- Top of page
- 1 Introduction
- 2 Material and methods
- 3 Results
- 4 Discussion
- 5 References
Inorganic arsenic is a well-documented human carcinogen after both oral exposure and inhalation. While environmentally relevant arsenic species include inorganic as well as organic arsenicals, elevated cancer incidences have been attributed to inorganic arsenic, and arsenic research has focused on identifying the underlying mechanisms [1-3].
In the general population, diet is the primary source of arsenic intake [4, 5]. Whereas total arsenic content of terrestrial foods is generally low, foods of marine origin usually have a much higher (100-fold) total arsenic content, and this arsenic is mostly present in organic forms [4-6].
Among the water-soluble organic arsenicals identified in marine food, arsenobetaine, which predominates in fish and crustaceans , and arsenosugars are the most widespread compounds. Arsenosugars are the major arsenical constituents of marine algae (typically 2–50 mg arsenic/kg dry mass) [8, 9]. They are also present at significant concentrations in animals feeding on algae (e.g. mussels and oysters; typically 0.2–5 mg/kg dry mass) and in smaller amounts in other marine foods . Although more than 20 arsenosugars have been reported as natural products, most of the arsenic bound as arsenosugars is associated with just four compounds [7, 10, 11]. These arsenosugars are 5-deoxy-5-arsinoyl-β-d-riboside derivatives with variable side chain substitution at the C1 position, namely glycerol (-CH2CHOHCH2OH), glycerylsulfate (-CH2CHOHCH2OSO3H), glycerylsulfonate (-CH2CHOHCH2SO3H), or glycerylphosphate diester (-CH2CHOHCH2OP(O)(OH)OCH2CHOHCH2OH) resi-dues .
A risk assessment of arsenic in foods must be based not only on the type and quantities of the arsenicals present, but also on the metabolism of these compounds in humans. In the case of inorganic arsenic ingestion, arsenate can be reduced to arsenite, which can then be further transformed through a series of reductive methylation and conjugation reactions, some of which involve re-oxidation of trivalent to pentavalent arsenic ; these processes are believed to take place primarily in the liver. In cultured cells, some of the inorganic arsenic metabolites, including the trivalent mono- and dimethylated arsenicals (e.g. [14-21]) as well as thio-dimethylarsinic acid (thio-DMAV) (e.g. [22, 23]) have been shown to exert higher cytotoxicity and genotoxicity than arsenite. The major inorganic arsenic metabolite, dimethylarsinic acid (DMAV), exerts generally lower cellular toxicity but, in contrast to all other arsenicals tested so far, has been shown to be a complete carcinogen to the rat bladder [24, 25]. Data from epidemiological studies suggest that biomethylation of inorganic arsenic contributes to inorganic arsenic-induced carcinogenicity (e.g. [26-28]).
In contrast to the numerous studies on inorganic arsenic, there have been relatively few human metabolism studies on the organoarsenicals present in fish and seafood. Although arsenobetaine is bioavailable to humans, it is not metabolized and is rapidly excreted unchanged in urine . Consequently, arsenobetaine is widely assumed to be of no toxicological concern [30-33]. In contrast, arsenosugars are metabolized by humans to a multitude of arsenic metabolites, which are subsequently excreted in the urine. Thus, after a single ingestion of synthetic 2′,3′-dihydroxypropyl 5-deoxy-5-dimethylarsinoyl-ß-d-riboside (DMAV-sugar-glycerol, AsS I) humans efficiently metabolized this arsenosugar [8, 34, 35]. Thereby, the excretion patterns indicated a strong individual variability in human metabolism [8, 35]. In human urine, traces of the intact DMAV-sugar-glycerol and more than ten metabolites were detected, including DMAV, the major metabolite, oxo-dimethylarsenoacetic acid (oxo-DMAAV), thio-dimethylarsenoacetic acid (thio-DMAAV), oxo-dimethylarsenoethanol (oxo-DMAEV), thio-dimethylarsenoethanol (thio-DMAEV), and thio-DMAV [34, 35]. Additionally, thio-DMAAV and thio-DMAEV were detected in blood serum. Besides humans also sheep chronically exposed to arsenosugars via seaweed consumption seem to efficiently metabolize arsenosugars .
Regarding the in vivo toxicity of naturally occurring arsenosugars, there is only one very recent paper available indicating the induction of oxidative stress, DNA damage, and neurobehavioral impairments in mice after high-dose DMAV-sugar-glycerol ingestion (≥20 mg/kg b.w. and day) . In vitro studies showed that DMAV-sugar-glycerol exerted no toxic effect up to the millimolar concentration range [38-40]. An IC50 value > 6 mM, determined by the neutral red uptake assay, was obtained for DMAV-sugar-glycerol in human keratinocytes; genotoxic potential was not observed in a DNA nicking assay nor was it observed in the Ames test. . Likewise, cytotoxicity as measured by the Alamar Blue test occurred only in millimolar concentrations in BALB/c 3T3 mouse fibroblast cells and murine alveolar macrophages (IC50 = 6 mM, IC50 = 8 mM, respectively) . Oya-Ohta et al. observed a significant increase of chromosomal aberrations in human fibroblasts after 24-h incubation with 12 mM DMAV-sugar-glycerol .
In their recently published Scientific Opinion on Arsenic in Food, the European Food Safety Authority Panel on Contaminants in the Food Chain concluded that a risk assessment for arsenosugars is currently not possible, largely because of the lack of relevant toxicological data. To address this issue, we report an extensive toxicological characterization of two arsenosugars (DMAV-sugar-glycerol and DMAV-sugar-sulfate) and their metabolites DMAV, thio-DMAV, oxo-DMAAV, thio-DMAAV, oxo-DMAEV, and thio-DMAEV (Fig. 1) in cultured human bladder cells. Moreover, for the first time intestinal bioavailability of the arsenosugars was assessed applying the Caco-2 intestinal barrier model.
- Top of page
- 1 Introduction
- 2 Material and methods
- 3 Results
- 4 Discussion
- 5 References
This study investigated the cellular toxicity and assessed the intestinal bioavailability of two arsenosugars naturally present in marine foods. For the first time, cellular toxicity of the arsenosugars was studied in direct comparison to the effects of six arsenosugar metabolites as well as the reference substances, first being arsenite (iAsIII), which is classified as a human carcinogen , and second being arsenobetaine, which is known to be harmless to humans .
In the applied concentration range, up to 500 μM the two arsenosugars, DMAV-sugar-glycerol and DMAV-sugar-sulfate, as well as arsenobetaine exerted neither cytotoxicity nor genotoxicity, even though they were bioavailable to UROtsa cells. Regarding DMAV-sugar-glycerol, the observed lack of cellular toxicity in the micromolar concentration range is in accordance with earlier studies [38-40]. Additionally, our data indicate that this lack of toxicity is not due to low cellular bioavailability of arsenosugars. In direct comparison to DMAV-sugar-glycerol, DMAV-sugar-sulfate, which has never been toxicologically characterized before, shows about a fourfold higher bioavailability to UROtsa cells. The trivalent arsenosugar analogue DMAIII-sugar-glycerol has been reported to show cellular toxicity in cultured human keratinocytes , however the existence of this arsenical in biological samples is yet to be analytically proven.
DMAV has been identified as the major human urinary metabolite after both inorganic arsenic and DMAV-sugar-glycerol intake; moreover, thio-DMAV has been identified as a common metabolite. In contrast, oxo-DMAAV, thio-DMAAV, oxo-DMAEV, and thio-DMAEV are exclusive metabolites of arsenosugars [8, 34, 35, 55]. In the present study, these four arsenosugar metabolites behaved similarly to the parent arsenosugars – they were bioavailable to the cells but did not show cellular toxicity in any of the investigated cytotoxicity and genotoxicity endpoints. With respect to effects on cellular dehydrogenase activity, this lack of toxicity is in accordance with data published in cultured human hepatocarcinoma (HepG2) cells applying the WST-8 test . Other cytotoxicity endpoints, cellular bioavailability and genotoxicity of oxo-DMAAV, thio-DMAAV, oxo-DMAEV and thio-DMAEV have not been studied before. A direct comparison of the effects of the organoarsenicals with arsenite in this study clearly indicates that cellular toxicity is not only related to the cellular total arsenic concentration but also depends on the arsenic species incubated. Thus, 48-h incubation with 500 μM DMAV-sugar-sulfate, arsenobetaine or thio-DMAAV resulted in a similar or even higher cellular total arsenic content than 48-h incubation with 1 μM arsenite. Nevertheless, incubation with just 1 μM arsenite caused a decrease in colony-forming ability as well as a strong increase in micronuclei frequency.
The common inorganic arsenic and arsenosugar metabolites DMAV and thio-DMAV caused moderate and strong cellular toxicity in UROtsa cells, respectively. Depending on the endpoint studied, thio-DMAV exerted up to 100-fold higher cytotoxicity than DMAV; indeed, thio-DMAV was slightly more cytotoxic than arsenite. In terms of cytotoxicity, this order of the arsenicals has been shown before in cultured human lung, liver, and bladder cells [22, 23], and is strongly related to the cellular bioavailability of these arsenicals. For all three arsenicals, colony-forming ability turned out to be the most sensitive cytotoxicity endpoint investigated, suggesting an indirect mode of toxic action for these arsenicals.
As previously reported (e.g. [23, 56, 57]), our reference compound arsenite induced micronuclei, whereas both DMAV and thio-DMAV increased the frequency of bi- and multinucleated cells, indicating cell-cycle arrest and disturbance in mitosis. On the cellular level, thio-DMAV and DMAV seem to exhibit genotoxicity in a similar manner, which is most likely different from that shown by arsenite. When additionally taking into account cellular bioavailability of the arsenic species, effects induced after thio-DMAV and DMAV incubation occur at comparable total cellular arsenic concentrations. These facts suggest that after incubation with either thio-DMAV or DMAV, a common arsenic species is formed inside the cell causing the respective cytotoxic and genotoxic effects. A possible candidate is DMAIII, which has recently been discussed to be formed inside the cell from DMAV and thio-DMAV . However, this assumption needs further experimental support including rigorous analytical confirmation. Genotoxic effects have been described for all three dimethylated metabolites, DMAV, DMAIII and thio-DMAV, with DMAIII often being the most potent species [23, 58-62].
When humans consume arsenosugars, whether present in seafood [33, 63, 64] or ingested as a pure synthesized compound [8, 34, 35], they efficiently metabolize these organoarsenicals. Six arsenosugar metabolites have been identified in humans so far, with DMAV being the major metabolite. Moreover, traces of arsenosugars themselves have also been identified in human urine [8, 34, 35], indicating that at least a fraction of the ingested arsenosugars are absorbed in the form of intact arsenosugars via the gastrointestinal tract. In vitro simulation studies provided evidence that arsenosugars might be stable during stomach passage . Regarding the rate of absorption of arsenosugars, there are conflicting data in the literature. In vitro digestion studies indicate a strong bioaccessibility (>80%) . In an early in vivo study, approximately 80% of ingested DMAV-sugar-glycerol (equivalent to 1220 μg arsenic) was excreted in urine in one male volunteer during 4 days after ingestion , giving evidence of almost complete absorption in humans. However, recent data based on urinary excretion after single dose oral intake of these arsenosugars suggest considerable individual variability in the absorption and metabolism of arsenosugars . Among the six volunteers, 4-98% of the ingested arsenic was excreted via urine within 4 days. Our data obtained from the Caco-2 intestinal bioavailability studies show a rather low transfer of the arsenosugars across the intestinal barrier model. Thereby, the applied sulfate arsenosugar was about 1.6-fold more bioavailable than the glycerol sugar that had been ingested by volunteers in earlier experiments [8, 34, 35]. Thus, within 48 h, 1.7 ± 0.4 and 2.8 ± 0.5% of the applied DMAV-sugar-glycerol and DMAV-sugar-sulfate crossed the intact Caco-2 barrier, probably via paracellular transfer. Nevertheless, paracellular transfer of the arsenosugars might be underestimated by the Caco-2 model. Paracellular transfer is often underestimated by the Caco-2 model because of the strong tightness of tight junctions in comparison to the in vivo situation . In vivo tight junctions of the intestinal mucosa are much more permeable [68, 69]. The transfer rate of the arsenosugars was slightly lower than the crossover of arsenobetaine and much lower compared to the transfer rates of arsenite; the observed transfer rates of both arsenite and arsenobetaine are similar to the respective Caco-2 transfer rates that have been published before [47, 49, 51]. Together with the observed high total arsenic absorption of the ingested DMAV-sugar-glycerol in some volunteers, these data point to the possibility that arsenosugars are at least partly biotransformed before absorption. Accordingly, Conklin et al. demonstrated bioconversion of DMAV-sugar-sulfonate to its thio-analogue applying mouse cecal microflora and tissue ; also for other arsenicals presystemic metabolism by cecal tissue and microbial flora to higher methylated or thiolated arsenic species has been postulated [70-73]. Presystemic metabolism is likely to affect toxico Kinetics as well as toxicity of the arsenosugars and interindividual differences in presystemic metabolism may also contribute to the great individual variability in arsenosugar metabolism observed by Raml et al. .
Asian population groups in general consume comparably high amounts of seaweed, some of which contain more than 100 mg of arsenic/kg dry mass , with arsenosugars being the predominant arsenic fraction. In Japan, exposure to arsenosugars from seaweed consumption can be as high as 1 mg/day . Furthermore, in the United States and Europe seaweed consumption, usage of marine algae-based food supplements, and food additives based on marine algae are increasing and therewith arsenosugar exposure. Additionally, the consumption of mussels and oysters might strongly contribute to arsenosugar intake. Immediately after the discovery of the metabolism of arsenosugars to DMAV in sheep by the Feldmann group , concern was voiced that DMAV might pose a risk to human health , particularly since DMAV has been shown to be a complete carcinogen in the rat [24, 25]. These fears have been strengthened by the arsenosugar metabolism studies in humans, which clearly indicated that after arsenosugar intake, likewise after inorganic arsenic intake, DMAV is the major (>50%) human urinary metabolite. To compare amounts of DMAV formation after arsenosugar or inorganic arsenic intake by humans, a worst case scenario can be assumed, based on the volunteers that excreted up to 98% arsenic of the ingested arsenosugar dose . When eating 10 g algae (dry weight) with a content of 25 mg arsenic/kg dry weight per day, an individual consumes approximately 250 μg arsenic mostly in the form of arsenosugars and might be expected to excrete approximately 120 μg DMAV. To reach similar amounts of DMAV from inorganic arsenic intake would require consumption of ca. 20 L of water containing the maximum permissible arsenic content of 10 μg/L . Arsenosugars might also show toxic effects via the formation of thio-DMAV, which exerts cellular toxicity in about 100-fold lower concentrations compared to DMAV, as shown in the present study. Moreover, thio-DMAV has been shown to affect the cellular oxidative stress response in cultured human lungs cells in the ultralow, picomolar concentration range . Additionally, the thio-analogues of the arsenosugars, which might be formed in the gastrointestinal tract , are likely to exert higher bioavailability and toxicity.
In summary, we demonstrate that the sulfate and glycerol arsenosugar investigated in this study, as well as four arsenic metabolites of these two arsenosugars, exert no cellular toxicity up to 500 μM exposure in cultured human bladder cells, even though they are bioavailable to the cells. However, DMAV and in particular its S-analogue thio-DMAV are toxic to bladder cells in the micromolar concentration range. Thus, it is likely that in a cellular system that metabolizes arsenosugars, especially to thiolated metabolites, cellular toxicity might arise. Moreover, arsenosugars are intestinally bioavailable, both in vivo [8, 34, 35] as well as in the Caco-2 model. Therefore, in strong contrast to arsenobetaine, arsenosugars cannot be categorized, as nontoxic for humans and a risk to human health cannot be excluded.