Selenium and Selenoproteins in Inflammatory Bowel Diseases and Experimental Colitis

Abstract

Inadequate dietary intake of the essential trace element selenium (Se) is thought to be a risk factor for several chronic diseases associated with oxidative stress and inflammation. Biological actions of Se occur through low-molecular weight metabolites and through selenoproteins. Several key selenoproteins including glutathione peroxidases; selenoproteins M, P, and S; and selenium-binding protein 1 have been detected in the intestine. Interestingly, Se and antioxidant selenoproteins are known to modulate differentiation and function of immune cells and contribute to avoid excessive immune responses. This review discusses the role of Se and intestinal selenoproteins in inflammatory bowel diseases, based on data from human, animal, and in vitro studies. In humans, Se deficiency is commonly observed in patients with Crohn's disease. In animal models of experimental colitis, the Se status was negatively correlated with the severity of the disease. While the cause–effect relationship of these observations remains to be clarified, the beneficial outcome of dietary Se supplementation and an optimization of selenoprotein biosynthesis in murine inflammatory bowel disease models have led to investigations of targets and actions of Se in the gastrointestinal tract. The Se status affects gene expression, signaling pathways, and cellular functions in the small and large intestine as well as the gut microbiome composition. This data, particularly from animal experiments, hold promise that adequate dietary Se supply may counteract chronic intestinal inflammation in humans.

Crohn's disease, ulcerative colitis, intestinal inflammation, GPx, Sepp1

Multiple factors contribute to the pathogenesis of Crohn's disease (CD) and ulcerative colitis (UC), the 2 main manifestations of the inflammatory bowel diseases (IBD). These include over 160 genes, which modify both overall risk and the pattern of disease, as well as the microbiome in the human gut and environmental risk factors such as smoking and diet.^1,–3 Recent reviews on the role of nutrition in IBD have concluded that dietary habits such as high-caloric diets and over consumption of sugars, saturated fat and meat, characteristics of a “western diet,” increase the risk of IBD.² In addition, micronutrients are of pivotal importance to gut physiology, because (assumed) beneficial effects of vitamins (e.g., vitamin D) and trace elements (e.g., zinc, selenium [Se]) have been explored in IBD and related animal models of intestinal inflammation.^4,–7

The essential trace element Se has received much attention over the last decades because of its antioxidative and anti-inflammatory properties and its use in cancer prevention.^8,9 Dietary requirements to optimize Se status in humans have been estimated to range from 47 to 105 μg Se per day, depending on the examined biomarker and the form of dietary Se.^10,11 Currently, European and U.S. health authorities recommend daily intakes of 55 μg Se per day.⁸ Overt Se deficiency is rarely observed in humans but large differences in Se intake have been reported, which are due to variations in dietary habits, Se bioavailability, and Se content of the food. The mean Se intake in the most European countries and in New Zealand ranges from 27 to 61 μg/d, compared with 93 and 134 μg/d for males and females, respectively, in the United States.^8,12,13 An inadequate Se status is associated with increased mortality, risk of cardiovascular disease, certain kinds of cancer and chronic inflammatory diseases, and impaired immune function.^8,9 Regarding IBD, a low Se status was found to be associated with CD,¹³ and Se deficiency exacerbates experimental colitis.^7,14 Se exerts biological functions through dietary and metabolic low-molecular weight Se compounds (e.g., Se-methylselenocysteine, methylselenol, and selenide) and through selenoproteins, which contain the rare amino acid selenocysteine (Sec), inserted through Sec-specific Sec-tRNA^[Ser]Sec. The mRNA-encoded incorporation of Sec into selenoproteins is a complex process that requires cis-regulatory elements and several accessory proteins (see Fig. 1 for a schematic illustration). Selenoprotein gene transcripts contain at their 3'-end a characteristic hairpin structure, termed selenocysteine insertion sequence and bound by selenocysteine insertion sequence-binding protein 2 (SBP2) during ribosomal translation.⁹ Defective SBP2 protein causes selenoprotein deficiency in individuals with mutations in the SBP2 gene.¹⁵ Interestingly, the complex and extremely rare phenotype of SBP2 mutation carriers includes colitis in one of the only 6 subjects described so far; however, the number of cases is too small to reason a link between selenoprotein deficiency and IBD in humans. Selenoproteins such as glutathione peroxidases (GPx), thioredoxin reductases (TrxR), selenoprotein K, selenoprotein P (Sepp1), and selenoprotein S (SelS) act as antioxidant enzymes, regulate redox-sensitive and inflammatory signaling pathways and are involved in maintaining endoplasmic reticulum (ER) homeostasis.⁹ Their antioxidant and anti-inflammatory functions support the idea of selenoproteins as mediators of beneficial effects of Se in the gut. Indeed, selenoproteins were required for Se-mediated chemoprevention of colon cancer in mice subjected to azoxymethane-induced aberrant crypt formation¹⁶ and have been linked to chronic intestinal inflammation by gene deletion studies.^17,18 This review discusses the role of Se and selenoproteins in IBD and related animal models, and it gives an overview of expression, functions, and regulation of selenoproteins in the intestine.

Figure 1.

Selenoprotein biosynthesis pathway. The metabolism of dietary Se compounds converges to hydrogen selenide, which is converted to selenophosphate by selenophosphate synthetase-2 (SPS-2). A specific seryl residue of tRNA^[Ser]Sec is phosphorylated by O-phosphoseryl-tRNA^[Ser]Sec kinase (Pstk). O-Phosphoseryl-tRNA^[Ser]Sec and selenophosphate react to selenocysteyl (Sec)-tRNA^[Ser]Sec, catalyzed by Sec synthase (SecS). SBP2, Sec-specific eukaryotic elongation factor (eEFSec), and other translation factors promote Sec insertion at UGA codons during ribosomal translation.

Se Status in Patients with IBD

Assessment of Se Status

Several biomarkers are in use to assess the Se status, including Se concentrations in blood/serum, hair, nails, and urine as well as levels of selenoproteins in serum, erythrocytes, and tissues.⁸ Most commonly, Se concentrations are measured in plasma or blood/serum, and concentration or activity of the selenoproteins Sepp1 and GPx are determined in plasma, serum, and/or erythrocytes. Se in plasma is mainly present as Sepp1, GPx-3, and in a low-molecular weight Se pool containing selenomethionine; GPx-1 is found in erythrocytes.¹⁹ Expression/activity of both GPx and Sepp1 correlate well with plasma Se concentrations, but a considerably higher Se intake than what is required to maximize GPx activity (approximately 40–47 μg/d)^10,20 is required to maximize Sepp1 levels. Plasma Se concentrations of 124 μg/L (related to a daily intake of 105 μg Se)¹¹ are associated with a plateau in Sepp1 levels and considered as optimal for human health with respect to lower total mortality and protection against some cancers.⁸

Se Status in Patients with Crohn's Disease

Most of the studies on the Se status in patients with IBD have reported values far below the requirements for optimization of Sepp1.^13,21,–30 In particular, patients with CD are often Se-deficient in comparison to healthy controls with their Se status assessed by measurements of plasma Se concentration,²⁴ serum Se concentration,^13,22,23,26 plasma, serum or erythrocyte GPx activity,^21,24,25,27 and serum Sepp1 concentration²³ (see Table1 for a listing). The cause–effect relationship of a low Se status in patients with CD has received much attention because it is pivotal to uncover selenium's role in the disease. Factors, suggesting that Se deficiency is rather a result than a cause of CD, are malabsorption and malnutrition of Se. The length of the resected small bowel correlated inversely with plasma and erythrocyte Se concentration and GPx activity in patients with CD who had undergone small bowel resection.²¹ Se status was severely impaired in patients with resection of >200-cm small bowel. Moreover, the length of the small bowel correlated with the uptake of radiolabeled [⁷⁵Se] selenite in patients suffering from short bowel syndrome.³¹ These observations reinforce that small bowel-resected CD patients under enteral nutrition (EN) are at risk of Se deficiency because of malabsorption through a decreased absorption surface area.

Table 1.

Se Status in Patients with IBD

Influence of Enteral Nutrition on the Se Status of Patients with CD

In the past, Se malnutrition was likely the cause of low Se status reported in patients with CD receiving formulated diets with minimal Se content.^23,28,29 In a cohort of patients with CD in Japan, an inverse correlation of serum Se concentrations with the duration and daily dose intake of formulated EN diets, which contained only 1.6 and 7.3 μg Se/1000 kcal, was found by Kuroki et al²⁸ The mean serum Se concentrations of the EN group (71 ± 35 μg/L) were significantly lower than in an age- and sex-matched non-EN group with similar range of Crohn's disease activity index (118 ± 14 μg/L, P < 0.0001). In another study by Andoh et al,²³ mean serum Se concentrations in patients with CD who took EN (78 ± 20 μg/L) were also significantly lower than in the non-EN control group (105 ± 20 μg/L, P < 0.01). Supplementation of EN formulas and normal diets with Se efficiently increased the Se status of patients with CD in remission.^25,29 CD patients with Crohn's disease activity index <150 who were on home EN with Elental formula received a combination of 100 μg of Se (in undefined chemical form) and 10 mg of zinc, given daily over 2 months.²⁹ After the supplementation period, serum Se concentrations had risen significantly from approximately 80 μg/L to approximately 95 μg/L and serum GPx activities from approximately 250 μmol·min·L⁻¹ to approximately 270 μmol·min·L⁻¹. Similarly, a randomized and double-blinded study, in which patients with CD on remission were either given placebo or a cocktail of nutritional antioxidant supplements in addition to their normal diet, showed that intake of the supplement mix with a daily dose of 24.8 μg of Se, but not placebo, increased mean serum Se concentrations slightly but significantly from 0.92 μM (72.6 μg/L) before supplementation to 0.96 μM (75.8 μg/L) after 3 months of supplementation.²⁷ Another intervention study showed that the Se status can also be improved in patients with active CD. Infants with active CD (pediatric Crohn's disease activity index >12 for each of the 15 patients) received exclusive EN in form of 2 elemental diets with normal and increased glutamine contents and undefined Se content for 4 weeks, after which the mean plasma Se concentration had increased significantly from 74.8 μg/L to 90 μg/L.³⁰ We conclude from these studies that the (low) Se status of patients with active or remissive CD can be improved by dietary intervention, and that EN formulas can, depending on their Se content, cause an increase or decrease of Se status. Nowadays, most EN formulas contain sufficient amounts of Se to provide adequate Se intake levels.

Influence of Subtype and Duration on the Se Status of Patients with CD

Gentschew et al¹³ compared serum Se levels between subphenotypes of CD: ileal versus only colonic involvement as well as stricturing and inflammatory (nonstricturing, nonpenetrating) versus penetrating (fistulizing) behavior was associated with significantly lower Se levels. Geerling et al²⁵ analyzed blood micronutrient concentrations in recently diagnosed patients with CD in remission and in those who had the disease for >5 years²⁷ and >10 years²⁶ and who were also currently in remission. Irrespective of the duration of the disease, Se status was lower in patients with CD compared with the respective control groups of the 3 studies. Patients with CD assessed within 6 months after initial diagnosis exhibited mean serum Se concentrations (and GPx activities) of 72.6 μg/L (786 U/mmol Hb) versus 78.2 μg/L (949 U/mmol Hb) in controls (age- and sex-matched volunteers with no history of IBD).²⁵ These differences between disease and control group were very similar to those found in patients with long-standing CD and their respective control groups: serum Se 71.9 μg/L (serum GPx activity 778 U/mmol Hb) for CD >5 years versus 79.7 μg/L and 972 U/mmol Hb in controls (age- and sex-matched volunteers with no history of IBD),²⁷ and 67.9 μg/L (768 U/mmol Hb) for CD >10 years versus 81.3 μg/L (967 U/mmol Hb) in healthy controls.²⁶ This suggests that a low Se status is manifested early in CD and not a mere consequence of the progression of the disease. However, one study reported significantly lower serum Se levels of patients with CD who were diagnosed before the age of 40 years were compared with those diagnosed after the age of 40 years.¹³ This report however did not specify whether the CD subgroups were matched for age at time of serum sampling.

Se Status in UC

Reports of Se status in UC are less consistent than for CD, but generally, the Se status seems to be less affected or unaffected in patients with UC (Table 1). One study found slightly but significantly lower serum Se concentration in recently diagnosed patients with UC in remission compared with controls (P < 0.05),²⁵ whereas 2 other case–control studies conducted in Norway and in Japan reported similar levels of serum Se^22,23 and serum Sepp1²³ in UC cases compared with controls. Furthermore, serum Sepp1 levels were not affected by UC disease activity in the latter study. Former patients with UC who had undergone ileal pouch-anal anastomosis surgery, which removes most of the large bowel, did not significantly differ in their mean plasma Se levels and erythrocyte GPx activities from healthy controls.³² Together with the observations in patients with CD, this reinforces that most of the dietary Se is absorbed through the small bowel.

Evidence for Beneficial Effects of Se and Selenoproteins on Intestinal Inflammation

Animal Studies

Se supplementation has been shown to be beneficial in a variety of different experimental settings of colitis in rodent animal models (Table 2).^7,14,33,34 In these experiments, wild-type or selenoprotein knockout animals were fed with diets of different Se content ranging from deficient-to-adequate and supranutritional levels. Thereafter, colitis was induced, and the cohorts were compared for colitis scores, markers of inflammation, extent of tissue damage, etc. The above-mentioned studies differ in the Se status of their animal cohorts, the Se compounds used for supplementation, the duration of supplementation, and the protocols for induction of colitis.

Table 2.

Effects of Dietary Se Supplementation or Deletion of GPx in Studies of Experimental Colitis

Two recent articles that used relevant doses of dietary Se for intervention periods sufficient to properly adjust the Se status of wild-type mice before colitis induction are discussed here in detail. Krehl et al¹⁴ supplemented the diet of mice for 4 to 5 weeks, using selenomethionine (SeMet) as Se source and employing 3 dietary levels of Se (“poor” 0.086 ppm Se, “adequate” 0.15 ppm Se, and “supranutritional” 0.64 ppm Se). When intestinal inflammation was assessed 1 week after treatment with azoxymethane and dextran sodium sulfate (DSS), the total inflammation score in mice fed with supranutritional Se was significantly lower than in the poor Se group. In cohorts that had received sulforaphane, a food-derived inducer of several antioxidant enzymes, both Se-supplemented groups had significantly lower inflammation scores compared with the Se poor group after azoxymethane/DSS treatment. Barrett et al⁷ followed up on these results and induced a severe Se depletion, with diets containing 0.01 and 0.25 ppm Se as sodium selenite fed for 12 weeks. Mice on the Se-deficient diet had only 18% of colonic Se levels and 9% of plasma GPx activities of mice fed with the Se-sufficient diet. On DSS treatment, Se-deficient mice showed more colitic damage and weight loss, decreased survival, and increased stool score. These results demonstrate that Se counteracts (chemically induced) intestinal inflammation. However, Se deficiency per se does not cause colitis in mice, and similarly, severe Se deficiency in humans does not necessarily coincide with IBD, but additional triggers are required. The results by Barrett et al and Krehl et al suggest that Se elicits a dose-dependent protection against intestinal inflammation, and that even moderate Se deficiency, as seen in human populations, seems to have negative effects therein. Although the above-mentioned studies do not allow conclusions on molecular mechanisms underlying protective actions of Se, the involvement of individual selenoproteins in chronic intestinal inflammation has been delineated by the use of transgenic mice. Chu's group generated mice with a knockout of the GPx isoforms, GPx-1 and GPx-2. Homozygous GPx-1/2 double knockout mice, but not mice with an intact allele of either GPx-1 or GPx-2, spontaneously developed ileocolitis similar to IBD.¹⁷ Deletion of GPx-3 did not cause colitis, but DSS-induced colitis was significantly exacerbated in GPx-3^−/− compared with wild-type mice.¹⁸ Other animal models of selenoprotein deficiency have not yet been employed in colitis studies.

Human Studies

In a prospective case–control study, patients with CD in New Zealand were divided into tertiles based on their serum Se levels.¹³ A significantly higher risk of having CD (odds ratio = 2.61) was found for participants in the low Se group (<100.05 μg/L) compared with those in the medium (100.05–118.4 μg/L) and high (>118.4 μg/L) Se groups.¹³ Nevertheless, intervention studies addressing the potential use of dietary Se supplementation for prevention of IBD have not been performed yet. Very few studies have examined associations of selenoprotein gene variants with the risk to develop IBD. Genome-wide association studies have implicated loci containing the GPx-1 and the GPx-4 gene in CD.³ A significant interaction of one SNP in the SEPSECS gene (rs1553153) with serum Se levels and CD has been reported.¹³ This is an interesting finding because the product of this gene, the tRNA:Sec (selenocysteine) tRNA synthase, is required for selenoprotein biosynthesis; however, it remains to be examined whether the tested allele of the rs1553153 SNP modulates selenoprotein biosynthesis.

Effects and Targets of Se Supplementation in the Gut

Se concentration in the gastrointestinal tract is affected by dietary intake of Se, as it has been shown for the colon and small intestine of rats and mice fed with Se-supplemented diets.^35,36 In the following, we discuss studies on the influence of various dietary Se compounds on gene expression, cellular functions, and microbiome composition in the gut.

Effects of Se on Global Gene Expression in the Intestine

Kipp et al³⁷ fed mice for 6 weeks with diets considered as Se-adequate (0.15 ppm Se as selenomethionine) and moderately Se-deficient (0.086 ppm); this relatively small difference was designed to mimic the range of Se intake in many human populations. Plasma Se concentrations and colon GPx activities were higher in the Se-supplemented group, and 4 selenoproteins (SelH, SelM, SelW, and GPx-1) significantly responded to Se supplementation. Global gene expression in the colon was assessed by DNA microarray and revealed a total of 952 genes to be differentially expressed between the 2 groups. Pathway analysis of the microarray data showed effects of supplementary Se on pathways that regulate protein biosynthesis, response to stress, inflammation, carcinogenesis, and the Wnt pathway. Also, splenic leukocytes from Se-deficient mice showed significantly lower GPx and TrxR activity and mRNA expression of many selenoproteins.³⁸ Dietary Se affected gene expression of 40 proteins involved in regulation of inflammatory responses, including myeloperoxidase and lysozyme.³⁸ The authors concluded that moderate Se deficiency impairs the inflammatory response in mice and their ability to cope with infections. A mechanistic link between Se status and inflammatory responses that is partly attributed to the regulation of NF-кB signaling through selenoproteins has been confirmed in studies with cultured cells in vitro.^39,40 Barger et al³⁶ have shown that Se compounds strikingly differ in their impact on gene expression in the mouse small intestine. Mice were fed for 100 days with either a Se-deficient diet (<0.03 ppm) or diets supplemented with 1 ppm Se in form of sodium selenite, Se-enriched yeast, or SeMet. Five hundred twenty-three of 12,445 examined genes were differentially expressed in the small intestine of Se-supplemented mice, many of them by only 1 Se source. The sodium selenite and Se-enriched yeast diets shared 86 genes, and only 7 genes were shared between the 3 Se diets. Several selenoprotein genes were consistently upregulated, first of all SelW and GPx-1, and protein biosynthesis pathways were upregulated by SeMet, confirming earlier findings from Kipp et al.³⁷ The majority of the affected signaling pathways responded differently to the 3 Se compounds, possibly because of differences in the catabolic pathways of the ingested Se species and their utilization for selenoprotein biosynthesis.

Se and the Intestinal Immune System

In silico modeling of intestinal inflammation has suggested that the interaction between inflammatory M1 macrophages and T cells is a critical driver of the immunopathology in IBD.⁴¹ M1 macrophages are key players for initiation and progression of a proinflammatory response, whereas alternatively activated anti-inflammatory M2 macrophages contribute to its resolution. These actions are partly mediated by arachidonic acid–derived prostaglandins (PG) and thromboxanes, which are secreted by macrophages and act as regulators of other immune cells. Se has been shown to increase production and secretion of anti-inflammatory Δ¹²-PGJ₂ and 15d-PGJ₂ and to decrease proinflammatory PGE₂ in macrophages.⁴² These anti-inflammatory effects were exerted through selenoprotein-dependent regulation of hematopoietic PGD₂ synthase, PGE₂ synthase, and NFκB. Thus, the switch from the M1 to the M2 phenotype of macrophages in the resolution phase of inflammation might depend on the sufficient availability of Se. In this regard, Se deficiency impaired the innate immune response in mice infected with pathogenic bacteria Listeria monocytogenes and Citrobacter rodentium.^43,44 Several selenoproteins have been shown or proposed to affect immune functions. Sepp1 was identified as an effector of the M2 phenotype, providing host protection during an infection with the parasite Trypanosoma congolense.⁴⁵ Selenoprotein K is required for the oxidative burst of activated macrophages.⁴⁶ We recently detected SelS in lamina propria macrophages.⁴⁷ Studies in the RAW264.7 cell line suggested that SelS modulates release of cytokines from macrophages⁴⁸; and therefore, SelS might be involved in regulation of the intestinal immune response. GPx-1 and TrxR1 have been detected in RAW264.7 cells, but their role in macrophages remains to be established.⁴² For an overview of Se and selenoproteins in T-lymphocytes, we refer to a recent review.⁴⁶

Interplay Between Dietary Se and Gut Microbiota

Dietary habits and specific food components may influence the gut microbiome and consequently affect the immune response in the gut mucosa.^2,49,50 Two studies have shown that the level of dietary Se intake affects the gut microbiome composition, and vice versa the gut microbiome affects Se status and expression of selenoproteins in the host. Under conditions of limited dietary Se supply, germ-free mice had higher levels of Se and some selenoproteins (TrxR and GPx-1/2) in the small and large bowel and higher GPx activity and Se levels in liver and plasma than bacteria-exposed mice.⁵¹ These differences were abrogated in mice fed diets with adequate Se content. The authors assumed that intestinal bacteria compete with the host for luminal Se, which can compromise the host's Se status when luminal Se availability is limited. However, dietary Se (in form of sodium selenite) increased the diversity of the gut microbiome and affected the relative abundance of specific phylotypes in mice.⁵² Several phylotypes in the Bacteroidetes and Firmicutes phyla were identified to respond significantly to dietary Se; for example, the relative abundance of Parabacteroides phylotype 3 and Alistipes phylotype 1 was lower, whereas Akkermansia phylotype 1 and Tanerella phylotype 2 was increased in mice fed with Se-supplemented versus Se-deficient diets.⁵² Lesser diversity and altered composition of the gastrointestinal microbiome is associated with chronic inflammatory gut disorders, and there is also clinical evidence for an involvement of the microbiome in CD etiology.⁵³ Interestingly, a decrease of 2 phylotypes belonging to the Porphyromonadaceae has been observed in patients with CD,² and the Porphyromonadaceae phylotypes 5 and 10 levels were lowered in Se-deficient mice.⁵² Deregulation in CD and significant Se-mediated effects in mice were also observed for Ruminococcaceae and Clostridiales phylotypes.^2,52 Therefore, manipulation of the gut microbiome composition might underlie some of the beneficial actions of Se in experimental colitis. It remains to be determined whether Se has similar effects on the microbiome composition in the human gut. Given that patients with CD are relatively Se-deficient, it seems worthwhile to test for an impact of supplementary Se on microbiota in the inflamed and in the healthy gut. This might offer novel and cost-effective therapeutic opportunities in the treatment and/or prevention of the disease.

Expression, Regulation, and Functions of Selenoproteins in the Inflamed and Healthy Gastrointestinal Tract

The full selenoprotein transcriptome of 24 murine selenoprotein genes is expressed in the mouse intestine.⁵⁴ Selenoprotein gene transcripts vary in their relative expression levels by approximately 3 orders of magnitude, with highest levels found for Sepp1, GPx-4, selenoprotein K, SelW, and GPx-2, whereas other well-characterized selenoproteins such as 15-kDa selenoprotein, GPx-1, SelS, and TrxR1 were expressed at medium levels.⁵⁴ Genes with very low or undetectable expression in the murine colon included deiodinases Dio2 and Dio3, SelN, and SelV.³⁷ In the human rectum, Sepp1, GPx-2, GPx-1, and TrxR1 transcripts were found to be expressed at similar levels as in the murine intestine.⁵⁵ An earlier article by Behne et al⁵⁶ examined the selenoprotein proteome in the rat intestine by in vivo labeling with ⁷⁵Se and autoradiography. Numerous individual protein bands were detected by this method but could not readily be assigned to specific selenoproteins. Table 3 lists selenoproteins detected at the protein level in the human and/or murine gut.

Table 3.

Localization and Functions of Intestinal Selenoproteins

In the following, we discuss selenoproteins that have been shown or are likely to be involved in Se-mediated protection against gut inflammation. For a more detailed presentation of beneficial and detrimental effects of selenoproteins in prevention and progression of colorectal carcinoma, we refer the reader to another recent review.⁷²

Glutathione Peroxidases

Aberrant expression of GPx isoenzymes has been observed in colorectal cancers,⁷³ and their roles in colorectal carcinogenesis has been reviewed recently.⁷⁴ Regarding the functions of GPx, we also refer the reader to other comprehensive reviews.^9,75 The 4 major GPx isoenzymes are expressed in the gut. GPx-1, GPx-2, and GPx-4 are localized in epithelial cells, whereas Gpx-4 is also found in lamina propria cells.⁶³ GPx-2 is mainly localized in cells near crypt bases in the small and large bowel and in Paneth cells of the small bowel, whereas GPx-1 and GPx-4 are more abundant in mature enterocytes located at the top of villi and crypts.^57,63 The extracellular (plasma) GPx-3 has been detected in both the large and the small bowel; GPx-3 levels were higher in mature enterocytes than in cells near the crypts.⁵⁸ Human colon explants and Caco-2 cells synthesize and secrete GPx-3.^58,60 Another study detected GPx-3 exclusively at the basolateral side of epithelial cells in the duodenum and colon of mice, and the authors suggested that extracellularly located GPx-3 in the gut is derived from the kidney.⁶¹ Biosynthesis and activity of the GPx isoforms in the gut are dependent on sufficient Se availability, although to a different extent. GPx-1 gene and protein expression is highly responsive to Se intake throughout the bowel.⁵⁷ Consequently, total GPx activity in the murine gut increases with higher dietary Se intake.^57,76 Protein expression of GPx-2 in the mouse ileum and colon as well as in Caco-2 cells is slightly lowered during Se deprivation, but to a much lesser extent as GPx-1.^14,57,77 Ileal GPx-4 activity is resistant to moderate Se deficiency and slightly higher at supranutritional Se intake, whereas GPx-4 activity in colon significantly decreased in moderate Se deficiency.⁵⁷

The expression of GPx-1 in colitis has not been examined yet. GPx-2 is consistently upregulated at the mRNA level in experimental colitis (DSS- and TNBS-treatment, CD4⁺ CD45RB^high) and in IBD (colonic UC and ileal CD).⁷⁸ A beneficial role of both GPx-1 and GPx-2 in intestinal inflammation has been delineated through studies in knockout mice. Double knockout of GPx-1 and GPx-2 causes ileocolitis but mice carrying a deletion of only one of these 2 GPx genes lack a phenotype.¹⁷ This result has been explained by compensation, as GPx-1 expression in crypt epithelial cells of the small and large bowel was elevated in GPx-2 knockout mice fed with adequate or supranutritional Se.⁵⁷ Upon azoxymethane/DSS treatment, Se-deficient Gpx-2 knockout mice showed increased crypt cell apoptosis^14,57 and higher total inflammation scores in the colon, and this was abrogated under a Se-supranutritional diet.¹⁴ Redundancy of GPx-1 and GPx-2 is also due to overlapping substrate (hydroperoxide) specificity of the 2 isoenzymes.⁵⁹

Like GPx-1 and GPx-2, GPx-3 exerts a protective function in gut inflammation. Deletion of GPx-3 did not cause colitis but significantly exacerbated colitis in GPx-3^−/− mice treated with DSS.¹¹ The localization of GPx-3^58,61 suggests that it may protect the basolateral membrane of epithelial cells in the small and large intestine from oxidative damage. This is supported by in vitro data: Deletion of GPx-3 increased oxidative stress, DNA damage, and apoptotic cell death in intestinal epithelial Caco-2 cells treated with hydrogen peroxide.¹⁸

GPx-4 stands out from the above-mentioned GPx isoforms, as it reduces a broader range of substrates including phospholipid hydroperoxides and thymine hydroperoxides.^79,80 GPx-4 thereby maintains integrity of cellular membranes and protects against oxidative DNA damage. Deletion of GPx-4 causes embryonic lethality.^81,GPx4^+/− had no phenotype, despite decreased expression of GPx-4 in many tissues. Murine embryonic fibroblasts generated from the heterozygous mice showed increased sensitivity to oxidative stress.⁸¹ GPx-4 plays an essential role in prevention of 12/15 lipoxygenase-derived lipid peroxidation, which can not be compensated by other GPx isoforms.⁶² Thus, GPx-4 may contribute to the Se-mediated protection of barrier integrity of polarized Caco-2 cells challenged with 13-hydroperoxy octadecadienoic acid⁷⁷; and in vivo, it might be required to protect the intestinal epithelial cell lining against damage from luminal hydroperoxides.

Selenoprotein S

We have recently characterized the expression of SelS in the mouse and human gut.⁴⁷ High SelS levels were found in Paneth cells and in macrophages. SelS expression was induced in vitro by Se supplementation and on induction of ER stress. Consistently, SelS was induced in vivo together with the ER stress marker protein GRP78 in DSS-treated mice and in Muc2 mutant mice, models of induced and spontaneous colitis associated with ER stress. In humans, SelS and GRP78 levels were elevated in inflamed versus noninflamed ileal tissue of patients with CD.⁴⁷ Persistent ER stress is linked to inflammatory signaling, and ER stress can trigger chronic intestinal inflammation. It is tempting to speculate that ER-resident selenoproteins such as SelS might contribute to resolve ER stress, thereby mediating anti-inflammatory effects of Se supplementation. Although depletion of SelS by RNA interference did not cause or modulate an ER stress response in colon-derived enterocyte- and goblet-like cell lines, the role of SelS in the gastrointestinal tract and especially in Paneth cells remains to be established. Paneth cells are crucial for the host defense because they protect the small intestinal mucosa against infiltration of microbes by secretion of antimicrobial peptides, e.g., defensins and lysozyme. We hypothesize that selenoproteins in Paneth cells (e.g., GPx-2, SelS, and SelM) might link Se status and protection against intestinal inflammation. Ten common polymorphisms in the SELS gene have been detected.⁸² The variant rs28665122 was associated with the occurrence of proinflammatory markers in plasma⁴⁸ and with the risk for gastric cancer,⁸³ but it was not associated with IBD in a white population.⁸⁴

Selenoprotein P

The plasma Se transport protein Sepp1 is mainly expressed and secreted by the liver, but it has also been detected in many other tissues, including the human and rat intestine.^65,73 Sepp1 is strongly downregulated in colorectal tumor tissue and possibly modifies colorectal carcinogenesis.⁷² We have recently characterized Sepp1 localization in the rat jejunum and colon; epithelial Sepp1 was uniformly distributed in colonic crypts, whereas its expression increased along the crypt-to-villus axis in the jejunum.⁶⁵ This is consistent with in vivo and in vitro data, showing significantly higher Sepp1 mRNA levels in villus epithelial versus crypt epithelial cells of the human ileum⁸⁵ and an induction of Sepp1 biosynthesis in the course of enterocytic differentiation of Caco-2 cells.⁸⁶ Colonic Sepp1 expression is downregulated in experimental colitis, probably through proinflammatory cytokine-dependent induction of inducible nitric oxide synthase-2.⁸⁶ The precise function of Sepp1 in the gastrointestinal tract is not known. We have recently reported that polarized Caco-2 secretes Sepp1 to their basolateral side.^65,86 As Sepp1 may also act as a phospholipid hydroperoxide reductase,⁶⁶ extracellularly located Sepp1 might protect the intestinal epithelium against oxidative membrane damage. Sepp1 was also detected to be bound to the plasma membrane of CD138-positive plasma cells (antibody-secreting B-cells) in the rat colon.⁶⁵ A role of Sepp1 in Se delivery to lymphocytes or in their differentiation is intriguing but remains to be established.

Selenium-binding Protein 1

In contrast to the above-discussed selenoproteins, selenium-binding protein 1 (Selenbp1) does not contain mRNA-encoded selenocysteine, and how it interacts with Se is unknown. Selenbp1 is abundantly expressed in mature enterocytes; it is a marker and possibly also a regulator of enterocyte differentiation.⁸⁷ It is downregulated in many cancers, including those of the GIT,⁸⁷ and has been shown to act as a tumor suppressor.⁶⁷ Two studies reported downregulation of Selenbp1 gene expression in UC, comparing inflamed tissue (rectum and sigmoid) from patients with UC versus healthy control tissue (sigmoid)⁶⁸ and inflamed versus noninflamed colonic UC tissue.⁶⁹ Hsieh et al⁷⁰ assessed colonic proteome expression in patients with UC (n = 4) and healthy control subjects (n = 5) by 2D gel electrophoresis and mass spectrometry and reported downregulation of Selenbp1 protein expression in UC disease versus normal colon mucosa. Selenbp1 protein expression was also downregulated in mice treated with azoxymethane/DSS to mimic UC-associated carcinogenesis.⁷¹ Silencing of Selenbp1 expression in colon cancers and colon cancer-derived cell lines occurs, at least partially, through promoter hypermethylation,⁶⁷ but it remains to be clarified whether this also happens in patients with UC.

Conclusions and Future Directions

The exacerbation of experimental colitis under condition of Se and selenoprotein deficiency can be attributed to various effects of Se on gut physiology, including modulation of inflammatory pathways and oxidative stress response. An overview of biological effects of Se in the gut is depicted in Figure 2. Altogether, the currently available data from human and animal studies suggest that dietary Se supplementation might be beneficial in IBD, particularly in patients with CD with relative Se deficiency. The cause–effect relationship of a low Se status in patients with IBD remains to be clarified, and intervention studies are required to answer the question whether Se status and intake might affect risk, progression, and/or remission of IBD. Lessons from studies examining the usefulness of Se for cancer prevention indicate that individuals with low baseline Se status are likely to benefit most from dietary Se supplementation.⁷² The desirable Se intake level to improve gut health may depend on determinants such as variants in selenoprotein genes and genes involved in the metabolization of Se, interactions of Se compounds with other components of the diet, and the gut microbiome, medical conditions, and drug use. Therefore, our knowledge regarding the impact of different Se compounds on the human gut and its commensal bacteria needs to be broadened to allow a more targeted use of Se in IBD. Polymorphisms in selenoprotein genes need to be tested for associations with IBD, and interactions between multiple loci and with Se status should be considered. Mechanistic in vitro studies using the different intestinal cell types and in vivo studies employing selenoprotein-deficient mice with single or global selenoprotein depletion will help to further understand actions of selenoproteins in the GIT and to identify mediators of beneficial Se effects therein. The signal transduction pathways that regulate the intestinal biosynthesis of individual selenoproteins beyond Se availability need to be characterized in detail to target relevant selenoproteins more directly.

Figure 2.

Effects of selenium in the gut that are related to the beneficial outcome of dietary Se supplementation in (experimental) intestinal inflammation.

Author notes

Reprints: Bodo Speckmann, PhD, Institute for Biochemistry and Molecular Biology I, Heinrich-Heine-University Düsseldorf, Universitätsstrasse 1, Geb. 22.03, D-40225 Düsseldorf, Germany (e-mail: ).

Supported by grants of Deutsche Forschungsgemeinschaft (DFG, Bonn, Germany) to B. Speckmann (SP 1333/2-1) and to H. Steinbrenner (STE 1782/2-2).

The authors have no conflicts of interest to disclose.