Clinical Toxicology
ISSN: 1556-3650 (Print) 1556-9519 (Online) Journal homepage: http://www.tandfonline.com/loi/ictx20
Rethinking mercury: the role of selenium in the pathophysiology of mercury toxicity
Henry A. Spiller
To cite this article: Henry A. Spiller (2017): Rethinking mercury: the role of selenium in the pathophysiology of mercury toxicity, Clinical Toxicology, DOI: 10.1080/15563650.2017.1400555 To link to this article: http://dx.doi.org/10.1080/15563650.2017.1400555
Published online: 10 Nov 2017.
Submit your article to this journal
Article views: 3
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=ictx20
Download by: [American Academy of Clinical Toxicology] Date: 13 November 2017, At: 08:04 CLINICAL TOXICOLOGY, 2017 https://doi.org/10.1080/15563650.2017.1400555
REVIEW ARTICLE Rethinking mercury: the role of selenium in the pathophysiology of mercury toxicity
Henry A. Spillera,b aCentral Ohio Poison Center, Columbus, OH, USA; bDepartment of Pediatrics, College of Medicine, Ohio State University, Columbus, OH, USA
ABSTRACT ARTICLE HISTORY Introduction: There is increasing evidence that the pathophysiological target of mercury is in fact sel- Received 17 July 2017 enium, rather than the covalent binding of mercury to sulfur in the body’s ubiquitous sulfhydryl Accepted 31 October 2017 groups. The role of selenium in mercury poisoning is multifaceted, bidirectional, and central to under- Published online 9 November standing the target organ toxicity of mercury. 2017 Methods: An initial search was performed using Medline/PubMed, Toxline, Google Scholar, and Google KEYWORDS for published work on mercury and selenium. These searches yielded 2018 citations. Publications that Mercury; methylmercury; did not evaluate selenium status or evaluated environmental status (e.g., lake or ocean sediment) were selenium; poisoning; excluded, leaving approximately 500 citations. This initial selection was scrutinized carefully and 117 of selenoprotein; toxicity the most relevant and representative references were selected for use in this review. Binding of mercury to thiol/sulfhydryl groups: Mercury has a lower affinity for thiol groups and higher affinity for selenium containing groups by several orders of magnitude, allowing for binding in a multifaceted way. The established binding of mercury to thiol moieties appears to primarily involve the transport across membranes, tissue distribution, and enhanced excretion, but does not explain the oxidative stress, calcium dyshomeostasis, or specific organ injury seen with mercury. Effects of mercury on selenium and the role this plays in the pathophysiology of mercury tox- icity: Mercury impairs control of intracellular redox homeostasis with subsequent increased intracellular oxidative stress. Recent work has provided convincing evidence that the primary cellular targets are the selenoproteins of the thioredoxin system (thioredoxin reductase 1 and thioredoxin reductase 2) and the glutathione-glutaredoxin system (glutathione peroxidase). Mercury binds to the selenium site on these proteins and permanently inhibits their function, disrupting the intracellular redox environ- ment. A number of other important possible target selenoproteins have been identified, including sele- noprotein P, K, and T. Impairment of the thioredoxin and glutaredoxin systems allows for proliferation intracellular reactive oxygen species which leads to glutamate excitosis, calcium dyshomeostasis, mito- chondrial injury/loss, lipid peroxidation, impairment of protein repair, and apoptosis. Methylmercury is a more potent inhibitor of the thioredoxin system, partially explaining its increased neurotoxicity. A second important mechanism is due to the high affinity of mercury for selenium and the subsequent depletion of selenium stores needed for insertion into de novo generation of replacement selenopro- teins. This mercury-induced selenium deficiency state inhibits regeneration of the selenoproteins to restore the cellular redox environment. The effects of selenium on mercury and the role this plays in biological response to mercury: Early research suggested selenium may provide a protective role in mercury poisoning, and with limita- tions this is true. The roles selenium plays in this reduction of mercury toxicity partially depends on the form of mercury and may be multifaceted including: 1) facilitating demethylation of organic mer- cury to inorganic mercury; 2) redistribution of mercury to less sensitive target organs; 3) binding to inorganic mercury and forming an insoluble, stable and inert Hg:Se complex; 4) reduction of mercury absorption from the GI tract; 5) repletion of selenium stores (reverse selenium deficiency); and 6) res- toration of target selenoprotein activity and restoring the intracellular redox environment. There is con- Downloaded by [American Academy of Clinical Toxicology] at 08:04 13 November 2017 flicting evidence as to whether selenium increases or hinders mercury elimination, but increased mercury elimination does not appear to be a major role of selenium. Selenium supplementation has been shown to restore selenoprotein function and reduce the toxicity of mercury, with several signifi- cant limitations including: the form of mercury (methylmercury toxicity is less responsive to amelior- ation) and mercury dose. Conclusions: The interaction with selenium is a central feature in mercury toxicity. This interaction is complex depending on a number of features such as the form of mercury, the form of selenium, the organ and dose. The previously suggested “protective effect” of selenium against mercury toxicity may in fact be backwards. The effect of mercury is to produce a selenium deficiency state and a direct inhibition of selenium’s role in controlling the intracellular redox environment in organisms. Selenium supplementation, with limitations, may have a beneficial role in restoring adequate selenium status from the deficiency state and mitigating the toxicity of mercury.
CONTACT Henry A. Spiller [email protected], www.linkedin.com/in/henryspiller/ Central Ohio Poison Center, 700 Childrens Dr, Columbus, OH 43205, USA ß 2017 Informa UK Limited, trading as Taylor & Francis Group 2 H. A. SPILLER
Introduction selenocystine, and selenocysteine. These searches yielded 2018 citations. Publications that did not evaluate selenium There has been an important shift in the understanding of status or evaluated environmental status (e.g., lake or ocean the mechanisms of toxicity of mercury both at the cellular sediment) were excluded, leaving approximately 500 cita- and organism level. The shift in a large part has occurred tions. This initial selection was scrutinized carefully and 124 from a long-held focus on the covalent binding of mercury of the most relevant and representative references were to sulfur in the body’s ubiquitous sulfhydryl groups [1,2]. selected for use in this review. There is increasing and convincing evidence that the patho- physiological target of mercury is not the in vivo binding of sulfur, but rather selenium [3–5]. This may not be surprising Binding of mercury to thiol/sulfhydryl groups since the binding affinity between mercury and selenium is several orders of magnitude greater than the affinity of mer- Decades of work have focused on the interaction between cury and sulfur [5,6–9]. Previous reviews of mercury toxicity mercury and sulfhydryl moieties, including cysteine, homo- have been published and this review is not intended to cysteine, s-adenosylmethionine, metalothionine, and glutathi- readdress well established kinetics, clinical effects or out- one [1]. The primary outcome of the binding to sulfhydryl comes. Rather, this review will focus on growing understand- groups appears to be increased transport across membranes and enhanced excretion. Early work on the monovalent orga- ing of the role selenium plays in the pathophysiology of þ mercury toxicity. nomercury methylmercury (MeHg ) suggested binding of the The role of selenium in mercury poisoning is multifaceted, sulfur moiety on cysteine allowed the MeHg-cysteine com- bidirectional and central to understanding the target organ plex to become a substrate for the L-type large neutral toxicity of mercury. An addition critical feature of mercury amino acid transporters 1 and 2 (LAT-1 and LAT-2) [13,14]. toxicity and of the interaction with selenium is the chemical The MeHg-cysteine complex mimics methionine as a sub- state of mercury (e.g., Hg0,Hgþ, and Hgþ2). The mechanism strate for L-type large neutral amino acid transporters allow- of toxicity of mercury is based on its ability to bind to and in ing it to readily cross the blood brain and placental barrier certain cases inhibit moieties containing sulfur and selenium and cell membrane [15]. [3–5,10]. Mercury has a lower affinity for thiol groups and However, the divalent inorganic form of mercury (e.g., þ2 higher affinity for selenium containing groups, allowing for mercuric chloride, Hg ) has difficulty penetrating the CNS or binding in a multifaceted way. The relative ease with which crossing cell membranes suggesting L-type large neutral þ2 mercury can move from one thiol group to another allows amino acid transporters are not involved in Hg transport. for movement from one binding group to another, transport Kageyama et al. [16] suggested intracellular sulfhydryl group across membranes, and potentially temporarily impairment of inhibition based on indirect evidence of increased lympho- þ multiple proteins [10]. The binding of mercury to selenium cyte death from MeHg . However, in this same, in vitro on selenoproteins and selenoenzymes is of higher affinity by model established direct sulfhydryl group inhibitors did not several orders of magnitude and more stable when com- cause cell death. Additionally, they did not evaluate the þ pared to its binding with thiol groups, making these the final effect of MeHg on lymphocyte glutathione peroxidase, a or primary target of mercury [5,11,12]. selenoenzyme that is known to be critical to lymphocyte This review will be divided into three sections: activity and far more likely to be responsible for the effects seen [17]. A brief discussion of thiol-mercury binding. This is provided Kung et al. [18] showed only moderately small effects of þ because of its effect on mercury distribution and had previ- MeHg on sulfhydryl group enzymes, suggesting some other ously been suggested as a mechanism of toxicity; biological mechanism altering enzyme activity. In vitro work Effects of mercury on selenium and the role this plays in the by Magour [19] suggested binding to the sulfhydryl groups on pathophysiology of mercury; and synaptosomal Na/K ATPase producing inhibition of Na/K Effects of selenium on mercury and the role this plays in bio- ATPase. However, the clinical significance of this finding is Downloaded by [American Academy of Clinical Toxicology] at 08:04 13 November 2017 logical response to mercury. unclear as evidence of alterations of sodium, potassium, or the ATPase pump has not been found at the organism level. It has been suggested that the mechanism for elevated circu- Methods lating catecholamines (producing the hypertension An initial search was performed using Medline/PubMed, and tachycardia seen in acrondynia) may be inhibition of cat- Toxline, Google Scholar, and Google for published work on echolamine-O-methyltransferase by binding to the thiol moi- selenium and mercury. As mercury and selenium are some- ety s-adenosylmethionine, a co-factor of catecholamine- times listed in unique forms (e.g., dimethylmercury, diphenyl O-methyltransferase [20,21]. However, specific inhibition of diselenide, selenoprotein, etc.) multiple searches were per- catecholamine-O-methyltransferase has not been documented formed using combinations of text words that included a and in fact incubation of placental tissue with HgCl showed form of mercury and a form of selenium: mercury, mercury an increase in catecholamine-O-methyltransferase activity [22]. toxicity, mercury poisoning, inorganic mercury, organic mer- S-adenosylmethionine is widely available in all cells with cury, methylmercury, dimethylmercury, mercuric chloride, sel- continuous rapid regeneration [23]. From a stoichiometric enium, selenide, selenite, diphenyl diselenide, selenoprotein, perspective, it is unlikely that circulating mercury could cause inorganic selenium, organic selenium, selenomethionine, such depletion for the long duration this effect has been CLINICAL TOXICOLOGY 3
seen. An alternate mechanism for elevated circulating cate- The mechanism of toxicity of mercury is based its ability to cholamines is inhibition of selenium stores in adrenal chro- impair control of intracellular redox homeostasis and the sub- maffin cells and subsequent oxidative damage (see later sequent increased intracellular oxidative stress. Recent work discussion mercury effects on selenium). Mercury, as either by several groups has provided convincing evidence that the þ þ MeHg or Hg 2, bound to cysteine or homocysteine (thiol primary cellular targets are the selenoproteins of the thiore- moieties) may act as a substrate for the organic ion trans- doxin system (thioredoxin reductase 1) in the cytosol and thi- porter 1 (OAT1) in the renal tubular cells, allowing for the oredoxin reductase 2 in the mitochondria) and the uptake into the kidneys and the high renal mercury burden glutathione–glutaredoxin system (glutathione peroxidase) [24]. An important thiol that has been proposed as the mech- [4,11,12,46]. However, a number of other important possible anism of toxicity of mercury in both the kidneys and brain is target selenoproteins have been identified, including seleno- glutathione [24,25]. protein P, K, and T [29,41,47–49]. Multiple studies have shown reduction of glutathione in renal, hepatic, and neural cell studies [12,26–28]. This occurs – as glutathione is mobilized in response to the oxidant stress Thioredoxin system and the glutathione glutaredoxin effects of mercury and the glutathione binding required for system transport of mercury (liver and kidneys). Additionally, gluta- The thioredoxin system together with the – thione may be used in the process of formation of the inert glutathione glutaredoxin system (glutathione reductase and Hg:Se moiety [9,29]. MeHgþ and to a much lesser extent glutathione peroxidase) control the cellular redox environ- inorganic mercury (Hgþ2) binds to glutathione in the liver ment [50]. Impairment of the thioredoxin and glutaredoxin and are excreted in the bile by carriers for glutathione [30]. systems allows for proliferation of cytosol and mitochondrial Excretion via bile into the feces is the primary route of elim- reactive oxygen and nitrogen species which lead to mito- ination for MeHgþ [31,32]. chondrial injury/loss, lipid peroxidation, calcium dyshomeo- A second elimination pathway for MeHgþ that increases stasis, impairment of protein repair, and apoptosis [50–52]. þ þ2 over time is demethylation and clearance via the kidneys as Both MeHg and Hg bind to the selenocysteine binding inorganic mercury [31,32]. In the kidneys, glutathione binding site of thioredoxin reductase 1 in the cytosol and thioredoxin of inorganic Hgþ2 (or co-transport of glutathione with Hgþ2) reductase 2 in the mitochondria significantly inhibiting their is required for excretion from the proximal tubules, likely via function [11]. The non-selenium containing enzymes in these the multidrug resistance protein (MRP2) [33,34]. redox groups, such as glutathione reductase, show no þ Metallothionein (a thiol moiety) plays an important role in changes in activity after MeHg exposure [12,53]. the initial binding of mercury in astrocytes in the brain, hepa- The inhibition of thioredoxin reductase impairs the cyclical tocytes and renal tubular cells in the kidney and may be the regeneration of thioredoxin activity, as thioredoxin remains first protective mechanism [1,35–38]. Additionally metallo- in the oxidized state [1,50]. Thioredoxin can be reduced to thionein plays a role in Hg detoxification of mercury in the restore function by glutathione peroxidase, which acts as a liver and kidney, primarily as a scavenger [38]. However, with backup redundant system [12,50,54]. Unfortunately, glutathi- increases in dose and over time, mercury shifts from these one peroxidase is also a selenoprotein target that is inacti- low weight thiol proteins to the higher affinity selenium con- vated by increasing concentrations of mercury [12,54–56]. taining selenoproteins [35,39–41]. In summary, the estab- An important cellular response to increased oxidative lished binding of mercury to thiol moieties appears to stress is upregulation and transposition of nuclear factor E-2- primarily involve the transport across membranes (e.g., related factor 2 (Nrf2) to the nucleus to stimulate increased þ Human L-type amino acid transporter 1 and 2 (LAT1, LAT2) synthesis of new thioredoxin reductase production. MeHg is and glutathione carriers), tissue distribution and enhanced a more potent inhibitor of the thioredoxin system (both thio- þ excretion (e.g., metallothionein, multidrug resistance protein), redoxin reductase 1 and thioredoxin reductase 2) than Hg 2 but does not explain the oxidative stress, calcium dyshomeo- [3]. Additionally, upregulation and transposition of Nrf2 þ þ2 Downloaded by [American Academy of Clinical Toxicology] at 08:04 13 November 2017 stasis or specific organ injury seen with mercury. occurs significantly slower with MeHg than Hg [3]. Branco et al. [3] measured cytosol nuclear factor E-2-related factor 2 concentrations at 12 and 24 h in an in vitro hepatic cell line Effects of mercury on selenium and the role this plays in þ þ comparing response to Hg 2 and MeHg . In this hepatic cell the pathophysiology of mercury toxicity line and measured at 12 and 24 h after selenium supplemen- A critical feature of mercury toxicity and of the interaction tation as selenite, thioredoxin reductase activity increased (de with selenium is the chemical state of mercury (Hg0,Hgþ, novo production) and Nrf2 transposition had occurred by þ and Hgþ2). The chemical state and valance of mercury influ- 24 h, but not 12 h, in the Hg 2 line. However, under the þ ences kinetics, target organs, clinical patterns, and outcomes. same conditions but with MeHg , no evidence of nuclear Additionally, there are noted differences between kinetics factor E-2-related factor 2 transposition was detected at 12 or and tissue distribution in the animal models frequently used 24 h. The mechanism for the slower nuclear factor E-2-related for research and in humans [42]. factor 2 response has not been elucidated. However, the The injury from mercury on the organism is a result of oxi- combined effects of greater inhibition of thioredoxin reduc- dative stress and the primary target organs are the brain, kid- tase activity and reduced nuclear factor E-2-related factor neys, and to a lesser extent the liver [10,43–45]. induced regeneration of de novo thioredoxin reductase may 4 H. A. SPILLER
Figure 1. Inhibition of de novo thioredoxin reductase and glutathione peroxidase production.
be a factor in the increased neurotoxicity of MeHgþ com- dietary MeHgþ exposure (9 weeks and 18 weeks) three pared with inorganic mercury. groups were compared; controls (MeHg-free and Se-free), A final step in regeneration of thioredoxin activity is gen- MeHgþ added (Se-free), and MeHgþ with Se supplementa- eration of de novo thioredoxin reductase which requires tion. Comparing the MeHgþ poisoned group (Se-free) with insertion of selenium (as a selenocysteine moiety) into the controls there was a decrease in brain selenium concentra- new enzyme from an available pool of selenium. Mercury tion of 60% (60 micromol MeHg/kg/day for 9 weeks) and binds to available intracellular selenium to form an insoluble 43% (50 micromol MeHg/kg/day for 18 weeks) and clear evi- Hg–Se, MeHg–Se, or Hg–Se–cysteine complex, producing an dence of MeHg-induced neurotoxicity in the MeHgþ group, intracellular selenium deficiency state and reducing availabil- as measured by loss of muscle control and early death. ity of selenium for de novo thioredoxin reductase and gluta- Comparing the MeHgþ with selenium supplementation thione peroxidase production, further inhibiting restoration group (50 micromol MeHg/kg/day and 10 micromol Se/kg/ of thioredoxin and glutaredoxin systems [45,57,58](Figure 1). day) with controls showed a 300% increase in brain selenium In the presence of supplemental selenium, de novo regen- content and no evidence of MeHg-induced neurotoxicity eration of thioredoxin reductase 1 concentrations can be [57,60]. This “selenium-stripping” effects was not seen in the restored via increased synthesis, although to a lesser degree blood, kidney, or liver. in MeHgþ compared with Hgþ2 [3,11,59]. Potential direct Downloaded by [American Academy of Clinical Toxicology] at 08:04 13 November 2017 reactivation of thioredoxin reductase by selenium (as 2 Mercury induced mitochondrial injury selenide as opposed to selenocysteine) has been proposed, with the selenide acting as a chelate to remove Hg from the A key cellular injury of mercury secondary to thioredoxin active thioredoxin reductase binding site, subsequently form- reductase and glutathione peroxidase impairment is the ing an Hg–Se chelate and restoring enzyme function [59]. This reduction/loss of mitochondria and fusion of remaining mito- effect was only seen with inorganic mercury (Hgþ2). Selenium chondria organelles to preserve energy and respiratory func- supplementation did not improve de novo regeneration of thi- tion [51,61,62]. Supplementation with selenium, either as oredoxin reductase 2 (mitochondrial thioredoxin reductase) diphenyl diselenide or sodium selenite, preserved, or restored – [3]. MeHgþ may need to be demethylated to remove its inhib- mitochondrial content and function [51,52,63 65]. Glaser ition of thioredoxin reductase2 regeneration [59]. et al. [51] showed a reduction of 60% of mitochondrial con- tent in a murine model using MeHgþ. In a 21 day, murine model of MeHgþ toxicity comparing (1) controls (no MeHgþ Mercury-induced selenium deficiency state or Se), (2) with MeHgþ (40 ppm MeHgþ in freely Support for a mercury-induced selenium deficiency state was available water), and (3) MeHgþ with diphenyl diselenide reported by Ralston et al. [57,60] In a rat model of chronic supplementation (40 ppm MeHgþ in freely available water CLINICAL TOXICOLOGY 5
Figure 2. Inhibition of glutamate uptake by astrocytes and neurons.
and 5 mmol//kg diphenyl diselenide SQ daily injection) brain protective mechanism with their high content of metallothio- mitochondrial content was reduced 60% by MeHg compared nein (a thiol); however, this process is saturable and excess with controls, while brain mitochondrial content was mercury moves from the thiol to the selenium bonds (e.g., increased 57% over controls in the MeHg and Se group [51]. thioredoxin reductase and glutathione peroxidase). Mercury One proposed mechanism for the increased mitochondrial localization in the brain corresponds closely with regions biogenesis post-selenium supplementation was evidence where metallothionein protein concentrations are highest showing the increased production and transposition of and may help explain the distribution in the brain post-mer- nuclear factor E-2-related factor 1 and 2 [51,52]. In addition cury exposure [36,66]. to regulating the production of cytoprotective selenoen- zymes thioredoxin reductase and glutathione peroxidase, nuclear factor E-2-related factor 1 and 2 regulate mitochon- Impairment of cellular redox control Downloaded by [American Academy of Clinical Toxicology] at 08:04 13 November 2017 drial biogenesis. Selenium alone, in the absence of mercury The first stage in the cascade of neurologic injury is the poisoning, stimulates nuclear factor E-2-related factor upre- increased accumulation of reactive oxygen species (i.e., H2O2) gulation and mitochondrial biogenesis [52]. The proposed in the cell secondary to impairment of the cellular redox con- mechanism for the upregulation of nuclear factor E-2-related trol via thioredoxin reductase and glutathione peroxidase factor 1 and 2 is the direct phosphorylation of protein kinase inhibition. The increased intracellular H2O2 or superoxide B on the genome either by selenium or selenocysteine [52]. anion produces a redox sensitive inhibition of glutamate uptake by astrocytes and inhibition of vesicular glutamate reuptake by neurons [68–70](Figure 2). In an in vitro model Cascade producing neurologic injury using rat astrocyte cultures, the inhibition of glutamate The cascade producing the neurologic injury at the organ uptake by astrocytes is primarily a Hgþ2 effect at physiologic- level from mercury exposure has been proposed by Farina ally relevant concentrations of mercury [70]. However, inhib- et al. [45]. In the brain, MeHgþ is preferentially taken up by ition of vesicular glutamate reuptake by neurons occurs with þ astrocytes [36,66]. Hgþ2 does not cross the blood–brain bar- MeHg [69]. The persistence of glutamate in the synapse rier well, but once in the brain appears to be taken up by after reuptake inhibition ultimately leads to overstimulation neurons [67]. This uptake by astrocytes may be an initial of N-methyl-D-aspartate type receptors (NMDA). 6 H. A. SPILLER
Figure 3. Calcium dyshomeostasis and overproduction of reactive oxygen species. Downloaded by [American Academy of Clinical Toxicology] at 08:04 13 November 2017
Calcium dyshomeostasis and overproduction of reactive Of additional interest in this pathological cycle is seleno- oxygen species protein K, an endoplasmic reticulum protein involved in cal- Downstream, as a consequence of NMDA receptor overstimu- cium homeostasis [49]. Impairment of selenoprotein K lation there is an increase in neuronal intracellular and mito- produces increased intracellular free calcium from the endo- chondrial free calcium concentrations leading to plasmic reticulum [49]. Interestingly, in a cell culture line, sel- overproduction of reactive oxygen (ROS) and nitrogen spe- enium supplementation as selenomethionine increased cies (in an already challenged redox environment) with the selenoprotein K expression and restored calcium homeostasis þ þ subsequent pathological cycle from glutamate exocytosis [49]. It is unknown to what extent mercury (MeHg or Hg 2) [71,72](Figure 3). Overproduction of ROS, in the presence of impairs this selenoprotein, but impairment of selenoprotein K loss of redox control by the inhibited selenoproteins, results would produce an exaggerated calcium dyshomeostasis after in mitochondrial loss (Figure 3), lipid peroxidation, protein NMDA receptor overstimulation, potentially worsening the oxidation, and if severe, neuronal loss. The distribution of pathological cycle (Figure 3). Additional indirect support neurological injury and neuron loss is consistent with areas comes from Tan et al. [73] and Limke et al. [74]. In in vitro of high density of NMDA receptors [36]. cell cultures of T-lymphocytes and cerebellar neurons, CLINICAL TOXICOLOGY 7
MeHgþ increased cytosolic free calcium concentrations from showed an increase in catecholamine-O-methyltransferase the endoplasmic reticulum [73,74]. activity [22]. Selenoprotein P is a Se-rich glycoprotein (10 selenocys- teines) produced in the liver and found mainly in the serum Mercury induced renal injury that functions as the primary Se transport/supplier to cells The kidney is the second target organ, with more mercury [84,85]. Selenoprotein P production is upregulated after mer- accumulation per organ weight in the kidneys than any other cury exposure in the presence of adequate or supplemental organ [1]. The toxic moiety in the kidneys is primarily inor- selenium supply. This upregulation may be a biological þ2 ganic mercury in the form of Hg but methylmercury- response to increased brain demand for selenium after mer- induced injury may also occur [1]. The primary renal cury-induced selenium deficiency [53]. Blood and other tis- responses to mercury are thiol based and are protective. sues such as liver and muscle preferentially redistribute After entry into tubular epithelial cells (via Organic ion trans- selenium to the brain and endocrine tissues in periods of sel- porter 1), mercury binds to metallothionein, due to its prefer- enium deficiency [57,86]. Selenoprotein P is preferentially ence for divalent cations, and with glutathione for excretion absorbed by the brain but no other tissues in selenium defi- [75]. This is initially a protective mechanism, but appears to cient animals [85]. þ be a saturable process, as metallothionine production is not In MeHg poisoned rats and mice without selenium sup- readily upregulated [37]. With increasing mercury concentra- plementation, a decrease in Selenoprotein P occurs (likely an tion these thiol-based responses are overwhelmed and mer- early marker for onset of selenium deficiency) prior to onset cury is available to bind to and deplete the selenoproteins of oxidative stress and neurodegeneration [40,53]. In a mur- þ þ thioredoxin reductase 1, thioredoxin reductase 2, and gluta- ine MeHg poisoning model comparing MeHg versus þ þ thione peroxidase initiating the subsequent oxidative stress MeHg with low dose Se and MeHg with high dose Se, and cellular injury including renal tubular necrosis [55,76–78]. serum mercury increased 53 and 75%, respectively [40]. With selenium supplementation there was an increase in both serum mercury and selenoprotein P, with the increased Selenoproteins K, P, and T serum mercury bound to the selenoprotein P as an Hg:Se There are several other selenoprotiens that deserve mention: complex [40]. However, an important feature is the mercury selenoprotein K (brain) selenoprotein P (serum), and seleno- bound to selenoprotein P is in the form of an Hg:Se or protein T (endocrine). Selenoproteins K and T are both MeHg:Se complex, as selenoprotein P does not bind free Hg located on the membrane of the endoplasmic reticulum and [40,87–89]. In a rat model evaluating the effects of selenium þ are involved with calcium release and maintaining intracellu- supplementation administered after 4 weeks of MeHg lar calcium homeostasis [47,49]. Impairment of selenoproteins administration had ceased and compared with controls þ K and T both result in increased cellular free calcium. The (MeHg but no SE supplementation) serum Hg concentra- effects of impairment of selenoprotein K in the brain was dis- tions increased 22 and 32% at 30 and 90 days, respectively cussed previously, with potential exacerbation of calcium [35]. Similar effects have been seen in humans [85]. This sug- dyshomeostasis occurring after NMDA receptor overstimula- gests an additional role for selenoprotein P of redistribution tion. Selenoprotein T has been identified on the membrane of Hg away from target organs of the brain and kidney and “ ” of the endoplasmic reticulum in the chromaffin cells of the possible sink effect [35,40,85]. adrenal medulla. Inhibition of selenoprotein T results in increased cellular free calcium and increased catecholamine release [47,79,80]. Increases in catecholamine release appear The effects of selenium on mercury and the role this to occur between >1 and 10 mmol of Hg [81]. This mechan- plays in biological response to mercury ism appears to be a better explanation of the tachycardia Ganther and Parizek [43,90] initially suggested selenium may and hypertension (with increased circulating catecholamines) be protective of mercury toxicity. In dietary MeHgþ exposure Downloaded by [American Academy of Clinical Toxicology] at 08:04 13 November 2017 seen in acrodynia. in cats and quail, groups with selenium supplementation in Additionally, a decrease in glutathione peroxidase activity the diet showed reduced incidence of toxicity and none of – – þ in the adrenal cortex impairs the hypothalamic pituitary a- the fatalities seen with the MeHg alone groups [43]. Similar drenal axis (HPA axis) [79,82]. Selenium deficiency reduced “protective” effects have been reported in additional animal adrenal glutathione peroxidase (selenoenzyme) activity by models and humans [57,90–93]. In healthy marine mammals, >80% [83]. Inhibition of thioredoxin reductase 1, thioredoxin marine fish, and humans, the selenium:mercury molar ratio reductase 2 and glutathione peroxidase has been docu- approximates 1:1 [39,94–97]. This equimolar balance appears mented in renal and neuroendocrine chromaffin cells. The to occur in vivo, despite molar differences in dietary Hg and previous speculation on the mechanism for the tachycardia Se consumption [39,98]. and hypertension was inhibition of catecholamine-O-methyl- The role of selenium may play in this reduction of mer- transferase by binding to the thiol moiety s-adenosylmethio- cury toxicity partially depends on the form of mercury and is nine, a co-factor of catecholamine-O-methyltransferase likely to be multifaceted including: [20,21]. However, specific inhibition of catecholamine-O- methyltransferase has not been documented in mercury facilitating demethylation of organic mercury to inorganic exposure and in fact incubation of placental tissue with HgCl mercury; 8 H. A. SPILLER