Effects of Selenium and Cadmium on Breast Muscle Fatty-Acid Composition and Gene Expression of Liver Antioxidant Proteins in Broilers

The present work was part of a project intended to evaluate whether organic selenium (Se) has the potential to protect against toxic effects exerted by cadmium (Cd). For this reason, 300 as-hatched, one-day-old broiler chickens were randomly allocated in four dietary treatments with five replicate pens per treatment. Chickens in T1 treatment, were offered a diet supplemented with 0.3 ppm Se (as Se-yeast), without added Cd; in T2 treatment, they were offered a diet with 0.3 ppm Se and 10 ppm Cd; in T3 treatment, they were offered a diet with 0.3 ppm Se and 100 ppm Cd; in T4 treatment, chickens were offered a diet supplemented with 3 ppm Se and 100 ppm Cd. Cadmium was added to the diets in T2, T3, and T4 as CdCl2. On the fourth and sixth weeks, liver and breast samples were obtained from two broilers per replicate pen. Relative gene expression levels of catalase (CAT), superoxide dismutase 1 (SOD1) and 2 (SOD2), methionine sulfoxide reductase A (MSRA) and B3 (MSRB3), iodothyronine deiodinase 1 (DIO1), 2 (DIO2), and 3 (DIO3), glutathione peroxidase 1 (GPX1) and 4 (GPX4), thioredoxin reductase 1 (TXNRD1) and 3 (TXNRD3), and metallothionein 3 (MT3) were analyzed by real-time quantitative PCR in liver, whereas the fatty-acid (FA) profile of breast muscle was determined by gas chromatography. Broilers supplemented with 0.3 ppm Se could tolerate low levels of Cd present in the diets, as there were no significant changes in the breast muscle FA profile, whereas excess Cd led to decreased polyunsaturated fatty acids (PUFAs), and in particular n-6 PUFA. Furthermore, treatments mainly affected the messenger RNA (mRNA) expression of SOD2, TXNRD3, and MT3, while age affected CAT, MSRB3, DIO2, DIO3, GPX4, TXNRD1, and MT3. In conclusion, dietary Se may help against the negative effects of Cd, but cannot be effective when Cd is present at excessive amounts in the diet.


Introduction
Cadmium (Cd) is a non-essential heavy metal and one of the most toxic environmental pollutants. Its occurrence in feed constitutes a considerable issue in agriculture and animal production. In some circumstances, Cd levels surpass the maximum permitted limits. There are mainly two routes of animal exposure to Cd including the application of feed mineral premixes with high-Cd residues, and the use of animal manure as an organic fertilizer that contains high levels of Cd [1]. Cadmium is known for its comprehensive toxicity to mammals, and numerous experiments demonstrated that Cd causes various

Animals, Diets, and Experimental Procedures
A total of 300, one-day-old, as-hatched Cobb broiler chickens from a commercial hatchery were used. Five replicate pens of four dietary treatments were employed, namely T1, T2, T3, and T4, randomly allocated in the house (the pen was the experimental unit). Each replicate was assigned to a clean concrete floor pen (2 m 2 ), and chickens were raised on a litter of wheat straw shavings. There were 15 broilers per pen and 75 per treatment. In treatment T1, birds were fed a diet supplemented with 0.3 ppm Se (Se-yeast), without added Cd; in treatment T2, broilers were fed a diet with 0.3 ppm Se and 10 ppm Cd; in treatment T3, they were fed a diet with 0.3 ppm Se and 100 ppm Cd; in treatment T4, broilers were fed a diet with 3 ppm Se and 100 ppm Cd. The Cd was added to diets T2, T3, and T4 in the form of CdCl 2 (Sigma-Aldrich, St Louis, MO, USA). Supplemented Se was obtained from a yeast source, Sel-Plex ® (Alltech Inc., Nicholasville, KY, USA).
The trial protocol was approved by the Institutional Committee for Animal Use and Ethics of the Faculty of Animal Science and Aquaculture of the Agricultural University of Athens (Ethical protocol code: . During the whole trial, chickens were handled in compliance with national and European Union (EU) regulations and laws and in accordance with the guidelines and principles for the care of animals in experimentation [23].
The duration of the experiment was 42 days and broilers were raised, according to Cobb's management manual, in a house where ventilation and light were controlled. The lighting program was a 23-h/1-h of light/dark cycle, while heat was provided using a heating lamp in each pen. The birds were fed a starter diet up to the 10th day of their life, then a grower diet up to the 20th day, and a finisher diet up to the 42nd day (Table 1). Water and food were provided ad libitum. At the end of the 28th and 42nd days, two chickens per pen were killed in order to obtain tissue samples from breast muscle and liver, which were rapidly frozen in liquid nitrogen and kept at −80 • C until analyzed.

Determination of Fatty-Acid Profile
The feeds were prepared by grinding through a 1-mm screen, whereas breast samples were carefully freed from any visible adipose and connective tissues, and then blended in short bursts in a domestic food processor. Extraction and methylation of FA in feed and breast muscle samples were performed according to the method of O'Fallon et al. [24]. Then, 1-g (± 0.05) duplicate samples were hydrolyzed for 1.5 h at 55 • C in 6 mL of a mixture containing 5.3 mL of methanol and 0.7 mL of 10 N KOH. A known amount (ca. 0.5 mg) of internal standard (C13:0) was added to the samples prior to hydrolysis. Subsequently, 0.58 mL of 24 N H 2 SO 4 was added to the samples, which were then incubated at 55 • C for 1.5 h to prepare the fatty-acid methylesters (FAME). Hexane (3 mL) was added to the reaction tube, which was vortex-mixed and centrifuged at 1100× g. The supernatant hexane layer containing the FA methyl esters was evaporated to dryness under nitrogen stream, rediluted in 0.5 mL of clear hexane, and kept at −20 • C until analyzed by gas chromatography.
A temperature-programmed run was followed on a Perkin Elmer Autosystem XL gas chromatograph equipped with a 30 m × 0.25 mm × 0.25 µm internal diameter HP-Innowax capillary column (Agilent Technologies, J&W GC columns, Santa Clara, CA, USA) and a flame ionization detector (FID; Perkin Elmer, Waltham, MA, USA), as described elsewhere [22]. Commercial standards (FAME 37 Component; Sigma-Aldrich Co. Supelco, St. Louis, IL, USA) were used to identify the FAs, which were quantified using the amount of internal standard added prior to hydrolysis. The sum of areas for all FA peaks compared to that for the internal standard was used to calculate the total weights of FAs (in mg/100 g) in diets and breast muscle. Individual FAs were expressed as percentage by weight of total FA.

RNA Isolation and Reverse Transcription
Total RNA was extracted from approximately 50 mg of tissue using TriReagent (Sigma-Aldrich Co., St. Louis, IL, USA) according to the manufacturer's instructions. Extracted RNA was treated with DNAseI (New England Biolabs, Beverly, MA, USA) according to the manufacturer's instructions. The quality and quantity of the RNA extracted were confirmed by spectrophotometry, as well as gel electrophoresis. Total RNA (500 ng) was reverse transcribed with the PrimeScript First-Strand complementary DNA (cDNA) Synthesis Kit (Takara Bio Inc, Kusatsu, Shiga 525-0058, Japan), according to the manufacturer's instructions, using a mix of random hexamers and oligo-dT primers.

Statistical Analysis
Data for FA were analyzed using the SPSS statistical package (version 17.0, IBM Corp., Armonk, NY, USA.). A generalized linear model (GLM) for repeated measures with treatment and age as fixed effects and their interactions was carried out. Post hoc tests were performed for diet when appropriate, using Tukey's multiple range test, and significance was set at 0.05. All data for FAs are presented as least-squares (LS) means.
Statistical analysis of gene expression was performed using Prism version 6.0 (GraphPad Software, Inc., La Jolla, CA, USA). Normal distribution of data values was assessed using the D'Agostino-Pearson omnibus test. For variables that did not exhibit a normal distribution, common logarithmic (MSRB3, DIO3, GPX1, GPX4, TXNRD1) or cube-root (MSRA, CAT, SOD2, DIO2, MT3) transformation was used in order to meet ANOVA assumptions. Data were analyzed using two-way analysis of variance (ANOVA). "Age" and "treatment" were the two independent variables, while further analysis of the interaction between the two independent variables was also carried out. Tukey's honest significant difference (HSD) post hoc tests were performed to analyze the effect of "treatment" within "age". Data are presented as means ± standard error for variables that did not require transformation to meet ANOVA assumptions. Data for transformed variables are presented as back-transformed means with back-transformed upper and lower confidence limits (confidence interval 95%). The level of statistical significance was set at p < 0.05.

Breast Muscle Fatty-Acid Composition
Significant differences in the breast muscle FA composition between treatments were observed in the present study. Total weights of FA were lower (p < 0.05) in T3 and T4 when compared to T1 and T2 broiler chickens. Regarding the FA profile of breast meat, total n-6 FAs were lower (p < 0.05) in T3 and T4 broilers compared to T1 and T2 broiler chickens, mainly as a result of the decrease in 18:2n-6 ( Table 3). Total n-3 FAs were not affected by the dietary treatment; however, changes in individual FA occurred. Most notably, 18:3n-3 decreased (p < 0.05) and 20:5n-3 increased (p < 0.05) in breast muscle of broilers from T3 and T4, respectively, compared to T1 and T2. Dietary treatments did not affect total saturated fatty acids (SFAs). In contrast, total PUFAs were strongly affected by the dietary treatments; T3 and T4 had significantly lower (p < 0.05) PUFAs and, as a result, a lower PUFA/SFA ratio (p < 0.05), in comparison with T1 and T2 broiler chickens (Table 3). Sampling time (age of broilers) had significant effects on monounsaturated fatty acids (MUFAs), PUFAs, and n-3 FAs. Total MUFAs (in particular cis9-18:1) were lower (p < 0.001), whereas PUFAs and n-3 FAs were greater (p < 0.005) at 42 days compared to 28 days of age (Table 3). No significant differences in the breast muscle FA composition between T1 and T2, or between T3 and T4 were found.

Gene Expression of Liver Antioxidant Enzymes
Overall, age of broilers (28 vs. 42 days) was a major factor significantly affecting gene expression of CAT (p < 0.001), MSRB3 (p < 0.05), DIO2 (p < 0.05), DIO3 (p < 0.0001), GPX4 (p < 0.05), TXNRD1 (p < 0.0001), and MT3 (p < 0.0001) (Figure 1). Similarly, treatment effects were found to be statistically significant for gene expression of SOD2 (p < 0.05), TXNRD3 (p < 0.0001), and MT3 (p < 0.0001). Furthermore, post hoc multiple comparisons within the same age group revealed that the effect was pronounced on the 42-day-old broilers. In detail, there was a downregulation of SOD2 expression in the T2 treatment when compared to the control T1 group, with no statistically significant differences for the other two treatments (T3 and T4). Regarding TXNRD3, T4 treatment samples exhibited lower levels of expression when compared to those of the control T1 group. As for MT3 expression, it was highly upregulated (5-10-fold) in T3 and T4 treatments when compared to treatments T1 and T2. Additionally, DIO3 expression in the liver increased incrementally (T1 < T2 < T3 < T4), although those differences were found to be statistically significant only between T4 and T1 treatments. Finally, another important finding was the observed interaction of the two main factors in the expression of CAT and TXNRD1. More specifically, CAT expression was upregulated in groups T2 and T3 in 42-day-old broilers when compared to 28-day-old broilers, but not in groups T1 and T4. The exact same effect was observed for TXNRD1. Regarding the expression of SOD1, MSRA, DIO1, and GPX1, no statistically significant effect was observed for any of the main factors or their interaction (Figure 1).    , and TXNRD3 are presented as means ± standard error of the mean (SEM). Data for all other genes are presented as back-transformed means with back-transformed upper and lower confidence limits. Significances of main factor effects "age" (P A ) and "treatment" (P T ), as well as their interaction (P Int ), are reported for each variable. Columns with different superscripts within the same age differ significantly (p < 0.05).

Discussion
The study was designed taking into account that (i) the EU limit of Cd in feed is 5 ppm [29], (ii) in some cases, Cd levels in feed exceed maximum permitted limits [30], (iii) increased essential element supplementation may alleviate the negative effects of toxic metal contamination [15,31,32], and (iv) the maximum permitted inclusion level of Se is 0.3 ppm (in the United States) [33]. The present study investigated the interactions between Cd, Se at low or high concentrations (well below toxic levels) [34], and gene expression of several enzymes that affect the antioxidant/prooxidant balance, finding that gene expression is altered in the presence of prooxidants and that organic Se can protect the body during stress conditions induced by Cd.
It is known that selenomethionine, a large constituent of Se-yeast, is non-specifically incorporated into proteins replacing methionine [35], and stored tissue Se amounts can be mobilized and employed in circumstances of oxidative stress. Cadmium is absorbed from the gastrointestinal tract and lungs and is mainly accumulated in the liver and kidney, where it is bound to metallothionein (MT) [36,37]. Metallothioneins are among the most well-known antioxidants that protect against metal toxicity [38,39]. They are low-molecular-weight proteins (∼7 kDa) that bind to and are induced by free cytosolic metal ions, especially Cd, Zn, Cu, and Hg, and are implicated in the defense against metal toxicity [40]. These properties indicate that MT is an important factor affecting metal toxicity in vertebrates as it was previously shown in invertebrates [41]. Indeed, MT3 expression was highly increased after exposure to 100 ppm Cd, and this increase persisted even after high Se addition. Moreover, MT3 expression was significantly increased with age (28 vs. 42 days of experimental time), i.e., when Cd was accumulated in the liver [19]. Selenium was shown to lower the accumulation of toxic elements by shifting the distribution of tissue elements from MTs toward high-molecular-mass proteins [42,43]. Generally, it seems that Cd exposure markedly increases the expression of MTs in various species, with some differences depending on the Cd concentration and the duration of exposure [44]. Our findings collectively support the hypothesis that Cd increases the synthesis of MT, which plays a role in heavy-metal detoxification by reducing oxidative stress [45].
Thioredoxin reductase (TXNRD) enzymes are oxidoreductases that use NADPH to catalyze the reduction of oxidized thioredoxin (Trx) [46]. There are three mammalian TXNRDs: cytoplasmic/nuclear TXNRD1 which reduces Trx1, mitochondrial TXNRD2 which reduces Trx2, and testes-specific TXNRD3. Their functions are essential part of the Trx system, redox regulation, antioxidant defense, and cell signaling [12,47]. It was shown that increased Se intake results in increased expression of both the TXNRD1 and TXNRD2 proteins. However, the levels of Se are not the only regulator of expression of TXNRDs. Their expression may be affected, under certain conditions, by oxidative stress or as a response to growth factors [47]. In our study, TXNRD3 mRNA expression was significantly decreased after Cd addition, and this effect could not be reversed by Se.
Superoxide dismutase, as well as GPX, CAT, and metal-binding proteins (transferrin, metallothionein, albumin, myoglobin, haptoglobin, hemopexin, ceruloplasmin, ferritin, and lactoferrin), belongs to the first line of defense developed by the antioxidant system of the cell in order to prevent the initial formation of free radicals. Superoxide dismutase 2 inactivates catalysts or removes precursors of free radicals [14]. Addition of 10 ppm Cd caused a downregulation of SOD2 expression in T2 treatment when compared to the control T1 group. It is possible that the absence of an effect of high Se addition between T3 and T4 on SOD2 gene expression was related with its use either for the formation of Se-Cd complexes [19,48] or for antioxidant protection against the induced oxidation by Cd [49]. Similar results were obtained by previous studies in chicken kidney [50] and liver [7], where significantly decreased SOD activities were demonstrated after 60 days of dietary co-treatment with Se and Cd.
Interestingly, the results of this study also pointed that, with experimental time and the accompanying increased levels of Cd in the liver [19], mRNA expression of CAT, DIO3, GPX4, and TXNRD1 was elevated, probably as a protective attempt of the antioxidant defense system against oxidative stress caused by Cd [14].
In relation to the above, Se contribution on metabolic pathways associated with the protection of the organism against oxidative stress was shown to induce changes in the activity of selenoproteins. Their expression is regulated, among others, by the concentration of this element [51]. However, Se concentration does not obligatorily affect the rate of transcription of selenoprotein genes. The observed differences in protein expression may be the result of changes in mRNA translation or reduced stability (increased degradation). Depending on Se levels, various effects were observed on cellular functions (energy metabolism, immunity) [52].
In this work, the protective mechanism of Se could not be sufficiently clarified by mRNA expression alone, since Cd itself could have induced the synthesis of antioxidant proteins [53]. Previously, it was shown that Cd was able to modify the activity of some enzymes, both with antioxidant and other activities, interacting directly with the enzymatic proteins [54]. In this case, the observed Se-induced expression differences at the mRNA level of the proteins in question could have no effect on the toxic action of Cd. The results from these studies, however, are not easily comparable with the results of the present study, since they refer to different examined proteins, different animal species (rats vs. chickens), different routes of administration (intraperitoneal (i.p.) injection vs. dietary addition), and different levels of administration, among others.
Administration of selected trace elements, such as Se, as a remedy to ameliorate Cd toxicity, was lately evaluated in several studies [55][56][57]. Selenium exhibited protective effects against Cd toxicity by counteracting the immunosuppressive, as well as hepatic and renal oxidative, damage [58]. Previous works showed that supranutritional levels of supplemented Se were able to prevent the peroxidation of PUFAs in broiler chicken meat, thereby increasing their intramuscular levels [22]. It is known that Cd present in feeds results in oxidative deterioration of lipids [4] and, therefore, it is not unlikely that it can reduce the PUFAs in meat. In the present study, meat PUFA composition decreased in broiler chickens fed with a 100 mg Cd/kg diet; however, it cannot be attributed to Cd-induced oxidation alone. Broiler chickens fed with 100 mg Cd/kg diet had significantly lower feed intake, as found in our previous work [1], which resulted in a lower energy, as well as PUFA intake and deposition in the breast muscle. The lower energy intake was clearly demonstrated by the lower body weight [1] and the reduction in breast muscle fatness, as indicated by the total weights of FA herein, which may explain the lower PUFA composition in broiler chickens fed with 100 mg Cd/kg. However, muscle fatness plays an important role in defining the FA profile [59], and the differences in the breast muscle FA composition between treatments were evaluated by removing the effects of fatness using the total weights of FA as a covariate in the statistical analysis. Breast meat PUFAs, and in particular n-6 PUFAs, remained lower in broiler chickens fed 100 ppm Cd, when compared to those fed the diet unsupplemented with Cd (T1) or that supplemented with 10 ppm Cd. These results indicate that an oxidative stress imposed by dietary Cd addition at very high levels cannot be dismissed. Hence, it appears that the reduction in PUFA composition may be caused by both the lower feed and PUFA intake and the increased oxidation processes induced by the presence of Cd at very high dietary levels. Furthermore, a protective role of dietary Se on FA composition was observed. The addition of 10 ppm Cd to the diet (T2 diet containing 0.3 ppm Se) did not alter the FA profile of breast meat in comparison with the diet unsupplemented with Cd (T1). This indicated that the dietary Se addition at the recommended levels (0.3 ppm) in T2 was adequate to alleviate any potential negative effects of low amounts of Cd. However, dietary Se addition either at the recommended (0.3 ppm) or at supranutritional levels (3.0 ppm) could not ameliorate the impact of high Cd levels. These results are in agreement with our previous findings, which showed that broilers fed supplemental Se can tolerate low levels of Cd present in the diet; however, they cannot tolerate high levels of Cd in the diet regardless of the level of dietary Se. Most notably, it was shown that dietary 10 ppm Cd addition (T2) did not result in any significant change in chicken body mass compared with that of chickens fed diets without added Cd (T1), while addition of 100 ppm Cd significantly reduced (p < 0.001) chicken body mass compared with that of chickens fed no added Cd [1]. The body mass of chickens fed a diet with 100 ppm Cd and 3 ppm Se (T4) did not differ compared with that of chickens fed 100 ppm Cd and 0.3 ppm Se, and it was lower compared with that of chickens fed no or low levels of added Cd.

Conclusions
It seems that a direct implication of the oxidative stress status of the liver on the regulation of MT3, SOD2, and TXNRD3 gene expression, along with FA profile changes, in breast muscle exists; however, the exact mechanisms of these interactions need to be further investigated. Based on previous reports from our group and on the present findings on the effect of Se supply on the oxidative stress balance in chickens exposed to Cd, our data would suggest that dietary Se can ameliorate the negative effects of Cd, but cannot be efficient in the presence of excessive dietary amounts of Cd.
Author Contributions: E.Z. designed and coordinated the project, carried out part of the transcriptional analyses, and was a major contributor to drafting the manuscript; G.P. carried out FA and statistical analyses; A.C.P. designed the project, collected the data, and contributed to writing the manuscript; G.T. assisted in the design of the project, and supervised and carried out most of the transcriptional analyses; K.F. designed and coordinated the project and reviewed the manuscript. All authors read and approved the final manuscript.