Dietary Selenium Supplementation and Whole Blood GeneExpression in Healthy North American Men

Wayne Chris Hawkes & Diane Richter & Zeynep Alkan

Received: 2 November 2012 /Accepted: 6 August 2013 /Published online: 18 August 2013# Springer Science+Business Media New York (outside the USA) 2013

Abstract Selenium (Se) is a trace nutrient required in micro-gram amounts, with a recommended dietary allowance of55 μg/day in humans. The nutritional functions of Se areperformed by a group of 25 selenoproteins containing theunusual amino acid selenocysteine at their active sites. Theselenoproteins with known activities are oxidation–reductionenzymes with roles in antioxidant protection, redox homeo-stasis and signaling, and thyroid hormone metabolism. Bothdeficiencies and excesses of Se are associated with impairedinnate and adaptive immune responses. We supplemented 16healthy men for 1 year with 300 μg Se/day as high-Se yeast orplacebo yeast and measured whole blood gene expressionwith DNA microarrays before and after supplementation.Protein phosphorylation was the main biological process incommon among the Se-responsive genes, which included aprominent cluster of protein kinases, suggesting that proteinphosphorylation in leukocytes is sensitive to Se supplementa-tion. We found highly ranked clusters of genes associated withRNA processing and protein transport, suggesting that dietarySe may regulate protein expression in leukocytes at both theposttranscriptional and posttranslational levels. The mainfunctional pathway affected by Se supplementation was FASapoptosis signaling, and expression of genes associated with Tcell and natural killer cell cytotoxicity was increased. At thesame time, the numbers of circulating natural killer and Tcells

expressing activation markers decreased. These changes areconsistent with an anti-inflammatory effect of Se supplemen-tation exerted through regulation of protein phosphorylation.

Keywords Selenium . Supplementation . Gene expression .

T lymphocyte . Natural killer cell . Microarray

Selenium (Se) is an essential trace element required in micro-gram amounts by all animals. Signs of Se deficiency includeliver necrosis in rats, pancreatic atrophy in chickens, nutrition-al muscular dystrophy in sheep, and skeletal myopathy inpatients receiving total parenteral nutrition without supple-mental Se [1]. The essential functions of Se in mammals aremediated by a group of 25 selenoproteins that containselenocysteine (Sec), the Se homolog of cysteine, at theiractive sites. Enzymatic activities have been assigned to 12 ofthese selenoproteins: four glutathione peroxidases, threeiodothyronine deiodinases, three thioredoxin reductases, onemethionine sulfoxide reductase B, and one selenophosphatesynthetase. Most of these selenoproteins are expressed inwhite blood cells [2–4].

Se deficiency has pleiotropic effects on humoral and cellu-lar immunity [5]. Se is essential for development of the bursaand supports development of the thymus in chicks [6]. Seprevented myocardial lesions from coxsackievirus B3 infec-tion in mice [7] and increased microbicidal activity in rats andmice [8, 9]. Se raised serum IgM and in vitro lymphocyteproliferation in lambs and increased chemotaxis in goat neu-trophils [10–12]. Se supplementation improved polio vacci-nation response and poliovirus handling in UK subjects withmarginal Se status [13], and supplementation of elderly sub-jects with Se-enriched yeast increased in vitro lymphocyteproliferation in response to pokeweed mitogen [14], whereasSe deficiency was associated with increased mortality in HIV-positive populations [15–17]. However, despite the numberand variety of reported effects of Se on immune

Electronic supplementary material The online version of this article(doi:10.1007/s12011-013-9786-5) contains supplementary material,which is available to authorized users.

W. C. Hawkes (*) : Z. AlkanUnited States Department of Agriculture, Agricultural ResearchService, Western Human Nutrition Research Center, University ofCalifornia Davis, 430 West Health Sciences Drive, Davis,CA 95616, USAe-mail: wayne.hawkes@ars.usda.gov

D. RichterDepartment of Nutrition, University of California Davis, Davis,CA 95616, USA

Biol Trace Elem Res (2013) 155:201–208DOI 10.1007/s12011-013-9786-5

function, little is understood about the underlyingmechanisms.

In addition to direct antioxidant protection, some of thehealth effects of Se may be mediated through its anti-inflammatory actions in the immune system. Se has signifi-cant immunoregulatory effects [18]. As little as 0.2 ppm Seinhibited the in vitro cytotoxic activity of natural killer cells[19] and supplementation of chicken diets with as little as1 mg Se/kg diet depressed levels of circulating T cells [20],raising concerns that Se intakes near the upper safe limit (UL)might impair immune function. Because of its immunosup-pressive properties, Se has found application as an anti-inflammatory treatment for critically ill patients with systemicinflammatory response syndrome [21]. There is a growingappreciation of the role of inflammation in the developmentof chronic diseases such as type 2 diabetes mellitus (T2DM),insulin resistance, atherosclerosis, and nonalcoholic fatty liverdisease [22]. Inflammation is initiated, and to a large extentregulated, by monocytes and other immune cells in the circu-lation. Thus, the anti-inflammatory effects of Se are likelymediated, at least in part, by circulating leukocytes.

The Dietary Reference Intake Committee of the US Na-tional Academy of Sciences has established an UL of 400 μgSe/day to prevent Se toxicity [23]. However, concerns remainabout the potential adverse health effects of long-term Sesupplementation at or near the UL. Although overt Se toxicityis rare, high serum Se was associated with increased preva-lence of T2DM in the National Health and Nutrition Exami-nation Survey 2003–2004 [24] and Se supplementation wasassociated with increased T2DM incidence in a cancer pre-vention trial [25]. On the other hand, a prospective study of3,535 men and 3,630 women in the Nurses’ Health Study andthe Health Professionals Follow-Up Study found a significantinverse association between toenail Se and incidence of T2DM[26]. Insulin signaling is redox-sensitive: overexpression ofglutathione peroxidase 1 induced insulin resistance and obesity[27], whereas glutathione peroxidase 1 null mice had improvedinsulin sensitivity [28].

We used DNA oligonucleotide microarrays to analyzewhole genome gene expression in leukocytes to identify genesand pathways influenced by high-level Se-yeast supplements.The most prominent leukocyte pathway altered by Se supple-mentation was FAS apoptosis signaling, with T cell and nat-ural killer cell cytotoxicity genes upregulated and antigen-specific receptor genes downregulated.

Materials and Methods

Subjects and Experimental Protocol The present study waspart of a larger investigation into the health effects of long-term, high-level Se supplementation in healthymen aged 18 to45 [29, 30]. Inclusion criteria were self-reported absence of

disease and clinically normal blood count and blood chemis-tries. Exclusion criteria were tobacco smoking, positive bloodtest for HIV, hepatitis B, or syphilis, or positive urine tests fordrugs of abuse or Se supplements providing more than 50 μg/day. The study protocol was reviewed and approved by theInstitutional Review Board of the University of California atDavis School of Medicine, and informed consent wasobtained in writing from all subjects. Two subjects at a timewere randomized to treatment from July 2000 to November2002, with one subject from each pair randomly assigned toreceive either placebo or high-Se yeast tablets for 48 weeks(300 μg Se, SelenoPrecise™, Pharma Nord, Denmark). Sub-jects visited the center for duplicate blood draws 2 days apartevery 6 weeks and then again at 72 and 96 weeks. Bloodsamples were collected in the mornings after an overnight fast.Forty-two subjects completed the entire study.

DNA Microarray Analysis of Whole Blood RNA Microarrayanalysis was performed on baseline and 48-week blood sam-ples from eight randomly selected subjects in the controlgroup and eight randomly selected subjects from the Se-supplemented group. RNA was purified from fresh wholeblood using QIAGEN columns supplied with the RNeasy PlusMini Kit (QIAGEN) per manufacturer’s protocol. RNA sam-ples from duplicate blood draws within the same week werepooled into a single baseline sample and a single 48-weeksample for each subject. Five micrograms of total RNA wasused for the preparation of biotin-labeled target complemen-tary RNA (cRNA), using the standard Affymetrix protocol[31] and the Enzo BioArray High Yield RNA TranscriptLabeling Kit (Enzo Diagnostics, Farmingdale, NY). Twentymicrograms of the cRNA product was fragmented and thenhybridized for 16 h to Affymetrix Human Genome U-133Aarrays, which were processed with the Affymetrix FluidicsStation 450 and then scanned with the Affymetrix GeneChipScanner 3000. Expression values were calculated withAffymetrix GCOS software using MAS 5.0 algorithms [32].

Flow Cytometry Cells were labeled in fresh whole blood witha “lyse and wash” procedure using four-color fluorescentantibodies, and reagents supplied by Becton-Dickinson (BDBiosciences, San Jose, CA) according to the manufacturer’sprotocols and analyzed on a BD FACSCalibur flow cytometeras described previously [33]. The cell counts from duplicateblood draws within the same week were averaged and report-ed as a single value for each subject.

Statistical Analysis The mean within-subject changes in ex-pression values for each probe set were compared betweengroups using unpaired two-sided t tests. The set of 339 geneswith p values <0.05 comprised the set of Se-responsive genes(Table S1, Supplemental Materials). The set of Se-responsivegenes was analyzed with the DAVID gene ontology (G.O.)

202 Hawkes et al.

software system [34] to identify overrepresented G.O. termsand clusters of functionally related genes. Flow cytometrydata were analyzed with an unpaired two-sided t test of thefinal values at the end of the 48-week supplementation period.Statistical calculations were performed with SigmaStat(version 2.03, SPSS, Chicago, IL). A probability of 0.05 orless was considered significant.

Results

In the larger trial from which the subjects of the present studywere drawn, Se supplementation increased plasma Se from142±19 to 228±63 μg/L (1 μg Se=12.7 nmol) and raisederythrocyte Se from 261±35 to 524±141 μg/L, but glutathi-one peroxidase enzyme activity was not changed in eithercompartment [30]. Se intake from the diet was estimated at138±39 μg/day in the controls and 137±42 μg/day in thesupplemented group, well above the RDA of 55 μg/day.

A set of 349 differentially expressed probe sets (p <0.05, ttest of within-subject changes) representing 339 Se-responsive genes were identified for gene ontology analysis(Table S1, Supplemental Material). More than half the Se-responsive genes (183 genes, 54.0 %) were associated withprotein phosphorylation. The low probability of finding thismany protein phosphorylation genes by chance, p =6.7×10−12 (Table 1), supports the validity of the Se-responsivegene selection procedure. The G.O. analysis identified a clus-ter of 26 functionally related kinase genes affected by Se (p =0.00027, Supplemental Table S2), most of which functionnear the top of protein phosphorylation cascades. The

prominence of protein phosphorylation in the G.O. analy-sis is particularly interesting in light of our recent findingthat selenoprotein W (SEPW1), a selenoprotein increasedby Se supplementation [35], regulates activation of mito-gen-activated protein (MAP) kinases in cultured humancells [36].

The second largest cluster of functionally related genes wasassociated with the G.O. term “alternative splicing.” Alterna-tive splicing enhances RNA and protein diversity in speciesfrom bacteria to humans, particularly in mammals where itfulfills an important regulatory role. Additionally, RNA pro-cessing plays a special role in immune cells, where it isessential for V(D)J recombination and immunoglobulinisotype switching. Almost half of the Se-responsive genes inleukocytes (165 genes, 48.7 %) were associated with alterna-tive splicing (p =8.8×10−6, Table 1). We found a cluster of 14Se-responsive genes directly involved in RNA processing(p =0.0018, Supplemental Table S3), suggesting that Se sup-plementation might alter RNA processing in leukocytes. Sub-sequent analysis of the genes associated with acetylation (80genes), cytoplasm (88 genes), and nucleus (104 genes) did notreveal any further clusters of functionally related genes. How-ever, the sixth-ranked item in Table 1, “protein transport,”corresponded exactly to a cluster of 31 genes involved inautophagy, vesicle trafficking, intramitochondrial transport,and nuclear-cytoplasmic transport (Supplemental Table S4).

The top-ranked pathway in the G.O. analysis was FASsignaling (p =0.00022, Table 2), suggesting that Se supple-mentation may affect apoptosis in immune cells. Apoptosis isvital in the development and maintenance of the immunesystem, and Fas-mediated cell killing is an important mecha-nism of cytotoxicity employed by immune cells. The FASsignaling gene cluster consisted of seven genes involved inapoptosis (Supplemental Table S5), including two caspases,two actin-modifying enzymes, and c-Jun N-terminal kinase 2(JNK2). JNK2 was recently shown to mediate the cell cyclearrest caused by silencing SEPW1 [36], an interesting link ascancer cells tend to undergo apoptosis when arrested. The nextranked pathway was a cluster of genes associated withHuntington’s disease (p =0.00205, Table 2), which was pri-marily a subset of the protein transport genes (SupplementalTable S4). The cluster of genes in the Tcell activation pathwayand the decreased numbers of activated T cells are consistentwith our earlier observation that Se supplements retardeddevelopment of T cell-mediated delayed-type hypersensitivity(DTH) anergy [33].

The genes whose expression showed the largest absoluteincreases with Se supplementation are listed in Table 3. Mostof the genes upregulated by Se are involved with T cell andnatural killer (NK) cell cytotoxicity. Granzyme H and perforinare lytic agents deployed by T cells and NK cells to kill theirtargets. RAC2 is a small GTPase involved in release of thesecretory granules containing the lytic agents. Upregulation of

Table 1 G.O. functional annotation analysis of white blood cell geneschanged by selenium

Keyword Count Percent Probability

phosphoprotein 183 54.0 6.7×10−12

acetylation 80 23.6 1.3×10−7

alternative splicing 165 48.7 8.8×10−6

cytoplasm 88 26.0 9.1×10−6

nucleus 104 30.7 3.7×10−5

protein transport 31 9.1 0.00015

kinase 26 7.7 0.00027

methylated carboxyl end 4 1.2 0.00052

prenylated cysteine 5 1.5 0.00054

nucleotide-binding 47 13.9 0.00071

P-loop 9 2.7 0.00077

The 339 genes called “present” in at least four cases and whose expres-sion differed significantly between groups (p <0.05, t test of within-subject changes) were analyzed with the DAVID gene ontology softwaresystem [34]. The terms overrepresented at p <0.001 are shown

Dietary Selenium Supplementation and Whole Blood Gene Expression 203

actin regulators cofilin and ACTR3, plus two forms of tubulinconcords with the essential role of the cytoskeleton in T celland NK cell cytotoxicity. The upregulated genes in Table 3 arein large part responsible for the significance of the FASapoptosis pathway in Table 2.

Table 4 shows the genes with the largest decreases inexpression due to Se supplementation. The top five mostdownregulated probe sets represented just three genes: immu-noglobulin kappa variable 1–5, immunoglobulin heavy con-stant mu, and Tcell receptor alpha constant. These genes sharethe unusual property of being subject to somatic V(D)J re-combination, a process that depends on RNA processing inorder to proceed. Thus, the downregulation of V(D)Jrecombination-dependent genes may be related to the highranking of the “alternative splicing” G.O. term (165 genes,8.8×10−6) and the clustering of RNA processing genes

(Supplemental Table S3). Downregulation of these cell sur-face receptors may be related to the decreased activation of Tand NK cells (see below).

Flow cytometry (Table 5) revealed no differences betweengroups in lymphocytes, B cells, T cells, T-helper cells, or T-suppressor cells. However, there was a 55 % decrease in NKcell count in the Se-supplemented subjects that was accompa-nied by a 42 % decrease in activated T cells and a 79 %decrease in activated NK cells. These results are generallysimilar to the results previously reported from the clinicalstudy, except the decrease in NK cell count was not statisti-cally significant in the larger cohort. Overall, Se supplemen-tation seemed to decrease the numbers and activation states ofeffector immune cells.

Discussion

To control for between-subject variation in this small group ofsubjects, we selected the Se-responsive genes using t tests ofwithin-subject changes. We applied a low-stringency p valueof 0.05 to yield an adequate number of Se-responsive genesfor gene ontology analysis. The analysis set was composedmostly of genes associated with protein phosphorylation, ahighly unlikely occurrence (p =6.7×10−12) that supports thevalidity of the selection method and demonstrates this geneexpression profile contains physiologically significant infor-mation. The Se-responsive genes in leukocytes were not sig-nificantly enriched with cell division cycle or mitosis genes aswe previously observed in cultured human breast cells [37].This probably reflects cell type-specific differences as well asdifferences between homeostatically regulated immune cells

Table 2 G.O. pathway analysis of white blood cell genes changed byselenium

Pathway Count Percent Probability

FAS signaling pathway 7 2.1 0.00022

Huntington disease 12 3.5 0.00205

T cell activation 10 2.9 0.00274

p38 MAP kinase signaling pathway 6 1.8 0.00395

Viral myocarditis 7 2.1 0.00820

Fc gamma R-mediated phagocytosis 8 2.4 0.00919

The 339 genes called “present” in at least four cases and whose expres-sion differed significantly between groups (p <0.05, t test of within-subject changes) were analyzed with the DAVID gene ontology softwaresystem [34]. The pathways represented at p <0.01 are shown

Table 3 Leukocyte genes most upregulated by selenium supplementation

Gene symbol Gene title Low-Se placebo yeasta High-Se yeasta Se effect (p)

Baseline 48 weeks Baseline 48 weeks

CFL1 Cofilin 1 10,908±438 10,528±495 7,948±1,165 10,742±572 0.021

RPL23A Ribosomal protein L23a 13,003±842 12,754±887 8,729±1,177 9,979±472 0.042

OAZ1 Ornithine decarboxylase antizyme 1 5,642±90 5,649±91 4,185±544 5,203±171 0.015

PRF1 Perforin 1 2,410±311 2,184±162 2,902±407 3,767±241 0.043

RAC2 Ras-related C3 botulinum toxin substrate 2 5,036±193 5,079±185 3,978±301 4,738±155 0.038

ACTR3 ARP3 actin-related protein 3 homolog (yeast) 3,707±146 3,655±157 2,742±384 3,489±124 0.022

TUBA1A Tubulin, alpha 1a 2,275±161 2,342±170 1,508±167 2,186±66 0.017

TUBA1B Tubulin, alpha 1b 3,336±126 3,307±131 2,702±392 3,349±145 0.041

PECAM1 Platelet/endothelial cell adhesion molecule 2,370±235 2,435±226 1,363±223 1,989±243 0.027

ZFP36L2 Zinc finger protein 36, C3H type-like 2 2,613±125 2,546±100 2,260±199 2,864±77 0.049

FCER1G Fc fragment of IgE, receptor for; gamma polypeptide 2,085±150 2,071±153 1,482±216 1,922±84 0.038

MAX MYC-associated factor X 1,884±158 1,838±141 1,385±197 1,696±100 0.026

GZMH Granzyme H 554±171 464±119 714±112 1,019±170 0.025

a Values represent group means±SEM

204 Hawkes et al.

in the circulation and rapidly proliferating epithelial cells inculture.

Previous publications from this study reported that Seprevented DTH anergy and suppressed T cell and NK cellactivation, while differential blood cell counts, leukocyte phe-notypes, serum immunoglobulins, and complement factorswere unaffected [33]. Similar trends were observed in thissmaller cohort: Se supplementation decreased NK cell countsand suppressed activation markers in T cells and NK cells.Two cell-killing mechanisms used by both T cells and NKcells are (a) exocytosis of secretory granules that release lyticagents to directly attack and breach the target cell’s membraneand (b) engagement of the Fas/L receptor to induce caspase-mediated programmed cell death in the target cells. Both ofthese cytotoxic mechanisms were prominent features in thegene expression profile. FAS signaling was the most highlyranked biological pathway in the gene expression profile andthe most upregulated genes were associated with cytotoxicity.Perforin and granzyme are lytic agents packaged into secreto-ry granules whose concerted action is essential for effective

cell killing in vivo [38]. Exocytosis of the lytic granulesinvolves dynamic rearrangements of the actin cytoskeletonthat require RAC2 [39]. Other cytoskeletal proteinsupregulated by Se included two tubulin subunits and theactin-interacting proteins ACTR3 and cofilin. FCER1G is acomponent of the low-affinity IgE receptor required to initiatethe cytotoxic response in NK cells [40] and for phagocytosisin monocytes, whereas PECAM1 is involved in tethering ofthe dying cells to phagocytes. Thus, 9 of the 13 mostupregulated genes are involved in T cell and NK cellcytotoxicity.

The most significant gene clusters in the G.O. analysiswere also associated with functions vital to T cell and NK cellcytotoxicity. Protein phosphorylation is an essential processinvolved in all parts of the immune system [41], such asregulation of cytokine secretion by p38 MAP kinase [42],macrophage activation by receptor tyrosine kinases [43], andregulation of NK cytotoxic activity by protein kinase A [44].RNA processing is altered during granzyme-induced apopto-sis [45] and RNA processing enzymes like Dicer make micro-RNAs that regulate development and function of T cells andNK cells [46], whichmay be related to the theme of alternativesplicing identified in the G.O. functional annotation analysis.Similarly, the high ranking of protein transport in the G.O.analysis may be due to the central role of protein transport inlytic vesicle maturation and exocytosis.

Selenoproteins regulate reactive oxygen species (ROS),which are intermediates in many signaling pathways, frominsulin [47] to growth factors [48]. For example, glutathioneperoxidase 1 and thioredoxin reductase are antioxidants thatdampen ROS signaling. On the other hand, sperm glutathioneperoxidase 4 uses hydrogen peroxide to form disulfide bondsthat stabilize protamines during DNA condensation [49]. Ex-ogenously added ROS activate immune cells and activatedimmune cells produce ROS to kill invading pathogenic mi-crobes. Neutrophils, eosinophils, macrophages, and

Table 4 Leukocyte genes most downregulated by selenium supplementation

Gene symbol Gene title Low-Se placebo yeasta High-Se yeasta Se effect (p)

Baseline 48 weeks Baseline 48 weeks

IGKV1-5 Immunoglobulin kappa variable 1-5 (214836_x_at) 335±63 334±63 641±135 265±42 0.026

IGHM Immunoglobulin heavy constant mu 715±87 703±91 943±254 625±104 0.032

TRAC T cell receptor alpha constant (209671_x_at) 937±122 903±116 1,188±168 937±129 0.037

TRAC T cell receptor alpha constant (210972_x_at) 1,007±92 986±91 1,270±167 1,084±105 0.030

IGKV1-5 Immunoglobulin kappa variable 1-5 (214768_x_at) 102±14 101±14 230±54 86±9 0.032

HLA-DQB1 Major histocompatibility complex, class II, DQ beta 1 505±95 500±94 664±115 554±86 0.025

NBEAL2 Neurobeachin-like 2 444±37 465±42 453±36 386±36 0.036

MPRIP Myosin phosphatase Rho interacting protein 303±31 322±28 343±47 295±16 0.020

PSMA1 Proteasome (prosome, macropain) subunit, alpha type, 1 499±13 500±13 547±83 510±18 0.039

a Values represent group means±SEM

Table 5 Flow cytometry analysis of lymphocyte subpopulations

Cell population Placeboyeasta

High-Seyeasta

p value

Lymphocytes (CD45+) 2,172±189 1,647±243 0.110

B lymphocytes (CD45+CD19+) 278±44 249±44 0.655

T lymphocytes (CD45+CD3+) 1,440±162 1,208±180 0.355

Helper/inducer T cells (CD4+) 783±88 745±118 0.802

Suppressor/cytotoxic T cells (CD8+) 598±92 434±68 0.173

NK cells (CD16+CD56+) 379±53 169±23 0.003

IL2R+ T cells 6.2±0.5 3.6±0.4 0.002

IL2R+ NK cells 2.9±0.5 0.6±0.1 0.0006

a Values represent mean number of cells per microliter whole blood±SEM

Dietary Selenium Supplementation and Whole Blood Gene Expression 205

neighboring tissue cells produce pro-inflammatory ROS in-termediates as part of the signaling cascades activated byreleased cytokines. Protein phosphorylation is a prominenttarget of ROS [50]. Several molecules that regulate cellgrowth, differentiation, and apoptosis by affecting proteinphosphorylation are regulated by ROS, including protein ki-nase A, protein kinase C, extracellular-signal-regulated ki-nase, T cell p59fyn kinase, protein tyrosine phosphatase-1B,protein kinase A, and mitogen-activated protein kinase kinasekinase 1 [51]. Thus, selenoprotein regulation of ROS mayunderlie the dominant protein phosphorylation theme in theSe-responsive genes. Protein phosphorylation was identifiedas a major target of Se in cultured mouse and rat cells [52] andselenoproteins modulate redox regulation of protein kinases[53]. Se can also affect gene expression in leukocytes by othermechanisms. For example, thioredoxin reductase activateszinc finger transcription factors by reducing the zinc-bindingcysteine residues in the regions that bind DNA [54] and theselenodeiodinases regulate thyroid hormone, which controlsexpression of a wide variety of genes [55]. Thus, there areseveral independent and overlapping mechanisms by whichselenoproteins can potentially modulate the numbers, pheno-types, and activation states of circulating leukocytes.

SEPW1 was upregulated by Se in a British supplementa-tion trial [35], in sheep [56], and in cultured human cells [37],suggesting that it is responsible for at least some of the effectsof Se supplementation. We recently found that SEPW1 mod-ulates MAP kinase signaling, providing a potential mecha-nism linking dietary Se and protein phosphorylation [36].Although we detected no changes in selenoprotein mRNA inleukocytes, we did not measure selenoproteins per se andcannot rule out changes in their concentrations or activities.Total bodySewas approximately doubled during the 48weeksof supplementation [30], which could have affected leukocytedevelopment or maturation indirectly through effects ontissue-resident lymphoid cells without affecting selenoproteinexpression in circulating leukocytes. The total Se intake of thesubjects taking the high-Se yeast was approximately 435 μg/day, which exceeded the UL. However, no sign of Se toxicity(garlic breath, abnormal hair, or nail growth) was observed orreported by the subjects. The highest individual blood Seobserved in this study was 673 μg/L, considerably below thethreshold of chronic toxicity of 1,000 μg/L that was consid-ered by the DRI Committee [23].

Overall, the major results of this study center around T celland NK cell cytotoxicity. The numbers of activated Tcells andNK cells were suppressed by Se supplementation (Table 5)and T cell activation was a highly ranked pathway (Table 2).This is consistent with our prior study in which high-Se foodsaltered T cell-mediated vaccine responses [57]. A microarraystudy in British subjects found that Se supplementation affect-ed T cell genes [58] and a microarray study of Se-supplemented mice found increases in several T cell-specific

genes [59]. Similar to the high ranking of the FAS apoptosissignaling pathway in the present study, apoptosis genes werealso affected by Se supplements in individuals with premalig-nant skin lesions [60]. Our results are also consistent with astudy in Se-replete US subjects in which supplementationwith 200 μg/day Se as sodium selenite for 8 weeks increasedthe in vitro tumor cell killing activity of natural killer cells andcytotoxic T lymphocytes [61, 62]. It might at first seemparadoxical that Se supplements increased expression of Tand NK cell cytotoxicity genes while at the same time de-creasing the activation of T and NK cells. However, this isconsistent with the fact that dietary Se is not only required tomount an effective immune response but also plays an impor-tant role in regulating and limiting immune reactions. Muchmore work will be needed to identify the roles of individualselenoproteins in immune cells and how they contribute tomaintenance of a healthy immune system.

Acknowledgments U.S. Department of Agriculture CRIS Project Nos.5306-51530-009-00D and 5306-51530-018-00D supported this research.The authors gratefully acknowledge the excellent technical assistance ofthe Human Studies Unit of WHNRC and the clinical staff of the UCDavis Cowell Student Health Center for their assistance with the conductof this study. The UC Davis Cancer Center Gene Expression Resourcesupported byNCI Cancer Center Support Grant P30 CA93373 performedthe microarray labeling, hybridizations, and scanning. Mention of tradename, proprietary product, or specific equipment does not constitute aguarantee or warranty by the U.S. Department of Agriculture or theUniversity of California, nor does it imply approval to the exclusion ofother products that may be suitable. The opinions expressed hereinrepresent those of the authors and do not necessarily represent those ofthe U.S. Department of Agriculture or the University of California.USDA is an equal opportunity provider and employer.

References

1. Combs GF Jr, Combs SB (1986) The role of selenium in nutrition.Academic, San Diego

2. Gladyshev VN, Stadtman TC, Hatfield DL, Jeang KT (1999) Levelsof major selenoproteins in T cells decrease during HIV infection andlow molecular mass selenium compounds increase. Proc Natl AcadSci U S A 96:835–839

3. Liu Q, Lauridsen E, Clausen J (1998) Different selenium-containingproteins in the extracellular and intracellular media of leucocytescultivated in vitro. Biol Trace Elem Res 61:237–252

4. Kakiuchi C, IshiwataM, Nanko S, Ozaki N, Iwata N et al (2007) Up-regulation of ADM and SEPX1 in the lymphoblastoid cells of pa-tients in monozygotic twins discordant for schizophrenia. Am J MedGenet B Neuropsychiatr Genet 14:14

5. Arthur JR, McKenzie RC, Beckett GJ (2003) Selenium in the im-mune system. J Nutr 133:1457S–1459S

6. Marsh JA, Combs GF Jr, Whitacre ME, Dietert RR (1986) Effect ofselenium and vitamin E dietary deficiencies on chick lymphoid organdevelopment. Proc Soc Exp Biol Med 182:425–436

7. Ilback NG, Fohlman J, FrimanG (1989) Protective effect of seleniumon the development of coxsackievirus B3-induced inflammatorylesions in the murine myocardium. J Trace ElemExpMed 2:257–266

206 Hawkes et al.

8. Boyne R, Arthur JR,Wilson AB (1986) An in vivo and in vitro studyof selenium deficiency and infection in rats. J Comp Pathol 96:379–386

9. Boyne R, Arthur JR (1986) The response of selenium-deficient miceto Candida albicans infection. J Nutr 116:816–822

10. Reffett JK, Spears JW, Brown TT Jr (1988) Effect of dietary seleniumand vitamin E on the primary and secondary immune response inlambs challenged with parainfluenza3 virus. J Anim Sci 66:1520–1528

11. Turner RJ, Wheatley LE, Beck NF (1985) Stimulatory effects ofselenium on mitogen responses in lambs. Vet ImmunolImmunopathol 8:119–124

12. Aziz ES, Klesius PH, Frandsen JC (1984) Effects of selenium onpolymorphonuclear leukocyte function in goats. Am J Vet Res45:1715–1718

13. Broome CS, McArdle F, Kyle JA, Andrews F, Lowe NM et al (2004)An increase in selenium intake improves immune function andpoliovirus handling in adults with marginal selenium status. Am JClin Nutr 80:154–162

14. Peretz A, Neve J, Desmedt J, Duchateau J, Dramaix M et al (1991)Lymphocyte response is enhanced by supplementation of elderlysubjects with selenium-enriched yeast. Am J Clin Nutr 53:1323–1328

15. Baum MK, Shor-Posner G, Lai S, Zhang G, Lai H et al (1997) Highrisk of HIV-related mortality is associated with selenium deficiency. JAcquir Immune Defic Syndr Hum Retrovirol 15:370–374

16. Campa A, Shor-Posner G, Indacochea F, Zhang GY, Lai H et al(1999) Mortality risk in selenium-deficient HIV-positive children. JAcquir Immune Defic Syndr Hum Retrovirol 20:508–513

17. Kupka R, Msamanga GI, Spiegelman D, Morris S, Mugusi F et al(2004) Selenium status is associated with accelerated HIV diseaseprogression among HIV-1-infected pregnant women in Tanzania. JNutr 134:2556–2560

18. Huang Z, Rose AH, Hoffmann PR (2012) The role of selenium ininflammation and immunity: from molecular mechanisms to thera-peutic opportunities. Antioxid Redox Signal 16:705–743

19. Nair MPN, Schwartz SA (1990) Immunoregulation of natural andl ymp h o k i n e - a c t i v a t e d k i l l e r c e l l s b y s e l e n i um .Immunopharmacology 19:177–184

20. Peng X, Cui H, Cui Y, Deng J, Zuo Z et al (2011) Lesions of thymusand decreased percentages of the peripheral blood T-cell subsets inchickens fed on diets excess in selenium. HumExp Toxicol 30:1972–1978

21. Hardy G, Hardy I, Manzanares W (2012) Selenium supplementationin the critically ill. Nutr Clin Pract 27:21–33

22. Jin C, Henao-Mejia J, Flavell RA (2013) Innate immune receptors:key regulators of metabolic disease progression. Cell Metab 17:873–882

23. Panel on Dietary Antioxidants and Related Compounds, Sub-committee on Upper Reference Levels of Nutrients, Subcom-mittee on Interpretation and Uses of DRIs, Standing Commit-tee on the Scientific Evaluation of Dietary Reference Intakes,Food and Nutrition Board (2000) Dietary reference intakes forvitamin C, vitamin E, selenium, and carotenoids. NationalAcademy Press, Washington, DC, 529 pp

24. Laclaustra M, Navas-Acien A, Stranges S, Ordovas JM, Guallar E(2009) Serum selenium concentrations and diabetes in U.S. adults:National Health and Nutrition Examination Survey (NHANES)2003–2004. Environ Heal Perspect 117:1409–1413

25. Stranges S, Marshall JR, Natarajan R, Donahue RP, Trevisan M et al(2007) Effects of long-term selenium supplementation on the inci-dence of type 2 diabetes: a randomized trial. Ann Intern Med147:217–223

26. Park K, Rimm EB, Siscovick DS, Spiegelman D, Manson JE et al(2012) Toenail selenium and incidence of type 2 diabetes in U.S. menand women. Diabetes Care 35:1544–1551

27. McClung JP, Roneker CA, Mu W, Lisk DJ, Langlais P et al (2004)Development of insulin resistance and obesity in miceoverexpressing cellular glutathione peroxidase. Proc Natl Acad SciU S A 101:8852–8857

28. Loh K, Deng H, Fukushima A, Cai X, Boivin B et al (2009) Reactiveoxygen species enhance insulin sensitivity. Cell Metab 10:260–272

29. Hawkes WC, Keim NL, Richter BD, Gustafson MB, Gale B et al(2008) High-selenium yeast supplementation in free-living NorthAmerican men: no effect on thyroid hormone metabolism or bodycomposition. J Trace Elem Med Biol 22:131–142

30. Hawkes WC, Richter BD, Alkan Z, Souza EC, Derricote M et al(2008) Response of selenium status indicators to supplementation ofhealthy North American men with high-selenium yeast. Biol TraceElem Res 122:107–121

31. Affymetrix (2004) Affymetrix GeneChip expression analysis techni-cal manual. Affymetrix, Santa Clara

32. Liu WM, Mei R, Di X, Ryder TB, Hubbell E et al (2002) Analysis ofhigh density expression microarrays with signed-rank call algo-rithms. Bioinformatics 18:1593–1599

33. Hawkes WC, Hwang A, Alkan Z (2009) The effect of seleniumsupplementation on DTH skin responses in healthy North Americanmen. J Trace Elem Med Biol 23:272–280

34. Dennis G Jr, Sherman BT, Hosack DA, Yang J, Gao W et al (2003)DAVID: Database for Annotation, Visualization, and Integrated Dis-covery. Genome Biol 4:P3

35. Goldson AJ, Fairweather-Tait SJ, Armah CN, Bao Y, Broadley MRet al (2011) Effects of selenium supplementation on selenoproteingene expression and response to influenza vaccine challenge: arandomised controlled trial. PLoS One 6:e14771

36. Hawkes WC, Alkan Z (2012) Delayed cell cycle progression inselenoprotein W depleted cells is regulated by a mitogen-activatedprotein kinase kinase 4 (MKK4)-p38/c-Jun NH2-terminal kinase(JNK)-p53 pathway. J Biol Chem 287:27371–27379

37. Hawkes WC, Wang TTY, Alkan Z, Richter BD, Dawson K (2009)Selenoprotein W modulates control of cell cycle entry. Biol TraceElem Res 131:229–244

38. Pardo J, Aguilo JI, Anel A, Martin P, Joeckel L et al (2009) Thebiology of cytotoxic cell granule exocytosis pathway: granzymeshave evolved to induce cell death and inflammation. Microbes Infect11:452–459

39. Eitzen G, Lo AN, Mitchell T, Kim JD, Chao DV et al (2011)Proteomic analysis of secretagogue-stimulated neutrophils implicatesa role for actin and actin-interacting proteins in Rac2-mediated gran-ule exocytosis. Proc Natl Acad Sci U S A 9:70

40. Kikuchi-Maki A, Catina TL, Campbell KS (2005) Cutting edge:KIR2DL4 transduces signals into human NK cells through associa-tion with the Fc receptor gamma protein. J Immunol 174:3859–3863

41. Dong C, Davis RJ, Flavell RA (2002) MAP kinases in the immuneresponse. Annu Rev Immunol 20:55–72

42. Schieven GL (2009) The p38alpha kinase plays a central role ininflammation. Curr Top Med Chem 9:1038–1048

43. Tietzel I, Mosser DM (2002) The modulation of macrophage activa-tion by tyrosine phosphorylation. Front Biosci 7:d1494–d1502

44. Raskovalova T, Lokshin A, Huang X, Jackson EK, Gorelik E (2006)Adenosine-mediated inhibition of cytotoxic activity and cytokineproduction by IL-2/NKp46-activated NK cells: involvement of pro-tein kinase A isozyme I (PKA I). Immunol Res 36:91–99

45. Rajani DK, Walch M, Martinvalet D, Thomas MP, Lieberman J(2012) Alterations in RNA processing during immune-mediatedprogrammed cell death. Proc Natl Acad Sci U S A 109:8688–8693

46. Zhou L, Park JJ, ZhengQ, Dong Z,Mi Q (2011)MicroRNAs are keyregulators controlling iNKT and regulatory T-cell development andfunction. Cell Mol Immunol 8:380–387

47. Bashan N, Kovsan J, Kachko I, Ovadia H, Rudich A (2009) Positiveand negative regulation of insulin signaling by reactive oxygen andnitrogen species. Physiol Rev 89:27–71

Dietary Selenium Supplementation and Whole Blood Gene Expression 207

48. Mesquita FS, Dyer SN, Heinrich DA, Bulun SE, Marsh EE et al(2010) Reactive oxygen species mediate mitogenic growth factorsignaling pathways in human leiomyoma smooth muscle cells. BiolReprod 82:341–351

49. Pfeifer H, Conrad M, Roethlein D, Kyriakopoulos A, Brielmeier Met al (2001) Identification of a specific sperm nuclei selenoenzymenecessary for protamine thiol cross-linking during sperm maturation.Faseb J 15:1236–1238

50. Corcoran A, Cotter TG (2013) Redox regulation of protein kinases.Febs J 280:1944–1965

51. Dalle-Donne I, Rossi R, Giustarini D, Colombo R, Milzani A (2007)S-Glutathionylation in protein redox regulation. Free Radic BiolMed43:883–898

52. Stapleton SR, Garlock GL, Foellmi-Adams L, Kletzien RF (1997)Selenium: potent stimulator of tyrosyl phosphorylation and activatorof MAP kinase. Biochim Biophys Acta 1355:259–269

53. Lenart A, Pawlowski K (2013) Intersection of selenoproteins andkinase signalling. Biochim Biophys Acta 1834:1279–1284

54. Sun QA,WuYL, Zappacosta F, Jeang KT, Lee BJ et al (1999) Redoxregulation of cell signaling by selenocysteine in mammalianthioredoxin reductases. J Biol Chem 274:24522–24530

55. Larsen PR (1997) Update on the human iodothyronineselenodeiodinases, the enzymes regulating the activation andinactivation of thyroid hormone. Biochem Soc Trans 25:588–592

56. Yeh JY, Ou BR, Forsberg NE, Whanger PD (1997) Effects of sele-nium and serum on selenoprotein W in cultured L8 muscle cells.Biometals 10:11–22

57. Hawkes WC, Kelley DS, Taylor PC (2001) The effects of dietaryselenium on the immune system in healthymen. Biol Trace ElemRes81:189–213

58. Pagmantidis V, Meplan C, van Schothorst EM, Keijer J, Hesketh JE(2008) Supplementation of healthy volunteers with nutritionally rel-evant amounts of selenium increases the expression of lymphocyteprotein biosynthesis genes. Am J Clin Nutr 87:181–189

59. Hooven LA, Butler J, Ream LW, Whanger PD (2006) Microarrayanalysis of selenium-depleted and selenium-supplementedmice. BiolTrace Elem Res 109:173–179

60. KibriyaMG, Jasmine F, ArgosM, VerretWJ, Rakibuz-ZamanM et al(2007) Changes in gene expression profiles in response to seleniumsupplementation among individuals with arsenic-induced pre-malignant skin lesions. Toxicol Lett 169:162–176

61. RoyM,Kiremidjianschumacher L,Wishe HI, CohenMW, Stotzky G(1994) Supplementation with selenium and human immune cellfunctions. 1. Effect on lymphocyte proliferation and interleukin 2receptor expression. Biol Trace Elem Res 41:103–114

62. Kiremidjian-Schumacher L, Roy M, Wishe HI, Cohen MW, StotzkyG (1994) Supplementation with selenium and human immune cellfunctions. 2. Effect on cytotoxic lymphocytes and natural killer cells.Biol Trace Elem Res 41:115–127

208 Hawkes et al.