no evidence of a significant increase in hype\rplasia of the urinary bladder at 10 ppm, which is equivalent to an approximate dose of 0.79 mg/kg/day in females and 0.73 mg/kg/ day in males. The female rat appeared to be more sensitive to the urothelial effects of DMA^sup V^ than the male. The role of estrogen in this difference in response between sexes has not been investigated. Given the effects of estrogen demonstrated in the mouse transplacental model of inorganic arsenic carcinogenesis, this appears to be an important issue for further investigation. The study demonstrated that increases in female rat bladder tumors occurred only at 100 ppm in feed, which is approximately equivalent to 8 mg/kg/day. Wei et al. (2002) showed that DMA^sup V^ administered to male rats in drinking water over a period of 104 weeks resulted in an increase in bladder tumors only in the high dose groups (i.e., 50 and 200 ppm). There was no effect at 12.5 ppm. Female rats were not tested in this study. These two bioassays are consistent with the hypothesis that the dose response relationship of the DMA^sup V^-induced carcinogenicity in rats is nonlinear; tumors occur only at relatively high doses.

Administration of DMA^sup V^ in the diet at doses up to 500 ppm for 2 years had no carcinogenic effects in B6C3F1 mice at any site (van Gemert and Eldan, 1998). There was a slight increase in subcutaneous sarcomas, but the increase was not statistically significant, and it was later determined by a pathology working group that the lesions were not treatment related (Arnold et al., submitted). In addition, there was no evidence of urothelial hyperplasia in the mice at any dose.

In other studies, researchers have administered DMA^sup V^ concurrently with or sequentially after administration of other tumorigenic compounds or in animals genetically designed to be susceptible to arsenic toxicity (Kitchin, 2001 ; Wanibuchi et al., 2004). For example, Yamamoto et al. (1995) reported that after pretreatment of rats with five DNA-reactive carcinogens, DMA^sup V^ (100 ppm in drinking water) acted as a promoter, significantly increasing the incidence of bladder, kidney, liver, and thyroid gland tumors. No effect was detected with DMA^sup V^ alone, without pretreatment with the five carcinogens. The use of five different DNA-reactive chemical carcinogens to initiate the process, however, constitutes a highly unusual study design, the significance of which is very difficult to interpret with respect to human exposures. In other studies, DMA^sup V^ administered for short periods of time in rats has caused changes in liver (Wanibuchi et al., 1997) or kidney (Murai et al., 1993), but not tumor formation.

Similarly, no preneoplastic lesions or tumors were observed in rats treated with DMA^sup V^ alone by Wanibuchi et al. (1996). However, this study was relatively short, approximately one-half year. In contrast, DMA^sup V^ (50 or 200 ppm in drinking water) promoted bladder tumors in male rats that were pretreated with the genotoxic bladder carcinogen BBN (Chen et al., 1999; Wanibuchi et al., 1996). No effect was detected at a dose of 12.5 ppm DMA^sup V^.

TABLE 3

Key carcinogenicity and mode of action studies with DMA^sup V^

Yamanaka et al. (1996) administered 200 or 400 ppm DMA^sup V^ in the drinking water to ddY mice after pretreatment with the DNA reactive carcinogen 4-nitroquinoline. N-oxide (NQO). The only significant effect was an increase in the number of lung tumors per mouse at the highest dose. There was no statistically significant increase in the incidence of mice with lung tumors.

Hayashi et al. (1998) reported that DMA^sup V^ caused lung tumors in AJ) mice, a strain highly susceptible to lung tumorigenesis. The only statistically significant endpoint was the number of tumors per mouse at the highest dose (400 ppm in drinking water). Dose-related trends in the number of tumor-bearing mice, total number of tumors, and size of tumors were not statistically significant. The authors of the study concluded that their results were not definitive, and that "Further studies using a larger number of animals, including other strains of mice and other species, are required to conclusively demonstrate the carcinogenic potential of DMA^sup V^."

Morikawa et al. (2000) suggested that DMA^sup V^ acted as a skin tumor promoter in K6/ODC transgenic mice, accelerating the induction of 7,12-dimethylbenz[a]anthracene(DMBA)-induced skin tumors in mice. The K6/ODC transgenic mouse strain is very sensitive to skin tumor induction. In fact, all of the mice pretreated with DMBA developed skin papillomas, a type of benign tumor, within 20 weeks of the study, whether or not the mice were subsequently treated with DMA^sup V^. DMA^sup V^ treatment resulted in a decrease in latency for the papillomas by only a few weeks and only a slight increase in the number of tumors compared to DMBA alone. Furthermore, DMA^sup V^ alone had no effect even though the daily amount of DMA^sup V^ administered to mouse skin was very high (3.6 mg).

Yamanaka et al. (2000, 2001a) administered DMA^sup V^, 400 or 1000 ppm in drinking water, to female hairless HR-I mice for 50 weeks, after pretreatment with DMBA. Although the incidence of skin tumors was increased, the effect was not dose dependent, and ultraviolet B (UVB) radiation administered with DMA^sup V^ did not greatly enhance the response. These doses of DMA^sup V^ are extraordinarily high compared to other studies. In another study, DMA^sup V^ in drinking water did not induce lung cancer in rats when administered after N-bis(2-hydroxypropyl)nitrosamine (DHPN), even at levels as high as 200 mg/L (Seike et al., 2002).

No significant difference was observed in tumor incidence (neither site specific nor overall tumor incidence) when DMA^sup V^ was administered for 80 weeks at 50 or 200 ppm in the drinking water to male p53(+/-) mice with a C57BL/6J background (Salim et al., 2003). The authors note a decrease in latency for all tumors combined, but not for individual tissue sites. However, combining tumor sites is considered inappropriate for evaluating carcinogenesis studies as discussed by the NTP (Haseman et al., 1986). No significant tumorigenic findings were detected in this same study for the wild type, parental strain. In both the wildtype and knockout strains, no effect was observed on the lower urinary tract, skin, or lungs.

In conclusion, based on the collective scientific literature, the carcinogenic properties of inorganic arsenic, MMA^sup V^, and DMA^sup V^ are distinct. Based on epidemiological data, inorganic arsenic is carcinogenic in humans, but long-term treatment with high doses of inorganic arsenic has not been tumorigenic in experimental animals. MMA^sup V^ is not carcinogenic either in rats or in mice in 2-year bioassays. DMA^sup V^ is carcinogenic only in the rat urinary bladder after administration of relatively high doses in the diet or in the drinking water, but is not carcinogenic in mice in a 2-year bioassay.

IV. MECHANISTIC STUDIES OF DMA^sup V^-INDUCED CARCINOGENICITY IN THE RAT

A. Description of Mechanistic Studies With DMA^sup V^

Mechanistic studies (Table 3) have clarified the mode of action for DMA^sup V^-induced bladder tumors in rats as involving cytotoxicity with necrosis of the bladder urothelium, followed by a regenerative process and leading to sustained increased cell proliferation within the affected tissue (i.e., hyperplasia) (Arnold et al., 1999; Cohen et al., 1999, 2001, 2002b). In the study by Arnold et al. (1999), female rats exposed to 40 or 100 ppm DMA^sup V^ in their diet for 10-20 weeks exhibited urothelial cytotoxicity and hyperplasia. In addition, the bromodeoxyuridine (BrdU) labeling index, which allows for detection of newly synthesized DNA, was significantly increased, indicating increased cell proliferation. The urothelial cytotoxicity and hyperplasia found in female rats treated with 100 ppm DMA^sup V^ for 10 weeks was reversible following a recovery period of 10 weeks without DMA^sup V^ in the diet (Arnold et al., 1999). Shortterm urothelial changes were greater in the female rat compared to the male rat, a finding consistent with the 2-year bioassay results (van Gemert and Eldan, 1998). The short-term effects occurred at the same doses that caused hyperplasia and tumors in the 2-year bioassay. Importantly, the toxicity was not due to alteration of urine composition or formation of urinary solids, indicating that the toxicity was due to the DMA^sup V^ and/or its metabolites.

In a separate experiment, urothelial cytotoxicity was detected morphologically by scanning electron microscopy (SEM) as early as 6 hours after treatment of rats with 100 ppm DMA^sup V^ in the diet. Focal cellular necrosis followed after one to three days, and widespread urothelial necrosis after seven days (Cohen et al., 2001). Increased cell proliferation in response to the cell death was detected by an increased BrdU labeling index and also observed by SEM beginning 7 days after the start of dietary treatment. These findings demonstrate that at sufficiently high doses DMA^sup V^ causes urothelial cytotoxicity with necrosis, followed by sustained regenerative hyperplasia in the rat bladder epithelium, and that these events occur in a relatively short timeframe.

When DMPS, which inactivates trivalent arsenicals via chelation (Aposhian, 1998), was coadministered in the diet to female rats with DMA^sup V^ for 2 weeks, urothelial necrosis and regenerative proliferation were inhibited (Cohen et al., 2002b). In the urine of rats administered 100 ppm DMA^sup V^(without DMPS), the investigators subsequently found 0.8-5.1 M of DMA^sup III^, the immediate metabolite of DMA^sup V^. These concentrations were shown to be cytotoxic to both rat and human bladder urothelial cells in culture (Cohen et al., 2002b). DMA^sup V^ was also present in the urine but at concentrations that would not be expected to cause cytotoxicity. In additio\n, two unidentified, stable arseniccontaining metabolites were detected at low concentrations. The contribution of these compounds to the urothelial cytotoxicity cannot be determined until they are chemically identified. Two unidentified urinary metabolites were also present in rat urine after DMA^sup V^ administration in drinking water (Wanibuchi et al., 1996). Based on these findings (Cohen et al., 2002b), it is likely that the trivalent metabolite of DMA^sup V^ (i.e., DMA^sup III^) was responsible for the cytotoxicity observed in the rat bladder epithelium. This hypothesis is consistent with the observation that rats methylate DMA^sup V^ to TMAO more than other species (see section HE), generating DMA^sup III^ as an intermediate, and are also more susceptible than other species to DMA^sup V^-induced bladder cytotoxicity and carcinogenicity. The urothelium is the target tissue presumably because the concentration of DMA^sup III^ is higher in urine compared to blood. These findings are also consistent with observations demonstrating that urothelial cytotoxicity and regeneration is a frequently observed mode of action for chemically induced bladder cancer in rats by chemicals that are not DNA reactive (Cohen, 1998b).

Following administration of 2, 10,40, or 100 ppm of DMA^sup V^ in feed, urinary concentrations of DMA^sup III^ increased in a dose- dependent manner, with a possible threshold of toxicity based on the presence of urinary DMA^sup III^ levels (Arnold et al., 2003a). Previously established doses of DMA^sup V^ that cause bladder hyperplasia (i.e., 40 ppm and 100 ppm) resulted in urine DMA^sup III^ concentrations consistent with those that cause cytotoxicity in the in vitro experiments. Conversely, DMA^sup V^ doses that did not result in DMA^sup III^ urinary concentrations that were cytotoxic in human and bladder urothelial cell cultures (i.e., <0.1 M), also did not lead to significant bladder alterations (Arnold et al., 2003a). Okina et al. (2004) verified that after dosing rats with a carcinogenic regimen of DMA^sup V^ (i.e., 200 ppm in drinking water), urinary concentrations of DMA^sup III^ were above the LC^sub 50^ values established in the in vitro studies. These investigators also found levels of MMA^sup III^ above the LC^sub 50^ values, and concluded that MMA^sup III^ and DMA^sup III^play significant roles in the toxic and carcinogenic effects of DMA^sup V^ on urinary bladder in rats.

B. Cytotoxicity of Inorganic and Organic Arsenic Compounds

Mechanistic studies have clarified that the mode of action of DMA^sup V^-induced rat bladder tumors involves bladder epithelial cell necrosis followed by regenerative hyperplasia, presumably due to the presence of high urinary concentrations of cytotoxic metabolites (Cohen et al., 2002b). Table 4 presents a summary of the key in vitro studies that demonstrate the higher potency of inorganic arsenic compared to MMA^sup V^ and DMA^sup V^ and of trivalent arsenicals compared to pentavalent arsenicals.

Studies aimed at the determination of the specific metabolites which cause the cytotoxicity observed in DMA^sup V^-treated rats are ongoing. DMA^sup III^, the trivalent metabolite of DMA^sup V^, induces comparable or greater cytotoxicity than trivalent inorganic arsenic (Cohen et al., 2002b; Kitchin, 2001). Furthermore, there is evidence to suggest that the trivalent metabolites MMA^sup III^ and DMA^sup III^ can indirectly induce DNA damage, possibly via the generation of oxidative damage (see next section). Demethylation of DMA^sup V^ and MMA^sup V^, which may occur in certain species of bacteria, does not occur in mammalian systems (Cullen and Reimer, 1989; Kenyon and Hughes, 2001). Thus, it is unlikely that cytotoxicity associated with methylated pentavalent or trivalent arsenicals is due to conversion to inorganic arsenic.

In evaluating the effects of arsenic compounds in cell culture studies, consideration of cell type is critical. Results in epithelial cell systems are more likely to be relevant to arsenical carcinogenesis in vivo than results from non-epithelial cell systems, since the tumors produced in humans and animals by arsenicals are exclusively of epithelial origin. Epithelial cells differ biologically significantly from other cell types, such as fibroblasts or lymphocytes, especially in vitro, and therefore, studies of arsenicals in nonepithelial cell systems should be interpreted with caution.

Several studies have demonstrated that inorganic arsenic is more toxic than the pentavalent methylated metabolites (Table 4). For example, Klimecki et al. (1997) examined the relative cytotoxicity of pentavalent arsenic compounds (As^sub i^^sup V^, MMA^sup V^, and DMA^sup V^) in a human keratinocyte cell line. They demonstrated that As^sub i^^sup V^ was significantly more cytotoxic than MMA^sup V^ or DMA^sup V^. In fact, DMA^sup V^ was nontoxic up to, and including, concentrations in the millimolar range. Results of other studies indicate that trivalent forms of arsenic (inorganic and organic) are significantly more cytotoxic, in some cases by as much as several orders of magnitude, than the respective pentavalent forms (e.g., Cohen et al., 2002b; Petrick et al., 2000; Styblo et al., 2000). Specifically, the methylated trivalent arsenic compounds are acutely cytotoxic at micromolar concentrations or less, whereas the pentavalent forms are acutely cytotoxic only at about millimolar concentrations or higher. For example, the higher cytotoxic potency of trivalent arsenic has been observed in human liver cells (Petrick et al., 2000); in cultured rat liver cells; in human liver, skin, bladder, and lung cells (Styblo et al., 2000); in human skin cells (Vega et al., 2001); and in rat and human bladder epithelial cells (Cohen et al., 2002b). LC^sub 50^ values from urothelial cytotoxicity studies with various arsenic compounds are presented in Table 5. As Table 5 shows, the methylated trivalent arsenic compounds are three orders of magnitude more toxic than the respective methylated p

Source: Critical Reviews in Toxicology

More News in this Category