Maternal Seafood Consumption is Associated with Improved Selenium Status: Implications for Child Health

Abstract

Selenium ($\text{Se}$) is required for synthesis of selenocysteine (Sec), an amino acid expressed in the active sites of $\text{Se}$-dependent enzymes (selenoenzymes), including forms with essential functions in fetal development, brain activities, thyroid hormone metabolism, calcium regulation, and to prevent or reverse oxidative damage. Homeostatic mechanisms normally ensure the brain is preferentially supplied with $\text{Se}$ to maintain selenoenzymes, but high methylmercury (CH₃Hg) exposures irreversibly inhibit their activities and impair Sec synthesis. Due to $\text{Hg}’\mathrm{s}$ high affinity for sulfur, CH₃Hg initially binds with the cysteine (Cys) moieties of thiomolecules which are selenoenzyme substrates. These CH₃Hg-Cys adducts enter selenoenzyme active sites and transfer CH₃Hg to Sec, thus irreversibly inhibiting their activities. High CH₃Hg exposures are uniquely able to induce a conditioned $\text{Se}$-deficiency that impairs synthesis of brain selenoenzymes. Since the fetal brain lacks $\text{Se}$ reserves, it is far more vulnerable to CH₃Hg exposures than adult brains. This prompted concerns that maternal exposures to CH₃Hg present in seafood might impair child neurodevelopment. However, typical varieties of ocean fish contain far more $\text{Se}$ than CH₃Hg. Therefore, eating them should augment $\text{Se}$-status and thus prevent $\text{Hg}$-dependent loss of fetal selenoenzyme activities. To assess this hypothesis, umbilical cord blood and placental tissue samples were collected following delivery of a cohort of 100 babies born on Oahu, Hawaii. Dietary food frequency surveys of the mother’s last month of pregnancy identified groups with no (0 g/wk), low (0-12 g/wk), or high (12+ g/wk) levels of ocean fish consumption. Maternal seafood consumption increased $\text{Hg}$ contents in fetal tissues and resulted in ~34% of cord blood samples exceeding the EPA $\text{Hg}$ reference level of 5.8 ppb (0.029 μM). However, $\text{Se}$ concentrations in these tissues were orders of magnitude higher and ocean fish consumption caused cord blood $\text{Se}$ to increase ~9.4 times faster than $\text{Hg}$. Therefore, this study supports the hypothesis that maternal consumption of typical varieties of ocean fish provides substantial amounts of $\text{Se}$ that protect against $\text{Hg}$-dependent losses in $\text{Se}$ bioavailability. Recognizing the pivotal nature of the $\text{Hg:Se}$ relationship provides a consilient perspective of seafood benefits vs. risks and clarifies the reasons for the contrasting findings of certain early studies.

Graphical Abstract

1. Introduction

The toxicity of mercury ($\text{Hg}$) has been recognized for millennia (Krebs, 2006), but its mechanisms of action, the brain tissue specificity of its effects, prolonged latency, differences in maternal vs. fetal vulnerability, and contrasting outcomes observed in different study populations were poorly understood. These aspects were particularly challenging because the metabolic pathways affected by $\text{Hg}$ were unknown to earlier toxicologists. While they were aware that supplemental dietary selenium ($\text{Se}$) prevented or ameliorated the toxic effects of high $\text{Hg}$ exposures (Pařízek and Oštádalová, 1967), they imagined $\text{Se}’\mathrm{s}$ high binding affinity with $\text{Hg}$ (Dyrssen and Wedborg, 1991) protected the brain from harm by forming insoluble $\text{HgSe}$ precipitates that immobilized $\text{Hg}$. Before the importance of $\text{Se}$ physiology in the brain was recognized (Chen and Berry, 2003; Reeves and Hoffman, 2009; Nicholson et al., 2022), it was impossible to realize that $\text{Se}$-dependent enzymes (selenoenzymes) of the brain were the molecular target of $\text{Hg}$ toxicity and that supplemental dietary $\text{Se}$ simply enabled these normal enzyme activities to continue without interruption (Ralston and Raymond, 2018; Raymond and Ralston, 2020; Spiller et al., 2018).

All vertebrate cells express selenoenzymes which incorporate selenocysteine (Sec), the 21^st genetically encoded amino acid (Turanov et al., 2011). Selenoenzymes demonstrate tissue-specific patterns of expression and abundance, and high conservation in expression and function across vertebrate species (Santesmasses et al., 2020; Mariotti et al., 2012). The human genome includes 25 genes for selenoproteins that incorporate Sec as the catalytic mediator of their enzyme activities (Hatfield and Gladyshev, 2002; Hatfield et al., 2006) with almost half of these are involved in control, prevention, or reversal of oxidative damage, while others perform further vital functions such as controlling thyroid hormone metabolism and regulating calcium (Ralston, 2021). The selenol of Sec is the most powerful intracellular nucleophile identified to date and its high reduction potential makes it more efficient in catalyzing redox reactions than analogous Cys containing enzymes. (Arnér, 2009). Due to the brain’s high metabolic rate, spontaneously generated reactive oxygen species (ROS) and various byproducts of respiration would cause damage and death in these tissues if specific selenoenzymes were absent (Behne et al., 2000; Chen and Berry, 2003; Reeves and Hoffman, 2009; Nicholson et al., 2022, Pillai et al., 2014; Schweizer and Fabiano, 2022). The cerebral cortex, hippocampus, cerebellum, and olfactory bulb have exceptional selenoprotein expression patterns (Zhang et al., 2008). In addition to neurons, the glial cells, parvalbumin-expressing interneurons, and cerebellar Purkinje cells also rely on selenoproteins during development (Schweizer and Fabiano, 2022).

Severe or prolonged shortfalls in dietary $\text{Se}$ intakes do not diminish essential selenoenzyme activities in brain, placenta, or endocrine tissues. This is due to expression of LDL Receptor-related Protein 8 (LRP8) on their cell surfaces which selectively bind and internalize selenoprotein P (SELENOP), the plasma $\text{Se}$-transporter (Dlugosz and Nimpf, 2018). SELENOP possesses 10 Sec residues in its primary structure and thus efficiently delivers $\text{Se}$ to LRP8 expressing tissues to ensure they are preferentially supplied with the $\text{Se}$ they need (Burk et al., 2007; Burk et al., 2014). Mice with Selenop or Lrp8 genetic deletions experience sensorimotor dysfunctions, degeneration of brain tissues, and death unless $\text{Se}$-enriched diets are provided to preserve their brain selenoenzyme activities (Burk et al., 2007). LRP8 expression is also regulated by developmental stage since its constitutive mRNA expression by fetal brain is over 9 times greater than in adult brains (Burk et al., 2014; Pitts et al., 2012). Although dietary $\text{Se}$ deficiency in laboratory animals rapidly depletes liver, muscle, and blood $\text{Se}$ contents to ~ 2% of normal (Prohaska and Ganther, 1977; Behne et al., 2000), SELENOP-LRP8 homeostatic mechanism redistributes $\text{Se}$ from somatic reservoirs to ensure placenta, brain, and endocrine tissues are supplied. As a result, even animals continuously fed $\text{Se}$-deficient laboratory diets for several generations are able to maintain brain $\text{Se}$ at ~60% of normal in their offspring (Behne et al., 2000) and thus preserve brain selenoenzyme activities at near normal levels. Mice with genetic deletion of LRP8 suffer severe neurodegeneration in brain regions associated with auditory and motor functions (Burk et al., 2014) exhibiting pathologies similar to that observed following high methyl-Hg (CH₃Hg) exposures. While the loss of coordination (ataxia) and cerebellar damage that occurs as a result of genetic deletion of LRP8 could involve additional causes, the loss of selenoenzyme activities is clearly pivotal since supplemental dietary $\text{Se}$ prevents these consequences. Postmortem examination of the brains of victims of CH₃Hg poisoning show similar patterns of neuronal cell loss in the sensory regions of the cortex, cerebellar granular cells, and primary motor cortex (Castoldi et al., 2003). Regardless of the proximal cause, once selenoenzyme activities falter, damage from spontaneous production of reactive oxygen species (ROS) will begin to accumulate and lead to cytotoxicity and apoptosis as noted in $\text{Hg}$ toxicity (Sarafian and Verity, 1991: Clarkson and Magos, 2006) as ferroptosis of neurons and astrocytes is accompanied by precipitates of $\text{HgSe}$ which accumulate and demonstrate long term retention in brain parenchyma (Korbas et al., 2010). Mercury toxicity can result in progressive development of neurofunctional deficits, sensory dysfunctions, and further consequences that can lead to permanent disability or death. However, if sufficient $\text{Se}$ is provided soon enough, impaired motor, sensory, and cognitive functions are able to recover (Spiller et al., 2017; Spiller et al., 2021), .

CH₃Hg in fish and tissues is primarily bound with cysteine (Cys) as CH₃Hg-Cys (Harris et al., 2003), a molecular mimic of methionine that is readily absorbed and selectively transported across cell membranes (Aschner and Clarkson, 1989). This enables it to bypass placental and blood-brain barriers and preferentially distribute into rapidly growing tissue compartments where it may impair the synthesis and activities of selenoenzymes (Watanabe et al., 1999a; 1999b). Since the Cys moieties of many cellular antioxidants are substrates for selenoenzymes, these thiomolecules can function as suicide substrates by delivering CH₃Hg in the proper orientation to bring $\text{Hg}$ into close proximity with the $\text{Se}$ of the active site Sec. While $\text{Hg}$ has a high affinity for the sulfur of Cys, its affinity for the $\text{Se}$ of Sec is orders of magnitude higher (Dyrssen and Wedborg, 1991). As a result, $\text{Hg}$ transfers its association from Cys to form CH₃Hg-Sec (Raymond and Ralston 2004; Ralston and Raymond, 2018; Ralston, 2021a), thus inactivating the enzyme. The CH₃Hg-Sec eventually degrades to form insoluble $\text{HgSe}$ precipitates that permanently retire $\text{Se}$ from further participation in selenoenzyme synthesis. Therefore, by biochemical definition, CH₃Hg is a highly specific irreversible selenoenzyme inhibitor that can directly induce a localized $\text{Se}$-deficiency and prevent selenoenzyme synthesis in brain tissues (Ralston et al., 2008; Ralston 2021a). High $\text{Hg}$ exposures cause $\text{Se}$ attrition and if $\text{Hg}$-dependent depletion rates exceed dietary $\text{Se}$ intakes, may deprive the brain of biologically available Se. While other electrophile exposures can impair selenoenzyme activities in other tissues (Ozoani et al., 2023;), a high CH₃Hg exposure is the only environmental insult known to deprive the brain of $\text{Se}$ and seriously impair selenoenzyme synthesis and activities (Ralston and Raymond, 2015; ).

To characterize risks of $\text{Hg}$-dependent depletion of placental and fetal $\text{Se}$ in association with maternal consumption of ocean fish, $\text{Hg}$ must be assessed in relation to tissue $\text{Se}$. Based on the high $\text{Se}$ contents of ocean fish from Hawaii (Kaneko and Ralston, 2007; Ralston et al., 2019) and elsewhere (Robinson and Schroff, 2004; Burger and Gochfeld, 2013), we hypothesize that increasing maternal consumption of ocean fish will improve rather than diminish $\text{Se}$-status in fetal tissues. If so, ocean fish consumption would prevent instead of increase risks of $\text{Hg}$-dependent diminishments in $\text{Se}$-availability. To evaluate this hypothesis, this study examined the effects of maternal ocean fish consumption on fetal $\text{Hg}$ and $\text{Se}$ as reflected in cord blood and placental tissues.

2. Methods

This study was approved by the Western Institutional Review Board and informed consent was obtained from all participants and all data depersonalized. Methods used in this study were fully described in Gilman et al., (2015), but are briefly stated below.

2.1. Subject enrollment

Pregnant women who presented to the Kapiolani Medical Center Labor and delivery suite between June 2010 and March 2011 were approached about participating in this study and assessed for eligibility to participate. Eligibility criteria included: age 18 to 45 years old, arriving in labor with a full-term singleton fetus, gestational age ≥ 37 weeks at the Kapiolani Medical Center for Women and Children in Honolulu, Hawaii, with no chronic medical conditions, no tobacco, alcohol, or illicit drug use. The first 100 eligible participants to consent were included in this cohort. Participants completed a questionnaire that recorded demographic information and a dietary survey of seafood consumption during their last month of pregnancy. Participants reported their ethnicity and quantified their seafood intakes. Seafood consumption was calculated by multiplying the number of times they reported eating specific varieties in the past month by the amount they typically ate per meal. Seafood consumption was almost exclusively ocean fish with only 5 individuals reporting an occasional meal that included either lobster, crab, shrimp, mussel, squid, or octopus. Seafood consumption rates were used to separate subjects into three groups that ate either: no (0 oz/wk), 0-12 oz/wk, or 12 or more oz/wk of seafood in the past month.

2.2. Sample Collection

A whole blood sample (~14 mL in standard EDTA anticoagulated vacutainer) was collected post-delivery from the umbilical cord. Within two hours of delivery, 2 mL aliquots were stored at −25°C until shipping for elemental analysis. Placental tissue samples were collected from regions proximal and peripheral to the umbilical cord within 2 hours of delivery, frozen in liquid nitrogen and stored at −80°C until shipping for elemental analysis.

2.3. Selenium and Mercury Analysis

Umbilical cord blood and placental tissue samples were shipped to the University of North Dakota for elemental analysis. Individual (~0.25 g) samples of blood and placental tissue were weighed into single-use, trace element-free 50-mL digestion tubes (Environmental Express, Mt. Pleasant, SC) prepared in a series of digestion sets along with analysis blanks, certified reference materials (UTAK) and quality control samples. Samples (and identically treated controls) were treated with 5 mL of HNO₃ (Fisher Trace Metal Grade, Fisher Scientific, Waltham, MA), capped, and heated at 85°C in deep cell hot blocks (Environmental Express) for 24 hours. Samples were cooled, then 1.5 mL of 30% H₂O₂ (Fisher Certified A.C.S., Fisher Scientific) was added prior to returning to the dry block at 85°C for 8 hours. Samples were cooled, treated with 15 mL of 12 N HCl (Fisher Trace Metal Grade, Fisher Scientific) and heated at 90°C for 90 minutes to reduce SeVI to SeIV. Samples were cooled and diluted to 25 mL with double-distilled water. Digested and diluted samples were analyzed for total $\text{Hg}$ by cold-vapor atomic absorption spectrophotometry using a CETAC M-6000A (CETAC Technologies, Omaha, NE). Total $\text{Se}$ was analyzed by hydride generation atomic absorption spectroscopy using a PS Analytical Dual Millennium Excalibur (PS Analytical, Deerfield Beach, FL). Analytical blanks and whole blood controls (UTAK Laboratories, Valencia CA) for mercury and selenium were digested and run alongside samples in all batches. The certified values and expected ranges for $\text{Hg}$ (14.7 ± 2.2 μg/kg) and $\text{Se}$ (214 ±32 μg/kg) for UTAK 2 coincided well with the observed values: 13.6 ± 1.1 μg $\text{Hg}$/kg and 202 ± 13.1 μg $\text{Se}$/kg, respectively.

Upon confirming analytical quality control results were consistent with expectations, the data from each digestion batch were qualified for inclusion in the database. Mass concentrations in samples in μg/kg were converted to molar concentrations (μM) using 1.06 g/mL as a density correction for umbilical cord blood and 1.05 g/mL for placental tissues. Descriptive statistics and data distributions for mass and molar concentrations of $\text{Hg}$ and $\text{Se}$ were calculated and graphed in relation to relative fish consumption as described above. Elemental concentration results in cord blood and placental samples were segregated into groups based on maternal fish consumption levels as shown in Table 1. To conform with the Standard International system and for clarity of understanding of biochemical interactions, concentrations were preferentially reported in molar units, but since mass concentration units are commonly used, those are shown for comparison. Molar concentrations of $\text{Hg}$ and $\text{Se}$ were uniformly plotted in log-log scale graphs for cord blood (Figure 1) and placental samples (Figure 2). The Benchmark Dose lower confidence limit (BMDL) level of 58 ppb $\text{Hg}$ (~0.29 μM) and cord blood $\text{Hg}$ reference level (National Research Council, 2000) of 5.8 ppb (~0.029 μM) corresponding to the US EPA 0.1 μg $\text{Hg}$/kg/bw reference dose (RfD) regulatory criteria are shown for comparison in Figure 1.

Table 1.

Fish consumption effects on mass and molar concentrations of $\text{Hg}$ and $\text{Se}$ in fetal tissues^*

Cord Blood Fish Intake	n	μg $\text{Hg}$/kg	μg $\text{Se}$/kg	μM $\text{Hg}$	μM $\text{Se}$
No Fish	14	2.32 ± 1.81a	142.87 ± 37.20	0.011 ± 0.008a	1.707 ± 0.444
Low Fish	76	5.54 ± 3.84b	153.91 ± 27.00	0.026 ± 0.018b	1.839 ± 0.323
High Fish	10	6.24 ± 3.15b	165.08 ± 25.04	0.029 ± 0.015b	1.972 ± 0.299
Placenta Fish Intake	n	μg $\text{Hg}$/kg	μg $\text{Se}$/kg	μM $\text{Hg}$	μM $\text{Se}$
No Fish	14	3.50 ± 3.11a	246.18 ± 29.84	0.017 ± 0.015a	2.969 ± 0.360
Low Fish	76	6.99 ± 4.76b	257.27 ± 32.66	0.033 ± 0.023b	3.102 ± 0.394
High Fish	10	7.84 ± 3.92b	238.34 ± 29.11	0.037 ± 0.019b	2.875 ± 0.051

^*Fish consumption histories were assessed in participating mothers for partitioning into groups with either no (0 oz/wk), low (0-12 oz/wk), or high (12+oz/wk) ocean fish intakes. Means and standard deviations are shown. Statistically significant differences (p<0.05) as established by unpaired Student’s t-test are indicated by different superscripts.

Fig 1.

Molar concentrations of $\text{Hg}$ and $\text{Se}$ in umbilical cord blood samples from the current study are shown in relation to umbilical cord blood data reported in the Faroe Islands. Data from the Faroe Islands was acquired from graphs in Choi et al. (2008), using Image J (http://rsbweb.nih.gov/ij/) as described in Ralston et al. (2015). The dashed line indicates the benchmark dose lower confidence limit (BMDL = 58 ppb, ≈ 0.29 μM) established using data from the Faroe’s Study and the dotted line shows the reference cord blood $\text{Hg}$ level = 5.8 ppb, (≈ 0.029 μM) associated with the US EPA $\text{Hg}$ RfD of 0.1 μg/kg-bw/d after the 10-fold uncertainty factor had been applied. For reference purposes, the diagonal line indicates the equimolar $\text{Hg:Se}$ stoichiometry.

Fig. 2.

Concentrations of $\text{Hg}$ and $\text{Se}$ in placenta samples from Hawaii did not approach equimolar stoichiometry and were not correlated. Elemental concentrations observed in placenta samples reported from the Genoa study (Capelli and Minganti, 1986) are shown for comparison.

Comparison data from previous studies were examined to provide context. Umbilical cord blood $\text{Hg}$ and $\text{Se}$ concentrations from the Faroe Islands study were acquired from graphs shown in Choi et al. (2008), using Image J (http://rsbweb.nih.gov/ij/) particle recognition software to identify x and y coordinates (See Ralston et al., 2015) and displayed for comparison with data in Figure 1. Placental $\text{Hg}$ and $\text{Se}$ from Genoa (Capelli and Minganti, 1986) were adjusted from dry to wet weight values and converted to molar concentrations.

2.4. Statistical analysis

A standard unpaired Student’s T-test was used for comparisons of seafood consumption-dependent differences as indicated in the accompanying table and figure legends. Differences with p values less than 0.05 were considered significant. Regressions were performed to assess relationships between $\text{Hg}$ and $\text{Se}$ in cord blood and placenta. If significant linear relationships (p < 0.05) were noted between their elemental concentrations, the slope, intercept, and trendlines for the data were calculated and reported in comparison to reference data. To facilitate comparisons of $\text{Hg}$ and $\text{Se}$ molar concentrations observed in cord blood and placental tissues, the observed and reference data sets were plotted using log-scale graphs which included an index indicating the 1:1 molar ratio.

3. Results

3.1. Umbilical cord blood $Hg$ and $Se$ concentrations

Umbilical cord blood and placental $\text{Se}$ concentrations were in substantial excess of $\text{Hg}$ in all samples from the Hawaii cohort (Table 1). The cord blood $\text{Se}$ concentration of 1.707 ± 0.444 μM observed in the group whose mothers did not eat ocean fish and the regression intercept for cord blood $\text{Se}$ of 1.703 μM (see Figure 1) indicates $\text{Se}$ status among this cohort in Hawaii is adequate even without seafood consumption. Average cord blood $\text{Hg}$ rose from 0.011 among those whose mothers consumed no ocean fish to 0.029 μM in babies whose mothers consumed more than the recommended amount. The $\text{Hg}$ contents of all but one cord blood sample were under 0.100 μM while all but one $\text{Se}$ concentration was in excess of 1.000 μM. The $\text{Se}$ concentrations in these samples were orders of magnitude higher, respectively averaging 1.707 and 1.972 μM in these groups. Maternal seafood consumption increased placental $\text{Hg}$ accumulation but did not affect placental $\text{Se}$. Between groups with no vs high fish consumption, placental $\text{Hg}$ concentrations doubled from 0.017 to 0.037 μM. Placental $\text{Se}$ was not affected by maternal seafood intakes but was approximately 2 orders of magnitude higher (~3 μM) than the $\text{Hg}$ levels from mothers with high fish consumptions.

Cord blood $\text{Hg}$ and $\text{Se}$ data had been requested from the Faroe Islands but was not made available. In lieu of having the data provided, the x and y coordinate data was acquired from graphs in Choi et al. (2008) using Image J (http://rsbweb.nih.gov/ij/) particle recognition software to identify $\text{Hg}$ and $\text{Se}$ elemental concentrations as described in Ralston et al. (2015). Cord blood $\text{Se}$ status was higher in the Hawaiian cohort (1.703 μM) than in the Faroe Islands (1.374 μM), but increased in direct relation (p<0.0001) with $\text{Hg}$ in both cohorts (See Figure 1). However, $\text{Se}$ contents increased >9 μM for every μM of increase in cord blood $\text{Hg}$ in samples from Hawaii, but rose less than 0.5 μM per μM increase in $\text{Hg}$ in cord bloods from the Faroe Island’s cohort. Therefore, predicted trends for $\text{Hg:Se}$ molar ratios among samples from Hawaii do not intersect the 1:1 molar ratio associated with increasing risk. However, in the Faroes, cord blood $\text{Hg}$ rose in relation to $\text{Se}$ and approached/exceeded equimolar stoichiometries at the higher levels of exposure.

3.2. Placental $Hg$ and $Se$ concentrations

Placental $\text{Hg}$ concentrations were significantly (p < 0.001, adjusted R² = 0.280) associated with cord blood $\text{Hg}$ contents: y = 0.672x + 0.014 and placental $\text{Se}$ was significantly (p < 0.001, adjusted R² = 0.113) related to cord blood $\text{Se}$: y = 0.398x + 2.334. The $\text{Hg}$ contents of all but one placenta sample were under 0.100 μM while all placental $\text{Se}$ concentrations were in excess of 2.000 μM. While maternal seafood consumption resulted in an increase of ~0.020 μM $\text{Hg}$ in placental tissues from this cohort, their $\text{Se}$ contents averaged ~3.232 ± 0.417 μM and did not show any tendency to increase as $\text{Hg}$ concentration rose. The average $\text{Se}$ concentrations in placentas from the Genoa study were lower (~1.54 ± 0.33 μM), as were their $\text{Hg}$ contents (0.06 ± 0.04 $\text{Hg}$ μM). In contrast to cord blood, no correlation between $\text{Se}$ and $\text{Hg}$ was observed in placenta. The highest placental $\text{Hg}$ observed in the Hawaiian cohort was just over 0.100 μM, but since the $\text{Se}$ in that sample was ~3.000 μM, there little risk of its availability being compromised. Although placental $\text{Se}$ contents were considerably lower in samples from the Genoa Study, their $\text{Hg}$ contents were similarly unlikely to diminish $\text{Se}$ bioavailability.

4. Discussion

4.1. Key findings

The concentrations of $\text{Se}$ in fetal tissues from the Hawaii cohort are substantially higher than $\text{Hg}$ levels regardless of the amount of ocean fish mothers consumed. Since ocean fish generally contain far more $\text{Se}$ than $\text{Hg}$, maternal seafood consumption resulted in fetal cord blood $\text{Se}$ increasing ~10 times faster than $\text{Hg}$. Therefore, instead of increasing risks of $\text{Hg}$-dependent diminishments of $\text{Se}$ availability, ocean fish consumption protects against loss of selenoenzyme activities. While seafood consumption is accompanied by small increases in cord blood $\text{Hg}$ relative to $\text{Se}$ and resulted in ~34% of cord blood samples in this cohort exceeding the EPA reference level of 5.8 ppb (0.029 μM), even the highest level of $\text{Hg}$ measured in these samples (0.112 μM $\text{Hg}$) would have negligible effects on availability of the 3.227 μM $\text{Se}$ present in that sample. Likewise, even the highest $\text{Hg}$ level observed in placenta tissues (0.100 μM) would hardly diminish availability of the 4.015 μM $\text{Se}$ present. Unlike umbilical cord blood, placental $\text{Se}$ is not correlated with $\text{Hg}$, possibly reflecting maintenance of adequate $\text{Se}$ levels by high LRP8 expression augmenting SELENOP uptake. Similar homeostatic mechanisms would be expected to maintain optimal $\text{Se}$ in fetal brain and endocrine tissues.

4.2. The biochemical context of CH₃Hg-dependent inhibition of brain selenoenzyme activities

Due to the high affinities of thiols for $\text{Hg}$ and the high tissue concentrations of Cys, CH₃Hg initially forms CH₃Hg-Cys. Since thiomolecules are cofactors and/or substrates for selenoenzymes such as glutathione peroxidases and thioredoxin reductases, CH₃Hg bound to these forms function as suicide substrates that directly deliver $\text{Hg}$ into selenoenzyme active sites in the proper orientation and proximity to react with the active site Sec. The greater affinity of $\text{Hg}$ for $\text{Se}$ and the additional potentiation of reactivity afforded by coordinated actions within the tertiary structure of the enzyme itself result in CH₃Hg transferring its covalent association from Cys to Sec. Since CH₃Hg is selectively delivered into the active site where it directly binds to the primary catalytic mediator, it is by definition a highly specific irreversible inhibitor. But in contrast to all other irreversible inhibitors in which the inhibitor is eventually released, $\text{Hg}$ binding to $\text{Se}$ is unique in that the bond that is formed is truly irreversible. Once $\text{HgSe}$ forms, aqua regia is the only acid that can decompose it. Much of the $\text{HgSe}$ that forms is deposited in cellular lysosomes where crystalline precipitates accumulate and demonstrate long term retention.

Because $\text{Hg}’\mathrm{s}$ affinity for $\text{Se}$ orders of magnitude higher than its affinity for thiols, it might be expected that CH₃Hg would sequester equimolar amounts at equilibrium, but this is not the case. Although $\text{Hg}’\mathrm{s}$ affinity for $\text{Se}$ is far greater, its affinity for thiols is still very high and their intracellular concentrations of ~300 mM are ~five orders of magnitude higher than the ~1 μM concentrations of $\text{Se}$. Therefore, thermodynamics compete with mass action effects to result in $\text{Hg}$ being almost equally distributed between cellular $\text{Se}$ and thiols at equilibrium. Since the approach to equilibrium is very slow under these conditions, the majority of $\text{Hg}$ which enters the body will exit without binding $\text{Se}$.

Selenocysteine is unique among genetically encoded amino acids in that it is not reused in subsequent cycles of selenoprotein synthesis. Instead, each cycle of their synthesis requires a sequence of steps in which existing Sec residues are degraded to release inorganic $\text{Se}$ in preparation for de-novo Sec synthesis concurrent with its incorporation in the primary structure of the selenoprotein being expressed (Berry et al., 2001; Xu et al., 2006; Labunskyy et al., 2014; Ha et al., 2019). As $\text{Se}$ is transported through maternal, placental, and fetal tissue compartments, this sequence of Sec degradation and resynthesis may be repeated many times. Selenium from dietary sources or residing in maternal tissues is used to create SELENOP which is carried in the plasma until it is acquired by the LRP8 expressed by the placenta and selectively internalized. The SELNOP eventually enters the fetal cord blood where it is taken up by tissues which express LRP8 (e.g., fetal brain) to ensure the $\text{Se}$ needs of these rapidly growing tissues are met. Any interruption of selenoenzyme synthesis would leave these tissues without the essential functions they require (e.g., preventing and/or reversing oxidative damage). In contrast, adult brain tissues have endogenous reserves of $\text{Se}$ and do not create new cells at the same rate as fetal brain. Therefore, $\text{Hg}$ exposures that cause serious impairments in fetal brain selenoenzymes will often leave the maternal brain unaffected (Watanabe 1999a; Watanabe 1999b).

An important aspect of CH₃Hg exposures is its ability to severely diminish $\text{Se}$ distribution through the placenta (Pařízek et al., 1971) and its unique capacity to cross placental and blood-brain barriers (Aschner and Clarkson, 1989) that enable it to impair selenoenzyme activities in fetal brain tissues (Watanabe 1999a; Watanabe 1999b). Other than genetic knockouts of LRP8 or SELENOP, toxic CH₃Hg exposures are the only environmental insult capable of inducing a $\text{Se}$-deficiency in the brain. Genetic knockouts and high CH₃Hg exposures both reduce the activities of key brain selenoenzymes such as glutathione peroxidase 4 (GPX4) and thioredoxin reductase (TXNRD), jeopardizing their ability to prevent and reverse oxidative damage. Impaired $\text{Se}$ distribution to the brain arising from either cause will result in $\text{Se}$-deficiencies in brain tissues which cannot survive without selenoenzyme protection. The loss of brain selenoenzyme activities result in adverse effects including sensorimotor dysfunctions, neurological damage, and cell death. Providing supplemental dietary $\text{Se}$ is able to prevent these consequences in genetic knockout (Burk et al., 2007; Burk et al., 2014) as well as CH₃Hg toxicity models (Ralston & Raymond, 2015; Stringari et al., 2008; Watanabe et al., 1999a). Supplemental $\text{Se}$ was recently applied to successfully treat a patient hospitalized with severe health effects as a result of a sustained high exposure to elemental $\text{Hg}$ (Spiller et al., 2017). The patient responded rapidly and recovered completely.

Evidence supporting the “$\text{Se}$-sequestration mechanism of $\text{Hg}$ toxicity” has developed through the past 50+ years and has become increasingly well established (Watanabe et al., 1999a; 1999b Seppänen et al., 2004; Stringari et al., 2008; Ralston et al., 2007; Ralston et al., 2008; Ralston and Raymond, 2015). While it had earlier been assumed that $\text{Hg}$ caused oxidative damage through direct reactions with biomolecules, this was disproven by Seppänen et al. (2004) who demonstrated $\text{Hg}$-dependent oxidative damage was the result of $\text{Hg}$-inhibition if the activities of selenoenzymes that prevent and reverse oxidation caused by ROS and related products of normal respiration. Soon afterwards, Carvalho et al. (2008) demonstrated CH₃Hg potently inhibited TXNRD activities (IC₅₀ of ~19.7 nM) and numerous in vitro and in vivo studies have since confirmed these findings (Carvalho et al. 2011; Usuki et al., 2011; Branco et al., 2014; 2017; Ralston & Raymond, 2015; Rodrigues et al., 2015; Branco et al., 2022).

Therefore, high CH₃Hg exposures during pregnancy are a significant concern if dietary $\text{Se}$ intakes are too low to offset losses due to $\text{Hg}$ sequestration. Homeostatic mechanisms readily mobilize $\text{Se}$ from tissue reserves as well as from dietary $\text{Se}$ intakes to continuously ensure the brain is provided with adequate $\text{Se}$. This controlled redistribution of $\text{Se}$ to supply the brain reveals the biochemical basis for the previously inexplicable latency period (Weiss et al., 2002) between $\text{Hg}$ exposure and the onset of signs and symptoms of toxicity.

Biochemical interactions between certain nutrients are already well described and synergistic effects are likely. For example, docosahexaenoic acid (DHA) is a fatty acid which is essential in the brain. Since DHA possesses 6 unsaturated double bonds which are vulnerable to oxidation, tissues with high DHA contents are highly dependent on selenoenzymes that prevent and reverse oxidative damage which spontaneously arises due to ROS and other reactive products of normal cell respiration. Similarly, $\text{Se}’\mathrm{s}$ vital role in deiodinase enzymes suggest iodine from seafoods may assist in ensuring healthy thyroid hormone balance is maintained during fetal growth. Another $\text{Se}$-containing nutrient, selenoneine, is a novel molecular form of $\text{Se}$ whose abundance in ocean fish was discovered just over a decade ago (Yamashita et al., 2010). Its physiology and roles in counteracting $\text{Hg}$ toxicity (Usuki et al., 2011) are being studied, but if its beneficial effects arise independently or if they occur through selenoenzyme metabolism remains unknown. It may be that degradation pathways release inorganic $\text{Se}$ for Sec synthesis in a pathway analogous to selenomethionine. Since the nutritional benefits of ocean fish consumption are substantial, there is a need to establish whether these effects are due to individual nutrients in ocean fish or are the aggregate effects of the total package of nutrients that ocean fish consumption provides (Liu and Ralston 2021).

4.3. The selenium perspective reveals the underlying consistency of otherwise “conflicting” study findings

Mercury toxicity associated with $\text{Hg:Se}$ exceeding equimolar stoichiometries is clearly observed in the umbilical cords saved from births occurring in the years before and during the Minamata poisoning incident as reported by Nishigaki and Harada, (1975). Samples collected prior to the poisoning incident had $\text{Hg:Se}$ molar ratios averaging 0.03:1.00, but those born during the peak $\text{Hg}$ exposure period contained ~4.66 μM, nearly 5 times more than their $\text{Se}$ contents (0.97 μM). Such high $\text{Hg}$ exposures would have retired tissue $\text{Se}$ as $\text{HgSe}$ and greatly reduced the availability of $\text{Se}$ required for selenoenzyme synthesis and activities. The effects of $\text{Hg}$ exposures in Minamata caused concern regarding exposures from other seafoods, so studies were initiated in sentinel populations with high $\text{Hg}$ intakes in New Zealand, the Faroe Islands, and the Seychelle Islands.

High maternal $\text{Hg}$ exposures from eating great white shark meat prepared as fish and chips (Mitchell et al., 1982) in New Zealand occurred at a time when its population had notably low $\text{Se}$-status (Thomson and Robinson, 1980). This would have enhanced their vulnerability to the effects of high exposures to $\text{Hg}$ and other persistent bioaccumulative toxicants PBTs that accumulate in great white shark meats (Crump et al., 1998). Similarly high exposures to $\text{Hg}$ along with other PBTs arose from maternal consumption of pilot whale blubber, muscle, and organ meats in the Faroe Islands (Grandjean et al., 1997; Grandjean et al., 1998). Subtle diminishments in child performance outcomes (Debes et al., 2006) were noted in association with increasing $\text{Hg}$, but it important to note that cord blood $\text{Hg}$ in the Faroe Islands approached equimolar stoichiometries with $\text{Se}$ (See Fig 1). While dietary $\text{Se}$ intakes in the Faroes were better than in New Zealand at the time of their seafood consumption study, both are poor compared to the United States where the RDA of 50 μg $\text{Se}$/day (~633 μmol $\text{Se}$/day) is typically met and usually exceeded, even by consumers that do not eat $\text{Se}$-rich ocean fish. The US EPA advisory is based on findings in the Faroe Islands, but it is important to note that pilot whale blubber, muscle, and liver provided 3.6, 68.9, and 69.6 μM $\text{Hg}$/day, resulted in ~85% of their total $\text{Hg}$ exposures with only ~15% arising from ocean fish (Ralston et al., 2015). However, the ocean fish they ate provided over 80% of their total $\text{Se}$ intake from seafood.

Pilot whales are apex marine predators that live 40-60 years and reach body weights of 5,000 pounds (~2,300 kg) at maturity (NOAA, 2023). Eating them not only results in high exposures to $\text{Hg}$, but also PCBs, dioxins, cadmium (Cd), and all other PBTs which accumulate in their tissues. Since Cd is also a soft electrophile that inhibits selenoenzymes (Splittgerber and Tappel, 1979; Ralston 2021b), it is worth noting that pilot whale blubber, muscle, liver, and kidney respectively contained: 7.1, 9.8, 106.8, and 55.1 μM Cd (Julshamn et al., 1987). These are in molar excess of $\text{Se}$ in all tissues except liver. It should be noted that the $\text{Hg}$, Cd, and related PBT exposures among pilot whale consumers in the Faroes are among the highest measured in any human population (Sharma et al., 2015; Anderson et al., 1987; Undeman, et al., 2018).

Since the biochemical mechanism of $\text{Hg}$ toxicity had not been identified, the BMDL for $\text{Hg}$ (58 ppb ≈ 0.29 μM) was based on data (Budtz-Jorgensen et al., 2000) which omitted consideration of $\text{Se}$. At the time, it was unclear whether the risks of $\text{Hg}$ exposures from whale meat were any different from eating ocean fish so the unexplained reasons for the contrasting findings of the major studies caused concern. In recognition of the numerous unknown aspects of the $\text{Hg}$ issue, a 10-fold uncertainty factor was applied to the BMDL as a precaution to ensure adequate protection of vulnerable subpopulations (National Research Council, 2000). The US Environmental Protection Agency (EPA) established 5.8 μg $\text{Hg}$/L as the reference level and pregnant and nursing mothers and young children were encouraged to eat no more than 12 oz (340 g) of ocean fish per week. While well intentioned, this warning has caused widespread anxiety regarding risks associated with eating ocean fish and diminished seafood consumption during pregnancy (Oken et al., 2003, Bloomingdale et al., 2010).

4.4. The Health Benefit Value in relation to mercury risk assessments

The need to provide a more reliably accurate and easy to interpret index of risk associated with $\text{Hg}$ exposures prompted development of the Health Benefit Value ($\text{HBV}$). The molar concentrations of $\text{Hg}$ and $\text{Se}$ in samples are used to determine the $\text{HBV}$ using the equation reported in Ralston and Raymond (2016):

$$\text{HBV}=((\text{Se}-\text{Hg})∕\text{Se})•(\text{Se}+\text{Hg})$$ (Equation 1)

Apex marine predators such as great white sharks, mako sharks, and pilot whales accumulate $\text{Hg}$ and other PBTs throughout their long lives and are among the rare types of “seafood” which contain more $\text{Hg}$ than $\text{Se}$ (Ralston et al., 2019) as shown in Figure 3. These species have $\text{HBVs}$ that become increasingly negative as they grow older and larger, and they should be avoided during pregnancy. While these seafoods have not been shown to have adverse effects on adults, their PBT burdens are notably high so consistent consumption of these forms is inadvisable. In contrast, the majority of ocean fish are considerably smaller, have shorter lives, and contain far more $\text{Se}$ than $\text{Hg}$ (see examples shown in Figure 3), and therefore have uniformly positive $\text{HBVs}$. It should also be noted that not all sharks have negative $\text{HBVs}$ since less predatory shark varieties such as threshers do not accumulate $\text{Hg}$ in excess of $\text{Se}$ and therefore have positive $\text{HBVs}$.

Fig 3.

Molar relationships between $\text{Hg}$ and $\text{Se}$ in various seafood types. The fish data represent varieties of commonly consumed fish collected from Hawaii fisheries (Kaneko and Ralston, 2007; Ralston et al., 2019) and pilot whale data are from Julshamn et al., (1987) collected around the time of the Faroes Study. The diagonal line indicates the 1:1 $\text{Hg:Se}$ molar ratio. (Note: 1 mg $\text{Se}$/kg = ~12.6 μM $\text{Se}$ and 1 mg $\text{Hg}$/kg = ~5 μM $\text{Hg}$).

The $\text{Se}$-perspective explains why maternal consumption of ocean fish with positive $\text{HBVs}$ are associated with beneficial outcomes while eating seafoods with negative $\text{HBVs}$ has been consistently associated with $\text{Hg}$-dependent risks. The $\text{HBVs}$ of great white sharks ($\text{HBV}=\text{-}120$) eaten in New Zealand or the pilot whales ($\text{HBV}=\text{-}80$) eaten in the Faroe Islands explain the adverse effects observed in those early studies. Although the average $\text{Hg}$ exposures reported in the Seychelles study (Myers and Davidson, 1998; Davidson et al., 2011; van Wijngaarden et al., 2013; van Wijngaarden et al., 2017) were higher than in the Faroes, beneficial effects were associated with increasing seafood intakes. Since nearly all of the ocean fish eaten in the Seychelles contain far more $\text{Se}$ than $\text{Hg}$ (Robinson and Schroff, 2004) and thus have positive $\text{HBVs}$, their findings are consistent with expectations.

Increasing maternal consumption of typical $\text{Se}$-rich varieties of ocean fish which are also notable sources of omega-3 fatty acids and other nutrients is associated with substantial improvements in neurodevelopmental outcomes (Myers and Davidson, 1998; Hibbeln et al., 2007; Hibbeln et al., 2019; Davidson et al., 2011; Julvez et al., 2016; Llop et al., 2017; Golding et al., 2017; van Wijngaarden et al., 2013; van Wijngaarden et al., 2017). Reviews of multiple large scale epidemiological studies have shown that increasing ocean fish consumption during pregnancy is associated with improvements in child IQs, scholastic abilities, and social skills (Avella-Garcia and Julvez, 2014; Starling et al., 2015; Spiller et al., 2019; Hibbeln, et al., 2019; Dietary Guidelines Advisory Committee, 2020). The most recent review of the effects of maternal fish consumption on child neurodevelopment describes findings from over 200,000 mother-child pairs which collectively report 51 beneficial associations with neurodevelopmental outcomes (Spiller et al., 2023).

Many health agencies continue to warn against $\text{Hg}$-exposure risks from eating typical varieties of ocean fish based on the outcomes of studies involving great white shark and pilot whale consumption, seafoods which never were available to most consumers. Just as great white sharks are no longer commonly consumed in New Zealand (Shark Facts, 2023), pilot whale meats are no longer consumed by pregnant women in the Faroe Islands (Weihe and Grandjean, 2012). However, consumers continue to face uncertainty and confusion arising from unclear and contradictory advice because of findings based on their consumption. This misattribution error contributes to confusion among health advisors and maternal avoidance of seafood during pregnancy (Oken et al., 2003, Bloomingdale et al., 2010) even though studies performed in the Seychelles and elsewhere (Myers and Davidson, 1998; Hibbeln et al., 2007; Davidson et al., 2011; Julvez et al., 2016; Llop et al., 2017; Golding et al., 2017; van Wijngaarden et al., 2013; van Wijngaarden et al., 2017) confirm ocean fish consumption is beneficial to child outcomes. To remedy these misattribution errors and the misplaced restrictions imposed by the current advisory, there is an urgent need to update maternal seafood consumption advisories to ensure future generations will benefit from improved prenatal access to essential nutrients provided by ocean fish.

4.5. Conclusions

This study confirms the hypothesis that maternal ocean fish consumption provides substantial amounts of $\text{Se}$ to protect against the risk of $\text{Hg}$ exposure diminishing fetal $\text{Se}$ bioavailability and selenoenzyme activities. The effects high $\text{Hg}$ exposures on $\text{Se}$ metabolism provide a consilient perspective of the seafood $\text{Hg}$ issue that explains the contrasting findings of previous studies. Since eating ocean fish will prevent rather than cause $\text{Hg}$-dependent diminishments in fetal $\text{Se}$ availability, maternal concerns regarding supposed risks associated with seafood consumption can be alleviated. However, in $\text{Se}$-poor regions where exposures to $\text{Hg}$ and other electrophiles from freshwater fish are likely to pose greater risks, they may currently be unheeded and contribute to pathologies which may not be adequately considered (Ralston, 2023).

The 10-fold uncertainty factor applied to findings from the Faroes study which appeared justified by the multiple unknowns regarding $\text{Hg}’\mathrm{s}$ mechanism of toxicity is no longer required now that its effects on normal $\text{Se}$ physiology in brain health and fetal development are understood. Furthermore, now that it is clear that studies indicating maternal $\text{Hg}$ exposures from eating pilot whale meats provide appropriate guidance for protecting whale eating populations, but do not apply to populations which eat ocean fish, it is time to revise public health policies to ensure future generations benefit from the essential nutrients which ocean fish provide.

Highlights:

• Selenium-dependent enzymes prevent and reverse oxidative damage in the brain and other tissues.

• Mercury inhibits selenium-dependent enzymes, resulting in neurological damage and dysfunction.

• Consumption of certain seafoods which contain more mercury than selenium may harm child health.

• Ocean fish consumption caused cord blood selenium to increase ~10 times faster than mercury.

• Therefore, eating selenium-rich ocean fish will directly counteract mercury toxicity.