Expression of Selenoprotein Genes Is Affected by Obesity of Pigs Fed a High-Fat Diet^,,

Abstract

Background: Relations of the 25 mammalian selenoprotein genes with obesity and the associated inflammation remain unclear.

Objective: This study explored impacts of high-fat diet-induced obesity on inflammation and expressions of selenoprotein and obesity-related genes in 10 tissues of pigs.

Methods: Plasma and 10 tissues were collected from pigs (n = 10) fed a corn-soy–based control diet or that diet containing 3–7% lard from weanling to finishing (180 d). Plasma concentrations (n = 8) of cytokines and thyroid hormones and tissue mRNA abundance (n = 4) of 25 selenoprotein genes and 16 obesity-related genes were compared between the pigs fed the control and high-fat diets. Stepwise regression was applied to analyze correlations among all these measures, including the previously reported body physical and plasma biochemical variables.

Results: The high-fat diet elevated (P < 0.05) plasma concentrations of tumor necrosis factor α, interleukin-6, leptin, and leptin receptor by 29–42% and affected (P < 0.05–0.1) tissue mRNA levels of the selenoprotein and obesity-related genes in 3 patterns. Specifically, the high-fat diet up-regulated 12 selenoprotein genes in 6 tissues, down-regulated 13 selenoprotein genes in 7 tissues, and exerted no effect on 5 genes in any tissue. Body weights and plasma triglyceride concentrations of pigs showed the strongest regressions to tissue mRNA abundances of selenoprotein and obesity-related genes. Among the selenoprotein genes, selenoprotein V and I were ranked as the strongest independent variables for the regression of phenotypic and plasma measures. Meanwhile, agouti signaling protein, adiponectin, and resistin genes represented the strongest independent variables of the obesity-related genes for the regression of tissue selenoprotein mRNA.

Conclusions: The high-fat diet induced inflammation in pigs and affected their gene expression of selenoproteins associated with thioredoxin and oxidoreductase systems, local tissue thyroid hormone activity, endoplasmic reticulum protein degradation, and phosphorylation of lipids. This porcine model may be used to study interactive mechanisms between excess fat intake and selenoprotein function.

Introduction

Obesity is a major public health problem worldwide (1, 2). Etiologically, a high percentage of calorie intake derived from dietary fat represents an important contributor to this metabolic disorder (3, 4). Because pigs are an appropriate model for human nutrition and medicine (5, 6), we have produced obesity in pigs by feeding them a high-fat diet from weanling to finishing (7). Compared with the controls, pigs fed the high-fat diet had heavier body weight (146 vs. 121 kg), thicker back fat (3.5 vs. 2.3 cm), and greater abdominal fat mass (3.5 vs. 1.9 kg) and eye muscle area (59.1 vs 50.5 cm²). These obese pigs exhibited hyperlipidemia, hyperglycemia, and hyperinsulinemia. Although obesity may alter hormones other than insulin and cause chronic inflammation via macrophage infiltration in tissues, especially adipose tissue (8), our initial study (7) did not measure responses of major energy metabolism-related hormones such as thiiodothronine (T₃)⁹ and thyroxine (T₄) and cytokines, including TNFα, IL-6, leptin, and leptin receptor (LEPR) (9, 10). Previously, our group revealed impacts and mechanisms of dietary selenium and expression of selenium-dependent glutathione peroxidase-1 (GPX1) on glucose and fat metabolism and developments of insulin resistance, diabetes, and obesity in several species (11–15). However, reciprocal metabolic impacts of high-fat diet-induced obesity on inflammation and expression of selenoprotein genes in various tissues of pigs or other species were not studied or recognized.

Among the 25 selenoprotein genes identified in humans (16) and pigs (14, 17), GPX1, selenoprotein P plasma 1 (SEPP1), and selenoprotein S (SELS) were shown to be involved in diabetes, obesity, or impaired energy metabolism (18). Global overexpression of Gpx1 induces type 2 diabetes-like phenotypes in mice (11, 19). Elevated erythrocyte and hepatic GPX1 activities are associated with insulin resistance in pregnant women (20), gestating rats (13), and finishing pigs (14). From a rat model (Psammomys obesus) of insulin resistance, obesity, type 2 diabetes, and dyslipidemia, a Tanis gene (VCP-interacting membrane protein gene) was identified to be homologous to human SELS (16, 21, 22). Further study found this protein to be a glucose regulatory protein (23) and involved in the production of inflammatory cytokines (24, 25). Expression of Sels can be activated by endoplasmic reticulum stress, an important factor in inflammatory diseases, including obesity (26). Three studies showed elevated plasma selenoprotein P concentrations in patients with type 2 diabetes mellitus compared with healthy humans (27–29). Circulating selenoprotein P concentrations were increased in subjects with nonalcoholic fatty liver disease and in subjects with visceral obesity (30). Despite all these intriguing findings and strong effects of dietary selenium and (or) vitamin E (14, 17, 31), regulation of the selenogenome expression by dietary fat was not explored in humans or higher animals.

The high-fat diet-induced obesity in our pig model (7) was presumably associated with expression of obesity-related genes such as agouti related protein (AGRP) (32) and LEPR (33). The subsequent question was, therefore, to find which of these obesity-related genes in which tissues were affected most by the dietary treatment and how changes in these genes were related to the expression profiles of selenoprotein genes and the metabolic phenotypic indicators. Therefore, this study was conducted to determine 1) impacts of high-fat diet feeding on the expression of mRNA of the 25 selenoprotein and 16 selected obesity-related genes in 10 tissues of pigs, 2) interactions between these 2 groups of genes and their potential relations with the obesity phenotypic measures, and 3) responses of major energy metabolism-related hormones and cytokines to high-fat diet feeding and their relations with the expression of selenoprotein and obesity-related genes.

Methods

Animal, diet, and sample collection.

The animal experiment protocol was approved by Sichuan Agricultural University, Chengdu, China. The husbandry, diets, and physical characteristics of the experimental pigs were reported previously (7). To summarize, 20 Duroc × Landrace × Yorkshire crossbred, castrated boars (body weight: 19.7 ± 0.26 kg) were divided into 2 groups (n = 10) and fed a corn-soy control diet (0.3 mg Se/kg; total fat: <0.82%) or that diet with added lard as follows: 3% (20–50 kg), 5% (50–80 kg), and 7% (>80 kg) for 180 d. All nutrient concentrations in the 2 diets (Supplemental Table 1) were matched. The pigs were reared in individual pens and had free access to feed and water. At the end of the study, 8 pigs from each group were killed by exsanguination after stunning to collect blood and tissue samples. Plasma samples were prepared as described by Liu et al. (14) and stored at −80°C. After liver, kidney, skeletal muscle, thyroid, pituitary, hypothalamus, heart, subcutaneous fat, perirenal fat, and pancreas were rapidly removed, they were immediately dissected on an ice-cold surface, perfused with ice-cold isotonic saline, minced with surgical scissors, divided into aliquots, snap-frozen, and stored in liquid nitrogen until use.

Plasma biochemical measures.

Plasma T₃ and T₄ concentrations were determined with an automated clinical chemistry analyzer (Model 800; Roche). Plasma TNFα and IL-6 concentrations were determined with respective porcine ELISA kits (R&D Systems), and plasma leptin and LEPR concentrations were determined with respective human ELISA kits (the human leptin and LEPR proteins as the standards; R&D Systems), according to the manufacturer’s instructions.

Real-time quantitative PCR analyses of mRNA abundance.

RNA extraction and quality control, real-time quantitative PCR procedure, and relative mRNA abundance quantification were described elsewhere (14, 17, 31) with the exception that total RNA of subcutaneous and perirenal fat tissues was extracted with RNeasy Lipid Tissue Mini Kit (Qiagen). Melting curve and cycle threshold analyses were performed to confirm the specificity of amplification. The primers for the 25 selenoprotein genes, 16 obesity-related genes (selected from a pilot microarray analysis), and 2 reference genes (β-actin and GAPDH) were designed with Primer Express 3.0 (Applied Biosystems) and are presented in Supplemental Table 2.

Statistical analysis.

Student’s t test (SPSS for Windows 13.0; SPSS Inc.) was used to examine effects of the high-fat diet on plasma biochemical measures and mRNA levels of each gene within a given tissue. Data are presented as means ± SEs, and significance level was set at P ≤ 0.05 unless indicated otherwise. In a few cases of unequal variances for the variables, the t test was conducted with that option. Stepwise regression analyses were conducted with the program of PROC REG (SAS 8.2; SAS Institute), and the data sets were derived from the same pigs collected from the present and the previous (7) (metabolic phenotypes and plasma biochemical measures) studies. When individual selenoprotein gene mRNA levels within a given tissue were regarded as the independent variables, there were 3 types of dependent variables as follows: metabolic phenotypes, plasma biochemical measures, and obesity-related gene mRNA levels in the same tissue (as the selenoprotein genes). The reciprocal regression analyses were also done with the obesity-related and selenoprotein gene mRNA levels in the same tissue as independent and dependent variables, respectively. The significance was also set at P ≤ 0.05 unless indicated otherwise.

Results

Plasma biochemical measures.

Pigs fed the high-fat diet had plasma concentrations of TNFα, IL-6, leptin, and LEPR that were 29.2%, 32.7%, 41.6%, and 34.9% greater (P < 0.05) than the control diet, respectively (Table 1). However, plasma T₃ and T₄ concentrations were similar between the 2 groups of pigs.

TABLE 1.

Effects of the high-fat diet on plasma concentrations of hormones and cytokines in pigs compared with pigs fed the control diet from weanling to finishing for 180 d¹

Measures	Control diet	High-fat diet	P
T3, μg/L	0.45 ± 0.06	0.45 ± 0.05	0.99
T4, μg/L	5.23 ± 0.27	5.37 ± 0.36	0.76
TNFα, pg/mL	44.5 ± 2.3	57.5 ± 3.5	0.01
Leptin, pg/mL	805 ± 71.3	1140 ± 54.9	0.01
IL-6, pg/mL	501 ± 22.7	665 ± 22.6	0.00
LEPR, pg/mL	41.3 ± 2.8	55.8 ± 2.5	0.01

¹Values are means ± SEs, n = 8. LEPR, leptin receptor; T₃, triiodothronine; T₄, thyroxine.

Abundance of selenoprotein mRNA.

Of the 25 selenoprotein genes assayed, 2 genes (thioredoxin reductases 2 and 3) provided results that were too close to background to be interpreted or reported. Overall, expression of the remaining 23 genes in the 10 tissues examined responded to high-fat diet feeding in 3 ways. First, mRNA levels of 12 selenoprotein genes in 6 tissues were up-regulated (P < 0.05–0.1) by the high-fat diet (Figure 1). Specifically, thyroid had the highest number (n = 5) of selenoprotein genes enhanced with the largest magnitude of change (1.8- to 13-fold), whereas muscle had 4 genes increased (70–90%) over the controls. The high-fat diet elevated both SELM and GPX4 mRNA levels in pituitary but only selenoprotein 15 kDa gene (SEP15) in liver, SELO in kidney, and GPX3 in hypothalamus. Only SELV and GPX4 mRNA levels were enhanced by the high-fat diet in >1 tissue (2 tissues). Second, mRNA levels of 13 selenoprotein genes in 7 tissues were down-regulated (P < 0.05–0.1) by the high-fat diet (Figure 2). Pancreas had the highest number (n = 7) of the genes decreased, followed by perirenal fat (n = 5) and liver (n = 4). There was an ∼50% decrease in 2 or 3 genes each due to high-fat diet feeding in subcutaneous fat (SELI, SELM, and SELX), hypothalamus [SELH, SELI, and thioredoxin reductase 1 (TXNRD1)], and pituitary (SELI and TXNRD1). Only SELM was down-regulated in the kidney. Among the 13 down-regulated genes, 1 gene (SELI) was decreased in 4 tissues, 3 genes (SELH, SELM, and TXNRD1) in 3 tissues, and 3 genes [SELK, selenoprotein W 1 (SEPW1), and GPX3] in 2 tissues. Third, the high-fat diet produced no statistically significant or any effect on mRNA levels of 5 genes (GPX1, GPX2, selenoprotein N gene, selenoprotein T gene, and selenophosphate synthetase 2 gene) in any of the 10 assayed tissues, of 4 genes (SEP15, SELO, SEPP1, and SELS) in 9 tissues, or 7 genes [SELK; SEPW1; SELX; iodothyronine deiodinases 1, 2, and 3 (DIO1, DIO2, and DIO3); GPX3; and GPX4] in 8 tissues (Supplemental Table 3).

FIGURE 1.

Effect of the high-fat diet on relative mRNA levels of selenoproteins in thyroid (A), liver (B), kidney (C), muscle (D), pituitary (E), and HT (F) of pigs compared with pigs fed the control diet from weanling to finishing for 180 d. Data are presented as means ± SEs (n = 4). Asterisks indicate different from control: *P < 0.05. DIO1, iodothyronine deiodinase 1 gene; DIO2, iodothyronine deiodinase 2 gene; GPX3, glutathione peroxidase 3 gene; GPX4, glutathione peroxidase 4 gene; GPX6, glutathione peroxidase 6 gene; HT, hypothalamus; SELI, selenoprotein I gene; SELM, selenoprotein M gene; SELO, selenoprotein O gene; SELS, selenoprotein S gene; SELV, selenoprotein V gene; SEP15, selenoprotein 15 kDa gene; TXNRD1, thioredoxin reductase 1 gene.

FIGURE 2.

Effect of the high-fat diet on relative mRNA levels of selenoproteins in pancreas (A), HT (B), subcutaneous fat (C), perirenal fat (D), liver (E), pituitary (F), and kidney (G) of pigs compared with pigs fed the control diet from weanling to finishing for 180 d. Data are presented as means ± SEs (n = 4). Asterisks indicate different from control: *P < 0.05, **P < 0.01. DIO1, iodothyronine deiodinase 1 gene; DIO3, iodothyronine deiodinase 3 gene; GPX3, glutathione peroxidase 3 gene; GPX6, glutathione peroxidase 6 gene; HT, hypothalamus; SELH, selenoprotein H gene; SELI, selenoprotein I gene; SELK, selenoprotein K gene; SELM, selenoprotein M gene; SELV, selenoprotein V gene; SELX, selenoprotein X gene; SEPP1, selenoprotein P plasma 1 gene; SEPW1, selenoprotein W 1 gene; TXNRD1, thioredoxin reductase 1 gene.

mRNA levels of obesity-related genes.

Among the 16 obesity-related genes assayed, 5 genes in 3 tissues were up-regulated (P < 0.05) (Figure 3). These included agouti signaling protein gene (ASIP), AGRP, and resistin gene (RETN) in skeletal muscle; uncoupling protein 3 gene (UCP3) in thyroid; and uncoupling protein 2 gene in pituitary. In contrast, 11 genes were down-regulated (P < 0.05–0.01) in 6 tissues (Figure 4). These changes included AGRP in liver, kidney, and perirenal fat; LEPR in liver, kidney, and subcutaneous fat; adiponectin receptor 2 gene in kidney, perirenal fat, and subcutaneous fat; and FA binding protein 4, adipocyte, gene in perirenal fat, pituitary, and hypothalamus. In addition, adiponectin gene (ADIPOQ) was down-regulated in the 2 fat tissues, whereas the remaining genes were affected only in single tissues. The high-fat diet had no substantial effect on mRNA levels of other obesity-related genes in different tissues (Supplemental Table 4).

FIGURE 3.

Effect of the high-fat diet on relative mRNA levels of obesity-related genes in muscle (A), thyroid (B), and pituitary (C) of pigs compared with pigs fed the control diet from weanling to finishing for 180 d. Data are presented as means ± SEs (n = 4). Asterisks indicate different from control: *P < 0.05, **P < 0.01. AGRP, agouti related protein gene; ASIP, agouti signaling protein gene; RETN, resistin gene; UCP2, uncoupling protein 2 gene; UCP3, uncoupling protein 3 gene.

FIGURE 4.

Effect of the high-fat diet on relative mRNA levels of obesity-related genes in liver (A), kidney (B), HT (C), perirenal fat (D), subcutaneous fat (E), and pituitary (F) of pigs compared with pigs fed the control diet from weanling to finishing for 180 d. Data are presented as means ± SEs (n = 4). Asterisks indicate different from control: *P < 0.05, **P < 0.01. ADIPOQ, adiponectin gene; ADIPOR2, adiponectin receptor 2 gene; AGRP, agouti related protein gene; ASIP, agouti signaling protein gene; FABP1, FA binding protein 1, liver, gene; FABP3, FA binding protein 3, muscle and heart, gene; FABP4, FA binding protein 4, adipocyte, gene; FASN, FA synthase gene; HT, hypothalamus; LEPR, leptin receptor gene; RETN, resistin gene; UCP2, uncoupling protein 2 gene.

Stepwise regression analysis.

There were 62 regressions (P < 0.05) between phenotypic and plasma measures and tissue selenoprotein mRNA abundance (Supplemental Table 5). The strongest regressions (R² ≥ 0.90, P < 0.01) are depicted in Figure 5. Specifically, body weight of pigs was affected by 5 selenoprotein genes (SEP15, GPX3, SELV, SELH, and SELI) in liver and hypothalamus, whereas plasma TG concentrations were correlated negatively with hypothalamus SELH and TXNRD1 and with thyroid SELS and SELV. The rest of the 7 measures were correlated with 1–3 selenoprotein genes in thyroid, pancreas, or liver. Reciprocally, SELV in liver and thyroid was involved in 7 of these regressions, whereas SELI in pancreas, hypothalamus, and thyroid was involved in 4 regressions. The rest of the 5 selenoprotein genes (GPX3, SEP15, TXNRD1, SELH, and SELS) were involved in 1 or 2 regressions each.

FIGURE 5.

Correlation analysis of dependences (dependent variables) of phenotypic and plasma measures (left) and obesity-related gene expression (right) on selenoprotein gene expression (independent variables, middle) by tissue of pigs. The correlation was analyzed by stepwise regression, and only relations with R² > 0.9 are shown. ADIPOQ, adiponectin gene; AGRP, agouti related protein gene; ASIP, agouti signaling protein gene; DIO1, iodothyronine deiodinase 1 gene; DIO3, iodothyronine deiodinase 3 gene; GPX3, glutathione peroxidase 3 gene; HT, hypothalamus; L, liver; LEPR, leptin receptor gene; M, muscle; P, pancreas; PF, perirenal fat; RETN, resistin gene; SELH, selenoprotein H gene; SELI, selenoprotein I gene; SELS, selenoprotein S gene; SELV, selenoprotein V gene; SEP15, selenoprotein 15 kDa gene; TH, thyroid; TXNRD1, thioredoxin reductase 1 gene; UCP3, uncoupling protein 3 gene.

There were 18 regressions (P < 0.04) for the mRNA levels of the 12 obesity-related genes (dependent variables) with those of selenoprotein genes (independent variables) in 7 tissues (Supplemental Table 6). Among these regressions, 5 selenoprotein genes (DIO1, DIO3, SELH, SELI, and SELV) had 6 strong correlations (R² ≥ 0.90, P < 0.01) with 5 obesity-related genes (ASIP, AGRP, UCP3, ADIPOQ, and RETN) in 4 tissues (liver, muscle, perirenal fat, and thyroid) (Figure 5). All these 6 regressions were positive, and 4 of them were derived from single genes and tissues. Only DIO3 was correlated with both ASIP and AGRP in liver.

There were 51 statistically significant (P < 0.05) regressions between tissue mRNA levels of obesity-related genes (independent variables) and the phenotypic and plasma biochemical measures (dependent variables) (Supplemental Table 7). Among them, 7 equations associated with body weight, back fat thickness, plasma TG, and plasma IL-6 showed strong correlations (R² = 0.90–0.98) with 9 genes in 4 tissues (Figure 6). Likewise, body weight and plasma TG concentration displayed the highest number (n = 6 or 7) of regressions among the dependent variables.

FIGURE 6.

Correlation analysis of dependences (dependent variables) of phenotypic and plasma measures (left) and selenoprotein gene expression (right) on obesity-related gene expression (independent variables, middle) by tissue of pigs. The correlation was analyzed by stepwise regression and only relations with R² > 0.9 are shown. ADIPOQ, adiponectin gene; ADIPOR2, adiponectin receptor 2 gene; AGRP, agouti related protein gene; ASIP, agouti signaling protein gene; DIO1, iodothyronine deiodinase 1 gene; DIO3, iodothyronine deiodinase 3 gene; FABP1, FA binding protein 1, liver, gene; FABP3, FA binding protein 3, muscle and heart, gene; FABP4, FA binding protein 4, adipocyte, gene; FASN, FA synthase gene; GPX6, glutathione peroxidase 6 gene; K, kidney; L, liver; LEPR, leptin receptor gene; M, muscle; PF, perirenal fat; RETN, resistin gene; SELH, selenoprotein H gene; SELI, selenoprotein I gene; SELM, selenoprotein M gene; SELV, selenoprotein V gene; SEPP1, selenoprotein P plasma 1 gene; SEP15, selenoprotein 15 kDa gene; SF, Subcutaneous fat; TH, thyroid; UCP2, uncoupling protein 2 gene; UCP3, uncoupling protein 3 gene.

There were 23 statistically significant (P < 0.05) regressions between tissue mRNA levels of the obesity-related genes (independent variables) and tissue mRNA levels of the selenoprotein genes (dependent variables) (Supplemental Table 8). The correlation coefficients for SELH and SEPP1 in perirenal fat; SELI in thyroid; SELM in subcutaneous fat; GPX6, DIO1, and SELV in muscle; and SEP15 and DIO3 in liver were strong (R² = 0.90–0.99, P < 0.01) (Figure 6). Muscle SELV and liver SEP15 each were affected by 3 obesity-related genes, whereas ASIP, ADIPOQ, and RETN each were correlated with 3 selenoprotein genes.

Discussion

The present study has clearly demonstrated that the prolonged high-fat diet feeding elevated plasma leptin and LEPR concentrations and affected mRNA abundances of 16 obesity-related genes in various tissues of pigs. These responses were consistent with the high-fat diet-induced physical and biochemical phenotypes of obesity in the pigs, including hyperglycemia, hyperlipidemia, and elevated body fat accretion (7). The overall metabolic responses of pigs to the high-fat diet were similar to those of rodents (34) and humans (35–37). Intriguingly, our study showed an increase in plasma LEPR in the obese pigs fed the high-fat diet. Although this observation is contrary to a well-accepted notion that soluble LEPR decreases with obesity (38, 39), several previous reports may help partially explain our discrepancy. First, plasma soluble LEPR was shown to be essential in mediating leptin’s weight-reducing and other biological effects, and its concentration was elevated by ∼20-fold in Zucker diabetic fatty rats (40). Second, hypocaloric diet decreased leptin and LEPR concentrations in type 2 diabetic patients (41). Third, serum LEPR concentrations and parallel values of the molar ratio between serum LEPR and leptin were higher in diabetic than in healthy children (42). Seemingly, the response of plasma soluble LEPR concentration to obesity may be more complicated than generally perceived. Furthermore, the present study has also shown elevated plasma concentrations of TNFα and IL-6 in the pigs fed the high-fat diet. Apparently, the high-fat diet resulted in obesity associated with a chronic, mild inflammation status (8). Therefore, the newly observed changes in hormones, adipocytokines, and obesity-related gene expressions in the present study, along with the previously described phenotypes (7), indicate the experimental suitability and physiologic relevance of this porcine model for studying interactions and mechanisms between obesity and selenoprotein expression and function.

One of the most important findings from the present study is the illustration of the 3 patterns of regulation by dietary fat on the selenogenome expression. First, 12 selenoprotein genes (GPX4, GPX6, DIO1, SELV, DIO2, SELI, SELS, TXNRD1, SELM, SELO, SEP15, and GPX3) were up-regulated in 6 tissues, in particular in thyroid and muscle. Although mRNA changes of these and other genes do not explain the metabolic phenotypes of the porcine obesity in a cause–effect fashion, a systematic analysis of these data may offer functional implications. The increased expression of GPX4 and GPX6 might suggest an association with membrane oxidative events and possibly control of bioactive phospholipid metabolism (43–45). Likewise, enhanced SELI and SELO gene expression, through mediating the pertaining phosphokinase activities, have the potential to control phospholipid metabolism in response to the increased fat intake and consequent inflammation (46, 47). Up-regulation of SELS (21), a component of endoplasmic reticulum protein degradation, may also be related to the inflammatory responses induced by high fat intake (48), because a promoter polymorphism of SELS, -105G-A, enhanced proinflammatory cytokine expression (26). The proteins encoded by SELV, SELM, SEP15, and TXNRD1 all have potential to bind with thioredoxin or have thioredoxin folds and could all mediate protective oxidoreductase responses to the high-fat diet (49). Furthermore, enzymes encoded by DIO1 and DIO2 are involved principally in the activation of thyroid hormones (50, 51). Because local T₃ production is an essential regulatory mechanism of metabolism within a tissue (52), the up-regulation of DIO1 and DIO2 may indicate a need for an increased local T₃ production due to high fat intake. The local elevation of T₃ would not necessarily be reflected in the total circulating concentrations of the hormone (52), which may explain the lack of difference in plasma T₃ or T₄ between the high-fat diet and control groups.

The second pattern was down-regulations of 13 selenoprotein genes in 7 tissues by the high-fat diet. Pancreas had the most genes (7 genes) altered, followed by perirenal fat of 5 and liver of 4 genes. Although high fat intake down-regulated 4 genes (SELI, SELH, SELM, and TXNRD1) in 3 or 4 tissues, the rest of the 9 genes (GPX3, GPX6, DIO1, DIO3, SELK, SEPP1, SEPW1, SELX, and SELV) showed decreases in only 1 or 2 tissues. Seemingly, these down-regulations may reflect a general adverse oxidative effect of high-fat diet-induced obesity on tissue redox control or metabolism because most of these genes encode redox-associated selenoproteins (53). This notion is supported by the number and type of selenoprotein genes affected in pancreas and its key role in producing endocrinal hormones and exocrine enzymes to cope with high fat intake (54). Changes in SELI and SELK may be connected to membrane phospholipids and endoplasmic reticulum protein degradation mechanisms that also play roles in the regulation of more general gene expression (22, 55).

The third pattern of regulation included a lack of effect by the high-fat diet on 5 genes (GPX1, GPX2, selenoprotein O, selenoprotein T, and selenophosphate synthetase 2) in any of the 10 assayed tissues, along with no effect on another 11 genes in 8 or 9 tissues. Strikingly, most of these genes, in particular GPX1, are highly responsive to dietary selenium changes in different tissues of various species (14, 18, 56, 57). Their lack of responses to the high-fat diet may imply a physiologic necessity for a constant production of their encoded selenoproteins to cope with the obesity-related metabolic stress. These selenoproteins cover a wide range of metabolic functions associated with selenium (53). The fact that many such functions were unaffected highlights that the up- and down-regulation of specific selenoprotein genes discussed in the previous two paragraphs were likely to be specific responses to the stress of a high fat intake. Clearly, our present study has revealed a completely different regulation of selenogenome by dietary fat from dietary selenium. It will be fascinating to unveil the underlying molecular mechanism for this differential regulation between these 2 types of nutrients.

Among the 17 affected selenoprotein genes, 5 genes (GPX4, SEP15, SELO, SELS, and DIO2) were only up-regulated and 5 genes (SELK, SEPW1, SEPP1, SELX, and SELH) were only down-regulated, whereas 7 genes (GPX3, GPX6, DIO1, TXNRD1, SELV, SELI, and SELM) showed both responses to the high-fat diet. As shown in Figure 1 and 2, thyroid and pancreas had the most selenoprotein genes up- and down-regulated, respectively, among all the tissue assayed. These differential regulations by dietary fat may be partially explained by tissue-specific expression and hierarchy of selenoprotein expression (58). The mechanisms yet elucidated ensure that those proteins with the most essential functions for a particular cell type are expressed at the expense of those whose loss may have a less detrimental effect (59). Hence, the main changes in selenoprotein gene expression resulted from feeding the high-fat diet may reflect crucial roles for the respective selenoproteins in the tissues affected and point to how they may interact with proinflammatory and prodiabetogenic challenges of the diet.

The high-fat diet also exerted 3 types of impacts on the 16 obesity-related genes assayed. These included up-regulations of 5 genes (ASIP, AGRP, RETN, UCP3, and uncoupling protein 2) in 3 tissues (muscle, thyroid, and pituitary), down-regulations of 11 genes in 6 tissues, and no effect on 3 genes in 9 tissues. These divergent patterns of responses are consistent with the belief that obesity is not a hereditary disease caused by monogenic mutations, but it is associated with >200 genes (60, 61). Because many prior studies have determined regulations of these genes by dietary fat (32, 62–67), our interest herein is to explore their correlations with high-fat diet-induced obesity and changes of selenogenome expression. Collectively, there were 154 regressions with tissue selenoprotein gene or obesity-related gene mRNA abundances as the independent variables. Body weights and plasma TG concentrations of pigs turned out to be the most responsive dependent variables to changes of these 2 types of independent variables induced by high fat intake. Likely, body weights and circulating TGs were highly regulated by both selenoprotein and obesity-related genes in the circumstance. Although correlations between the latter such as FA binding protein 4, adipocyte, gene and body weights were reported (68), no such link was identified for the former. Although SELV and SELI seemed to be the strongest independent variables of selenoprotein genes for the regression of phenotypic and plasma measures, ASIP, ADIPOQ, and RETN represented the strongest independent variables of obesity-related genes for the regression of tissue selenoprotein gene expression. Most of these regressions were associated with the gene expression in thyroid, pancreas, liver, and hypothalamus. All these organs are central in the control of metabolism and thus may, through regulating selected selenoprotein and obesity-related gene expression, provide key steps for the mechanisms that act to limit effects of excess fat deposition and adverse inflammatory responses.

In contrast, the relations of selenoproteins to the obesity-related genes are less clear cut than the above-mentioned links to body physical and plasma variables. However, the associations between DIO1 and DIO3 and obesity genes in liver and muscle again highlight how a large proportion of metabolism via thyroid hormones could be changed by high fat intake. It is our current interest to explore how selenoprotein genes involved in antioxidant/lipid mediator and thioredoxin and oxidoreductase systems are linked to genes for uncoupling proteins and FA binding proteins. An exciting question is whether or not high-fat diet-induced porcine obesity and inflammation are mediated by the regulation of selenoprotein genes via the function of obesity-related genes or vice versa. The answer may help explain a wide range of potential adverse effects of high-fat diets and how this predisposes one toward the eventual adverse consequences associated with the diet. Identification of the families of selenoproteins that are most crucial in the response to high-fat diets may require the use of gene knockout models. This would then point to dietary or drug-related strategies that are potentially effective in the prevention and treatment of the diseases that occur because of high fat intake. The pig model used in this study would be ideal for such research.