From experiments in mice in the 1970s, we learned that not only was selenium (Se) supplementation able to affect the immune response(Reference Spallholz, Martin and Gerlach1) but data from the 1990s also showed that Se status influenced the ability of mice to deal with viral infection(Reference Beck, Handy and Levander2). Part of the reason for those responses rests with the fact that Se, the only trace element to be specified in the genetic code, is an integral part of the large selenoprotein family; the mouse genome contains twenty-four genes that encode selenoproteins while the human genome contains twenty-five(Reference Rayman3,Reference Santesmasses, Mariotti and Gladyshev4) . Selenoproteins carry out the major nutritional roles of Se and have a range of effects from antioxidant, redox-active, anti-inflammatory, anti-ferroptosis, transport of Se, reduction of endoplasmic reticulum (ER) stress, to the production of active thyroid hormone(Reference Rayman3). Many of these responses are relevant to the role of Se in dealing with viral infection as shown in Table 1(Reference Zhang, Saad and Taylor5).
Table 1. Multiple selenoprotein functions relevant to viral infection(Reference Zhang, Saad and Taylor5)
ER, endoplasmic reticulum; GPX, glutathione peroxidase; MSRB1, methionine sulfoxide reductase B1; selenoproteins F–W are represented by SELENOF – SELENOW; TXNRD, thioredoxin reductase.
In this review, we explore the ways in which selenoproteins and redox-active Se species interact with RNA viruses, including severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), to reduce adverse health effects such as coronavirus disease (COVID-19).
Se intake and status
Se is present in trace amounts in both organic and inorganic forms in marine and freshwater systems, soils, biomass and in the atmosphere(Reference Winkel, Vriens and Jones6). In the terrestrial environment, rocks and minerals at the Earth's surface are a primary source of plant Se and depend on the physico-chemical properties that control the movement of Se in the food chain, e.g. soil pH, organic content and speciation(Reference Johnson, Fordyce and Rayman7). However, the atmosphere is also an important and dynamic reservoir of Se supplying agricultural soils and terrestrial and marine ecosystems(Reference Winkel, Vriens and Jones6). In the Se-poor belt in China which ranges from north-east to south-west, Se status is decisively affected by monsoonal precipitation(Reference Blazina, Sun and Voegelin8).
There is an extremely wide range of intake of Se seen across the world(Reference Winther, Rayman and Bonnema9); it is high in Venezuela, Canada, the USA and Japan and much lower in Europe(Reference Rayman3). China has areas of both Se deficiency and excess where plasma concentrations of Se range from 22 to 550 μg/l(Reference Xia, Hill and Byrne10,Reference Yang and Xia11) . Recommendations for Se intake, which are based on optimising plasma glutathione peroxidase (GPX3) activity, average 60 μg daily for men and 53 μg daily for women(Reference Rayman12).
Organ meats (e.g. kidney and liver) are a rich source of Se. In addition, seafood and muscle meat are also good sources(Reference Rayman13). The Se content of cereals and grains, however, varies widely, ranging from very low in much of Europe (e.g. mean values of 0⋅025–0⋅033 mg/kg dry weight in the UK) to as much as 30 mg/kg in high-Se areas of the USA(Reference Rayman13).
The effect of Se deficiency on RNA viruses
Se-deficient environments favour replication, virulence and mutation of RNA viruses(Reference Guillin, Vindry and Ohlmann14). Viral infections produce reactive oxygen species (ROS) and disturb the balance between the generation of ROS and their scavenging systems(Reference Guillin, Vindry and Ohlmann14). Selenoproteins such as the GPXs and thioredoxin reductases (TXNRD) are an important part of those scavenging systems(Reference Guillin, Vindry and Ohlmann14). In Se deficiency, GPX will fall and viral replication will increase. The more the virus replicates (e.g. in an immunosuppressed person, or in a Se-deficient population), the more viral copies will be generated per unit time, which, applying the multiplier from the basal mutation rate, leads to an increased pool of mutants because of the larger viral population size, thus providing further opportunity for selection of a more fit viral strain(Reference Moya, Elena and Bracho15,Reference Domingo16) . Perhaps the best known example of this effect is the often-stated hypothesis, supported by abundant evidence, that some of the newer highly mutated variants of SARS-CoV-2 may have developed in the body of a single immunocompromised individual; such observations were being made even before the emergence of the hypermutated Omicron variant(Reference Corey, Beyrer and Cohen17).
Specific instances of an effect of host Se deficiency on RNA viruses or associated diseases
Coxsackie virus and Keshan disease
In the 1990s, Beck and colleagues showed that Se deficiency in mice, with its associated low or absent activity of protective GPX1, caused mutations in a non-virulent stain of coxsackie virus B3 that lead to the development of a virulent strain that caused myocarditis, even in a subsequent mouse that was Se-sufficient(Reference Beck, Shi and Morris18). Those findings showed that host nutritional status could influence not only the host response to the pathogen, but also the genetic make-up of the viral genome(Reference Beck, Shi and Morris18). They were also able to explain the appearance of an endemic cardiomyopathy known as Keshan disease in a Se-deficient area in north-east China detected in the early 1930s which mainly affected infants, children and women of childbearing age(Reference Li19). In the 1970s, Se supplementation of the diet was given to the people of those areas completely eradicating this disease in those provinces of China(Reference Guillin, Vindry and Ohlmann14). Coxsackie virus was subsequently isolated from the archived blood and tissues of patients with Keshan disease showing that Se deficiency in human subjects affects the viral genome in the same way as in mice resulting in viral mutation and the development of heart pathology(Reference Huang, Xia and Jin20).
Influenza virus
Significantly more harmful lung inflammation developed in Se-deficient mice infected with a mild strain of influenza, influenza A/Bangkok/1/79 (H3N2), than in those with adequate Se(Reference Beck, Nelson and Shi21). Lungs from infected mice were examined for histopathological changes at days 4, 5, 6, 10 and 21 post-inoculation. Mice fed the Se-deficient diet had significantly more lung inflammation than those on the Se-adequate diet. This increase in pathology was associated with significant alterations in mRNA levels for cytokines and chemokines involved in pro-inflammatory responses(Reference Beck, Nelson and Shi21). Following infection in the Se-deficient mice, there were twenty-nine mutations in one segment of the viral genome(Reference Beck, Nelson and Shi21).
These results demonstrate that adequate Se nutrition is required for protection against viral infection and suggest that nutritional deprivation may be one of many factors that increase the susceptibility of individuals to influenza infection(Reference Beck, Nelson and Shi21).
Polio virus
Sixty-six UK participants with plasma Se ≤95 μg/l were supplemented for 15 weeks in a double-blind study with 100 or 50 μg/d Se (as sodium selenite) or placebo(Reference Broome, McArdle and Kyle22). All subjects received an oral, live, attenuated poliomyelitis vaccine after 6 weeks. Those given selenite, particularly at the higher dose, cleared the virus more quickly as measured at 7, 14 and 21 d and with fewer mutations in the viral genome(Reference Broome, McArdle and Kyle22). The results suggested that plasma Se ≤95 μg/l represented a functional Se deficit with suboptimal immune status and a deficit in viral handling.
HIV and AIDS
Human subjects infected with HIV are under chronic oxidative stress(Reference Guillin, Vindry and Ohlmann14). A dramatic consequence of this oxidative stress is the fatal decrease in the number of CD4 T-cells by apoptosis and ultimately a failure of the immune system leading to death(Reference Guillin, Vindry and Ohlmann14). Low Se status is associated with a lower number of CD4 T-cells, faster progression of AIDS and 20 % increase in the risk of death(Reference Guillin, Vindry and Ohlmann14). Se deficiency in HIV disease has been associated with disease progression and mortality, whether antiretroviral therapy has been initiated or not(Reference Martinez, Huang and Acuna23).
A number of Se supplementation trials in HIV-infected individuals have been conducted: two in the US with Se-enriched yeast(Reference Burbano, Miguez-Burbano and McCollister24,Reference Hurwitz, Klaus and Llabre25) and three with selenomethionine in Tanzania(Reference Kupka, Mugusi and Aboud26,Reference Kupka, Mugusi and Aboud27) , in Botswana(Reference Baum, Campa and Lai28) and in Rwanda(Reference Kamwesiga, Mutabazi and Kayumba29). In Botswana, 878 antiretroviral therapy-naïve HIV-infected adults were supplemented daily for 24 months with multivitamins and 200 μg Se(Reference Baum, Campa and Lai28). There was a significant reduction in the risk of immune decline and morbidity. Multivitamins alone and Se alone had no effect(Reference Baum, Campa and Lai28). In Rwanda, 24-month Se supplementation with 200 μg Se significantly reduced the rate of CD4+ cell-count decline among antiretroviral therapy-naïve patients(Reference Kamwesiga, Mutabazi and Kayumba29). It therefore appears that Se supplementation can delay HIV progression through the maintenance of CD4 cell counts(Reference Martinez, Huang and Acuna23).
Hepatitis B and hepatocellular carcinoma
In Qidong county near Shanghai, hepatocellular carcinoma is highly prevalent(Reference Yu, Zhu and Li30,Reference Li, Zhu and Yan31) . Eleven to fifteen per cent of adults are infected with hepatitis B and are 200 times more likely to develop hepatocellular carcinoma(Reference Yu, Zhu and Li30,Reference Li, Zhu and Yan31) . Two thousand and sixty-five hepatitis B antigen-positive men from Qidong were randomised into two groups: 1112 received 0⋅5 mg Se/d (as sodium selenite) and 953 men received a placebo(Reference Li, Zhu and Yan31). After 3 years treatment and follow-up, thirty-four cases of new hepatocellular carcinoma had developed in the Se group v. fifty-seven cases in the placebo group (P < 0⋅01) suggesting that the relatively large daily dose of Se had reduced the risk of hepatocellular carcinoma(Reference Li, Zhu and Yan31).
Hantavirus and ‘epidemic haemorrhagic fever’
Hantaviruses are RNA viruses that are zoonotic pathogens(Reference Kalaiselvan, Sankar and Ramamurthy32). Transmission among rodents and from rodents to human subjects generally occurs through inhalation of aerosolised excreta. There was a six-times higher incidence of hantavirus infection in severe Se-deficient regions of China than in non-deficient regions(Reference Fang, Goeijenbier and Zuo33). An epidemic of haemorrhagic fever caused by hantavirus in Henan province, China, was successfully treated with multiple oral doses of sodium selenite (≤2 mg/d for the first 9 d), giving an overall 80 % reduction in mortality(Reference Hou34).
Other viruses
Further information on the link between Se, selenoprotein and other viral infections prior to the appearance of SARS-CoV-2 can be found in Guillin et al.(Reference Guillin, Vindry and Ohlmann14).
SARS-CoV-2 and COVID-19
Evidence for an effect of Se status on the SARS-CoV-2-related COVID-19 was first noticed in China in 2020 when the cure rate in cities with up to fifty COVID-19 patients appeared to be different and to relate to the Se status of those cities(35). Cumulative data on the cure rate in Chinese cities were observed from the specific date of 18 February 2020(Reference Zhang, Bennett and Saad36). With regard to Se status, we found that seventeen cities with more than forty cases (outside Hubei, the epicentre province in China) had documented data on hair Se(Reference Zhang, Bennett and Saad36); hair Se was found to be highly correlated with Se intake(Reference Zhang, Saad and Taylor5,Reference Li, Banuelos and Wu37) . We found a significant association between cure rate and background Se status, as represented by hair Se concentration, in cities outside Hubei (R 2 0⋅72) (see Fig. 1)(Reference Zhang, Bennett and Saad36). No correlation analysis was done for cities inside Hubei because Se status was only available for two cities. Though our study has an important number of limitations(Reference Zhang, Bennett and Saad36), our results show an association between the reported cure rates for COVID-19 and Se status, consistent with the evidence of the antiviral effects of Se from the studies discussed earlier.
Fig. 1. Correlation between COVID-19 cure rate in seventeen cities outside Hubei on 18 February 2020, and city population Se status (hair Se concentration) analysed using weighted linear regression. Each data point represents the cure rate, calculated as percentage of patients hospitalised with SARS-CoV-2 deemed to be cured*. The size of the marker is proportional to the number of cases (adapted from Am J Clin Nutr with permission(Reference Zhang, Bennett and Saad36)). From the graph of Se intake v. hair Se concentration, Seintake = 232⋅98Sehair − 44⋅521, allowing the calculation of corresponding values of Se intake and hair concentration(Reference Zhang, Saad and Taylor5). Thus value A represents the hair concentration corresponding to an intake of 55 μg/d where platelet GPX1 activity is maximised, value B represents the hair concentration corresponding to an intake of 105 μg/d where SELENOP concentration is maximised, and value C is the hair Se concentration (1⋅0 mg/kg) at the maximum cure rate in the investigated cities which corresponds to an intake of 188 μg/d(Reference Zhang, Saad and Taylor5).*Cured patients are those in whom temperature has returned to normal for more than 3 d, respiratory symptoms are significantly improved, lung imaging shows significant reduction of inflammation, negative nucleic acid test of respiratory pathogen on two consecutive occasions with a sampling interval of at least 1 d. COVID-19, coronavirus disease; GPX, glutathione peroxidase; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; Se, selenium.
Interestingly, the final point on the correlation corresponded with a hair-Se status corresponding to an intake of 188 μg/d, significantly higher than that relating to the optimal Se intake for the hair concentration corresponding to an intake of 55 μg/d where platelet GPX1 activity is maximised(Reference Hurst, Armah and Dainty38) or for the production of SELENOP at which all selenoproteins are optimised(Reference Hurst, Armah and Dainty38). This suggests that a factor other than a selenoprotein may be relevant to the mechanism by which Se affects SARS-CoV-2 and COVID-19.
A study corroborating these findings(Reference Zhang, Bennett and Saad36) was published the following year; it showed that the case fatality rate of COVID-19 was linked to Se deficiency in soils and crops at the city level in China(Reference Zhang, Zhang and Lu39). Fourteen thousand and forty-five cases of COVID-19 were reported from 147 cities with >20 cases/city from December 2019 to December 2020. Cities were grouped by Se content in crops which was consistent with Se content in topsoil to show that 15 % of cities were severely-Se-deficient, 51 % that were moderately-Se-deficient and 34 % that were non-Se-deficient(Reference Zhang, Zhang and Lu39). Co-variates included population density, gross domestic product per capita, proportion of population >60 years. Using zero-inflated negative binomial regression and correcting for the co-variates, the incidence rate ratio (95 % CIs) for case fatality was significantly different between non-Se-deficient, moderately-Se-deficient and severely-Se-deficient soils and crops, e.g. for soils 1, 2⋅38 (1⋅14–4⋅98), 3⋅06 (1⋅49–6⋅27)(Reference Zhang, Zhang and Lu39). The highest COVID-19 case fatality rates fell within the Se-deficient belt in north-east and central China(Reference Zhang, Zhang and Lu39).
A number of studies have assessed the effect of Se status in patients with COVID-19. A German study looked at serum Se and SELENOP concentrations in COVID-19 patients(Reference Moghaddam, Heller and Sun40). The patients showed a pronounced deficit in total serum Se and SELENOP concentrations when compared to reference data from a cross-sectional analysis of the European EPIC study(Reference Moghaddam, Heller and Sun40). The Se status was significantly higher in samples from surviving COVID patients than in non-survivors (Se, 53⋅3(sd 16⋅2) v. 40⋅8(sd 8⋅1) μg/l; SELENOP, 3⋅3(sd 1⋅3) v. 2⋅1(sd 0⋅9) mg/l). Se status recovered with time in survivors while remaining low or even declining in non-survivors(Reference Heller, Sun and Hackler41,Reference Hackler, Heller and Sun42) . Apart from the studies conducted in Germany(Reference Moghaddam, Heller and Sun40,Reference Heller, Sun and Hackler41,Reference Hackler, Heller and Sun42) , some six other studies have assessed Se in serum/plasma (μg/l) in COVID-19 patients and found lower Se status in those with COVID than in healthy persons(Reference Du Laing, Petrovic and Lachat43,Reference Fakhrolmobasheri, Mazaheri-Tehrani and Kieliszek44,Reference Younesian, Khodabakhshi and Abdolahi45) .
These observations raise the question of whether serum/plasma Se was lower at baseline in these patients or was it simply lowered by SARS-CoV-2 or COVID-19? Indeed, it may have been lower at baseline but under inflammatory conditions (systemic inflammatory response) as in COVID, hepatic SELENOP biosynthesis is diminished and plasma or serum Se falls(Reference Heller, Sun and Hackler41). Infection by SARS-CoV-2 may reduce selenoprotein expression. However, when selenoprotein biosynthesis by hepatocytes recovers, SELENOP secretion and improved systemic Se transport occurs and SELENOP will rise in serum/plasma(Reference Heller, Sun and Hackler41). If selenoprotein biosynthesis does not recover, weak recovery and high mortality risk is indicated as seen in non-recovery from COVID-19(Reference Heller, Sun and Hackler41). The only way to be sure that Se status was low before disease occurs is to measure erythrocyte Se which, with a half-life of 120 d, will only vary after a longer period of inflammation(Reference Stefanowicz, Talwar and O'Reilly46).
Mechanisms by which Se affects SARS-CoV-2 and COVID-19
There is no doubt that selenoproteins are capable of antiviral effects as explained in Table 1 and in the studies previously cited and we will expand on those effects later. However, the data that show that the maximum intake of Se that had the greatest effect on COVID-19 cure rate in China(Reference Zhang, Bennett and Saad36) suggests that an intake above that required to optimise selenoproteins may be relevant to the mechanism by which Se affects SARS-CoV-2 and COVID-19. We believe that this includes redox-active Se species (Fig. 2)(Reference Zhang, Saad and Taylor5).
Fig. 2. Potential mechanisms by which selenoproteins or redox-active Se might suppress the life cycle and mutation to virulence of SARS-COV-2 while attenuating viral-induced oxidative stress(Reference Zhang, Saad and Taylor5) (published with permission of Redox Biology). M2, M2 macrophages; Mpro, main protease; PLpro, papain-like protease; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; Se, selenium; Th1, Th1-type cytokines.
Selenoprotein mechanisms
Antioxidant effects
The balance between ROS and their scavenging systems is disturbed by viral infection which increases the number of ROS(Reference Guillin, Vindry and Ohlmann14). As previously explained, selenoproteins, particularly GPX and TXNRD, can reduce ROS thus helping to maintain the balance(Reference Guillin, Vindry and Ohlmann14). Loss of those selenoproteins in Se deficiency allows more viral copies to be generated per unit time giving more opportunities for the emergence of mutations that are the basis for evolutionary selection of more virulent strains(Reference Moya, Elena and Bracho15,Reference Domingo16,Reference Corey, Beyrer and Cohen17) ; hence Se adequacy is important.
Immune effects
The immune system needs adequate Se(Reference Huang, Rose and Hoffmann47). IL-2 is a necessary growth factor for most T-cells and acts via the IL-2 receptor(Reference Ma and Hoffmann48). Se supplementation was able to induce up-regulation of the IL-2-receptor enabling T and B lymphocytes to respond to IL-2(Reference Roy, Kiremidjian-Schumacher and Wishe49,Reference Kiremidjian-Schumacher, Roy and Wishe50) . An example of the need for adequate Se for immune function is shown by a UK study in which men were supplemented with 50 or 100 μg Se as sodium selenite or placebo daily for 15 weeks in a double-blind study(Reference Broome, McArdle and Kyle22). All subjects received an oral, live, attenuated poliomyelitis vaccine after 6 weeks. Those given Se had improved immune response through an earlier peak T-cell proliferation and an increase in T-helper cells(Reference Broome, McArdle and Kyle22). They showed more rapid clearance than those without extra Se. The data show that these subjects whose Se status was relatively low (plasma Se ≤95 μg/l) had a functional Se deficit with sub-optimal immune status and a deficit in viral handling(Reference Broome, McArdle and Kyle22).
Roles identified for selenoproteins in T-cells have been nicely shown by Ma and Hoffmann in 2021(Reference Ma and Hoffmann48). Selenoproteins involved are GPX1, GPX4, TXNRD1, TXNRD2, SELENOK and SELENOP in the manner specified in Fig. 3(Reference Ma and Hoffmann48).
Fig. 3. Roles identified for selenoproteins in immune cells. Functions of the selenoproteins (blue background) are explained(Reference Ma and Hoffmann48): SELENOP delivers Se to the T-cell; SELENOK (with IP3R, STIM and the CRAC channel) raises the cytosolic Ca2+ content to μM levels, activating pathways for optimal T-cell activation; GPX1, GPX4, TXNRD1, TXNRD2 control redox regulation during T-cell receptor signalling; GPX4 prevents the accumulation of lipid-based ROS which drives the ferroptosis of T-cells (published with permission from Seminars in Cell & Developmental Biology 2021(Reference Ma and Hoffmann48)). Cys, cysteine; ER, endoplasmic reticulum; GPX, glutathione peroxidase; NOX2, NADPH oxidase 2; OXPHOS, oxidative phosphorylation; ROS, reactive oxygen species; TCR/CD3, T cell receptor/CD3 complex plays a key role in antigen recognition; TXN, thioredoxin; TXNRD, thioredoxin reductases.
Reduction of inflammation: inflammatory cytokines and NF-κB
Evidence from clinical studies in human subjects and animals has linked the increased systemic levels of IL-6 with the exacerbation of clinical outcomes involving viral pathogens(Reference Velazquez-Salinas, Verdugo-Rodriguez and Rodriguez51). We also know that NF-κB is crucial for transcription of inflammatory cytokines associated with severe COVID-19(Reference Hariharan, Hakeem and Radhakrishnan52). Some data on Se/selenoproteins affect both those outcomes: (i) deficiency of Se is known to increase IL-6(Reference Tseng, Ho and Hsu53,Reference Jaspers, Zhang and Brighton54) ; (ii) in acute respiratory distress syndrome patients, serum IL-1 and IL-6 correlated inversely with serum Se while intravenous selenite moderated the inflammatory response and meaningfully improved the respiratory system(Reference Mahmoodpoor, Hamishehkar and Shadvar55); (iii) NF-κB has been shown to be specifically inhibited by selenite in cell culture studies(Reference Kim and Stadtman56); (iv) optimal Se status (100 nm/l in cell culture) is critical for alternative macrophage activation, leading to attenuated expression of proinflammatory mediators(Reference Nelson, Lei and Prabhu57).
In another anti-inflammatory function, Se/selenoproteins aid in the shunting of arachidonic acid towards endogenous anti-inflammatory mediators as an adaptive response to protect cells against pro-inflammatory gene expression induced by oxidative stress(Reference Rayman, Hatfield, Berry and Gladyshev58). Thus, Se supplementation in macrophages increases the production of the PG, 15-deoxy-Δ-12,14-PG J2 (by the COX-1, cyclooxygenase pathway), an endogenous inhibitor of a key kinase of the NF-κB cascade, IκB-kinase β(Reference Vunta, Davis and Palempalli59) (see Fig. 4). This effect results in decreased activation of NF-κB and down-regulates the expression of inflammatory genes such as COX-2, TNF-α, IL-6 and VCAM-1, vascular cell adhesion molecule 1(Reference Vunta, Davis and Palempalli59,Reference Vunta, Belda and Arner60) . In a second Se-dependent anti-inflammatory mechanism acting through 15-deoxy-Δ-12,14-PG J2, Se-supplemented macrophages activate the peroxisome PPAR-γ, repressing inflammatory gene expression (Fig. 4)(Reference Ricote, Li and Willson61,Reference Touyz and Schiffrin62) .
Fig. 4. Mechanisms by which inflammation is reduced by Se/selenoproteins(Reference Rayman, Hatfield, Berry and Gladyshev58). AP-1, activator protein 1; COX-1, cyclooxygenase 1; 15d-PGJ2, 15-deoxy-Δ-12,14-PG J2; ER, endoplasmic reticulum; IKKβ, IκB-kinase β; Se, selenium; SELENOS, selenoprotein S.
Lastly, SELENOS is an ER membrane protein involved in the control of inflammation(Reference Shchedrina, Zhang and Labunskyy63,Reference Stoedter, Renko and Hög64) . SELENOS has a role in the removal of stressor-induced misfolded proteins from the ER, preventing the accumulation of these proteins and the subsequent stress response that leads to activation of NF-κB, pro-inflammatory cytokine gene transcription and the inflammation cascade (Fig. 4). Genetic variation in SELENOS has been shown to influence the inflammatory response(Reference Curran, Jowett and Elliott65). Impairment of SELENOS is directly associated with increased cellular cytokine production and release(Reference Curran, Jowett and Elliott65). Other selenoproteins located in the ER that may also contribute to the reduction of ER stress are SELENOF, SELENOM and SELENOK(Reference Wang, Huang and Sun66).
The synthesis of host selenoproteins is targeted at the RNA level by SARS-CoV-2
Higher Se status is associated with a better clinical outcome of SARS-CoV-2(Reference Zhang, Bennett and Saad36,Reference Moghaddam, Heller and Sun40) , and some of that effect is likely to be mediated through increased biosynthesis of host selenoproteins, as reviewed earlier. However, it is also possible that a partial basis of this effect is that SARS-CoV-2 infection can interfere with the synthesis of a subset of host selenoproteins so that increased dietary Se intake helps to counteract that effect. One mechanism by which the virus could impair host selenoprotein synthesis is by antisense targeting of host selenoprotein mRNA. This would require that regions of the viral genomic or mRNA have antisense complementary to regions of a host mRNA sequence. By pairing up with a complementary host mRNA sequence, the viral antisense strand could prevent translation of the host mRNA, thus suppressing host selenoprotein gene expression (Fig. 5)(Reference Robinson67). An example of this type of potential antisense interaction is shown as RNA secondary structures for SARS-CoV-2 v. human TNXRD3 in Taylor et al.(Reference Taylor, Ruzicka and Premadasa68). The possibility of antisense interactions between viral RNAs and host selenoprotein mRNA was first proposed for HIV-1 and Ebola virus based on computational analysis(Reference Taylor, Ruzicka and Premadasa68) and has been supported experimentally for HIV-1 and Zika virus(Reference Premadasa, Dailey and Ruzicka69,Reference Dailey, Premadasa and Ruzicka70) .
Fig. 5. How SARS-CoV-2 inhibits the synthesis of host selenoproteins. Antisense is a single strand of RNA complementary to a target mRNA sequence. By pairing up with it, the viral antisense strand prevents translation of the host mRNA (graphic modified from: Robinson(Reference Robinson67)). SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.
We looked at the effect of infection by SARS-CoV-2 on the expression of selenoprotein mRNA in Vero cells(Reference Wang, Huang and Sun66). SARS-CoV-2 triggered an inflammatory response as evidenced by increased IL-6 expression. Of the selenoproteins, SARS-CoV-2 significantly suppressed the mRNA expression of GPX4, TXNRD3 and ER-resident SELENOF, SELENOK, SELENOM and SELENOS(Reference Wang, Huang and Sun66). Using a gel mobility shift assay and synthetic oligonucleotides, we assessed the potential antisense interaction previously mentioned between SARS-CoV-2 and a region of human TXNRD3 that was identified computationally, corresponding to the predicted antisense pair(Reference Wang, Huang and Sun66,Reference Taylor71) . DNA oligos corresponding to those specific regions of SARS-CoV-2 and TXNRD3 mRNA hybridised quantitatively to form a single slower moving band corresponding to the expected duplex DNA. This interaction was sequence specific, because a negative control consisting of a randomly shuffled SARS-CoV-2 oligo of identical composition still moved as a single band in the presence of the TXNRD3 oligo(Reference Wang, Huang and Sun66). These data present a plausible mechanism for a direct inhibitory effect of SARS-CoV-2 replication on the expression of a specific set of selenoprotein mRNA which may be relevant to the correlations between dietary Se status and the outcome of SARS-CoV-2 infection(Reference Wang, Huang and Sun66).
The possible consequences of SARS-CoV-2-induced down-regulation of selenoprotein genes that we demonstrated at the mRNA level are shown in Fig. 6. Very importantly, SARS-CoV-2 suppresses GPX4 gene expression; GPX4 is vital in the survival and expansion of recently activated T-cells by prevention of lipid peroxidation and ferroptotic cell death(Reference Matsushita, Freigang and Schneider72). SARS-CoV-2 suppresses the expression of SELENOF, SELENOM, SELENOK and SELENOS, causing ER stress(Reference Pitts and Hoffmann73). TXNRD3 is mainly expressed in the testis(Reference Sun, Kirnarsky and Sherman74). Down-regulating TXNRD3 may help explain COVID-19-associated orchitis(Reference Xu, Qi and Chi75) or gastrointestinal manifestations since suppression of TXNRD3 promotes colitis(Reference Liu, Du and Zhu76).
Fig. 6. SARS-CoV-2 suppresses GPX4 gene expression; GPX4 is vital in the survival and expansion of recently activated T-cells by prevention of lipid peroxidation and ferroptotic cell death(Reference Matsushita, Freigang and Schneider72). SARS-CoV-2 causes ER stress and suppresses the expression of SELENOF, SELENOM, SELENOK and SELENOS. Down-regulating TXNRD3 may help explain COVID-associated gastrointestinal manifestations(Reference Liu, Du and Zhu76) and testicular function(Reference Xu, Qi and Chi75). COVID-19, coronavirus disease; GPX, glutathione peroxidase; LPO, lipid peroxidation; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; TXNRD, thioredoxin reductases.
Host selenoproteins may be targeted at the protein level by the SARS-CoV-2 main protease (Mpro)
Another mechanism by which SARS-CoV-2 could interfere with host selenoprotein function is by targeting it at a post-translational stage, via proteolytic degradation. Since RNA viruses typically encode one or more protease enzymes to process their own polyproteins, the targeting of specific host proteins by viral proteases is a well-known mechanism exploited by other RNA viruses including HIV-1, and which has been reported for SARS-CoV-2 by several independent research groups(Reference Gordon, Jang and Bouhaddou77,Reference Taylor and Radding78,Reference Miczi, Golda and Kunkli79) . One of these studies identified amino acid sequences in several human selenoproteins matching known Mpro cleavage sites in the SARS-CoV-2 polyprotein(Reference Taylor and Radding78). The most obvious of these was an almost exact match between a region of SELENOF and the Mpro target site at the non-structural protein12/13 junction in SARS-CoV-2, the sequences of which are TVLQ/AVSA and TVLQ/AVGA respectively, where / corresponds to the protease cleavage site. Other strong candidates for proteolysis by Mpro included sites in TXNRD1 and SELENOP. Synthetic peptides corresponding to all three of these putative cleavage sites in human selenoproteins have been found to be cleaved in vitro by recombinant Mpro using synthetic peptides and MS to identify the cleavage products(Reference Gallardo80).
Significantly, viral knockdown of SELENOF via proteolysis is consistent with the observed knockdown of SELENOF mRNA, which was the most decreased of all the human selenoprotein mRNAs in SARS-CoV-2 infected cells(Reference Wang, Huang and Sun66). This two-pronged assault on a host selenoprotein suggests that the virus must benefit significantly by the disruption of SELENOF activity. The results reviewed earlier further imply that TXNRD1 and SELENOP may also be actively degraded by the virus during SARS-CoV infection. This could help to explain the results of Moghaddam et al.(Reference Moghaddam, Heller and Sun40), who found that a progressive decline in blood SELENOP levels was characteristic of COVID-19 patients who died, despite Se levels being low but unchanged, whereas SELENOP levels increased during the recovery of patients who survived(Reference Moghaddam, Heller and Sun40). The possible proteolytic deactivation of TXNRD1 by SARS-CoV-2 would be consistent with the hypothesis that RNA viruses may benefit by inhibiting DNA synthesis, in order to conserve the ribonucleotide pool for viral RNA synthesis(Reference Taylor71).
Redox-active Se species inhibiting SARS-CoV-2 Mpro
The Mpro of SARS-CoV-2, critical for viral replication, is a key target for therapeutic development(Reference Amporndanai, Meng and Shang81,Reference Mengist, Mekonnen and Mohammed82) . Redox-active Se species have the potential to react with Mpro. After entering into the host cell, SARS-CoV-2 releases its RNA. Two overlapping polyproteins, pp1a and pp1ab are encoded by the replicase gene that constitutes two-thirds of the RNA genome(Reference Amporndanai, Meng and Shang81). They are proteolytically digested into fifteen non-structural proteins by the two viral proteases, Mpro and papain-like protease that are critical for viral replication. The released non-structural proteins form the viral RNA polymerase complex allowing the replication and transcription of fresh virus in the host(Reference Amporndanai, Meng and Shang81). Therapeutic development has produced compounds which are Mpro inhibitors and can bind to the Mpro binding cleft resulting in inactivation with failure of virion assembly; the host cells cannot release the new virions and new infection is inhibited(Reference Mengist, Mekonnen and Mohammed82).
The GPX mimic, ebselen, one of 10 000 compounds examined for inhibiting the Mpro activity of SARS-CoV-2, powerfully inhibited it, suppressing its life cycle(Reference Jin, Du and Xu83). It showed a low concentration for maximal effect and a high concentration for cytotoxicity. From an in silico analysis of ebselen bound to the main protease Mpro of SARS-CoV-2, the Se atom of the open structure of ebselen establishes a covalent interaction with the Mpro catalytic cysteine-145 (see Fig. 7)(Reference Sies and Parnham84). His-41 of Mpro forms an interaction with the aromatic ring of ebselen and a polar interaction with its carbonyl group(Reference Sies and Parnham84). A chemical mechanism for the inhibition of SARS-CoV-2 Mpro cysteine-145 by ebselen and its derivatives has been published(Reference Amporndanai, Meng and Shang81).
Fig. 7. In silico analysis of ebselen bound to the main protease Mpro of SARS-CoV-2(Reference Sies and Parnham84) (In silico analysis was performed using MOE and ACEMD software. Kindly provided by Prof. G. Cozza, Padova; produced with permission of Free Radic Biol Med(Reference Sies and Parnham84)). Cys-145, cysteine-145; His-41, histidine 41; Mpro, main protease; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.
Redox-active Se compounds could similarly inactivate Mpro and papain-like protease by binding to the sulphur of cysteine and contributing to the anti-SARS-CoV-2 effect of Se(Reference Gordon, Jang and Bouhaddou77,Reference Jin, Du and Xu83,Reference Sies and Parnham84,Reference Węglarz-Tomczak, Tomczak and Talma85,Reference Nogara, Omage and Bolzan86) . The non-stoichiometric nature of redox-active Se compounds in oxidising thiols with the involvement of oxygen implies that these redox-active Se compounds may play an important role in inhibiting replication of SARS-CoV-2(Reference Fernandes, Wallenberg and Gandin87).
Another strongly redox-active Se species is known to be produced in the atmosphere from the metabolism of Se in plants, i.e. dimethyl-diselenide. Indeed, Se concentration in forage-crop alfalfa leaves has been used as a proxy for regional Se exposure(Reference Cowgill88). Se in plant, soil, water and bacteria can be transformed into volatile dimethyl-diselenide and can be inhaled by, and excreted from, the lung(Reference al-Bayati, Raabe and Teague89). We have hypothesised that atmospheric dimethyl-diselenide levels and lung Se exposure may be different in US states with different Se concentrations in alfalfa and have explored the interesting question of whether the strongly redox-active dimethyl-diselenide in the atmosphere may play a role in affecting COVID-19 mortality(Reference Zhang, Will Taylor and Bennett90).
Conclusion
This review presents information supporting a role for Se/selenoproteins in reducing the effect of SARS-CoV-2 and COVID-19. There are some credible mechanisms relying on roles for selenoproteins and redox-active Se. However, as yet, there are no randomised controlled trials using Se, although one is planned in Texas entitled Selenium as a potential treatment for moderately-ill, severely-ill, and critically-ill COVID-19 patients (SeCOVID) (ClinicalTrials.gov identifier NCT04869579), though it is not yet recruiting. Until and unless randomised controlled trial information becomes available, proof of the importance of Se/selenoproteins in SARS-CoV-2/COVID-19 may never be confirmed.
Financial Support
None.
Conflict of Interest
None.
Authorship
The authors were solely responsible for all aspects of preparation of this paper.
Open access
