Selenium Health Benefit Values: Updated Criteria for Mercury Risk Assessments

Selenium (Se)-dependent enzymes (selenoenzymes) protect brain tissues against oxidative damage and perform other vital functions, but their synthesis requires a steady supply of Se. High methylmercury (CH3Hg) exposures can severely diminish Se transport across the placenta and irreversibly inhibit fetal brain selenoenzymes. However, supplemental dietary Se preserves their activities and thus prevents pathological consequences. The modified Se health benefit value (HBVSe) is a risk assessment criterion based on the molar concentrations of CH3Hg and Se present in a fish or seafood. It was developed to reflect the contrasting effects of maternal CH3Hg and Se intakes on fetal brain selenoenzyme activities. However, the original equation was prone to divide-by-zero-type errors whereby the calculated values increased exponentially in samples with low CH3Hg contents. The equation was refined to provide an improved index to better reflect the risks of CH3Hg exposures and the benefits provided by dietary Se. The HBVSe provides a biochemically based perspective that confirms and supports the FDA/EPA advice for pregnant and breast-feeding women regarding seafoods that should be avoided vs. those that are beneficial to consume. Since Se can be highly variable between watersheds, further evaluation of freshwater fish is needed to identify locations where fish with negative HBVSe may arise and be consumed by vulnerable subpopulation groups.


Introduction
Selenium (Se)-dependent protection against otherwise lethal effects of high mercury (Hg) exposures was first described in 1967 [1], and soon afterwards, it was shown that high Hg or methyl-Hg (CH 3 Hg) exposures severely diminished Se transport across the placenta [2,3]. Although the nutritional essentiality of Se has been known since 1957 [4], the importance of tissue Hg:Se molar ratios in relation to Hg toxicity was not described until it was recognized that Se-dependent enzymes (selenoenzymes) were inhibited by Hg [5]. It was subsequently noted that when Se and Hg are coadministered, insoluble and biologically unavailable HgSe complexes formed in blood and tissues [6,7]. Although this mechanism was initially misinterpreted as Se sequestering Hg, thereby rendering Hg unable to impose harm, numerous studies have since confirmed that Hg sequesters Se [8][9][10][11] and thereby inhibits the activities of selenoenzymes, which are vitally important for brain health and functions [12][13][14]. Animal studies of maternal CH 3 Hg exposures have revealed that fetal brain selenoenzyme activities are far more sensitive to CH 3 Hg inhibition than those of adults [15,16], and once fetal brain selenoenzyme activities are inhibited, they are not readily restored [17]. Irreversible inhibition of selenoenzyme activities and its biochemical sequelae are well characterized [14,18,19] and appear to be the primary mechanism of CH 3 Hg toxicity [14,20].
The Se health benefit value, or Se-HBV [21], is a recently developed risk assessment criterion that was developed to enable concurrent consideration of CH 3 Hg exposures and dietary Se intakes, particularly in regard to maternal consumption during pregnancy. Dietary Se and Hg have opposing effects on Se status and brain selenoenzyme activities. Therefore, this equation is used to provide an index (Se-HBV) to predict effects of maternal CH 3 Hg exposures from seafood consumption [22]. The equation was modified to reflect variances in CH 3 Hg and Se concentrations and eliminate disproportionality otherwise encountered in samples with low Hg levels. Because the equation employs Se:Hg molar ratios, the Se-HBV can become disproportionate when Hg concentrations are very low in relation to Se. The modified equation has the virtue of accurately indicating the amount of Se in excess of CH 3 Hg present in that food and is designated HBV Se to distinguish it from the original equation. Since Hg does not quantitatively sequester Se, the HBV Se provides a highly conservative index for establishing food safety considerations. This article details the enhanced reliability of the updated HBV Se index and compares the Se-HBV and HBV Se of ocean fish and other seafoods for reference purposes.

Modification of the Selenium Health Benefit Value Equation
Neurological effects in children have been associated with CH 3 Hg exposures from maternal consumption of seafoods that contain CH 3 Hg in sufficient excess to induce a conditioned Se deficiency in placental and fetal tissues. This does not occur when adequate amounts of maternal Se are available for transport to the fetus. Assessments based only on CH 3 Hg exposures [23] may indicate risks in situations where they do not exist and do not indicate risks that are accentuated by poor dietary Se intakes. The Se-HBV was developed as a more accurate index of the relative risks or benefits expected in association with seafood or freshwater fish consumption because it considers the absolute and relative molar amounts of CH 3 Hg and Se that are present [21]: This equation yields positive values when the amount of Se present in the fish is in excess of Hg, thereby indicating health risks that might otherwise accompany CH 3 Hg exposures are negated. Negative values indicate the seafood's CH 3 Hg concentrations are in excess of Se; thus, maternal consumption of that food does not offer protection from its CH 3 Hg content but could instead induce a temporary interruption or decrease in Se transport to the fetus. However, assessing the Se-HBV becomes problematic when the Hg content of the sample is at or below the detection limit. In such cases, the Se:Hg molar ratio approaches infinity and the Hg:Se ratio approaches zero, resulting in an erroneously high value that exaggerates the health benefits of increased dietary Se. Furthermore, since excessive Se intakes can be associated with health consequences, it is essential to have an index that appropriately reflects the amounts of dietary Se provided. The effects associated with consumption of a seafood with a negative Se-HBV or HBV Se depend on the CH 3 Hg in excess of Se but are also dependent on the absolute amount of Se available. For example, the adverse effects of eating seafood containing 5.5 μmol CH 3 Hg/kg with 0.5 μmol Se/kg would be greater than eating seafood containing 15.5 μmol CH 3 Hg/kg with 10.5 μmol Se/kg. Although both instances involve a Se deficit of 5 μmol/kg, the second example involves a higher CH 3 Hg exposure albeit less associated risk due to the additional Se available for distribution to fetal tissues. In certain circumstances, continual high intakes of Se might have adverse effects, so an index that reflects the amount of Se that is biologically available also needs to be reflected by the HBV Se of the fish being consumed. Thus, both the CH 3 Hg and the Se concentration are crucial aspects of this index. For that reason, the equation for calculating the index was refined in order to (1) incorporate relative and absolute amounts of Hg and Se, while eliminating the molar ratios that can result in disproportionately high values as a consequence of very low Hg concentrations, and (2) provide an indication of the net Se surplus or deficit. This approach provides a straightforward assessment of Se availability and provides a value that indicates the magnitude of the relative Se deficit or surplus for such seafoods or fish.
To determine whether the amounts of CH 3 Hg and Se present in the seafood would potentially result in a Se deficit or a net surplus, it is necessary to incorporate the difference in their molar concentrations. Through the use of Se in the denominator, the absolute molar concentration present in the food is recognized, while the result also provides an indication of the relative amount of Se available: However, in order to reflect the amount of physiological Se that is potentially provided or lost in respect to sequestration by the associated Hg, the relative amount of Se available is multiplied by the total amount of Hg and Se present in the food. To differentiate this index from that provided by the original Se-HBV equation, the improved criterion is designated as HBV Se [24]: The sign indicates whether the food would improve or diminish Se status while the scale of the value proportionately reflects the Se surplus or deficit associated with eating that seafood.
To demonstrate how these indices are affected by CH 3 Hg molar concentrations, a comparison of the calculated Se-HBV vs. HBV Se was performed using the range of CH 3 Hg concentrations that have been observed in various types of seafood. For purposes of this comparison, Se contents were maintained constant at 10.0 μmol Se/kg (approximating the average Se content of ocean fish) in relation to a range of CH 3 Hg increasing from 0.125 to 9.971 μmol/kg (0.025 to 2.0 mg/kg), shown in the log scale of Fig. 1a, and from 9.971 to 34.9 μmol Hg/kg (2.0 to 7.0 mg/kg), as shown in the linear scale of Fig. 1b.

Comparative Evaluation of Selenium Health Benefit Values of Seafoods
The Se and CH 3 Hg contents of various types of seafood were used to calculate the Se-HBV and HBV Se for each sample, along with their means and standard deviations. The molar concentrations of CH 3 Hg and Se present in yellowfin tuna (Thunnus albacares), bigeye tuna (Thunnus obesus), blue marlin (Makaira mazara), albacore (Thunnus alalunga), swordfish (Xiphias gladius), thresher shark (Alopias vulpinus), mako shark (Isurus oxyrinchus), and pilot whale (Globicephala melas) were used to perform side-by-side comparisons of the Se-HBV vs. HBV Se . The data for the ocean fish samples were originally reported in Kaneko and Ralston [21], but the results of additional repeat analyses are included in this assessment. The pilot whale data for samples collected in 1977 and 1978 were reported by Juhlshamn et al. [25]. The 1978 data were selected by Grandjean et al. as reflective of pilot whale Hg exposures by the Faroese mothers during their study [26]. The Se-HBV vs. HBV Se results for these seafoods are graphically compared in Fig. 2 and shown in Table 1.

Comparison of Se-HBV and HBV Se
At low CH 3 Hg concentrations, the Se-HBV increases exponentially as Hg diminishes (Fig. 1a). This increases bias as Se:Hg molar ratios asymptotically approach infinity when CH 3 Hg concentrations approach zero. For that reason, the Se-HBV fails to accurately reflect the moderate nutritional benefits associated with excess Se. In contrast, the calculated HBV Se asymptotically approaches the actual Se concentration of the seafood as Hg contents diminish toward zero; thus, it accurately reflects the net amount of Se available. The outcomes calculated for seafoods with negative Se-HBV or HBV Se similarly reflect the diminishment in Se status potentially associated with excess of maternal CH 3 Hg intakes (Fig. 1b).

Comparison of Se-HBV and HBV Se of Seafoods
Although the Hg contents of ocean fish species such as yellowfin tuna, bigeye tuna, blue marlin, albacore tuna, and thresher shark vary dramatically (Table 1), their HBV Se indicates they are all a net source of surplus Se and are thus predicted to protect against risks associated with CH 3 Hg exposures. However, swordfish do not consistently provide Se in excess of CH 3 Hg and therefore are not advised for mothers to consume during pregnancy. The negative HBV Se consistently observed for mako shark and pilot whale meats indicates that their consumption could compromise fetal Se supply. Thus, consumption of these seafoods should be limited during pregnancy.
The standard deviations of the Se-HBV and HBV Se for the seafood examples shown in Table 1 and Fig. 2 indicate a much higher variability of Se-HBVs in comparison to HBV Se results. Variability between the two indices was primarily driven by disproportionately high Se-HBVs calculated for seafoods that had low CH 3 Hg contents relative to Se (Table 1). For example, the Se-HBV of the bigeye tuna samples was uniformly positive but had a standard deviation that was greater than their sample mean and a coefficient of variability (CV) of 122 % (ranging from 8.6 to 594). In contrast, the HBV Se for these same samples ranged from 2.4 to 36.5, with a CV of 33 %. Since the ocean food web is rich in Se and tissue Se contents are homeostatically regulated, few seafoods have Se concentrations below 2 μmol Se/kg. For that reason, negative Se-HBVs are not prone to the exponential increases due to divide-by-zero-type errors such as those that occurred for certain seafoods with positive Se-HBVs. Thus, seafoods that contain more Se than CH 3 Hg tend to have negative Se-HBV and HBV Se values that are more or less equivalent. To summarize the comparisons of these seafoods, the differences between Se-HBV and HBV Se were greatest for samples with highly positive values, but differences decreased as the magnitude of their calculated values diminished and were negligible for seafoods with negative values (Fig. 1). The Se-HBV and HBV Se for the pilot whale data from 1977 shown in Table 1 and Fig. 2 reflect the results based on the mean CH 3 Hg and Se contents that were reported for these samples. For that reason, the standard deviations for those samples were not established.
The HBV Se results were uniformly positive for all ocean fish other than mako shark and swordfish. Because Se is homeostatically regulated in vertebrates while Hg bioaccumulates in relation to increasing age and weight, the HBV Se of most varieties of fish and other forms of aquatic life tend to diminish as they grow larger. Blue marlin was a unique exception. The amount of Se in its fillets remained in excess of CH 3 Hg at a near-constant amount, and their HBV Se remained consistent (11.46±4.18) even though their CH 3 Hg contents ranged from <1.0 μmol/kg to more than 60 μmol/kg. The concentration of CH 3 Hg in the fillets approached equimolar stoichiometries with Se (~10 μmol/kg), but the Se concentrations consistently remained in excess of CH 3 Hg by 6.00± 4.05 μmol/kg. The HBV Se of mako shark samples were uniformly negative but demonstrated a downtrend in HBV Se that accompanied increasing CH 3 Hg bioaccumulation. The HBV Se of swordfish diminished with increasing body weight and particularly with increasing CH 3 Hg (F =199, p = 9.8 × 10 −19 ). However, the highest HBV Se was not observed in the smallest swordfish, nor were the most negative values observed in the largest specimens.
Among pilot whales, only calves had positive values, while the HBV Se of meats from adults were uniformly negative. CH 3 Hg concentrations tended to increase in relation to body weight while tissue Se concentrations remained constant or diminished slightly. Therefore, the Se deficit potentially associated with pilot whale meats were significantly (F=9.1,   [21]; pilot whale data are from Julshamn et al. [25] b Since only the means±standard deviations (SD) were available for Hg and Se of the 1977 pilot whale data, the approximate Se-HBV and HBV Se for those samples were calculated based on mean values p<0.01) related to the weight of the animal. Therefore, increasing CH 3 Hg contents resulted in increasingly negative HBV Se (F=31.9, p<0.0001).

Discussion
Epidemiological and toxicological studies of CH 3 Hg exposures omit consideration of Se as the biochemical Btarget^of Hg, thus introducing statistical bias, confounding, and imprecision to their assessments. Beneficial effects of improved intakes of nutrients that counteract the adverse effects of maternal CH 3 Hg exposures on fetal outcomes are well recognized [27,28]. However, the pivotal importance of dietary Se's biochemical role in the mechanism of CH 3 Hg toxicity [14,19] was generally misunderstood and often overlooked.
Predictions of risk based only on CH 3 Hg exposures are inaccurate. The HBV Se reflects the Se surplus or deficit in a seafood compared to its CH 3 Hg contents, providing a more reliable index for assessing CH 3 Hg exposure risks. This was evident in a recent animal study that found predictions based on Se-HBV were far more consistent with observed effects than predictions based only on CH 3 Hg exposures [23]. In that study, Se-HBV's relation to toxic effects of CH 3 Hg exposures was highly significant (F=161.0, p<0.0001) and consistent (adjusted R 2 =0.735). Predictions based only on CH 3 Hg exposures were less consistent (adjusted R 2 =0.158), and their statistical significance was less robust (F=10.9, p<0.001). The crucial difference was the ability of the Se-HBV index to differentially recognize CH 3 Hg exposures that would induce Se deficits potentially severe enough to impair brain selenoenzyme activities from those that would not. In another animal study, HBV Se , Se-HBV, and CH 3 Hg exposures were compared as indices of risk. The statistical strength of HBV Se and Se-HBV regressors were virtually identical, and both indexes identified adverse effects of CH 3 Hg exposures sooner and with higher p values than assessments performed using only the CH 3 Hg regressor [24].

Role of Background Diet
Differences in Hg exposure levels or dietary Se intakes that minimally affect physiological Se status are unlikely to have clinical consequences. However, individuals with poor dietary Se status are more susceptible to the adverse effects from consuming foods with negative HBV Se than Se-rich populations. This can explain why studies have reported negative effects from high CH 3 Hg exposures in populations with low dietary Se intakes. For example, a study in New Zealand reported that high CH 3 Hg exposures from maternal consumption of seafoods during pregnancy resulted in negative effects in children [29]. However, this population was known to have an extremely poor Se status [30] making it especially vulnerable to adverse effects from eating foods with a negative HBV Se . The study indicated that shark fillets with CH 3 Hg contents as high as 4.4 mg/kg (~22 μmol/kg) and an estimated HBV Se as low as −120 were frequently consumed in the form of fish-and-chips [31]. Eating such high-CH 3 Hg fillets would not be recommended for any population, but the reported adverse effects were especially predictable since Se availability to fetal tissue was already compromised by the mothers' extremely low Se status. Conversely, the adverse effects of high CH 3 Hg exposures have been shown to be alleviated or eliminated when diets containing seafoods with a negative HBV Se are complemented by Se-rich diets (e.g., from consuming Serich ocean fish) [32,33].

The Contrast Between Hg Exposures from Ocean vs. Freshwater Fish
Although most ocean fish contain excess Se over their CH 3 Hg contents [21,34], top predators in freshwater with particularly poor Se availability have been shown to accumulate more CH 3 Hg than fish of the same species and size from Se-rich watersheds [35]. This situation is especially notable in areas with high Hg inputs from local point sources or with inputs of acidic material, which greatly decrease Se bioavailability. Therefore, the fish that have the least amount of Se tend to bioaccumulate the most CH 3 Hg. Likewise, increases in amounts of bioavailable Se have been shown to increase CH 3 Hg efflux from fish [36][37][38][39][40][41][42][43][44] and rapidly diminish their CH 3 Hg body burdens. This mechanism of depuration is augmented by production of insoluble HgSe in tissues of prey animals at each level of the food web. Because HgSe is highly stable, it passes through the digestive tract unabsorbed and is eliminated, resulting in essentially permanent retirement in the sediments.
Watersheds with low-Se fish occur in various regions of the world. This arises due to Se's poor bioavailability at low pH [45], poor geological abundance in soils from igneous parent rock materials, or extensive leaching of porous soils by high rainfalls [46]. Since increased CH 3 Hg burdens are associated with lower Se contents in fish [34], regions with freshwater fish potentially having negative HBV Se need to be identified. Fish from low-Se watersheds that are concurrently exposed to high CH 3 Hg inputs and acidic waste drainage are therefore expected to have negative HBV Se . Eating fish from such areas would pose greater risks than consuming Se-rich fish that contain the same amount of CH 3 Hg. Because the reference dose and fish consumption advisories are based on CH 3 Hg levels alone, the extent of risk associated with high CH 3 Hg exposures due to eating fish from Se-poor watersheds is currently overlooked. In the absence of dietary Se intakes sufficient to compensate for losses due to Hg sequestration, high CH 3 Hg exposures are more likely to diminish maternal and fetal Se status. Therefore, consumption of fish with high CH 3 Hg contents that arise in areas with poor Se availability is an issue that deserves further study. Fortunately, restoring fish Se concentrations to optimal levels comes with the added benefit of diminishing their CH 3 Hg contents [37][38][39]. The combined effects of diminishing CH 3 Hg contents while improving the Se status of the aquatic ecosystem would improve the HBV Se of the fish. In Se-deficient areas, CH 3 Hg remediation can easily be achieved by augmenting environmental Se to adequate levels.

Conclusions
Since the HBV Se is based on the biochemical mechanism of CH 3 Hg toxicity, it provides an objective index for assessing the relative effects of CH 3 Hg exposures and dietary Se intakes on Se status. Seafoods with negative values (i.e., pilot whale, certain types of shark, some individual swordfish) are differentiated from ocean fish varieties with positive values. Consumption of seafoods with positive HBV Se would negate risks otherwise associated with CH 3 Hg exposures. It is important to note that intermittent CH 3 Hg exposures are unlikely to compromise maternal/fetal Se status, but consistent consumption of negative HBV Se seafoods could pose this risk, especially among mothers with poor Se intakes. The HBV Se provides a biochemically based perspective that confirms and supports the FDA/EPA advice for pregnant and breastfeeding women regarding seafoods that should be limited vs. those that are beneficial to consume. Since maternal consumption of seafoods has repeatedly been shown to benefit child neurodevelopment, the use of the HBV Se provides a reliable, easily understood, and consistent index for identifying healthy seafood choices.
While erring on the side of caution is entirely appropriate when protecting public health, the HBV Se may be overly cautious regarding the potential risks of CH 3 Hg exposures from fish consumption. The HBV Se conservatively considers only the Se from the fish itself, but dietary CH 3 Hg would also interact with Se from all other dietary sources as well as from host tissue Se reserves. Furthermore, the equation presumes that CH 3 Hg from fish consumption will unfailingly sequester an equivalent amount of Se, but the majority of the Hg that enters the body will remain bound to thiomolecules during its entire time of residence in the body without encountering or binding cellular Se. This fundamental aspect of CH 3 Hg biochemistry contributes to the prolonged latency between acquiring a toxic dose and the initial onset of signs and symptoms of toxicity [47]. The HBV Se is unique in being applicable for assessing risks associated with high exposures to CH 3 Hg as well as in rare circumstances when excessive Se contents of fish is a concern.
The reference dose established for assessing risks associated with CH 3 Hg exposures omits consideration of Se and is based on effects that were observed in a population which consumed Se-rich diets. Therefore, the reference dose may not be applicable to health consequences that may be associated with elevated CH 3 Hg exposures in Se-poor populations. For that reason, the HBV Se of freshwater fish in Se-poor regions warrants study to help identify populations that may experience accentuated risk from consistently consuming fish with negative HBV Se . A thorough evaluation of HBV Se of freshwater fish will enable recognition of locales with varieties that should be avoided or whose consumption should be limited among susceptible subpopulations. Such studies would also indicate where Se augmentation to accomplish CH 3 Hg remediation and restore Se to optimal concentrations would be appropriate.
Acknowledgments The research described in this article was funded by the US Environmental Protection Agency (EPA) National Center for Environmental Research (NCER) Science to Achieve Results (STAR) grant number RD834792-01: Fish Selenium Health Benefit Values in Mercury Risk Management. Funding to cover costs of preparation of this article was provided by the National Fisheries Institute. The sponsors had no role in the study design, collection, analysis, and interpretation of the data, or the decision to submit the article for publication. This article has not been subjected to their review and may not reflect their perspectives. Thus, no official endorsements are implied.
Compliance with Ethical Standards The work described in this manuscript did not involve studies of humans or live animals.

Conflict of Interest
The US EPA funded~98 % of the work described in this article while National Fisheries Institute (NFI) funded~2 % costs for preparation of the article for publication and travel expenses for the corresponding author to present the findings of this project at the Society of Toxicology (SOT) Meeting in 2013.