S15 (I) Mercury and selenium interactions: Biogeochemistry and human health

Selenium (Se) is an essential element incorporated into several human proteins needed for metabolism. Toxicities have been observed in populations either deficient in Se or exposed to Se in excess. The possibility that co-exposure to Se may ameliorate adverse effects of mercury (Hg) has been raised in experimental studies and reviews of the literature. Such a possibility has been embraced by the popular press and several websites, but there is no scientific consensus as to the likelihood and generalizability of selenium reduction of mercury toxicity. At the Eighth International Conference on Mercury as a Global Pollutant (2006) the following consensus statement was developed: “There is some evidence from animal studies showing that selenite protects against inorganic mercury toxicity. However, there is almost no evidence showing protection against methylmercury toxicity by organo-selenium compounds, such as selenomethionine or selenocysteine, the forms of selenium commonly found in human diet. There is [sic] no human data demonstrating a protective role for selenium against the neurotixicity of mercury including developmental neurotoxicity”. Since 2006 there have been a number of studies investigating the potential for nutrients in fish to decrease measured effects of methylmercury; other studies have observed that co-exposure to contaminants, including methylmercury, could be considered to reduce the salutary effects of fish consumption. This session reviews both published and new data on Hg-Se exposed populations. A useful approach in evaluating a wide variety of perhaps conflicting data is weight of evidence. In this process data quality objectives are established; data or endpoint hierarchies are described; all studies (positive and non-positive) are considered; and a judgment is rendered as to the likelihood that an effect occurs in humans. A proposed hierarchy for evaluating potential Hg-Se interactions is this: human effects data are preferred to those from animal models; measured exposure to relevant forms of Hg and Se are preferred; in vivo data are preferred to in vitro. The opinions in this abstract are those of the author and do not necessarily reflect the policies of U.S. EPA.

Background:
Fish and other seafood may contain beneficial nutrients as well as harmful contaminants, and the balance of risks and benefits on brain development are not clear. The prenatal period appears to be a time of particular susceptibility to the adverse influence of mercury (Hg) as well as the potentially beneficial influence of nutrients such as the elongated n-3 fatty acids docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA), and of selenium (Se). Selenium binds Hg and may protect against its toxic effects. Several well-designed longitudinal cohort studies have examined the potential neurotoxicity of prenatal Hg exposure, but less information is available about the overall influence of dietary fish intake, including Hg, Se, and fatty acids.

Methods:
In 1999-2002 we enrolled pregnant women in Massachusetts into the Project Viva cohort, and have followed their children since birth. Using a semiquantitative food frequency questionnaire administered at 24-28 weeks gestation, we estimated weekly fish consumption and intake of DHA+EPA. We collected maternal blood at the same timepoint and assayed stored maternal erythrocytes for total Hg (Direct Mercury Analyzer 80) and Se (DRC-ICP-MS). When children reached age 7 years, we administered the Kauffman Brief Intelligence Test (KBIT), which includes verbal and nonverbal subscales. We present here preliminary results from 291 children (approximately half of those with stored prenatal blood samples). We performed multivariable linear regression analyses adjusting for maternal and child characteristics including home environment and maternal IQ.

Results:
Mean (SD) values were: DHA+EPA intake 160 (136) mg/d, erythrocyte Hg 4.5 (4.2) ppb, erythrocyte Se 286 (67) ppb, KBIT verbal 115 (13) points, KBIT nonverbal 107 (16) points. We saw no evidence that age 7 KBIT scores were related to prenatal Se [per 25 ppb 0.02 (95% CI: -0.49, 0.53) verbal; -0.46 (-1.19, 0.27) nonverbal], Hg [per ppb 0.24 (-0.11, 0.59) verbal; -0.25 (-0.75, 0.24) nonverbal], or the molar ratio Se/Hg [0.46 (-2.37, 3.29) verbal, -1.46 (-5.56, 2.63) nonverbal]. Prenatal dietary intake of DHA+EPA was associated with higher verbal [per 100 mg/d 1.15 (0.12, 2.18)] but not nonverbal [0.59 (-0.91, 2.09)] scores. Mutual adjustment did not substantially change estimates.

Conclusion:
Preliminary results suggest a beneficial association of prenatal fatty acid intake with age 7 year verbal intelligence, but no association of mercury or selenium levels with intelligence. Analysis of remaining stored biosamples will be completed by the time of the ICMGP meeting.

Experimental studies suggest that selenium (Se) may decrease methylmercury (MeHg) toxicity under certain experimental regimens. In epidemiological studies, fish intake may seem to compensate for some MeHg-associated effects. Although no evidence was found that Se protected against MeHg neurotoxicity in two Faroese birth cohorts, little is known about the potential protective effects of dietary Se against the adverse MeHg effects on cardiovascular function in humans. We examined 713 Faroese residents aged 70-74 years (64% of eligible population) to assess the possible interaction. Dietary habits in this fishing population included frequent consumption of seafood and whale meat, which is high in mercury. Both Hg and Se were measured in whole blood. Outcome measures include heart rate variability (HRV), blood pressure (BP), and common carotid intima-media thickness (IMT). Multiple regression analyses were carried out to determine the confounder-adjusted effect of mercury exposure. Each outcome was modeled as a function of Hg and Se interactions (with adjustments for potential risk factors) by expressing the effects of log₁₀(Hg) within the lowest 25 percent, the middle 50 percent, and the highest 25 percent of the Se distribution. Surplus Se was present in whole blood, the average being a ten-fold molar excess above MeHg. Regression analyses failed to show any statistically significant effects of Se, or interaction terms between Se and MeHg. For example, the effect of a 2-fold increase in mercury exposure in the low, middle, and high Se groups on mean IMT (indicator of carotid atherosclerosis) was 0.32 (95% CI -12.5,13.1), 2.68 (95% CI -7.89,13.1), and 4.58 (95% CI: -8.37,17.4) respectively (p-interaction, 0.90). Overall, no evidence was found that Se was a significant protective factor against adverse mercury effects on cardiovascular function. Preventive methods, therefore, need to address MeHg exposures rather than Se intakes. Furthermore, because of the benefits associated with fish intake during pregnancy, additional research is needed to determine the identity of the nutrients conferring these effects.

Amazonian riverside communities have the highest reported mercury (Hg) exposure in the world today. Selenium (Se) is an essential element and anti-oxidant involved in several body functions through selenoprotein expression. It has been suggested that selenoproteins may be key targets of Hg toxicity and dietary high Se intake may restore selenoproteins involved in mitigating Hg-mediated oxidative stress, although animal and epidemiological studies are still inconsistent. Elevated plasma Se (P-Se) concentrations have been associated to type-2 diabetes, hypercholesterolemia and/or hypertension, and very high blood Se (B-Se) status (B-Se ≥1000µg/L) has been associated to hair and nail loss, gastrointestinal problems and paraesthesia. However, other studies have shown a negative association between B-Se and blood pressure, and no association with type-2 diabetes prevalence or motor outcomes.

In our study conducted in 2006 in the Lower Tapajós Region, Se status ranged from normal to high (B-Se median: 228µg/L, range 103-1500µg/L, N=448). Brazil nuts constituted the most important source of Se (10X more than fish); other foods, including some local freshwater fish species, eggs, meat, chicken, and game meat also contributed to Se intake. Very few participants reported diabetes (1.1%), and despite high levels of Se in some individuals, no signs and symptoms of Se toxicity were observed. P-Se concentrations were associated with beneficial outcomes: lower prevalence of age-related cataracts and improved performance on tests of motor functions, and these associations were stronger when taking into account blood Hg (B-Hg) exposure. When stratifying multiple regression analysis by risk groups (low and high P-Se and B-Hg), P-Se was positively associated to health outcomes but only when B-Hg was high, while B-Hg was negatively associated to several outcomes, and this, also when P-Se is high. In this population where we see deleterious effects of Hg, the median B-Se to B-Hg molar ratio was 22.7 and ranged from 1.7 to 137.

There are increasing evidences showing that a high dietary Se intake may play a role in offsetting some deleterious effects of Hg, and this, without evidence of Se-induced toxicity. However, the beneficial effects of Se may not be necessarily observed in a population with low Hg exposure and Se may not offset all Hg-induced toxic effects in fish-eating populations. Public health advisories should avoid simplifying the Se-Hg interaction debate since the risks and benefits of Se on human health are still complex questions to answer.

Mercury compounds exert toxic effects by interacting with proteins. Seleno-enzymes, such as thioredoxin reductase (TrxR), which are involved in vital cellular functions (e.g. protein repair, stress response), were demonstrated to be particularly sensitive to mercury compounds. Recently, we showed that in vivo exposure to methylmercury inhibits TrxR but not glutathione reductase (GR), an homologous enzyme without selenium in its active site. In this study, our aim was to understand how sodium selenide supplementation interferes with the interaction between mercury compounds and enzymatic systems involved in anti-oxidant regulation in vertebrates. Juvenile zebra-seabreams (Diplodus cervinus) were divided in six groups under different exposure conditions: control, Se control, Hg(II), MeHg, MeHg and Se co-exposure, Hg(II) and Se co-exposure. Exposure lasted 28 days followed by a 14-day depuration period with clean seawater. Additionally, fishes exposed to MeHg and Hg(II) were split in two groups during the depuration period, one kept in clean seawater and the other supplemented with Se. Exposure concentrations in tanks were 2 µg L^-1 for MeHg and Hg(II) and 10 µg L^-1 for Se. Fishes were sampled at days 14, 28 and 42 and the brain, liver and kidney were analyzed for total mercury, MeHg and Se contents. Activities of TrxR, thioredoxin (Trx), seleno-dependent glutathione peroxidase (SeGpx) and GR were also determined.

The accumulation of mercury in the organs studied was much higher (10-fold) in fishes exposed to MeHg than in those exposed to Hg(II). Selenium co-exposure had no influence in the accumulation of Hg(II) but decreased to half the accumulation of MeHg. Exposure to both mercury compounds decreased significantly the activity of TrxR (>50%) in all the organs analyzed. A protective effect of the thioredoxin system was observed with Se supplementation, in the liver of fishes co-exposed to Hg(II) and Se. Supplementation of Se during the depuration phase had no visible effects. The activity of SeGpx was only affected in the brain of fishes exposed to MeHg and co-exposed to MeHg and Se. GR activity in organs was independent of the exposure conditions. These results will be discussed considering the importance of these antioxidant systems, the structure of the enzymes involved and their contribution to clarify the molecular mechanisms of mercury toxicity.

There is a long-standing literature on selenium’s ability to protect against methylmercury’s neurotoxicity. In recent studies, our lab has examined the role of methylmercury dose, selenium concentration, and exposure window (adult-onset vs gestational) to obtain a full picture of selenium-mercury interactions. Adult rats consumed a purified diet containing 0.06 or 0.6 ppm of selenium, levels that are lean or rich in selenium but still nutritionally appropriate. In a factorial design, the rats were also exposed chronically to 0, 0.5, 5, or 15 ppm of mercury (as methylmercury) in drinking water to provide approximately 40 to 1200 ug/kg/day of exposure. Exposures lasted up to 18 months. Concentrations of 5 and 15 ppm of methylmercury diminished grip strength, somatosensory function, survival and attenuated age-related increases in spontaneous running in an dose dependent fashion. Selenium substantially delayed the appearance of methylmercury’s neurotoxicity by about a month in the 15 ppm group and about 5 months in the 5 ppm group. By itself, selenium increased running in the older rats. In the developmental model using a similar design, but with only 0, 0.5, and 5 ppm of mercury, rats were exposed to methylmercury and selenium during gestation. Methylmercury exposure ended at birth but the selenium diets were throughout life. Gestational exposure at both doses impaired reversal learning, enhanced sensitivity to dopamine, and impaired reward function in the rats when tested as adults. On no task did selenium confer protection against gestational methylmercury exposure. To test theories about selenium-methylmercury interactions, the ability of selenium:mercury ratios and health-benefit values will be compared against simple mercury dose for their ability to predict neurotoxicity [SUPPORTED BY ES10865 from NIEHS]

The apparent protective effect of selenium(Se) on mercury(Hg) toxicity was well-recognized by 1980. Many toxic effects of mercury have been ascribed to its affinity for sulfur (S) and its ability to disrupt enzyme functions by breaking disulfide (cysteine) bonds. Recent years have witnessed a rapid growth in identification of selenoproteins and documentation of their critical functions. These are enzymes containing seleno-cysteine (for example, glutathione peroxidases). One suggested mechanism is that Hg toxicity results from blocking selenoprotein function or tying up Se rendering it unavailable for protein synthesis. Hence the benefit of excess Se. There are far more S-containing enzymes (virtually all) than Se-enzymes, but the latter occur at lower concentrations so are particularly vulnerable. The protean manifestations of Hg toxicity probably involve multiple mechanisms affecting multiple endpoints, arguing against a straight forward application of Se:Hg molar ratios in food for predicting risk or benefit. This paper will consider 1) the state of knowledge on the relative role of Se vs S affinity for Hg and disruption of selenoproteins versus non-selenoproteins, 2) how much and in what form does Se protect against Hg toxicity, and 3) how clarifying the protective vs. toxic mechanism will inform risk assessment for methylmercury exposure from fish consumption., particularly with respect to Se:Hg molar ratios.

Selenium (Se) is an essential trace element required at the active sites of enzymes that protect the brains of all higher organisms from oxidative damage. Mercury (Hg) binds to Se with affinities (K_d 10^-45) a million times greater than Hg’s affinity for sulfur (K_d 10^-39). As a result, Hg and methylmercury (MeHg) become increasingly potent neurotoxicants as their tissue concentrations approach those of Se (~1µM). MeHg covalently binds to the Se present at selenoenzyme active sites. Therefore, MeHg is, by biochemical definition, a highly specific irreversible inhibitor of selenoenzymes. When Se is available to replace Se lost to binding with MeHg, no increase in oxidative damage occurs. However, if the rate of Se-sequestration by MeHg exceeds the rate of dietary Se replacement, increased oxidative damage and other effects occur due to loss of selenoenzyme activities.

>The intracellular abundance sulfur is >100,000 fold greater than that of Se, so binding to sulfur is kinetically favored. However, intracellular thiols are constantly acted upon by selenoenzymes. Since the Se in the active sites of these enzymes is the most potent intracellular nucleophile, Hg that is initially bound to thiols inexorably transfers to Se. Physiological and environmental interactions between Hg and Se reflect the reciprocal effects these elements have on each other’s abundance:

Physiological interactions involve Hg-dependent Se-retirement. These occur through irreversible inhibition of selenoenzymes and secondary mechanisms due to loss of selenoenzyme functions. Therefore, toxicity of MeHg is accentuated when Se status is poor, but diminished when Se status is enriched.

Environmental interactions involve Se-dependent Hg-retirement. These effects occur through food chain mechanisms related to formation of HgSe in the tissues of prey species that is poorly absorbed by predators. Therefore, Hg bioaccumulation is enhanced in freshwater fish from regions with low-Se status and diminished in fish from regions with enriched environmental Se availability.

In summation, MeHg risk criteria need to consider dietary Se and Hg:Se molar ratios present in fish to accurately assess risks vs. benefits of maternal fish consumption by humans and wildlife. Regulatory policies that ignore the importance of Se in brain physiology and overlook the pivotal role of Se in the Hg issue may fail to adequately protect the public. Improved public health policies will require clear understanding of Hg-Se biochemical interactions and their physiological and environmental implications.