Four different breeds and crosses of various origin were chosen with characteristically different body fat content and body composition: purebred Blonde MAN (n= 12), Mangalica × Duroc (MD) (n= 12), HLW (n= 10), and Pietrain × Duroc (PD) (n= 4). Backfat (subcutaneous fat from the fourth rib) and muscle (m. levator scapulae) samples were collected from gilts at a local abattoir and then put into RNase-free freezer vials within 45 min after slaughter. Filled vials were immediately submerged in liquid N2 to avoid extended RNase exposure and degradation. Samples were transported and stored in liquid N2 containers pending processing. Traits (BFT: backfat thickness; LD: loin diameter; LW: live weight; ADG: average daily gain during fattening; and age) of the experimental animals were also recorded (Table 1). When experimenting with MAN and modern commercial breeds, an ever-emerging question is when to compare the animals – at identical weight or age – because it is practically not feasible to acquire both at the same time, even under identical housing and feeding conditions. In this study, animals were selected to be of similar live weight rather than of similar age; however, age can also be an important factor affecting gene expression and protein levels (Christoffersen et al., 2010; Ren et al., 2005). Further studies with modified experimental design are needed to assess age-dependent changes in MAN and other breeds.

Total RNA was isolated using TRIzol Reagent (Thermo Fisher Scientific, USA) and 1-bromo-3-chloropropane (VWR International, USA). Three isolations were carried out for every experimental animal and for each tissue. Samples (150–200 mg) were processed with TissueLyser LT (Qiagen, Germany) for each isolation. Concentration of RNA was determined by means of a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific). Integrity of RNA diluted in DEPC-treated nuclease-free water was verified by agarose gel electrophoresis and ethidium bromide staining (Fig. S1 in the Supplement). The RNA yields greatly varied between tissues; backfat RNA yields were typically below 200 ng µL-1, whereas muscle RNA yields were approximately 1.5 µg µL-1. To avoid potential DNA contamination, isolated RNA was treated with RQ1 RNase-free DNase (Promega, USA) following the manufacturer's instructions. According to the manufacturer's recommendations, 1 µg of each total RNA sample was reverse-transcribed using an iScript cDNA synthesis kit (Bio-Rad, USA) containing a blend of oligo (dT) and random hexamer primers. Gene expression was quantified by qPCR using SsoFast EvaGreen Supermix (Bio-Rad) and the 2-ΔΔCt method. Reactions were performed in triplicates on a CFX96 real-time PCR detection system (Bio-Rad) using clear plates. In the experiments, β-actin (ACTB) was used as a reference gene (Luo et al., 2009). The sequences of the primers applied, product lengths, accession numbers, relevant annealing temperatures, and efficiency of PCRs are shown in Table 2. For each gene, the efficiency was determined by 10-fold serial dilutions (“standards”) of the PCR products. Every run contained no template controls (NTCs) for the analyzed and the reference genes as well. The NTCs were accepted as negative with threshold cycles over 35, as quantitation cycles were always prior to this cycle. Thermal profile was as follows: one initial denaturation cycle at 95 ∘C for 5 min, followed by 40 two-step cycles of 95 ∘C for 30 s and gene-specific annealing temperature (Table 2) for 30 s. After the last cycle, melting curve analysis was performed (from 65 to 95 ∘C, with 0.5 ∘C increments) to verify specificity of the amplified products.

Expression data were analyzed by one-way ANOVA (Tukey's) test in SPSS v.16 for Windows (SPSS Inc.) and means were considered significantly different at P < 0.05. Pearson's correlation between gene expression and production traits was also determined. Throughout the study, animal handling and sampling were conducted in accordance with the standards recommended by Directive 2010/63/EU.

In every breed and cross, adipocyte fatty-acid-binding protein (A-FABP) (or FABP4) was predominantly active in backfat and less abundant in muscle (P < 0.05). Similar distribution between muscle and adipose tissue has formerly been described in MD crossbred pigs (Tempfli et al., 2015). The four analyzed groups appeared to be clearly separated based on the A-FABP transcript level: higher (P < 0.05) expression was detected in adipose as well as in muscle tissues of MAN and MD than in those of HLW and PD (Table 3). A-FABP plays a central role in intracellular desorption and transport of fatty acids from the plasma membrane to the cytoplasm (the site of triacylglycerol and phospholipid synthesis) and to the mitochondria (the site of beta oxidation) and is considered a candidate gene for fat deposition in swine (Chmurzynska, 2006). Consistent with the present results, Zhao et al. (2009) detected higher A-FABP transcript levels in muscle of the local Chinese Wujin, a fatty pig breed compared to the leaner Landrace, indicating that leaner breeds can transport fewer fatty acids through intracellular trafficking, resulting in less extended IMF. Expression of A-FABP was positively correlated with IMF content in purebred and Large White crossbred Chinese Laiwu Black pigs, where A-FABP transcript levels increased rapidly with the body weight until 60–70 kg and then remained at high levels in both breeds analyzed (Chen et al., 2013). In the present study, no measurements were taken to determine IMF content; however, the MAN is well known for marbled pork production and is generally characterized by largely elevated IMF compared to other breeds (Holló et al., 2009; Koncz et al., 2014). Furthermore, BFT is positively correlated with IMF content in pigs (Jacyno et al., 2015), and greatest BFT was measured in the MAN group (Table 1).

A-FABP was found to be upregulated in a selected high-fat line of Duroc pigs when compared to the low-fat Duroc group (Canovas et al., 2010). Differential A-FABP expression was also detected between Berkshire and Yorkshire pigs, where highest levels were measured in Berkshire, a breed well known for marbled pork production and IMF deposition (Cho et al., 2011). No such correlations were identified in Large White × Landrace animals (Gerbens et al., 2001); however, A-FABP content in muscle was remarkably greater in pigs with increased IMF than in those with low IMF content; furthermore, positive correlation was identified between A-FABP level and adipocyte number and lipid content (Damon et al., 2006). Similarly, a moderately strong positive correlation was observed between BFT and A-FABP levels in the present study (Table 4), whereas LD and ADG were negatively correlated to A-FABP expression.

In cases where a lack of correlation was detected between A-FABP mRNA and protein levels, an occurrence of post-transcriptional protein–protein interaction mechanisms was hypothesized (Damon et al., 2006). In the present study, A-FABP mRNA levels of purebred and Duroc crossbred MAN were highest (P < 0.05) in both backfat and muscle; however, protein abundance was not investigated. Further experiments are needed to analyze the relationship between gene expression and actual protein levels in MAN and MD.

Of all six genes analyzed, adiponectin (ADIPOQ) was most abundant in backfat, while its expression in muscle was found to be the lowest (Table 3); this is not surprising since ADIPOQ is secreted almost exclusively by adipocytes. ADIPOQ was previously shown to have elevated expression in MD fat compared to muscle (Tempfli et al., 2015). No significant differences were detected between groups in ADIPOQ backfat level; however, muscle expression was remarkably higher (P < 0.05) in MAN and MD than in the other groups (Table 3). Regarding the distinctive muscle ADIPOQ expression, the role of intramuscular fat cells is to be emphasized. According to Ding et al. (2004) ADIPOQ can be hardly detectable in muscle samples of breeds that do not deposit excessive IMF, as it can be an exclusive source of ADIPOQ transcripts in muscle. Daniele et al. (2008) detected significant differences in subcutaneous adipose tissue ADIPOQ expression of Large White and fat-type Casertana pigs, with higher levels in lean-type Large White. This is consistent with human studies, where ADIPOQ hormone levels have been inversely related to body fat content (Matsubara et al., 2002). Lord et al. (2005) observed similar patterns in adipose tissue ADIPOQ expression of Upton-Meishan (excessive fat content) and Large White pigs. Interestingly, in the present study this difference was not observed between backfat ADIPOQ mRNA levels of MAN and MD and leaner HLW or PD pigs. However, muscle ADIPOQ expression was in positive correlation with BFT, and in negative correlation with LD and ADG; no significant correlation was detected between backfat ADIPOQ expression and BFT, LD, or ADG (Table 4).

A potential explanation for this is that fatter pigs express higher levels of ADIPOQ in visceral fat than in subcutaneous fat (Lord et al., 2005). By means of muscle transcriptome profiling, Canovas et al. (2010) detected elevated ADIPOQ expression in high-fat Duroc compared to low-fat Duroc animals, which is consistent with the findings reported in this study. As Canovas et al. (2010) also indicated, these results are not in agreement with the observations in humans that fatty acid oxidation is reduced in obese individuals. Paradoxically, ADIPOQ seems to be more expressed in muscle tissue of obese pigs without decreasing lipid accumulation via beta oxidation, which can be partly explained by the lower expression of ADIPOQ receptors (Lord et al., 2005; De Rosa et al., 2013).

The fatty acid synthase gene (FASN) was most active in adipose tissue in all groups. Its expression in backfat was highest in MAN, moderate in MD, and lowest in HLW and PD (Table 3). Muscle expression of MAN and MD did not differ significantly; however, both were higher (P < 0.05) compared to HLW and PD. Correlation analysis showed similar patterns to expression differences, as fat and muscle FASN levels were correlated positively to BFT and negatively to LD and ADG (Table 4).

FASN encodes for a crucial enzyme complex of lipogenesis that catalyzes the synthesis of palmitate (a long-chain saturated fatty acid) from acetyl-CoA and malonyl-CoA. Based on the differences in FASN expression, it was concluded that lipid synthesis is more active in MAN and MD than in HLW or PD. Higher muscle lipid synthesis can largely contribute to IMF deposition. Similar patterns of FASN expression were observed when Wujin pigs were compared to the leaner Landrace (Zhao et al., 2009), and when Italian Duroc was compared to Italian Large White (Braglia et al., 2014), underlining the greater potential of some breeds for accumulation of fat in muscle.

Differential fat mass and obesity-associated gene (FTO) expression was observed in backfat of MAN and MD as well as HLW and PD (Table 3). Muscle FTO levels did not differ significantly (P > 0.05). The level of FTO in fat was significantly correlated to BFT but not to LD or ADG, whereas muscle levels were in negative correlation with ADG but not with BFT or LD (Table 4).

FTO is known to be involved in the regulation of body weight and adiposity, and was found differentially expressed between Göttingen minipigs and lean production pigs (Cirera et al., 2014); however, expression was not as largely differentiated between subcutaneous backfat and muscle tissues compared to present breeds and crosses. Similar to the results presented here, more definite differences were observed in the expression of FTO in muscle and backfat tissues of the Taihu-based Chinese pig breed Suzhong (Fu et al., 2013). Higher FTO expression in MAN and MD adipose tissue confirmed that elevated FTO level contributes to the development of the fattier phenotype. Targeted sequencing of the MAN FTO gene is needed to identify the causes and the genetic background (e.g., potential specific alleles modifying transcription factor binding sites) of elevated expression.

Regarding backfat tissue, highest (P < 0.05) leptin (LEP) expression was detected in MAN, followed by MD and PD, whereas HLW was characterized by the lowest expression levels (Table 3). In muscle, MAN and MD produced more (P < 0.05) LEP compared to HLW and PD. Correlation analysis revealed positive correlation between BFT and LEP levels in fat and muscle and negative correlation with LD and ADG (Table 4).

Georgescu et al. (2014) observed similarly increased LEP expression in MAN compared to purebred Duroc and other commercial lean production pigs. LEP (a hormone mainly secreted by adipocytes) plays pivotal roles in controlling feed intake and energy homeostasis of pigs through hypothalamic areas associated with the regulation of appetite (Barb et al., 2001). A continuous increase in LEP levels can lead to LEP resistance and consequently to impaired appetite suppression and unmitigated food intake, which is a potential mechanism responsible for developing the obese phenotype in MAN. Furthermore, elevated LEP levels can lead to disturbances in luteinizing hormone (LH) secretion, ovulatory failure, and deterioration in reproductive functions (Brüssow et al., 2008; Mitchell et al., 2005). High LEP expression and subsequent resistance are potential factors behind the poor reproductive performance of MAN, with average litter sizes of 5–7 piglets (Tempfli et al., 2011).

Of the analyzed genes, only the PLIN2 (perilipin 2 or adipocyte differentiation-related protein) gene was detected with higher normalized expression in muscle compared to backfat. Muscle PLIN2 levels were in positive correlation with BFT but in negative correlation with LD and ADG (Table 4). There was no significant (P > 0.05) difference between PLIN2 expressions in backfat of any of the analyzed groups. Nevertheless, muscle levels were different (P < 0.05) for MAN and MD as well as for HLW and PD (Table 3). Elevated PLIN2 expression in muscle was therefore concluded to associate with IMF accumulation. Similar patterns of PLIN2 activity were presented by Davoli et al. (2011); however, in their study a slightly higher PLIN2 level was observed in backfat compared to that in muscle, which can be attributed to the different selection of experimental muscle tissues.